MEMOIRS

OF THE

QUEENSLAND MUSEUM

BRISBANE VOLUME 34
1 AUGUST, 1994 PART 3

In Memorian

NOEL CHARLES GILLESPIE
14.12.48 — 29.06.94

Noel died in Bnsbane after a 2 year battle with cancer:
Noel meant a lot of things to a lot of people. His
warmth, generosily and humour set him apart. Noel was
a scientist par excellence, a larrikan, a loyal friend, a
gifted leader, a loving family man anda real human
being, warts and all. He was someone very special. His
drive and enthusiasm for life were unique. These quali-
tics never Jeft him. Noel was a pioneer in ciguatera and
aquaculture in Australia, Without his clear vision —“his
dréam’ — for ciguatera research in Australia, this work-
shop Would not have happened, You were our hero. and
we treasure your memory in our hearts.

Clive, John, Mike and Richard

Noel in search of the origin of Ciguatera,
Flinders Reef, c, 1986.

Sponsored by the Fishenes Research & Development Corporation, Queensland Department of Primary Industries
and the Queensland Museum,

(QUEENSLAND |

museum

DEPARTMENT OF
PRIMARY INDUSTRIES
Fe

© Queensland Museum, PO Box 3300, South Brisbane, Q. 4101, Australia.
Phone (07) 840 7555. Fax (07) 846 1918
Nauonal Library of Australia card number ISSN 0079-8835

NOTE
Papers published in this volume and in all previous volumes of the Memairs of the Queensland Museum may be reproduced to
scientific research, individual study. or other educational purposes. Properly acknowledged quotations may be made but querie
regarding the republication of any papers should be addressed to the Director, Copies of the joumal can be purchased from the
Queensland Museum Bookshop.

A Queensland Government Project
Typeset at the Queensland Museum,
Printed by Merino Lithographics,

18 Baldock Street. Moorooka, Queensland 4105,

CONTENTS
PART 1 (Issued 24 December, 1993)

STANISIC, J.
The identity of Helicarion semoni Martens, 1894: a large semi-slug from the Wet Tropics, northeastern
Queensland (Pulmonata: Helicarionidac) ......... 00... eee eee ee ete eve e sd eeaeunes 1
STANISIC, §,
Danielleilona gen, nov., from the Wet Tropics, northeastem Queensland (Pulmonata: Charopidae) . ... . 1]
STANISIC, J.

Lenwebbia paluma sp. nov., from the Wet Tropics, northeastem Queensland (Pulmonata: Charopidae), . 21
STANISIC, J.
Eungarion mcdonaldi gen. et sp.nov., a montane semi-slug from mideasterm Queensland rainforests

(Pulmonata: Helicarionidac)..... 2.2.22 ee eee a hte tp ebeceeeeretnere 27
BRAILOYSKY, H,
A revision of the Tribe Colpurini from Australia (Hemiptera-Heteroptéra-Coreidae)............ bsete 35
SHORT, 5.W,
Caridina zebra, a new species of freshwater atyid shrimp (Crustacea: Decapoda) from northeastern Queens-
Jand rainforest ........ Suge Spee Peis Pep Rant the Beret hye hele ie él

SHORT, J.W.& DAVIE, P.J.F.
Two new species of freshwater crayfish (Crustacea: Decapoda: Paraslacidae) from northeastern

Queensland rainforest ,,.,.., shog tax atejaa sie sletete pleat lete dave} abinele: tr pbibicscele ttelaittetanadasteetethte | 65
RAVEN, RJ,
The biodiversity of Australian mygalomorph spiders, J. Two new species of
Namirea (Araneae:Dipluridac).. 6... 6.2 c ete en enn etree eben eny BI
RICHARDS, S.J.
Functional significance of nest construction by an Australian rainforest frog: a preliminary analysis .... 89

COUPER, P.J,, COVACEVICH, J.A. & MORITZ, C.
A review of the leaf-tailed geckos endemic to castern Australia: a new genus, four new species, and
other new data ,...... 0.0... ccc eee cere eer eri PA Vas atl als tala Pte sate Cy ied Alsi alle se jo 95
WHITTIER, J.M.
Ecological notes on Carlia rostralis in rainforest and associated habitat in the southem Wet Tropics , . . 125
CUNNINGHAM, M.
Reproductive biology of the Prickly Forest Skink, Gnyperoscincus queenslandiae, an endemic
species from northem Queensland. 22.0.0... 6.0. eee renee eee yee 13)
SADLIER, R.A., COLGAN, D.J. & SHEA, GM,
Taxonomy and distribution of the scincid lizard Saproscincus challengeri and related wingeks in

southeasiem Australia. ...... 2-22 ec ee ee eee cece yeep penueuas 139

COVACEVICH, J.A., COUPER, PJ. & JAMES, C.

A new skink, Nangura spinosa gen, ct sp,noy., from a dry rainforest of southern Queensland, -.,.,,., 159
JAMIESON, B.G.M. & SCHELTINGA, D.M.

The ultrastructure of spermatozoa of Nengura spinesa (Scincidae, Reptilia) ...........-..-..... , 169
INGRAM, G.J. & COVACEVICH, J.A.

Two new species of sinped blindsnakes .. 0... 22. eee cette eee e ee eben ee nes 18]
COVACEVICH, J.A., COUPER, PJ. & INGRAM, GJ.

New reptile records from rainforests of south and mideastem Queensland .............-..-..-.--- 185
COVACEVICH, J.A. & MCDONALD, K.R.

Distribution and conservation of frogs and reptiles of Queensland rainforests .... 5-2... 00.00.4000. 189

JOSEPH, L., MGRITZ, C. & HUGALL, A.
A mitochondrial DNA perspective on the histoncal biogeography of mideastem Queensland

rpdinfomeg® Biss 5. 5) joe 5) 62 «eee elfen oop nif ctege bla enc al anudlene vaalicle Geos tlele gle yet e adele peta plele pala ale 201

HORSUP, A., JAMES, C. & PORTER, G,

Vertebrates of dry rainforest of south and mideastern Queensland, ...... 02... 060. een eee ee 215
WERREN, GL.

Conservation strategies for rare and threatened vertebrates of Australia's Wet Tropics Region ....... - 229
NOTES
RICHARDS, 5.J., MCDONALD, K.R, & INGRAM, G.J.

Recognition of Litoria eucnemis (Lonnberg) in Australia ......... ste tictactole totett ote: rlect wlettefett oferiot 94

McDONALD, ER. & STORCH, D.R.
A new reproductive mode for an Australian hylid frog...........-.-.--------- 22-22 ee eee -.-. 200

WERREN, G.L. & TRENERRY, M.P.

Size and diet of Bufo marinus in rainforest of northeastem Queensland...,..,....)..)..)-2)2-a-. 240
WHITTIER, J.M. & MOELLER, D.R.
Varanus prasinus (the Emerald Goanna) on Moa Island, Torres Strait, Australia. ,.,-.,..,-.),. .. 130

TRENERRY, M.P. & WERREN, GL.
Possum assemblages in rainforest of the Carbine Yeates, NEQ, « with aa, reference to
Hemibelideus lemuroides . 2 eel teetertaetests-2---.-— IRB

PART 2 (Issued 1 March, 1994)

LAWRENCE, D.
Customary exchange across Torres Strait... 0 a oe ce eect epee peed ve bera ts weriate. 2b

PART 3 (Issued ) August, 1994)
BABINCHAK, J.A,, MOELLER, P.D.R., VAN DOLAH, F.M., EYO, P.B. & RAMSDELL, IS.

Production of. ‘ciguatoxins in cultured Gambierdiscus POXICUS 1 2 og peewee vee te oa eletbe od eed ee ae 447
BAGNIS, R,

Natural versus anthropogenic disturbances to coral reefs:

comparison in epidemiological patterns of ciguatera. .. 2.6.4.2... 0052020 g eee aeee ees 455

BENOIT, E. & LEGRAND, A.M.

Gambierioxin-induced modifications of the membrane potential of myelinated nerve fibres ©... -... 461
BLYTHE, D.G., FLEMING, L-E., AYYAR, D.R., DESYLVA, D,, BADEN, D. & SCHRANK, K.

Mannitol therapy for acute and chronic: ciguatera fish poisoning ...-..-.-..-..-.--.---------- . 465
DALZELL, P.

Management of ciguatera fish poisoning in the south Pacific... 5-2... 2. ey oe ee ee eee eee 471

DICKEY, R.W., GRANADE, ELR. & MCCLURE, F.D
Evaluation of a sol id-phase immunobead assay for detechon of ciguateraelated biotoxins in
Caribbean Tnbisht 2.5.5): .o)jecee ein lee eyewear hy (we tas Sa ene FS SU eS | ote eho poet 48]
HOKAMA, Y., ASAHINA, A.Y., TITUS, E., ICHINOTSUBO, D., CHUN, S.. HONG, T.L.WP,, SHIRAI, JL,
ASUNCION, DA. & MIYAHARA, J.T,
Assessment of ciguateric fish in Hawaii by lens an

mouse toxicity and guinea pig atrialassays. - E8556 es Se a ABR

HOLMES, MJ. & LEWIS, RJ,

The origin of ciguatera..-. 2.1... eee eg eens TERRY EST Las ae Ure fe ccs eee 497
HOLMES, M.J., LEWIS, R.J., SELLIN, M. & STREET, R.

The origin of ciguatera in Platypus Bay, Australia . 305
ICHINOTSUBO, D., ASAHINA, A.Y., TITUS, E., CHUN, S., ‘HONG, T.L.W.P. SHIRAI, ‘UL, |. & HOKAMA, ¥ ng

Survey for ciguatera fish poisoning in west Hawaii ..-, 0... 0.055.400 us $13
KALY, ULL. & JONES, G.P.

Test of the effect of disturbance on ciguatera in Tuvalu ....---..-.......,, Lepieytegrtietayey 225

LANG, R,J., VOGALIS, F., HOLMES, M.J, & LEWIS, R.J.
Maiiotoxin induces muscle contraction and a non-selective cationic current

in single smooth muscle cells of the guinea-pig proximal colon ..,,.,.....0....00.se.-22-- 533
LEWIS, RJ.
Immunological, biochemical and chemical features of ciguatoxins: implications
for the detection of Ciguateric Ash. 5. ii eve eae cence teed cede ee sedeeternees . S54
LEWIS, R.J.
Impact of a validated, cost effective screen for ciguateric fish ....-..-.,--. 2-0 2.0 2 eee eee 549
LEWIS, RJ. & BRERETON, 1M.
Inverse-detected NMR of ciguatoxin: quaternary carbon locations confirmed in CTX-] 0... 0.00.00. 335

LEWIS, R,J., HOLMES, M.J. & SELLIN, M.
Invertebrates implicated in the transfer of gambiertoxins
to the benthic carnivore Pomadasys maculatus... .... 22 eo ee ey eyes 361
LEWIS, R.J., SELLIN, M., {GILLESPIE, N.C., HOLMES, M.J., KEYS, A., STREET, R.. SMYTHE, H.,
THAGGARD, H. & BRYCE, S.
Ciguatera and herbivores: uptake and accumulation of ciguatoxins

in Clenochaetus striatus on the Great Bamier Reef . Sa WAR ESAS WARY 5 ee |
MANGER, R.L., LEJA, L.S., LEE, 5.Y.. HUNGERFORD, IN. & WEKELL, M.M_
Cell bioassay for the detection of ciguatoxins, brevetoxins, and saxitoxins --,-., 0. .,.,,see0-.., ST

MOLGO, J., JUZANS, P. & LEGRAND, A.M.
Confocal laser scanning microscopy: a new tool for studying the effects of ciguatoxin (CTX- IB) a and
D-mannitol at motor nerve terminals of the neuromuscular junction in sith ...,......-..-.-, 577

PARK, DL.

Reef management and seafood monitoring programs for ciguatera ... 6,46... 00s scene -ee-eeeee-s SBT
PAYNE, J.

Ciguatera poisoning: current issues in law ..... 0. cee cece eee eee eben tebe e ens 595
PEARN, J.

Ciguatera: dilemmas in clinical recognition, presentation and management ...........-....-..-04- 601
PEARN, J. & LEWIS, R.

Ciguatera: risk perception and fish ingestion... 2.02... 20.00... 0c c eee eee pet eee ee ee »- 605
RUFF, T.A. & LEWIS, R.J.

Clinical aspects of ciguatera; an overview... 2.60.6 cece eee et cee etre enna been enna 609
TERAO, K., ITO, E., OHKUSU, M. & YASUMOTO, A.

Pathological changes in murine hearts induced by intermittent administration of ciguatoxin.......... 621
VERNOUX, J.P.

The mouse ciguatoxin bioassay: directions for use to control fish for consumption ......-........45 625
VERNOUX, J.P. & LEJEUNE, J.

Ciguatera in the french West Indies... 22-26. eect eben eet e ee eee cere treet ents 631
ABSTRACTS
BROCK, J.A., JOBLING, P., McLACHLAN, E.M. & LEWIS, R.J.

Effects of Ciguatoxin-1 on electrical activity recorded intracellularly from rat tail artery in vitro... .., 454
CAPRA, M.F., CAMERON, J., FLOWERS, AE. & PURCELL, C.E.

Responses of vertebrate nerves to Ciguatoxin. .... 0.0... 0... cece eee ee ee eee rete acs 454
CHALOUPKA, M.Y., LEWIS, R.J. & SELLIN, M,

The changing face of ciguatera prevalence .. 22.2... ete tne tence eee beens 554

HAWN, S.T., CAPRA, M.F. & MILLER, D.M.
Oral and intraperitoneal administration studies of toxins derived from fish tissues and extracts of
cultured G, toxicus in the humbug (D. aruanus), Damsel-fish (P. wardi) and the Stripey

(Lc CAP PONOTAIUS) 65) 6 oe 8 eo hea te tog tafe el ote e efor ete del we) ere lesen angle “deeleceselp le eiate pale pare ste 554
HALLEGRAEFF, G.M.
On the global increase of harmful algal blooms .........-.. 000 c- eevee cece tee e seen ee neeees 560
HAMBLIN, P., McLACHLAN, E.M. & LEWIS, R.J.
Ciguatoxin-1 induces spontaneous synaptic activity in isolated sympathetic ganglia of guinea pigs .... 560

LEGRAND, A.F, & LOTTE, C.J.
Detection of ciguatoxic fish by using the binding property of ciguatoxins to voltage-dependant

sodium channels... 2.2.6... ccc cece eee eee tenet ened tenn e een beet ee bene 576
PALAFOX, N.
Evaluation of intravenous mannitol for treatment of actute ciguatera fish poisoning................. 576
PURCELL, C.E., CAMERON, J. & CAPRA, M.-F.
Modification of nerve conduction in the rat by brevetoxin (PBTX-3) ,. 2.2.2.0. 2.000000 50050 pees 586
SCHEUER, P.J.
Ciguatera research - an historical perspective . . sth ete sca swear terbeasuat ae tuliatss tears eRe

YASUMOTO, T., SATAKE, M., MURATA, M. & NAOKI, H.
Structures of maitotoxin and ciguatoxin congeners isolated from cultured Gambierdiscus toxicus ..... 600

Preface

This issue of the Memoirs of the Onfonitanes Musewrn 1s devoted to the International Workshop on
Ciguatera Management tha) was held on Bribie Island near Rasbane of 12-!6 Apni, 1993, The
Workshop was sponsored by the Australian Fisherics Research and Development Corporation and
the Queensland Department of Pamary Industries (QDPI), Scientists, medical practitioners and
fisheries managers with an interest in ciguatera attended the Workshop which focussed on current
research having implications for the management of ciguatera.

Fifty six registrants from Japan, USA, France, French Polynesia, New Caledonia, Germany and cach
of the eastem sea-board states of Australia attended. The Wi wkshop comprised talks, posters.and two
discussion sessions which specifically addressed (i) the detection of ciguateric fishes and (ii) the
management of ciguaicra cases, P, Scheuer opened the scientific pengram with an historical perspec-
live of modem viguatera research initiated by the Jatc A,H, (Hank) Banner and outlined some of the
challenges for the future.

Major themes of the Workshop were: |,Chemical and immunological aspects of the detection of loxins
involved in ciguatera. 2, Pharmacology and teajment of ciguatera, 3, Origin of the toxins involved
in ciguaters. 4, Clinical aspects and epidemiology of ciguatera.

Detection of eiguaterte fish

A cost-effective screen for ciguateric fish was recogaiscd as an important management tool able to
reduce the adverse effects of ciguatera on public health, fishenes, trade and tounsm (R. Lewis, D.
Park). CTX-] was widely considered the major target for screens for ciguateric fish (R, Lewis),
Several different approaches tu the detection of ciguateric fish were presented. Two approaches
measured the interaction between ciguatoxin and the voltage-dependent sodium channel through
either {i) the inhibition of brevetoxin hypding to sodium channels in 4 rut brain synaptosome
preparation (A.-M. Legrand) or (11) the cytotoxic effects of ciguatoxin on sodium channe -containing
cells pre-exposed 10 ouabain and veratridine (R. Manger). Both assays were more sensitive than the
mouse bioassay and may replace jn vive assays in laboralones possessing the specialised equipment
required. Nowever, itis unlikely that these approaches, as they siand, could he used fur the routine
screening of suspect fish prior lo consumption.

Amibody-based screens stil uppear to hold most promise for the cost-effective detection of caguateric
fish (R. Lewis), This a proach is the basis of a potential commercial test lo detect ciguateric fish being
developed by Hawaii emtect. D, Park presented a sunimary of the V sadsighen eg Of this solid-phase
immunobead assay (Ciguatect™) which was claimed to be able to detect most, if nat all erguatenc
fish. However, the test was reported to be unsuitable for detecting toxins involved in cigustera if the
fish flesh being sercened was slightly acidic (pl] = ~6.5), a factor that may considerably limit the
usefulness of the lest. Y. Hokama commented that the test may not work satisfactonly because the
solid-phase used in the Ciguatect™ test may not be as efficient at extracting ciguaioxins from Fish as
ihe correction fluid used for the solid-phase in the anginal stick lest he developed (apparently the
same antibody was used for both testis), When compared with the results of a well conducteid! mouse
bioassay, prediclive indices from 5% to 7595 were obtained with the Ciguatect™ test in an inde-
pendent study of ciguateric fish fron the Canhbean (R. Dickey). This result Supgcsts thatthe test may:
not be responding to the major toxins (as yet unidentified) present in these Canbbean fish. Lack of
ready access 10 samples of pure ciguatoxin or its analogues and gn inability to independently validate
the levels ot ciguatoxing present in the fish samples heing screened hinder attempts to validate (or
otherwise) the Ciguatect™ rest,

Pharmacology and rrearnent of clewarere

Major advances are being maue on how ciguatoxins cause humin poisoning (J, Moloo, & McLachlan.
J. Brock, P, Hamblin, E. Benoit, K. Terao, C. Purcell, M. Cupra), This research highlighted the
usefullness of ciguatexin as a tool enabling physiologists to understand physiological processes, The
precise mechanism by which mannitol, the treatment of choice for acute ciguatera, acts to relieve the
symptoms of cigualera remains unelear, A deuble-hlynd clinical study of the mannitol treatment is
being conducted (N, Palafox) bul results were not available at the meeting, Clinical experiences with
the mannitol therapy (D. Blythe, N. Palafox) continue to be positive and mannitol should remain the
tecatment of choice for acute ciguatera, Confirmation of the clinical findings would be assisted by the
Hevelopment of an animal model for ciguaters thay responds to mannitol,

Clinical aspects and epidemiology of ciguatera
While most of the clinical features of csguatera are well documented, the longaenn effects of crzuatern

and how frequently these occur are tly understood (T. Ruff). Follow-up research on victinus is
required to establish the true extent of long-term effects, especially the allengy-like reactions thal can
last.afier a single exposure to toxic lish. J, Pear reported thal ciguatera remains typically a poorly
recognised an blr disease, despite the introduction of the mannitol Giga 3 PT maintams a
database that covers 27 years of ciguatera cases reported in Queensland. An analysis of this dalabuse
using recently developed Slalistica meacelligg approaches revealed major shifts in the species and the
niture of the poisoning in Queensland (M_ Chaloupka). P. Dalzell provided imponant insights into
how ciguatera could he managed in the Pacific where it can he hi hly prevalent

The Iegal situation with regard to cigualera was.alsv discussed at the orkshop by J. Payne. Duty of
¢are issues and the Queensland Workplace Health and Safety Act could be pursued fora possible
successful court action against suppliers of toxic fish, Legal opinion was that duty of eare rssues would
not be adequately addressed jf the problem of ciguatera was not continually monitored and if upto
Gils Management options were not being implemented.

Origin and identification of toxins invelved in ckeuatena

Gamibrerdiscus toxics ts now widely accepted as the organism that produces the gambicrtoxins which
are the precursors to toxins involved in ciguatera. Indeed this organism tnay be the only source of
toxins involved in ciguatera. T. Yasumvte reported the structure for the gambiertoxin GTX-4A (S2
ep-GTX-4B). the major gambiertoxin produced by a Rangiroa Atoll strain of G. reaicws grown in
culture. This study contirms that G. toxicns is indeed the origin of the ciguatoxins isolated [rom fish,
Long and short range inveise detected NMR of CTX-1 confined the structure orginally proposed
for cyguatoxin (R. Lewis), The structure of a maitotoxin produced by cullured G. faxicus was also
presented by T. Yasumoto. Maitotoxin is the largest non-repeating unit compound for which a
structure is Known. Elucidation of jis suructure represents a significant milestone in natural product
chemistry. Maitotoxin’s structure is only distantly related to the ciguatoxins, thereby ending specu-
lation that matlotoxin may be precursor of the ciguatoxins.

The cnvironmental factors that cause the upsurges of ciguatera remain poorly understood (M, Holmes,
R, Lewis, ¥, Hokama, D, [chinotsubo, R. Bagnis. U. Kaly). The ongin of ciguatera was question by
J-P. Vernoux based on studies conducted in French Polynesia. A rapid extraction method for isolating
ciguatoxins from G. toxicus may assist studies On Cuxin production in culture (J, Babinchak), Studies
on the vectors transfernng the gambjentoxins to camiverous Nshes revealed that invertebrates
(shrimps) may he involved in some community strictures, whereas in other communities herbivorous
fish, such as Crenechaetus striatus, may be the key vector involved (R. Lewis).

As well as ciguatera, a minge’ of other toxic algae which produce biotoxins that cause diurrhetic,
paralytic, neurotoxic and amnesic shellfish peysoning could become a problem im tropical and
sub-tropical waters through ballast water introductions and/or environmental degradation (G. Halle-
eracfh). These biotoxins have the potential to severely damage fisheries such as the shellfish fisheries
in Qucensland which are presently unaffected by such toxins.

Future directions for eiguatera reseuteh

The Workshop identified wt least four areas for further Australian research that could result Ww
improved management of ciguatera and related seafood poisonings. These were in order of priority:
{t) development ofa cus-effective sereen forciguate re fish: an uncoordinated meremational effort has
to-date failed to develop a useful sereerang test for ciguilenc fish. Results presented al the Workshop
indicate that a biosensor could be developed based on ciguatoxins exceptionally high specificity and
affinity for the sodium channels found in nerve ancl musele tissues:

{i ydetermindtonaf the enviranmental factors responsible for flare-ups of ciguatera: lille as presently
Snown of the precise environmental factors that resol in the proliferation of the toxins that callse
eiguaicra bul fuman activities including “pollution” have been implicated The ciguaivra “hop sput'
in Hervey Bay tepresents an edeal Jocalion te pursue studies on the contnbuting Factors.
(ii) assessment of the potential for orher toxic dinoflagellates to enter rropicsil and Sab-trepical waters
throveh batlast-water introduciony: studics are required ley determine if toxic dinnflage(late species
have already beech introduced inte tropical waters und to assess Ihe potential for these ty bloom and
fornew ballast water introductions of toxic dinoflagellate cysts.

(tv) assess the extent and nature of adverse reactions Te seafoud con Vupiron’ a survey is required (6
determine the extent and nature of the allergy-like reactions to seafood, including such reactions thar
fallow etguatera, Such reactions have not been well qaantified in Australia but may have a consuler

ahle inypuct on the marketabiliry of seafood,

Richard J. Lewis, Chainncin of the Workshey, Scientufie and Organizing Cominrees
Devepition Bays fut, (994

PRODUCTION OF CIGUATOXINS IN CULTURED GAMBIERDISCUS TOXICUS

JOHN A. BABINCHAK, PETER D.R. MOELLER, FRANCES M, VAN DOLAH,
PAMELA B, EYO AND JOHN 8. RAMSDELL

Babinchak, J.A., Moeller, P-D.R., Van Dolah, F_M., Eyo, P.B. & Ramsdell, J.S, 1994.08 O1;
Production of ciguatoxins in cultured Gambierdiscus toxicus. Memoirs uf the Queensland
Museum 34(3); 447-453, Brisbane, ISSN 0079-8835,

Production of ciguatoxin congeners (CTX) from mass cultured dinoflagellates appears. to be
the only source of CTX that has the potential of providing sufficient quantities of purified
toxins for studies on biosynthesis, structural analysis, pharmacology, biotransformation and
detection. Established Gambierdiscus toxicus clones and recent isolates from Tahiti, Guam
and Grand Cayman Island were mass cultured and toxins separated by step-wise clution on
a silica gel column, CTX was identified by its binding competition with [5H]-brevctoxin for
sodium channel receptor sites in rat synaptosomes. MTX was identified by its ability to
induce calcium flux activity in rat pituitary cells. Although the silica gel column separated
CTX from MTX, general toxicity of the CTX congeners decreased afict separauon. A sample
partially purified CTX from G, toxieus clone MO2 was used as a standard for evaluating
ie assays.

John A. Babinchak, Peter D.R. Moeller, Frances M. Van Dolah, Pamela B. Eye and John
5S. Ramsdell. National Marine Fisheries Service, P.O. Box 12607, Charlesion, SC 29412

USA; 28 March, 1994.

Two classes of cyclic polyethers produced by
the epiphytic dinoflagellate, Gambierdiscus
toxicus, are non-polar ciguatoxins (CTX), the
principal toxins causing ciguatera, and polar
maitotoxins (MTX). CTX was first isolated from
the Pacific red snapper (Scheuer et al.,1967) and
later purified from moray ee) liver and viscera.
Multiple congeners of CTX (CTX-1, CTX-2 and
CTX-3) have been identified in finfish (Lewis et
al,,1991), CTX-1 being identical to CTX from
moray eel liver. Several CTX congeners have
been identified in wild and cultured G. toxicus
(Legrand et al..1992; Holmes et al.,1991; Satake
etal..1993) although none identical to that found
in the moray eel. Based on their structural
relutionships, Lewis et al. (1991) propased that
CTX-1 and CTX-2 in finfish represent oxidation
products of two different toxins im G. toxicus.

Even in the moray eel, which is the most toxic
species, CTX content is extremely low, usually
only several ppb in whole bodies (Murata et
al.,.1990). Thus purification of large quantities of
CTX from finfish has not been successful. How-
ever, obtaining large quantities of the dinoflagel-
late CTX congeners is limited only by the ability
io identify and mass culture a G. toxicus clane
capable of producing high levels of these toxins.
While MTX levels greatly exceed thuse of CTX,
strain-dependent differences in the composition
of toxins produced by cultured G. toxicus have
been described (Holmes et al.,1991). Toxin
profiles appeared to be stable, suggesting a

genetic basis for the production of different toxin
profiles. However, no isolate has yet produced
substantial quantities of CTX in culture, with
reported yields Jess than one mouse unit per mil-
lion cells (Holmes et al..1991: Satake et al. 1993).

We proposed to screen a collection of 44 G.
fexiciés clones from 12 geographically distinct
locations worldwide for their production of CTX
in mass culture. The screening of clones has
traditionally been a laborious task, requiring the
use of mouse bioassay to identify toxins. Since
CTX and MTX elicit similar symptoms in the
mouse bioassay, the identification of CTX versus
MTX in dinoflagellate extracts has required cx-
tensive punfication. Our laboratory 1s capable of
distinguishing and quantifying CTX and MTX in
crude extracts of G. soxicus using a battery of
established bioassays. These include a rapid,
high-throughput in virra cytotoxicity assay for
total toxicity, and a Ca** flux assay as well as
Teceptor binding competition between CTX and
(PH]-hrevetoxin to distinguish MTX from CTX
(Van Dolah et al., in press). Maitotoxin causes an
increase in Ca** permeability, possibly through
voltage dependent Ca?* channels (Takahashi et
al.,1982; Gusovsky & Daly,1990), Ciguatoxin
promotes Na* channel opening by binding to site
5 on the voltage dependent sodium channel
(Lewis & Endean, 1984: Bidard et al,,1984), This
sile is also recognized by the brevetoxins, a re-
lated class of dinoflagellate polyether toxins that
can be displaced by CTX {Lombet et alL..1987;

448

Baden, 1989). The objective of this study was to
develop a rapid, single column technique to
separate CTX from MTX in crude dinoflagellate
extracts. This would expedite identifying G.
toxicus clones which produce high levels of CTX
congeners in mass culture.

MATERIALS AND METHODS

STOCK CULTURES

Isolation procedure for clonal cultures fol-
lowed Babinchak et al. (1986). Stock cultures of
clonal strains of Gambierdiscus toxicus (Table 1)
were maintained at 27°C under an illumination of
30-40nEM7S" and a 16:8 hour light:dark cycle
without aeration. Illumination was provided from
above by a 50:50 mixture of Cool White (North
American Phillips Lighting Corp.) and Vita-Lite
(Duro-Test Corp.) fluorescent bulbs. K medium
(Keller et al.,1987), an enriched seawater
medium used in all culturing, was modified by
eliminating CuSO,, Tris buffer, silica and using
ES vitamin concentrations (Guillard & Keller,
1984). Seawater was collected from a saltwater
well in Vero Beach, FL, (Florida Institute of
Technology field station). The seawater (35-36
°/oo) was filtered through 0.45pm cartridge filters
into sanitized polycarbonate carboys and
refrigerated in the dark. The seawater was
autoclaved in 10 litre borosilicate glass bottles.
The vitamin mixtures and enrichments for K
medium were prepared in concentrated stocks,
filter-sterilized and autoclaved respectively, and
stored frozen. The enrichment and vitamins were
added aseptically to autoclaved seawater which
was then used immediately for culturing. G.
toxicus clones were harvested by filtration
through 12pm polycarbonate membranes,
washed 3 times with sterile seawater and inocu-
lated at 200-300 cells/ml into 2.8 litre Fernbach
flasks containing 1 litre of medium. Stock cul-
tures of G. toxicus were harvested for transfers
and mass culture inoculum at 10-14 days.

Mass CULTURE

Twenty-nine clones (Table 1) were selected for
mass culture. Micro-carrier spinner flasks (Bellco
Glass, Inc.), designed for suspension cell culture
systems, were selected as mass culture vessels.
The 8 litre model was of a design, wide mouth
with two access ports, and weight that could be
easily handled and cleaned. The maximal work-
ing volume of these flasks for dinoflagellate cul-
ture was increased to 12 litre. Magnetic stirring
units (Bellco Glass, Inc.), designed for gentle

MEMOIRS OF THE QUEENSLAND MUSEUM

agitation, maintained a stirring speed of 20 RPM.
Shelving units accommodated 6 culture vessels
and stirrers and provided illumination from above
with 4ft Vita-Lite fluorescent bulbs, and from
behind with 2ft Cool White fluorescent bulbs at
404EM?S" using a 16:8 hour light:dark cycle.
Two shelving units were installed in each of two
walk-in environmental rooms (1800 cu. ft. and
600 cu. ft.). These rooms were maintained at 27°C
and provided space for 24 culture vessels. Vessels
were inoculated at a concentration of 500-1000
washed cells/ml. Air was supplied at | litre min"!
by Whisper 800 aquarium air pumps (Willinger
Bros., Inc.) and filtered through AQ microfiber
disposable filter tubes, (Balston Filter Products).
Aeration was initiated after 10-14 days incuba-
tion and bubbled into the culture vessels through
sterile plastic air diffusers, (Lee’s Aquarium
Products).

HARVESTING PROCEDURE

Harvesting the micro-carrier spinner flasks re-
quired minimal handling of the vessels. After
21-28 days of incubation, the stirring and aera-
tion apparatus were removed, the flask swirled
and a 10ml sample taken for determining cell
counts. The cells were allowed to settle and the
supernatant removed with a peristaltic pump and
filtered through a 142mm stainless filter holder
(43u.m polyester membrane, Spectrum Medical
Industries, Inc.). The settled cells from all the
vessels were combined and collected on a 43m
polyester membrane in a 90mm glass filter holder
which produced a cake of wet cells. The cells
were stored frozen at -20 or -90°C until extracted
for toxin content.

CELL COUNTS

Cell counts on individual micro-carrier flasks
were determined at harvest using natural
chlorophyll autofluorescence and direct
epifluorescence microscopy. Duplicate 0.5ml
volumes of each micro-carrier sample were col-
lected on 5m black polycarbonate membranes
and observed at 320x on a video display monitor.
A Leitz Dialux 20 microscope was used,
equipped with a 150W xenon lamp, fluorescence
vertical illuminator, KG1 heat filter, BG23 blue
filter and a Leitz 12 filter block. Twelve fields or
a minimum of 400 cells per sample were counted.

PREPARATION OF CTX STANDARD

A 150g (wet weight) sample of clone MQ2 was
disrupted in 100% methanol at 0°C with a Tekmar
500 watt sonic disrupter and filtered through a

PRODUCTION OF CIGUATOXINS IN CULTURED G. TOXICUS

Table 1. Gambierdiscus toxicus Culture Collectiont,
National Marine Fisheries Service, Charleston
Laboratory

ISOLATOR

*CI03, *C104,
*CI05, C108,
C109, C110, C112,
C113, C114, C115,
*CI16, *CI18

Grand Cayman Is. Babinchak

CZ2, CZ3,CZA |Cozumel, Mexico Babinchak

MQI1, *MQ2 Martinque Babinchak

G03, *G05, Guam Babinchak
#G06, *G15,

*G16, *G17

Palau Babinchak
*G01, G20, *G23 | Pohnpel, FSM Babinchak

*MR-1 Babinchak
*TO1B, *T02B, Babinchak
#T03B, *TO4B,

*T11B, *T15B
| Tahiti | Bagnis |
Australia
St. Barthelemy

SBO1, SBO3, Durand-Clement

SB04

Tahiti Durand-Clement
*HIT-10, Tahiti Legrand
*HIT-25
#TP125B Dry Tortugas Tomas
*T39 Hawaii Withers
* Clones successfully mass cultured in micro-carrier
system
# Clones not successful in mass culture

+G. toxicus clones isolated by Babinchak are available
from NMFS

0.22,.m polycarbonate membrane. This crude ex-
tract was first fractionated on a G-10 Sephadex
size exclusion column (0.5m x 3.8cm i.d.), using
gravity flow with a 100% methanol mobile-
phase. The Sephadex CTX fraction, as deter-
mined by PbTx competition assay, was further
fractionated on a C18 (30ym particle size)
reverse phase silica column (1.5m x 3.8cm i.d.).
The sequence of elution solvents was 100%
water, 50%, 25%, 12% and 6% water:methanol
using gravity flow.

IATROBEAD FRACTIONATION

A crude extract of 5g (wet weight) from each
of 6 (MQ2, T04, G15, Hit25, G17 and T11B)
mass cultured G. toxicus clones was prepared as

449

for the CTX standard. The methanol extract was
evaporated to dryness and taken up to a final
volume of 5-10ml chloroform. This solution was
introduced to a Michel-Miller column (350mm x
40mm i.d.) equipped with a pre-column (130mm
x 22mm i.d.). Both columns were packed with
lIatrobeads, (porous, beaded silica; pH 6.8, 60.m
particle size, 80A pore size, Iatron Laboratories,
Inc.). The solvent scheme used to fractionate the
sample at pump flow rates of 5-8 ml/min was
100% chloroform, 2%, 5%, 10%,15%, 25% and
$0% methanol:chloroform, 100% methanol and
finally 15% water:methanol.

MOUSE BIOASSAY

Column fractions were evaporated to dryness
and brought up in methanol and dilutions made
up in 0.9% saline containing 1% Tween 20.
Toxicities were determined by injecting 0.25-
0.5ml of appropriate toxin concentrations in-
traperitoneally into female ICR mice weighing
c.20g. Three mice were initially injected per dose
which was increased to at least 5 mice for doses
used to determine the LDso toxicity. Total
lethality is expressed in mouse units (MU),
defined as the LDso dose for a 20g mouse over
48hr.

CELL CULTURE

Rat pituitary tumor cells (GH4C1) were main-
tained at 37°C and 5% COz in Hams F10 nutrient
mixture supplemented with 2.5% fetal bovine
serum and 15% horse serum in the absence of
antibiotics (Hams F10+). The cells were passed
at weekly intervals to maintain their exponential
growth, for a maximum of 10 passages. These
cells were used in the cytotoxicity assay and
calcium uptake experiments.

CYTOTOXICITY ASSAY

Cytotoxicity was determined by a modification
of the procedure of Mosmann (1983).GH4C) cells
were plated using 0.1ml Hams F10+ in 96-well
tissue culture plates at a concentration of 5.0x 10°
cells/ml. Column fractions were evaporated to
dryness and serially diluted in methanol. Dupli-
cate wells of fraction-cell mixtures were then
incubated for 18hr at 37°C. For determination of
viability, 15j.1 of 3-(4,5 dimethylthiazol-2-yl)-
2,5-diphenyl tetrazolium bromide (MTT, 5
mg/ml in PBS) were added to each well and the
cells incubated 4hr at 37°C. Mitochondrial
dehydrogenases in live cells convert the MTT to
an insoluble formazan crystal. After incubation,
the cells were solubilized by addition of L1O% SDS

450

in 0.1N HC! and absorbance at 570nm was
recorded using a Titre Tek 96-well plate reader.
Non-specific absorbance due to media and non-
converted MTT were subtracted to yield a cor-
rected absorbance value. Cytotoxicity was
considered positive if the reading was within 10%
of the positive control, negative if within 20% of
the negative control and partial if in-between, In
this study, a minimal lethal concentration (MLC)
of toxin was defined as the last dilution to give
only positive results.

CALCIUM UPTAKE EXPERIMENTS

Calcium uptake experiments were performed
by a modification of the method of Enyeart et al.
(1986), GH4C\ cells were plated in 96-well plates
in O.1ml of Hams F10+ and allowed to attach
overnight. For the assay, the medium was
replaced with Hams F10+ containing 5pCi/ml
SCa?+ and test treatments. Cells were incubated
with treatments for 10min at 37°C. To terminate
the assay, the extracellular “Ca** was aspirated
off and the cells rinsed 3 times with ice cold Hams
FHHO+. Cells were solubilized by the addition of
scintillation cocktail] and the plates counted
directly in a 96-well format microplate scintilla-
tion counter (Wallac). The calcium channel
agonist. Bay K 8644, was used as a positive
control, Samples were considered positive for
values greater than 2 times the negative control.

BREVITOXIN (PbTx) DISPLACEMENT ASSAY
Assay of binding competition for the PhTx Site
on sodium channel receptors was carried out in
96-well plates by a modification of the method of
Poli et al, (1986). All assays were carried out in
the presence of binding buffer [50 mM Hepes,
(pH 7.4), 130mM choline chlonde, 5.5mM
glucose, 0.8mM magnesium sulfate, 5.4mM
potassium chlonde, 1.Ome/ml BSA, and 0.01%
Emulphor-EL 620). To cach sample well, 0 reac-
tion mixture of 35] of FH]PbTx3 (SnM) and
35ul of test treatment were added in binding
buffer. To this reaction mixture, 135.1 rat brain
synaptosomes in binding buffer were added.
Plates were incubated at 4°C for Uhr, then recep-
tor bound [SH)PbTx3 was trapped onto a glass
fiber filter pad using a 96-well filtration apparatus
(Milliblot, Millipore). The filter pad was dried
and impregnated with solid scintillant und bound
PH]PbTx3 detected by liquid scintillation
spectroscopy ina 96-well format microplate scin-
ullation counter (Wallac). Fractions which
caused >50% decrease in [°H]PbTx3 binding
were considered positive. PbTx-3(luM) was

MEMOIRS OF THE QUEENSLAND MUSEUM

used as a positive control and caused complete
inhibition of FHJPbTx3 binding at the receptors.

RESULTS

At an inoculum level of 500 cells/mi, the cel
yield of Martinique G. toxicus clone, MQ2, in
mass culture (4532+526 cells/ml. N=44) was
equivalent to the yield obtained in 11 of culture
using 2.81 Fernbach flasks (473541105 cells/ml,
n=26). Before an aeration system was introduced
into the mass culturing system, cell yields with
MQ? (3325+£502 cells/liter, n=52) were 30% less
than produced in Fernbach culture. With aeration,
the media pH rose to 8.8 units, but without aera-
lion the rise was to pH 9.9. When the inoculum
size for MQ2 was doubled, the final biomass
yield increased 60% (0:85—1.3 g/l, wet weight).

Four clones that produce copious amounts of
mucoid material in culture caused a problem in
mass culture when aeration was introduced
(Table 1). Bubbling created by the aeration drove
the mucoid secretions and entrapped cells onto
the wall of the flasks above the medium, separal-
ing the cells from their nutnents. When this oc-
curred, mass culturing was terminated. Final
yields in mass culture of the remaining 25 clones
ranged from 0.6 to 1,5 g/l (wet weight).

While preparing the CTX standard, 2 suites of
maitotoxin eluted within thr dunng fractionation
on G-10 Sephadex. The CTX family of toxins
eluted from the column between 10-48hrs, but
peaked at 15hr. When the CTX fraction was run
on the C18 column in the next stage of purifica-
tion, fractions collected during elution with 100%
water and 50%-25% water;methanol were weak-
ly cytotoxic. These represent fractions where any
remaining MTX not separated from CTX on G-
10 would be expected to elute off the CIR
column. No toxicity was observed again until
CTX containing fractions were eluted with 12%-
6% water:imethanol.

The CTX fractions used as the CTX standard,
9_Oml total volume, had a mouse LDsy toxicity of
D.OO7p/MU for a total of 1.3 million MU
produced from the 150g of MQ2 processed. The
mice displayed intense lumbar contractions and
progressive paralysis from hind to frontJegs. The
foxin output for MQ2 was equivalent to one
mouse unit per 2000 cells. The CTX standard had
seytoloxicity of 0.005nUMLC and was negative
ur weakly positive in calcium Mux and competi-
tively inhibited labeled PbTx binding. The MTX
fractions were calcium flux positive abd did not
inhibit PbTx receptor binding.

PRODUCTION OF CIGUATOXINS IN CULTURED G. TOXICUS

The toxin elution profiles for the 6 G. toxicus
clones processed on Iatrobeads were similar and
followed the solvent elution scheme. The first
fraction isolated from the Iatrobead column with
100% chloroform was unique in that it displayed
calcium flux activity, but was not toxic to mice or
GH«Ci cells. CTX eluted in the next 1-3 fractions,
2%, 5% and occasionally 10% methanol:
chloroform, inhibited PbTx binding, but did not
affect mice or display cytotoxicity. After 1 or 2
fractions with no activity, 10% and 15%
methanol:chloroform, MTX eluted with 25% and
50% methanol:chloroform and 15% water:
methanol. The MTX fractions had high cyto-
toxicity, calcium flux activity, but did not inhibit
PbTx binding on sodium channels receptors.

DISCUSSION

Dinoflagellates are generally considered sensi-
tive to stirring in culture displaying cellular
damage and reduced growth rates (White, 1976;
Galleron,1976; Tuttle & Loeblich,1975). How-
ever, the agitation of the micro-carrier flask sys-
tem maintained G. foxicus in suspension without
any detrimental effect on its growth rate, and with
aeration, the pH of the medium stabilized during
the final growth phase and provided final yields
equivalent to Fernbach culture. Silicates which
dissolve off the walls and crystallize while
autoclaving seawater in borosilicate bottles
(Brand etal.,1981) provided fine attachment sites
for the epiphytic G.toxicus in the medium. Aera-
tion supplied the carbon needed for growth and
also stabilized the pH in seawater by keeping it
from rising too high through the interaction of the
carbonate system with pH (Guillard & Keller,
1984), A pH-stat system (Goldman et al.,1982),
would provide finer control of the pH as well as
inorganic carbon for photosynthetic uptake. This
should produce higher biomass yields and the
reduction in creation of bubbles would also allow
culturing of dense mucoid producing clones.

The initial mass culturing of MQ2 was a
balance between using available incubator space
for growing inoculum and mass culturing vessels.
Because of this, inoculum size was limited to 500
cells/ml. By doubling the inoculum size of MQ2
in later studies, a 60% increase in the final yield
was achieved which would be expected to occur
with other clones if optimal inoculum were used.
Limited culturing (n=3) using 12 litre polycar-
bonate culture vessels (Nalgene Corp.), with 18
litre working capacity, provided a 50% increase
in culturing capacity with the same biomass

451

yield:medium volume ratio. The flasks occupied
the same space as the borosilicate micro-carrier
flasks, but were lighter and safer to handle.

It was critically important to filter and wash
each inoculum and harvest cells in mid-Log
growth phase while mass culturing. Neglect of
these parameters increased the bacteria load and
decreased growth vigor which resulted in reduced
biomass harvest. Phenotypic changes that oc-
curred during mass culturing that typified non-
compliance to these parameters were increased
mucoid production and adhesion of the cells to
the vessel walls.

Wet weight yields at harvest were empirically
related to the cellular volume of the clone.
Generally, the larger the clone, the greater
biomass produced per unit volume of medium.
Acclimation to in vitro culture parameters was
another determinate of biomass yield. Sixteen of
the mass cultured clones were isolated less than
9 months previous to this study. G. toxicus can
take up to one year to acclimate to culture condi-
tions after isolation (Bomber et al.,1989). An
increase in final yields would be expected for new
isolates not yet acclimated to culture conditions.

The amount of CTX (purportedly) isolated
from clone MQ2 using size exclusion G-10
Sephadex, followed by C18 reverse phase silica
columns was 500-fold more on a per cell basis
than previously isolated from this clone. How-
ever, previous separation of CTX from MTX was
with a silicic acid column and a step-wise elution
with chloroform and methanol (Tachibana, 1980),
a separation technique recommended for use in
ciguatera research (Anderson & Lobel,1987). In
the present study, the toxicity of CTX appeared
to be unstable to contact with unprotected silica
or the combination of silica and chloroform. This
would explain the loss of toxicity, but retention
of PbTx competitive binding activity when frac-
tionated on Jatrobeads. To test this possibility, a
sample of the CTX standard was fractionated on
Jatrobeads. CTX was collected apart from the
lipophilic, active calcium flux material using
step-wise 100% ethyl acetate to 5% water/
methanol eluents. The CTX standard lost 50% of
its mouse toxicity and greater than 50% of its
cytotoxicity. A hypothesis for this observation
could be that CTX toxins exist in nature as
epoxides. Though possibly still toxic after such
an epoxide is opened, much of the toxicity would
be lost on silica which may facilitate epoxide ring
opening. Though only a hypothesis, the CTX
structures (and MTX) that are known, would
provide excellent sites for epoxide formation.

452

Since CTX toxins were not stable on an
latrobeads packed column, this was not an ideal
one-column procedure useful for rapid screening
of G. texicus clones for CTX jon, How-
ever, they have proved to be excellent column
packing for MTX punifications providing excel-
lent separation with no loss in toxicity. Unlike
other silica media, Iatrobeads allow higher con-
centrations of water to be used as eluant without
destroying or dissolving the silica.

When separating CTX from MTX by normal
phase chromatography, a non-polar, non-toxic,
calcium flux active fraction, not previously iden-
tified, was included with CTX. This led us to
believe that MTX was still present in these frac-
tions. If hexane:methanol partitioning was used
before column purification, this lipophilic com-
pound was extracted in the hexane phase. unlike
CTX or MTX.

Additional CTX extractions from clone MQ2
are in progress to determine if the high yield of
CTX in this study was a function of the separation
procedure or a serendipitous production of CTX
by clone MQ? in culture. Pacific G. toxicus clone,
G15, which produces several interesting fractions
(isolated from the [atrobead column) that com-
pelitively displaced PbTx binding to sodium
channels, is now in mass culture, Although the
identity of the CTX standard was confirmed with
a negative calcium flux assay and positive com-
petition for PHJPbTx3 binding, its yield from
cultured G. toxicus ts 500 times that previously
reported (Balthrop & Herring,}990), which
leaves. lingering doubts that the CTX standard
contained only CTX congeners. To eliminate this
uncertainty, further production and punfication
of the CTX standard using HPLC and definitive
confirmation of the compounds with mass
spectroscopy and NMR are in progress.

ACKNOWLEDGEMENTS

We wish to acknowledge the technical assis-
tance of Charleston NMFS staff members: B,
Lanouc, B. Haynes, S. Knoepp, and A. Sperr.

LITERATURE CITED

ANDERSON, DM, & LOBEL, P.S. 1987, The comtinu-
ing cng of cieuatera. Binlogecal Bulletn 172:
49-107,

RABINCHAK, J.A., JOLLOW, DJ, VOEGTLINE,
M.S. & HIGERD, T.B, 1986, Toxin prodyection
hy Gambierdiccuy tuxicus isolated from the
Dora Keys. Manne Fisheries Review 48: 53-

nu

MEMOIRS OF THE QUEENSLAND MUSEUM

BADEN, D.G. 1989, Brevetoxins: unique polyether
dinoflagellate toxins The FASEB Journal 3:
1807-1817.

BALTHROP, J.E. & HERRING T.L. 1990, The inter-
action of brevetoxin with rat brain synapeosomes.
Abstracts of the Third International Conference on
Ciguatera Fish Poisoning, Lajas, Puerto Rico,

BIDARD, I.N,, VISVERBERG, H.P.M., FRELIN, C,
CHUNGUE, E., LEGRAND, A.M.,BAGNIS, R.
& LAZDUNSKI, M, 1984. Ciguatoxin is a novel
type of Nat channel toxin. Joumal of Biological
Chemistry 259: 8353-8357.

BOMBER, J.W., TINDALL, D.R. & MILLER, D.M.,
1989, Genetic variability in toxin potencies
among seventeen clones of Gambierdiscus
toxicus (Dinophyceae). Journal of Phycology 25;
617-625,

BRAND,L.E., GUILLARD, R.R.L. & MURPHY, LS.
198L. A method for the rapid and precise deter-
mination of acclimated phytoplankton reproduc-
lion rates. Journal of Plankton Research 3:
192-201.

ENYEART, J.),, AIZAWA T. & HINKLE, P.M. 1986.
Interaction of dihydropyridine Ca** agonist Bay
K $644 with normal and transformed pituitary
cc American Journal of Physiology 259: C95-

CL02.

GALLERON, C. 1976. Synchronization of the manne
dinoflagellate Amphidinium carteri m dense cul-
tures, Joumal of Phycology 12: 69-73.

GOLDMAN. J.C., AZOY, Y., RILEY, C.B. & DEN-
NETT, MLR. 1982. The effect of pH in intensive
microalgal cultures. 1. Biomass regulation. Jour-
nal of Experimental Marine Biology and Ecology
§7: 1-13.

GUILLARD, R.R.L, & KELLER, M.D. 1984. Cultur-
ing dinoflagellates. 391-442. In D. Spector, (ed)
*‘Dinoflageliates," (Academic Press: New York),

GUSOVSKY, F. & DALY, J.W, 1990. Maitotoxin: a
unique phanmacological tool for research on cal-
cium-dependent mechanisms. Biochemical Phar-
macology 39: 1633-1639. :

HOLMES, J.H., LEWIS, RJ, POLI, M.A, & GIL-
LESPIE. N.C. 1991. Strain dependent production
of ciguatoxin precursors (Gambiertoxins) by
Gamiblerdiseus toxieus (Dinophyceae) in culture.
Toxicon 29: 761-775-

KELLER, M.D., SELYIN, R.C., CLAUS W. & GUIL-
LARD, R.R.L. 1987. Media for the cultare of
oceanic uluaphytoplankton. Joumal of Phycology
23: 633-638

LEGRAND, A-M., FUKUI, M., ISHIBASHI, Y. &
YASUMOTO, T, 1992. Characterization of
ciguatoxins from differem fish species and wild
Gambierdixcas toxicus, 25-32. In T.R, Tosteson
(ed) “Proceedings of the third intemational con-
ference on ciguatera fish poisoning, Puerto Rico
1990." (Polyscience Publications: Quebec).

LEWIS, R,J.& ENDEAN, 1984. Ciguatoxin in the flesh
and viscera of the barracuda, Sphyraena jello.
Toxicon 2): 19-24.

PRODUCTION OF CIGUATOXINS IN CULTURED G. TOXICUS

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae). Toxicon 29; 1115-1127.

LOMBET, A., BIDARD, J.N. & LAZDUNSKI, M.
1987. Ciguatera and brevetoxins share a common
receptor site on the neuronal voltage-dependent
Na* channel. FEBS Letters 219: 355-359.

MOSMANN, T. 1983. Rapid colorimetric assay for
cellular growth and survival: Application to
proliferation and cytotoxicity assays. Journal of
Immunological Methods 65: 55-63.

MURATA, M., LEGRAND, A-M., ISHIBASHI, Y..,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin and its congener
from the moray eel, Gymnothorax javanicus, and
its likely precursor from the dinoflagellate Gam-
bierdiscus toxicus. Journal of the American
Chemical Society 112: 4380-4386.

POLI, M.A., MENDE, T.J. & BADEN, D.G. 1986.
Brevetoxins, unique activators of voltage-sensi-
tive sodium channels, bind to specific sites in rat
brain synaptosomes. Molecular Pharmacology
30: 129-135.

SATAKE, M., ISHIMARU, T., LEGRAND, A-M. &
YASUMOTO, T. 1993, Isolation of a ciguatoxin

453

analog from cultures of Gambierdiscus toxicus.
p.575-579, In T.J. Smayda and Y. Shimizu (eds)
‘Toxic phytoplankton blooms in the sea.’ (El-
sevier: North-Holland).

SCHEUER, P.J., TAKAHASHI, W., TSUTSUMI, J. &
YOSHIDA, T. 1967. Ciguatera: isolation and
chemical nature. Science 155: 1267-1268.

TACHIBANA, K. 1980. Structural studies on marine
toxins, Ph.D. Thesis, University of Hawaii.(Un-
publ.)

TAKASHASHI, M., OHIZUMI, Y. & YASUMOTO,
T. 1982. Maitotoxin, a Ca** channel activator
candidate. Journal of Biological Chemistry 257:
7287-7289.

TUTTLE, R.C. & LOEBLICH, A.R. 1975. An optimal
growth medium for the dinoflagellate Cryp-
tothecodinium cohnii, Phycologia 14: 1-8.

VAN DOLAH, F.M., FINLEY, E.L., HAYNES, B.L.,
DOUCETTE, G.J,, MOELLER, P.D. &
RAMSDELL, J.S. in press. Development of rapid
and sensitive high throughput pharmacologic as-
says for marine phycotoxins. Natural Toxins.

WHITE, A. W. 1976, Growth inhibition caused by
turbulence in the toxic marine dinoflagellate
Gonyaulax excavata. Journal of the Fisheries Re-
search Board of Canada 33: 2598-2602.

454

EFFECTS OF CIGUATOXIN-1 ON ELECTRICAL AC-
TIVITY RECORDED INTRACELLULARLY FROM
RAT TAIL ARTERY IN VITRO. Memoirs of the
Queensland Museum 34(3): 454. 1994:—Ciguatoxin-1 (CTX-
1) is a lipid soluble toxin from the benthic dinoflagellate,
Gambierdiscus toxicus, it is responsible for the disease
ciguatera which exhibits a range of symptoms inyolving the
peripheral nervous system. CTX-1 has been credited with a
selective action on tetrodotoxin (TTX)- sensitive Na* chan-
nels and induces spontaneous nerve action potentials due to
opening of sodium channels at normal resting potential. In this
study the effects of CTX-1 on the rat tail artery have been
investigated, Intracellular recordings were made from isolated
sections of rat tail artery. Application of 0,002-0.2nM CTX-L
increased the occurrence of spontaneous excitatory junction
potentials (SEJPs) and duration of the evoked excitatory
junction potential (EJP), the decay phase no longer being

RESPONSES OF VERTEBRATE NERVES TO
CIGUATOXIN, Memoirs of the Queensland Museum 34(3):
454, 1994:— Electrophysiological studies were performed on
a variety of nerve preparations from both mammals (rats and
humans) and fish, Responses of human peripheral nerves to
ingested ciguatoxin were assessed in the Sural nerves of 15
victims of ciguatera. In rats responses of the ventral coccygeal
nerve in anaesthetised animals were studied after intoxication
was induced by intraperitoneal injection of sub-lethal doses
of ciguatoxin. In fish, isolated segments of the spinal nerves
and the lateral line branch of the Vagus nerve of both ‘carmers'
and ‘non carriers’ of ciguatoxin were exposed to solutions of
ciguatoxin in fish Ringer,

In all nerve preparations there were significant changes in
arange of nerve conduction parameters including conduction
velocity, amplitude, and the duration of refractory periods and
the supernormal period. In all preparations there was a sig-
nificant prolongation of an increase in the magnitude of the
supermmormal period. These changes conform with studies on

MEMOIRS OF THE QUEENSLAND MUSEUM

fitted by a single exponential function. At 0.2nM CTX-] also
produced a large (25-30m¥V) maintained depolarization. The
effects of CTX-1 were abolished by tetrodotoxin (0.3.4.M) and
were calcium dependent. In addition EJPs and SEJPs were
blocked by the purinoceptor antagonist suramin (1mM) and
the maintained depolarization was blocked by the -adrenocep-
lor antagonist phentolamine (114M). These data suggest the
actions of CTX-1 are due solely to activation of the sym-
pathetic nerves innervating the rat tail artery.

James A. Brock, Medical Faculty, University of Newcastle,
NSW 2308, Australia, Phillip Jobling, Elspeth M. McLachlan,
Department of Physiology and Pharmacology, University of
Queensland, S| Lucia 4072, Ausfralia & Richard J. Lewis,
Southern Fisheries Centre, Department of Primary In-
dustries, Deception Bay, Qld 4508, Australia; 12 April, 1993.

isolated cells that suggest a fundamental action of ciguatoxin
on Na* gating mechanisms.

Both rat and fish preparations have been used to assess the
efficacy of a range of potential antagonists of the ciguafoxin
response. In rats, ciguatoxin-induced changes in supernor-
mality are unaffected by mannitol but significantly an-
tagonised by lignocaine, In fish, lignocaine and tetrodotoxin
antagonise the responses induced by ciguatoxin.

Ithas also been established that the nerves of fish respond
to ciguatoxin in a similar manner to those of mammals and it
is suggested that fish may have evolyed some degree of
protection against ciguatoxin by mechanisms that do not
involve the Na* channel.

Michael F. Capra, John Cameron, Andrew E, Flowers &
Christine E. Purcell, School of Life Science, Queensland
University of Technology, Brisbane 4000, Australia; 12 April,
1993.

NATURAL VERSUS ANTHROPOGENIC DISTURBANCES TO CORAL REEFS:
COMPARISON IN EPIDEMIOLOGICAL PATTERNS OF CIGUATERA.

RAYMOND BAGNIS

Bagnis, R. 1994.08 01: Natural versus anthropogenic disturbances to coral reefs: comparison
in epidemiological pattems of ciguatera. Memoirs of the Queensland Museum 34(3): 455-
460, Brisbane. ISSN 0079-8835.

Patterns of ciguatera fish poisoning vary from one location to another. In French Polynesia
surveys have shown that outbreaks of the disease are associated with disturbances to live
coral reefs. Anthropogenic damages such as undersea works, dumping of wastes, wreckage
of ships, crashing of ship anchors may result in a fare up of poisonings in areas with no
previous history of ciguatera; in this patlem, detritus-feeding herbivorous fish or invert-
ebrates and carnivorous fish may become toxic in a confined area over variable periods of
time, Natural catastrophes such as hurricanes, tsunamis, massive coral bleaching may be
associated with a pattern of diffuse continuous risk of ciguatera poisoning from large
predaceous fish, with periodic outbreaks inyolying fish from primary and secondary trophic
evels. Seasonal disturbances such as storms, heavy swells, high freshwater drainage, red
tides, seem to be consistent with a pattern of ciguatera poisoning in which the overall picture
is stable, with the same fish species, most of the time large predators, toxic in well defined,
exlensive areas.

Raymond Bagnis, Institut Territorial de Recherches Médicales Louis Malardé, BP 30
Papeete and Université Francaise du Pacifique. BP 4635, Papeete, Tahiti, Frenck

Palynesia; 2 February, 1994.

In French Polynesia most islanders depend
greatly on seafood; increases in toxicity of local
fish does net go unnoticed. For decades,
knowledge of ciguatera fish poisoning in most
areas was based on information of variable
reliability gathered from a diverse range of per-
sons who had been in contact with the disease.
These included medical staff treating it. persons
afflicted with it, fishermen and administrators.

Fish poisoning was listed among the notiftable
diseases to South Pacific Commission in 1974,
Although the reporting is for all types of fish
poisoning, it may be assumed that most cases deal
with ciguatera (the pathognomonic cold-to-hot
sensory reversal dysthesia makes it easier to dis-
tinguish from other types of fish poisoning).
Reports from French Polynesia are completed by
the Public Health Department staff (PHD) in
Papeete and mailed, each month, to the South
Pacific Epidemiological and Health Information
Service (SPEHIS). Reporting is compulsory for
each territorial Health Department Unit and
yoluntary for private medical practitioners. In
both cases, it relies on the person in charge to
provide accurate and timely monthly reports.
Delays in reporting cases are beyond the control
of both PHD and SPEHIS. Due to forwarding and
transmission of mail, an outbreak occurnng today
may be notified one or several months later. So
seasonal evaluation from rough official data is

questionable. The number of cases reported
depends on several factors. Severe cases that
occur near a health unit are most accurately
reported, whereas mild poisoning on the various
islands without medical or paramedical staff is
probably less accurate. Cases treated according
to traditional medicine may also not be notified.
Moreover, simplified official data do not provide
information on the name or species of the accused
fish, nor on the location of its capture,

Cases reported during 1974-1990 give an in-
dication of the magnitude of the ciguatera prob-
Jem in French Polynesia compared to other
Pacific countries reporting in the same way, Data
available at the PHD show that no island was
completely immune fram it, but morbidity is not
evenly distributed and the overall data are not
always significant for the region as a whole. For
instance, data recorded in Tuamotu Islands are
less accurate than those from the Gambier Is-
lands. The 12,500 inhabitants of the Tuamotus
live on 46 atolls separated by a few or by many
miles of ocean and are spread over 400,000
square miles; they are served by two mobile
medical practitioners, and have at their disposal
one hospital in the main atoll of Rangiroa and five
infirmaries located in Ansa, Fakarava, Hao,
Makemo, Reao, notwithstanding the military
medical staff in Hao and Moruroa. On the other
hand, the Gambiers consist of 10 small volcanic

456

islands in one 600 square mile lagoon; the 600
inhabitants of the archipelago are grouped in one
island, Mangareva, with an infirmary in the main
village.

To assess accurately the qualitative, quantita-
tive and time related aspects of evolution of
ciguatera in the French Polynesian islands, a
monitoring programme was developed from
1965 to 1990 by the author and the Malardé
Institute staff.

MATERIAL AND METHODS

From 1965, a daily follow-up of cases in the
whole territory was set up using the medical and
paramedical staff of the Public Health Units, with
the filling in of 15,000 standard questionnaires
consisting of 26 medical and epidemiological
parameters.

From 1967, nearly 20 tons of fish, caught in 30
different French Polynesian Islands exposed to
ciguateric risk, and belonging to 30 families from
various trophic levels, were bioassayed for
ciguatoxicity. From 1976, 10,000 grams of
macro-algae, among 40 species, were sampled
for research and counting of Gambierdiscus
toxicus cells.

Follow up of these various analyses, over 15—-
25 years, resulted in 3 indexes to monitor
ciguateric risk (Bagnis et al.,1985a): the number
of cases per 1,000 residents (CIR), percentage of
toxic individuals in a group of fish from a given
species, family or trophic level (PCI), density of
the populations of the toxic dinoflagellate G.
toxicus per gram of algae (GTD).

In addition to these research programs, the
author had many discussions on fish poisoning
problems with officials from fisheries and health
departments, old natives and longtime residents,
to assess what events, according to them, could
be associated with ciguatera.

RESULTS AND DISCUSSION

Within the context of the monitoring
programme, in each inhabited atoll or large val-
ley, a person was officially appointed to fill in the
questionnaires and, from 1965 to 1990, the
ciguatera reporting system in French Polynesia
has gradually become one of the most com-
prehensive in all of the Pacific area. It has been
estimated that 70-80% of cases are reported (in-
stead of 10-20% registered by the PHD in
Papeete, from 1960 to 1965). During that time,
about 30,000 cases of ciguatera were recorded

MEMOIRS OF THE QUEENSLAND MUSEUM

(CIR: 8 °/oo), caused by some 100 fish species
from various trophic levels. Gambier Islands
(11% of cases; CIR: 60 °/o0), Marquesas Islands
(19% of cases; CIR: 30 Yoo), and Tuamotu Islands
(17% of cases; CIR: 25 °/oo) were the most in-
volved, compared to Society Islands (52% of
cases; CIR: 5 oo) and Austral Islands(<1% of
cases; CIR °/oo). As patterns of ciguatoxicity may
vary in time and space, from one island to another
and from place to place in the same island, only
data provided by the questionnaires give an ac-
count of the situation in each separate island.
Thus, in the Tuamotu Islands, the average overall
CIR was influenced by several flare ups that
occurred in Hao (primarily) but also in Hikueru,
Moruroa, Takaroa, Takapoto, Manihi, Reao,
Fakarava, Mataiva, Anaa, Makemo from 1964—
1975. In the Gambier Islands, from 1971 to 1980,
the CIR never decreased below 300, reaching 560
at its highest point in 1975. In the Marquesas
Islands, the CIR grew from about 10 in 1963 to
80 in 1971, and started decreasing gradually in
1976. In the Society Islands an outbreak occurred
from 1964-1970 in Bora Bora, while at the same
period a significant increase of the ciguatera
poisonings by fishes caught in the area of Faaa,
was reported in Tahiti, with a high surge in 1966—
1967. In the Austral Islands, the least affected of
the Polynesian archipelagoes, a few cases are
episodically reported from each island but Rapa.
In French Polynesia, there are also islands where
for decades the ciguateric risk seems to be limited
to some species of fish and to some well defined
areas, with a level of ciguatoxicity apparently
unchanged (Bagnis et al.,1985a).

The most significant data about the PCI con-
cern Tahiti, Hao, Gambier and Marquesas Is-
lands. In Tahiti, the study was carried out from
1967 (at the peak of the surge) to 1984, on the
most frequently poisoned fish, the surgeon fish
maito, Ctenochaetus striatus, and showed that the
average PCI was cut by half during that period.
In Hao, the PCI increased from c.0 (in July 1966)
to c.70%( within 2 years) before decreasing from
1970 to reach 7.5% in 1982. In Gambier Islands,
from 1969 to 1971, PCI increased five-fold to a
maximum of 67% during 1975-1978, for
detritus-feeding herbivorous fish and tenfold to a
maximum of 80% during 1978-1979 for car-
nivorous fish, before decreasing gradually. The
data related to Marquesas Islands indicate, from
1968-1984, a regular increase of PCI for her-
bivorous and carnivorous fish, from respectively
4.5 to 26% and 5.3 to 22% (Bagnis et al.,1985a).

A close relationship between PCI and GTD

CORAL REEF DISTURBANCES AND CIGUATERA PATTERNS

could be observed in the Gambier Islands (Bagnis
et al.,1985a) and on Hitiaa Reef in Tahiti (Bagnis
et al., 1985b). In the Gambiers, where data were
the most significant, the yearly average of GTD
decreased gradually from 45,000 in 1977 (a year
that fits roughly with the highest values of the
PCI) to 40,000 in 1978, 10,000 in 1978-1980,
1,500 in 1982, <150 in 1984 to c.100 in 1989-
1990. Elsewhere, the follow up was not accurate
enough to allow comparative evaluation between
the two indexes (PCI and GTD), within the same
time span.

Previous data and several events associated
with ciguatera, either from the monitoring of
some flare-ups, or noted empirically by natives,
such as various well identified anthropogenic and
natural disturbances in the coral reef ecosystem
(Bagnis,1987), support the hypothesis that the
appearances of poisonous fish are directly linked
to changes in some factors of their environment
(Randall,1958). Helfrich & Banner (1968) sug-
gested that the patterns of ciguatera in the Pacific
fall into one of three main categories or oc-
casionally a combination of these. These patterns
are discussed below with reference to ciguatera
in French Polynesia.

EPIDEMIC PATTERN

This pattern of poisoning concerns areas with
no or a rare history of ciguatera. In such areas,
after the initial outbreak, there was a general
increase in toxicity reaching a peak of severity
(based on the number of species affected and the
level of the PCD) within 2-5 years. After some
time, the toxicity began to decline. In the islands
of French Polynesia where such flare-ups have
been observed, this occurred c.5—10 years after
the onset of toxicity. The decline may start
sooner, but its exact timing is often obscured by
i) the fact that the local population is reluctant to
begin eating a fish after most of its members were
stricken in the period following the initial out-
break (several months or years) and ii) by fre-
quent clinical hypersensitization features which
can be mistaken for true poisonings. The rate of
decline in ciguatoxicity varies also according to
areas. Grazing or browsing herbivores and
detritus feeders, such as some acanthurids (C.
striatus or C. strigosus), scarids (Scarus gibbus
or Scarops rubroviolaceus) or mugilids
(Crenimugil crenilabis), become toxic first, fol-
lowed within a few months by the carnivores at
higher trophic levels (serranids, lethrinids, lut-
janids, carangids, labrids, muraenids). Thus the
various links of the food web are progressively

457

involved. At the peak of the flare up, an island
that had not harboured any toxic fish in man’s
memory, may have all fish contaminated within
a year, according to the natives. In the declining
toxicity phase, herbivores become significantly
less toxic first, followed by some species at
higher trophic levels, until only a few of the large
carmivores remain toxic.

Most flare-ups in the Tuamotu Islands, since
investigations started, proceed with this pattern
(Bagnis,1969,1982; Bagnis et al.,1973,1985a).
Similar features have been observed in Society
Islands, Austral Islands and Gambier Islands
(Bagnis et al., 1988).

All these flare ups followed various anthropo-
genic damage to coral reefs, linked to public or
military works (Bagnis,1987). The main distur-
bances were: tearing, blasting, disrupting, scour-
ing of pieces of coral reef (to deepen or widen
passes in the barrier reef, to open channels in
lagoons, to drill shafts in the basaltic layer for
nuclear tests), crashing of heavy ship anchors and
ploughing caused by dragging anchor cables,
dredging and shifting of sand, dumping of wastes
or debris (chiefly metallic ones), wreckage of
ship, building of piers, wharfs, roads, protective
sea-walls on live coral reefs and any other
damage less evident and more traditional, like
massive diving for pearl-oysters, or usual surf-
landing of whaleboats on the same outer reefs of
atolls without a pass. Usually, human damages
induce geographically limited flare ups (Bagnis,
1969,1981; Bagnis et al.,1973). In the Pacific
area, outbreaks of fish poisoning during the
Second World War were very likely related to
man-made damages on coral reef (Halstead,
1967).

ENDEMO-EPIDEMIC PATTERN

This pattern of poisoning does concern areas
with a history of continuous ciguateric risk for the
local population, made up of periods of stable,
relatively low fish toxicity, referred to as the
“quiescent stages’ by Cooper (1964), alternating
with periods of major outbreaks, usually only
within the memory of the older inhabitants.

In this pattern, toxicity remains confined,
during long periods, to large carnivorous fish:
some snappers, groupers, jacks, emperor fish and
moray eels. Periodically, such areas experience
flare ups similar in form and duration to that
described for the Epidemic Pattern. Thus, the
detritus feeding herbivores (like some surgeon-
fish, parrotfish, mullets) become toxic, followed
by some carnivores at higher trophic levels,

i58

which were previously safe. Finally, all the food
chain may be more or less affected by cigua-
toxicity during a period of 5-20 years tn extensive
parts of the coast (Bagnis, 1974). This pattern was
first observed in Marquesas Islands where data
obtained by the author (1965-1973) from inquiry
with old natives, pointed out 4 flare ups since the
beginning of the century. After multiple cross-
checking, their dates and durations could be
Toughly established: 1905-1915, 1925-1935,
1953-1958, 1965-1985. The first broke out, ac-
cording to the people interviewed, after the pas-
sage of some cyclones. Such an assumption
prompted the author to look for a chronological
relationship between the alleged flare ups and
cataclysmic events, in these islands not protectest
by a barmer reef and occasionally exposed to
tsunamis. Informations collected from the
Geophysical Laboratory of Tahiti, the French
Meteorological Service and the Hydrographic
Department of the French Navy about tsunamis,
Strong stomms, hurricanes which have affected
Marquesas Islands since the beginning of the
century, pointed out a close relationship between
both sets of events (Bagnis,]980}. Some data
from Majuro in the Southem Marshall Islands
(Bartsch & McFarren,1962), in certain Kinbati
Islands (Cooper,1964), could also illustrate this
pattern.

The episodic resurgence of toxicity involving
detritus- feeding herbivorous fish, would indicate
that G. toxics (or another ciguatoxic microor-
fanism) was recently (or is still) proliferating
again and actively manufacturing the ciguatoxin
or a precursor of it, Very likely amongst the
widespread macro-algal turf covering the many
coral colonies damaged hy the passage nf a
tsunami or a cyclone. This new toxin production
would increase the overall ciguatoxicity of the
food chain for years.

The long quiescent periods, during which only
afew large carnivores ure toxic, en explained
by the longevity of the fish themselves and their
ability to retain toxin. The available data about
longevity of tropical reef fish, indicate that some
may live more than 30 years (moray eels). If the
ciguatoxin were restricted to a ‘pool’ in the
ecosystem bound in the organisms at trophic
levels above the secondary one, and the reduction
of toxin in the pool would be only by means of
natural mortality and break-down by reducer or-
ganisms, assimilation by other consumer or-
ganisms would merely retain and recirculate it
within the pool. As the overall level of the
ciguatoxin in the pool slowly declines through

MEMOIRS OF THE QUEENSLAND MUSEUM

loss of toxic individuals Dy natural mortality,
reduction by microorganisms and possibly by
slow natural excretion, only those animals with
very great longevity would be expected to he
toxic, and to contain a large quantity of
ciguatoxin (Helfrich & Banner,1968).

Another example of potential recycling of the
ciguatoxin in the pool of carnivores in a restricted
area is frequently observed in Marquesas Islands
(and elsewhere), Many lutjanids, serranids and
Icthrinids, considered as toxic in some areas, are
thrown back inte the water when caught. Such a
practice would essentially act as a feedback
mechanism in the ecosystem. In this condition,
one can think that, rather than a gradual ac-
cumulation of the toxin at the higher trophic
levels by the carnivores with the greatest lon-
gevity, as has been noted previously, the activity
of the fishermen may continually recycle the
toxin among the carnivores, favouring those with
non-specialized food habits, such as the
scavengers (some lethrinids for instance).

Another explanation of a resurgence in the
ciguatoxicity (in both the level and the species
involved) of reef fish dunng the ‘quiescent
stages’ lies also in some anthropogenic disturban-
ces, Thus, extensive blasting and dredging may
release some of the ciguatoxin concentrated in the
pool of sedentary eels, swrasses and dis-
tribute it to individual usually safe camivorous
fishes at various trophic levels in the community.
Such events may result in a senes of scattered
cases of ciguatera incriminating fish other than
detritus-feeding herbivorous fishes,

ENDEMIC PATTERN

This pattern occurs in areas in which the overall
picture of toxicity seems to be quite stable, with
the same species (most of the time predators)
exhibiting the same Jevel of toxicity within well-
defined geographical areas, The condition is said
to be unchanged as far as the local people can
remember, Confirmation of the continued
toxicity of the fish usually occurs when an out-
sider, such as a tourist or a member of a ship’s
crew, is poisoned or when a native who cannot
resist the temptation offered by a meal of a fat,
succulent grouper, snapper or eel, gamble on il
being nontoxic and lose. According to nid ts-
landers, approximately the sanve species (most of
the time carnivorous, but also detritus-feeding
herbivorous fish) that Were toxic in some reefs of
Society, Marquesas, Tuamotu, Austral and Gam-
bier Islands 50 years ago, may still be toxic today-
The same stability is observed outside of French

CORAL REEF DISTURBANCES AND CIGUATERA PATTERNS

Polynesia in some parts of New Caledonia,
Vanuatu, Fiji, Samoa, Tonga, Guam, Marshall Is-
lands, (for the Pacific), Reunion, Mauritius (in
Indian Ocean) and in most Caribbean Islands.

Seasonal natural disturbances like heavy
swells, abundant rains with freshwater drainage
and soil runoffs, red tides, slight coral bleaching,
crown of thorns proliferation, and any other kind
of insidious disturbance, not taken into considera-
tion by man, could be the most frequent causes of
this pattern of ciguatera poisoning, with recycling
of the toxin by the fishermen. Every disturbance
destroying coral colonies and creating conse-
quently new surfaces available for macro-algal
colonization results either in a significant in-
crease of genetically toxic strains of G. toxicus or
in the development of toxicity by previously non
toxic strains, because of some changes, very like-
ly due to the microflora associated with G.
taxicus,

From today's knowledge on the origin of
ciguatera, natural disturbances would explain its
antiquity and its occurence in many uninhabited
islands and shoals distant from land. Through
information provided by old islanders, one may
safely say that no island of French Polynesia, save
perhaps Rapa, is completely immune from it.

CONCLUSIONS

To understand the dynamics of the evolutive
patterns of ciguatera, one should remember that
the reef community is probably the most complex
in the sea. A delicate balance must exist among
competing organisms, so much so that a subtle
change in environment can result in temporary

liferation of one at the expense of the others.

‘hatever the disturbances in lagoon, pass, fing-
ing or barrier reef, the result is the same: occur-
rence of dead coral beds and new surfaces
available for opportunist populations. We know
that toxic G. toxicusis very poor in reef, innonnal
environmental conditions; but it may increase
rupidly and significantly among the filamentous
or calcareous algae growing on new or denuded
surfaces, in normal ecological succession, fixin
itself on the thalh of the algae, primarily the re
ones. If we postulate G. toxicus as the single or
main source of ciguatoxin, with herbivore inges-
tion as the chicf means of transfer of the toxin to
higher trophic Jevels, then the more widespread
the macroalgal colonization, the more the toxic
dinoflagellate bloom can be important and the
higher the amount of toxin introduced and stored
in the food-web. The detritus-feeding her-

459

bivorous fishes can be expected to remain
poisonous, after the ciguatoxin producing
dinoflagellates decrease in number, disappear or
cease fo be toxic, by a period of time equal to the
maximum longevity of the fish species involved.

Discussion on patterns of ciguatera develop-
ment is based on empirical data, from French
Polynesia, It is admittedly tenuous and conjec-
tural in some cases. Nevertheless. it supports
Randall's (1958) hypothesis, resumed by
Helfrich & Banner (1968), on the evolution of
ciguatera-

The epidemic pattern seems to be most of the
time associated with anthropogenic damuge
(Bagnis,1969.1982; Bagnis et al.,1973). The en-
demo-epidemic pattem may be related to natural
catastrophes, as in the Marquesas Islands with
their fringing patches of coral without the protec-
tion of barrier reefs amd many shoals away from
the shore, especially exposed to variations of
oceanic hydrodynamism (Bagnis, 1980). The en-
demic pattern may be associated with natural
seasonal dlisiurbances in most of volcanic islands
with fringing or barrier reef where the detrimental
action of man or cataclysms in the marine cn-
vironment are negligible (as observed in many
pans of French Polynesia). The combination of
the three pattems is observed in Gambier, where
ull the islands are edged with a fringing reef and
hounded with the same continuous half emerged
168 miles long barrier reef: the latter is bordering
arelatively shallow lagoon, with a bottom rugged
with many patch reefs, knoll reefs and coral
ndges. In this archipelago, a close relationship
between ciguateric fishes and anthropogenic and
natural damage 19 reefs was pointed out (Bagnis
et al.,1 988).

ACKNOWLEDGEMENT

The author wishes to thank Gérard and Laetitia
Nardou for their attemptat proof-reading the final
draft.

LITERATURE CITED

BAGNIS, R. 1969, Naissance et développement d'une
Nambée de ciguatera dans un atoll de l’archipe|
des Tuamotu. Revue des Corps de Santé des
Armmées 10: 115-127,

BAGNIS, R,, THEVENIN, S. & BENNETT, J. 1973.
Pollution marine et ciguatera dans l‘atoll de
Manihi (Tuamotu). Actes du Séminaire de la
Commission du Pacifixque Sud sur la polluuon des
lagons. (Guam, 15 May, 1973).

BAGNIS, R- 1974. Evolution d'une flambée dc

ciguatera aux Iles Marquises. Médecine et Armées
2: 115-122.

BAGNIS, R. 1980. Agressions sur les édifices coral-
liens des Iles Marquises et ciguatera. Médecine
Océanienne 12: 42-50.

BAGNIS, R. 1981. L’ichtyosarcotoxisme de type
ciguatera: phénoméne complexe de biologie
marine et humaine. Oceanologié Acta 4: 375-387.

BAGNIS, R. 1982. La ciguatera dans les atolls des
Tuamotu. Médecine Océanienne 17:1-9.

BAGNIS, R., BENNETT, J., BARSINAS, M.,
CHEBRET, M., JACQUET, G., LECHAT, I.
MITERMITE, Y., PEROLAT, Ph. &
RONGERAS, S. 1985a. Epidemiolgy of ciguatera
in French Polynesia from 1960 to 1984. Pp, 475-
482. In C. Gabrie & B. Salvat (eds), ‘Proceedings
of the 5th International Coral Reef Congress,
Tahiti, vol.4’, (Antenne Museum-Ephe: Moorea).

BAGNIS, R., BENNETT, J., PRIEUR, C. &
LEGRAND, A.M. 1985b. The dynamics of three
toxic benthic dinoflagellates and the toxicity of a
ciguateric surgeonfish in French Polynesia. Pp.
177-182. In D.M. Anderson, A.W. White & D.G.
Baden (eds), “Toxic dinoflagellates’. (Elsevier:
Oxford).

BAGNIS, R. 1987. Ciguateric fish poisoning: an objec-
tive witness of the coral reef stress. Pp. 241-253.

MEMOIRS OF THE QUEENSLAND MUSEUM

In B. Salvat (ed.), ‘Human impacts on coral reefs:
facts and recommendations 18”

BAGNIS, R., BENNETT, J., BARSINAS, M., DROL-
LET, J.H., JACQUET, G., LEGRAND, A.M.,
CRUCHET, Ph. & PASCAL, H. 1988. Correla-
tion between ciguateric fish and damage to reefs
in Gambier Islands. Pp. 195-200. In H. Choat et
al. (eds), ‘Proceedings of the 6th International
Coral Reef Symposium, Townsville, vol.3’. (6th
International Coral Reef Symposium Executive:
Townsville).

BARTSCH, A.F. & MCFARREN, E.F. 1962. Fish
poisoning: a problem in food toxication. Pacific
Science 16: 42-56.

COOPER, M.J. 1964. Ciguatera and other marine
poisoning in the Gilbert Islands. Pacific Science
18: 411-440,

HALSTEAD, B.W. 1967. ‘Poisonous and venomous
animals of the world, vol.2’. (US Government
Printing Office: Washington D.C.).

HELFRICH, Ph & BANNER, A.H. 1968. Ciguatera
fish poisoning, 2, General patterns of development
in the Pacific. Occasional papers of the Bernice P.
Bishop Museum 23: 371-382.

RANDALL, J.E. 1958. A review of ciguatera tropical
fish poisoning, with a tentative explanation of its
cause. Bulletin of Marine Science of the Gulf and
Caribbean 8: 236-267.

GAMBIERTOXIN-INDUCED MODIFICATIONS OF THE MEMBRANE POTENTIAL

OF MYELINATED NERVE FIBRES
EVELYNE BENOIT AND ANNE-MARIE LEGRAND

Benoit, E, & Legrand, A.-M.. 1994 0% 01. Gambiertoxin-induced modifications of the
membrane potential of myelinated nerve fibres. Memioirs of the Queensland Museum
34(3):461-464. Brisbane. ISSN 0079-8835.

The effects of external application of 1.2-24nM of gambiertoxin (CTX- 4B), extracted from
the dinoflagellate Gambierdiscus loxicus, were studied on the membrane potential of
myelinated nerve fibres isolated from the frog, At concentrations of 12 and 24nM, CTX-4B
induced spontaneous action potential discharge at a frequency of 30-100Hz. In the presence
of 24nM of CTX-4B, the amplitude and duration of these spontaneous action potentials were
respectively decreased and increased compared {a control action potentials, Toxin-induced
spontaneous action potentials were suppressed by increasing the external calcium concentra-
tion or by lidocaine, It is concluded that the action of CTX-4B on membrane potential, in
some respects, resembles that of moray-eel ciguatoxin previously reported (Benoit et al.,
1986).

Evelyne Benoit, Laboratoire de Physiologie Cellulaire, URA CNRS 1121, bat, 443,
Université Paris-Sud, 91405 Orsay Cedex, France; Anne-Marie Legrand, U.R.
d'Océanographie Médicale, Institul Territorial de Recherches Médicales Louis Malardé,

BP 30, Papeete, Tahiti, Polynésie Francaise; 2 February, 1994.

Ciguatera toxins are responsible for the most
common human food poisoning associated with
the consumption of various tropical and subtropi-
cal fishes. The major source of these toxic com-
pounds is the dinoflagellate, Gambierdiscus
toxicus (Adachi & Fukuyo,1979).

Studies with ciguatoxin extracted from the
moray-eel concluded that the toxin specifically
interacts with Na channels of yarious prepara-
tions. In particular, in the frog node of Ranvier,
the toxin induces spontaneous action potentials
due to specific modifications of Na channels
(Benoit et al.,1986).

In this work, we have investigated effects of the
main toxic compound extracted from G. foxicus,
pambiertoxin (CTX-4B), on the membrane
potential of the frog myelinated nerve fibres.

MATERIALS AND METHODS

The experiments were carried out on single
myelinated nerve fibres isolated from the sciatic
nerve of the frog Rana esculenta. The membrane
potential was recorded under current clamp con-
ditions as previously described (Benoit ct
4l.,1986). The Ringer’s solution had the follow-
ing composition (mM) : NaCl 111.5; KCI 2.5;
CaClz 1,8, HEPES 10; pH 7,4. The fibre ends
were cut in a solution containing 120mM KCL.
The temperature was 15-16°C.

Gambiertoxin (CTX-4B) was extracted from
Gambierdiscus toxicus and purified. Samples

were kept al -1B°C and diluted immediately
preceding experiments in Ringer's solution to
give the final toxin concentrations indicated,

RESULTS

EFFECTS OF CTX-4B ON MEMBRANE
POTENTIAL

CTX-4B at concentrations 1.2-6nM, applied
for 12-22min, had no effect on the membrane
potential, even when K currents were suppressed
by replacing the solution bathing the cut fibre
ends, 1.e., the interna) solution, with | 10mM CsCl
+10mM NaCl and by adding tetraethylam-
monium to make an external solution with a final
concentration of 1OmM. This was consistently
observed on six different fibres, whether they
were sensory or motor fibres.

In contrast, 2-Smin after the addition of 12 or
24nM of CTX-4B to the Ringer’s solution, spon-
laneous action potentials appeared at a frequency
of 30-1 00Hz (Fig.1A). The toxin-induced spon-
taneous action potentials were often separated by
silent periods and were less regular in amplitude
and frequency than the spontaneous action poten-
tials induced by moray-eel ciguatoxin (Benoit et
al.,1986). Moreover, the resting membrane
potential of fibres was apparently not modified by
CTX-~4B and maintained depolarizations to near
-20mV were never observed, in contrast to the
previously reported effects of moray-eel
ciguatoxin (Benoit et al.,1986).

462

MEMOIRS OF THE QUEENSLAND MUSEUM

50 msec

FIG.1. Control action potentials recorded in Ringer’s solution evoked by 0,.5msec depolarizing stimuli (left
traces). Spontaneous action potentials recorded about 10min (middle traces) and about 12min (nght traces) after
bathing the nerve fibre in Ringer’s solutions containing 12nM (middle traces) or 24nM (right traces) of CTX-4B.
Note the difference in horizontal scales between (A) and (B).

In the presence of 12nM of CTX-4B, the
amplitude of the spontaneous action potentials
was only slightly reduced to 8646% (n=3) of
control on average and their duration, measured
at the 50% level repolarization, was not sig-
nificantly modified (mean 1.05+0.08%, n=3)
compared to the control action potential elicited
by a depolarizing stimulus (Fig.1B). This was not
the case in the presence of 24nM of toxin, where
the amplitude and duration of spontaneous action
potentials were respectively reduced (mean
52+3%, n=4) and increased (mean 1.78+).06%,
n=4) compared to the control action potential
(Fig. 1B). Finally, it should be noted that the spon-
taneous action potentials induced by 12 or 24nM
of CTX-4B (applied for 8-42min) were not sup-
pressed after 18-36min wash out of toxin with
control solution.

EFFECTS OF TETRAETHYLAMMONIUM ON
SPONTANEOUS ACTION POTENTIALS

Under control conditions, after the addition of
tetraethylammonium to make the Ringer’s solu-

tion with a final concentration of 10mM, the
duration of elicited action potentials was
1.8340.05 fold (n=3) greater than the duration of
control action potential. Similar results were ob-
tained when tetraethylammonium was added to a
solution containing 12nM of CTX-4B, i.e., spon-
taneous action potentials were prolonged. In con-
trast, the subsequent addition of tetraethyl-
ammonium to the solution containing 24nM of
CTX-4B or the addition of 24nM of toxin to the
solution containing tetraethylammonium, did not
noticeably further modify the duration of action
potentials. As tetraethylammonium is well
known to specifically suppress K current, these
results strongly suggest that the nodal K current
was not significantly affected by 12nM of CTX-
4B, whereas it was reduced in the presence of
24nM of toxin-

EFFECTS OF INCREASING EXTERNAL CALCIUM
CONCENTRATION ON SPONTANEOUS ACTION
POTENTIALS

The effects of increasing the external con-

GAMBIERTOXIN EFFECTS ON MYELINATED NERVE FIBRES

24 nM CTX-4B_ A
-— 1.8 mM calcium —— t+— 5.4 mM caicium —— +— 1.8 mM caicium ——

463

50 msec

25 mV

FIG. 2. Spontaneous action potentials recorded in a 24nM solution of CTX-4B made up in Ringer’s solution
containing successively 1.8mM (left trace), 5.4mM (middle trace) and 1.8mM of calcium (right trace).

centration of calcium were studied on spon-
taneous action potentials recorded in the presence
of 12nM or 24nM of CTX-4B, or during wash out
of CTX-4B with a Ringer’s solution. Increasing
the external calcium concentration from 1.8 to
5.4mM suppressed spontaneous action potentials
in <30sec (Fig.2). In the presence of 5.4mM of
calcium, an action potential could be elicited by
a depolarizing stimulus. The effects of increasing
external calcium concentration were reversed
within 1—2min by superfusing the fibre with solu-
tions containing 1.8mM of calcium (Fig.2).

WU,

24 nM CTX-4B ——_—————oH«u
+— 50 uM lidocaine ——

EFFECTS OF LIDOCAINE ON SPONTANEOUS
ACTION POTENTIALS

Addition of 50uM of lidocaine to the Ringer’s
solution containing 12nM or 24nM of CTX-4B,
inhibited spontaneous action potentials in less
than 30sec (Fig.3). However, under such condi-
tions, as previously described in the presence of
5.4mM of calcium, an action potential could be
observed when the myelinated nerve fibre was
stimulated by a depolarizing current. It should be
noted that in the presence of CTX-4B, the fre-
quency of appearance of spontaneous action
potentials was 30-100Hz (Fig.1A) whereas the
depolarization-induced action potentials were

WU

FIG. 3. Spontaneous action potentials recorded in a 24nM solution of CTX-4B made up in a Ringer’s solution,
in the absence (left trace), presence (middle trace) and after wash out (right trace) of 50M of lidocaine.

usually elicited at a frequency of 1Hz. The in-
hibitary action of lidocaine has been shown to be
more effective when the frequency of events is
increased (Hille,1977). This may explain in part
why lidocaine was more effective at blocking
spontaneous action potentials than the elicited
action potentials. The effects of lidocaine were
reversed by a 2—3min wash with a Ringer's solu-
tion containing CTX-4B (Fig.3).

DISCUSSION

These results show that CTX-4B induces spon-
taneous action potentials in frog sciatic nerves
which are suppressed by increasing the external
concentration of calcium or by lidocaine.

Appearance of spontaneous action potentials
was also observed in the presence of moray-eel
ciguatoxin (Benoit et al.,1986). However,
ciguatoxin was active at concentrations as low as
0.22nM, whereas in the present experiments con-
centration of at least 12nM of CTX-4B was
needed to induce spontaneous action potentials.
Thus, CTX-4B appears to be about 50 fold less
effective than ciguatoxin on the membrane poten-
tial of myelinated nerve fibres. Finally, the effects
of ciguatoxin were completely reversed upon

MEMOIRS OF THE QUEENSLAND MUSEUM

washing out of toxin with a Ringer’s solution, in
contrast to those induced by CTX-4B.

It is concluded that, in some respects, the action
of CTX-4B on membrane potential of myelinated
nerve fibre resembles that of moray-eel
ciguatoxin.

ACKNOWLEDGEMENTS

This work was supported by a grant from In-
stitut National de la Santé et de la Recherche
Médicale (CRE 91 0906).

LITERATURE CITED

ADACHI, R. & FUKUYO, Y. 1979. The thecal struc-
ture of a marine toxic dinoflagellate Gambierdis-
cus toxicus gen, and sp. novo collected in a
ciguatera endemic area. Bulletin of the Japanese
Society of Scientific Fisheries 45: 67-71.

BENOIT, E., LEGRAND, A.M. & DUBOIS, J.M.
1986. Effects of ciguatoxin on current and voltage
clamped frog myelinated nerve fibre. Toxicon 24:
357-364.

HILLE, B. 1977. Local anaesthetics: hydrophilic and
hydrophobic pathways for the drug-receptor reac-
tion. Journal of General Physiology 69: 497-515.

MANNITOL THERAPY FOR ACUTE AND CHRONIC CIGUATERA FISH POISONING

DONNA G. BLYTHE. LORA E. FLEMING, D. RAM AYYAR, DON DE SYLVA,
DANIEL BADEN AND KATHY SCHRANK

Blythe, D.G., Fleming, L.E., Ayyar, D.R., deSylva, D., Baden, D. & Schrank, K. 1994 08
QO); Mannitol therapy for acute and chronic ciguatera fish poisoning. Memoirs of the
Queensland Museum 34(3): 465-470. Brisbane. ISSN 0079-8835.

Ancevaluation of Intravenous (1V) mannitol therapy for treatment of the marine toxin disease,
Ciguatera Poisoning, in 107 persons from the South Florida/Caribbean area, 70 patients with
ciguatera poisoning received IV mannito] treatment (1 g/kg) within hours to 1000 days from
exposure, and 37 patients with ciguatera poisoning received only supportive therapy, if any.
The treated and non-treated groups were comparable, excep! for prolonged time until
presentation of the untreated group. 29 out of 32 (91%) patients treated with mannitol within
the first 48 hours from exposure had complete reversal of symptoms. Although not a formal
randomized clinical trial, this case series does provide yaluable information and support for
the use of intravenous mannitol in the treatment of acute and chronic ciguatera poisoning.

Denna G. Blythe, University of Miami School of Medicine, Miami, Florida; Lara E. Fleming
and D. Ram Ayyar, University of Miami Schoel of Medicine, Miami, Flarida; Don de Sylva
and Daniel Baden, Rosenstiel School ef Marine and Atmospheric Sciences, University of
Miami, Miami, Florida; Kathy Schrank, University of Miami School of Medicine, Miami,

Florida; 18 June 1994,

Dinoflagellates in the marine genus Gambier-
discus elaborate a number of toxins which are
bioconcentrated in the food chain through reef-
feeding herbivores to larger predatory fish. When
these larger species are eaten by humans. they
lead to Ciguatera Poisoning, Ciguatoxin (CTX)
is responsible for the majority of human illness
associated with ciguatera poisoning (Carmichael
et al.,1986; ILO,1984).

Ciguatoxin is a lipid soluble, heat stable and
acid resistant neurotoxin (Carmichael et al.,1986,
Sakamoto et al.,1987). It causes no adverse ef-
fects to the fish, and cannot be detected by dif-
ferences in smell or taste, nor is it eliminated by
cooking, freezing or other preparation procedures
(Lewis,1986). The mechanism of action of
ciguatoxin is as a sodiom channel toxin (Lewts,
1986; Baden et al., 1990).

In the past, a variety of bioassays {including
feeding and injections into cats. mongoose and
mice) have been used to test for CTX in fish.
Intraperitoneal injection in mice has been one of
the most widely accepted bioassays, and more
recently, rat brain synaptosome (Lange,1987;
TLO,1984). In addition, to being impractical for
routine use in the fish industry, there has been no
test available for the evaluation of human
ciguatera in clinical practice. Several new tests
have been developed. One of these is a radioim-
munoassay for ciguatoxin, aso-called ‘stick test’,
which can be used to test for ciguatoxin in fish
and has been widely used in Hawaii (Hokama,

1985). A highly sensitive ELISA test for assay in
human biologic fluids is currently being trialled
(Trainer & Baden,1990). Until these assays are
established in human populations, diagnosis of
ciguatera can only be arrived at clinically.

In the United States, nearly half of the reported
foodborn disease outbreaks of chemical origin
are due to marine toxins, with CTX causing al
least one third of these outbreaks (Lange.1987).
Ninety percent (90%) of the reported cases of
cigualera poisoning come from Florida and
Hawaii (Lange, !987). In Miami, an average an-
nual incidence of at least 5 cases/10,000 persons
was eslimated by reports to the Public Health
Department and based on clinical diagnosis
(Lawrence et al.,!980). In certain islands of the
South Pacific up to 43% of the population his
experienced ut least one episode of Ciguatera
(Rodgers & Muench,1986) and in Puerto Rico,
up to 7% of the residents (Holt et al., 1984).

The human disease entity of ‘Ciguatera Potson-
ing’ isa direct result of the stimulation of adrener~
gic and cholinergic nervous system due to the
opening of the sodium-dependent channels by
ciguatoxin (ILO,1984; Lange,1987). It presents
as an acute syndrome characterized by a variety
of gastrointestinal, neurologic and cardiovascular
symptoms within a few hours of contaminated
fish ingestion, Most commonly, patients exper-
jence acute nausea, vomiting, diarrhea, gastro-
intestinal cramping, paresthesias, and brady-
cardia. Fatality, usually due to respiratory

366

MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 1. sas 3a characteristics in mannitol treatment sey

107 39y

No Rx 37

failure, cardiac arrhythmias and possibly cerebral
edema, is reported to range from 0.1-12% of
cases (Lange,1987; Morris et al.,1982; Bagnis et
al., 1979). In addition to the acute illness, the
chronic symptoms of ciguatera poisoning, espe-
cially the paresthesia, can persist in varying
severity for months to yeurs after the acute illness,
with significant long term disability as a result.
Chronic effects of ciguatera poisoning have been
largely ignored tn the literature, probably due to
inaccurate diagnosis by mexpenenced healthcare
workers and lack of available human diagnostic
testing (Lange,1987; Blythe & deSylva,1992).
A yariety of treatment modalities have been
tried for intervention in ciguatera poisoning.
These include antihistamines, corticosteroids,
calcium supplements. amitriptyline, fluoxetine,
and lidocaine derivatives (Lange,1987; Berlin e1
al.,1992; Pearn et aJ.,1989; Gillespie et al., 1986).
None of these therapies have withstood the test of
ume. 23 cases of clinically diagnosed acute cig-
uatera poisoning in the Marshal Islands were
treated with an intravenous infusion of 20% man-
nitol (lg/kg at a rate of 500ce/hr) over 30 mins
‘piggy backed’ on an intravenous infusion at 30
cc/hr of either 5% dextrose in Ringer's or saline
solution; there was complete resolution of
symptoms within 48 hours in 17/23 patients
(Palafox etal. 1988). Pearn etal. (1989) published
a case series of |2 patients treated with [V man-
nitol (0.5-Ig/ke over 30mins); there were
dramatic results in the 5 acutely il] patients. They
postulated that mannito] might reduce axonal
edema and/or act as a scavenger of hydroxy!
radicals located on the ciguatoxin molecule.

We present 2 case series of 107 subjects with
clinically diagnosed ciguatera poisoning from
south Florida and the Caribbean collected since
1985, Seventy of these were treated with TV man-
nitol and 37 were not ireated because they either

| Sesty_| Member Ae (enn Sex Siete [ties |__|. ee

<a

aa 27% grouper
76% neura 25king —
9296 ne uro 27 Sogroupe

16%GI
67% neuro

16%G1 63%grouper

(0. +70) 42% neuro L1%barr acuda
12.5d 16%GI 234-king
0.3-365) | — 76%neuro__—| 14 per |

presented prior to the Palafox publication, or
mannitol was not offered or was declined,

METHODS

Patients of all ages and both sexes were diag-
nosed clinically with ciguatera poisoning if they
gave a history of a) consuming reef fish from
South Florida or the Canbbean and b) the onset
of gastrointestinal symptoms within 6—24 hours
of consumption and c) when relevant, subsequent
onset of neurologic symptoms, usually after 24—
48 hours from consumption. The gastrointestinal
symptoms include nausea, vomiling, abdominal
pain and cramps, and diarrhea; these gastraintes-
tinal symptoms rarely last more than 48 hours
from exposure with or without treatment. The
neurologic symptoms reported include pares-
thesia (in the extremities and around the mouth)
and weakness; these symptoms can persist upto 3
years in the case of one patient without treatment.

The patients were collected by two clinicians
(DB, DRA) with known interest in ciguatera
poisoning in the Miami (Florida) area from their
clinical practice since 1985. A Ciguatera Net-
work with referral telephone number had been set
up by the authors to give advice and recommend
treatment. Due to an ongoing research study, the
following information was collected for each
patient at presentation: age, sex, lime from ex-
posure until time of presentation, symptoms at
presentation, and type of fish implicated.

Patients with clinical features consistent with
either acute (ie. within 48 hours of exposure) or
chronic ciguatera poisoning seen after the publi-
cation by Palafox ctal_ (1988) were offered man-
nitol treatment. Both clinicians gave mannitol in
a dose of 1g/Kg; one clinician administered it
over 3—4 hours (Slow) and the other over 30
minutes (Rapid).

Patients were asked to rate their response im-

MANNITOL THERAPY FOR CIGUATERA

467

Response to Mannitol Treatment

Response
to mannitol

(Days 0-365)

Time from Exposure until treatment

FIG.1. Response to Intravenous Mannitol Treatment (Days 0-356)

mediately after receiving the mannitol in-
travenous treatment (i.e. after 3-4 hours with the
‘Slow’ treatment and after 30 minutes with the
‘Rapid’ treatment). The responses were rated on
a scale of 0 to 44+, in which 4+ meant recovery to
normal without further treatment needed and 0
meant no change in their symptoms. In any case
in which the reportedly positive effects of man-
nitol treatment did not last and neurologic
symptoms returned, repeat treatments were of-
fered under the same protocol until a 3+ to 4+
response was obtained and maintained.

Statistical analysis was performed on SE 30
MacIntosh using STATVIEW (Brainpower)
statistical program.

RESULTS

Of 107 subjects, 70 (65%) were treated with
mannitol (Table 1). 48% were women and 52%
men with a mean age of 39 (+ 15.9) and range of
1-79 years. The mean time from exposure to
presentation was 46 days (+ 149.5) with a range
to 0.3-1000 days. Symptoms consisted of purely
gastrointestinal in 12%, neurologic in 76% and

combination of the two types in 12%. The pathog-
nomonic symptom of cold to hot reversal was
reported by the majority, but not all patients. The
fish types reported were 27% grouper, 25%
kingfish, 9% amberjack, 8% barracuda, 8% snap-
per, 8% other, and 15% unknown.

The symptoms at presentation had the follow-
ing relationship to time from exposure: gastro-
intestinal symptoms alone were reported by those
presenting within the first 24 hours, then both
neurologic and gastrointestinal symptoms were
reported by those presenting 24 hours from ex-
posure until day 22, while neurologic symptoms
alone were reported by those presenting after day
1 and up to day 1000 from initial exposure.

Of the 37 patients who did not receive any
mannitol, there were 57% men and 43% women
with a mean age of 4] (+15.24), range 1-67
years. Their mean time to presentation from ex-
posure was 111 days (+235.3) with a range of
1—1000 days. The symptoms reported at the time
of presentation were 5% gastrointestinal, 92%
neurologic and 3% both; mannitol was not of-
fered and no other intervention (ie. supportive)
relieved these symptoms. Fish types identified

A5R

were 40% kingfish, 27% grouper, 11% snapper.
3% amberjack, 8% other, and 11% unknown.

70 patients were treated with 1Y mannitol. 50%
were males and 50% females with a mean age of
39 years (= 16.4), range 2-79 years. The mean
Ome to presentation from exposure was | 1.5 days
(+44.5), range 3-365 days. The symptoms
reported at the time of presentation were 16%
gastrointestinal, 67% neurologic, and 17% both,
Fish types reported were 27% grouper, 17%
kingfish, 13% amberjack, 11% barracuda, 7%
snapper, 10% other, and 14% unknown.

Of the 70 treated patients, $1 (73%) were
treated with the ‘slow’ mannitol treatment and 19
(27%) by ‘rapid’ treatment. The mean overall
response to mannitol treatment was a score of
3.3+ (+0,94) with arange of O-4+. There were
no adverse side effects reported to receiving man-
nitol treatment, either Rapid or Slow treatment.

32 (46%) of the 70 individuals treated within
the first 2 days from exposure. 91% had +4
response, and 100% had 3+ or 4+ (Fig.1). There
were 31 (44%) of the 70 individuals treated after
day 2 through day from exposure; 23% had a 4+
response, 31% had a 3+ response and 35% hada
2+ respanse to mannito) treatment. Finally, 7
(10%) of the 70 treated individuals were 1519365
days from exposure: 33% had a 4+ response, 33%
had a 3+ response and 33% had a 2+ response
while the individual who was 1 year from ex-
posure had no response to mannitol treatment.

As mentioned by Peam et al_ (1989), multiple
treatments (mean 2 treatments, but up to 4 treat-
ments) were required in five cases to maintain the
initial positive response to treatment. All the mul-
liple trealment persons presented within two days
from exposure. There were no further reports of
symptom recurrence or necessity for further
medication from successfully treated patients
after completing the mannitol Ireatments.

32 persons were treated within the first two
days from exposure; 50% were male and 50%
were female with a mean age of 31 years (range
2-60 years). Of these 32 persons, 28% reported
purely gastrointestinal symptoms, 38% reported
a mixture of gastrointestinal and neurologic
symptoms, and 34% reported purely neurologic
symptoms at the time of presentation. The
majority (89%) of those persons reporting purely
gastrointestinal symptoms at the time of presen-
tation were 24 hours or less from exposure, while
neurologic symptoms were reported after 24
hours from exposure.

Of the 51 patients treated by ‘slow’ mannitol
Ireatment, there were 53% men and 47% women,

MEMOIRS OF THE QUEENSLAND MUSEUM

with a mean age of 38 years (+ 17,7), range 2-79
years. The mean time to presentation (and treat-
ment) was 12.46 days (+50.7), range 0,3-365
days. The presenting symptoms dd were
16% gastrointestinal, 76% neurologic and, 3%
both. The fish types identified were 23% kingfish,
14% grouper, 14% amberjack, 12% barracuda,
10% other, and 19% unknown. The mean overall
response to “slow' mannitol treatment was rated
3.1+(+0.995)_ range 0-4.

Of the 19 patients treated with ‘rapid’ mannitol
treatment, 42% were men and 56% were women
with a mean age of 36 years (+121), range
12-68 years. The mean time to presentation (and
treatment) was 9.04 days (+ 19.3), range 04-70
days. The symptoms reported were 16%
gastrointestinal, 42% neurologic, and 42% both,
The fish types identified were 63% grouper, 11%
snapper, 11% barracuda, | 1% amberjack, and 3%
other, The mean overall response to “rapid” man-
nitol treatment was rated +3.7 (0,56), range 24+.

By ANOVA analysis (Table 2), there were no
differences between those with and without treat-
ment with respect to sex (F=0.437, p=0.51), age
(F=0.37, p=0.54), type of fish (F=1.67. p=0.19),
and type of symptom (F=0.167, p=0.68). There
was a Stalistically significant difference between
those with (11.5 ~ 16.4 days} and without treat-
ment (11] +235.2 days) with regards to time ta
presentation from exposure (F=11.8, p=0.0008).

There was no difference by ANOVA analysis
{Table 2) between the two treatment groups with
Tespect to age (F=0.57, p=0.45), sex (F=0.638,
p=.43), type fish (F=1.37, p=0.246), time to
presentation (F=0.081, and p=0.776). There was
a statistically significant difference between
Slow and Rapid treatment groups with regards to
response to treatment (F=6.9, p=0.01) and type of
symptoms at presentation (F=5.14, p=0,03), with
Rapid treatment resulting in a better response.

By correlation analysis, positive response to
mannitol treatment (ie. by the rating scale 0 to 4+)
was correlated with the type of treatment (ie,
Rapid vs Slow mannitol} (r=0.31) and time until
presentation from exposure to contaminated fish
(r=-0.442). By regression analysis in a mexicl
with the varjables of type of treatment and time
until treatment, both vanables were statistically
significant and predictive of successful response
to treatment (F=12.6, p= 001),

DISCUSSION

Although not a random controlled clinical tral,
treated and untreated patients with ciguatera

MANNITOL THERAPY FOR CIGUATERA

469

TABLE 2: Results of ANOVA ir ceaers of Treatment Group Outcomes

poisoning from south Florida were identified
which are comparable with respect to age, sex,
type of symptoms, and type of fish consumed.

The only significant difference between the
treated and untreated groups was that the treated
group were more likely to present to the authors
earlier in the course of their disease than untreated
patients, even though the symptoms and disease
entity were the same. We believe that the treated
patients as a group presented earlier in the course
of their illness due to new expectations for suc-
cessful treatment and improved early diagnosis
thanks to the community work in south Florida
on ciguatera by these investigators. The majority
of the persons in the untreated group had
presented early in the course of their illness to
other healthcare facilities and were treated unsuc-
cessfully with a variety of treatments prior to
being seen by the authors.

With regards to IV mannitol] treatment, it ap-
pears to be effective in all ages and both sexes.
There were no reported side effects to mannitol
treatment. It is most effective if given within the
first 48 hours from exposure. Mannitol treatment
was moderately effective if given from 3-14 days
from exposure (responses 2+ to 4+). In addition,
based on data derived only from single in-
dividuals, moderate success was seen with treat-
ment of individuals upto 70 days from exposure;
the one individual treated 1 year from exposure
had no response to mannitol treatment.

Multiple treatments (upto 4 additional treat-
ments) were necessary in 5 individuals; there was
complete resolution of symptoms with repeat
treatments. The absence of interviewer blinding
and of an objective measurement of response are
weaknesses in this study. However, persons who
reported a maintained successful response to
mannitol treatment did not return for further treat-
ment of any kind, while those who were not
treated by mannitol continued to report symp-
toms even 3 years after exposure.

The prolonged symptoms with accompanying
significantly increased time from exposure to
presentation (mean 111 +235.3 days) among the
untreated control group support the need to con-
sider mannitol treatment for ciguatera poisoning,

Treatment mean) (%Female Present

No Rx » =0.54 »=0.51 =O 008* p=0 68 p=() 19
F=1.37

Rapid Rx vs F=6.9 F=0.57 F=0.64 F=0.08 F=5.14
—Slow Rx__|_ p= 01%) p0.45__ |p =0.43_ | pe 077_ |p .034*

p=0.25

especially acutely. Although only a few in-
dividuals were treated after 14 days from ex-
posure, in our experience it is worthwhile
attempting mannitol treatment because it may
relieve or even eliminate the debilitating
neurologic symptoms of chronic ciguatera.

It appears that the more rapid administration
(30 minutes) may be slightly more effective, al-
though there were only 19 patients who received
this treatment and as a group, they presented
earlier in the course of the illness (9.04 vs 12.46
days) which correlates with a better response to
treatment. However, the slow intravenous ad-
ministration (over 3-4 hours) may be more ap-
propriate with small children and others who
cannot tolerate a heavy fluid load.

As opposed to the symptom course described
in Pacific ciguatera poisoning (Bagnis et al.,
1979), in our Atlantic experience the gastrointes-
tinal symptoms universally preceded the
neurologic symptoms. As such, it would be im-
portant to consider treatment with IV mannitol of
any person presenting with the acute onset of
gastrointestinal symptoms within 6-24 hours of
consuming a large reef fish from a tropical area,
even though the more classic neurologic
symptoms have not yet presented.

Multiple fish types were reported in these
ciguatera cases, although all were large reef fish
species from tropical areas. Ciguatera poisoning
has been associated with over 400 species
worldwide. Also of interest to clinicians and
epidemiologists, the ciguatera cases often
presented in clusters due to sharing of fish among
family and friends.

The social and economic impact of ciguatera
poisoning, due both to the threat and the actual
disease, is enormous. For example, in several
highly endemic areas, local fish are avoided as a
food source, as in south Florida where the sale of
barracuda (a major source of ciguatera poisoning
in the past) has been banned (Lawrence et al.,
1980). Fear of ciguatera poisoning has lead to
depression of local and exporting fishing in-
dustries and of tourism, and indirectly on human
health due to avoidance of fresh fish consumption
(despite its nutritional value) (Lewis,1986). The

470

impact of ciguatera and other marine toxins on
the fishing industry is evidenced by the FDA
Recommendations before Congress to require
mandatory testing for marine toxins in all marine
fish imported and sold in the United States.

Further work is needed on diagnosis, treatment
and epidemiology of ciguatera poisoning. First,
biomarkers in human fluids, as well as fish, are
needed for the diagnosis and management of
acute and chronic ciguatera poisoning- Animal
studies of IV mannitol treatment are necessary to
understand the mechanism of action. Random
controlled double blind tnals of 1V mannitol in
humans with biomarker-diagnosed ciguatera
poisoning are needed. Biomarkers would be use-
ful in determining the extent of acute and chronic
ciguatera poisoning in humans worldwide. Final-
ly, education of healthcare workers in endemic
areas i8 crucial for the correct recognition and
early intervention in ciguatera poisonmig.

ACKNOWLEDGEMENTS

We acknowledge Ms. Susanne Cramer AB for
her dedicated coordination of this project and Mr.
Mauricio Ortiz, MS for his statistical assistance,
We would like to thank the patients who par-
ticipated in this study.

LITERATURE CITED

BADEN, D,G., GAWLEY, R.E., KINOSHITA, M. &
REIN, K_S. 1990. Computational modelling of the
polyether ladder toxins: Brevetoxin and
Ciguatoxin. In Tosteson, TLR. (ed), “Proceedings
of the Third International Conference on
Ciguatera Fish Poisoning, Puerto Rico’ (Polys-
cience: Québec).

BAGNIS, R., KUBERSKI, T. & LAUGIER, S, 1979.
Clinical observations on 3,009 cases of ciguatera
(fish poisoning) in the South Pacific. American
Journal of Tropical Medicine and Hygiene 28;
1067-1073.

BERLIN, R.M., KING, S.L. & BLYTHE, D.G. 1992.
Systemic Improvement of chronic fatigue with
fluoxetine in Ciguatera fish poisoning (Letter).
Medical Journal of Australia 157: 567.

BLYTHE, D.G. & DA SYLVA, D, 1992, Ciguatera
Fish Poisoning. Miami Medicine 63: 31-32,

CARMICHAEL, W., JONES, C.L.A., MAHMOOD,
N.A. & THEISS, W.C. 1986. Algal toxins and
water-based diseases. CRC Critical Reviews in
Environmental Control 15(3):275-993.

GILLESPIE, N.C.. LEWIS, R.J., PEARN, J.H.,
BOURKE, A.T.. HOLMES, MJ.. BURKE, J.B.

MEMOIRS OF THE QUEENSLAND MUSEUM

& SHIELDS, W.J. 1986. Ciguatera in Australia:
occurrence, clinical features, pathophysiology
and management. Medical Joumal of Australia
145: 584-590.

HOKAMA, Y. 1985. A rapid, simplified enzyme im-
munosssay stick test for the detection of
ciguatoxin and related polyethers from fish tis-
sues, Toxicon 23: 939-946.

HOLT, R.J., MIRO, G. & DEL VALLE, A, 1984. An
analysis of poison control center reports of
ciguatera toxicity in Puerto Rico for one Year.
Jounal of Toxicology and Clinical Toxicology
22: 177-185.

INTERNATIONAL LABOUR ORGANIZA-
TION(UILO)-World Health Organization 1984,
Aquatic (Marine and Freshwater) Biotoxins, En-
vironmental Health Criteria 37. Geneva: World
Health Organization,

LANGE, W.R, 1987, Ciguatera toxicity. Amencan
Family Practice Journal 35; 177-182,

LAWRENCE, D.N., ENRIQUEZ, M.B,, LUMISH,
R.M. & MACEO, A. 1980. Ciguatera poisoning
in Miami. Jourmal of the Amencam Medical As-
sociation 244; 254-258.

LEGRAND, A.M., GALONNIER, M, & BAGIS, R.
1982. Studies on the mode of action of ciguatera
toxin, Toxicon 20;311-315,

LEWIS, N.D. 1986, Disease and Development:
ciguatera fish poisoning, Society for Scientific
Medicine 23(10):983-93.

MORRIS, J.G., LEWIS, P., SMITH, C.W., BLAKE,
P.A. & SCHNEIDER, R. 1982, Ciguatera fish
poisoning: cpidemiology of the disease in St.
Thomas, US Virgin Islands. American Journal of
Tropical Medicine and Hygiene 31: 574-578,

PALAFOX, N.A., JAIN, L.G., PINANO, A.Z.,
GULICK, T.M., WILLIAMS, R.K. & SCHATZ,
LJ, 1988. Successful treatment of ciguatera fish
poisoning with intravenous mannitol. Journal of
the American Medical Association 259: 2740—
2742.

PEARN, J.H.. LEWIS, RJ., RUFF, T., TAIT, M.,
QUINN, J., MURTHA, W., KING, G., MAL-
LETT, A, & GILLESPIE, N.C. 1989. Ciguatera
and mannitol: experience with new treatment
regimen. Medical Journal of Australia 151: 77-80.

RODGERS, D.L. & MUENCH, C. 1986, Ciguatera:
Scourge of seafood lovers. Sea Frontiers/Sea
Secrets 32: 338-346.

SAKAMOTO, Y_, LOCKEY,F.R. & KRZANOWSKI,
J,J, 1987, Shellfish and fish poisoning redated to
toxic dinoflagellates. Southem Medical Journal
80:866-72.

TRAINER, V.L. & BADEN, D.G. 1990. Enzyme im-
munoassay of Brevetoxins. Pp.430-435. In
Graneli, E., Sundstrom, B., Edler, L. & Anderson,
D.M. (eds), ‘Toxic marine phytoplankton’ {EI-
sevier:Amsterdam),

MANAGEMENT OF CIGUATERA FISH POISONING IN THE SOUTH PACIFIC
PAUL DALZELL

Dalzell, P. 1994 08 01: Management of ciguatera fish poisoning in the South Pacific.
Memoirs of the Queensland Museum 34(3):471-479. Brisbane. ISSN 0079-8835.

Catches of near-shore or coastal fish are a major source of animal protein in the South Pacific
region. Nominal landings amount to about 90,000 tonnes/yr, about half of which is reef fish.
Ciguatoxic fishes are ubiquitous but in some areas, such as Niutao in Tuvalu, there is a higher
risk of intoxication associated with eating reef fish. Commonly ciguateric fishes are normally
avoided, but where alterative food sources are not available, they are readily consumed. The
risk of poisoning is considered acceptable by the community and ciguatera is not regarded
as a significant health problem in most island countries. This risk of ciguatera poisoning is
reflected in the attitude of the medical community which is reluctant to see ciguatera given
priority over other public health problems. Although ciguatera is of relative low priority as
a medical problem in the region, fisheries development initiatives in high risk areas should
aim to reduce the incidence of ciguatera by improving supplies of non-toxic deep reef slope
and pelagic species. Elsewhere, ciguatera management is mainly required to limit the impact
of fish poisonings on tourism and reef fish exports. Initiatives to improve the management
of ciguatera such as the South Pacific Commission’s seafood poisoning database are
discussed.

Paul Dalzell, Inshore Fisheries Research Project, South Pacific Commission, BP D5

Noumea Cedex, New Caledonia; 2 February, 1994.

Dalzell (1992) presented an overview of
ciguatera in the South Pacific from the perspec-
tive of fisheries development. In common with
statistics on other illnesses, ciguatera cases are
incompletely documented in the region. The
South Pacific Epidemiological and Health Infor-
mation Service (SPEHIS) records numbers of
cases of fish poisoning reported by national
health departments, which it is assumed represent
mostly ciguatera intoxications. The annual num-
ber of reported cases fluctuates (Dalzell,1992,
fig.1). Lewis (1986) suggested that reported cases
account for c.20% of incidence of ciguatera in the
region. The increased incidence of fish poisoning
from 1985 on (Dalzell, 1992, fig.1) may be due to
improvements in diagnosing ciguatera and in data
collection rather than a rise in the occurrence of
ciguatera.

Landings of coastal fishes in most of the islands
of the South Pacific come from coral reefs and
lagoons (Dalzell,1992, table 1). In many loca-
tions fish is still a principle source of animal
protein and forms a major component of subsis-
tence diets. In the smaller countries of the region,
fish stocks represent one of the few viable
economic resources that have potential for
development. Outbreaks of ciguatera, besides
being a health hazard, may have detrimental ef-
fects on fish production and marketing through
adverse publicity and litigation.

Key issues that face health and fisheries

workers in the South Pacific are:1, How big a
problem is ciguatera to community health and the
economy? 2, If ciguatera is a serious problem,
how and where should resources be focused to
manage it?

Dalzell (1992) summarized Pacific Island
fisheries and discussed possible effects of
ciguatera on development fisheries. A similar
summary is given here as a preliminary to assess-
ing steps that may limit ciguatera outbreaks.

SOUTH PACIFIC REEF FISHERIES

Statistics on the composition and distribution
of landings from reef and other coastal fisheries
in the South Pacific, are poorly developed. Coas-
tal fishing is characterised by small scale methods
catching a wide variety of species and landing
fish at many locations. This, plus the limited
resources of most fisheries departments in the
region, results in poor coverage of fisheries
production.

The nominal coastal fin-fish landings for the
South Pacific region (Dalzell,1992, table 1)
should not be confused with absolute values,
rather this is the best estimate that can be made
given data sources available. Estimated fisheries
production for the region is about 90,000 ton-
nes/yr (Dalzell,1992, table 1), although this is
almost certainly an underestimate. Assuming a
regional average value of US$2.00/kg for these

Av?

Acute rasp Infection

Conjunctivitis
Matnutrition 4
Measles
Fish poisoning
Gonorrhea/NSU
Dengue 4—
Tuberculosis f=" 5}
Tetanus fo =]
Hepaptitis f=
Meningitis" ]
Leprosy 2."

aw ah ‘ iw Loe

Occurence (cases per 1000)

FIG.1. The 15 most commonly reported illnesses from
the SPEHIS database in ranked order of occurrence
(averaged between 1988 and 1992),

landings, then the potential nominal value is
c.US$174,000,000/¥r.

Given the quality of the data on fishenes land-
ings and on the incidence of ciguatera, itis dif-
ficult to draw any firm conclusions (Dalzcil,
1992, table 1). The volume of fish production per
capita tends to be highest in the smaller, less-
developed islands and atolls with limited land
area and long traditions of fishing. These loca-
tions also record the highest incidence of
ciguatera. Fresh fish production (and consump-
tion) is lowest in the large islands of Melanesia.
where some of the population live in the interior
(especially PNG) and people traditionally look to
the land as a principal source of food.

The percentage of the total fish landings that
comprise teef fish ranges from 7-88% with a
mean of 51% (Dalzell,1992, table 1), About half
the families of reef fish landed in the Pacific
contain species known te haye been ciguateric
(Dalzell,1992, table 2). Certam species such as
the small surgeon fish Crenochaetus striatus, the
snapper Lutjanus bohar and barracudas
(Sphyraena spp.) are known health hazards but
are still occasionally consumed. In some areas,
these species are known not to be ciguaioxic and
are safely consumed.

Fishing for reef fish in the South Pacific is
accomplished mainly with hand-lines, traps, nets
and spears, deployed from dinghies or canoes.
Lock (1986) presented observations of the com-

MEMOIRS OF THE QUEENSLAND MUSEUM

position of the catch taken by differemt fishing
gears deployed on the South Papuan Barrier Reef
off Port Moresby. Hand-lines select mostly for
carmivorous species, which may be ciguatoxic
through the ingestion of already contaminated
prey species. Spearing and net fishing take a
wider range of species, including reef herbivores
that become ciguatoxic through ingestion of the
dinoflagellate, Gambierdiscus toxicus, Catch
rates of such smal] scale artisanal gears are
modest at best — on average 1—4kg/man-hour,
depending on the gear (Dalzell & Wright,1986)
and there may be a disincentive for fishermen to
retum a portion of the catch to the sea simply
because of suspected ciguatoxicity.

Food accounts for c.20% of imports into Pacific
Island countries as opposed to <10% of imports
into the metropolitan countries (SPC,1992),
Domestic food production, and hence fisheries
production, is important to diminish the reliance
of the Pacific nations on imported foods. How-
ever, as island populations increase, fish harvests
will also increase, both in subsistence and com-
mercial sectors, In some countries of the region.
harvest levels from coastal fisheries are already
very high and fishing pressure is unlikely to
diminish. The main objectives for fisheries
managers und admunistrators are developing fish
production and maintaining it at a sustainable
level,

CIGUATERA AS A REGIONAL
HEALTH ISSUE

A range of different illnesses and afflictions are
recorded in the SPEHIS database (Fig.1). The
level of under-reporting is unknown and some
countries may report few medical statistics. Fish
poisoning (which it is assumed represents mainly
ciguatera) is ranked eighth, The mformation is
misleading as the high incidence of yaws is based
on cases reported mainly from the Solomon Is-
lands, Similarly, the incidence of malaria is con-
fined to Papua New Guinea, the Solomon Islands
and Vanuatu. However, acute respiratory infec-
lions and gastro-intestinal infections are more
evenly spread throughout the Pacific and are the
major health issue in the region, Dalzell & Gawell
(1989) reported that respiratory tract infections
and pastro-intestinal diseases are the main con-
cem of the Health Department in the Federated
States of Micronesia, particularly amongst the
young who compnise the majority of the popula-
tion.

The non-communicable nature and relatively

MANAGEMENT OF CIGUATERA IN THE SOUTH PACIFIC

low mortality rates from ciguatera mean that
while the risk of fish poisoning is acknowledged.
health departments place higher priorities on
other health problems. The most effort health

ments are likely to devote to ciguateta is
publicising those species which are dangerous to
consume and sometimes restricting marketing of
those species.

Despite the possible impact of ciguatera on the
fisheries in the South Pacific, there appears to be
litle concern about this form of intoxication from
fisheries departments in the region, Fisheries of-
ficers May express mleres! in ciguatera, par-
ticularly if there is an outbreak in their country,
but in general itdocs not appear to be givena very
high priosity,

Ciguatera was discussed at the first SPC
fisheries conference (SPC.1952); was the subject
of an SPC technical paper (Banner et al.,1963);
and has been the topic of several workshops and
conferences (SPC,1968,1978,1981,1988). The
Commission has also produced a handbook on
ciguatera (Bagnis,1973) and the Inshore
Fisheries Research Project formed a Ciguatera
Special Interest Group in 1991 which has pub-
lished two bulletins.

Regular international mectings, symposia and
workshops have addressed the subject of
ciguatera. Despite these meetings, reports and
publications there has been little in the way of
requests for assistance from the various member
countries of the Commission to deal with
ciguatera.

THE SPC SEAFOOD POISONING
DATABASE

In the absence of any reliable information on
ciguatera from most of the South Pacific (French
Polynesia, Hawaii and Australia are the excep-
tions), the South Pacific Commission established
a regional seafood poisoning database in late
1990.

Both fisheries and health workers have been
contacted to report (Dalzell,1992, fig.3) cases of
fish poisoning, The database has been publicised
repeatedly in SPC publications and in national
newspapers and by radio. Despite repeated at-
tempts to encourage the reporting of cases, the
response has been vanable.

Over 400 case histories have been collected
from eight countnes in the South Pacific (Table
1). Case histories from elsewhere are limited or
non-existent. even from countries known from
the SPEHIS database to have a high incidence of

473

ciguatera such as Kiribati, Tokelau and the Mar-
shall Islands (Dalzell,1992: table 1). This may be
due to the priority accorded to ciguatera by na-
tional medical departments and the lack of train-
ing for medical and fisheries personnel in the
collection of case history dala

Not all intoxications (Table 1) are likely to be
due to ciguatera and case histories involving skip-
jack juna (Katsuwonus pelamis), herring
(Herklorsichthys quadrimaculatus) and snake
mackerel (Promethichthys prometheus) are
probably the result of some other form of paison-
ing. Several invertebrates such as crahs, lobsters,
clams and sea cucumber, have also been involved
in a number of polsonings. although again,
another form of toxin other than ciguatera
probably was responsible for causing illness.

Given the complexities of coral reef fish
taxonomy and the majority of case histones
recorded by medical personnel, itis not surprising
that many reports do not identify precisely the
fish responsible for intoxications. Thus it is com-
mon, forexample, for the person poisoned to give
a local genenec term for surgeonfishes. groupers
and parrotfish. Occasionally some local words for
particular fish are for individual species. Good
examples are the blue-line surgeon fish, Acar-
thurus lineatus, and the convict lang, Acanthurus
iriostegus. which in Tuvalu are Known as
‘ponelolo” and ‘manini’ respectively. Indeed,
manini is 2 common name for A. triestegus in
much of Polynesia, where it js a common food
fish. However, in many of the poisonings
reported from Tuvalu involving surgeonfish, the
term ‘pone’ 3s used a8 a generic term for all
acanthurids,

Intoxications involving surgeenfish (20.7%)
and parrotfish (12%) together account fore. 1/3 of
case histories in the database, Other families
responsible for 10% of intoxications include the
groupers (15%) and the snappers (10%). Besides
A, lineatus and A. triestegus, other species cons-
monly implicated in poisonings are the snappers
Lutjanus bohay and L. monostiomus, the groupers
Plectropoma spp. and Cephalopholis argus, ancl
the soapfish, Granunisves sexlineatus,

Surgeunfish and pattothish are two of the com-
mon families of fishes on coral reefs and often
form major fractions of the landings in a reef
fishery. At two locations in Papua New Guinea,
they fonmed c.12% of landings from reef fisheries
in the north and south of the country (Dalzell &
Wright,1986), Parotfish and surgeonfish ac-
counted for 1/3 of commercial landings ef reef
fish in Palau during 1976-1990 (Kitalong & Dal-

474 MEMOIRS OF THE QUEENSLAND MUSEUM

Species Fiji Federated Kiribati Marshall Nauru New Niue Tuvalu
States of Is Caledonia

Micronesia

FINFISH

Abudefduf spp.

Acanthuridae

Acanthurus leucopareius

Acanthurus lineatus

Acanthurus triostegus

Balistidae

Carangoides ferdau

Caranx ignobilis

Carcharinus longimanus

Cephalopholis argus

Chaetodontidae

Chelinus undulatus

Ctenochaetus striatus

Decapterus macarellus

Epinephelus cyanopodus

Epinephelus fuscoguttatus

Epinephelus melanostigma

Epinephelus microdon

Epinephelus spp.

Grammistes sexlineatus

Gymnothorax javanicus

Herklotsichthys

quadrimaculatus

Holocentridae

Katsuwonus pelamis

Kyphosidae

Lethrinidae

Lethrinus elongatus

Lutjanidae

Lutjanus argentimaculatus

Lutjanus bohar

Lutjanus gibbus

Lutjanus monostigmus

Lutjanus sebae

Monocanthidae

Monotaxis grandoculus

Mugilidae

Mullidae

Naso brevirostris

Naso unicornis

Plectropoma spp.

Promethichthys prometheus

Sargocentrum spiniferum

Scaridae

Serranidae

Shark

Siganidae

Soleidae

Sphyraena jello

Sphyraenidae

Symphorus nematophorus

Unknown fish

MANAGEMENT OF CIGUATERA IN THE SOUTH PACIFIC

Federated
States of
Micronesia

Species Fiji

CRUSTACEA
Carpilius maculatus
Panulirus spp.
Scylla serata

MOLLUSCS
Chama pacifica

ECHINODERMS
Holothuroidea

Kiribati

475

Marshall Nauru New Niue Tuvalu

Is Caledonia

TABLE 1. Species implicated in fish poisoning by country in the South Pacific Commission seafood poisoning
database. Fish that could not be identified to species are grouped under family headings.

zell,1994). In the Philippines, parrotfish and sur-
geonfish formed 16 and 27% of landings from
two coral reef fisheries (Bellwood,1988; Alcala
& Russ,1990). Sims (1988) estimated that im-
ports of fish into Rarotonga from Palmerston
Atoll were composed mainly of parrotfish (70—
80%) and that annual landings of parrotfish
ranged from 15-20t. Smith & Dalzell (1993)
found that surgeonfish and parrotfish comprised
about 74% of landings from community spear
and net fishing on Woleai Atoll in Micronesia and
that the Scaridae and Acanthuridae accounted for
60-90% of the fishable biomass on lagoon back
reefs.

Grazing herbivores such as the parrotfish and
surgeonfish are likely to be toxic if a ciguatera
outbreak occurs on a reef and thus a sizeable
fraction of the fishable stocks on reefs may be-
come a health hazard. The same is true of the
snapper Lutjanus bohar, which is one of the most
common predatory species on coral reefs and is
widely distributed throughout much of the tropi-
cal Indian and Pacific Oceans. This species may
account for up to half the fish caught by handlines
in some areas (Dalzell & Preston,1992) but may
be rejected due to its reputation for toxicity. Sale
of L. bohar is prohibited in Mauritius due to its
toxic reputation, although this species is a
dominant feature of handline catches at the banks
and islands of the western Indian Ocean (Wheeler
& Ommanney,1953).

The five most common maladies from
ciguatera are joint aches, headache, temperature
reversal, diarrhoea and muscle cramps (Fig.2).
The five most common symptoms from eating
herbivorous fishes were joint aches, headache,
diarrhoea, temperature reversal and vomiting
(Fig.2). Eating carnivorous fishes most common-
ly caused joint aches, headache, temperature

reversal, muscle cramps and tingling & numbness
(Fig.2). Gastro-intestinal maladies such as vomit-
ing and diarrhoea occur but with less frequency.

Collection of case history data will continue
and further efforts will be made to increase the
rate of reporting, particularly from those
countries where ciguatera is relatively common
but case histories are not forthcoming.

MANAGEMENT OF CIGUATERA

If ciguatera is not perceived as a major health
hazard in most places, then this should guide the
approaches taken to manage this problem. In
locations such as Papua New Guinea (PNG), the
incidence of ciguatera is very low although cases
do occur intermittently. The PNG Medical Jour-
nal (1950s—1993) did not uncover a single case of
ciguatera poisoning, although reports were given
of turtle meat poisoning (Dewdney,1967),
paralytic shell fish poisoning (Rhodes et al.,1975)
and scombroid fish poisoning (Barrs, 1985). Ac-
cording to the Department of Fisheries and
Marine Resources (DFMR) 5 outbreaks of
ciguatera occurred between 1971 and 1981 at
Port Moresby, Finschhafen, Milne Bay and New
Hanover (Anonymous 1988b).

The species involved in the poisonings were
coral trout (Plectropoma sp.), red bass (Lutjanus
bohar), parrotfish (Scarus strongylocephalus)
and barracuda (Sphyraena sp.). In all cases no
fatalities were recorded and most people
recovered in 3-5 days. The only instance of a
commercial ban on the purchase of species on the
basis of suspected ciguatera is that of the Milne
Bay Fishing Authority in regard to the red bass,
L. bohar. The general consensus from the DFMR
summary is that ciguatera is a minor problem and
that the importance of fish to the economy and

476

Total

Joint aches

Headache

Temp reversal

Diarrhoea

Muscle cramps
Tingling numbness
Fever & chills
Pricking sensation

Vomiting

Strange taste

Symptoms

Excess sweating
Itching & redness
Discomfort urinating
Difficulty breathing
Excess Salivation
Eye irritation
Difficulty walking
Difficulty talking

MEMOIRS OF THE QUEENSLAND MUSEUM

Herbivores Carnivores

it) 5 10 15
Percentage

FIG. 2. Ranking of ciguatera symptoms from the consumption of all fishes, and separately the herbivores and
carnivores in the South Pacific reported to the SPC seafoods poisoning database (N=380).

diet of the country is such that it would be impru-
dent to ban all or any species of which a few
individuals may be dangerous to eat.

There are, however, places where ciguatera is
a serious problem. These are usually small atolls
or coral islands where per-capita fish and seafood
consumption is very high (Dalzell, 1992: table 1).
A good example is Niutao Island in the Tuvalu
archipelago. Since the establishment of the SPC
database in 1990, over 200 cases of ciguatera
have been reported from Niutao Island. For a
population of almost 1000 people this is an in-
cidence equal to about 10% of the population per
year. Niutao is a coral island of 226ha, that

encloses a small landlocked brackish lagoon and
is surrounded by a fringing reef of c.108ha. Kaly
et al. (1991) reported 80 case histories from the
1988 Niutao ciguatera outbreak, some months
prior to a programme of subtidal reef blasting to
create boat passages. Some blasting was con-
ducted on the reef at Niutao in 1981 to create a
boat channel but this was not followed by a
serious outbreak of ciguatera. According to Kaly
et al., there were on average fewer than three
cases of ciguatera per year prior to 1988 resulting
from the consumption of susceptible fish such as
L. bohar. The outbreak, which begun in 1988,
shows no signs of abating. Monthly occurence of

MANAGEMENT OF CIGUATERA IN THE SOUTH PACIFIC

20
34s
=
3
3
=
rf
B io
=
&
z
5
DIitWMaMl | AS OARIEMAMS TASOND:
1991 1992

Month

FIG.3. Number of case histories of ciguatera fish
poisoning recorded in Niutao, Data smoothed by a
running average of 3.

Ciguatera cases (Fig.3), shows that intoxications
occur throughout the year but with much varia-
tion between months,

Given the isolation and size of Niutao animal
protein necessarily comes from the sea. However,
reef fish are a risk on this island and persons are
likely to be exposed to ciguatoxins on a regular
basis through the consumption of reef fish. The
most effective approach to managing the problem
1s to increase the supply of non-toxic fish. This
means targeting pelagic species that live beyond
the reef, and demersal snapper and grouper of the
deep reef slope. Most pelagic and deep slope
species are rarely ciguatoxic.

Deep reef slope fishes include mainly large
snappers, groupers, emperors and a mix of other
species including oilfish, jacks and barracuda.
The deep slope dwelling snappers (Flelinae) have
not been implicated in fish poisoning, as opposed
to the shallow reef snappers (Lutjaninae). Al-
though carmivorous, the diet of the Etelinae is
mainly restricted to fish and benthos from the
immediate environment (Parrish,1987) and not
shallow reef dwelling herbivores. The same is
probably true for groupers on the deep reef slope.

Stocks of deep slope species are more Jimited
then pelagic fishes, particularly around small is-
lands where the steep submarine gradient off the
shelf limits the deep slope habitat. However, sub-
sistence catches of deep slope species could play
their part in displacing reef fish from the diet
where ciguatera is a severe problem. The South
Pacific Commission has been instrumental in

477

successfully introducing the simple technology
required to catch deep slope fishes in a number of
countries.

Deep reef slope species may range between the
shallow reef and deeper waters away from the
reef. In Hawaii the amberjack (Seriola dumeril!)
has been implicated in a number of ciguatera
outbreaks, despite being caught predominantly
from the deep reef slope. Outbreaks of ciguatera
led to the demise of the fishery for this species
which amounted to about 30 tonnes annually
(Humphreys,1986). Similarly, catches of deep
slope fishes in the South Pacific include the snap-
per Lutianus bohar, Untike the eteline snappers,
this species is found throughout the water column
and is an opportunistic predator (Wright et al.,
1986) with w diet that includes reef fish. L bohar
comprised 8% by weight of the catches of deep
slope species in the South Pacific (Dalzell &
Presion,1992). For the atolls and coral islands
only, catches of L bohar amounted to about 15%
of tolal landings, Thus, although deep slope
species can be an important supplement to the
diet, potentially toxic fish may still be involved,

Similar strategies can be adapted for other
countries where small islands are severely af-
fected by ciguatera. In larger coral atolls where
there is a substantial lagoon and reef area, sites
that are known to produce ciguatoxic fish can be
avoided in favour of safe fishing grounds,
Development of pelagic and deep slope fisheries
should be pursued, however, so that the resource
base is widened and reef fish ts not the sole source
of fish protein. However, people may persist in
fishing in known ciguateric sites and eating fishes
which are recognised as being potentially
dangerous. It will be difficult to control consump-
tion of fish in most countries and there is an
accepted risk in eating reef fishes even in those
islands where they may be ciguatoxic. Manage-
ment under these circumstances should be
restricted to publicising those species which are
known to be implicated in causing ciguatera,

A number of countries and ferritones of the
South Pacific are heavily dependant on aid to
maintain their cconomies. The typical island stare
is small and it has few natural resources, Trade
between other islands is also limited because
what one island can offer, the products of subsis-
tence farming and fishing, can be found on most
other islands, Islands that produce cocoa, palm oil
and copra are competing against much larger
producers elsewhere in the tropics. Revenues are
accrucd from permitting foreign fishing vessels
to catch tuna within the 200 n.mi exclusive

478

economic zones (EEZ) of some of the Pacific
Islands.

Some manufacturing enterprises do exist but
are small and generally serve limited domestic
markets. Further, imports of raw material tend to
be costly because relatively small quantities have
to be shipped over long distances. Probably the
greatest economic potential for many of the South
Pacific Islands, particularly those with limited
natural resources, is tourism (Anon 1991). In
some countries such as Fiji, Vanuatu, and parts of
Micronesia, tourism is a major source of revenue
and employment. Other countries in the region
are trying to develop a better tourist industry
infrastructure and increase earnings from this
sector of the economy.

Ciguatera can be a threat to the hotel and res-
taurant business and to the tourist trade in general.
The results of intoxication may include loss of
business for the individual establishment and
potentially for a particular location if the problem
is severe enough. Further, there is the added risk
of litigation brought by persons who have been
poisoned. Hotels and restaurants in countries
where tourism is important should be clearly in-
formed about which fishes are in the high risk
category. This is particularly important in institu-
tions that have chefs and kitchen staff from over-
seas and who may not be familiar with potentially
toxic fish, and where fishermen may tend to sell
suspect fish that might be rejected by local con-
sumers.

Most nations in the South Pacific have expand-
ing commercial fishing industries. As described
earlier, increasing volumes of reef, pelagic and
deep slope fishes are being sent to markets in
Japan, Hawaii, New Zealand and Australia from
Pacific Islands to take advantage of high market
prices. Outbreaks of ciguatera poisoning can have
adverse effects on all coastal fisheries since, in
the mind of retailers and consumers, even fish
which are perfectly safe may be feared. As with
tourism, litigation may be brought against fisher-
men and exporters with adverse consequences for
an expanding fishing industry.

If countries manage to establish exports of reef
fish, then high risk species such as Lutjanus
bohar, moray eels, barracuda and some surgeon-
fish (eg. Acanthurus lineatus, Acanthurus trios-
tegus, Ctenochaetus striatus) might be prohibited
from export. Unlike subsistence production, the
scale of fish exports is likely to be such that
control and management can be effectively im-
plemented. Fishes that are sent for export should
be scrutinised and checked as to the location of

MEMOIRS OF THE QUEENSLAND MUSEUM

the catch, particularly if reef areas with
ciguatoxic fish are known to exist. Affordable
tests for ciguatera (Park,1992) should be routine
for export fishes.

ACKNOWLEDGEMENTS

I thank Dr Tim Adams and Mr Peter Cusack
(SPC Fisheries Programme), and Dr Steve Ter-
rell-Perica (SPC Health Programme) for review-
ing this paper and for their helpful comments and
suggestions which greatly improved the final
draft.

LITERATURE CITED

ALCALA, A.C & RUSS, G.R. 1990. A direct test on
the effects of protective management on abun-
dance and yield. Cons. Int Exploration Mer 46:
40-47.

ANON. 1988. ‘Ciguatera in Papua New Guinea’. Infor-
mation Paper 13, Twentieth Regional Technical
Meeting on Fisheries, 15 August, Noumea. South
Pacific Commission, Noumea, New Caledonia,
3p.

ANON. 1991. The Pacific Idea. The Economist 318:
17-20.

BAGNIS, R. 1973. ‘Fish poisoning in the South
Pacific’. (South Pacific Commission: Noumea,
New Caledonia) 112p.

BANNER, A.H., SHAW, S.W., ALEXANDER, C.B.
& HELFRICH, P. 1963. Fish intoxication: notes
on ciguatera, its mode of action and a suggested
therapy. South Pacific Commission Technical
Paper 141: 1-17.

BARRS, P. 1985. Scombroid fish poisoning at Alotau.
Papua New Guinea Medical Journal 28: 131.

BELLWOOD, D.R. 1988. Seasonal changes in the size
and composition of the fish yield from reefs
around Apo Island, central Philippines, with notes
on the methods of yield estimation. Journal of
Fisheries Biology 32: 881-893.

DALZELL, P. 1992. Ciguatera fish poisoning and
fisheries development in the South Pacific region.
Bulletin de la Société de Pathologie Exotique 85:
435-444.

DALZELL, P. & WRIGHT, A. 1986. An assessment of
the exploitation of coral reef fishery resources in
Papua New Guinea. Pp.477—-48 1. In J.L. Maclean,
L.B. Dizon & L.V. Hosillos (eds), ‘The first Asian
fisheries forum’ (Asian Fisheries Society:
Manila).

DALZELL, P. & GAWELL, M.J. 1989. A proposed
sampling protocol for ciguatoxic reef fishes in the
Federated States of Micronesia. South Pacific
Commission Inshore Fisheries Research Project,
Noumea, (New Caledonia and Department of
Resources and Development: Pohnpei, FSM).

DALZELL, P. & PRESTON, G.L. 1992. Deep slope
fishery resources of the South Pacific. South

MANAGEMENT OF CIGUATERA IN THE SOUTH PACIFIC

Pacific Commission, Inshore Fisheries Research
Technical Document 2: 1-299

DEWDNEY, J.C.H. 1967. Turtle meat poisoning the
New Ireland epidemic, 1965. Papua New Guinea
Medical Journal 10 (2): 55-58.

HUMPHHREYS, R.L. 1986. Carangidae: greater am-
berjack. Pp.100—-101. In R.N. Uchida & J.H.
Uchiyama (eds) ‘Fishery atlas of the northwestern
Hawaiian Islands’. NOAA Technical Report
NMFS 38.

KALY, U.L., JONES, G.P. & TRICKLEBANK, K.
1991. Preliminary assessment ofa severe outbreak
of ciguatera at Niutao, Tuvalu. South Pacific Jour-
nal of Natural Science 11: 62-81.

KITALONG, A.H. & DALZELL, P. 1994. ‘A prelimi-
nary assessment of the status of inshore coral reef
fish stocks in Palau’. South Pacific Commission
Inshore Fisheries Project Technical Document 6:
1-37.

LEWIS, N.D. 1986. Epidemiology and impact of
ciguatera in the Pacific: a review. Marine Fisheries
Review 48(4): 6-13.

LOCK, J.M. 1986. Study of the Port Moresby artisanal
reef fishery. Department of Primary Industry,
Fisheries Division, Technical Report 86-1: 1-56.

PARK, D.L. 1992. Rapid facile solid-phase im-
munobead assay for screening ciguatoxic fish in
the market place. Paper presented at IVth Interna-
tional Conference on Ciguatera Fish Poisoning,
4-7 May, Papeete, Tahiti.

PARRISH, J.D. 1987. The trophic biology of snappers
and groupers. Pp.405-464. In J.J. Polovina & S.
Ralston (eds), “Tropical snappers and groupers:
biology and management’. (Westview Press:
Boulder, Colorado).

RHODES, F.A., MILLS, C.G. & POPEI, K. 1975.
Paralytic shellfish poisoning in Papua New
Guinea. Papua New Guinea Medical Journal
18(4): 197-202.

479

SIMS, N.A. 1988. ‘Cook Islands fisheries resource
profiles No. 5: Parrotfish’. (Ministry of Marine
Resources: Rarotonga, Cook Islands), 14p.

SMITH, A. & DALZELL, P. 1993. ‘Fisheries resources
and managemenmt investigations in Woleai Atoll,
Yap State, Federated States of Micronesia’. South
Pacific Commission Inshore Fisheries Project
Technical Document 4: 1-64.

SPC 1952. ‘Report of meeting. Fisheries conference,
1422 May 1952’. (South Pacific Commission:
Noumea, New Caledonia), 46p.

SPC 1968. ‘Report of meeting. Seminar on ichthyosar-
cotoxism, 1622 August, Papeete’. (South Pacific
Commission: Noumea, New Caledonia), 13p.

SPC 1978. ‘Report of meeting. Expert committee meet-
ing on ciguatera fish poisoning, 2226 May,
Papeete’. (South Pacific Commission: Noumea,
New Caledonia), 15p.

SPC 1981. ‘Report of meeting. Expert committee meet-
ing on ciguatera, 26 February, Suva’. (South
Pacific Commission: Noumea, New Caledonia)
26p.

SPC 1988. ‘Report of meeting. Twentieth regional tech-
nical meeting on fisheries, 15 August, Noumea’.
(South Pacific Commission: Noumea, New
Caledonia), 26p.

SPC 1992. ‘Coastal fisheries statistics in the South
Pacific. Twenty fourth regional technical meeting
on fisheries, 26 August, Noumea, Working Paper
9’. (South Pacific Commission: Noumea, New
Caledonia), 7p.

WHEELER, J.G.F. & OMMANEY, F.D. 1953. Report
on the Mauritius — Seychelles fisheries survey,
1948-49. Fisheries Publications 1(3): 1-145.

WRIGHT, A., DALZELL, P.J. & RICHARDS, A.H.
1986. Some aspects of the biology of the red bass,
Lutjanus bohar (Forskal), from the Tigali Islands,
Papua New Guinea. Journal of Fish Biology 28:
533-544,

EVALUATION OF A SOLID-PHASE IMMUNOBEAD ASSAY FOR DETECTION OF

CIGUATERA-RELATED BIOTOXINS IN CARIBBEAN FINFISH
ROBERT W, DICKEY, H. RAY GRANADE AND FOSTER D. MCCLURE

Dickey, R.W.,, Granade, H.R. & McClure, F.D, 1994 08 O1: Evaluation of a solid-phase
immunobead assay for detection of ciguatera-related biotoxins in caribbean finfish, Memoirs
of the Queensland Museum 34(3): 481-488, Brisbane. ISSN (079-8835.

The Ciguatect™ solid-phase immunobead assay for detection of cigualera-related palyether
biotoxins in finfish was evaluated for consistency with the mouse bioassay. Fifty finfish from
ciguatera-endemic waters of St. Thomas, U.S. Virgin Islands, and one fish remnant from a
confirmed case of ciguatera poisoning were mouse bioassayed. The 51 specimens were then
assayed using the Ciguatect’ assay with 3 different methods of tissue sampling: single
exposure, triple exposure and single exposure to solvent extract from Alesh (RE ‘rapid
extract method). Qualitative statistical analyses ascertained false positive and false negative
rates. Positive matches for the single, triple and REM™ methods of tissue sampling were
58, 8S and 94%, respectively, and negative matches were 17, 22 and 12%, respectively-
Coresponding false negative rates were 82, 55 and 50%, and false positive rales were 44,
33 and 33%. Predictive indices for Ciguatect ™ performance under ciguatoxin contamina-
lun tates ranging from 5 to 75% project that high false negative und false positive values
might be expected in market situations.

Robert W. Dickey & H. Ray Granade, Gulf Coast Seafood Laboratory, Office of Seafood,
Food and Drug Administration, P.O. Box 158, Dauphin Island, AL 36528: Foster D.
MeClure, Division ef Mathematics, Food and Drug Administration, 200 C Strees, SW.

Washington, DC 20204; 28 March, 1994.

The major impediment to management strat-
egies for the ciguatera public health hazard is the
difficulty in detecting highly potent ciguatoxins
{and maitotoxins) in seafood matrices. Cigua-
toxin and maitotoxin structures have been eluc-
idated from Pacific sources (Murata et al., 1989,
1990,1992,1993; Lewis etal.,1991; Yokoyama et
al.1988; Holmes et al., 1990), but Caribbean
forms have not been resolved and analytical
methods to detect and quantify them are not yet
available. Several potential assessory toxins have
also been characterized (Murakami et al..1982:
Torigoe et al.,1988; Nagai et al.,1992; Fukui et
al., 1987). Difficulties with developing ciguatera
detection methods are due to lack of analytical
standards. An immunochemical assay for detec-
tion of ciguatera-related polyethers was devel-
oped (Hokama,1985,1990: Hokama et al.,1990,
1992) and modified into a ‘kit’ format,
*‘Ciguatect™” (Park et al.,1992), We compare
this assay with the mouse bioassay, the most
widely recognized method for identification of
ciguatoxic finfish.

EXPERIMENTAL
APPARATUS

Eberbach 8017 explosion-proof blender (Eber-
bach Corp., Ana Arbor, MI}; Brinkmann/Buchi

rotary evaporator Model RE-111 (Brinkmann In-
struments, Inc., Westbury, NY}; Mettler Model]
PM2000 top-loading balance (Mettler Instrument
Corp., Hightstown, NJ); Sartonus Model R180D
analytical balance (Sartorius Corp., Bohemia,
NY); Savant Model §S-2 centrifugal evaporator
(Savant Instruments, Inc., Farmimgdale, NY},

REAGENTS AND MATERIALS

Sclid-phase immunobead assay (SPIA)
materials were purchased from Hawaii Chemiect
Intemational, Inc. (San Diego, CA) on 24 May
1992 {immunobead lot number 051492) and
again on 6 October 1992 (immunobead lot num-
ber 100592). All solvents were of reagent grade:
or better (J-T. Baker Chemical Co., Phillipsburg.
NJ). Silica gel solid-phase extraction columns
were obtained from Varjan Associates, Inc, (Sun-
nyvale, CA). Swiss white mice (Crl:CFW
(SW)BR) were obtained from Charles River
Laboratories (Wilmington, MA).

SPECIMEN COLLECTION

Potentially ciguatoxic finfish were collected
4-13 June,1992 from ciguaters-endemic waters.
8 of St. Thomas, U.S. Virgin Islands (McMillan
et al.,1989). Fish were solicited from local sports
fishermen. Fish were collected by spearftshing.
Participants in an annual fishing toumament were

notified of the need for specimens, and donations
of fish were accepted at the weigh-in station.
Larger fish of suspect species from areas noted
for ciguatera were procured as they have the
highest probability of being ciguatoxic. This ap-
proach was taken knowing that a significant per-
centage of finfish are non-ciguatoxic even from
high incidence areas. Specimens were identified.
weighed, tagged. frozen and air-freighted to the
FDA Gulf Coast Seafood Laboratory (GCSL),.
Dauphin Island, AL. All specimens arrived
frozen with no evidence of thawing or decom-
position. Fish specimens were stored at -20°C.

TEST SAMPLE PREPARATION

Fifly Caribbean fish were selected, including
Sphyraena barracuda (70%), Caranx tatus (20
4), C. hippos (2%), Seriola dumerilt (4%), Seri-
ola rivoliana (2%) and Scomberomerus cavalla
(25%). From each specimen two muscle tissue
samples were taken from the anterolateral part of
the body immediately behind the head: 10g for
use in the SPIA procedure and 450g for extraction
and mouse bioassay. Samples also were taken
from a fillet portion of barracuda (specimen VI-
108: courtesy of Dr. Norbert Mantor, St. Thamas
Hospital) that was responsible for ciguatera
poisoning in St. Thomas and from a Cynoscion
arenarius (white trout: specimen 92-{17-1; not
included in the tables) obtained from the retail
seafood market of Dauphin Island, AL, as.a nega-
tive control, A portion of the latter specimen was
consumed without toxic effect. The 10g samples
were frozen at -20°C. The 450g samples were
weighed to the nearest 0.1g, and extracted and
partitioned (McMillan et al., 1983; Yastumoto et
al.,1984). The dried products from the extraction
and partitioning procedures were dissolved in
chloroform and applied to 60mL (10g) silica gel
solid-phase extraction (SPE) columns that had
been preconditioned by washing with 60mL of
the same solvent. The SPE columns were washed
with an additional 120mL chloroferm and then
eluted with 120mL 10% methanol in chlorofonn-.
The eluting solvent mixtures were removed by
rotary evaporation. and residues transferred in
methanol (c.2—3mL) to tared 13x 100mm Teflon-
capped test tubes. Solvent was removed again by
vacuum centrifugation and residue weights were
recorded to the nearest 0.)mg for each final
product. Final SPE products were dissolved in
methanol {c.1—2mL) and stored at -20°C.

MOouSE BIOASSAYS
Methanol was removed from the muscle Gssue

MEMOIRS OF THE QUEENSLAND MUSEUM

SPE products by vacuum centrifugation, The
residues were redissolved in known total volumes
of saline (0.15M NaCl) containing 1% Tween-60.
pe ine beg aliquots of the saline solutions were

en such that the dosing of mice was equal to
muscle tissue fresh weight equivalents (MTE) of
45, 90 or 180g per mouse. Residue weights ad-
ministered to mice did not exceed 20mg. and
ranged as low us 3])8pg when adjusted to the
chosen MTE. All bioassay solutions were acd-
justed to 4 total injection volume of O.5mL prior
to bioassay. The 0.SmL solutions were ad-
ministered by intraperitoneal injection into Swiss
white mice (Cri:CFW(SWIBR) weighing 18-
21g. Control mice were administered saline-
Tween solution only. The mouse bioassays were
performed in duplicate for controls and for each
MTE of the final SPE products. Mice were ob-
served post-injection for a period of 48hr. Signs
of toxicity were noted and when death occurred,
times from injection were recorded. Mouse bivas-
Says were conducted according to the principles
provided in the Guide for the Care and Use of
Laboratory Animals, Institute of Laboratory
Animals Resources, National Research Council,
NIJH Pub. No. 85-24.

Cicuatect ™ SPtA

A 9x10mm white membrane, affixed to a
90x9mum plastic strip is exposed to muscle tissue
taken from behind the head of the fish. The
membrane is permitted to air-dry and is then
immersed in methanol for no more than Isec.
After the membrane is completely dry, it is im-
mertsed in an anti-ciguatoxin-antibody-latex bead
suspension (gently mixed prior to use) for 5 or
10min. The latter solution is blue. The membrane
is then dipped in and ont of a phosphate-buffered
saline solution 3 times, placed flat on an absorb-
ent towel and gently blotted to remove excess
saline solution. Adsorption of ciguatera biotoxins
to the membrane and subsequent conjugation
with the anti-ciguatoxin antibody bound to the
surface of the blue latex beads should produce 2
blue color (cast) in the membrane affixed to the
plastic strip if the fish contains ciguatera
biotoxins, Kil posiftive-contral membrane strips
are processed from the point of exposure to an-
tibody-latex bead solution. Kit negative-control
membrane strips are processed from the & point of
exposure to methanol, All Ciguatect '™ SPLA
materials are stored at 4°C_

Instructions for the first set of materials
(received 24 May 1992) directed the user to ex-
pose the membrunes to fish tissue only ene time

IMMUNOBEAD ASSAY FOR BIOTOXINS IN CARIBBEAN FISH 483

Specimen | 45g MTE | 45g MTE | 90g MTE | 90g MTE
dose survival dose survival

VI-67

TABLE 1. Mouse bioassay dosing, fish muscle tissue
equivalents (MTE) and survival times. * indicates
dose of 180g MTE administered to mice.

90g MTE
survival

484

Ss ba [REM™| Mouse |

imen duplic-| bioas-

ates* | say

Vi- =i bcd
Vi-I7

So

<
ae
i]
a

1s 0.4 os | 3

Aer tas irae aa eet CN |
rien aetl oo Iola Le ot |
fviss| os | o4 | 33 | ia | ser] T |

[ves] 22 | o4 | 27 | i | sor] n_|
peso | 20 | oo fae {13 {ss} nN

| vis2| 10 | 00 |

2.3
rvess [2s [os lar tori)
pust{ us os fou fos fas {|
oz [os [03
ra fortis
yet sp Loos U=aa | ae [ed Oe
rvea [1s [oe | as | o7}as | on |

on

ivi73| 08 | o« | ao | vo | ss | 7
lvi-7s| 13 | os | 12 | oa | aa | N
vere on [ae T9023 tr |

| vi79| 20 | oo | i3 | os | ss
puis | os | os 1.0 Loo | 4s fn |
vies | 17 | o7 | 20 | 06 | 55 | 7 |
| 25 | os | 55 | T |
Pee ee ae ae Pear trac
[ vido | 03 | os | is | o4 | 33
pwesi{ a3 fas [is | is tas 1
vt | as | 4
04 a {as [a os

VI-ST nop aaa
ris 5 01
Ls

MEMOIRS OF THE QUEENSLAND MUSEUM

Mouse
wile esponse FREM bioas-
ates* | say

os Los [22 Los | ss]

TABLE 2. Ciguatect™ SPIA mean scores (n=5) and
standard deviations (SD) for three variations of pro-
cedure, Boldface values indicate false positive of
false negative scores relative to mouse bioassity
results (T =ciguatoxic; N = survival beyond 48hr), *
indicates data set provided by Dr D.L. Park, Univ. of
Arizona. ftindicales apparent inconsistency with
regard to immunoassay scoring scale.

(single exposure) and to allow the membrane to
dry before proceeding. Weak color development
prompted the manufacturer to modify the proce-
dure for the second set of materials (received 6
October 1992). The latter procedure instructed
the user to expose the membranes by contact with
fish tissue three times (triple exposure) with
drying periods. following each exposure. Conse-
quently, Ciguatect™ immunoassays were per-
fonmed twice on each 10g test sample. The
instructions for immunoassay setup, fish assay
and use of posilive and negative kit controls were
followed precisely. Positive and negative control
test strips (provided with SPIA materials) were
developed prior to fish tissue assays. The SPIA
color developments for controls and fish tissues
were read and interpreted independently by 6
analysts. Participants were asked to.assign a value
of 0 (no color) to 5 (strong color) to color dey-
elopment of each test strip by comparison with a
standard series of positive and negative control
test strips supplied by the kit manufacturer.

The 102 test samples that were used for
Ciguatect’™ SPIA evaluation at the FDA GCSL
were frozen and delivered “blind’ (1.c., each test
pertion was identified only by code number) to
the laboratory of the kit manufacturer fora repeti-
thon of the SPIA procedures.

STATISTICAL ANALYSIS

The Ciguatect™ SPLA method was assessed.
using procedures for evaluating screening tests
(Fleiss, 1981). Method performance is described
through four interrelated rates: sensitivity,
specificity, false positive, and false negative. The
sensitivity rate (P+) is the proportion of correctly

IMMUNOBEAD ASSAY FOR BIOTOXINS IN CARIBBEAN FISH

TABLE 3.Ciguatect™ SPIA performance rates rela-
live to mouse bioassay results. * indicates data set
provided by Dr D.L. Park, University af Anzona.

pa era Single- Triple- REM™

Sper exposure exposure | immuno
immunoassay | immunoassa assay*
PENSITIVITY 19/33 (58%) | 28/33 (85%) | 31/33 (94%)
3/18 (179%) | 4/18(22%) | 2/17 (12%)

FALSE 14/17 (82%) | 5/9 (55%) 2/4 (50%)
NEGATIVE

FALSE | 15/34(449%) | 14/42 (33%) | 15/46 (33%)
|_POSITIVE | __ :

Classified ‘known’ positive test samples. Spec-
ificity rate (P_) is the proportion of correctly
classified ‘known’ negative test samples. False
positive rate (PF,) 1s the proportion of positive
classified test samples that are misclassified
‘known’ negatives. False negative rate (PF_) is
the proportion of negative classified test samples
that are misclassified ‘known’ positives. A
‘known’ positive or ‘known’ negative test sample
is atest Sample that has been classified as positive
or negative by the reference method (i.c., mouse
bioassay). False positive and false negative rates,
us defined by Fleiss. are predictive rates. These
were used to assess the Ciguatect™ SPTA method
when used on populations of fish with different
proportions of ciguatoxic specimens,

RESULTS AND DISCUSSION

MOUSE BIOASSAY

Characteristic signs of ciguatoxicity (Hoffman
el al.,1983; Kimura et al.,1982) and death within
48hr were chosen as the determinants for the
presence of ciguatera-related biotoxins. The first
three fish SPE products bioassayed were ad-
ministered at doses of 180g MTE. Survival times
were short for the three products (S0.36hr). All
subsequent bioassays were conducted at an MTE
of either 90 or 45g in order to record the presence
er absence of characteristic signs of cigua-
Joxicity, The mouse bioassays resulted in 33/51
(65%) mortalitics (Table 1). A total of 79 pairs of
mice were injected (duplicate bioassay for cach
MTE). In 10 of the 79 pairs 1 mouse expired
within 48hr, and the other survived beyond 48hr.
Signs of ciguatoxicity in mice included inac-
tivity; piloerection; vasodilation in cars; cyanosis
of tatl, feet and muzzle; lacrimation: salivation;
tharthea; dyspnea; unsteady gait; temor/conyul-

sive jumping; straub tail; convulsions; and death.
Not all mice displayed each of these signs of
toxicity. Some characteristic signs of toxicity
were not observed in mice that expired within Lhe
(15/33: 48%). In 50% of short survivals the
dosage was reduced to 45g MTE (lowest dose
level administered) and longer survival times or
survivals oor 4 48hr were noted with charac-
tenstic signs of toxicity, The mouse bioassay of
specimen VI-LO8 resulted in survival beyond
48hr at 452 MTE, survival times of 0.83 and 1.83
al 90g MTE, and survival times of 0.35 and (),22
at 180g MTE. Specimen 92-07-1 resulted in sur-
vival beyond 48hr at 180g MTE without any sign
of toxicity. Fish specimens from which cnr-
responding SPE products produced mouse death
with characteristic signs of toxicity at (g MTE
were considered ciguatoxic

CrcuaTect™ SPLA

SPIA eatings of the fish specimens were scaled
from 0 to 5 by comparison of color development
with a standard senes of pre-developed mem-
branes supplied by the manufacturer, The
manufucturer specified that ratings of 0 or 1 were
to be considered non-ciguatoxic and ratings
above 1 through 5 ciguatoxic at progressively
ereater levels. The means and standard deviations
of ratings for SPIA responses were recorded
(Table 2). However, for comparison with mouse
bioassays the specimens were scored as either
posilive (ciguatoxic) or negative (non-cigua-
toxic) by SPIA. Results from single exposure of
test-strip membranes to muscle tissues indicated
that 34/51 (67%) of fish specimens were
ciguatoxic. Based upon the precedent (McMillan
etal., 1983; Yasumoto etal..1984; Hoffmanetal,,
1983; Kimura et al.,1982; Vernoux et al., 19845)
that the reference mouse bioassay accuritely
reflects ciguatoxicity, results of the 3 sample
treatment methods with the immunoassay technt-
que are shown in Table 3. Single-exposure SPIA
scored the human illness case sample (specimen
VI-108) negative (mean 0.5, SD=0.5) and the
non-toxic white tout specimen positive (mean
2.7, SD=0.7), The SPIA results following triple
exposure of test-strip membranes to muscle tis-
sues indicated that 42/51 (82%) of the fish
specimens were ciguatoxic. Triple-exposure
SPIA scored V1-]08 (mean 2.2, SD=0.4) and the
white trout specimen (mean 4.0, SD=0,5) bath
positive. Mean scores for the kit internal centrels
for single exposure of test strips included 2/2
positive matches and 2/2 negative matches.
Scores for kit internal controls for triple exposure

486

TABLE 4.Predictive indices of Ciguatect™ perfor-
mance, given ciguatoxins contamination rates from 5
to 75%. PF; = false positive; PF. = false negative.

Predictive index
[Single exposure | Triple exposure] __REM™ _|

|_5 | 0.9649] 0.1181 | 0.9457 | 0.0346 | 0.9486 | 0.0264 |
0.5284 | 0.3581 | 0.5426
| 55 | 0.5422 | 0.7566 | 0.4286 | 0.4545 | 0.4426 | 0.3864
0.4380 | 0.8254 | 0.3305 | 0.5587 | 0.3432
[0.3254 | 0.8842] 0.2340 | 0.6716 | 0.2444 [0.6071

of test strips included 2/2 positive matches, 1/2
negative matches and 1 false positive (mean of
1.2 on 5-point scale). False negatives were not
observed among the kit control test strips.

The Ciguatect ™ SPIA results obtained by the
manufacturer’s laboratory for the fish test
samples evaluated by FDA GCSL do not cor-
respond to results obtained by GCSL. The
manufacturer laboratory first performed extrac-
tion and partitioning procedures (REM™ :rapid
extraction method) on each test sample. The
SPIA was then performed on the REM™
products (Tables 2, 3). SPIA results obtained by
following the REM*™ modification of the proce-
dure indicated that 46/50 (92%) of the fish
specimens were ciguatoxic. VI-108 scored posi-
tive (duplicate scores of 5 and 5).

The sensitivity and specificity test performance
rates for each variation of the SPIA procedure
(Table 3) were used to evaluate how the
Ciguatect’™ SPIA would be expected to perform
when used on fish populations with different
proportions of ciguatoxic specimens. Table 4
presents a summary of the predictive values for
each SPIA method by population ciguatoxicity
rate. Assuming that a true ciguatoxins contamina-
tion rate of 55% is encountered in a hypothetical
lot of tropical fish, the predictive indices gener-
ated from the present study indicate that single-
exposure Ciguatect™ SPIA would produce a
false negative rate of 76% and a false positive rate
of 54%. Triple-exposure Ciguatect™ SPIA
would produce a false negative rate of 45% and
a false positive rate of 43%, and the REM
Ciguatect™ SPIA would produce a false nega-
tive rate of 39% and a false positive rate of 44%,

The Ciguatect™ SPIA was examined for inter-
ference from potential fish decomposition

MEMOIRS OF THE QUEENSLAND MUSEUM

products and for possible non-specific toxicity
unrelated to ciguatera in the mouse bioassay. A
repeat sampling of 13 muscle tissues (25%) from
the original 51 fish provided a cross-section of
tissue that were SPIA positive and negative,
mouse bioassay positive and negative, and SPIA
false positive and false negative. The tissues were
analyzed for putrescine and cadaverine by the
method of Staruszkiewicz & Bond (1981) and for
histamine by the fluorometric method (AOAC,
1990). Determination of putrescine was not pos-
sible because of matrix interference. Cadaverine
levels were below the lowest calibration standard
of 0.5p.g/g in 10 of the 13 tissues. Three of the
tissues contained 3.7, 7.2 and 8.4.2/g cadaverine
(Table 5). Cadaverine levels above 6g/g in
tunafish indicate decomposition. Tolerance
levels for other species of fish (including those
used in the present study) are not established.
Histamine levels did not exceed 0.1mg% in 11 of
the 13 tissues, and did not exceed 0.2mg% in the
two remaining tissues. The defect action level for
histamine in the United States is 20mg%, and
50mg% poses a human health hazard. There was
no correlation between cadaverine or histamine
and SPIA or mouse bioassay findings.

Thirteen tissue extract products (corresponding
to the specimens analyzed for decomposition
products) were evaluated for sodium channel ac-
tivity by using a tetrazolium-based neuroblas-
toma cell bioassay for neurotoxins active on
sodium channels (Manger et al.,1993). Sodium
channel potentiating activity is indicative of the
ciguatoxins. The bioassay indicated that sodium
channel activity was present in all 13 tissue
products and provided a general ranking of
sodium channel activity for those tested (Table
5). The activity ranking correlated well with
mouse bioassay survival times. Specimen extract
products shown to be highly toxic by the mouse
bioassay also produced significant reductions in
cell viability through sodium channel potentiat-
ing effects. Inhibitory doses which reduced cell
viability by 90% after 22 hr exposure (IDo0)
ranged from 2 to 6mg MTE per 200microliter cell
culture (96L well format). Specimen extract
products classified non-toxic by the mouse bioas-
say were found to possess sodium channel effects
at [Doo dosing that ranged from 7 to 40mg MTE.
The latter finding, if considered in conjunction
with mouse bioassay results from the ciguatera
case specimen V]-108 (i.e., mouse survival at 45g
MTE and expiration at 90g MTE), suggests that
utilization of 90g MTE in the mouse bioassay
may approximate a correlate for human toxicity.

IMMUNOBEAD ASSAY FOR BIOTOXINS IN CARIBBEAN FISH

487

TABLE 5. Sodium channel activity of specimen solid-phase extraction products (ID99 =inhibitory dose which
reduces cell viability by 90% at 22hr exposure). Values expressed in milligram muscle tissue equivalents per

well; mouse bioassay survival times; Ciguatect

SPIA scores and decomposition indicators. * indicates data

provided by Dr D.L. Park, University of Arizona. + indicates apparent inconsistency with regard to scoring
scales. # indicates dose of 45g MTE administered to mice.

>48

1.8(0.4)

Specimen Sodium 90g MTE Ciguatect™ score Decomposition products
channel mouse : . A . :
‘ Single- Triple- REM* Cadaverine Histamine
D
MTEwell) |" cn) | SxRRSUE | exposure a ia
VI-13 >40 >48 1.5(0.5) 0.2(0.4) 3,3 <0.5 0.1
>48
VI-59 32 >48 2.2(0.4) 2.7(1.1) 5,67 <0.5 0.1
>48
VI-63 22 >48 2.5(0.5) 2.7(0.7) 1,1 <0.5 0.1
>48
VI-32 20 >48 1.8(0.4) 3.8(1.2) [ 3,3 <0.5 0.1
>48
VI-81 9 >48 0.8(0.4) 1.0(0.0) 45 7.2 0.2

2.5(0.5)

0.8(0.4)

0.0(0.0)

CONCLUSION

Ciguatect™ SPIA performance with cigua-
toxic Caribbean finfish may be characterized by
low specificity rates and high false positive and
false negative values. Extrapolating these perfor-
mance characteristics to a market situation im-
plies that a proportion of wholesome fish might
falsely be identified as ciguatoxic and an equally
significant proportion of ciguatoxic fish might
reach the marketplace undetected. Conclusions
of the present study cannot be extended to
Ciguatect™ SPIA performance with Pacific
Ocean finfish (where the immunoassay
originated) without separate evaluation of the
method using fish from the Pacific region.

ACKNOWLEDGEMENTS
We thank GCSL staff, Mr Don Mowdy, Mr Ed

0.2(0.4) 0.3(0.5)
>48
VI-53 2 2.50 1.8(0.4) 2.7(1.1) 3,3 8.4 0.2
ae = 6.40 =~ I

VI-85 2 1.70.7) 2.0(0.6) 5,5 <0.5 0.1
VI-77 2 1.7(0.7) 1.2(0.4) 3,3 <0.5 0.1
0.5(0.5) 3.7(0.5) 5,5 <0.5 0.1

| 1.0(0.0) 2.7(1.2) 5,5 3.7 0.1

Jester, Ms Della Clausen, Dr Susan McCarthy, Dr
Antony Guarino and Ms Kathy El Said for assis-
tance in preparation of extracts, and reading of
results. For assistance in fish collection we thank
Mrs Ann Seiler and staff of the Division of Fish
and Wildlife, St. Thomas, U.S. Virgin Islands, Mr
Louis Greaux, President of Frenchtown-St.
Thomas Fishing Tournament, and Mr Mark
Marin and Mr Sid Smith of the St. Thomas Spear-
fishing Club. Appreciation is extended to Drs
Ron Manger, Jim Hungerford and Marleen
Wekell, Ms Linda Leja and Ms Sue Lee, Seafood
Products Research Center, FDA, Bothell, WA,
for performing the bioassays for Na channel ac-
tivity.

LITERATURE CITED

AOAC, 1990. ‘Official methods of analysis’, 15th ed.,
AOAC Intemational, Arlington, VA, section

48s
Y77_L3: Histamine in seafood, Muorometric
method, pp, 876-877.

FLEISS, J.L. L. ‘Statistical methods for rates and

proportions’, 2nd ed. (Wiley & Sons: New York).

FUKUI, M., MURATA, M., INOUE, A., GAWEL, M.
& YASUMOTO, T. 1987, Occurrence of
palytoxin in the iigeertish Melichtys vidua.
Toxicon 25; 1121-1124,

HOFFMAN, P.A,, GRANADE, H.R. & McMILLAN,
J.P. 1983. The mouse ciguatoxin bioassay: a dose-
response Curve and symptomatology analysis.
Toxicon 21: 363-369.

HOKAMA, Y. 1985. A rapid. simplified enzyme im-
munoassay stick test for the detection of
ciguatoxin and related polyethers from fish tis-
sues, Toxicon 23: 939-946,

HOKAMA, Y_ 1990. Simplified solid-phase im-
munobead assay for detéction of ciguatoxin and
related polyethers, Journal of Clinical Laboratory
Analysis 4; 213-217.

HOKAMA, Y., ASAHINA,. A.Y,, HONG, T.W,P.,
SHANG, ES. & MIYAHARA. J.T. 194),
Evaluation of the stick enzyme immunoassay in
Caranx sp. and Seriola dumerili associated with
Ciguatera. Journal of Clinical Laboratory Analysis
4: 3632-366.

HOKAMA, Y., HONG, T.W.P.. ISOBE, M.,
ICHIKAWA, Y. & YASUMOTO, T. 1992.
Cross-reactivily of highly purified okadaic acid
(OA), synthesic, spiroketal east sphere of OA and
ciguatoxin. Journal of Clinical Laboratory
Analysis 6: 54-58.

HOLMES, M.J., LEWIS, R.J, & GILLESPIE N.C.
1990, Toxicity of Australian and French Poly-
nesian strains of Gambiendiscus texicus Dino-
phyceae) grown in culture: characterization of a
new type of maitotoxin. Toxicon 28: 1159-1172.

KIMURA, L.H., HOKAMA, Y., ABAD, M.A.,
OYVAMA, M. & MIYAHARA, LT, 1982. Com-
parison of three different assays for the assessment
of ciguatoxin in fish tissues: radioimmunoassay,
mouse bioassay and in vitro guinea pig atnum
assay. Toxicon 20: 907-912,

LEWIS, RJ, SELLIN, M., POLI, M.A,, NORTON,
R,S., MACLEOD, J.K. & SHEIL, M.M. 1991,
Purification and characterization of cigualoxins
from moray eel (Lycodontis javanicus,
Muraenidac). Toxicon 29; 1115-1127,

MANGER, R.L., LEJA, L.S., LEE, S.Y., HUNGER-
FORD, J.M. & WEKELL, M.M. this memoir.
Detection of ciguatoxin, brevetoxin, and saxitoxin
by cell bioassay.

McMILLAN, J.P., GRANADE, H.R. & HOFFMAN.
P.A. 1983. Ciguatera fish poisoning in the United
States Virgin Islands: preliminary studies. Journal
of the College of the Virgin Islands 6: 84-107.

MURAKAMI, Y,, OSHIMA, Y. & YASUMOTO, T.
1982. Identification of okadaic acid as a toxic
component of a marine dinoflagellate
Proroceatrum Jima, Bulletin of the Japanese
Society of Scientific Fishenes 4%: 69-72.

MEMOIRS OF THE QUEENSLAND MUSEUM

MURATA. M., TWASHITA, T., YOKOYAMA, A.,
SATAKI, M. & YASUMOTO, T. 1992, Panial
structures of maitotoxin, the most potent marine
toxin from dinoflagellate Gambierdiscus toxteus,
Joumal of the American Chemical Society 114:
6594-8596.

MURATA, M., LEGRAND, A.M., ISHIBASHI, ¥.,
FUKUI, M. & YASUMOTO, T, 1999. Stractures
and configurations of ciguatoxin Erom the moray
eel Gymuathorax javanicus and its likely precur-
sor from the dinoflagellate Gamblerdiscus
loxtews. Journal of the American Chemical
Society 112; 4380-4386.

MURATA, M., LEGRAND, A.M. ISHIBASHI, Y, &
YASUMOTO, T. 1989. Structures of ciguatoxin
and its congener. Journal of the American Chemi-
cal Society 111: 8929-8931,

MURATA, M,, NAOKI, H., IWASHITA, T,, MAT-
SUNAGA, S., SASAKI, M., YOKOYAMA, A.
& YASUMOTO, T. 1993. Structure of
maitotoxin, Journal of the American Chemical
Society 115: 2060-2062.

NAGAI, H., TORIGOE, K., SATAKE, M., MURATA,
M. & YASUMOTO, T, 1992, Gamberic acids;
unprecedented potent antifungal substances iso-
laled from cultures of a marine dinoflagellate
Gambierdiscus toxicus, Journal of the American
Chemical Society 114: 1103-1105.

PARK, D.L., GAMBOA, P.M. & GOLDSMITH, C.H.
1992. Rapid facile solid-phase immunobead assay
for screening ciguatoxic fish in the market place,
Bulletin de la Societe de Pathologie Exotique 85:
504-507,

STARUSZKIEWICZ, W.F.Jr & BOND, IF. 1981, Gas
chromatographic determination of cadaverine,
Pulresciné, and histamine in foods. Journal of the
Association of Official Analytical Chemists 64:
584-591,

TORIGOE, K,, MURATA, M,, YASUMOTO, T. &
IWASHITA, T, 1988, Prorocentrolide, a toxic
nitrogenous macrocycle from a marine dinaflagel-
late, Prorocentrum lima. Journal of the American
Chemical Society 110: 7876-7877.

VERNOUX, J.P., LAHLOU, N., ABBAD EL
ADALOUSSI, §., RIYECHE, N. & MAGRAS,
L.P. 1985. A study of the distribulion of
ciguatoxin in individual Caribbean fish. Acta
Tropica 42: 225-233.

YASUMOTO, T., RAJ, U.. BAGNIS, R., INOUE, A.,
KAMIYA, H,, OSHIMA, Y¥,, FUKUYO, Y. &
KOTAKE, Y, 1984, ‘Scafood poisonings in tropi-
cal regions’. (Symposium on Seafood Toxins
Tropical Regions: Lab Food Hygiene. Faculty
Agne,, Tohoku Univ.) 74p,

YOKOYAMA, A., MURATA, M., OSHIMA, Y_,
IWASHITA, T. & YASUMOTO, T. 1988. Some
chemical properties of maitotoxin, a putative cal-
cium channel agonist isolated from a marine
dinoflagellate. Journal of Biochemistry 104: 184—
187.

ASSESSMENT OF CIGUATERIC FISH IN HAWAII BY IMMUNOLOGICAL, MOUSE
TOXICITY AND GUINEA PIG ATRIAL ASSAYS

YOSHITSUGI HOKAMA, AUDREY Y. ASAHINA, ERIC TITUS, DANA ICHINOTSUBO, STEWART
CHUN, TERRI-LYNN W.P. HONG, JULIE L. SHIRAI, DAISY A, ASUNCION AND JAMES T.

MIYAHARA

Hokama, Y., Asahina, A.Y., Titus, E., Ichinotsubo, D., Chun, S., Hong, T.-L.W.P., Shirai,
J.L., Asuncion, D.A. & Miyahara, J.T. 1994 08 01. Assessment of ciguateric fish in Hawaii
by immunological, mouse toxicity and guinea pig atrial assays. Memoirs of the Queensland
Museum 34(3): 489-496. Brisbane. ISSN 0079-8835.

Ciguatera studies were determined at the Waianae Boat Harbor, Oahu when 12 or more
individuals became ill after eating freshly caught mullet (Mugil cephalus, amaama) in
January—March,1991. Typical clinical manifestations of ciguatera were shown by the
patients. Immunological assay for ciguatoxin and polyethers with monoclonal anti-
ciguatoxin (MAb-CTX) showed 80% of the mullet to be borderline to positive. The
herbivores, Ctenochaetus strigosus (kole), Acanthurus sandvicensis (manini) and other
Acanthurus sp. (palani), showed high levels of toxins. The mackerels showed little or no
toxic levels, while the carnivores (jack, amberjack) showed borderline to positive toxicity.
Abundant green algae (Bryopsis), 30-60cm below the seawater surface, found at all five sites
examined, contained Gambierdiscus toxicus in moderate numbers. At 2 sites, when Bryopsis
disappeared (summer - early winter), no Gambierdiscus toxicus was found. Fish extracts of
mullet and other herbivores (palani, manini, kole-surgeonfishes) were highly toxic to mice.
Guinea pig atrium analysis of the wild Gambierdiscus toxicus and fish extracts showed
typical ciguatoxin-like inotropic response strongly inhibited by tetrodotoxin.

Yoshitsugi Hokama, Audrey Y. Asahina, Eric Titus, Dana Ichinotsubo, Stewart Chun,
Terri-Lynn W.P. Hong, Julie L. Shirai, Daisy A. Asuncion and James T. Miyahara, John A,
Burns School of Medicine, University of Hawaii, Honolulu, Hawaii, 96822; received 2

February, 1994.

In January—March,1991, an outbreak of
ciguatera poisoning occurred among individuals
eating mullet (Mugil cephalus, amaama) from
Waianae Boat Harbor. Approximately 12 people
exhibited classical symptoms of ciguatera
poisoning (Hokama,1988) with an older couple
hospitalized for several days. Based on
knowledge of the ciguatera food chain (Randall,
1958; Hokama et al.,1986; Hokama et al.,1993),
a systematic study of Waianae Boat Harbor was
carried out and included:

1.Examination of algae species;

2.Examination ‘of ciguatoxin producing Gam-
bierdiscus toxicus; 3.Immunological testing of as
many species of fish from the boat harbor;

4.Analysis of alga and G. toxicus chemical
extracts for ciguatoxin by mouse toxicity bioas-
say and by the guinea pig atrium assay;

5.Analysis of fish extracts using the mouse
toxicity and guinea pig atrium assays; and

6.Water quality data (from the Department of
Health).

Basis for this systematic assessment is that of
Yasumoto et al., (1979,1980,1984) and Hokama
et al. (1993).

METHODS

SURVEY AREA

The Waianae Boat Harbor (117057m?) in-
cludes 4 ramps and 3 docking areas for boats.
Two water inflow outlets are the entrance channel
to the ocean, and a tunnel located between the
boat docks (fresh water and seawater runoff).
Five areas are surveyed within the boat harbor
(Fig.1).

ANALYSIS FOR G, TOXICUS

Algae specimens (0.5kg) were collected in a 5
litre ziplock bag containing 1 litre of seawater.
The contents were shaken for 2 minutes to loosen
epiphytic dinoflagellates from the alga. The salt
water—algal suspension was passed through a
125j.m sieve to remove larger algal fragments
and then through a 371m sieve. This residue was
backwashed with a filtered seawater media, trans-
ferred into a 100m! sterile glass bottle and capped
loosely to provide aeration. After gently shaking
the algal sample bottle, 1.0ml was removed and
transferred onto a Sedgewick Rafter Cell Count-
ing Slide. Cell counts were carried out in tripli-

490

PARRINGTON

WAIANAE

MEMOIRS OF THE QUEENSLAND MUSEUM

HIQNUWAT

REGIONAL

~ TH.
-
“Ses-ne-nae
weed ewe —=\

0

metres

FIG. 1. Map of the survey area of Waianae Boat harbor (WBH). Tunnel is in the area of Section 2.

cate and the average number of cells determined
per millilitre.

COLLECTION OF FISH

Fish samples (Table 4) were taken by divers
mainly using spear-guns but also nets and lines
and were identified from Tinker (1982).

FISH GUT SMEARS

An incision was made in the belly of the fish,
in particular mullet (amaama), surgeonfish (kole)
and Sandwich Island surgeon fish (manini) to
extract gut contents which were mixed with 0.1 ml

of 0.85% NaCl and one drop was smeared onto a
slide. The mixture was air dried and analyzed for
G. toxicus with a microscope (Zeitz) at x400.
Presence of other algal fragments were noted.

STICK ENZYME IMMUNOASSAY (S-EIA)

The S-EIA procedure comprised: a, make in-
cision in fish tissue; b, insert skewered liquid
paper-coated end of bamboo stick into flesh; c,
air dry coated end of stick; d, immerse coated end
of stick into absolute methyl! alcohol for 0.5 sec;
e, air dry; f, immerse in 1ml monoclonal anti-
ciguatoxin-horse radish peroxidase (MAb-CTX-

ASSESSMENT OF CIGUATERIC FISH IN HAWAII

TABLE 1. Mouse toxicity assay scoring

Description of Visible Clinical Symptoms in
Mouse After Extract Injection
No ill effect

15-60 min: muscle contraction in lower back
area (flexion), increased respiration, immobile

Recover in 12-24 hours: same as 2, muscle
contraction, paralysis in extremities (usually
hind legs), rapid and irregular breathing,
immobile, closed eyes, pilo-erection, slight

cyanosis (tail).

Symptoms as in 3, but death within 24*-48 hrs.
_|_ Symptoms of 3 and 4, death in less than 6 hrs.

*] mouse unit, death within 24hrs of 20gm mouse;
contains 7-9ng ciguatoxin in the sample of 100mg of
crude extract/mouse injected (IP) (Hokama, personal
estimation from Department of Health confirmed
ciguateric fish extracts).

HRP); g, wash end of coated stick twice in buf-
fered saline; h, immerse stick in horse-radish
peroxidase (HRP) substrate; i, score color inten-
sity of substrate as previously reported
(Hokama, 1988; Hokamaetal.,1989). Final color:
Q—1.2 is scored as negative, 1.3—1.9 borderline
and 2.0 as positive. All fish scoring 1.3 are
reported as rejections (not edible). S-EIA values

49]

of 1.3 generally represent 0.4ng ciguatoxin or
related polyether per gram of fish tissue.

MOUSE BIOASSAY (Kimura et al.,1982)

Swiss Webster mice weighing 20—25g were
utilized in this study to assess the toxicity of fish
extracts. 100mg of crude fish extract was
resuspended in 1 ml of 1% Tween 60 in saline and
injected intraperitoneally (IP) into mice.
Symptoms displayed by the mouse were ob-
served at 30min, 1, 2,4, 6, 8, 24 and 48hours after
injection and rated on a scale of 0—S5 according
to toxicity (Table 1). One mouse unit equals
7—9ng of ciguatoxin per 100mg of crude extract
which kills a mouse within 24hours. This is based
on our estimation of ciguatoxin from known
cases of ciguateric fishes obtained from the
Department of Health.

GUINEA PIG ATRIAL ASSAY (Miyahara et al.,
1989)

100mg of each fish extract was resuspended in
Iml of Krebs-carbonate solution. 1001 of the
suspension was tested on the guinea pig atria.
Subsequent inotropic and chronotropic actions
were noted in addition to its pharmacological
response to TTX (tetrodotoxin), verapamil and
the adrenergic blockers (propranolol and phen-
tolamine). The inhibitors, TTX, verapamil and
adrenergic blockers, were given after the in-

TABLE 2.WBH Water Quality Physical Analysis and G. toxicus in 1992 (Bloom Year’s Analysis).

— Sections*
1 2 3 4 5
Temp(°C) | meantS.D. | 263+ 0.85 | 262408 26.3+0.8 26.2+0.7 26.1+0.7:
range 25.1-27.2 25.1-27.2 24.9-27.2 25.1-27.0 25.1-26.9
pH mean +S.D. 8.1402 | 81+02 8.1+0.3 8.140.2 8.1+0.2
range 8.0-8.6 | 8.0-8.6 7.9-8.6 7.9-8.6 7.9-8.4
DO (mg/l) | mean+S.D. 6.2413 6.11.7 6341.6 5.9413 5920.8
range 5.2-8.6 4.8-.5 4.1-9.4 4.3-8.6 5.0-7.1
Salinity (%) mean +S.D. 3.5+0.4 3.520.3
3.5-3.6 3.5 3.4-3.5
Conductance 53.620.3 53.6+0.4
range 52.5-54.0 52.4-54.0 52.9-54.2
G.toxicus | mean+S.D. | _23.4426.4 10+10.9 18.32@37.0
(cells/gm range 0.9-65.6 0-26.5 0-120.3 1,3-90.7 0-268.9
weight of alga)

“1.No relationship of physical properties of seawater to G. toxicus growth.
2.All values are means+ S.D. for 9 months evaluation of each section (1 to 5).
3.The salinity in % of WBH generally below 4.0% maximum of seawater is due to the underground fresh water

along the Leeward coast of Oahu.

492

otropic responses at 12.541] of a 10°M concentra-
tion.

WATER QUALITY

Throughout Sections 1-5 (Fig.1), the salinity,
temperature, pH, dissolved oxygen and conduc-
tance were measured utilizing a Surveyor 2
(Hydrolab Corp., Austin, Texas) (Table 2).

RESULTS

G. TOXICUS — DATA FOR 1992

Bryopsis growth was most prominent in Sec-
tions 3, 4 and 5 throughout the years. Bryopsis
was generally sparse at Sections 1 and 2, between
which the rainfall and wash water outlet is
present. If non seawater enters, it is expected to
lower the salinity and may interfere with algal
growth. However, the lowering of salinity has not
been observed in Sections 1 or 2 (Table 2). G.
toxicus cell counts were best in all Sections on
June 25,1992, especially Sections 3 and 5. Oc-
tober 29,1992 also showed moderate to high
counts of G. toxicus in Sections 4 and 5.

It is suggested that constant presence of Bryop-
sis and G. toxicus helps to maintain the high
toxicity level of herbivore fishes in the WBH
(Table 4).

GUT SMEAR ANALYSIS

Gut smear analysis by microscope at x400 mag-
nification revealed on several occasions, G.
toxicus in smears of Mugil cephalus (mullet).
This involved 3 samples of fish gut.

TABLE 3.Number of Gambierdiscus toxicus/gm of
Alga in Waianae Boat Harbor at Various Time
Periods and Stations

Station Number*

Dates 1 2 3 4 5
(1992)
Mar. 27 1,4 2.1 0.0 3.2 4.0
Apr. 28 6.6 27.0 9.8 4.0 27.0
May 29|__0.9 0.0 0.0 | 1.3 0.8
Jun. 25 | 15.0 20.0** | 120.0 10.0 269.0
Jul. 24 NA NA 3.2 | 2.5 8.0

2.7

“NA, no Bryopsis available
“20 cells/gm Ulva, not Bryopsis

MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 4.Stick Enzyme Immunoassay (S-EJA)
Evaluation of the Fish in Waianae Boat Harbor for
Ciguatera (1992) _

No [ 22 [isisf <2 |

é
line

Lis] 4 | 12 | 2 |

Ctenochaetus 4 12 2
(22.2) | (66.7) | (11.0)
53_ | 33

hawaiiensis
(Hawaiian kole)
Acanthurus

sandvicensis
(Manini)

Mugil cephalus
(amaama, mullet)

Kuhlia sandvicensis
(aholehole)

Savotherodon sp. 10 1 8 1
(Tilapia)

Acanthurus
dussumieri (palani)

Abudefduf
abdominalis (mamo)

IRE
pos |
fo | i |

Trachiurops
crumenophthalmus
(akule, halalu)

Caranx sp. (papio)

Mulloidichthys
auriflamma (weke)

TOTALS

[319 | 109 | 127 | 93_|
_| | 642) | G28 | 26.0) |

*positive plus borderline represent 75% of fish rejected
(non-edible)

STICK ENZYME IMMUNOASSAY OF FISH

There is a high rejection rate of Mugil cephalus
(Table 4) which is the species that caused the
outbreak of ciguatera poisoning in the early
months of 1992. Halalu (Trachiurops crumen-
ophthalmus) was mostly negative (Table 4), and
has never been implicated in ciguatera poisoning.
This species has been continuously caught and
eaten throughout 1991 and 1992 with WBH with
no known incidence of ciguatera poisoning.
Aholehole has shown a large number in the rejec-
tion category (Table 4), but no known toxicity has
been reported. All the other species are not
evaluated individually because of the numbers.
Evaluation of all fish (67 in 1991) suggests a high
prevalence of fish in the rejection category (bor-
derline plus positive = 77.6%). This may be due
to the continuous presence of G. toxicus in the

ASSESSMENT OF CIGUATERIC FISH IN HAWAII

Mullet Flesh (pooled)

/ |

100 pt Tm

(100 mp/enl) 12.5 ut 12.5 12.5 ysl
gosM) = (t0-S My) (0S M)
A B c BD £

Mullet Gut

Mailet Flesh

Mallet Gut

A E

FIG, 2. Guinea pig atrium assay with two mullet flesh
and gut samples from WBH, 1992. 1-R; A, S-EIA
positive pooled mullet flesh extract: B, addition of
TTX; C, addition of verapamil; D, addition of
adrenergic inhibitors; and EB, rinse of physiobath with
medium. 2-L; A, S-Bl1A positive pooled mullet gut
extract; B, TTX; C, verapamil; D, adrenergic in-
hibitors; and E, rinse. 3-R; A, S-ELA positive pooled
mullet flesh extract; B, TTX; C, verapamil; D,
adrenergic inhibitors; and E, rinse, 4-L; A, S-ELA
positive pooled mullet gut extract, B, TTX; C,
verapamil; D, adrenergic inhibitors; and E, rinse.
R=nght atrium; L=left atrium,

enclosed WBH, which has only one outlet to the
ocean, The non-toxicity of halalu suggest the
significance of the food consumption habit and
source of this species is different from the her-
bivores feeding on algae (Bryopsis).

S-EJA data for 1992 (Table 4) show rejection
rates in descending order: mamo (100%), Caranx

sp. (100%), tilapia (90%), Hawaiian kole
(87.9%), manini (78.2%), aholehole (70.6%),
mullet (69%) and palani (62%). The samples of
mamo, weke and Caranx sp. are too smal) to
evaluate, though the high value of toxicity in
Caranx sp. may be significant.

Mullet flesh and gut were all of high toxicity in
all samples collected, except mullet flesh from
4/92. Surprisingly, extracts from negative mullets
from 6/92 and 9/92 were highly toxic to mice.
Similarly, Hawatian koles were all highly toxic
with mouse assay values of 5+. Manini flesh and
gut extract showed variable results, although the
majority of samples were highly toxic in mouse
(5+). The 5/92 samples of Caranx gave a high
toxicity value in the gut (5+) and a low toxicily
value (2+) in the flesh. This is generally the
pattern when examining flesh and gut separately
in carnivores. Palani appeared to be of high
toxicity in both flesh and gut extracts for samples
on 3/92 and 6/92. Mamo and aholehole extracts
gave comparable results with the flesh extract
showing low toxicity (2+) and the gut extracts
high toxicity (5+). Of interest is the high toxicity
level of whole tilapia extracts obtained from
mostly borderline tested fish in the S-ETA.
Whether this is truly ciguatoxin (congeners) or
other polyethers remains to be determined for the
tilapia. The highly toxic gut extracts were essen-
tially non-cigualoxin-like in the mouse assay.

GUINEA PIG ATRIAL ASSAY

The results of the mullet tissue analysis (Table
5, Figs 2,3) suggest ciguatoxins as noted by the
consistent inhibition of the inotropic effect by
tetrodotoxin (TTX). See also the inotropic pat-
terns presented for the mullet tissue extracts
(Fig.2), In some cases, verapamil (verap.) also
shows the inhibition of the Ca** ion suggesting a
maitotoxin-like toxin. The Hawaiian kole ex-
tracts showed variable inotropic response unlike
the mullet, but with slight inhibition by TTX and
verap. The manini flesh and gut extracts gave
moderate inotropic responses with inhibition by
both TTX and verap, This suggests possible
ciguatoxin and maitotoxin-like toxins in these
extracts. Though not shown here, a new toxin
with sodium channel inhibition was observed in
the guinea pig atrium assay. This toxin(s) was
unlike TTX or PSP in solubility.

DISCUSSION

The Waianae Boat Harbor appears to be an
excellent mode! for the solution of possible cn-

494

MEMOIRS OF THE QUEENSLAND MUSEUM

Intact Bryopsis (green algae)

Atrium

I-R

Washed Bryopsis (green algae)

+L

A A

B

FIG. 3. Ciguatoxin in Gambierdiscus toxicus ciguatoxin-like toxins or no ciguatoxin in green and red algae. 1-R;
A, 1001 of a 100mg/ml solution of intact Bryopsis acetone extract; B, addition of TTX (12.0pl of a 107 M
solution); C, addition of verapamil (12.0.1 of a 103 M solution); D, addition of a 12.0.1 of a 10° M solution
each of phentolamine and propranolol; and E, rinse with medium. 2-L; A, 100.1 of a 100mg/ml solution of G.
toxicus extract of a 22mg/ml solution; B, TTX added; C, verapamil added; E, rinse; F, return of inotropic
response initiated by A. This indicates the CTX in G. toxicus extract remains tightly bound to atrium and the
inhibitor (TTX) is removed by rinse. 3-L; A, 100p] of washed Bryepsis added (concentration 79 mg/ml, B,
TTX added; C, verapamil; D, adrenergic inhibitors; and E, rinse. Concentrations added same as in 2-L. 4-L; A,
100p1 added (50mg/ml solution); A, second addition of A; TTX added, same concentration as |, 2, and 3.

R=right atrium. L=left atrium,

vironmental control of G. foexicus and hence
possible control of the fish poisoning problem
associated with ciguatoxin and other polyethers:
for example, understanding the growth stimulant,
displacement of algae favourable to G. toxicus
growth with algae inhibitory to G. foxicus, and
control of the growth factors coming from the
inlet tunnel into the harbor. In Waianae Boat
Harbor, this means 1, diverting the tunnel outside
the harbor to prevent rainfall bringing soil ex-
tract; 2, determination of the life cycle of G.
toxicus; 3, inhibiting by appropriate compounds

the weak link of the G. toxicus life cycle and; 4,
assessing the toxic fish level and variability in
toxicity. But thus far, the data reveal no biological
cyclic effect. This probably results in the con-
tinuous presence of G. toxicus throughout the
year. The food chain concept has been verified in
this study: G. toxicus growth near Bryapsis is
eaten or taken in by mullet, kole, manini, etc.,
then travels up to the carnivores such as ulua and
kahala and finally to man, In the case of WBH,
the major ciguatera causative fish was mullet. A
few mullet revealed G. toxicus in the gut smear

4/92 N

ASSESSMENT OF CIGUATERIC FISH IN HAWAII 495
Table 5.Results of Crude Fish Extracts in the Guinea Pig Atrium and Mouse Toxicity Assays
Date Guinea Pig Mouse
Atria Inotropic TTX Verap Adren Toxicity
4/92 B&P L + sl sl 7 5
6/92 Mullet flesh ND R + - 5
6/92 Mullet gut B&P R
4/92 | Mullet flesh _| Neg L
6/92 Neg R
Mullet gut Neg R
3/92 Mullet flesh Pooled all fish R
5/92 Mullet flesh Pooled all fish L
3/92 Mullet gut Pooled L 5
4/92 | Mullet gut Pooled | iL 5
5/92 Mullet gut Pooled R 5
4/92 Hawn kole flesh B L 5
4/92 Hawn kole gut B R 5
5/92 Hawn kole flesh Pooled R 5
Hawn kole gut Pooled L
4/92 Manini flesh B&P R + - - 0
5/92 Manini flesh B&P L a Sl Sl - 5
6/92 Vsl Vsl -
6/92
2
5
5
5

5/92 Neg L
6/92 Manini flesh Neg L
Manini gut Neg L
Manini gut Pooled
a
5
2
5

*S-EIA Results: B, borderline; P, positives; ND, no data; Neg, negative; Pooled, pooled extracts of the same
species (negative, borderline and positive S-EIA results)
**Atria: L, left; R, right; TTX, tetrodotoxin; Verap, verapamil, Adren, adrenergic; Sl, slight; Vsl, very slight; +,

positive; -, negative (no response)

analysis. It is also suggested that until the G.
toxicus is diminished, herbivores and carnivores
in WBH should not be consumed. The only safe
fish have been the akule and halalu and they are
being caught and eaten by recreational fishermen
without ciguatera outbreaks. These findings are
essentially similar to those reported for Puako,
Hawaii (Hokamaetal.,1993), except for differen-
ces in the presences of algae species associated
with G. toxicus. Since the major outbreak in 1991,
no other reports of toxicity due to WBH fish have
been observed to date. This is to be expected,
since the species of potentially toxic fish in WBH
have been continuously monitored and reported

to the public by the Department of Health. These
species include the mullet and all herbivores and
carnivores within WBH. These reports of warn-
ing also include the presences of G. toxicus
among Bryopsis.

The guinea pig atrial analysis revealed at least
2 toxins in the fish obtained from WBH. These
included a ciguatoxin-like (inotropic response
inhibited by TTX) and a maitotoxin-like (in-
otropic response inhibited by verapamil). The G.
toxicus extract showed that it contained mostly
ciguatoxin-like toxin. In addition, a new sodium
channel inhibitory toxin(s) has been noted and is
being studied.

496

ACKNOWLEDGEMENTS

We thank the Hawaii State Legislature through
the Department of Health, the Asian-Pacific Re-
search Foundation and the Hokama-Yagawa
Memorial Fund (UHF) for financial assistance.

LITERATURE CITED

HOKAMA, Y., SHIRAI, L.K. & MIYAHARA, J.T.
1986. Seafood and ciguatera poisoning.
Laboratory identification methods. Laboratory
Managment 24: 29-40,

HOKAMA, Y. 1988. Ciguatera fish poisoning. Journal
of Clinical Laboratory Analysis 2: 44-50,

HOKAMA, Y., HONDA, S.A.A., KOBAYASHI,
M.N., NAKAGAWA, L.K., ASAHINA, A.Y., &
MIYAHARA, J.T. 1989. Monoclonal antibody in
detection of ciguatoxin (CTX) and related
polyethers by the Stick Enzyme Immunoassay
(S-EIA) in fish tissue associated with ciguatera
poisoning. Pp. 303-310. In S. Natori, K.
Hashimoto & Y. Ueno (eds), ‘Mycotoxins and
phycotoxins *88", (Elsevier: Amsterdam).

HOKAMA, Y., ASAHINA, A.Y., TITUS, E., SHIRAI,
J.L.R., HONG, T.W.P., CHUN, S., MIYAHARA,
J.T., TAKATA, D., MURANAKA, A., PANG,
E., ABBOTT, L.A, & ICHINOTSUBO, D. 1993,
A survey of ciguatera: assessment of Puako,
Hawaii, associated with ciguatera toxin epidemics
in humans. Journal of Clinical Laboratory
Analysis 7: 147-154.

KIMURA, L.H., HOKAMA, Y., ABAD, M.A.,

MEMOIRS OF THE QUEENSLAND MUSEUM

OYAMA, M., & MIYAHARA, J.T. 1982. Com-
parison of three different assays for the assessment
of ciguatoxin in fish tissues: radioimmunoassay,
mouse bioassay and Jn vitro guinea pig atrium
assay. Toxicon 20; 907-912.

MIYAHARA, J.T., KAMIBAYASHI, C,K. &
HOKAMA, Y. 1989. Pharmacological charac-
terization of the toxins in different fish species.
Pp.399—-406. In S. Natori, K. Hashimoto, Y. Ueno
(eds), ‘Mycotoxins and phycotoxins "88" (El-
sevier: Amsterdam).

RANDALL, J.E. 1958. A review of ciguatera tropical
fish poisoning with tentative explanation of its
cause. Bulletin of Marine Science in the Caribbean
Gulf 8: 236-267.

TINKER, S.W. 1982. ‘Fishes of Hawaii, a handbook of
marine fishes of Hawaii and the central Pacific
Ocean’. (Hawaii Service: Honolulu).

YASUMOTO, T., INOUE, A., BAGNIS, R. & GAR-
CON, M. 1979. Ecological survey on a dinoflagel-
late possibly responsible for the induction of
ciguatera. Bulletin Japanese Society for Scientific
Fisheries 445: 395-399,

YASUMOTA, T., INOUE, A., OCHI, D., FUJIMOTO,
K., OSHIMA, Y., FUKUYO, Y., ADACHI, R. &
BAGNIS, R. 1980. Environmental studies on a
toxic dinoflagellate responsible for ciguatera. Bul-
letin of the Japanese Society of Scientific Fisheries
46: 1397-1404.

YASUMOTO, T., RAJ, U. & BAGNIS, R. 1984.
‘Seafood poisoning in tropical regions’.
(Laboratory of Food Hygiene, Faculty of Agricul-
ture, Tohoku University).

THE ORIGIN OF CIGUATERA
MICHAEL J. HOLMES AND RICHARD J, LEWIS

Holmes, M.J.& Lewis, R.J..1994 08 01; The origin of Ciguatera. Memoirs of the Queensland
Museum 34(3), 497-504, Brisbane, ISSN 0079-8835.

Ciguatera is caused by eating fish contaminated with ciguatoxins. Ciguatoxins-1, -2 and -3
are the major ciguatoxins found in the flesh and liver of ciguateric fishes with ciguatoxin-1
the most toxic and most abundant. Gambiertoxin-4b is the likely precursor of ciguatoxin-3
which is in turn oxidatively metabolised in fishes to ciguatoxin-|. Consequently, gambier-
toxin-4b accounts for more than 90% of the toxicity of ciguateric fishes. Gambiertoxin-4b
has been extracted from biodetritus containing large numbers of the benthic dinoflagellate
Gambierdiscus toxicus, indicating that G. foxicus is the primary source of Loxins involved
in ciguatera. Putative gambiertoxins have also been detected in certain strains of cultured G,
toxicus. However, the link between G. toxicus and ciguatera remuins circumstantial since
pecnbiertspeln-4 has not yet been Unambiguously identified from cultures of this dinoflagel-
ale.

Michael J, Holmes & Richard J. Lewis, Southern Fisheries Centre, Queensland Department
of Primary Industries, PO Box 76 Deception Bay, Queensland 4508: 22 November, 1993,

Ciguateric fishes are poisonous because their
flesh and viscera contain elevated concentrations
of lipid-soluble polyether ciguatoxins (Murata et
al.1990; Lewis et al.,1991; Lewis & Sellin,
1992). Ciguatoxins-1, -2 and-3 (Fig.1) have been
isolated fromthe flesh of toxic carnivorous fish
with ciguatoxin-1 being most abundant and most
toxic (Lewis & Sellin, 1992). Some minor toxins
(presumably ciguatoxins) remain to be charac-
terised from carmivorous and herbivorous fishes
(Murata et al,,1990; Lewis et al.,1991; Lewis &
Sellin. 1992; Legrand et al., 1992). Ciguatoxins-2
and -3 do not have the secondary hydroxy] on
carbon 54 and are therefore less-polar than
ciguatoxin-1 (Lewis etal., 1991). Ciguatoxin-3 is
thought to be an intermediate in the oxidative
metabolism of a less-polar precursor, gambier-
toxin-th, to cigua- toxin-1 (Lewis et al., 1991)
whereas ciguatoxin-2 is 3 diastereomer of cigua-
toxin-3 (Lewis et al., 1993) which may originate
from a different precursor. The stereochemistry
at carbon 52 indicates that ciguatoxin-2 has a
different structural backbone to ciguatoxins-1
and -3 (Lewis et al., 1993). Ciguatoxins-2 and -3
induce similar bioassay signs in mice, including
hind limb paralysis, the only bioassay sign that
differentiates these less-polar ciguatoxins from
ciguatoxin-|] (Lewis et al.,1991). Precursors of
ciguatoxins-2 and -3 are thought to enter the
manne food web incidentally upon ingestion by
lower trophic level fishes (e.g. herbivores/
detritivores like surgeonfishes (Randall,1958;
Lewis et al,,1991)) or invertebrates (Kelly et
al,,1992; Lewis et al.,1994), These species are in
turn preyed on by carnivorous fishes.

Randal}’s (1958) hypothesis that the ciguatera
toxins onginate from a small benthic organism
received strang but circumstantial support when
Yasumoto et al. (1977a,b) extracted ciguatoxin-
like and maitotoxin-like toxins from a benthic
detrital sample containing large numbers of the
dinoflagellate Gambierdiscus toxicus Adachi &
Fukuyn. Yasumoto et al. (1979a) and Bagnis et
al. (1980) were able to repeat the extraction of
such toxins from biodetritus from the Gambier
Islands and to extract maitotoxin from cultures of
G. toxicus. However, the column and thin-layer
chromatography of the ciguatoxin-like toxin
found by Yasumoto’s group was not consistent
with the major ciguatoxin (ciguatoxin-1) found in
fishes but instead was indicative of a less-polar
ciguatoxin-like toxin (Lewis,1985). The role of
G, toxicus in ciguutera remuined in doubt sinve
meagre amounts of this ciguatoxin-like toxin, if
any, were produced by cultured G. texicus
(Yasumotoetal.,1979a: Bagnis etal..1985a; Hol-
mes et al.,1990; Murata ct al.,1990; Yasumoto,
1990) and this toxin could not be extracted from
wild G. toxicus outside of French Polynesia (Gil-
lespie et al.,1985a).]t was not until gambiertoxin-
4b (precursor of ciguatoxins-} and -3) was
extracted from biodetntus samples containing
large numbers of wild G. toxicus (collected in
1979 from the Gambier Islands and kept at -20°
for many years) and its structure compared with
that of ciguatoxin-l, that G. fexicus was once
again considered the likely origin of ciguatera
(Murata et al.,1990; Legrand et al.,1990,1992).
Nine gambiertoxins have since been extracted
from wild G. fexicus, including gambiertoxin-4c

d98

MEMOIRS OF THE QUEENSLAND MUSEUM

a
if]

1CH, = CH -

FIG. 1. Structures of ciguatoxins (CTX)-1, -2 and -3 and gambiertoxin-4b (GT-4b) (Murata et al.,1990; Lewis
eval.,1991). Ciguatoxin-2 is a diaslereomer of ciguatoxin-3 (Lewis et al.,1993).

(the major toxin in terms of mouse lethality) and
the isomers, gambiertoxin-4a and -4b (Legrand et
al,,1992), Of these, only the structure of gambier-
toxin-4b 1s known (Murata etal., 1990), However,
gambiertoxin-4b appears to be the most impor-
tant gambiertoxin contributing te ciguatera as its
oxidative products, ciguatoxin-1 and ~3, account
for more than 90% of the toxicity of ciguateric
fish (Lewis et al.,1991; Lewis & Sellin,1992).
The site of biotransformation of gambiertoxin-4b
to ciguatoxin-3 and of ciguatoxin-3 to
ciguatoxin-I remains to be established, but likely
occurs in the liver of fishes (Lewis et al.,1991).
Extraction of gambiertoxins from reef bio-
detritus containing wild G. roexicus is circumstan-
tial evidence that G. toxicus is the origin of the
toxins that cause ciguatera, Further support for
this hypothesis was obtained by Holmes et al.
(1991) and Holmes & Lewis (1992) who found
two putative gambiertoxins (called major and
minor based upon their relative contribution to
total lethality) in cultures of certain strains of G.
toxicus. These toxins were less-polar than
ciguatoxins-1,-2 or-3.but were considered close-
ly related to the ciguatoxins since: (i) both gam-
biertoxins produced bioassay signs in mice
similar to those produced by ciguatoxins-2 and
~3, (ii) the major (more-polar) toxin was shown
to be a Nat channel activator toxin (as are the
three ciguatoxins) that produced pharmacologi-
cal responses in isolated tissues similar to those
produced by the ciguatoxins (especially
ciguatoxin-3; Lewis & Wong Hoy, 1993) and (iii)
the major loxin competitively inhibited the bind-

ing of ?H)brevetoxin-3 to rat brain synap-
tosomes. The ciguatoxins (and brevetoxins) are
the only toxins known to competitively inhibit
brevetoxin binding to the Na* channel (Lombet
et al.,1987; Lewis et al..1991), However, gam-
biertoxin-4b has not been unambiguously iden-
tified from cultured G. toxicus and therefore the
origin of the precursor of ciguatoxins-1 and. -3
remains to be established. Many reports have
claimed to extract ciguatoxin or ciguatoxin-like
toxins from cultures of G. toxicus (Bergmann &
Alam.1981: Withers,1982; Shimizu et al.,1982;
Miller et al.,1984; Lechat et al.,1985; Durand et
al., 1985; Durand-Clement, 1987) butthese inves-
tigations relied upon a liquid-liquid partition (eg.
diethy] ether-water) to completely separate
ciguatoxin-like toxins from the considerable
amounts of maitotoxin present in crude extracts,
Since maitotoxin can partition into both the lipid-
and water-soluble phases (Yasumoto et al.,
1979a; Holmes et al.,1990) these former studies
are unlikely to have completely separated any
ciguatoxin-like material from the maitotoxin.
Production of gambiertoxins in cultured G,
toxicus appears to be strain-dependent, with mos
clones only producing maitotoxins (Holmes et
al,,1991), These authors found that only two of
13 cultured strains produced putative precursors
of the ciguatoxins, indicating that the aetiology
of ciguatera 1s likely restricted to genetic strains
of G. toxieus which can produce gambiertoxins.
The most striking evidence for this genetic
variability was that putative gambiertoxins were
produced by only one of four G. foxicus clones

ORIGIN OF CIGUATERA

isolated from the same site, with only one of two
clones isolated from the same site and at the same
lime producing gambiertoxins (Holmes et
a].,1991). This resule has imporiant implications
for ecological studies of G_ toxicus, as it indicates
that the size of G. fexicus populations does not
necessanly reflect the potential for these popula-
tions to cause ciguatera. Wild populations of G.
foxicus from the Gambier Islands, Kinbati and
Platypus Bay, Queensland have been found to
produce gambiertoxins (Legrand et al.,1990,
1992; Holmes et al..1991; Holmes et al,,1994)
whereas a large population of wild cells from
Flinders Reef, south Queensland did not produce
detectable levels of these toxins (Gillespie ct al.,
1985a; Lewis et al.,19838a).

Only relatively low concentrations of gambter-
toxins have so far been detected from cultured
compared with wild G. rexicus cells (Holmes et
4].,1991, 1994; Holmes and Lewis, 1992). Consjd-
erable variation can also occur in the concentra-
lion of gambiertoxins produced by wild G-.
fexicus {Holmes et al_, 1994). Hoimes etal. (1994)
have proposed the existence of “super-producing
strains” of G, toxics to explain some of this
variation in gambiertoxin production between
cultured and wild G. toxicus. However, environ-
mental factors are also likely to affect toxin
production since the concentration (or type) of
gambiertoxin produced can change in culture
(Holmes & Lewis,1992). Environmental condi-
tions obyiously effect the growth of G. toxicus as
they do for any other algae, bueitis not known if
environmental parameters can selectively affect
the growth and toxicity of super-producing
strains over non-producers. It is quite likely that
the conditions which enhance growth will not
necessanly be the conditions which enhance
toxin production. Future research could focus on
the effect of different combinations of genetic
and environmental parameters on the rate of gam-
hiertoxin production,

Ecological studies need to also consider the
effect of different rates of tumover Of G. roxtetes
populations. A large standing crop of G. toxicus
dees not necessarily indicate a greater potential
to cause ciguatera compared with a smaller
population, if the lower biomass is simply a
reflection of higher productivity and higher rates
of consumption by herbivores. Recent evidence
that fishes can excrete/metabolise the ciguatoxins
(Tosteson etal.,1988; Lewiset al.,1992) suggests
that considerably greater quantities of gambier-
toxins are entering the marine food web than
would otherwise be expected by the frequency of

499

ciguatera, Seasonal patterns of ciguatera in some
Pacific Island countries (Sorokin, 1975; Daw-
son,1977; Bagnis,1979; Bagnis et al.,1992;
Lewis, 1992) may also reflect seasonal pattems in
the abundance and/or gambiertoxin production
by wild G. foxiceus. There are numerous reports of
seasonal vanation of populations of G. fexicus
and other benthic dinoflagellates (Yasumoto et
al.,1979b; Carlson & Tindall,1985; Gillespie et
al.,1985b; Bagnis et al., 1985b; Ballantine et al,,
1985, 1988; Bomber ctal., L988; McCaffrey ctal,,
1992; Holmes et al.,19%4), However, seasonal
patterns of G. texicus abundance have been cor-
Telated with fish toxicity only in French Polynesia
(Bagnis et al. 1985bh).

Most strains of G. foxieus appear not lo
produce gambiertoxins, while most produce a
maitatoxin (Holmes et al.,1991), Maitotoxins are
generally referred to as water-soluble toxins al-
though they are soluble in a range of organic
solvents and a butanol-water liquid-liquid parti-
tion will recover nearly 100% of maitotoain im the
butanel phase (unpublished result). The
mattotoxins have a cyclic polyether structure as
do the gambiertoxins and ciguatoxins (Yoko-
yamaet al.,1988; Murataetal.,1991, 1992, 1993),
Interestingly, the type of maitotexin produced by
G. toxicus is dependent upon the strain being
cultured, with each strain apparenily producing
only the one type of maitoioxin (Holmes et
al.,1990}. Holmes & Lewis (in press) have recent-
ly found that large maitotoxins (maito-toxins- |
and -2, with molecular weights >3,000) were
produced by strains of G. taxicus whieh do not
produce gambiertoxins, whereas the small
maitotoxin-3 (molecular weight 1,060 for the dis-
edium salt) was produced by a clone which ulso
produces gambiertoxins, The molecular weight
of maitotoxin-3 is the same as gambiertoxins—4tu
and -4b (Murata et al.,1990; Legrand et al..1992).
Holmes & Lewis (in press) have suggested that
ihe biosynthesis of gambiertoxins and
maitoioxins may be linked in strains of G. toxicus
which produce both of these types of toxins,

Maitotoxin has been found in the gut contents
of surgeonfishes (Yasumoto et al..1976) but there
is lite evidence that maitotoxin is accumulated
in the flesh of these or other fishes. Water-
soluble, maitotoxin-like toxins haye been ex-
tracted from the flesh of fishes in Hawan and
Queensland (Iwanka et al..1993; Endean et al.
1993), However, these studies based the deiec-
lion of these toxins, at least in part, on in-
traperitoneal (i-p.) injections into mice of at lewst
100 mg of crude extracts (25 g extracl/ke mouse

500

body weight). Doses of crude fish extracts >1
g/kg i.p. can produce non-specific toxic effects
(Banner et al.,1961; Lewis et al.,1988a). Un-
saturated fatty acids extracted from shellfish have
also been shown to produce toxic effects when
injected i.p. into mice (Takagi et al.,1984). Cal-
culations of total toxicity based upon lethal doses
of such large amounts of extract would likely
result in the overestimation of the quantity of
toxin present. Additionally, the water-soluble ex-
tracts of fish flesh killed mice quickly (5 and 13
min, Iwaoka et al., 1993; 3—30min, Endean et al.,
1993). However, based upon these rapid deaths,
we conclude that any water-soluble toxins iso-
lated were not maitotoxins, since very large doses
of native maitotoxin (e.g. >100 lethal units)
would be required to produce such short death
times. The three maitotoxins characterised so far
are potent but slow acting toxins with the shortest
survival times (calculated according to Molinen-
go, (1979)) being greater than 41 min (Holmes et
al.,1990; Holmes & Lewis in press).
Maitotoxins are the most toxic toxins produced
by G. toxicus, often comprising more than 99%
of total toxicity (Holmes & Lewis,1994), Fishes
fed cultured G. toxicus cells display abnormal
swimming behaviour (Davin et al.,1986,1988;
Kelly et al.,1992) probably as the result of
maitotoxin intoxication. Maitotoxin poisoning of
fishes in the wild may result in these fishes being
preferentially preyed upon. This could be a
mechanism for concentrating gambiertoxins
through the food chain of fishes when her-
bivorous fishes ingest strains of G, toxicus which
produce both gambiertoxins and maitotoxins.
Toxins other than ciguatoxins-1, -2 and -3 have
been suggested as causes of ciguatera. Scaritoxin,
extracted from parrotfish (Scarus gibbus) from
the Gambier Islands (Chungue et al.,1977), may
be a less-polar form of ciguatoxin (Lewis et al.,
1991; Legrand et al., 1992). Vernoux & Talha
(1989) detected fast-acting ciguatoxins in fish
flesh; instability and quick death-times induced
by these toxins distinguish them from cigua-
toxins-], -2 and -3, which are stable and slow
acting (Lewis et al., 1991). Other toxins suggested
as causal agents of ciguatera include palytoxin,
okadaic acid and other toxins (predominately
water-soluble toxins) produced by the benthic
dinoflagellate species Ostreopsis spp. and
Prorocentrum spp., and toxins produced by the
planktonic cyanophyte (cyanobacterium) Oscil-
latoria (Tricodesmium) erythraea (Yasumoto et
al.,1980; Nakajima et al.,1981; Murakami et al.,
1982; Tindall et al.,1984,1990; Norris et al.,1985;

MEMOIRS OF THE QUEENSLAND MUSEUM

Holmes et al.,1988; Kodama et al., 1989; Dickey
et al.,1990; Juranovic & Park,1991; Hahn &
Capra, 1992).

Palytoxin is a potent water-soluble toxin iso-
lated from various Palythoa coral species (Moore
& Scheuer,1971; Habermann,1989). Palytoxin
(or one of its congeners) has been found in the
flesh and viscera of triggerfish Melichthys vidua
(Fukui et al.,1987a,b) and smoked mackerel
Decapterus macrosoma (Kodama et al.,1989).
Palytoxin is also thought to be responsible for
intoxications caused by eating parrotfish liver
(Ypsiscarus ovifrons) from western Japan
(Noguchi et al.,1987). The extent of human
poisoning from palytoxin is not known but we
believe it is a separate poisoning distinct from
ciguatera. Hospitalised cases present with signs
distinguishable from ciguatera including
elevated serum enzyme levels (Noguchi et
al.,1987; Kodama et al., 1989). However, mild
palytoxin poisoning may be mistaken for
ciguatera.

Okadaic acid is a lipid-soluble polyether toxin
with similar chromatography to ciguatoxins
(Yasumoto et al.,1980; Murakami et al.,1982).
Okadaic acid was originally isolated from the
black sponge Halichondria okadai (Tachibana et
al.,1981) but has been isolated from the benthic
dinoflagellates Prorocentrum lima (Murakami et
al.,1982) and P. concavum (Dickey et al.,1990)
and from the temperate dinoflagellates Dino-
physis (Lee et al.,1989). Okadaic acid is one of
the toxins that can accumulate in shellfish to
cause a disease known as diarrhetic shellfish
poisoning (Lee et al.,1988). However, the only
fishes from which okadaic acid has been ex-
tracted are barracuda from the Caribbean (Gam-
boa et al.,1992). This result requires confirm-
ation, including an estimate of whether the levels
detected were sufficient to cause human poison-
ing. The primary basis for linking okadaic acid
(and the other toxins produced by Ostreopsis spp.
and Prorocentrum spp.) with ciguatera is (i) the
dinoflagellates that produce these toxins would
likely be ingested by the same or similar her-
bivores that ingest G. toxicus, and (ii) the diverse
range of symptoms of ciguatera could result from
a combination of several toxins. However, there
is little evidence to indicate that the toxins
produced by these dinoflagellates accumulate in
fish flesh to cause human poisoning. Symptoms
of ciguatera could result from ingestion of dif-
ferent relative amounts of different ciguatoxins
(Lewis & Sellin,1992) and/or different doses of

ORIGIN OF CIGUATERA

ciguatoxin (Yasumoto et al,,1984; Lewis et al..
1988b).

Evidence linking O. erythraea to ciguatera is
similarly unconvincing. O. erythraea is a tropical
and sub-tropical planktonic, filamentous
cyanophyte common off the cast and west coasts
of Australia (Hallegraeff, 1991). However, this
Species 1s also common in areas where ciguatera
has not been reported (Lewis, 1988; unpublished
observations). Toxins have heen extracted from
Caribbean and Australian samples of O.
erythraea (Hawser et al.,1991; Hahn & Capra,
1992) but plankton-feeding fishes, which would
be the most likely to ingest this algae, apparently
do not cause ciguatera (Randall,1958). Hahn &
Capra (1992) showed that a toxic fraction could
be bnefly accumulated by filter-feeding bivalves
and that this toxin may be subsequently accumu-
latect in the viscera of the molluscivorous fish
Trachinotus blocki. However, only one case of
human poisoning by Trachinotur sp. has been
recorded in Australia since 1965 (unpubl. data).

Ciguatera is caused by eating fish which have
accumulated toxins from their diet, Nearly all
benthic dinoflagellates produce toxins but not all
of these toxins are accumulated to harmful levels
in the flesh of fishes. Evidence to-date suggests
that, of the benthic dinoflagellates, only G-.
toxicus produces toxins, especially gambier-
toxin-4b, responsible for the human poisoning
syndrome known as ciguatera. The ciguatoxins
are the major toxins found inciguateric fishes and
must be considered the primary cause of
ciguatera. Other toxins and other sources of these
toxins have been suggested as being involved in
ciguatera but their involvement, if any, remains
to be established.

LITERATURE CITED

BAGNIS, R.A. 1979. Leichtyosarvotoxisme de type
ciguatera cn Nouvelle-Calédonic, aspects clini-
ques ct épidémiologiques. Rey. Epidem. Sand
Publ, 27; 17-29,

BAGNIS, R., CHANTEAU, S., CHUNGUE, E., HUR-
TEL, J,.M,, YASUMOTO, T, & INOUE, A, 1980.
Origins of ciguatcra fish poisoning: a new
dinoflagellate, Ganbierdiscus toxicus Adachi and
Fukuyo, definitively involved as a causal agent.
Toxicon 18:199-208,

BAGNIS, R., BENNETT, J.. BARSINAS, M.,
CHEBRET, M., JACQUET, G., LECHAT, I.,
MITERMITE, Y., PEROLAT. P. &
RONGERAS, 8&. L98Sa. Epidemiology of
ciguatera in French Polynesia from 1960 to 1984.

, 475-482 In Gabrie, C. & Salvat, B. (eds),
dings of the Fifth International Coral Reef

Si)

Congress, Tahiti, vol.4’, (Antenne Museum
Ephe: Moorea).

BAGNIS, R., BENNETT, }., PRIEUR, C. &
LEGRAND, A-M. 1985b. The dynamics of uirce
toxic dinoflagellates and the toxicity of caguateric
surgeonfish in French Polynesia, Pp, 177-182 In
Anderson, D.M., White, A.W. & Baden, D.G.
feds), “Toxic dinoflagellates’. (Elsevier: New

Vosk),

BAGNIS, R.A., SPIEGEL, A,. NGUYEN, L.,
PLICHART, R, 1992, Public health,
epidemiological and socioeconomic patierns of
ciguatera in Tahiti. Pp. 157-168 In Tosteson, T_R.,
(ed.), ‘Proceedings of the Third Intemational Con-
ference on Ciguatera Fish Poisoning, Puerto Rico"
(Polyscience Publications: Québec),

BALLANTINE, D.L., BARDALES, A,T., TOS-
TESON, T. & DUPONT- DURST, H. 1985.
Seasonal abundance of Gambierdiscns toxicus
and Ostreapsis sp. in coastal waters of southwest
Pueno Rico. Pp. 417-422 In Gabnie, C. & Salvue,
B. (eds), “Proceedings of the Fifth Internaticnal
Coral Reef Congress, Tahiti, vol.4", (Antenne
Muscum-Ephe: Moorea).

BALLANTINE, D.L., TOSTESON, T, & BAR-
DALES, A.T. 1988, Population dynamics and
loxicity of natural populations of benthic
dinoflagellates in southweslerm Puerto Rico, Jour-
nal of Experimental Marine Biology and Ecology
119; 201-212.

BANNER, A.H., SASAKI, §., HELFRICH, P,,
ALENDER, C.B. & SCHEUER, P.J, 1961. Bixas-
say of ciguatera toxin, Nature 189: 229-230.

BERGMANN, J.S. & ALAM, M. 1981. On the toxicity
of the ciguatera producing dinoflagellate Gar-
bierdiscus toxicus Adachi and Fukuyo isolated
from the Florida Keys. Journal! of Environmental
Science. Series A: Health 16: 493-500.

BOMBER, J.W., GUILLARD, R.R.L. & NELSON,
W.G. 1988, Réles of temperature, salinity, and
light in seasonality of ciguatera-causing Gambier-
discus texicus Adachi ct Fukuyo (Dinophyceze).
Jourmal of Experimental Marine Biology and
Beology 115; 53-65,

CARLSON, R.D. & TINDALL, D.R. 1985. Distiba-
tion and periodicity of toxic dinoMugellates in the
Virgin Islands. Pp. 171-176 In Anderson, D.M_.,
White, A.W, & Baden, D,G. (eds), ‘Toxic
dinoflagellates’, (Elsevier: New York).

CHUNGLIE, E., BAGNIS, R., FUSETANI, N, &
HASHIMOTO, Y. 1977. Isolation of two voxins
from a parrotfish Scarus pibbus. Toxicon 15: 89-

J.

DAVIN, W.T., KOHLER, C.C, & TINDALL, DR,
1986. Effects of ciguatera toxins on the.bluebead.
Transactions of the American Fisheries Society.
115: 908-912,

DAVIN, W-T., KOHLER, Cc. & TINDALL, D.R-
1988, Ciguatera toxins adversely affect pis-
civorous. fishes, Transactions of the American
Pisherics Sociery 177; 374-384.

502

DAWSON, J.M. 1977. Fish poisoning in American
Samoa. Hawaii Medical Journal 36: 239-243.

DICKEY, R.W., BOBZIN, S.C., FAULKNER, D.J.,
BENCSATH, F.A. & ANDRZEJEWSKI, D.
1990. Identification of okadaic acid from a Carib-
bean dinoflagellate, Prorocentrum concayum.
Toxicon 28: 371-377.

DURAND-CLEMENT, M. 1987. Study of production
and toxicity of cultured Gambierdiscus toxicus.
Biological Bulletin 172: 108-121.

DURAND, M., SQUIBAN, A., VISO, A.C. & PESAN-
DO, D. 1985. Production and toxicity of Gambier-
discus toxicus effects of its toxins (maitotoxin and
ciguatoxin) on some marine organisms. Pp. 483—
487 In Gabrie, C. & Salvat, B. (eds), ‘Proceedings
of the Fifth International Coral Reef Congress,
Tahiti, vol.4’. (Antenne Museum-Ephe: Moorea).

ENDEAN, R., GRIFFITH, J.K., ROBINS, J.J. &
MONKS, S.A. 1993. Multiple toxins in a spec-
imen of the narrow-barred Spanish mackerel,
Scomberomorus commersoni. Toxicon 31: 195—
204,

FUKUI, M., MURATA, M., INOUE, A., GAWEL, M.
& YASUMOTO, T. 1987a. Occurrence of
palytoxin in the triggerfish Melichthys vidua.
Toxicon 25: 1121-1124.

FUKUI, M., YASUMURA, D., MURATA, M., AL-
CALA, A.C. & YASUMOTO, T. 1987b. Occur-
rence of palytoxin in crabs and fish, Pp. 477-482
In Gopalakrishnakone, P. & Tan, C.K. (eds),
‘Progress in venom and toxin research’. (National
University of Singapore: Singapore).

GAMBOA, P.M., PARK, D. & FREMY, J.-M. 1992.
Extraction and purification of toxic fractions from
barracuda (Sphyraena barracuda) implicated in
ciguatera poisoning. Pp. 13-24 In Tosteson, T.R.,
(ed.), ‘Proceedings of the Third International Con-
ference on Ciguatera Fish Poisoning, Puerto Rico’
(Polyscience Publications: Québec).

GILLESPIE, N.C., LEWIS, R., BURKE, J. & HOL-
MES, M. 1985a. The significance of the absence
of ciguatoxin in a wild population of G. toxicus.
Pp. 437-441 In Gabrie, C. & Salvat, B. (eds),
‘Proceedings of the Fifth International Coral Reef
Congress, Tahiti, vol.4’. (Antenne Museum-
Ephe: Moorea).

GILLESPIE, N.C., HOLMES, M.J., BURKE, J.B. &
DOLEY, J. 1985b. Distribution and periodicity of
Gambierdiscus toxicus in Queensland, Australia.
Pp. 183-188 In Anderson, D.M., White, A.W. &
Baden, D.G. (eds), ‘Toxic dinoflagellates’. (El-
sevier: New York).

HABERMANN, E. 1989. Palytoxin acts through Na‘,
K*-ATPase. Toxicon 27: 1171-1187.

HAHN, S.T. & CAPRA, M.F. 1992. The cyanobac-
terium Oscillatoria erythraea - a potential source
of toxin the ciguatera food chain, Food Additives
and Contaminants 9: 351-355.

HALLEGRAEFF, G.M. 1991. ‘Aquaculturists’ guide
to harmful Australian microalgae.’ (Fishing In-
dustry Training Board of Tasmania: Hobart),

MEMOIRS OF THE QUEENSLAND MUSEUM

HAWSER, S.P., CODD, G.A., CAPONE, D.G. &
CARPENTER, E.J. 1991. A neurotoxic factor
associated with the bloom-forming cyanobac-
terium Tricodesmium. Toxicon 29; 277-278.

HOLMES, M.J., GILLESPIE, N.C. & LEWIS, R.J.
1988. Toxicity and morphology of Ostreopsis cf
siamensis, cultured from a ciguatera endemic
region of Queensland, Australia. Pp. 49-54 In
Choat et al. (eds), ‘Proceedings of the Sixth Inter-
national Coral Reef Symposium, Townsville, vol.
3’. (6th International Coral Reef Symposium Ex-
ecutive Committee: Townsville).

HOLMES, M.J., LEWIS, R.J. & GILLESPIE, N.C.
1990. Toxicity of Australian and French
Polynesian strains of Gambierdiscus toxicus
(Dinophyceae) grown in culture: characterization
of a new type of maitotoxin. Toxicon 28: 1159-
1172.

HOLMES, M.J., LEWIS, R.J., POLI, M.A. & GIL-
LESPIE, N.C. 1991. Strain dependent production
of ciguatoxin precursors (gambiertoxins) by Gam-
bierdiscus toxicus (Dinophyceae) in culture.
Toxicon 29, 761-775.

HOLMES, M.J. & LEWIS, R.J. 1992, Multiple gam-
biertoxins (ciguatoxin precursors) from an
Australian strain of Gambierdiscus toxicus in cul-
ture. Pp. 520-529 In Gopalakrishnakone, P. &
Tan, C.K. (eds), ‘Progress in venom and toxin
research’. (National University of Singapore: Sin-

apore),

HOLMES, M.J. & LEWIS, R.J. in press. Purification
and characterisation of large and small
maitotoxins from cultured Gambierdiscus
toxicus. Natural Toxins.

HOLMES, M.J., LEWIS, R.J., SELLIN, M. &
STREET, R. this memoir. The origin of ciguatera
in Platypus Bay.

IWAOKA, W., HORITA, J., SHIMOJO, R. & TRAN,
T. 1993. Analysis of Acanthuris triostegus for
marine toxins by stick enzyme immunoasssay and
mouse bioassay. Toxicon 30: 1575-1582.

JURANOVIC, L.R. & PARK, D.L. 1991. Foodborne
toxins of marine origin: ciguatera. Review of En-
vironmental Contamination and Toxicology 117:
51-94,

KELLY, A.M., KOHLER, C.C. & TINDALL, D.R.
1992. Are crustaceans linked to the ciguatera food
chain. Environmental Biology of Fishes 33: 275-
286.

KODAMA, A.M., HOKAMA, Y., YASUMOTO, T.,
FUKUI, M., MANEA, S.J. & SUTHERLAND,
N. 1989. Clinical and laboratory findings im-
plicating palytoxin as cause of ciguatera poisoning
due to Decapterus macrosoma (mackerel).
Toxicon 27: 1051-1053.

LECHAT, I., PARTENSKI, F. & CHUNGUE, E. 1985.
Gambierdiscus toxicus: culture and toxin.produc-
tion. Pp. 443-448 In Gabrie, C. & Salvat, B. (eds),
‘Proceedings of the Fifth International Coral Reef
Congress, Tahiti, vol.4’. (Antenne Museum-
Ephe: Moorea).

ORIGIN OF CIGUATERA

LEE, J.S., TANGEN, K., DAHL, E., HOVGAARD, P.
& YASUMOTO, T. 1988. Diarrhetic shellfish
toxins in Norwegian mussels. Bulletin of the
pense Society of Scientific Fisheries 54: 1953-

LEE, J.S., IGARASHI, T., FRAGA, S., DAHL, E.,
HOVGAARD, P. & YASUMOTO, T. 1989.
Determination of diarrhetic shellfish toxins in
various dinoflagellates. Journal of Applied
Phycology 1: 147-152.

LEGRAND, A.-M., CRUCHET, P., BAGNIS, R.,
MURATA, M., ISHIBASHI, Y. &
YASUMOTO, T. 1990. Chromatographic and
spectral evidence for the presence of multiple
ciguatera toxins, Pp. 374-378 In Graneli, E.,
Sundstr6m, B., Edler, L. & Anderson, D.M.
(eds),.‘Toxic marine phytoplankton’. (Elsevier,
New York).

LEGRAND, A.-M., FUKUI, M., CRUCHET, P.,
ISHIBASHI, Y. & YASUMOTO, T. 1992. Char-
acterization of ciguatoxins from different fish
species and wild Gambierdiscus toxicus. Pp. 25-—
32 In Tosteson, T.R., (ed.), “Proceedings of the
Third International Conference on Ciguatera Fish
Poisoning, Puerto Rico’ (Polyscience Publica-
tions: Québec),

LEWIS, R. 1985. Ciguatera and ciguatoxin-like sub-
stances in fishes, especially Scomberomorus com-
mersoni from southern Queensland. PhD Thesis,
University of Queensland.

LEWIS, R. 1988. Ciguatera in southeastern
Queensland. Pp. 181-187 In Covacevich, J.,
Davie, P. & Pearn, J. (eds), “Toxic Plants and
animals, a guide for Australia.’ (Queensland
Museum, Brisbane),

LEWIS, R.J. 1992. Socioeconomic impacts and
management of ciguatera in the Pacific. Bulletin
de la.Société de Pathologie Exotique 85: 427-434.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fishes. Toxicon 30: 915-919.

LEWIS, R.J. & WONG HOY, A. 1993, Comparative
action of three major ciguatoxins on guinea-pig
atria and ilea. Toxicon 31: 437-446.

LEWIS, R.J., GILLESPIE, N.C., HOLMES, M.J.,
BURKE, J.B., KEYS, A.B., FIFOOT, A.T. &
STREET, R. 1988a. Toxicity of lipid-soluble ex-
tracts from demersal fishes at Flinders Reef,
southern Queensland. Pp. 61-65 In Choat et al.,
(eds), ‘Proceedings of the Sixth International
Coral Reef Symposium, Townsville, vol 3’. (6th
International Coral Reef Symposium Executive
Committee: Townsville).

LEWIS, R.J., CHALOUPKA, M.Y., GILLESPIE, N.C.
& HOLMES, M.J. 1988b. An analysis of the
human response to ciguatera in Australia. Pp.
67-72 In Choat et al., (eds), ‘Proceedings of the
Sixth International Coral Reef Symposium,
Townsville, vol 3’. (6th International Coral Reef
Symposium Executive Committee: Townsville).

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K., & SHEIL, M.M. 1991.

503

Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae). Toxicon 29: 1115-1127.

LEWIS, R.J., SELLIN, M., STREET, R., HOLMES,
M.J. & GILLESPIE, N.C. 1992. Excretion of
ciguatoxin from moray eels (Muraenidae) of the
central Pacific. Pp. 131-143 In Tosteson, T.R.,
(ed.), ‘Proceedings of the Third International Con-
ference on Ciguatera Fish Poisoning, Puerto Rico’
(Polyscience Publications: Québec).

LEWIS, R.J., NORTON, R.S., BRERETON, ILM. &
ECCLES, C.D. 1993. Ciguatoxin-2 is a
diastereomer of ciguatoxin-3. Toxicon 31: 637—

LEWIS, R.J., HOLMES, M.J. & SELLIN, M. this
memoir. Invertebrates implicated in the transfer of
gambiertoxins to the benthic carnivore
Pomadasys maculatus.

LOMBET, A. BIDARD, J.-N. & LAZDUNSKI, M.
1987. Ciguatoxin and brevetoxins share a com-
mon receptor site on the neuronal voltage-depend-
ent Na* channel, FEBS letters 219: 355-359,

MCCAFFREY, E.J., SHIMIZU, M.M.K., SCHEUER,
P.J. & MIYAHARA, J.T. 1992. Seasonal abun-
dance and toxicity of Gambierdiscus toxicus
Adachi et Fukuyo from O’ahu, Hawai’i. Pp. 145-
153 In Tosteson, T.R., (ed.), ‘Proceedings of the
Third International Conference on Ciguatera Fish
Poisoning, Puerto Rico’ (Polyscience Publica-
tions: Québec).

MILLER, D.M., DICKEY, R.W. & TINDALL, D.R.
1984. Lipid- extracted toxins from a dinoflagel-
late, Gambierdiscus toxicus. Pp. 241-255 In
Ragelis, E.P. (ed.), ‘Seafood toxins.’(American
Chemical Society: Washington).

MOLINENGO, L. 1979. The curve doses vs survival
time in the evaluation of acute toxicity. Journal
Pharmaceutical Pharmacology 31: 343-344.

MOORE, R.E. & SCHEUER, P.J. 1971. Palytoxin: a
new marine toxin from a coelenterate. Science
172: 495-498,

MURAKAMI, Y., OSHIMA, Y. & YASUMOTO, T.
1982. Identification of okadaic acid as a toxic
component of a marine dinoflagellate
Prorocentrum lima. Bulletin of the Japanese
Society of Scientific Fisheries 48: 69-72.

MURATA, M., LEGRAND, A.M. ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T, 1990. Structures
and configurations of ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386.

MURATA, M., GUSOVSKY, F., SASAKI, M.,
YOKOYAMA, A., YASUMOTO, T. & DALY,
JW. 1991. Effect of maitotoxin analogues on
calcium influx and phospoinositide breakdown in
cultured cells. Toxicon 29: 1085-1096.

MURATA, M., IWASHITA, T, YOKOYAMA, A.,
SASAKI, M. & YASUMOTO,-T. 1992. Partial
structures of maitotoxin, the most potent marine

Att

toxin from the dinoflagellate Gambjerdiscus
toxicus. Journal of the American Chemical
Society 114: 6594-8596.

MURATA, M,, NAOKI, H,, IWASHITA, T., MAT-
SUNAGA, S., SASAKI, M., YOKOYAMA, A.
& YASUMOTO, T. 1993, Structure of
maitotoxin, Journal of the American Chemical
Society 115: 2060-2064.

NAKAJIMA, |, OSHIMA, Y, & YASUMOTO, T.
1981. Toxicity of benthic dinoflagellates in
Okinawz. Bulletin of the Japanese Society of
Scientific Fisheries 47- 1029-1033. ‘

NOGUCHI, T.. HWANG, D.F.,. ARAKAWA, O.,
DAIGO, K., SATAO, S,, OZARI, H., KAWAL,
N,. ITO, M. & HASHIMOTO, K. 1987. Palytoxin
as the causative agent in the parrotfish Seig
Pp, 325-335 In Gopalakrishnakone, P. & Tan,
C.K. (eds), “Progress in venom and toxin re-
search’. (National University of Singapore: Sin-

gapore).

NORRIS, D.R., BOMBER, J.W. & BALECH, E. 1985.
Benthic dinoflagellates associated wilh ciguetera
from the Florida Keys [, Ostreapsis heptagene sp,
nov. Pp. 39-44 In Anderson, D.M., White, A.W.
& Baden, D.G, (eds), ‘Toxic dinoflagellates”. (El-
sevier; New York),

RANDALL, J.B. 1958. A review of ciguatera, tropical
fich poisoning, with a tentative explanation of ils
cause. Bulletin of Marine Science 8; 236-267,

SHIMIZU, Y.. SHIMIZU. H., SCHEUER, PJ.,
HOKAMA., Y., OYAMA, M. & MIYAHARA,
J.T. 1982. Gambierdiscus toxicus, a ciguatera-
causing dinoflagellate from Hawaii. Bulletin af
the Japenese Society of Scientific Fishenes 48:
811-813.

SOROKIN, M. 1975, Ciguetera poisoning in north-west
Visi Leva, Fiji Isiands. Hawaii Medical Journal
34: 207-210,

TACHIBANA, K., SCHEUER, I'.J., TSURITANAL,
Y., KIKUCHI, H., ENGEN, D-V_,CLARDY, ¥.,
GOPICHAND, Y. & SCHMITZ. F.J. 1981.
Okadaic acid, a cylotoxic polyether from two
marine sponges of the genus /alichondria, Jour-
nal of the American Chemical Society 103: 2469-
2471.

TAKAGL T,, HAYASHI, K, & ITABASHI, Y. 1984,
Toxic effect of free unsaturated fatty acids in the
mouse assay of diarrhetic shellfish toxin by in-
traperitoneal injection, Bulletin of the Japanese
Society of Scentific Fishenes 50; 1413-1418.

TINDALL, D.R., DICKEY, . R.W., CARLSON, R.D.
& MOREY-GAINES, G. 1984. Ciguatoxigenie
dinoflagellates from the Caribbean Sea, Pp, 225—
239 In Ragelis. E.P. fed.), ‘Seafood toxins.’
(Amencan Chemical Society: Washington).

TINDALL, D.R., MILLER, D.M. & TINDALL, P.M.
1990. Toxicity of Ostreopsis lentictlaris from the
British and United States Virgin Islands. Pp. 424—-
429 In Graneli, E., Sundstrém, B.. Edler, L. &
Anderson, D.M. (eds), ‘Toxic marine
phytoplankton." (Elsevier: New York).

MEMOIRS OP 'THE QUEENSLAND MUSEUM

TOSTESON, T.R.. BALLANTINE, D.L.. & DURST,
H_D. 1988. Seasonal frequency of ciguatoxic bar-
racuda in southwest Puerto Rico, Toxicon 26:
7193-801,

VERNOUX, J.-P. & TALHA, F. 1989. Fractionation
and purification of some muscular and visceral
ciguatoxins extracted from Caribbean Fish. Com-
oye Biochemistry and Physiology 94B: 499—

WITHERS, N. 1982, Toxin production, numbton, and
distnbution of Gambierdiscus toxicus (Hawaiian
strain). Pp. 449-451 In Gomez, E.D., Birkeland,
C.E,, Buddemeier, R.W,, Johannes, RE. Marsh,
JA. & Tsuda, R.T. (eds), ‘Proceedings of the
Fourth International Coral Reef Symposium,
Manilla, vol. 2° (Marine Sciences Center, Univer-
sity of the Philippines: Manilla).

YASUMOTO, T. 1990, Marine microorganisms toxins
- an overview. Pp. 3-8 In Graneli, E., Sundstrim,
B,. Edler, L, & Anderson, D.M. (eds), ‘Toxic
marine phyineaniin (Elsevier: New York),

YASUMOTO, T., BAGNIS, R. & VERNOUX, JP.
1976, Toxicity of the surgeonfishes-I}. Properties
of the principal water- soluble toxin. Bulletin of
the Japanese Society of Scientific Fisheries 42:
359-365.

YASLUMOTO, T,, BAGNIS, R., THEVENIN, 5. &
GARCOM, M. 1977a. A survey of comparative
toxicity in the food chain of ciguatera. Bulletin of
ihe Japanese Society of Scientific Fisheries 43:
1075-1019.

YASUMOTO, T., NAKAJIMA, L, BAGNIS, R. &
ADACHI, R. 1977b. Finding of a dinoflagetlate
as a likely culprit of cignatera. Bulletin of the
Japanese Society of Scientific Fisheries 43) 1021-
1026,

YASUMOTO, T., NAKAJIMA, L, OSHIMA, ¥. &
BAGNIS, R, 1979a, A new. toxic dinoflagellate
found in association with ciguatera. Pp.65—70 In
Taylor, D.L, & Seliger, H.H. (eds), “Toxic
dinoflagellate blooms.’ (Elsevier/ North-Holland:
New York).

YASUMOTO, T., INOUE, A., BAGNIS, R. & GAR-
CON, M. 1979b. Ecological survey on a
dinoflagellate possibly responsible lor the induc-
ion of ciguatera, Bulletin of the Japanese Society
of Scientific Fisheries 45: 395-399,

YASUMOTO, T., OSHIMA, Y., MURAKAML, Y.,
NAKAJIMA, L, BAGNIS, R. & FUKUYO, Y.
1980. Toxicity of benthic dinoflagellates found in
coral reef, Bulletin of the Japanese Socacty of
Screnufie Fisheries 46: 327-331,

YASUMOTO, T, RAJ, U, & BAGNIS, R. 1984,
‘Seafood poisonings in tropical regions.’ (Tohoku
University: Sendui),

YOKOYAMA, A., MURATA, M., OSHIMA, Y.,
IWASHITA, T, & YASUMOTO, T. 198K, Some
chemical propertics of maitotoxin, a putative cal-
cium channel agonist isolated from a marine
dinoflagellate. J. Biochemistry 104; 184-187,

THE ORIGIN OF CIGUATERA IN PLATYPUS BAY, AUSTRALIA

MICHAEL J. HOLMES. RICHARD J, LEWIS, MICHELLE SELLIN AND RAEWY'N STREET

Holmes, M.J., Lewis, R.J., Sellin, M. & Street, R, 1994 08 01: The origin of ciguatera in
Platypus Bay, Australia. Memoirs of the Queensland Museum 34(3), 505-5 |2. Brisbane.
TSSN 0079-8835.

Platypus Bay on the northwestern side of Fraser Island is the only site in Queensland known
to frequently harbour ciguateric fishes, Platypus Bay is not typical of areas normally
associated with ciguateric fishes as it contains no corals but has a sandy bottom covered with
an unattached green macroalgae (Cladophora sp.). Benthic biodetritus samples sieved from
the Cladophara during seven sampling trips between May 1988 and February 1990 contained
Gambierdiscus toxicus with mean population densities of 4-556 cells per gram of
Cladophora. Biodetritus samples from six of the trips were extracted tor toxins. Putative
major and minor gambiertoxins (precursors of the ciguatoxins) were detected, suggesting
that these G. toxicus populations are the origin of the toxins in ciguateric fishes caught in
Platypus Bay, However, gambiertoxins were detected from only one of the six samples. This
indicates that not all strains of G, toxics produce these toxins in the wild, The concentrations
of major and minor gambiertoxins produced by these wild G, toxics cun be considerably
greater than the highest levels found from cultured G. texicus clones isolated from Platypus
Bay. ‘Super-producing’ strains of G, fexicus are hypothesised to explain thé high concentra-
tions of these toxins,

Michael J. Holmes, Richard J, Lewis, Michelle Sellin and Raevews Street, Southern Fisheries
Centre, Queensland Department of Primary Industries, PD Box 76, Deception Bay,

Queensland 4508; 22 November, 1993.

Ciguatera is a disease caused hy eating fish
contaminated with ciguatoxins which are thought
to originate from Jess-oxidised precursors called
gambiertoxins that are produced by the benthic
dinoflagellate Gambierdiscus toxicus. Gambier-
toxins are produced by only a minority of cultured
G. toxicus strains (Holmes et al., 1991). In
Australia, G. toxicus has been found along the
east coast from Flat Rock, E of Brisbane, to
Alexandra Reef N of Cairns, co-occurring with
corals (Gillespie et al., 1985a; Gillespie, 1987).
G. toxicus is the most common benthic
dinoflagellate on Queensland coral reefs, except
on some reef flats where Ostreopsis spp.
dominates (Gillespie et al., 198Sa). Population
densities of G. taxicus on Queensland reefs are
generally <5 cells’g of substrate. Flinders Reef in
south Queensland js an exception which seas-
onally produces >1,800 G, toxjeus/g of macroal-
gal substrate (Gillespie et al., 1985a). However,
gambiertoxins and ciguatoxins have not been ex-
tracted from G. toxieus or fish from Flinders Reef
(Gillespie et al.,1985b; Lewts et aj.,1988a).

Platypus Bay, on the northwestern side of
Fraser Island (Fig.1), is the only site in Queens-
land known to frequently harbour ciguateric
fishes (Lewis & Endean,!1983,1984; Gillespie ce
al,, 1986; Lewis,1987), Mackerels (Scom-
lvromorus spp.), especially narrow-barred

Spanish mackerel (8. cammersoni) caught near
the mouth of Wathumba Creek, have caused most
of these poisonings (Lewis & Endean, 1983; Gil-
lespie et al. 1986, Lewis,1987; Lewis et al.,
1988b). Since prohibition, in 1987, on the capture
of S. cormmersoni and barracuda (Sphyraena jello
and Agriaposphyraena spp.) from Platypus Bay,
the majority of cases of ciguatera in Queensland
have been caused by demersal fish from the Great
Barner Reef (unpubl. data), However, small
demersal shes from Platypus Bay have caused
ciguatera, including the blotched javelin (Pamie-
dasys maculatus) (Lewis et al.,1988b) and less
often, the rabbitfish Siganus spinus (unpubl.
data), Ciguatera is typically associated with
fishes of coral reefs (Randall,1958: Bagms et al..
1989) so its occurrence in Platypus Bay where
corals are absent seems unusual. The benthic
biodetritus of Platypus Bay was examined for
dinoflagellates and extracted for toxins. Putative
gambiertoxins from a biodetnital fraction with
large numbers of G. toxicus indicates that the
origin of ciguatoxins in fishes caught in Platypus
Bay is likely to be G. roxicus in Platypus Bay,

MATERIALS AND METHODS

BIODETRITUS FROM PLATYPLS BAY
Large areas of the bottom of Platypus Bay

506

QUEENSLAND

Brisbane &

Wathumba
Xx Creek

Platypus Bay

H
1 nautical mile

Urangan
\

FIG. 1. Map of Platypus Bay, Fraser Island.
Cladophora samples were collected by scuba divers
from the site marked with a cross.

Fraser Island

(Fig.1) are perennially covered with several cen-
timetres of a non-attached green macroalgae
(Cladophora sp.). Samples of Cladophora were
collected by scuba divers during seven sampling
trips between April 1988 and February 1990 from
about 1.5 nautical miles E of the mouth of
Wathumba Creek in c.15 m of water (Fig.1).
Samples were collected in Bitran zip-lock plastic
bags, transported to Brisbane, stored at 4°C over-
night and then processed. Insufficient material
was collected during the July 1988 trip for
toxicity studies but this sample was examined for
benthic dinoflagellates. The contents of the plas-
tic bags from the remaining six trips were shaken
and poured onto 2mm, 500um, 250um and
45\.m diameter plankton mesh sieves. The filtrate
from the 451m sieve was vacuum filtered on
Whatman Number 1 filter paper; the filtrate ob-
tained after this was discarded. The fractions
retained by the 500um, 250%m and 45um
diameter sieves and the filter papers were ex-
tracted for toxins. A 590g sample of Cladophora
(retained on the 2mm sieve) collected in May

MEMOIRS OF THE QUEENSLAND MUSEUM

1988 was also homogenised and extracted for
toxins. Benthic dinoflagellates were identified
with a light microscope (Fukuyo,1981). Cell
numbers were calculated from counts of 1ml
subsamples of sieved plastic bag samples of
Cladophora (n=2- plastic bag samples) using a
Sedgewick-Rafter counting chamber. Scanning
electron micrographs of formalin (5%) fixed
biodetrital fractions were prepared (Holmes etal.,
1990). Coolia monotis was identified by SEM
examination of cultured cells isolated from a
Cladophora sample collected in July 1988.

SOLVENT EXTRACTION OF SIEVED FRACTIONS

Fractions were homogenised for 20 minutes in
acetone (3-times, 3:1, v:v, acetone:sample) using
a Ystral air powered homogeniser. The 45-
250pm fractions were additionally homogenised
with methanol (1-time, 3:1, v:v, methanol:
sample). The extracts for each size fraction were
pooled, vacuum filtered through Whatman GF/A
glass fibre filters and dried under vacuum. The
dried extracts were resuspended in 9:1 methanol-
water and then separated into hexane-, diethyl
ether-, butanol- and water-soluble fractions (Hol-
mes et al., 1990). Gambiertoxins (or ciguatoxins)
extracted from these samples would partition into
the diethyl ether fractions, whereas maitotoxins
partition into both the diethyl ether and butanol
fractions (Holmes et al.,1990,1991; Holmes &
Lewis,1992). Diethyl ether fractions which in-
duced gambiertoxin-like bioassay signs in mice
were further characterised after separation by
silicic acid column chromatography (Holmes et
al.,1990; Holmes & Lewis, 1992).

MOUSE BIOASSAY

Fractions were dried under vacuum and finally
freed of solvent under a stream of Nz, resus-
pended in 0.5 ml of 1% or 5% Tween 60 saline
and injected intraperitoneally (i.p.) into 18—21g
Quackenbush strain mice (either sex) at a maxi-
mum dose of 1g of dried fraction weight per kg
mouse body weight. One mouse was injected per
dose, with 2 or 3 mice injected per fraction.
Where appropriate, 10% and 90% of fractions
were injected (for example, where 10% was not
lethal and 90% was < maximum injectable dose).
Mice were observed intermittently over 24 hours
and signs and death-times recorded. Fractions
where considered non-toxic if injection of a max-
imal dose was not lethal. Lethal fractions were
characterised on the basis of the signs and death-
times displayed by mice compared with the
responses induced by authentic gambiertoxins,

CIGUATERA IN PLATYPUS BAY

507

FIG, 2. Scanning electron micrograph of sieved Platypus Bay biodetritus (45-250j.m) containing Gambierdiscus
toxicus. A G. toxicus is shown enclosed in the box. Scale bar = 50m.

ciguatoxins and maitotoxins. Gambiertoxins
were quantified using the dose vs death-time
equation for the major gambiertoxin; log (dose)=
3.2 log (14t7!), where dose is in mouse units
(MU) and t = death-time in hours (Holmes &
Lewis, 1992). One MU is defined as an i.p. LDso
dose.

Gambiertoxins, ciguatoxins and maitotoxins
are all slow acting (Holmes et al.,1990; Lewis et
al.,1991; Holmes & Lewis,1992). For each, i.p.
injection of a dose one-tenth of that which would
kill a mouse in 30 minutes, would also be lethal.
Fast-acting toxins were therefore defined as frac-
tions which killed mice in 30 minutes (or less) but
which were not lethal at one-tenth of this dose.
The limit of detection of gambiertoxins from
non-toxic fractions was calculated from the max-
imum amount of material that could be injected
into mice which was non-toxic, assuming 0.5 MU
of gambiertoxin can be detected from the bioas-
say signs displayed by a mouse.

RESULTS

Platypus Bay biodetritus contains (Fig.2) cells
of G. toxicus as described by Fukuyo (1981). G.
toxicus was found in sieved fractions collected
from all seven sampling trips with mean popula-
tion densities of 4-556 cells/g of Cladophora

(Fig.3). There was considerable variation in the
size of the G. toxicus populations, with the May
1988 sample having the highest cell densities and
May 1989 the second lowest. There was no ob-
vious relationship between G. toxicus population

(3) (4)
O . © Gambierdiscus toxicus
«7 500 - ~N (2) O Caolia monotis
3 4 Prorocentrum sp.
a
S 400 4
u
= (3)
od 300
: ;
a
>
© 200 4 “fe \
a /
o Jf
© 100 \ y, (3)
o- —a— La —3———4.

AMJJASONDJFMAMJJASONOJF

1988 1989
Date

FIG. 3. Population densities of G. toxicus, Coolia
monotis and Proracentrum sp. on Cladophora from
Platypus Bay. Samples collected May 1988 to
February 1990. Shown are means +1 standard error,
2 to 4 replicate samples as indicated by numbers in
parenthesis. Sieved fractions were tested for toxicity
for all samples except the July 1988 sample.

© Gambierdireuy tor(eur
& Water temperature
O Salinsty

250 - oe ps

Fd =

5 oo @ +33

=

eq 509 4 a) “ =

9 a of }a g

3 Py ae \ o 2 —

5 ad ae Sy \ i bal a 3 ] =
is if 4 “2 he

§ \ SA / €

6 \ tf 22 0

= ad \ 5

2 a

3 \ A 20 ze

= \ A

5 1807 \ vo \

AMIJASONDIFMAMI UV ASONDIF
1988 7989
Date

FIG, 4. Population densities of Gambierdiscus loxicus
and Platypus Bay sea water temperatures and
salinities. Samples were collected between May 1988
and February 1990, G. toxicus populations densities
indicated are means -! standard error.

densities and seawater temperature or salinity
(Fig. 4),

Benthic dinoflagellates C. monotis and
Prorecentrumsp. were also observed in biodetri-

MEMOIRS OF THE QUEENSLAND MUSEUM

tal fractions but with generally smaller popula-
tion densities than G, toxicus (Fig.3). A bloom of
C. monotis (176 cells/g of Cladophora) was ob-
served in July 1988 but this was less than half the
G, toxicus cell density in the sample.

The yields and toxicity to mice of hexane-,
diethy] ether-, butanol- and water-soluble frac-
tions extracted from Platypus Bay biodetritus
(Table 1) show that 28 (of 96) fractions were
lethal to mice. Half of the lethal fractions were
extracted from the 45-250j.m sized biodetritus
samples which encompasses the size range of G.
toxicus. The May and October 1988 diethy] ether
extracts of the 45-250um sized biodetrital
samples induced signs in mice that appeared as a
composite of gambiertoxin and maitotoxin ef-
fects (Table 1). The diethyl ether extracts from
the remaining four 45—-250p.m biodetrital frac-
tions (Table 1) were non-toxic or contained a few
MU of a fast-acting toxin(s). These four fractions
would each have contained no more than 5-6 MU
(assuming a limit of detection of 0.5 MU for the
mouse bioassay) and were not characterised fur-
ther, No other extracts induced gambiertoxin-like
signs in mice.

i. 1. Yield and ae of extracts 01 of sieved Cladophora Sp. samples from n Platypus Bay.

2.0-0.5 | Diethy! ether

Hexane
Diethyl ether
Butanol
Water

Hexane
Diethy! ether
Butanol
Water

Diethyl ether
Butanol
Water

. » Date of collection —

8-11-89
(8.1 ky)

ba Toxin

* Weight of Cladophora sp. sampled (shaken, sieved and hand squeezed free of water).

* Fraction weight.

* Toxicity (i,p.) to 18-21 |g mice (n =2-3 per fraction); blank, indicates the fraction was not lethal up toa maximum

dose of 1g fraction weight per kg mouse body weight;

MTX, indicates a lethal fraction that induced maitotoxin-like signs;
G/MTX, indicates a lethal fraction that induced a composite of gambiertoxin-like signs and MTX signs;
?, indicates a Jethal fracuion that induced neither ciguatoxin, gambiertoxin or maitotoxin-like signs;

F, indicates fast-acting toxin.

CIGUATERA IN PLATYPUS BAY

Silicic acid column chromatography of the
May 1988 diethyl ether extract of the 45-250j.m
sized sample revealed that the 97:3 and 9;1
chloroform-methanol fractions were lethal to
mice and induced signs identical to those from
gambiertoxins (including hind-limb paralysis).
Quantification of the 97:3 chloroform-methanol
eluent using the dose vs death-time equation for
the major gambiertoxin indicated 65+18 MU of
toxin (mean=standard error, n=3), equivalent to
1.3 x 10° MU/cell of G. toxicus. Approximately
2 MU of toxin was recovered from the 9:1
chloroform-methanol eluent (3.9 x 10°77 MU/cell
of G. toxicus). On the basis of mouse bioassay
signs and column chromatography, the toxins in
the 97:3 and 9:1 chloroform-methanol eluents
Were interpreted to be major and minor gambier-
toxins, respectively. The chloroform eluent from
the silicic acid column was non-toxic and the
methanol eluent contained a fast-acting toxin thal
induced maitotoxin-like signs in mice. Gambier-
toxins were not detected in eluents from a silicic
acid column of the diethy] ether extract from the
45-250};:m biodetritus sample collected in Oc-
tober 1988. The chloroform. 97-3 and 9:1
chloreform-methanol eluents of this sample were
non-toxic to mice, the methanol cluent contained
a maitotoxin-like toxin. The limit for detection of
gambiertoxins using the mouse bivassay was 2-3
MU of toxin for these 97:3 and 9:1 chloraform-
methanol eluents.

Only one hexane- and three water-soluble Frac-
lions were lethal to mice. Seventeen of the 28
lethal fractions (Table 1) induced maitotoxin-like
sighs in mice including all the butanol-soluble
extracts from the 45-250ym sized biodetrital
samples. Fast-acting toxins were detected in five
of the diethyl ether-soluble extracts. Nine of the
28 lethal fractions (including four containing
fast-acting toxins) produced signs in mice that
were not consistent with those produced by either
ciguatoxins, gambiertoxins or maitotoxins. The
hexane-, diethyl ether- and butanol-soluble frac-
tinns extracted from the sample of Cladophora
were non-toxic, the water-soluble fraction was
lethal and induced maitotoxin-like signs in mice.

DISCUSSION

Extraction of putative major and minor gam-
biertoxins from a Platypus Bay biodetritus frac-
tion containing large numbers of G. texicus
indicates that the origin of the ciguatoxins in
Platypus Bay fishes is likely to be the G. tacieus
in Platypus Bay, The exiraction of similar gam-

biertoxins from cultures of a G. foxicus clone
isolated from Platypus Bay (Holmes et al.,199];
Holmes & Lewis,]992) supports this hypothesis.
Presumably, the gambrertoxins are bictrans-
formed to ciguatoxins and accumulated in
Platypus Bay fishes. Lewis & Sellin (1992)
showed that the two known structural types of
ciguatoxin are accimulated in the flesh of two
species of ciguateric fish caught in Platypus Bay.
These results indicate that the toxins that cause
ciguatera do not originate only from coral reefs,
but can be produced wherever substrates exist for
G. toxicus and environmental conditions are
suitable for the growth of gambiertoxin- produc-
ing strains of this dinoflagellate, Gambiertoxins
have now heen detected from wild G. toxicus
harvested from the rhodophyle, /ania sp. (Murata
et al.,1990) and the chlorophytes Halimeda sp.
(Holmes ef al.,1991) and Cladephora (present
study), This indicates that gambiertoxin-produc-
tion by G. foxicus is not dependent upon a par-
cular class or species of macroalgal substrate.

The yield of the major gambiertoxin (1.3 x 105
MU/cell of G. taxicus) was similar to the yield
from wild cells from Kiribari (Holmes et al.
1991), and greater than that extracted from cul-
tured cells (Holmes et al.,1991; Holmes & Lewis,
1992). The yield of the minor gambiertoxin from
Platypus Bay wild cells (3.9 x 10°? MU/cell) was
also an order of magnitude greater than from
cultures of G. sexicus (Holmes & Lewis,1992),
Consequently, the ratio of major to minor gam-
bienoxin (MU:MU) was similar for wild and
cultured G. fexicus (1:0,03 and 1:0.04, respec-
tively). However, only the major gambiertoxin
was detected from wild cells from Karibati in-
dicating that the ratio of major:minor gambier-
(OXin is nel constant in all strains of G. toxicus
that produce gambicrioxins. The production of
different relative quantities of the two gambier-
toxins by G_ foxicus may partly explain the dif-
ferences in relative amounts of CTX-1,-2 and -3
that Lewis & Sellin (1992) found in the flesh of
ciguateric fishes caught from Platypus Bay, Sul-
ficient G, texicus were extracted in the October
1988 and November 1989 biodetrital samples to
have detected gambienioxins ifthe concentrations
of the major gambiertoxin per cell had been
similar to the levels in the May 1988 sample.
Holmes et al_ (1991) showed that not al! strains
of G, toxicus produce gambiertoxins in culture.
Our study shows that differences in toxin produc-
non between strains is not restncted to cultured
G. toxieus.

lt appears that some wild cells must be capable

510

of producing considerably greater amounts of
gambiertoxins than the mean cell toxicity ( 10°
MU/cell) of the toxic wild sample found in this
study, given that the wild sample would likely
contain a mixture of high and low gambiertoxin-
producing strains. The high gambiertoxin-
producing strains of G. toxicus (here after
referred to as ‘super-producing’ strains) may in-
crease the potential for ciguatera. Production of
gambiertoxins in these super-producers may be
partly controlled by environmental parameters.
Holmes & Lewis (1992) showed that the con-
centrations (or type) of gambiertoxins produced
by cultured clones of G. toxicus can change
during the time they are maintained in culture.
Therefore, the size of G. toxicus populations does
not necessarily reflect their potential to cause
ciguatera. The absence of ciguatera at Flinders
Reef in southern Queensiand, which seasonally
harbours more than 1,800 G. toxicus/g of macro-
algal substrate (Gillespie et al.,1985a,b; Lewis et
al.,1988a) could therefore be explained by an
absence of gambiertoxin (or super-gambiertoxin)
producing strains of G. toxicus .

A number of Prorocentrum species produce
toxins (Yasumoto et al.,1980; Nakajima et al.,
1981; Dickey et al.,1990) while C. monotis is
thought to be non-toxic (Yasumoto et al.,1980;
Nakajima et al.,1981; Tindall et al.,1984). Proro-
centrum and C. monotis are unlikely to have
contributed significantly to the toxicity of
Platypus Bay biodetrital fractions tested because
of the small numbers of cells (<9cells/g
Cladophora) present with G. toxicus. Additional-
ly, there is no evidence that toxins produced by
these species accumulate in fish to cause human
illness. Approximately 30% of biodetrital frac-
tions were lethal to mice and about 60% of these
fractions induced maitotoxin-like signs, includ-
ing all butanol-soluble extracts of the 45—250um
fractions. All of the 45—250,.m fractions con-
tained G. toxicus and therefore butanol-soluble
extracts of these fractions would likely contain
maitotoxins. Toxins which induce maitotoxin-
like signs in mice have been extracted from the
viscera of fishes (Yasumoto et al.,1976; Lewis et
al.,1988a; Lewis et al.,1991), but there is no
evidence for the bioaccumulation of maitotoxin
or maitotoxin-like toxins in the flesh of
Queensland fishes causing human illness. Mouse
bioassay signs alone are unlikely to be diagnostic
for maitotoxins as the benthic dinoflagellate
Ostreopsis siamensis also produces a toxin
(more-polar than maitotoxin) that induces similar
signs in mice (Holmes et al.,1988). Fast-acting

MEMOIRS OF THE QUEENSLAND MUSEUM

ciguatoxins have been reported from the flesh of
ciguateric fishes from the Caribbean (Vernoux &
Talha, 1989) but authentic ciguatoxins, gambier-
toxins and maitotoxins from Pacific Ocean sour-
ces have all proved to be slow-acting toxins in
mice (Holmes et al., 1990; Holmes & Lewis, 1992;
Lewis et al.,1991). The dose vs death-time
relationship in mice is therefore a useful method
for differentiating toxins that produce otherwise
similar bioassay signs in mice.

The G. toxicus populations in Platypus Bay are
the second highest reported from Queensland
(Gillespie et al.,1985a). However, these popula-
tions are smaller than the 4.5 x 10° cells/g of
macroalgae reported from the Gambier Islands
(Bagnis et al.,1985). Bagnis et al. (1990) found
100-fold increases in G. toxicus populations can
occur in less than two weeks. The environmental
parameters which control the size of G. toxicus
populations and the proportion of these cells
which produce gambiertoxins (and possibly
super-producers) need to be determined before
the potential risk of ciguatera can be predicted
from G. toxicus populations.

ACKNOWLEDGEMENTS

We thank Hazra Thaggard and Sarah Brice for
technical assistance and Noel Gillespie, Brett
Davidson, Jo Masel, Darren Smallwood, Stuart
Hyland, Brian Pascoe and David Brown for as-
sistance with collection of samples.

LITERATURE CITED

BAGNIS, R., BENNETT, J., PRIEUR, C. &
LEGRAND, A.M, 1985. The dynamics of three
toxic benthic dinoflagellates and the toxicity of
ciguateric surgeonfish in French Polynesia. Pp.
172-182. In D.M. Anderson, A.W. White & D.G.
Baden (eds), “Toxic dinoflagellates’. (Elsevier:
Oxford).

BAGNIS, R., LEGRAND, A.M., CRUCHET, P.H., DE
DEKKER, F., GENTHON, J.N. & PASCAL, H.
1989. Human intoxications by ciguateric
phycotoxins in French Polynesia: incidence: clini-
cal and epidemiological features in 1987. Pp. 277-
288. In S. Natori, K. Hashimoto & Y. Ueno (eds),
“Mycotoxin and phycotoxins’. (Elsevier: Amster-
dam).

BAGNIS, R., LEGRAND, A.M. & INOUE, A. 1990.
Follow-up of a bloom of the toxic dinoflagellate
Gambierdiscus toxicus in a fringing reef of Tahiti.
Pp. 98-103. In E. Graneli, B. Sundstr6m, L, Edler

CIGUATERA IN PLATYPUS BAY

& D.M. Anderson (eds), ‘Toxic marine
phytoplankton’. (Elsevier; New York),

DICKEY, RW, BOBZIN, 8.C.. FAULKNER, DJ.
BENCSATH, F.A, & ANDRZEJEWSKI, D,
1990. Identification of okadaic acid fram a Carib-
bean dinoflagellate Proraceninwn concavum.
Toxicon 28: 371—377..

FUKUYO, Y. 1981. Taxonomical study on benthic
dinoflagellates collected in coral reefs. Bulletin of
the Japanese Society of Scientific Fisheries 47:
967-978.

GILLESPIE, N. | 987. Possible origins of ciguatera, Pp,
171-179. in J. Covacevich, P. Davie & J, Pear
(eds), “Toxic plants and animals: a guide for
Australia’. (Queensland Museum: Brisbane).

GILLESPIE, N,C,.. HOLMES, M.J., BURKE, J.B, &
DOLEY, J. 19852. Distribution and periodicity of
Gambierdiscus toxicus in Queensland, Australia.
Pp. 183-188. In D.M. Anderson, A.W, While &
D.G, Baden (eds), ‘Toxic dinoflagellates’. (El-
sevier: Oxford).

GILLESPIE, N., LEWIS, R., BURKE, J. & HOLMES,
MJ. 1985b, The significance of the absence of
ciguatoxin in a wild population of G. fextcas. Pp.
337-441. In C. Gabric & B. Salvat (eds),
“Proceedings of the Fifth linemational Coral Reef
Congress, Tahiti, val.4°. (Antenne Museum-
Ephe: Moorea).

GILLESPIE, N.C., LEWIS, R.J., PEARN, J.,
BOURKE, A.T.C, HOLMES, M.J., BOURKE,
1B. & SHIELDS, W.J_ 1986. Ciguatera in
Australia: Occurrence, clinical features,
pathophysiology and management. Medical Jour-
nal of Australia 145; 584-590.

HOLMES, M.J., GILLESPIE, N.C. & LEWIS, RJ.
1988. Toxicity and morphology ot Ostreepsix ct
siamensis, cultured from a ciguatera endemic
region of Queensland, Australia, Pp. 49-54. Tn
Choat, J.H, etal. (eds). ‘Proceedings of the Sixth
International Coral Reef Symposium, Townsville,
vol.3". (6th International Coral Reef Symposium
Executive Committee: Townsville).

HOLMES, M.J,, LEWIS, R.J. & GILLESPIE, N.C.
1990. Toxicity of Australian and French
Polynesian strains of Gambierdiscus toxicus
(Dinophycease) erown in culture: characterization
of a new type of maitotexin. Toxicon 28: 1159-
1172.

HOLMES, M.J,, LEWIS, RJ, POLI, M.A, & GIL-
LESPIE, N.C, 1991. Strain dependent production
of ciguatoxin precursors (gambiertoxins) by Gam-
bierdiscus toxieus (Dinophyceae) in culture.
Toxicon 29: 761-775,

HOLMES, M.J. & LEWIS, RJ. 1992, Multiple gar-
biertoxins (ciguatoxin precursors) from an
Australian strain of Gambierdiscus toxicus in cul-
ture, Pp, 520-530, In P. Gopalakrishnukone &
C.K. Tan (eds), ‘Recent advances in toxinology
research, vol.2". (National University of Sine
gapore: Singapore).

LEWIS, R. 1987, Ciguaterain southern Queensland, Pp,

Stl

181-187. In J. Covacevich, P. Davie and J. Pearn
(eds), “Toxic plants and animals; a guide for
Australia’. (Queensland Museum: Brisbane),
LEWIS, RJ, & ENDEAN, R. 1983. Occurrence of a
Fgunonee like substance inthe Spanish mackerel
(Scomberamorus commersoni). Toxicon 21: 19-
2

LEWIS, RJ, & ENDEAN, R, 1984. Ciguatoxin from
the Mesh and viscera of the barracuda, Spkyraena
jello. Toxicon 22; 805-810,

LEWIS, RJ. & SELLIN, M. 1992, Multiple ciguatoxins
in the flesh of fishes. Toxicon 30; 915-919,
LEWIS, R.J,, CHALOUPKA, M-Y,, GILLESPIE, N.C,
& HOLMES, M.J. 1988b, An analysis of the
human response lo ciguatera in Australia. Pp.
67-72, In J,H, Choul et al. (eds), “Proceedings of
the Sixth Intemational Coral Reef Symposium,
Townsville, vol.3"_ (6th Imternational Coral Reel
Symposium Executive Committee; Townsville),

LEWIS, RJ,, GILLESPIE, N.C, HOLMES, M.J.,
BURKE, J.B.. KEYS, A.B.. FIFOOT, AT. &
STREET, R. 1988a. Toxicity of 5 eo ex-
tracts from demersal fishes at Flinders Reef,
southern Queensland. Pp, 61-65. In J.H. Choat et
al. (eds), ‘Proceedings of the Sixth Intemational
Coral Reef Sympossum, Townsville, vol.3", (oth
Intermational Coral Ree! Symposium Executive
Commiltec: Townsville).

LEWIS, RJ. SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K,, & SHELL, M.M, 199],
Purification aad characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidac)}. Toxicon 29: 1115-1127.

MURATA, M., LEGRAND, A.M. [SHIBASHI, ¥.,
FUKUI. M. & YASUMOTO, T. 1990. Structures
and configurations ol cigualoxin from the moray
cel Gyrmiorthorar pavenicus and its likely precur-
sor from the dinoflagetlate Gambierdiscus
foxiens, Journal of the American Chemical
Society 112: 4380-4386.

NAKAJIMA, L, OSHIMA, Y. & YASUMOTO, T,
1981. Toxicity of benthic dinoflagellates in
Okinawa, Bulletin of the Japanese Society af
Scientific Fisheres 47: 1029-1033.

RANDALL, JE, 1958. A review of ciguatera, tropical
fish poisoning, wilh a tentative explanation of its
cause. Bulletin of Marine Science 8: 236-267.

TINDALL, D.R., DICKEY, R.W.. CARLSON, R.D, &
MOREY-GAINES, G. 1984, Ciguatoxigenic
dinoflagellates from the Caribbean Sea, Pp, 225—
339. In E.P. Ragelis (ed), ‘Seafood toxins’.
{Amencan Chemical Sociely: Washington).

VERNOUX, J.-P. & TALHA, F. 1989. Fractionation
and purification of some muscular and visceral
ciguatoxins extracted from Caribbean Fish, Com-
parative Biochemistry and Physiology 94B: 499-
S04

YASUMOTO, T., BAGNIS, R. & VERNOUX, 1P.
1976. Toxicity of the surgeontishes-Il Properties
of the principal water soluble toxin. Bullen af

512 MEMOIRS OF THE QUEENSLAND MUSEUM

the Japanese Society of Scientific Fisheries 42: 1980. Toxicity of benthic dinoflagellates found in
359-365. coral reef. Bulletin of the Japanese Society of
YASUMOTO, T., OSHIMA, Y., MURAKAMI, Y., Scientific Fisheries 46: 327-331.

NAKAJIMA, I., BAGNIS, R. & FUKUYO, Y.

SURVEY FOR CIGUATERA FISH POISONING IN WEST HAWAII

DANA ICHINOTSUBO, AUDREY Y. ASAHINA, ERIC TITUS, STEWART CHUN,
TERRI-LYNN W.P. HONG, JULIE L. SHIRA] AND YOSHITSUGI HOKAMA

Ichinotsubo, D., Asahina, A-Y., Titus, E,, Chun, §., Hong, T.W.P., Shirai, J-L. & Hokama,
Y¥. 1994 O08 O1: Survey for ciguatera fish poisoning in west Hawaii, Memoirs of the
Queensland Museum 34(3): 513-522. Brisbane, ISSN 0079-8835,

Approximately 25-30 fishes have caused cigualera poisoning in > 100 individuals in Hawaii,
as reported annually by the State Department of Health. Generally, about 6-10 species are
involved including herbivores and carnivores. A specific site at Puako, on the island of
Hawaii was selected because of persistent outbreaks of fish poisoning in the first few months
of nearly every year due to Cheilinus rhodochrous (wrasse, po'ou), The survey consisted of
(1) algae and Gambierdiscus toxicus assessment; (2) fish analysis by immunological assay:
(3) following fish extraction testing in mouse toxicity assay, and (4) analysis with guinea pig
atrium for an effect on Na+ channels. The immunological assay showed borderline and
positive in more than 50% of species examined (herbivores and carnivores), Several species
of algae were found, including Jania sp. and Turbinaria ornate previously shown to be
associated with Gambierdiscus toxicus blooms. In 5 areas along 2 miles of shoreline,
Gambierdixcus toxicus was noted in 2 areas (9-291/g algae). Organic solvent extracts from
Crenochaetus strigosus and Acanthurus sandvicensis showed inhibition of the Na*+ channel
in the guinea pig atrium assay. The inhibition appears to be very similar in action to
tetrodotoxin.

Dana Ichinotsubo, Audrey Y, Asahina, Erie Titus, Stewart Chun, Terri-Lyin WP. Hone,
Julie L, Shirai and Yoshitsugi Hokama, John A. Burns School of Medivine, University af

Hawaii, Hawaii 96822, USA; 18 May, 1994,

Approximately 115 incidents of ciguatera fish
paisonings occurred in the South Kohala area
from 1971-1992, The number of illnesses is high-
ly underreported due to the flu-like symptoms of
ciguatera (Hokama,198&) which often Jead to its
misdiagnosis by physicians. On Oahu, at a dawn-
town restaurant in February 1991, approximately
21 individuals became ill (5 hospitalized) after
consuming rose colored wrasses Cheilinus
rhadochrous (po' ou). Cheilinus rhodochrous ex-
tracts have shown typical ciguatoxin-like activity
in the immunological, mouse toxicity and guinea
pig atrium analysis (Amra et al., 1990). The ac-
tive toxins of the po’ou flesh extract were found
in the 1:9 (methanol:chloroform) and 1:1
(methanol:chloroform) fractions in the silica gel
chromatography analysis. The 10% methanol
fraction contained 10M.U, in LOOme crude flesh
extract, while the 50% methanol fraction had
5M.U. in 100mg crude flesh extract (1 MLU.
(mouse unit), is the amount of fish extract that
kalls a 20g mouse in 24hrs). The origin of the
implicated fish was an area called Puako on the
South Kohala Coast of Hawaii (Fig.1)}.

Puako is on the leeward coast of Hawaii in the
district of South Kohala. Before 1957 there was
essentially no development along this coast.
However, over the years many residential dwell-

ings, a 3.5km accessible coastline road, heat
launching ramp and several public right of ways
have been developed. Although development has
progressed markedly there ure still no real
beaches in this area. Coarse sand and gravel make
up much of the ocean-land interface with its
underlying structure composed mostly of lava,
Therefore, the land tends to be very porous. Other
nearshore areas are bordered mostly by Prosopis
pallida (kiawe trees) and soil is sparse. The
limited soil probably leads to nutrients being dis-
tributed in small, restricted quantities along the
coast. However, as one nears the boat launching
area, soil leyels substanually increase which ac-
counts for the often turbid nature of the water as
one nears this location and this probably in-
creascs its nutrient content. Puakois an extremely
and region with a visible freshwater lens notice-
able during the diving expeditions along the
coast. This is the underground fresh water which
lends to lower the salinity of the ocean water in
some areas of Puako,

The Puako study area extends 3.4km from
Puako Bay to approximately lkm N of Pauoa
Bay; il was divided into areas A-D, cach 850m
long (Fig. 1). The study commenced in April 1992
and visits Were made every 2 months. Each visit
included collection and identification of fish and

MEMOIRS OF THE QUEENSLAND MUSEUM

a“ ao

ny
‘
*
.
=
.
4

20 (ADDRESS)
RT. OF WAY -~

(1992) ~ :
raat y
en ri H
02 (ADDRESS) -~ ;
/ RT. OF WAY :
PINE TREE ON LEFT Borow
BLUE HOUSE WITH Pie J; =
COCONUT TREE ON RIGHT age (
= (1989) Fis fan (1992) SZ. SF
| Lan @ Panive 152 (ADDRESS) eo,
RT. OF WAY at 4
teansoukalua : ordid a * 45 ca x
Py (1992) ae, ae

FIG, 1. Map of the puako area surveyed for ciguatera. The 2 mile shoreline is transected into 4 segments (A, B,

C and D) from north (A) to south (D),

algae, upon which were conducted: 1) identifica-
tion of various algae types, 2) analysis of fish by
the Solid Phase Immunobead Assay (S-PIA) for
ciguatoxin and related polyether compounds, 3)
analysis of algae for Gambierdiscus toxicus, 4)
analysis of fish extracts with the mouse bioassay,
5) analysis of fish extracts with the guinea pig
atrial assay, 6) water quality analysis. Part of the
concept of this approach followed Yasumoto et
al. (1980, 1984) and Hokama et al. (1993).

METHODS

IDENTIFICATION OF ALGAE

Algae from each station was placed into 50ml
conical centrifuge tubes containing 2% for-
malin/seawater solution. Samples were taken to
Professor Isabella Abbott (Department of
Botany, University of Hawaii) for identification.

ANALYSES FOR GAMBIERDISCUS TOXICUS
Algal specimens were collected either by hand

SURVEY OF CIGUATERA IN WEST HAWAII

A B
100 jl alcoholic 10°7M tetrodotoxin
flesh extract
(10 mg)

B

A B
100 il alcoholic 10°7M tetrodotoxin
fish viscera extract
(10 mg)

10°7M verapamil

10-7M verapamil

Cc D E

10°™M adrenergic Rinse

blockers (propranolol
and phentolamine)

Cc E

Rinse

FIG.2. Guinea pig atrial responses to Ctenochaetus strigosus (kole) alcoholic extracts and the effects of blockers.

or by gently scraping coral substrates throughout
areas A-D. Samples were then placed into | gal-
lon ziplock bags containing 0.5—11 of seawater.
The contents were shaken for 2min to remove any
possible epiphytic dinoflagellates present on the
algal substrate. The algae/seawater suspension
was then consecutively passed over a 125m
mesh sieve to remove any large algal fragments
then through a 38pm sieve. The residue on the
38j.m sieve was then backwashed with an en-
riched seawater media, collected into a 100ml
sterile glass bottle and loosely capped to provide
aeration. After gentle agitation, 1m] was placed
onto a Sedgewick Rafter Cell Counting Slide.
Counts were performed in triplicate to determine
the average number of cells per ml and the num-

ber of cells per gram of algae (Yasumoto et
al.,1984).

ANALYSIS OF FISH BY S—PIA

Various herbivorous and carnivorous fishes
were collected by spearing and line fishing. A
bamboo paddle coated with pentel correction
fluid was then inserted into an incision made near
the head of each sample and the residue retained
on the paddle allowed to air dry for 5—10mins.
The paddle was then immersed in methanol for
O.1sec and allowed to air dry. The paddle was
then put into a blue latex bead-antibody solution
(ciguatoxin antibody) for 5 or 10min intervals.
Readings obtained were then scored according to
the intensity of the coloring obtained on the pad-

516

A

100 jl alcoholic 10-7M tetrodotoxin

po'ou extract (10 mg)

B

A
100 jl alcoholic 10°7M tetrodotoxin

po'ou extract
(10 mg)

10°7M verapamil

MEMOIRS OF THE QUEENSLAND MUSEUM

ene

10-7M verapamil Rinse

Cc E

10°7M adrenergic
blockers (propranolol
and phentolamine)

Rinse

FIG.3. Guinea pig atrial responses to Cheilinus rhodochrous (po’ ou) alcoholic extracts and the effects of blockers.

dle and rated as negative (no color), borderline
(slight blue) or positive (distinct blue). The pro-
cedure followed Hokama (1990).

EXTRACTION OF CIGUATOXIN AND RELATED
POLYETHER COMPOUNDS

Fish were separated into flesh and viscera
samples and extracted for ciguatoxin and related
polyether compounds. Each portion (flesh and
viscera) was placed in acetone at a 1:2.5 ratio
(flesh:acetone) and allowed to soak overnight.
The acetone suspension produced was then
decanted over a Whatman 4 filter and the filtrate
allowed to collect into a flask. The acetone was
then evaporated utilizing a Buchi Rotavapor. The
residue retained in the flask was then washed with
chloroform and partitioned with a brine solution
(2x) at a 4:1 ratio (chloroform:brine) in a
separatory funnel. The chloroform was succes-

sively evaporated and the residue retained was
then brought up in a 80% methanol solution and
partitioned with hexane (3x) ata 1:2 ratio (80%
methanol:hexane). The 80% methanol was
evaporated and the residue was collected and
used for further analysis in the mouse bioassay
and the guinea pig atrial assay. The method of
Kimura et al. (1982), and Miyahara et al. (1989)
were used in this study.

ANALYSIS OF FISH EXTRACTS WITH MOUSE
BIOASSAY

100mg of the 80% methanol fraction fish ex-
tract was diluted in Im] of a Tween 60/saline
solution. The suspension was then injected in-
traperitoneally (IP) into a 20-25g Swiss Webster
mouse and signs observed. Readings were suc-
cessively taken at 30min, 1,2,4,6,8,24, and 48hrs
post-injection and rated on a scale of 0-5 accord-

SURVEY OF CIGUATERA IN WEST HAWAII 517

TABLE 1. Survey of G. toxicus at Puako, April to August, 1992. * areas refer to sections shown on Fig.1.

B APR 92 0 Sargassum sp.
JUN 92 0 Turbinaria ornata
Pterocladia caerulescens
AUG 92 Eupogodon iridescens

C_ |APR 92

AUG 92
D | APR 92

A | APR 92 291 Galaxaura fasciculata
Tolypiocladia glomerulata
JUN 92 7.2 Galaxaura marginata

AREA) DATE | G. TOXICUS ALGAL SPECIES TOTAL S-PIA S-PIA S-PIA
* CELLS/GM #FISH NEGATIVK BORDER | POSITIVE
ALGAE LINE

74 18% (13) | 55% (41)
76 | 43% (33) | 43% (33) | 13% (10)

86 30% (26) | 63% (54) | 7% (6)

Galaxaura fasciculata

Galaxaura marginata with large
clumps of Biddulphia aurita (diatom)
Galaxaura fasciculata
Turbinaria ornata
epiphytic Tolypiocladia glomerulata
Dictyota friabilis (brown)
Chondria polyrhiza, Jania sp.,

Cladophora sp. (green)

Schizothrix calcicola (bluegreen)
Centroceras clavulatum

Turbinaria ornata

Schizothrix calcicola

Turbinaria ornata with epiphytic

Jania

ing to the signs presented. Toxicity ratings are
scored according to Kimura et al. (1982).

ANALYSIS OF FISH EXTRACTS WITH GUINEA
PIG ATRIAL ASSAY

Hartley Guinea Pig weighing approximately
300—350g were sacrificed and hearts surgically
removed after anesthetization (Enflurane). The
atria was then isolated from the ventricles and
separated further into its left and right com-
ponents. Each piece of atrial tissue was placed
into a 25ml physiological bath (37°C) containing
Krebs-bicarbonate solution and constantly
aerated (95% O2, 5% COz). Electrical probes
stimulated the left atria while the right relied on
its sinoatrial node. Subsequently, 1001] of fish
extract (80% methanol fraction) at a concentra-
tion of 100mg/ml resuspended in Krebs bicar-
bonate solution was added to the physiological
bath chamber. The tissue was then observed for
both an inotropic and/or chronotropic responses
displayed on a polygraph as reactions to the ex-

JUN 92 3.4 Turbinaria ornata with Jania
Pterocladia caerulescens

JUN 92 1.2 Turbinaria ornata with Jania ,
bluegreen and Biddulphia aurita
(diatom) :
| AUG 92 0 Phormidium crosbyanum

tract. Additionally, pharmacological charac-
teristics were noted with the addition of 12.51 of
tetrodotoxin (sodium channel blocker), vera-
pamil (calcium channel blocker) and propranolol/
phentolamine (adrenergic blocking drugs) fol-
lowing the method of Miyahara et al. (1989). In
one experiment (Fig.5) Manini extract from the
viscera was used in the atrial analysis. The extract
acted as a blocking agent. This action though
resembling TTX or PSP was neither of these
toxins, since it had a non-polar lipid charac-
teristic.

WATER QUALITY ANALYSIS

Physical parameters of the Puako coastline
(Areas A-D) were taken by the Department of
Health Clean Water Branch in December 1989
and August 1992. Measurements including
salinity, nitrate-nitrite, ortho-phosphates, am-
monia, silica, total dissolved nitrogen, total dis-
solved phosphorous, temperature, conductance,
dissolved oxygen and pH were taken utilizing a

518

TABLE 2. Puako fish assessment with S-PIA (April
1992). * Area refers to sections shown in Fig.1.

SPECIES TOTAL] S-PIA RESULTS
= +/- +

A |Hawaiin kole 1 0

1 0
Kole 9 0 8

1
¥ 6 2 3 1
1

AREA
*

Percent
D_ |Kaku (S. barracuda)
Kole

Percent
41 15

TOTAL 74
(%) (69) | (20)

Surveyor 2 water quality monitoring instrument.
Testing areas for those parameters measured in
December 1989 and August 1992 were within the
areas of A through D (Fig. 1).

RESULTS

IDENTIFICATION OF ALGAE

Puako contains a variety of algal types (Table
1), some of which have been found associated
with G. toxicus previously.

ANALYSIS OF ALGAE FOR G. TOXICUS (Table 1)
The highest level of cells (per g of algae) oc-
curred in area A during April 1992. G. toxicus
was found at all stations (in limited quantities),
except for Area B which was devoid of any
dinoflagellates over the survey period. The
moderate to high levels of G. toxicus were espe-
cially shown in April of 1992 in Area A and this
was associated with 7. glomerulata. Area B had

MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 3. Puako fish assessment with S-PIA (June
1992). * Area refers to sections shown on Fig.l. #
Not defined in Fig.1, but represents an area of similar
size north of A. A is the first transect which originates
at the boat ramp.

~ SPECIES |TOTAL| _S-PIA RESULTS

| 2 | 0
ree to fo
Pest

(60)

Percent

B_ | Hawaiin kole
Kole
Po’ou

Table boss (spp.)

Total 16 5 7 4
Percent (31) | (44) | (25)
C_ |Hawaiin kole 5 1 4 0

Percent

an or Vee
Po’ou 4 i) 3
Roi 4 0 2

Total
Percent

no G. toxicus despite a variety of algae. Areas C
& D had mild levels of G. toxicus probably as-
sociated with the epiphytic Jania sp.

ANALYSIS OF FISH WITH S-PIA (Tables 2-4)
The targeted species in this Puako Survey in-

cluded a herbivore Ctenochaetus strigosus (kole)

and 2 carnivores Cephalopholis argus (roi);

SURVEY OF CIGUATERA IN WEST HAWAII

A C D E
100 jl alcoholic 10-7M tetrodotoxin 10-7M verapamil 10-7M adrenergic Rinse
roi extract blockers (propranolol
(10 mg) and phentolamine)

B

A B
100 jl alcoholic 107M tetrodotoxin
Toi viscera extract
(10 mg)

10°7M verapamil

D E

10°7M adrenergic Rinse
blockers (propranolol

and phentolamine)

FIG,4, Guinea pig atrial responses to Cephalopholis argus (roi) alcoholic extracts and the effects of blockers.

Cheilinus rhodochrous (po’ou). The highest
overall percentage of fish falling into the rejec-
tion category of S-PIA (borderline and positive)
occurred during April (1992) in Areas A, B, and
C. Area D had the greatest percentage of fish in
the rejection range in August 1992.

ANALYSIS OF FISH EXTRACTS WITH THE MOUSE
BIOASSAY (Table 5)

Both flesh and viscera contained compounds
that induce abnormal symptoms and death in
mice. Although all fish tested showed both high
and low levels of toxicity, kole viscera on the
average gave consistently the highest toxicity

ratings. In all species, toxicity seemed to be
higher in the viscera as compared to the flesh
extract. There is a slight difference in the flesh
and viscera toxicities in mice. The viscera ex-
tracts tended to be most consistently toxic and are
noted in all areas and periods examined. The
mouse toxicities of po’ou extracts from Areas A
and D appear less severe than kole extracts.
Again, within the species tested the viscera were
more toxic especially in Areas B andC. Similarly,
the roi extracts tended to be less toxic than the
kole extracts. Again, the flesh extracts of roi
appear to be less toxic than the roi viscera extract.

520

TABLE 4, Puuko fish assessment with S-PIA (August
1992), * Area refers to sections shown on Fig.1. #
Not defined in Fig.1, but represents an area of similar
size north of AA, as described in Table 3-

S-PIA RESULTS

4 | — | | + |

A |Kole 20 | 12 [ 2 |

ae
fe.
ha
|

Po'ou 3
Rai 3

Pp
Ss fo |S.
im 1o
> |e
au

tt

ta

ag te

iv

A)

wi

aa

-

=

g (A. achilles)
Total
Percent

!
a

~
PF

:

Nw

|
i]

Ww

Total
Percent

c

al
7
iS

| 1 | 2 | 0

Pte asso 0]

| Perce |_| (33) | (67) | (©) |
Kole | wo | o | 6 | 4 |
Povou =| a | oo | i | oo |
Roi S| ol | | 4 |
Total o |u| 4
| coy | 073.3) [ 26.7)

FH [Kole | ao | 4 [| 6 | o |

| Total | et

[Percent | (40) | (60

ze, le ladle
___(PERCENT) ___| 30.2) Q)

ANALYSIS WITH THE GUINEA PIG ATRIAL
ASSAY

Testing of fish extracts with the Guinea Pig
atrial Assay indicated at least 2 pharmacological-
ly different toxins. The inotropic effect elicited
by one of these compounds is blocked by
tetrodotoxin, indicating that the observed reac-
tion has characteristics similar to ciguatoxin by
allowing a greater influx of sodium ions through
the membrane channel, Likewise, the second
reaction observed is an inhibition by verapamil in
response to the inotropic effect elicited by the fish
extract. Both types of reactions (TTX and
verapamil inhibitions) are found in the herbivores
as well as the carnivores.

MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 5. Compilation of mouse toxicity of three
species of fish: viscera and flesh for the months of
April, June and August, 1992, * mouse toxicity
values ranged from I (little or no toxicity) to 5 (not
toxic). NS =no ) sample.

ee
Tape [sun [Aue
p+ ts

| Ctenochaetus Desh EuEe :

B Ctenochaetus eed eit Tne eet
strigosus vivcugti 515 |5
Crenochaerus [fen { 3 ts :
Ctenochaetus flesh eal :
strigosus Viseera|. 8 raz:
Cheilinus races
rhodachrous visceral _S_{ NS | 3
Cheilinus lus. | s_|
rhodochraus aT ih Ps fs [oI
Cheilinus
rhadochrous pase aS Tas [J
Cheilinus NS
rhadechrous

Meese Ne |e a
Cephalopholis argus Pea 2 3 1s I
5
Cephalopholis argus | fesh | 2 | 1 | 5 |
Cephalopholis argus bests ts ts

yviscera| 5

Cephalopholisargus| Mesh | 2 | 2 | 3 |

viscera] 4 | 5 | NS_

Inotropic patterns and their inhibition by
tetrodotoxin, verapamil or both are demonstrated
(Figs 2-4) for kole, po’ou and roi, respectively.
Kole flesh and viscera extracts show typical in-
hibition by TTX and verapamil (Fig,2A.B). How-
ever, insome kole viscera cxtracts only verapamil
or TTX showed inhibition.

Po'ou (whole fish) extract induced positive in-
otropy is inhibited by TTX and verapamil
(Fig-3A). The latter inhibitor appears to give a
stronger reaction. A pooled po'ou extract scored
negative to the S-PIA test (negative for
polyether); this is reflected by no inhibition with
TTX channel blocker, but inhibition by
verapamil(Figure 3b).

The major difference between the inotropic
responses of whole roi (Fig.4A) and separated
viscera extracts (Fig.4B) is in the initial rise. Both
extracts are affected by TTX and verapamil, al-

SURVEY OF CIGUATERA IN WEST HAWAII

DOH
5 mg/bath

fish extract
10 mg/bath

ae
107 M

FIG.5. Inotropic response of pooled Department of Health (DOH) fish extract implicated in ciguatera poisoning
and the negative inotropic effect of manini (Acanthurus sandvicensis). This DOH extract was toxic for mice
(Value of 5) and its inotropic pattern showed strong block with tetrodotoxin (pattern not shown).

though the viscera extract appears to be affected
less than the whole extract. There 1s a suggestion
that the toxin in the viscera is slightly different
from whole roi extract.

Fig.5 shows the inhibition of the inotropic ef-
fect of known ciguateric pooled fishes (Depart-
ment of Health samples) by a sample from
Acanthurus sp. extract (manini) giving a TTX or
verapamil like block. This has been shown on
occasion with kole extracts also.

WATER QUALITY ANALYSIS

The Puako area is associated with underground
freshwater infusion into the ocean (Kay et
al.,1977), especially in Areas B and C - Salinity
has varied from a low of 2.2% to 3.4% in Areas
B and C. This fluctuation in part may account for
the variation m G. toxicus numbers.

DISCUSSION

There appears to be no relationship between
high levels of toxicity in fish and presence or
absence of G. toxicus: even in the absence of G,
toxicus, levels of SPIA borderline and positive
fishes were noted. April,1992 showed the least
number of negative fish in all species and areas
(18%) followed by August (30%) and then June
(43%) (Tables 2,3,4). In general, the herbiyores
appeared to be more toxic (+and +) than the
carnivores in all three months examined.

The mouse toxicity showed that the herbivore
(kole) was the most toxic with higher number of
mouse values of 5. The carnivores roi and po'ou
were also toxic but less than the kole, except for
the viscera extracts. Analysis of the guinea pig
inotropic response suggested the ciguatoxin-like,
maitotoxin-like (Figs 2,3,4) and a sodium chan-
nel blocking toxin (polyether-like) similar to
TTX or verapamil (Fig.5).

It is difficult to correlate immunological results
with extracts used in the mouse and atrial assays,
since the immunoassay is carried out in in-
dividual fishes whereas the mouse and atrial as-
says are done with pooled extracts though of the
same species. Nevertheless in general, those
fishes with borderline and positive SPIA results
appeared to be more toxic than the negative
fishes.

LITERATURE CITED

AMRA, H., HOKAMA, Y., ASAHINA, A.Y.,
SHANG, E.S. & MIYAHARA, LT. 1990,
Ciguatera toxin in Cheilinus rhedochrous (Po'ou,
Wrasse). Food and Agicultural Immunology 2:
119-124,

CAMPBELL, B., NAKAGAWA, L.K.,
KOBAYASHI, M.N. & HOKAMA, Y. 1987.
Gambierdisows toxicus in gut content of
Ctenochgetus strigesuy (herbivore) and its
relationship to toxicity in tissue. Toxicon 25:
1125-1128.

HOKAMA, Y., SHIRAI, L.K.. IWAMOTO, L.M.,
KOBAYASHI, M.N., GOTO, C.S. &
NAKAGAWA, L.K. 1987, Assessment of a rapid
enzyme immunoassay stick lest for the detection
of ciguatoxin and relaicd polyether toxins in fish
tissues. Biological Bulletin 172; 144-153,

HOKAMA, Y. 1988. Ciguatera fish poisoning. Journal
of Clinical Laboratory Analysis 2: 44-56.

HOKAMA, Y. 1990. Simplified solid-phase im-
munobead assay for the detection of ciguatoxin
and related polyether in fish ussue, Journal of
Clinical Laboratory Analysis 7; 26-30.

HOKAMA, Y., ASAHINA, 4.Y., TITUS, E., SHIRAL,
J.L., HONG, T.W.P., CHUN, 8., M[IYAHARA,
J.T., TAKATA, B., MURANAKA,E., PANG. E.,
ABBOTT, I. & ICHINOTSUBO, D. 1993, A
survey of ciguatera: assessment of Puako, Hawali
associated with ciguatera toxin epidemics in
humans. Jourmal of Clinical Laboratory Analysis
7: 143-154.

™ Rinse

522 MEMOIRS OF THE QUEENSLAND MUSEUM

KAY, E.A., LAU, L.S., STROUP, E.D., DOLLAR, terization of the toxins in different fish species. Pp.
S.J., FELLOWS, D.P. & YOUNG, R.H.F. 1977. 399-406. In Natori, S., Hashimoto, K. & Ueno, Y.
Hydrologic and ecologic inventories of the coastal (eds), ‘Mycotoxins and phycotoxins ’88’. (El-
waters of west Hawaii. Water Resources Research sevier: Amsterdam).

Center, University of Hawaii, Honolulu, Hawaii, YASUMOTO, T., INOUE, A., OCHI, D., FUJIMOTO,
Technical Report 105. K., OSHIMA, Y., FUKUYO, Y., ADACHI, R. &

KIMURA, L.H., HOKAMA, Y., ABAD, M.A., BAGNIS, R. 1980. Environmental studies on a
OYAMA, M. & MIYAHARA, J.T. 1982. Com- toxic dinoflagellate responsible for ciguatera. Bul-
parison of three different assays for the assessment letin of the Japanese Society for Scientific
of ciguatoxin in fish tissues: radioimmunoassay, Fisheries 46: 1397-1404.
mouse bio-assay and in vitro guinea pig atrium YASUMOTO, T., RAJ, U. & BAGNIS, R. 1984.
assay. Toxicon 20: 907-912. ‘Seafood poisoning in tropical regions’.

MIYAHARA, J.T., KAMIBAYASHI, C.K. & (Laboratory of Food Hygiene, Faculty of Agricul-

HOKAMA, Y. 1989. Pharmacological charac- ture, Tohoku University; Tohoku) 74p.

TEST OF THE EFFECT OF DISTURBANCE ON CIGUATERA IN TUVALU
URSULA L. KALY AND GEOFFREY P. JONES

Kaly, U.L. & Jones, G.P. 1994 08 01: Test of the effect of disturbance on ciguatera in Tuvalu,
Memoirs of the Queensland Museum 34(3): 523-532. Brisbane. ISSN 0079-8835.

We describe a field study on the potential link between the occurrence and intensity of
ciguatera outbreaks and human disturbance of coral reefs. We focused on the atolls Niutao,
Nui and Nanumeain the Tuvalu group; each having a different history of ciguatera. Relatively
small-scale disturbances associated with ship wrecks; and moderate disturbance associated
with blasting for boat channels were examined. Increases in Gambierdiscus toxicus abun-
dances were detected around channels at Nanumea and Niutao (prior history of outbreak),
but not at Nui (historically ciguatera free). At Niutao, the overriding pattern of G. toxicus
density around the island was independent of either form of human disturbance. Fish toxicity
(Hokama Stick Test) data indicated a similar pattern unrelated to human disturbance. We
suggest that some forms of human disturbance might affect (or even precipitate) outbreaks
of ciguatera but that other factors are likely to play a larger role.

Ursula L. Kaly and Geoffrey P. Jones, Department of Marine Biology, James Cook

University, Townsville, Queensland 4811, Australia; 28 March, 1994.

Ciguatera poisoning from coral reef fishes is a
problem for Pacific Island countries, both for
expansion of subsistence fishing to support in-
creasing populations, and for inshore commercial
fisheries (Dalzell,1992; Lewis, in press). While
knowledge has increased on the toxins involved
(Murata et al.,1990; Lewis et al.,1991), the dino-
flagellates responsible for ciguatoxins (Gambier-
discus toxicus and others) (Lewis et al. 1991;
Holmes & Lewis,1991; Holmes et al.,1991), and
the fish responsible for intoxications (Anderson
& Lobel,1987; Dalzell,1992), the causes of out-
breaks elude us. The spatial and temporal pat-
terns in the abundance of G. toxicus and ciguatera
poisonings paint a complex picture. Some areas
appear never to have outbreaks. Other areas show
severe short-term but unpredictable outbreaks
lasting a few years (Kaly et al.,1991), hereafter
‘acute outbreaks’. Other ‘hot spots’ have had
ciguatera problems for decades, sometimes with
a marked seasonality (Carlson & Tindall, 1985;
Gillespie et al.,1985a), hereafter ‘chronic
outbreaks’. Field data of dinoflagellate numbers
relative to ciguatera outbreaks are rare.

Speculations on the cause(s) of ciguatera out-
breaks include natural and human-induced
phenomena. Anecdotal and correlative studies
have implicated high salinity (Yasumoto et al.,
1980b; Carlson, 1984; Taylor,1985), low rainfall
(Carlson & Tindall,1985), low storm or wave
activity (Taylor & Gustavson 1985), high levels
of nutrients (Carlson,1984), physical destruction
of reefs, both natural and man-made (Randall,
1958; Banner,1976) coral death or bleaching
(Yasumoto et al.,1980a,b; Kohler & Kohler,

1992) and even nuclear testing (Bagnis,1969).
However, no one factor provides a universal ex-
planation and there have been few attempts to test
specific hypotheses.

The notion that the physical destruction of reef
areas promotes ciguatera, is based on the idea that
freshly denuded surfaces are colonised by certain
opportunistic macroalgae that are the preferred
hosts for the epiphytic dinoflagellates responsible
for ciguatera (Randall,1958; Banner,1976). Sup-
port comes from studies on a variety of disturban-
ces including the construction of boat channels
(Kuberski,1979; Banner,1974; Tebano,1984,
1992), the locations of wrecks or anchorages
(Cooper, 1964), and storm-induced and other dis-
turbances to the reef system (Bagnis,1969; Bag-
nis et al.,1985). However, testing the importance
of these factors requires the simultaneous sam-
pling of disturbed and undisturbed sites, and
preferably, sampling both before and after the
disturbance has occurred.

Our study examines potential effects of boat
channels and shipwrecks on incidence of
ciguatera in Tuvalu. We used 2 field indicators to
test the potential effects of these disturbances.
Firstly, abundance of Gambierdiscus toxicus
was measured on a range of host algae. Secondly,
we used an index of toxicity of muscle tissue from
a surgeonfish (Crtenochaetus strigosus) impli-
cated in ciguatera, using the Hokama ‘stick’ test
(Hokama et al.,1983,1987). We assumed that
these techniques may independently provide a
measure of the levels of ciguatera at different sites
or times without there necessarily being a direct

524

P

MEMOIRS OF THE QUEENSLAND MUSEUM

(A) NIUTAO Muli Ne
\ | NP
nV TRE, 7 en /
? a ie
Z
Z
{
i, ‘
4 1
ay '
‘
,
\—- OW
Kulia~ \
i
aioe if
/ * ’
i ae ll WO eee ee
! ~s OO NW
H ‘ \
\ sc ss
\ \, 0.5km
' .
i ‘
H \
1 ‘ GRID NORTH
f \ MAGNETIC
'
\ H REEF NORTH
' ~ \
1 d 1
1 * . ,
4 ‘ \ + Pin
‘ i : i eons
' \ ‘ ‘ H os
1 \ ‘ ' \ Pie
f > ‘ ss, bo,
fF f ‘§\ €)h \. em
‘ , ms Y, ~
ee Gs eee et
' & i 1 ~~ ina ie
1 H \ Ce ™ ae
1 ‘ 1 4 Pe,
\ | \ t
! i 1 ££ —"
1 ' ' *
t { ‘ ‘

Nc! \ H ‘ 03
1 \ : 1 AMERICAN cunon
1 H H \

1 ' } 4
H i J H KENNEDY ohiloeke ~

Peneih A !

H zg |

1

‘

‘

F

1 KILOMETRE

(B) NUI

9 2 KLOMETRES
|

(C) NANUMEA

FIG. 1. Maps of (A) Niutao, (B) Nui and (C) Nanumea showing channels, controls and wrecks surveyed. Muli
and Kulia are channels, NC=North Control, WC=West Control, SC=South Control, SS=Sagasaga, NP=North
Point, NW=New Wreck, OW=Old Wreck, EW=East Wreck.

link between dinoflagellate numbers and fish
toxicity.

Sampling protocols to test the effect of distur-
bance were: (1) a comparison between areas ad-
jacent to pre-existing channels and shipwrecks
with control areas, for numbers of G. toxicus and
for fish toxicity; and (2) a short-term study of
changes occurring in response to the construction
of two new boat channels, and distant control

areas. By sampling different distances from these
boat channels and control areas over an 18 month
period, we attempted to demonstrate any succes-
sional effects on ciguatera that could be attributed
unequivocally to boat channels. Since an out-
break of ciguatera began just prior to the con-
struction of the channels on Niutao (Kaly et al.,
1991), the study provided a useful test of the

EFFECTS OF DISTURBANCES ON CIGUATERA IN TUVALU

BEFORE

/\

Channels Controls

525

AFTER

ee

Controls

\ “NA
AX. AN,

0 20 50 bis , 50 20 0 O 20

Muli Kulia NC WC S

FIG.2. Design of the sampling programme used to detect any impacts of channels on indicators of ciguatera at
Niutao. Numbers under each channel after construction, are distances away from the channel in metres.

effects of disturbance during a period when fish
toxicity was high.

METHODS

STUDY AREAS AND SAMPLING REGIMES

Construction of boat channels and ship wrecks
represent two of the most obvious forms of
human disturbance to coral reef platforms and the
shallow subtidal zone in Tuvalu. This study
focussed on pre-existing boat channels on
Nanumea (2 channels) and Nui (1 channel), pre-
existing ship wrecks on Nanumea (1 wreck) and
Niutao (2 wrecks), and two new boat channels
constructed on Niutao during the study. These
islands differed in history of recorded ciguatera
outbreaks, with none ‘ever recorded’ from Nui,
an area on Nanumea which appears to have been
toxic for many years, and a severe outbreak which
occurred on Niutao and recovered over a 2-4 year
period.

The Nui channel and two control sites on that
island were sampled during early September
1989 (Fig.1B). The two channels (American and
Kennedy), two control sites and the wreck site on
Nanumea were sampled in late September 1989
(Fig. 1C). On Niutao, 2 wreck sites (NW and OW)
and 4 control sites (NC, Muli, WC and NP) were
sampled prior to construction of any boat chan-
nels during May 1989 (Fig.1A).

The 2 future channel sites (Kulia and Muli) and

4 controls (NC, WC, SS, SC) (Fig.1A) were
sampled prior to construction of the channels
during January 1989 (Fig.2 for sampling design).
Sampling at these sites was repeated during chan-
nel construction (May 1989), 3 months after
channel construction (September 1989) and 15
months after channel construction (September
1990). Sampling was carried out adjacent to each
channel, 20m, 50m and 100m in both directions
away from the channels, to assess spatial extent
of any impacts on the abundance of G. toxicus.

MEASUREMENT OF GAMBIERDISCUS ABUN-
DANCE AND COVER OF HOST ALGAE

The distribution and abundance of Gambierdis-
cus toxicus, the only dinoflagellate potentially
responsible for the toxin that was found in ap-
preciable numbers in our study, was sampled
using a method similar to Yasumoto et al.
(1980a). This is a simple washing procedure to
separate dinoflagellate from host macroalga. It
has been used successfully in field studies, with
minor modifications, by other workers (Tebano
& McCarthy,1984; Tebano,1984). Weighed
samples of 100g of each algal species were placed
in a plastic container to which filtered seawater
was added. The contents of the container were
then shaken vigorously for 2 min (c.250 shakes)
before being sieved at Imm and 38m. The
residue on the 38.m sieve was then washed into
a 50m vial to which 5ml of concentrated formalin

526

A Nui

G. toxicus per 100g host

Caulerpa Halimeda Hypnea
B Nanumea

~
yw
°
£
D
f—)
o
Ta
a.
o
o
2)
=
2
bad
ie)
2
me

a eg ie eS

=

a

o

| Ww

Jania | Boodlea Blue-green

FIG.3. Mean abundance of G. toxicus at Channels and
Controls at (A) Nui and (B) Nanumea (Time 3 =
October 1989). Values are mean cells counts + stand-
ard error, Ch=Nui Channel; NC=North Control;
SC=South Control; Am=American Channel; K=Ken-
nedy Channel.

had been added. Samples were brought back to
the laboratory, quickly shaken to resuspend par-
ticles, and allowed to settle for several days. After
settlement, each vial was found to contain a basal
sediment layer, a layer of partially suspended
mainly organic matter and an uppermost layer of
transparent formalin and seawater solution.
Neither the formalin solution nor the sedimentary
layer were found to contain G. toxicus that would
have been alive at the time of collection (there
were some skeletons in the sedimentary layer).
To estimate the abundance of G. toxicus in a
sample, the thickness of the primarily organic
layer was measured in the pre-settled but undis-

MEMOIRS OF THE QUEENSLAND MUSEUM

turbed sample jar using vernier callipers (correct
to 0.1mm) for later determination of volume. Five
replicate 0.1ml subsamples of the organic layer
were drawn from each jar using a micro-pipette
and mounted individually on a microscope slide.
All cells of G. toxicus on each slide were counted
with the 5 replicates and volume measurement
being used to obtain an estimate of the total
number of cells in the organic layer of the sample
(and hence the whole jar). This value gave abun-
dances of cells per 100g of host alga.

Three replicate samples of 100g of 2—3 species
of intertidal host algae were collected from each
sampling site and time. The identity of the host
species varied among islands and sites according
to what was available. On Nui, Caulerpa,
Halimeda and Hypnea were sampled; on
Nanumea, Jania, Boodlea and an unidentified
blue-green alga were sampled. On Niutao, Jania,
Caulerpa and an unidentified green turfing alga
were collected. To convert estimates of cell num-
bers per 100g of individual host algae to relative
densities per 100g of all hosts, measurements of
the percentage cover of the hosts were obtained
for all sites at Nui and Nanumea. Five replicate
readings of cover were made using a rope which
was marked with 20 random points, recording the
algal species found under each point.

MEASUREMENTS OF CTENOCHAETUS STRIGOSUS
TOXICITY

Ten to fifteen Crenochaetus strigosus were
speared at the shallow subtidal areas at the site
and time of sampling on all islands, except Nui.
Fifty grams of muscle was dissected from each
fish and stored in 100% ethanol. Ciguatoxins in
these fish samples were confirmed by mouse
bioassay (undertaken by E. Shang, Sept,1990).
All tissue samples were analysed in the lab-
oratory by the ‘Stick-EIA’ immunoassay method
(Hokama et al.,1983,1987).

ANALYSES

Cell abundance and fish toxicity data were
analysed using analyses of variance (ANOVA)
with planned comparisons. The main factors ex-
amined were: Locations (which were subdivided
to examine differences among channels, among
controls, between channels and controls, between
the wreck and controls, and distances away from
channels); Island; Times and Host algae (all
fixed). Assumptions of homogeneity of variances
were tested using Cochran’ t_ (Winer, 1971)
and data were transformed V(x+1) as required.

EFFECTS OF DISTURBANCES ON CIGUATERA IN TUVALU

A Nui
2 18
3 16 :
= 14
m 12
o
= 10
o 8
3 6
® 4
~ 2
So —7
NC sc
B Nanumea
5]
3
<=
a
L°2]
co
2
a
Qa.
wo
8
*
2
S
K NC sc

Fig.4. Composite abundances of G. texicus at (A) Nui
and (B) Nanumea. Values are calculated as cells per
100g of all hosts (i.e. the relative abundance of each
macroalga is taken into account). Ch=Nui Channel,
NC=North Control, SC=South Control,
Am=American Channel; K=Kennedy Channel.

RESULTS

PRE-EXISTING BOAT CHANNELS AND WRECKS
Abundance of G. toxicus. There was no dif-
ference between the Nui channel and controls in
the abundance of G. toxicus on Caulerpa,
Halimeda and Hypnea (Fig.3A). The numbers of
cells recorded was low, with a maximum of 500
cells per 100g of algae on one host at one site
only. However, at Nanumea, both channels ex-
amined had greater densities of cells than the

327

10000

1000

100

10

G. toxicus per 100g Jania

ow Nc Muli WC NP

3.00

f
=]
Ss

Stick-test value
a
=]

0.00

NW

ow NC Muli Wwe NP

FIG.5. Abundance of G. toxicus and fish toxicity al
Niutao (Time 2): A comparison of existing wrecks
with controls (Note that ‘Muli’ is regarded as a control
for this comparison: at Time 2 channel construction
had not yet begun). (A) G. toxicus density; (B) Stick
test values for fish toxicity. NW=New Wreck,
OW=Old Wreck, NC=North Control, Muli=Future
aa of Muli Channel, WC=West Control, NP=North

oint.

control areas on Boodlea and Jania (Fig.3B). For
the American passage, densities of cells reached
7000 per 100g of algae. The only alga that could
be found at the wreck site was a blue green alga
not represented at the channels or controls. Only
a few G. toxicus were associated with this species
at the wreck site.

When cell densities per individual host were
adjusted to account for the different percent cover
of the algal species, and host algae pooled to
provide relative densities of cells per 100g of ‘all
hosts’, a similar pattern was evident. The channel
on Nui was intermediate between the two con-

528

Stick-test value

FIG.6. Fish toxicity readings from Nanumea (Septem-
ber 1989=Time 3). Values are means of 10 fish
+ standard errors.

trols (Fig.4A). On Nanumea, densities were con-
siderably higher at channel sites compared to
controls (Fig.4B). On average, densities were
higher on Nanumea than Nui.

No significant difference could be detected be-
tween the 2 wreck sites on Niutao, and 4 controls
prior to construction of boat channels (Fig.5A).
Cell numbers reached 5000 per 100g of Jania at
North Point (NP), a concentration similar to the
American Passage on Nanumea.

Fish toxicity. There was no effect of boat chan-
nels or the wreck on toxicity of C. strigosus at
Nanumea (Fig.6). Marginally higher toxicity was
recorded at the north control site (NC).

Fish collected at wreck sites were no more toxic
than control areas on Niutao (Fig.5B). If any-

TABLE 1. Analyses of variance of G. toxicus cell abundances at Niutao:
planned comparisons of channels, controls and distance from channels
during Time 3. NW=new wreck; D=distance; Ch=channels; C=con-
trols; NS=non-significant F test; *=p<0.05. Data transformed sqrt(x+1).

MEMOIRS OF THE QUEENSLAND MUSEUM

1200

900

600

# cells/100g Jania

300

eo0
soa a
z

Kulia Channel

$100

Muli Channel Controls

FIG.7. Mean abundance of G. toxicus at Channels and
Controls at Niutao (Time 3 = October 1989). Values
are mean cells counts + standard errors, At Muli and
Kulia, samples were taken at four distances on either
side of the channels. At Muli, these were to the East
and West of the channel and at Kulia to the North and
South. Hence E100 =100m to the East of the Muli
Channel etc.

thing, the means were lower at these sites. The
most striking feature of this survey was that the
highest level of fish toxicity was recorded at
North Point (NP), which corresponded to the
highest concentrations of G. toxicus in the host
alga Jania. This was the only site that we consis-
tently found fish giving a positive result to the
Hokama immunoassay.

EFFECTS OF NEW BOAT CHANNELS ON NIUTAO
No consistent effect of new
boat channels on the abundance
of G. toxicus was detected (Fig.7,
Table 1). While some of the
highest densities were recorded at
the Om and 20m sites on the down
current side of the channels, most
of the channel sites fell within the

fl,f2_| SIG | range recorded at the controls.
Allcells _| 62 | 11270.0 | 181.77| Res 62,240| * | Patterns of change away from
Allocations | 20 | 7350.68 | 367.53| S(L) | 3.94 | 20,42 | * | channels indicate that there are
Channels 1 191.75 | 191.75 | S(L) 1.42_| NS_| majorchanges in abundance over
Distance 7 | 4000.21 | 571.46| S(L) | 6.12 | 7,42 | * | a 200m distance, but no obvious
ChxD 7_| 1990.42 | 284.35| SL) | 3.05 | 7,42 | * | gradients which could be at-
Controls 3_ | 872.97 [290.99] sa) | 3.12 | 3,42 | * | tributed to channels. Patterns of
NWvsc | 1 | 163.43 | 163.43] say | 1.75 | 1,42 | Ns | abundance were so localised that
Ch vs NW vs C 131.94 | 6597 | sa) | 071 | 2,42 | ns | G-foxicus abundance on opposite
Sample(L) | 42 93.38 |_Res | 5.28 | 42,240| + | Sides of the channels (the Om dis-
Gesaea oat 7.68 tances, only 10m apart) differed
asa dae laasieah at both channels sites.
: No significant difference be-

EFFECTS OF DISTURBANCES ON CIGUATERA IN TUVALU

tween channel and control sites could be found in
the toxicity of C. strigosus (Fig.8, Table 2).
During channel construction (May 1989), 3
months after (September 1989) and 15 months
after channel construction, toxicity levels at chan-
nels were well within the range exhibited at the
controls. In terms of Hokama’s thresholds, most
fish were within the negative to borderline
categories in terms of edibility.

TEMPORAL PATTERNS IN THE OUTBREAK ON
NIUTAO

Major changes in the abundance of G. toxicus
were recorded over the period of this study
(Fig.9A, Table 3), being highest at December
1988, and dropping to near O at the September
1990. In contrast, fish toxicity levels were lowest
at the first sampling date and rose to a peak in
September 1989 (Fig.9B, Table 4). This coin-
cided with a severe outbreak of ciguatera on this
island (Kaly et al. 1991). Toxicity began to
decline again in 1991. Although we were not able
to continue sampling beyond this date, we have
received reports that the outbreak of ciguatera
had started to abate by the end of 1992 and fish
were being consumed by late 1993. Although
data are minimal, there is a suggestion of a lag
phase of approximately one year in peak G.
toxicus numbers and peak toxicity, at least in an
indicator herbivorous fish.

DISCUSSION

This study provided no clear support for the
view that physical disturbances, such as
shipwrecks and boat passages, promote out-
breaks of ciguatera fish poisoning. The boat chan-
nel on Nui, and the shipwrecks on Nanumea and
Niutao all exhibited Gambierdiscus toxicus den-
sities and Ctenochaetus strigosus toxicity in the

TABLE 2. Analyses of variance of fish toxicity: com-
parison of channels and controls during T3 at Niutao
and Nanumea (Cochran’s Q=0.1645, v=14, k=11,
NS); NS=non-significant F test; *=p<0.05.

Location
(Ix Ch vs C)

December 1988 3
2 Positive
Borderline
1 Negative
0
May 1989
Ww
n
+
+
o
=
o
>
_
September 1989 3
bo
x
2
‘=
no

September 1990

= so
23
=¢

FIG.8. Toxin levels in Ctenochaetus strigosus col-
lected at Niutao at 4 times between December 1988
and September 1990. Values are means of stick-test
results obtained from 10 fish tested at each site and
time + standard errors. Muli and Kulia are channels,
NW and OW are wrecks, remaining sites are controls.
Missing bars are unsampled sites, not zero values.

range observed at sites several km from these
disturbances. However, the results for Nanumea
boat channels and trends at Niutao suggest that
disturbance may play a role at some times. G.
toxicus numbers were elevated in the vicinity of
boat channels, particularly the American passage,
despite the fact that fish toxicity was not. It is
noteworthy that the channel area, although not
toxic at the time of this study, has been toxic at
irregular intervals over the last 20-30 years and
at those times was avoided by local fisherman.
The channel bisects the reef crest into what was
once a ponding lagoon and is a region of extreme
currents, and continual disturbance due to wave
action. An area considered extremely toxic at the
time of this study (on the eastern side of the
island) was well-removed from the boat channels.

The before and after study of two new channels

530

A G. toxicus

G. toxicus per 100g Jania

Dec-88 May-89 Sep-89 Sep-90

B Fish toxicity

Stick-test value

Dec-88 May-89 Sep-89

Sep-90

FIG.9. (A) abundance of cells and (B) levels of fish
toxicity through time at ‘New Wreck’, Niutao.

on Niutao did not indicate a successional change
in macro-algal hosts and an early successional
outbreak in G. foxicus, although numbers ap-
peared to be slightly elevated on the down-cur-
rent side of each channel. However, cell numbers
in the vicinity of channels were not unusually
high relative to control areas at any stage during

TABLE 3. Analyses of variance of G. toxicus cell abundances at
Niutao: comparison of 2 hosts (Jania sp. and unidentified green)
through time at New Wreck (NW). NS=Non-significant F test;
*=p<0.05. Data transformed sqrt(x+1).

MEMOIRS OF THE QUEENSLAND MUSEUM

the study. Channel construction on this island was
preceded by a bloom of G. toxicus which must
have been caused by some other phenomenon
(Kaly et al. 1991). This may have made any effect
of channels difficult to detect.

Physical disturbance to reefs may or may not
lead to ciguatera. ‘Disturbance’ encompasses a
host of phenomena which may affect the habitat
and population dynamics of G. toxicus in dif-
ferent ways. Whether or not ciguatera is induced
by disturbance may depend on the intensity,
timing, frequency and scale of the disturbance, all
factors known to affect the trajectory of succes-
sional patterns on hard substrata (Connell &
Keough, 1985). The time-scale for recovery may
depend on the regime of disturbance and chance
factors in the recolonisation of damaged areas.
Disturbance itself may interact with other
phenomena to explain the observed patterns of
outbreaks. Factors such as wave exposure, fresh
water run-off and coral destruction are likely to
be closely linked, making such interactions likely
and single-cause scenarios extremely unlikely.

Different intensities of disturbance may have
opposite effects on ciguatera. If coral habitat is
damaged and the growth of host algal species
promoted, then densities of G. toxicus could be
elevated. Seasonally high wave exposure and
wind at North Point (NP) on Niutao appeared to
correlate with high numbers of G. toxicus, which
also coincided with a higher average fish toxicity.
In other studies, moderate wave exposure has
acted to reduce cell densities, presumably by
dislodging cells without disrupting the habitat
(Anderson & Lobel 1987).

A promising area for research lies in the poten-
tial relationship between G. toxicus numbers and
fish toxicity. While there was some evidence that
the distribution of G. toxicus around Niutao cor-
related with fish toxicity, toxic fish
were caught all round the island
during the peak of the outbreak (Kaly
et al. 1991). Temporal monitoring
during this outbreak was limited but
suggested a lag phase of approximate-
ly one year between peak G. toxicus

numbers and the highest toxicity
levels. More comprehensive sam-

pling is needed to confirm this pat-
tern. However, if such a pattern
proves to be of general significance
there are obvious implications with

Times 1484.57
Host 1 | 968.58 | 968.58
a
Ti xH_| 2 | 500.31
Sample | 12| 228.91
| ix) | | ee eee

Residual | 69 | 380.35 5.51

Total 86 | 3718.46

regard to the forecasting and manag-
ing of ciguatera outbreaks.

EFFECTS OF DISTURBANCES ON CIGUATERA IN TUVALU

531

TABLE 4. Analysis of variance of fish toxicity: comparison through time of Kulia, New Wreck and West control
at Niutao (Cochran’s Q=0.1399, v=9, k=12, NS). NS=non-significant F test; *=p<0.05.

FACTOR DF SS MS DENOM F f1,f2 SIG

Time 3 14.66 4.89 Res 8.99 3,103 s

Location 2 0.94 0.47 Res 0.86 2,103 NS

TxL 6 4.70 0.78 Res 1.44 6,103 NS

Residual 103. 55.95 0.54

Total 114 976.25

Tukey's Test: T3 (Sep 89) T4(Sep 90) T2 (May 89) T1 (Dec 88)
ACKNOWLEDGEMENTS ciguatera endemic region in the Caribbean. PhD

This study was made possible by funding from
the New Zealand Ministry of Foreign Affairs and
Trade, Wellington. We thank the Tuvalu Depart-
ments of Fisheries, and Transport and Com-
munications, the Niutao Island Council and the
people of Niutao, Nui and Nanumea for their
co-operation and support. Special thanks to T.
Logomalie for his assistance at Niutao. K. Trick-
lebank, B. Jackson, K. Clements, R. Cole, V.
Staines and R. Prasad assisted with the collection
of samples from the field and/or with laboratory
analysis. Special thanks to Y. Hokama and E.
Shang who confirmed ciguatoxins in our fish,
taught us how to use the stick test and allowed us
to use their laboratory to test our fish tissues.

LITERATURE CITED

ANDERSON, D.M. & LOBEL, P.S. 1987. The continu-
ing enigma of ciguatera. Biological Bulletin 172:
89-107.

BAGNIS, R. 1969. Naissance et development d’une
flambee de ciguatera un atoll des Tuamotu. Revue
Corps Sante Armees 10: 783-787.

BAGNIS, R., BENNETT, S. PRIEUR, C. & A.M.
LEGRAND. 1985. The dynamics of three benthic
dinoflagellates and the toxicity of ciguateric sur-
geonfish in French Polynesia. Pp. 177-182. In
D.M. Anderson, A.W. White & D.G. Baden (eds),
Toxic dinoflagellates. (Elsevier: New York).

BANNER, A.H. 1974. The biological origin and trans-
mission of ciguatoxin. Pp.15—36. In Humm, H.J.
& Lane, C.E. (eds), ‘Bioactive compounds from
the sea’. (Marcel Decker: New York).

BANNER, A.H. 1976. Ciguatera: A disease from coral
reef fish, Pp.177-213. In Jones, O.A. & Endean,
R. (eds), ‘Biology and geology of coral reefs 3’.
(Academic Press: London).

CARLSON, R.D. 1984. Distribution, periodicity and
culture of benthic/epiphytic dinoflagellates in a

Thesis, Southern Illinois University. (Unpubl.)

CARLSON, R.D. & TINDALL, D.R. 1985. Distribu-
tion and periodicity of toxic dinoflagellates in the
Virgin Islands. Pp.171—-176. In D.M. Anderson,
A.W. White & D.G. Baden (eds), ‘Toxic
dinoflagellates’. (Elsevier: New York).

CONNELL, J.H. & KEOUGH, M.J. 1985. Disturbance
and patch dynamics of subtidal marine animals on
hard substrata. Pp 125-155. In Pickett, S.T.A. &
White, P.S. (eds), ‘The ecology of natural distur-
bance and patch dynamics’. (Academic Press:
New York).

COOPER, M.J. 1964. Ciguatera and other marine
poisoning in the Gilbert Islands. Pacific Science
18: 411-440.

DALZELL, P. 1992, Ciguatera fish poisoning and
fisheries development in the South Pacific. Bul-
letin de la Societé de Pathologie Exotique 85:
435-444.

GILLESPIE, N., HOLMES, M., BURKE, J. &
DOLEY, J. 1985Sa. Distribution and periodicity of
Gambierdiscus toxicus in Queensland, Australia.
Pp.183-188. In D.M. Anderson, A.W. White &
D.G. Baden (eds), “Toxic dinoflagellates’. (El-
sevier: New York).

GILLESPIE, N., LEWIS, R., BURKE, J. & HOLMES,
M. 1985b. The significance of the absence of
ciguatoxin in a wild population of Gambierdiscus
toxicus. Proceedings of the 5th International Coral
Reef Symposium 4: 437-441.

HOKAMA, Y., ABAD, M.A. & KIMURA, L.H. 1983.
A rapid enzyme immunoassay for the detection of
ciguatoxin in contaminated fish tissues. Toxicon
21: 817-824,

HOKAMA, Y, SHIRAI, M. KUROSAWA, N.,
IWAMOTO,M., GOTO, C.S., & OSUGI, A.M.
1987. Assessment of a rapid enzyme immunoas-
say stick test for the detection of ciguatoxin and
related polyether toxins in fish tissues. Biological
Bulletin 172: 144-153.

HOLMES, M.J. & LEWIS, R.J. 1991. Multiple gam-
biertoxins (ciguatoxin precursors) from an
Australian strain of Gambierdiscus toxicus in cul-

532

ture. 10th World Congress on Animal Plant and
Microbial toxins 1-10.

HOLMES, M.J., LEWIS, R.J., POLI, M.A. & GIL-
LESPIE, N.C. 1991. Strain dependent production
of ciguatoxin precursors (Gambiertoxins) by
Gambierdiscus toxicus (Dinophyceae) in culture.
Toxicon 29: 761-775.

KALY, U.L., JONES, G.P. & TRICKLEBANK, K.
1991. Preliminary assessment ofa severe outbreak
of ciguatera at Niutao, Tuvalu. South Pacific Jour-
nal of Natural Science 11: 63-81.

KOHLER, S,T. & KOHLER, C.C. 1992. Dead bleached
coral provides new surfaces for dinoflagellates
implicated in ciguatera fish poisonings. Environ-
mental Biology of Fishes 35: 413-416,

KUBERSKI, T. 1979. The chain of events in ciguatera
fish poisoning. SPC Fisheries Newsletter 19: 18—
19,

LEWIS, R.J. in press. Socioeconomic impacts and
management of ciguatera in the Pacific. Bulletin
de Ja Societé de Pathologie Exotique 85: 427-
434..

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae). Toxicon 29: 1115-1127.

MURATA, M., LEGRAND, A-M., ISHIBASHI, Y. &
YASUMOTO, T. 1990. Structures and configura-
tions of ciguatoxin from moray eel Gymnotherax
javanicus and its likely precursor from the
dinoflagellate Gambierdiscus toxicus. Journal of
the American ChemicalSociety 112: 4380-4386.

RANDALL, J.E. 1958. A review of ciguatera tropical
fish poisoning with a tentative explanation of its
cause. Bulletin of Marine Science of the Gulf and
Caribbean 8: 236-267.

TAYLOR, F.J.R. 1985. The distribution of the
dinoflagellate Gambierdiscus toxicus in the East-

MEMOIRS OF THE QUEENSLAND MUSEUM

ern Caribbean. Proceedings of the 5th Internation-
al Coral Reef Symposium 4: 423-428.

TAYLOR, F.J.R. & GUSTAVSON, MLS. 1985. An
underwater survey of the organisms chiefly
responsible for “ciguatera’”’ fish poisoning in the
eastern Caribbean region: The benthic dinoflagel-
late Gambierdiscus toxicus. In: Stefanon, A. &
Flemming, N.J. (eds), ‘Proceedings of the VIlth
International Diving Science Symposium’
(Padova, Italy),

TEBANO, T. 1984. ‘Population density study ona toxic
dinoflagellate responsible for ciguatera fish
poisoning on South Tarawa Atoll Republic of
Kiribati’, Report from the Atoll Research and
Development Unit, Tanaea, Tarawa, Republic of
Kiribati, 48p.

TEBANO, T. 1992, Ciguatera fish poisoning and reef
disturbance in South Tarawa, Kiribati, SPC
Ciguatera Information Bulletin 2: 7.

TEBANO, T. & McCARTHY, D. 1984. Ciguatera fish
poisoning and the causative organism in the Gil-
bert Islands, Kiribati. Report; of the Atoll Re-
search and Development Unit, University of the
South Pacific. 109p.

WINER, B.J. 1971. ‘Statistical principals in experimen-
tal design,’ 2nd ed. (McGraw Hill; New York).

YASUMOTO, T., FUJIMOTO, K., OSHIMA, Y.,
INOUE, A., OCHI, T., ADACHI, R. &
FUKUYO, Y. 1980a. ‘Ecological and distribu-
tional studies on a loxic dinoflagellate responsible
for ciguatera’. Report to Ministry of Education,
Japan, 49p.

YASUMOTO, T., INQUE, A., OCHI, T., FUSIMOTO,
K., OSHIMA, Y., FUKUYO, Y., ADACHL, R, &
BAGNIS, R. 1980b. Environmental studies on a
toxic dinoflagellate responsible for ciguatera. Bul-
Ietin of the Japanese Society for Science and
Fishenes 46; 1397-1404.

MAITOTOXIN INDUCES MUSCLE CONTRACTION AND A NON-SELECTIVE
CATIONIC CURRENT IN SINGLE SMOOTH MUSCLE CELLS
OF THE GUINEA-PIG PROXIMAL COLON

RICHARD J. LANG, FIVOS VOGALIS, MICHAEL J, HOLMES AND RICHARD J. LEWIS

Lang, R.J., Vogalis, F., Holmes, M.J. & Lewis, R.J. 1994 08 01: Maitotoxin induces muscle
contraction and a non-selective cationic current in single smooth muscle cells of the
guinea-pig proximal colon. Memoirs of the Queensland Museum 34(3), 533-540, Brishane,
ISSN 0079-8835.

We have investigated the mechanisms of action of maitotoxin-2 (MTX), a marine toxin
isolated from the toxic dinoflagellate, Gambierdiscus toxicus, on the contractility of the intact
circular smooth muscle of guinea-pig proximal colon and on the membrane currents recorded
in enzymatically-dispersed single cells trom this muscle, using standard contraction and
patch clamp recording techniques. MTX (0,005-5.0nM) induced an initial phasic contraction
and a subsequent cessation of all spontancovs contractile activity. which was sometimes
associated with a sustained increase in muscle tone. The initial contraction to MTX was
blocked by atropine (2.M), the muscarinic receptor antagonist. Contractions to acety!choline
(0.5j.M) were reduced approximately 75% after 7 minutes exposure to MTX (0.5nM) (n=5),
this blockade was resistant to washout. MTX (5nM) completely abolished the contractions
to acetylcholine, but reduced contractions to raised external concentrations of K+(40mM)
only 54 +9% (n=4).

Single colonic smooth muscle cells were perfused with K+- filled patch pipettes and voltage
clamped ata holding potential of -80 mV. MTX (SnM) added to the bathing soluuon indoced
a Jarge inward current (1-3 nA) after 15-45 minutes. Development of this current was not
prevented by a number of K+ channel blockers, including tetraethylammonium (TEA; 2-126
mM), 4-aminopyridine (SmM), quinidine (SmM) or ghbenclamide (104M); nor when the
Ca2+ was removed, or replaced with Ba?+ (7.5mM). This current was, however, blocked by
Cd2+ (0.1-1mM) and reduced by nifedipine (10,.M) and La3+ (1mM). The MTX-activated
current had an almost linear current-voltage relationship with a reversal potential near —30
and OmV when cells were respectively filled with K+ or Cs+, When most of the extracellular
Nat (126mM) was replaced with TEA+, this current reversed near -60mV. These results
suggest that MTX induces the appearance/opening of voltage-insensitive channels which
allow the flow of Nat, K+ and Cs*, but not TEA+, and which are blocked by Cd2*,

Richard J. Lang, Fivos Vogalis, Department of Physiolagy, Monash Universiry, Clayton,
Victoria 3168; Michael J, Holmes, Richard J. Lewis, Southern Fisheries Centre, Queensland
Department of Primary Indusiries, PO Box 76, Deception Bay, Queensland 4508; 22
November, 1993,

Ciguatoxins(CTX) and maitotoxins(MTX) are
potent marine toxins isolated from the dinoflagel-
late Gambierdiscus toxicus. Ciguatoxins are lipid
soluble and accumulate in the flesh and viscera
of reef fish; they are the principal toxins respon-
sible for cigatera. Maitotoxins are more polar and
extracted in a number of forms from cultures of
G. toxicus (Yokayama et al.,1988; Holmes et al.,
1990). They have positive inotropic effects on
cardiac muscle (Kobayashi etal.,1985) and cause
contraction in smooth muscle (Ohizumi &
Yasumoto,1983). They also induce a rise in the
internal Ca** levels in BC3H; muscle cells
(Sladeczek et al.,1988) and aortic smooth muscle
cells in culture (Berta et al..1988) associated with
phosphoinositide metabolism (Gusovsky ct al.,
1987,1988; Sladeczek ef al..1988: Meucci et al.,

1992), and promote release of transmitters from
neurones. and hormones from 4a sumber of
secretory cells (Kim etal.,1985; Gusoysky et al.,
1988). Almost all of these effects of MTX depend
on the presence of extracellular Ca** and can be,
depending on the tissue, blocked by both organic
(verapamil, some dihydroprydines) and inor-
ganic Ca*+-channel entry blockers (Cd**, Ni*+
and Co?*). This rise in intracellular Ca** induced
by MTX has therefore been suggested to arse
from (i) modulation of voltage-activated Ca’*
channels (Kobayashi et al.,.1987, Yokayama ct
al., 1988), (ii) the mobilization of Ca** from inter-
nal stores (Meucci et al.,1992), or (712) from the
influx of Ca?* through MTX-activated pores or
channels (Yoshii et al..1987; Sladeczek et al.,
1988), Here, we descnbe that MTX-2 inhibited

534

MEMOIRS OF THE QUEENSLAND MUSEUM

(5x10 °M)

MTX

Ach W

FIG. 1. Effects of acetylcholine and MTX-2 on the contractile activity of circular muscle strips of the guinea-pig
proximal colon. MTX (5nM) induced a transient contraction sensitive to atropine (2}.M). Contractions to
acetylcholine (Ach, 54M) were abolished irreversibly, even after extensive washout (W) of MTX.

the spontaneous and acetylcholine-induced con-
tractile activity in the intact guinea-pig proximal
colon. These effects are likely to be related to the
large inward current observed in voltage clamped
single cells exposed to MTX, arising from the
induction of MTX channels selective for K* and
Nat. Some of these results have been presented
previously in brief (Lang et al.,1992).

METHODS

MTX-2 used in this study was isolated from
cultures of the NQ1 strain of G. toxicus and
purified to homogeneity on HPLC as previously
described (Holmes et al.,1990). This MTX has a
molecular weight of 3290 Daltons for the sodium
salt, a LDso of 0.08,1g/kg in mice (applied intra-
peritoneally) and was dissolved in a small volume
of methanol: water (1:1).

CONTRACTION STUDIES

The proximal colon (6cm long, 1—2cm aboral
of the caecum) of the guinea-pig was excised and
cut into 0.5cm rings Preparations were placed in
5m] organ baths and suspended under 0.5g ten-
sion between two stainless steel rods to allow
recording of circular muscle contraction.

Isometric recordings (F-60 Narco Biosystems)
were made at 37°C. Tissues were maintained in
oxygenated (95% Oz: 5% CO2) physiological
saline solution containing (mM): NaCl 137; KCI
2.7; CaClz 1.8; MgCle 1.0; KH2PO4 0.5, NaHCO3
11.9; glucose 5.5. Preparations were equilibrated
for 30 min and then used after obtaining
reproducible responses to acetylcholine (54M).

CELL DISSOCIATION

The proximal colon (3—4cm long), 1—2cm
aboral of the caecum, of the guinea pig was
excised, cut open longitudinally and pinned out,
mucosal surface uppermost, in a dissecting dish
filled with a nominally Ca**-free physiological
saline (PS)(see below). After removal of the
mucosa, the circular muscle layer was peeled
from the underlying longitudinal layer, cut into
small pieces (2mm”) and rinsed in low-Ca?* (30
1{M) PS for 2 min (at 37°C). The muscle pieces
were then transferred to low-Ca”* PS containing:
collagenase Type 1(0.6mg/ml; Worthington);
bovine serum albumin (2mg/ml; Sigma) and tryp-
sin inhibitor (0.2mg/ml; Sigma). After a 60
minute incubation period, the muscle pieces were
re-suspended in low-Ca?* PS and gently agitated
for 10 min (at 37°C). Single cells were obtained
by gentle trituration with a wide-bore glass
pipette. Cells were allowed to settle for 5-10 min
to the glass bottom of the recording chamber
mounted on an inverted microscope; the solution
was then exchanged for normal Ca?* (1.5mM) PS
(Vogalis et al.,1993).

WHOLE-CELL AND SINGLE CHANNEL CURRENT
RECORDINGS

Patch pipettes were drawn from glass capillary
tubing (1.5—1.8mm; Kimax-51, Kimble, USA) on
a programmable micro-pipette puller (Sachs-
Flaming PC-84, Sutter Instruments) and their tips
fire polished (MF-84 Narishige). Pipettes resis-
tances ranged from 2-—7MQ. when filled with
pipette solution. Single channel and whole-cell
membrane currents were recorded at room
temperature using an Axopatch 200 (Axon In-

MATTOTOXIN INDUCES MUSCLE CONTRACTION

struments} and conventional patch-clamp techni-
ques (Hamill et al.,1980). Current and voltage
signals from the patch-clamp amplifier were
digitized with a Labmaster TM125 analog-to-
digital device (Scientific Sobutions) interfaced to
an Arrow-AT desktop computer using p-CLAMP
software (Axon Instruments). Digitized data
were stored and analyzed using this p-CLAMP
software.

The physiological saline (PS) was of the fol-
lowing composition (mM): NaC] 126; KCl 6:
HEPES 6; d-glucose 11; MgClz 1.2; CaCl: 1.5;
adjusted to pH 7.4 with SM NaOH. The pipette
solution contained (mM): KC] 126: HEPES 6;
NazATP 3; EGTA 3; MgCh 3; d-glucose 11; pH
Was adjusted to 7.4 with SM KOH. In some cells.
current flow through all K* channels was blocked
by replacing the KCI in the pipette solution with
an equimolar concentration of CsCl], pH was set
with SM NaOH (Vogailis et al.,19953).

RESULTS

CONTRACTION STUDIES

A typical isometric recording of the spon-
{aneous contractions in a stip of circular muscle
from the guinea pig proximal colon is illustrated
in Fig.l. Acetylcholine (Ach, 5j.m) caused a
transient increase in muscle tension. MTX (SnM)
produced a similar transient increase in tension
which was sometimes followed by a maintained
increase in muscle lone associated with a Joss of
the spontaneous contractile activity. This in-
¢rease in muscle tone decreased upon washout
(W) of the MTX, the spontaneous contractile
activity. however, did not return. These effects of
MTX were concentration dependent. Threshold
phasic contractions (0.07-0.22g) to MTX were
evoked with concentrations between 0.005 and
0.5nM MTX (n=3-5), substantial contractions
{>2g) were only recorded with 5 nM MTX (n=7).
In four muscle strips, these effects of MTX (SnM)
were blocked by atropine (2).M), the muscarinic
receptor antagonist.

The contractions te Ach (514M) were complete-
ly abolished by MTX (5nM) and never recovered,
even 145 minutes after the removal of MTX
(n=6). In contrast, contractions elicited by direct
muscle depolarization with 40mM K saline were
decreased only 54.6+8.7% (n=4) hy MTX
(5nM). These effects of MTX were also con-
centration dependent, the concentration of MTX
which half-maximally inhibited the Ach contrac-
tions was approximately 0.2nM MTX, These ef-
fects of MTX were mimicked in part by

monensin, the Na™ ionophore, which blocked the
acetylcholine contractions dose-dependently
(0.1-10p2M), half-maximal reduction was
achieved with approximately 0.2uM monensin
(n=4). Monensin, however. did not induce any
muscle contraction. These data suggest that
MTX, monensin and acetylcholine may well he
shanng a common mechanism of action. In view
of this, the expenments below describe our
preliminary investigations of the action of MTX
on the membrane channel] currents recorded in
single smooth muscle cells of the guinea-pig
proximal colon.

CONTROL WHOLE-CELL CURRENT RECORDINGS

Resting membrane potentials of 40 to -60 mV
were recorded when a 6mM: 136mM K* gradient
was established across the cell membrane of
single cells of the proximal colon. Depolarizing
currents tngvered action potentials which peaked
between -10 and Om¥V and had durations of 100-
200ms at their half-maximal amplitude. Unter
voltage clamp, depolarizations, from a holding
potential of -80m¥V, triggered comples
membrane current responses (Fig. 2A). At poten-
uals positive to -HOm'V. a rapidly-activating and
inactivating Outward current was triggered. At
more positive potentials (-20mV) the transient
current, in most cells, was followed by a second
slowly developing and decaying outward current.
The pharmacological identification of three K*-
channel currents and one Ca?*-channel current
underlying these current responses to membrane
depolarization has been demonstrated (Vogalis et
al.,1993) and will be briefly summarized.

A substantial portion of the second slowly-
decaying outward current recruited at positive
potentials is blocked by the addition of
tetracthylammonium (TEA) (2-5mM) and a
Ca’*- entry blocker (0.1 mM Cd**) to the bathing
solution, suggesting that this current flows
through the large conductance Ca?*-activated
(maxr or “BK*) K* channels (/xcu) which have
been recorded in all smooth muscles so far ex-
amined (Vogalis et al..1993). The Ca**-insensi-
tive current remaining inthe presence of low TEA
(2-SmM): Cd?* (0.1mM) activates rapidly and
then decays slowly to a sustained current after
400ms, and can be further divided into two K*
channel current components. A slowly-activat-
ing, non-inactivating K* current (eer), which has
characteristics similar to delayed rectifier K* cur-
rents found in many electrophysiological
preparations (Rudy, 1988), is revealed when 4-
aminopyridine (4-AP) (5 mM) blocked the initia!

536

MEMOIRS OF THE QUEENSLAND MUSEUM

Memb. Current (pA)

2000pA
30ms

—80-60-40-20 0 20 40
Memb. Potential (mV)

FIG. 2. Effects of MTX (5nM) on the whole-cell membrane currents recorded in single cells of the guinea-pig
proximal colon. Cells were voltage clamped at a holding potential of -80mV with a K*-containing pipette.
Stepped changes in potential (100-400ms in duration) were applied to cells to evoke voltage-gated membrane
channel currents in control solutions (A) and 34 minutes after exposure to MTX (5nM)(B); short lines on either
side of these panels represent the zero current level. Note the ten-fold change in the vertical scale in B. C, current
voltage plots of the initial peak amplitude (filled circles) in control saline and the end of pulse current before
(hollow circles) and after MTX exposure (filled triangles).

itial transient component of the Ca?*-insensitive
K* current. On the other hand, blockade of the
sustained component of the Ca**-insensitive cur-
rent with TEA (12—20mM) reveals the time
course of the rapidly activating and inactivating
transient outward current, (/kio) which was
blocked by 4-AP (5mM) (Rudy,1988). In cells
recorded with pipettes containing Cs-saline,
stepped depolarizations elicits an inward current
at potentials positive to 40mV which peaks near
+10mV, reverses in direction near +5OmV and
decays slowly over 400ms. This inward current
was increased when Ba?* replaced Ca?* in the
bathing solution and blocked by the Ca?*-entry
blockers, nifedipine (10p.M) or Cd?* (0.1mM),
indicating that it represents current flow through
‘high-voltage activated’ or ‘L-type’ Ca*+ chan-
nels (Ica) (Vogalis et al.,1993).

ACTION OF EXTERNALLY-APPLIED MTX

MTX (5nM) induced a massive increase in the
holding current (at a holding potential of —80
mV), from about 50-100pA to about 1000-3000
pA, but only after a delay of some 15-45 minutes
(n=14 cells). When the whole-cell membrane
currents (elicited every 20mV between —90 and
+30mYV) recorded before (Fig.2A) and 34 min-
utes after (Fig.2B) the application of MTX (5 nM)
are illustrated, note the ten-fold change in the
vertical scale in Fig.2B. The time-dependent
whole-cell currents in control saline were totally

swamped by the MTX-induced current (Jrx)
which shows little time dependence. During the
development of this inward current, however,
MTX had little effect on the three whole cell K*
channel currents or on the Ca?* channel current
which underlie the currents in Fig.2A (data not
shown). In fact, development of [Jyrx was little
affected by anumber of known K* channel block-
ers such as TEA (2—126mM), quinidine (0.5
mM), 4-AP (5mM), glybenclamide (10—-20.M)
or Ba2t (7.5mM).

The current-voltage (I-V) relationship of the
initial peak amplitude (/x:o) and the current at the
end of these depolarizing steps in control saline
(Uxca + Ikael) and 34min after MTX (5nM)
(Fig.2C) shows that Jrx has no voltage depend-
ence (linear) and reverses in direction near —20
mV, suggesting that MTX induces an increase in
the membrane conductance to the cations Na* and
K*, or to Cl, possibly by the opening of mem-
brane channels.

The ionic selectivity of these MTX-activated
channels was investigated by substituting the ex-
tracellular Na* and intracellular K* with other
cations known to have different permeating prop-
erties. In the TEA* (126mM) saline, [rx in-
duced by 5nM MTX (Fig.3A,top panel; B) was
linear between —90 and +40mV and reversed in
direction at -6O0mV, positive of Ex (—78mV),
suggesting that current flow was mainly carried
by K*. This Jmrx was blocked upon the addition

MAITOTOXIN INDUCES MUSCLE CONTRACTION

o Ga a
fone oh c
Test Ca Ss EAE
im cd?*
1NBOpAé
lsoms

=>
<

=
=
t
=
=
a
a
—
o
>

Ment. Guerent (pad)

7000

~£0-60840-20 GC 20 40
Memb. Potenticl (mV)

~AOb0~40—29 0 20 40
Memb Potentiol {mv}

FIG, 3. MTX-induced currenis in colonic cells
recorded with K* or Cs*-containing pipettes. MTX-
induced whole-cell currents (/yrx) every 20 mV be-
tween —90 and +30m¥V (from a holding potential of
~80mY) in colonic cells filled with K*-containing (A,
B hollow circles) or Cs*-containing (C, D hollow
circles) pipette solutions. Jjgrx was blocked by Cd**
(ImM) (A,C middle panels; C.D filled circles), these
effects of Cd** were quickly and readily reversible
upon Cd** removal (A.C Jower panels; B,D hollow
triangles). Short lines on either side of these panels
represent the zero current level. In A,8, the external
Na* concentration had been mostly replaced with the
impermeant TEA* (126mM), Iarx reversed in direc-
tion near 60m. In Cs*-filled cells /4y7 reversed in
direction near OmV (C,D).

of Cd?* (ImM) to the bathing solution
(Fig.3A,middle panel; B). These effects of Ca**
were. completely reversible upon its remoyal
(Fig.3A,B, lower panels; C,D), When the intracel-
Jular K* (126 mM) is replaced with Cs* (Fig.3C),
a cation known to be permeant through many
Ca** and non-selective cationic channels, but
mostly impermeant through K* channels

(Hille, 1984), the reversal potential of Jmrx was
near OmY (Fig.3D), a reversa) potential 10 to
20mV positive of the reversal potential obtained
when K* was the main intracellular monovalent
cation. The data suggest that Jmrx under normal
physiological gradients is carried by Na* and K*
and that Cs*, but not TEA, will also freely pass
through these MTX-induced channels

The channels activated by MTX may also allow
the flow of, or be modulated by Ca**. The
development of /wrx was not prevented if Ca?* in
the bathing solution was omitted, However, Jarry
increased in amplitude if the Ca** concentration
was raised to 6.25mM Ca**. Inrx was also
reduced but not blocked by nifedipine (1—10p,.M)
or La** (mM).

EFFECTS OF INTERNALLY-APPLIED MTX

When MTX (5nM) was added to the pipette
solution (containing 3mM EGTA) Isr
developed slowly after formation of the whole-
cell seal. Effects of MTX on the instantaneous
whole-cell current (Fig.4B) activated by a ramp
depolarization (ta potentials between —60 and
+100mV) showed this currentreversed near OmV
and was sensitive to blockade by Cd** (0,1mM)
(Fig.4C). This Jurx was only 200-300pA in
amplitude (at -60mV) compared with the 2000—
3000 pA current (at -80mV) observed when
MTX was applied externally (Fig.2), However, if
the Ca** chelating agent, EGTA, was omitted
from the pipette solution the day induced by
intermally applied MTX was larger and reversed
at more negative potentials (near -60mV)(n=2).
The addition of TEA (5mM) to bathing solution
substantially reduced this /wrx and shifted its
reversal potential to near OmV, suggesting that ir
the absence of EGTA the internal Ca** concentra-
tion is relatively high so that /cca contributes to
the measured increase in current. These data also
suggests that MTX can form/activate its channels
from the interna! surface of the membrane.

SINGLE CHANNEL RECORINNGS

Recordings of the current flow through single
MTX channels were made in the cell-attached
patch clamp mode with patch pipettes filled with
normal PS containing 10mM TEA and SaM
MTX. The membrane patches were depolarized
with ramp depolarizations to obtain the instan-
taneous I-V relationship of any open channels.
After seal formation the slope of the ramped-
evoked current increased with time until discrete
single channe) openings and closings were ob-
served. The current flow through these channels

538

MTX

+160mV

MT
16) -60

100pA
20ms

MEMOIRS OF THE QUEENSLAND MUSEUM

x MTX M
60

ee,
TX MTX
wP*| + 100mV

+100mV

200pA
100ms

200pA |
100ms

FIG. 4. Influence of pipette-applied MTX. A, cell-attached recordings of the development of MTX-activated
single channels. Pipette solution contained normal saline plus 1OmM TEA and SnM MTX. B, development of
whole-cell rx after introduction of MTX (5nM) into the cell interior. Pipette solution contained high K* saline
plus 3mM EGTA. /mrx activated by internal MTX was also sensitive to extracellular Cd?* (1mM) blockade

(C).

was inward at the resting membrane potential
(OmV added to the patch pipette). As the
membrane patch was depolarized, however, the
single channel amplitudes decreased until zero
current flow was recorded at a potential some
30mV positive of the resting membrane potential.
Outward current flow was recorded with further
depolarization of the membrane patch. Such a
reversal potential 30mV positive of the resting
membrane potential is consistent with the rever-
sal potential of Zmrx (-20 to -10mV) measured
under whole-cell voltage clamp (Figs 2A,3A).

DISCUSSION

The spontaneous contractions of the circular
muscle of the proximal colon and the contractions
to acetylcholine were inhibited concentration-de-
pendently by MTX (0.005-5nM). These effects of
MTX followed an initial transient contraction to
MTX (Fig.1) which was sensitive to blockade by
the muscarinic antagonist, atropine. Given that
MTX can stimulate rises in internal Ca’* levels
and neurotransmitter release from nerves and
glands (Kim et al.,1985; Gusovsky et al.,1988),
we suggest that this initial contraction arises from
the release of acetylcholine from cholinergic
motor neurones known to be present in this colon

preparation. The subsequent blockade of the
spontaneous activity and the contractions to
acetylcholine were mimicked by monensin (0.1-
10M), the Nat ionophore, suggesting that a rise
in the intracellular Na* 1s induced by MTX.
Atthe single cell level, MTX triggered a whole-
cell inward current, Iyrx, that was some 50-100
times larger than the holding current (at -30mV)
in control saline (Figs 2,3). The channels opened
by MTX appeared equally permeable to K* and
Na’ as the reversal potential of Iurx (-30 mV)
was midway between Ex and Ewa. Confirmation of
this reversal potential comes from the cell-at-
tached single channel data which showed that the
reversal potential of the single channel currents
was some 30mvV positive of the cell’s resting
membrane potential (likely to be -40 to -60mV)
(Fig.4A). Replacing most of the external Na*
with TEA* shifted the reversal potential of Jmrx
to near -60mV, suggesting that, under these con-
ditions, current flow was now mostly carried by
K*. Replacing the internal concentration K* with
Cs* moved the Jwrx reversal potential some
20mV positive (to OmV), even though it had a
greater driving force (no added Cs* in the bath
would mean that its Nernst potential would be
very negative), suggesting that Cs* does not flow
through these MTX channels as readily as K*.

MAITOTOXIN INDUCES MUSCLE CONTRACTION

Preliminary calculations from the amplitudes
of current flow through the MTX channels with
voltage suggest that these channels have a con-
ductance of up to 100pS. As MTX (5nM) ac-
Ovated an inward current of about 2000-3000 pA
(at -80mV), this suggests that 330-500 chan-
nels/cell were activated, These channel currents.
do not arise from the recruitment of normal volt-
age-operated ‘L-type’ Ca*+ channels since the
time course of Jcs (or any of the K* currents)
activated by membrane depolanzation was not
altered during the development of Jiy7x (Yoshii et
al.,1987); concentrations of nifedipine (10j4M)
which would block /ca. only slightly reduced
Inerx, and the reversal potential of /rx was ap-
proximately 50m¥V negative of the reversal
potential of Ica in Cs*-filled cells (Wogalis et
al,,1993). lt has been suggested, however, that
MTX may well modulate Ca* channels, remov-
ing their voltage sensitivities for activation and
inactivation and their ionic selectivity (Koba-
yashi ct al.,1987, Yoshii et al. 1987). Perhaps a

more attractive nd Sonnet is that MTX is trigger-
ing the opening of cation-selective channels nor-
mally opened by acetylcholine. These channels
allow the flow of small cations and Ca’*, have a
reversal potential near OmV and are opened by
muscarinic agonists which stimulate phos-
phoinositide hydrolysis and IP3-induced release
of stored Ca** (Sims, 1992). These channels, how-
ever, show a marked outward rectification at
negative potentials and have a single channe!
conductance of 20-25pS (Inoue et al., 1987). If
MTX is indeed opening these channels it must
also be modifying them, removing their rectify-
ing properties and perhaps inducing some sort of
‘channel clustering’. If this is correct, our results
indicate that 4-5 cholinergic channels would be
needed to form a ‘single’ conducting pore with a
conductance of 100pS. Other explanations are
that MTX is an ionophore, or that it induces a pore
in association with a membrane protein not
necessarily involved in ion conductance. The
Jeliry in the action of MTX may also suggests that
several MTX might be acting, cooperatively.

ACKNOWLEDGEMENTS

This work was supported by the National
Health & Medical Research Council of Australia.

LITERATURE CITED
BERTA, P., PHANEUF, S., DERNACOURT, J.,

CASANOVA, J.. DURAND-CLEMENT, M., LE
PEUCH, C., HAIECH, J. & CAVADORE J.-C.

539

1988, The effects of maitotoxin on phus-
phoinositides and calcium metabolism in a

culture of aortic smooth muscle cells,
Toxicon 26: 133-14).

GUSOVSKY, F.. YASUMOTO, T. & DALY, J.W.
1987, Maitotoxin stimulates phosphoinositide
breakdown in neuroblastoma hybrid NCB20 cells.
Cellutarand Molecular Neurobiology 7:317-222,

GUSOVSKY, F., DALY, I.W., YASUMOTO, T. &
ROJAS, B. 1988. Different effects of maitovoxin
on ATP secretion and on phosphoinositide break-
down in rat pheochromocytoma cells. FEBS Let:
ters 233: 139-142.

HAMILL, O.P.. MARTY, A.. NEHER, E., SAK-
MANN, B. & SIGWORTH, F.J. 1981, Improved
patch-clamp technique for high-resolution record-
ing from cells and cell-free membrane patches,
Pfliigers Archiv 391: &5—100.

HILLE, B. 1984, ‘Tonic channels of excitable
membranes’, (Sinayer Assoc. Sunderland: Mas-
sachusells).

HOLMES, ML. LEWIS, RJ. & GILLESPIE, N.C
1990, Toxicity of Australian and French
Polynesian strains of Gamnbierdiscus toxicus
(Dinophyceae) grown tn culture: Characterization
of anew maitotoxin, Toxicon 28: 1159-1172.

INOUE, R., KITAMURA, K. & KURIYAMA, H.
1987. Acetylcholine activates single sodium chan:
nels ip smooth muscle cells. Pfliigers Archiy 41>
674,

KIM. Y. LOGIN, LS. & YASUMOTO, T. 1985.
Muitotoxin activates quantal transmitter release at
the nev romuscular junction; evidence for elevated
intratenminal calcium in the motirnerve lerminal.
Brain Research 346: 357-362.

KOBAYASHI, M., GCHI, R. & OHIZUMI, Y. 1987.
Mailotoxin-activated single calcium channels in
guinea-pig cardiac cells. British Journal of Phar-
macology. 92: 665-671.

KOBAYASHI, M., OHIZUMI, Y, & OHIZUMI, Y,
1985. Ihe mechanism of action of maitotoxin in
relation to calcium movements in guinea-pig and
rat curdiae muscles, British Joumal of Pharmacol-
ogy 86;385-391

LANG, R.J.. VOGALIS, F. & LEWIS, RJ. 1992.
Maifotoxtn induces smooth muscle contracticn
and a non-selective cationic current in freshly
isolated smooth muscle cells of the guinea-pig
Bromaiat colon. Proceedings of the Austrahiitn
Physiology and Pharmacology Society 23; 1-107,

MEUCCI, O,, GIRMALDI, M., SCORZIELO, A.,
GOVONI, S., BERGAMASCHI, §&.,
YASUMOTO, T. & SCHETTINI, G. 1992,
Maitotoxin-induced intracellular calcium rise: iri
PC12 Cells: involvement of dihydropyridine-sen-
sitive and w-conotoxin-sensitive calcium chan-
nels and phosphoinositide breakdown, Journal of
Neurochemistry 59: 679-688.

OHIZUMI, Y. & YASUMOTO, T. 1983. Contractile
response Of the rabbit aorta to maitotoxin, the most

540

potent marine toxin. Journal of Physiology 337:
711-721,

RUDY, B. 1988 Diversity and ubiquity of K channels.
Neuroscience 25: 729-749,

SIMS, S.M. 1992. Cholinergic activation of a non-selec-
tive cation current in canine gastric smooth muscle
is associated with contraction. Journal of Physiol-
ogy 449: 377-398.

SLADECZEK, F., SCHMIDT, B.H., ALONSO, R.,
VIAN, L., TEP, A., YASUMOTO, T., CORY,
R.N. & BOCKAERT, J. 1988. New insights into
maitotoxin action. European Journal of
Biochenistry 174: 663-670.

VOGALIS, F., LANG R.J., BYWATER, R.A.R. &

MEMOIRS OF THE QUEENSLAND MUSEUM

TAYLOR, G.S. 1993. Voltage-gated ionic cur-
rents in smooth muscle cells of guinea pig
proximal colon. American Journal of Physiology
264: C527-C536.

YOKAYAMA, A., MURATAM, O.Y., IWASNITA, T
& YASUMOTO, T. 1988. Some clinical proper-
ties of maitotoxin, a putative calcium channel
agonist isolated from a marine dinoflagellate.
Journal of Biochemistry 104: 184-187.

YOSHII, M., TUSANOO, A., KURODA, Y., WU,
C.H. & NARAHASHI, T. 1987. Maitotoxin-in-
duced membrane current in neuroblastoma cells.
Brain Research 424: 119-125.

IMMUNOLOGICAL, BIOCHEMICAL AND CHEMICAL FEATURES OF

CIGUATOXINS: IMPLICATIONS FOR THE DETECTION OF CIGUATERIC FISH

RICHARD J, LEWIS

Lewis, RJ. 1994 08 01: Immunological, biochemical and chemical features of clguavoxins:
implications for the detection of ciguateric fish. Memoirs of the Queensland Museen $4(3)-
§41-548. Brisbane. ISSN 0079-8835.

A major advance in the management of ciguatera will come with the development of a
validated, cost-effective assay that detects ciguatoxins contaminating fish. Progress towards
such a goal is summarised and the implications for detection of the toxins involved in
ciguatera are discussed. Ciguatera results predominantly from CTX-1 which is present at
>D.Ippb (10-!9 mole/kg) in the flesh of carnivorous fish. Other toxins in ciguateric fish are
likely to have no more than a minor role in ciguatera. Consequently, CTX-1 should be the
principal target of any assay for ciguateric fish. However, significant levels of the less potent
ciguatoxins, particularly ciguatoxin-2 and -3, may also accumulate in fish and such toxins
could potentially interfere with the response of an assay, Ciguatoxins-!, -2 and -3 have an
affinity for voltage-dependant sodium channels (EDs9 = 0.2 -0.9nM) that is proportional to
their ip. LDsos in mice. Assays (biosensors) that measure this binding or perhaps the
ciguatoxin induced sodium channel opening may provide a sensitive assay for ciguatoxins
with a response proportional to toxin potency. Ciguatoxins also bind fo a range of other
proteins, which may interfere with the response of some assays. Alteratively, such interac-
tions may be utilised in the development of novel sandwhich-type assays. Ciguutoxins-L, -2
and -3 do not possess a useful chromophore for selective spectroscopic detection; however,
each possesses a relatively reactive primary hydroxy] through which a label could be attached
(after appropriate clean-up) prior to detection. Detectors (e.g. fluorescence or mass
spectrometry) coupled to optimised HPLC may provide the required sensitivity for analytical
detection of derivatised ciguatoxins in crude extracts of fish. Such assays could replace the
mouse bioassay for the validation of responses obtained by more rapid screening assays.

Richard J. Lewis, Southern Fisheries Centre, Queenstand Department of Primary Industries,

PO Box 76, Deception Bay, Queensland 4508; 22 November, 1993.

Ciguatera is a disease with a wide array of
gastrointestinal and neurological symptoms. It
stems mostly from the effects of ciguatoxin-1
(Fig.1), the most potent of the ciguatoxins
(Murata et al.,1990; Lewis et al..1991; Lewis &
Sellin, 1992). The disease can be debilitating and
slowly resolving but is seldom fatal. In many
Western countries outbreaks of ciguatera often
attract media attention with a consequent nega-
live impact on the marketing of seafood and
victims of ciguatera may seek compensation
through the courts. A cost-effective means of
detecting ciguateric fish prior ta consumption
remains one of the few management options that
can directly reduce the adverse impacts of
ciguatera, Antibody-based assays appear to hold
most promise since they are able to detect, under
favourable circumstances, targeted compounds
to 10°'? M and can be developed as cost-effective
screens (Gazzaz et al.,1992). This paper reviews
untibody-based assays for the identification of
ciguateric fish and discusses immunological,
biochemical and chemical features of the
ciguatoxins relevant to their detection.

DEVELOPMENT OF ANTIBODY-BASED
ASSAYS FOR CIGUATERIC FISH

The potential of an antibody-based screening
assay for detecting ciguatoxin in fish flesh was
first indicated by Hokama et al. (1977). Hokama
has since led efforts to develop a rapid screen fur
ciguateric fish (Hokama,1991). The original
radioimmunoassay screened 88% of moray ecl
and 38% of other fish as toxic (>3.5 x 10°cpm/g:
Kimura et al.,1982) despite these fish rating as
non-toxic by the mongoose assay. Despite this
high false positive result, all fish rating toxic by
the mongoose rated as toxic by the antibody
assay, indicating the potential for this approach
to detect ciguateric fish. This assay was sub-
sequently employed to screen 5,529 Seriola
dumerili captured in Hawaiian waters (Kimura el
al.,1982), This study found 13% of S. dumerili
tested positive, with the remaining fish, including
those >%kg which normally are not marketed
owing to their perceived higher nsk of ciguatera,
being consumed without incident. The quantity
of additional S. durnerili entering the market in-

342

creased 62% as the direct result of the study,
However, Kimura ef ai. (1982) found that 7% of
fish (3 of 42) clinically implicated in ciguatera
tested negative by the radioimmunoassay.

In 1984 the radioimmunoassay for ciguateric
fish was replaced by a simpler enzyme im-
munoassay that in binding inhibition assays was
sensitive to as little as Spg of free ciguatoxin
(Hokama et al., 1983,1984), The cross-reactivity
of this assay with other polyether toxins was also
documented. An enzyme labelled polyclonal an-
libedy was subsequently used to develop a further
simplified ‘stick-test’ that rapidly distinguished
toxic from non-toxic flesh samples (Hokama,
1985); however six tests per fish appeared neces-
sary for accurate determination of ciguateric fish
that tested close to the borderline level. With the
stick-test the rejection rate for S. durmerili was
only 11% and S. dumerili testing non-toxic were
consumed without incident, This assay res-
ponded directly to >1.0ng of pure ciguatoxin
(Hokama, 1985),

These carly studies all employed a polyclonal
antibedy raised to ciguatoxin in sheep with the
disadvantage that for long-term antibody produc-
tion a continual supply of antigen is required for
booster injections. Monoclonal antibodies, on the
other hand, can provide a continuous supply of a
selected antibody. Hokamia et al. (1985,1989b)
reported production of monoclonal antibodies to
a related polyether toxin okadaic acid as well as
to ciguatoxin (likely CTX-1). Using a mono-
clonal IgG to ciguatoxin in a stick enzyme im-
munoassay, Hokama et aj. (19898) found that
98% of fish implicated in ciguatera (50 of 51)
lested positive, while 55% of 4 random mix of
fish, 55% of Crenochaetus striatus and 44% of S.
dumerili tested positive. A further simplified
solid-phase immunobead assay for detection of
ciguateric fish (Hokama,1990) appeared more
sensitive than previous slick tests.

The monoclonal antibody to ciguatoxin used in
these studies bas been assessed for cross-reac-
livily to other polyether toxins (Hokami et ai.,
1989b, 1992). The assay employing the antibody
raised to ciguatoxin detected similar concentra-
tions of pure ciguatoxin, okadaic and a syn-
thesised fragment of okadaic acid (ECs =~ 0.5
ng of toxin per ml methanol). The croés-reactivily
of an antibody for ciguatoxin and okadaic acid is
somewhat unexpected, given recent molecular
modelling studies which show that in solution the
structure of okadaic acid is quite different from
that of ciguatoxin (Norte et al,,1991; Lewis and
Ramsdale, unpublished observations). It is also

MEMOIRS OF THE QUEENSLAND MUSEUM

intriguing that a monoclonal! antibody obtained
specifically to okadaic acid was less sensitive at
detecting okadaic acid (ECso ~1 Sng/ml) than the
ciguatoxin antibody using the same assay format.
Somewhat different cross-reactivity was reported
in an earlier study (Hokama et al.,1989b), Impor-
tantly, addition of pure ciguatoxin (ECso ~1
ng/ml) and okadaic acid (ECs ~3ng/ml) to a
ciguatoxin antibody inhibited the subsequent
binding of this antibody to stickscoated with an
extract from a fish implicated in ciguatera. This
result suggests these polyether toxins compete at
a specific, saturable site on the IgG. Since
ciguatoxin alone can bind to correction fluid
coated sticks (the Hokama poke stick method is
based on the whility of sticks dipped into an
alchohol-based typist’ s correction fluid to extract
and immobilise CTX prior to detection of this
immobilised CTX with labelled antibody) and
subsequently bind antibody but the preformed
antibody-CTX complex is no longer able to bind
to such sticks through the CTX link, it is possible
that CTX binding sites on the coated sticks are
also saturable,

Development of a commercial screening assay
for ciguatenc fish is being undertaken by Hawaii
Chemtec International who purchased develop-
ment nghts for the Hokama stick test from the
University of Hawaii. Recent findings are given
in other articles in this memoir.

IMMUNOLOGICAL FEATURES

An antibody response is elicited in an animal
following injection of an immunogenic antigen.
The ciguatoxins (srt/z = 1,094—1,110) are relative-
ly small haptens that are likely ta have only low
immunogenicity. The absence of any protection
in people repeatedly exposed to ciguatoxins
through their diet supports this suggestion. A
hapten will become immunogenic when
covalently conjugated to a carrier protein pos-
sessing high immunogenicity (Erlanger,1980).
However, non-covalent conjugation and/or
Frnends” complete adjuvant are unable to sig-
nificantly enhance the immunogenicity of native
haptens (Layton et al..1987; Gazzaz et al.,1992).
One draw -back with having to use a hapten con-
jugated to a carrier protein ts that the antibodies
obtained often do not have high specificity for the.
unconjugated (native) form of the hapten,

Ciguatoxins possess a relatively reactive
primary hydroxyl (Murata et al..1990; Lewis et
al.,1991; Lewis et al.,1993) which can be reacted
with succinic anhydride to yield a hemisuccinate.

FEATURES OF CIGUATOXINS IMPORTANT POR DETECTION

The hemisuccinate so formed has an available
carboxyl group through which ciguatoxin can
now be linked to a carer protein (e.g. bovine
serum albumin, keyhole limpet haemocyanin,
ovalbumin) using a water soluble carbodiimide
cross-linking reagent. Ciguatoxin cross-linked to
a4 Carrier protein in this Way is expected to have
considerably enhanced immunogenicity com-
pared with the native ciguatoxin, Such acomplex
may additionally have reduced potency, an im-
portant consideration for in vive immunisation.
The availability of only one casily accessible site
on ciguatoxin for conjugation to a protein limits
(Mandal & Latif, 1988) the possibilities for
producing a range of antibodies possessing dif-
fering selectivities for the various ciguatoxin
analogues. To-date antibodies to ciguatoxin have
not been obtained using an immunogen covalent-
ly attached to a carrier protein. However, the use
of such an immunogen is presently under con-
sideration in a number of laboratories. Attempts
in our laboratory to produce a hemisuccinate of
ciguatoxin-1] have met with little success, despite
the use of succinic anhydnde in dried pyridine at
80°C (Baden et al.,1984), a method that we suc-
cessfully used to produce a hemisuccinate of
brevetoxin-3 (Lewis unpubl, data),

To obtain a maximal in vive immune response
with small quantities of immunogen, the im-
munogen should be emulsified in Fruends’ com-
plete adjuvant prior to injection (Vaitukaitis,
1983; Smith et al.,1992). Blood from suitably
immunised animals is then collected and assessed
for selective antibody titre against the compound
of interest. For monoclonal antibedy production,
the spleen cells of immunised mice are fused with
a murine myoloma cel] line, with hybridomafs)
secreting ciguatoxin-specific immunoglobulins
subsequently isolated using an appropriate
screen. This approach follows well described pro-
cedures (Galfré & Milstein, 1981; Goding, 1Y86;
Peters & Baumgarten,1992). Alternatively,
monoclonal antibodies can be obtained through
in vitro immunisation procedures (Brazeau et al.,
1982; Van Ness etal., 1984; Buchman etal.,1985;
Borrebaeck,1986; Brams et al..1987: James &
Bell,1987; Borrebaeck & Glad,1989). This ap-
proach allows the production of high specificity
antibodies (including human antibodies) with
small quantities of immunogen and compared
with in vivo immunisation ts less susceptible to
immunogen toxicity. Modifications to standard
in vivo immunisation protocols can alsa reduce
the quantity of immunogen required
(Vaitukaitis, }981) Forest & Ross,1993), Such wn

543

vitra and in vive approaches may be useful for
prodoction of antiboches to ciguatoxin-

To allow detection of ttially useful an-
tibodies, a screen must be loped that is selec-
tive only for those antibodies that combine with
high affinity to the compound being targeted.
Such a screen may utilise a hapten-protein con-
jugate thats attached, through non-specific inter-
actions, to the plastic surface of microtitre plates.
To avoid potential problems of cross-reactivity to
camer protein or lo epitopes extending beyond
the hapten itself, a protein and cross-linking
reagent should be used in screening that 1s dif-
ferent from that used in the preparation of the
immunogen. Additionally, any method employed
to couple ciguatoxin to a carer protein should
notalter the structure of the targeted compound,
otherwise antibodjes may be produced that will
not recognise the native compound, This ts equal-
ly important for compounds Jabelled for use as
the competitor in competitive antibody binding
assays. The careful choice of blockers (Tween 20,
fetal calf scrum, albumins etc) must be made to
ensure thal the response is specific for the com-
pound of interest (e.g. ciguatoxin). One pitfall is
the ability of certain clones to express antibody
which binds promiscuously to plastic (Conger et
al..1988). We have noted that exposing plastic
microtitire plates to methanol-water (1:1) in-
creases the plate’s affinity for IgG, an effect thal
could not be blocked by traditional blockers and
gave Msc to a oumber of false postive results.

Ciguatoxin recognising antibodies obtained in
an appropriate manner, and with confirmation
that they are specific for only the CTX-class of
polycthers, could then be used tn the detection of
ciguateric fish. Screening assays employing an-
libodies that have both high affinity and selec-
tivity for an epitope on ciguatoxin-1 are essential
if routine detection of the low levels of
ciguatoxins (10' to 5 x 10% CTX-1/g) con-
taminating the flesh of ciguateric fish is to be
achieved. Cross- reactivily to other ciguatoxin
congeners may also be acceptable. Ideally the
chosen antibodies should have an affinity that is
directly proportional to the oral potency (to
humans) of the contaminating toxins and should
mot cross-react with compounds normally present
in non-toxic fish. The high cross-reactivity of
ciguatoxin antibodies to less patent and struc-
turally dissimilar polyether toxins such as
okadaic acid (Hokama et al_,1989b,1992) may in
part explain the high number of false positive
results obtained with assays employing such an-
tibodies, Recent studies using steroid-antibody

sae

EDsg (ng/ml tor Inhibition of
PbTx-3 binding to sodium channels

o os 1.0 5 20 a5

LDsq (ng/a) ip. in mice

FIG, |. Relationship between sodium channel binding
affinity and mouse lethality for three ciguatoxins.

interactions as 4 model, have revealed that an-
uibodies recognising apolar and functionally inert
molecules (such as ciguatoxin) can have a high
afinity binding site but that this site apparently
can’t be engineered with high specificity to avoid
its cross-reactivity with related ligands (Arevalo
et al.,1993). In contrast to antibodies that recog-
nise ciguatoxin, and somewhat surpnsingly given
the previous statement, antibodies raised to
brevetoxin and okadaic acid have been found to
possess low cross- sreactivity with other
polyethers including CTX-1, -2, ‘and -3 (Levine
et al,,1988: Lewis et al.,1991; Poli et al.,1992)
and it remains to be confirmed if it is possible to
obtain brevetoxin-antibodies that cross-react
with ciguatoxin. Non-selective bindings of
ciguatoxin to IgG and non-selective binding of
IgG to fish tissue (Parc et al .1979; Chanteau et
al.,1981; Emerson ct al..1983) may present addi-
tional obstacles to the development of a success-
ful screening assay for ciguateric fish. The
limited experience of other laboratories with
several prototype assays utilising polyclonal
(Berger & Berger,1979) and monoclonal an-
tibodies (Lewis, unpubl, data) developed by
Hokama have at times given less than satisfactory
results that are difficult to explain.

BIOCHEMICAL FEATURES

Ciguatoxin binding to sodium channels
Ciguatoxins are characterised by high affinity
binding (EDso=0.23-0.85ng/ml) for voltage sen-
sitive sodium channels (Lewis et al.,1991). The
binding affinity of each ciguatoxin for the sodium
channel (EDso) is p: ional to its i.p. LDso in
mice (Lewis et al.,1991) indicating that the lethal

MEMOIRS OF THE QUEENSLAND MUSEUM

effect of the ciguatoxins likely stem from their
action on sodium channels. Interestingly, this
relationship is found to be linear (EDso = 0.30 x
LDso + 0,16) for CTX-1, -2 and -3 (Fig. 1). That
this relationship has a slope of 0,3 indiates that
binding affinity is not directly proportional to
lethality and additional factor(s), presumably
pharmacokinetic (Lewis et al.,1991) attenuate the
lethality of the ciguatoxins with lower binding
affinity. Since these EDsos are a measure of the
on-rate for binding, differences in off-rate may in
part account for the above result. However, recent
pharmacological studies indicate that the off-
rates for ciguatoxins-1, -2 and —3 are similariy
slow (Lewis & Wong Hoy,1993).

Ciguatoxins are sodium channel activator
toxins that bind to site 5 on sodium channels, a
site overlapping the brevetoxin binding site
(Lombet et al.,1987; Lewis et al.,1991). Cigua-
toxin binding Jeads to the opening of sodium
channels which in tum results in an influx of
sodium tons, cell depolarisation and the ap-
pearance of spontaneous actions potentials. The
high affinity binding and subsequent alteration of
voltage dependent sodium channels by
ciguatoxins could form the basis of biosensor-
type assays able to screen for the ciguatoxins. The
technology for developing such biosensors is
progressing rapidly (Ogert et al.,1992; Malm-
qvist, 1993) but there are few commercial biosen-
sor products available (Griffiths & Hall,1993),
The advantage of biosensor technology is that the
response can be proportional to level of sodium
channel activator in a mixture. The mouse bioas-
say could be considered a crude form of biosen-
sor. This assay has been validated for detection
of ciguatoxins in up to 20mg of lipid extract from
fish flesh (Lewis & Sellin, 1993). Cell based
assays also show potential for detection of
sodium channel blocking toxins (Gallacher &
Birbeck,1992) and hold potential for detecting
sodium channel activator toxins (Hungerford,
1993). Such assays also have the potential to be
automated and miniaturised (Goguen & Keder-
sha, 1993).

Ciguartoxin binding to other proteins
Ciguatoxins also have affinity for various
proteins including IgG from a vanety of sources
and fish liver and fish flesh proteins (Parc et al.,
1979; Emerson et al.,1983; Vernoux et al., 1985;
Hahn & Capra,1992). The affinity of ciguatoxin
for these proteins has not been quantified and
these studies have failed to exclude the possibility
that the binding they are measuring is not simply

FEATURES OF CIGUATOXINS IMPORTANT FOR DETECTION

w
>
uw

CTX-1 A, = 0H
CTX-2 R,=H
CTX-3 R,=H

FIG, 2, Structures of the major ciguatoxins presentin the flesh of carnivorous fish in the Pacific, CTX-2 is 52-<pi
CTX-3 (stereochemistry shown for CTX-1 and CTX-3).

the binding of ciguatoxin to the sodium channels
present in the protein-containing extracts. Such
an interaction is expected to be considerable in
tissues rich in excitable cells (e.g. flesh). A puta-
tive ciguatoxin binding protein has been detected
in fish flesh and its appearance has been proposed
to result from the exposure of fish to ciguatoxin
(Hahn & Capra,1992). Such a protein in fish may
explain why fish are relatively less susceptible to
the ciguatoxins than the closely related breve-
toxins (Lewis,1992). An assay to detect such a
protein in fish could provide the basis for a novel
screening assay for ciguateric fish.

Any high affinity binding of ciguatoxins to
proteins (including sodium channels) present in
fish tissue could be used to immobilise ciguatoxin
to a solid phase prior to detection with a labelled
antibody specific for ciguatoxin-1. This approach
can be considered a type of sandwich assay and
could allow rapid detection of ciguatoxin in fish
flesh without the need for a solvent extraction
step. Such an approach warrants further inves-
tigation but requires an antibody that can ‘see’ the
portion of ciguatoxin molecule not obscured as a
result of its binding to the fish protein,

CHEMICAL FEATURES OF THE
CIGUATOXINS

Structural features of the ciguatoxins
Ciguatoxins (Fig. 2) are lipid-soluble toxins
consisting of 13 rings fused by ether linkages into

a mostly rigid, ladder-like structure (Murata et
al.,1990; Lewis et al.,1991,1993a); they are rela-
tively inert molecules that can accumulate in fish
to levels that are toxic to humans, a feature that
distinguishes them from chemically and
biochemically related brevetoxins which are
lethal to fish before levels can be accumulated
that cause human poisoning. Structurally related
ciguatoxins-], -2 and -3, are present in car-
nivorous fish (Lewis et al,,1991; Lewis & Sel-
lin, 1992; Lewis et al.,1993a). However, CTX-1
contributes most to the toxicity of ciguateric fish
(Lewis & Sellin,1992) so an assay selective for
CTX-1 would have general utility for detection
of ciguateric fish. Involvement in human poison-
ings of other toxins (CTX-2, -3, gambiertoxins
etc) in ciguateric fish remains unsubstantiated.
The role in ciguatera for toxins other than site 5
sodium channel activator toxins is remote.

Extraction of ciguatoxins

Ciguateric fish contain 0.1-5ppb CTX-2
(Lewis & Sellin, 1992; Lewis, 1992), Whilst direct
detection may be possible, extraction and clean-
up procedures can be used to concentrate the
ciguatoxins, The ciguatoxins are sufficiently
hydrophobic to be insoluble in water but can be
extracted from fish flesh with organic solvents
such as methanol, chloroform or acetone. How-
ever, even with a several step clean-up procedure
involving removal of low polarity lipids and
water-soluble material, the ciguatoxins are still

present in the remaining lipid minature al low
levels of 30-5,000ppb. The lipid solubility and
lack of a rapid, efficient clean-up procedure also
have implications for antibody-based screening
assays. Any organic solvents required to
solubilise crude extracts conisining ciguatoxins
dunng incubation with an antibody may alter the
three- dimensional structure of the antibody bind-
ing domain, thereby reducing the affinity of the
antibody for ciguatoxin.

Chemical detection of the ciguatoxins

Pure CTX-1 can be detected tn helow Sng by
monitoring HPLC eluants with a sensitive u.v.
deteclor (Lewis, unpubl. data). However, since
the ciguatoxins do not possess a distinctive uv.
chromophore. it is nol possible to develop a
method for selectively detecting ciguatoxin in a
ervde lipid extract from fish by employing 4 u.v.
detector to monitor the eluant from a HPLC run.
As discussed previously, the major ciguatoxins
found in fish each possess a relatively reactive
primary hydroxyl through which Jabels could be
attached to enhance detectability. Since
numerous compounds in a lipid extract from fish
might also react with such a label, detection of the
tubelled ciguatoxins must be done incombination
with 4 subsequent separation technique. HPLC
coupled to Fluorescence detection provides a high
sensilivily method that has the potential to detect
natural levels of ciguatoxins m crude extracts
from fish flesh. Such an analytical detechon
methodology could be used to validate positive
(or negative) responses obtained by antibody-
based assays, Dickey et al. (1992) reported en-
couraging results by labelling ciguaiaxin with
novel courmarin based fluorescent reagents sold
by Molecular Probes Inc, (Eugene, OR, USA).

Although sensitive detection of anthroy) nitrile
labelled ciguatoxin has been reported (Legrand et
al.,1992), these approaches have not been ex-
tended to detection of ciguatoxins in crude ex-
tracts from fish. HPLC linked to a fluorescence
detector was able to detect in relatively crude
extracts 2400ng diarrhetic shellfish toxins/¢
shellfish (Lee et al.,1987) and 21 3ng aflatoxin/g
peanut butter or corn (Park et al.,1990), These
approaches would require modification and sig-
nificant improvements in sensitivity if the
ciguatoxins in crude extracts from lish are lo be
detected. A simple clean-up procedure that min-
imises inference from lipids at both the labelling
ani detection steps is critical for the development
of a sensilive test for ciguatoxins, HPLC coupled
io sclective-ion monitoring lonspray mass

MEMOIRS OF THE QUEENSLAND MUSEUM

spectrometry has shown considerable potential
for the detection of labelled diarthetic shellfish
toxins {Pleasance et al,1992). Initial studies with
CTX-1 indicate thatsuch an approach could form
the basis of a confirmatory analytical assay for
ciguatoxins in fish (Dickey et al., in press; Lewis
etal..1999h). [sufficient unknown toxin is avail-
able (>25jug). NMR approaches can be utilised
for confirmation or for characterisation of un-
known toxins

In addition to the major ciguatoxins isolated.
loW potency forms arising from further
biotransformution of the ciguatoxins (Tosteson et
al.,£988; Legrand e1 al..1992; Lewis & Sellin,
1992; Lewis et al,,1992) may also accumulate in
fishes. While low potency of these ciguatoxins
makes detection by the mouse biogssay difficult,
such compounds may cross-ceact with antibodies
to CTX-1, thereby increasing the probability of
obtaining false positive results. This problem
arises because antibody-based assays respond
depending on the relative affinity (specificity) of
the antibody for cach form of the toxin in a way
that may only by chance be related to the potency’
of the different forms.

LITERATURE CITED

AREY ALO, J.H., TAUSSIG, M.J. & WILSON, LA.
1993, Molecular basis of crossreactivity and the
limits of antibody-antigen complementarity. Na-
ture 365: 859-863.

BADEN, D.G,, THOMAS, MENDE, TJ., WALLING,
J. & SCHULTZ, D.R. 1984. Specific anubodies
directed against toxins of Prychadiscus brevis
(Florida’s red tide dinoflageilate). Toxicon 22;
783-784.

BERGER, J.A, & BERGER, L.R. 1979. Stndies to
develop a colorimetric ELISA test to assay
ciguatoxin in fish tissue. Revue Intemationale
Océanographique Médicine 53-54: 23-32.

BORREBAECK, C.A.K. 1986, Jn vitra immunization
for production of murine and human monoclonal
antibodies: present status. Trends in Biotechnol-
ogy June: 147-153.

BORREBAECK, C.A.K. & GLAD, C, 1989, Cross-
reactivity of mouse monoclonal antibodies
produced by i vitra or in vive immunization.
Immunology Letters 20: 255-260.

BRAMS, P., PETTUOHN, D.E., BROWN, M. &
OLSSON, L. 1987. Jn vitro Blymphocyte antigen
priming against both non- immunogenic and 1m-
munogenic molecules requiring low ammounts of
antigen and applicable in hybridoma technology.
Journal of Immunological Methods 98: 11-22.

BRAZEAU, P., BOHLEN, & GUILLEMIN, R. 1982.
Moncelonal antibodies to hypothalmic growth

FEATURES OF CIGUATOXINS IMPORTANT FOR DETECTION

hormone-releasing factor with picomoles of an-
tigen. Science 218: 887-889.

BUCHMAN, D., MILLER, D. & KOCH, G. 1985.
Monoclonal antibodies to digoxin: comparison of
in vitro and in vivo immunization. Hybridoma 4:
173-177.

CHANTEAU, S., LECHAT, I., PARC, F. & BAGNIS,
R. 1981. Essai de détection de la ciguatoxine par
une méthode immunoenzymatique. Bulletin de la
Société de Pathologie Exotique 74: 227-232.

CONGER, J.D., PIKE, B.L. & NOSSAL, G.J.V. 1988.
Analysis of the B lymphocyte repertoire by
polyclonal activation, Hinderance by clones yield-
ing antibodies which bind promiscuously to plas-
tic. Journal of Immunological Methods 106:
181-189.

DICKEY, R.W., BENCSATH, F.A., GRANADE, H.R.
& LEWIS, R.J. in press. Liquid chromatographic-
mass spectrometric methods for the determination
of marine polyether toxins. Fourth International
Conference on Ciguatera Fish Poisoning, Tahiti.

EMERSON, D.L., GALBRAITH, R.M., MCMILLAN,
J.P. & HIGERD, T.B. 1983. Preliminary im-
munologic studies of ciguatera poisoning. Ar-
chives Internal Medicine 143: 1931-1933.

ERLANGER, B.F. 1970. The preparation of antigenic
hapten- carrier conjugates: a survey. Methods in
Enzymology 70(A): 85-104.

FORREST, J.M.S. & ROSS, H.A. 1993. Splenic im-
plantation of nitrocellulose pellets as a means of
raising antibodies to nanogram quantities of an-
tigen. Journal of Immunological Methods 159:
275-276.

GALFRE, G. & MILSTEIN, C. 1981. Preparation of
monoclonal antibodies: strategies and procedures.
Methods in Enzymology 73: 3-46.

GALLACHER, S. & BIRKBECK, T.H. 1992. A tissue
culture assay for direct detection of sodium chan-
nel blocking toxins in bacterial culture supernates.
FEMS Microbiology Letters 92: 101-108.

GAZZAZ, 8.S., RASCO, B.A. & DONG, F.M. 1992.
Application of immunochemical assays to food
analysis. Critical Reviews in Food Science and
Nutrition 32: 197-229.

GODING, J.W. 1986. ‘Monoclonal Antibodies: Prin-
ciples and Practice’. (Academic Press: London).

GOGUEN, B. & KEDERSHA, N. 1993. Clonogenic
cytotoxicity testing by microdrop encapsulation.
Nature 363: 189-190.

GRIFFITHS, D. & HALL, G. 1993. Biosensors - what
real progress is being made? Trends in Biotech-
nology 11: 122-130.

HAHN, S.T., CAPRA, M.F. & WALSH, T.P. 1992.
Ciguatoxin- protein association in skeletal muscle
of Spanish mackerel (Scomberomorus commer-
soni). Toxicon 30; 843-852.

HOKAMA, Y. 1985. A rapid, simplified enzyme im-
munoassay stick test for the detection of
ciguatoxin and related polyethers from fish tis-
sues. Toxicon 23: 939-946.

HOKAMA, Y. 1990. Simplified solid-phase im-

547

munobead assay for detection of ciguatoxin and
related polyethers. Journal of Clinical Laboratory
Analysis 4: 213-217.

HOKAMA, Y. 1991, Immunological analysis of low
molecular weight marine toxins. Journal of
Toxicology - Toxin Reviews 10: 1-35.

HOKAMA, Y., BANNER, A.H. & BOYLAN, D.B.
1977. A radioimmunoassay for the detection of
ciguatoxin. Toxicon 15; 317—-325.

HOKAMA, Y., ABAD, M.A. & KIMURA, L.H. 1983.
A rapid enzyme-immunoassay for the detection of
ciguatoxin in contaminated fish tissues. Toxicon
21: 817-824.

HOKAMA, Y., KIMURA, L.H., ABAD, M.A.,
YOKOCHI, L., SCHEUER, P.J., NUKINA, M.,
YASUMOTO, T., BADEN, D.G. & SHIMIZU,
Y. 1984. An enzyme immunoassay for the detec-
tion of ciguatoxin and competitive inhibition by
related natural polyether toxins. Pp. 307-320. In
Ragelis, E.P. (eds), ‘Seafood toxins’. (American
Chemical Society: Washington, D.C.).

HOKAMA, Y., OSUGI, M., HONDA, S.A.A., MAT-
SUO, M.K. 1985. Monoclonal antibodies in the
detection of ciguatoxin and other toxic polyethers
in fish tissues by a rapid poke stick test. Pp.
449-455. In Gabrie, C. & Salvat, B., (eds).
‘Proceedings of the Fifth International Coral Reef
Congress, Tahiti, Vol 4’. (Antenne Museum-
Ephe: Moorea).

HOKAMA, Y., HONDA, 8.A.A., KOBAYASHI,
M.N., NAKAGAWA, L.K., ASAHINA, A.Y.,
MIYAHARA, J.T. 1989a. Monoclonal antibody
(MAb) in detection of ciguatoxin (CTX) and re-
lated polyethers by the stick-enzyme immunoas-
say (S-EIA) in fish tissues associated with
ciguatera poisoning. Pp. 303-310. In Naton, S.
Hashimoto, K. & Ueno, Y. (eds), ‘Mycotoxins and
phycotoxins "88. (Elsevier: Amsterdam).

HOKAMA, Y., HONDA, S.A.A., ASAHINA, A.Y.,
FONG, J.M.L., MATSUMOTO, C.M. & GAL-
LACHER, T.S. 1989b. Cross-reactivity of
ciguatoxin, okadaic acid, and polyethers with
monoclonal antibodies. Food and Agricultural
Immunology 1: 29-35.

HOKAMA, Y., HONG, T.W.P., ISOBE, M.,
ICHIKAWA, Y. & YASUMOTO, T. 1992.
Cross-reactivity of highly purified okadaic acid
(OA), synthetic, spiroketal east sphere of OA and
ciguatoxin. Journal of Clinical Laboratory
Analysis 6: 54-58.

HUNGERFORD, J.M. 1993. Seafood toxins and
seafood products. Joumal AOAC Intemational
76: 120-130.

JAMES, K. & BELL, G.T. 1987. Human monoclonal
antibody production: current status and future
prospects. Journal of Immunological Methods
100: 540.

KIMURA, L.H., ABAD, M.A., & HOKAMA, Y. 1982.
Evaluation of the radioimmunoassay (RIA) for
detection of ciguatoxin (CTX) in fish tissues. Jour-
nal Fish Biology 21: 671-680.

348

LAYTON, G.T., STANWORTH, D.R. AMOS, H.E.
1987. Factors influencing the immunogenicity of
the haptenic drug chlorhexidine in mice - Part 1.
Molecular requirements for the induction of IgE
and IgG anti-hapten antibodies. Molecular Im-
munology 24: 133-141.

LEE, J.S., YANAGI, T., KENMA, R. & YASUMOTO,
T. 1987. Fluorometric determination of diarrhetic
shellfish toxins by high-performance liquid
chromatography. Agricultural Biological
Chemistry 51: 877-881.

LEGRAND, A-M., FUKUI, M., CRUCHET, P.,
ISHIBASHI, Y. & YASUMOTO, T. 1992. Char-
acterization of ciguatoxins from different fish
species and wild Gambierdiscus toxicus. Pp, 25-
32. In Tosteson, T.R. (ed.), ‘Proceedings of the
Third International Conference on Ciguatera Fish
Poisoning, Puerto Rico’. (Polyscience Publica-
tions: Québec).

LEVINE, L., FUJIKI, H., YAMADA, K., OJIKA, M.,
GJIKA, H.B. & VAN VUNAKIS, H. 1988.
Production of antibodies and development of a
radioimmunoassay for okadaic acid. Toxicon 26:
1123-1128.

LEWIS, R.J. 1992. Ciguatoxins are potent ich-
thyotoxins. Toxicon 30: 207-211.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fishes. Toxicon 30: 915-919.
LEWIS, R.J. & SELLIN, M. in press. Recovery of

ciguatoxin from fish flesh. Toxicon.

LEWIS, R.J. & WONG HOY, A.W. 1993. Comparative
action of three major ciguatoxins on guinea-pig
atria and ilea. Toxicon 31: 437-446.

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae). Toxicon 29; 1115-1127.

LEWIS, R.J., SELLIN, M., STREET, R., HOLMES,
M.J, & GILLESPIE, N.C. 1992. Excretion of
ciguatoxin from moray eels (Muraenidae) of the
central Pacific. Pp. 131-143. In Tosteson, T.R.,
(ed.), ‘Proceedings of the Third International Con-
ference on Ciguatera Fish Poisoning, Puerto
Rico’. (Polyscience Publications: Québec).

LEWIS, R.J., NORTON, R.S., BRERETON, I.M, &
ECCLES, C.D. 1993. Ciguatoxin-2 is a
diastereomer of ciguatoxin-3. Toxicon 31: 637-
643.

LEWIS, R.J., HOLMES, M.J., ALEWOOD, P.A. &
JONES, A. in press. lonspray mass spectrometry
of ciguatoxin-1, maitotoxin-2 and -3 and related
marine polyether toxins. Natural Toxins.

LOMBET, A., BIDARD, J.-N. & LAZDUNSKI, M.
1987. Ciguatoxin and brevetoxins share a com-
mon receptor site on the neuronal voltage-depen-
dant Na* channel. FEBS Letters 219: 355-359.

MALMQUIST, M. 1993. Biospecific interaction
analysis using biosensor technology. Nature 361:
186-187.

MANDAL, C. & LATIF, N.A. 1988. Production of

MEMOIRS OF THE QUEENSLAND MUSEUM

highly specific polyclonal and monoclonal an-
tibodies using estradiol-3-O-carboxymethy] ether
as hapten. Steroids 52: 551-560.

MURATA, M., LEGRAND, A.M., ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T, 1990. Structures
and configurations of ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386.

NORTE, M., GONZLEZ, R., FERNNDEZ, J.J. &
RICO, M. 1991. Okadaic acid: a proton and carb-
on NMR study. Tetrahedron 47: 7437-7446,

OGERT, R.A., BROWN, J.E., SINGH, B.R.,
SHRIVER-LAKE, L.S. & LIGLER, F.S. 1992.
Detection of Clostridium botulinum toxin A using
a fiber optic-based biosensor. Analytical
biochemistry 205: 306-312.

PARC, F., DUCOUSSO, S., CHANTEAU, E., CHUN-
GUE, E. & BAGNIS, R. 1979. Problems linked to
the ciguatoxin immunological detection. Toxicon
17, Supplement 1: 137.

PARK, D.L., NESHEIM, S., TRUCKSESS, M.W.,
STACK, M.E. & NEWELL, R.F. 1990. Liquid
chromatographic method for determination of
aflatoxins B}, B2, G; and G2 in corn and peanut
products: collaborative study. Journal of the As-
sociation Official Analytical Chemists 73: 260—
266.

PETERS, J.H. & BAUMGARTEN, H. (eds) 1992.
‘Monoclonal antibodies’. (Springer-Verlag: Ber-
lin).

PLEASANCE, S., QUILLIAM, M.A. & MARR, J.C.
1992. lonspray mass spectrometry of marine
toxins. IV. Determination of diarrhetic shellfish
poisoning toxins in mussel tissue by liquid
chromatography/mass spectrometry. Rapid Com-
munications in Mass Spectrometry 6: 121-127.

SMITH, D.E., O’BRIEN, M.E., PALMER, VJ. &
SADOWSKI, J.A. 1992. The selection of an ad-
juvant emulsion for polyclonal antibody produc-
tion using a low-molecular-weight antigen in
rabbits. Laboratory Animal Science 42: 599-601.

TOSTESON, T.R., BALLANTINE, D.L. & DURST,
H.D. 1988. Seasonal frequency of ciguatoxic bar-
racuda in southwest Puerto Rico. Toxicon 26:
795-801.

VAN NESS, J., LAEMMLI, U.K. & PETTUOHN, D.E.
1984. Immunization in vitro and production of
monoclonal antibodies specific to insoluble and
weakly immungenic proteins, Proceedings of the
National Academy of Science 81: 7897-7901.

VAITUKAITIS, J.L. 1981. Production of antisera with
small doses of immunogen: multiple intradermal
injections, Methods in Enzymology 73: 46-52.

VERNOUX, J.P., LAHLOU, N., ABBAD EL AN-
DALOUSSI, S., RIYECHE, N. & MAGRAS,
L.Ph. 1985. A study of the distribution of
ciguatoxin in individual Caribbean fish, Acta
Tropica 42: 225-233.

IMPACT OF A VALIDATED, COST EFFECTIVE SCREEN FOR CIGUATERIC FISH
RICHARD J. LEWIS

Lewis, RJ. 1994 O8 G1; Impact of a validated, cost effective screen for ciguateric fish.
Memoitrs of the Queensland Museum 34(3): 349-553. Brisbane. ISSN 0079-8835.

Cigualoxins contaminating ciguateric fish may be detected by a range of in vive (e.g. mouse,
cat, Masquito or chicken), in vitro(ELISA, atria) and chemical assays. Current research seeks
a selective screen to detect low levels of ciguatoxin-1 (0.05-5.0ppb) in fish flesh or in an
easy-lo-prepare extract, This review summarises requirements fora validated, cost effective
screen for ciguateric fish, Implementation of such a screen will reduce adverse health effects.
An attendant benefit will be the improved marketability of reef fish.

Richard J. Lewis, Southern Fisheries Centre, Queensland Department of Primary Industries,
PO Bex 76, Deception Bay, Queensland 4508; 22 November, 1993.

One important goal of present-day ciguatera
research is the development of a cost-effective
screen for the toxins contaminating ciguateric
fish (Lewis, 1993). The range of toxins involved
in ciguatera has been the subject of some debate,
It appears that only the ciguatoxins and analogues
(e.g. gambiertoxins) are involved (Murata et al.,
1990; Legrand et al.,1992; Lewis et al.,1991).
These toxins are closely related in structure and
all activate voltage-dependent sodium channels.
Of these, CTX-1 typically contributes ~90% of
the toxicity of ciguateric camivorous fish in the
Pacific (Legrand et al.,1992: Lewis & Sellin,
1992) and should be considered the primary tar-
get of a screen for ciguateric fish. Other
ciguatoxin analogues and toxins such as okadaic
acid, maitotoxins or Trichedesmiun: toxins are
likely to play a minor role in human illness.
However, these lower potency toxins could inter-
fere with the response of a screen. The challenge
for researchers is to develop a method that can
rapidly and selectively screen CTX-1 which is
present between 0.1 and Sppb in the flesh of fish
that cause ciguatera (Lewis, 1992). Using a 2-fold
risk factor to ensure public health is protected
necessitates that the screen be capable of reliably
detecting CTX-1 in fish flesh at 0.05ppb (SOppt)
and aboye.

A number of assays have been used to detect
ciguatoxin in fish. These include a range of iv
vivo assays (e.g. mouse, cat, chicken, mosquito:
Banner et al.,1960: Kimura et al.,1982; Lewis &
Endean, 1984, Bagnis et al.,1985; Vernoux etal...
1985), a number of in vitro assays utilising anti-
bodies (Hokama,1991) or isolated tissues
(Kimuraet al..1982) and chemical assays involv-
ing derivatisation and HPLC separation with
fluorescence detection (Legrand et al.,1992;
Dickey et al.,1992). Biosensor-lype assays are

also under development and this approach holds
much promise for the detection of ciguatenc fish.
These assays remain to be validated for their
ability lo selectively detect ciguatoxins in crude
extracts of fish. We validated the mouse assay for
ciguatoxin in up to 20mg of ether extract (Lewis
& Sellin, 1993). This assay was only able tu detect
CTX-1 at >0.5pph in fish flesh. Importantly, this
study found that 63+ 14% of spiked ciguatoxin
was recovered using a standard extraction proce-
dure, thereby establishing its suitability for con-
firming whether a fish sample was toxic or not
The cost of such an assay, as well as its insuffi-
cient sensitivity and ethical considerations,
preclude the use of the mouse assay for routine
seafood monitoring programmes. This paper
summarises requirements fora validated, cost-cf-
feclive screen for ciguateric fish and discusses
some of the impacts of such a test on fisheries and
fisheries products,

FEATURES OF A USEFUL SCREEN

A useful screen for detecting ciguateric fish
needs to passess the following features: 1, simple
implementation; 2, ready availability and long
shelf life; 3, toxin selectivity proportional to the
human oral potency of the ciguatoxin analogues
{ie highest affinity for CTX-1); 4, it should yield
a linear response over the range of toxin con-
centrations encountered in ciguateric fish; 5, av-
ceptable accuracy (+20%) at the level of 0.05ppb
CTX-1 that is independent of the fish tested; 6,
high recovery of CTX-] (>60%) during extrac-
tion and clean-up (>30% in exceptional cir-
cumstances); and 7, total cost of screen must be
acceptable to the consumer.

Prototype stick tests (Hokama,1985; Hokama
et al.,1985; Hokama,1991) have many of the

SSA)

above features, but this or modified versions of
the test are still not readily available. Published
results of screening of fish suggest (but do not
prove) that these tests have sufficient sensitivity
(ie, few false positive results reported). How-
ever. the apparent high frequency of false positive
results (i.e, non-toxic fish rejected) suggests that
the antibody employed may havea relatively low
specificity for CTX-1-

A screen that utilises high affinity binding of
ciguatoxin to a cheap proteim substrate (e.g. IgG)
and couples this interaction to a simply measured
response (e.g, a colour change) has a high likeli-
hood of achieving an adequate compromise be-
tween accuracy and cost. Such antibody-based
screens have the potential to detect bevels of
analyte as low as 1r'*M in food (Gazzaz et al.
1992), However, matnx effects associated with
the type of sample screened can often dramatical-
ly reduce assay sensitivity. The extent of such
matnx effects can also vary with the solubility
characteristics of the analyte. For solid food
matrices, enzyme linked immunosorbent assays
(ELISA) detected okadaic acid in shellfish of
10-300ppb (DSP-check, Ube Industries, Ltd,
Tokyo) and detected aflatoxins above l0ppb
(Domer & Cole, 1989; Trucksess et al.,1989) or
more recently above 1 ppb in a range of solid
foods (Agri-Screen test for aflatoxins, Neogen
Corporation, Michigan). The challenge is to
develop a rapid screen for ciguatoxins that has
one fo two orders of magnitude greater sensitivity
than presently available ELISA assays. A variety
of approaches may be used to improve the sen-
sitivity of antibody-based assays: (1) production
of higher affinity antibodies (ii) optimising assay
conditions in which the antibody is used {iii)
amplifying the assay signal (not always accom-
panied by improved signal to noise) (iv) optimis-
ing sample eXtraction and clean-up (can add 4
significant cost). The potential of the latter ap-
proach is indicated for ciguatoxin which can be
concentrated from the levels in flesh of 0.1-Sppb
to levels of ~30-5,000ppb in a cnide lipid extract
with a two-step clean-up procedure that has 63%
efficiency (Lewis & Sellin.1993),

Toxins responsible for ciguatera anse through
the biotransformation of a number of gambier-
toxins produced by Gambierdiscus foxicus
(Murata et al,,1990; Holmes et al..1991; Legrand
ct al.,1992). Soa range of low potency forms of
the ciguatoxins and gambiertoxins (including
CTX-2 and CTX-3) could accumulate in fish.
These low potency forms may be detected with
antibodies raised to CTX-1 and if so they could

MEMOIRS OF THE QUEENSLAND MUSEUM

give rise to false positive results.. To-date the
cross-reactivity of the antibodies in use for the
various ciguatoxin analogues has not been estab-
lished,

An acceptable cost for ciguatera screening has
not been determined. This will relate to the added
value screened fish will attract in the market-
place. An add-on cost of less than 10% of the
value of the product may be reasonable, The
“cost” of a screening programme should incor-
porate an estimate of the cost of discarding non-
toxic fish as a result of false positive results. Il
may be possible to reduce the cost of screening
by pooling fish samples prior to screening. Such
an approach requires a highly sensitive screen but
could work where ciguatera is a low level prob-
lem (e.g. Australia). This approach could not
work where non-toxic fish have levels of
ciguatoxin within an order of magnitude of levels
that cause human poisoning.

A blind screening of ciguateric fish from
Australia (specimens confirmed ciguateric by
human and mice assays) with a prototype of the
Ciguatect™ test kit (November,1991) revealed
that there were few, if any. false positive results.
However, there was a strong possibility that some
ciguatoxic samples went undetected by the test
In fact five of six of the ciguateric fish samples
are likely to have given a false negative result.
Further testing of these fish using later modifica-
tions of the Ciguatect™ test are not reported and
no satisfactory explanation has been given for the
conflicting results obtained, These results were
obtained at a ime when the test was being con-
sidered for commercial release. Since this time
modifications ta the test have been made but a
final format for the Ciguatect™ test remains
clusive.

TYPES OF ANTIBODY ASSAYS
AVAILABLE

Several approaches are available for incor-
perating antibodies (or similar proteins) into an
assay format for detecting haptens such as
ciguatoxin (Fig.1), In all cases a label, be it a
radioisotope, an enzyme or a luminescent or
fluorescent probe, is used to detect the targeted
compound. Each approach has strengths and
weaknesses but all require a high affinity an-
tibody that is selective and specific for the tar-
geted hapten. The first assay considered is the
indirect hapten assay which requires that the tar-
geted hapten (e.g. ciguatoxin) is first non-selec-
tively immobilised to a solid support (along with

SCREEN FOR CIGUATERIC FISH

INDIRECT HAPTEN ASSAY

non-selective adsorption of haptens to the solid phase

SANDWICH ASSAY

Q
5

unlabelled antibodies (IgG] selctively fixed to the solid phase

DIRECT COMPETITIVE ASSAY

unlabelled antibodies (IgG) selctively fixed to the solid phase

FIG.1. Antibody assays available for the detection of
haptens such as ciguatoxin (CTX). Either the IgG or
hapten are labelled (L) with probe to indicate the
presence of the targeted hapten. Other shapes repre-
sent the range of compounds present along with the
targeted hapten. The IgG could be replaced with other
proteins possessing a high affinity for the targeted
hapten (eg the sodium channel could be used for
detecting ciguatoxin). The large box in each diagram
surrounds the interaction responding to the presence
of targeted hapten, while interactions beyond this box
result in a reduced (false negative) or enhanced (false
positive) assay response that is unrelated to the
presence of the targeted hapten.

numerous other contaminants) prior to its detec-
tion with a labelled antibody specific for the part
of the hapten left exposed following binding. In
practice, it is often not possible to obtain an
antibody that can still ‘see’ small haptens bound
to a solid phase, but this approach can allow rapid
detection of a hapten without a time-consuming

551

extraction step. Tests developed for ciguatoxins
by Hokama (1991) and more recently by Hawaii
Chemtect (Pasadena) have used this approach in
the development of a screen for ciguateric fish.

The second approach is to develop a sandwich
assay. This requires development of two an-
tibodies which can mutually bind the targeted
hapten. This assay can be more selective than the
direct assay although binding reactions in two-
site immunoassays are complex and not easily
predicted (Boscato et al.,1989). This assay is
limited by the difficulties associated with obtain-
ing two antibodies which don’t interfere with
each others binding to a small hapten such as
ciguatoxin.

The third approach is the direct competitive
assay which requires that a second hapten is
available that competes with the targeted hapten
for binding to an antibody. With this approach
either the hapten or the antibody can be fixed to
the solid phase. Unlike the first two assays, the
response 1s inversely proportional to the targeted
hapten, but such a response is likely to be more
specific for the targeted hapten than the former
assays. The direct competitive assay likely holds
most promise for an accurate screen but requires
a competing hapten to be found. While it is pos-
sible that the sandwich and competitive assays
could be developed for the direct sampling of fish
flesh, it is likely is that extraction and clean-up
steps will be required. The solubility require-
ments of the hapten and stability of antibody in
solvent need to be considered if the assay is
designed to detect solubilised hapten.

SCREEN VALIDATION

A number of criteria need to be met before a
screen for ciguateric fish can be considered use-
ful, including: 1, the screen must detect all fish
samples confirmed ciguateric following human
consumption; 2, the screen must have an accept-
ably low rate of false positives (i.e. a false posi-
tive rate within an order of magnitude of the
reported ciguatera incidence for the species
tested) and the cost of false positive results must
be included in the cost of screening; 3, the screen
must detect spiked CTX-1 in crude fish extracts
at levels occurring naturally and should detect
CTX-1 spiked in a fish flesh homogenate; 4, the
screen should produce appropriately accurate
results for toxic and non-toxic fish both within
and between laboratories. Wherever appropriate,
negative controls should be run in a pair-wise
design and results for these should be negative.

552

Absence of readily available reference toxins
of the ciguatoxin class and lack of a validated
analytical method to quantify the level and types
of toxins present in samples of fish hinder at-
tempts to validate screening tests for ciguateric
fish. The use of the mouse bioassay to assess
ciguatoxin levels (Lewis & Sellin in press) is
presently the best available alternative to an in
vitro approach. It could prove misleading to use
other in vivo bioassays to validate screening tests
at this time, especially the unreliable brine shnmp
assay (RJ. Lewis unpubl. data) which has recent-
ly been used by DLL. Park to characterise the
toxins present in fish screened by the S-PIA ver-
sion of the Ciguatect™ test (Hungerford,1993),

Toensure reliability of a screen, any limitations
of a sercen with regard to species, sample
preparation and storage etc should be well docu-
mented. Ideally, the screen should ‘work’ for all
potentially ciguateric fish, irrespective of how
they are caught or handled prior to sale. Accuracy
must be regularly evaluated with reference to
toxic and non-toxic specimens to ensure
reliability over time. A sound basis needs to be
established for any variation in methodology be-
tween control and test procedures,

IMPACT OF A SCREEN FOR
CIGUATERIC FISH

Implementation of a useful screen will result in
improved marketability of seafood captured in
ciguatera endemic areas. Removal of toxic fish
before consumption will lead to improved com-
munity health standards. With the availability of
a screen come possibilities for opening fisheries
for species which are presently restricted because
of ciguatera, In Queensland new and potentially
lucrative fisheries for red bass and perhaps
chinaman fish and paddletai] could be established
once an effective screen is available,

Screening for ciguatera could be conducted at
anumbera levels in the chain of marketing of fish
that includes fisherpersons, wholesalers, com-
mercial companies, government agencies or con-
sumer, Problems are likely to exist for the
consumet seeking compensation to prove thal the
test was jndeed performed on the fish involved in
the poisoning according to the manufacturers in-
structions. It may even be necessary to show that
the toxin involyed was indeed a ciguatoxin. Who
ends up conducting the test will depend on the
final format of the test, government requirements
and the level of risk of ciguatera associated with
the fish being screened. Cost of screening will

MEMOIRS OF THE QUEENSLAND MUSEUM

increase if more than one potnt of Lesting Is re-
quired. It is interesting to speculate on the fate of
a screening product that is shown to have failed
to detect a toxic fish, especially if the fish results
in a poisoning episode.

Another issue to be considered ts what fish are
to be screened. In some areas it might be ap-
propriate to screen all potentially ciguateric reef
fish, whereas in other areas only the high risk
species presently marketed may need Ic be
screened. Other classes of fish that may require a
differential approach include (1) fish presently
banned as a result of the threat of ciguatera they
pose {ii} fish destined for export and (ii) large
whole fish in one of the above categories.

LITERATURE CITED

BAGNIS, R., CHANTEAU, S.. CHUNGUE, E.,
DROLLET, J.H., LECHAT, L, LEGRAND,
A.M, POMPON, A., PRIBUR, C., ROUX, J. &
TETARIA, C. 1985, Comparison of the cat bioas-
say, (he mouse bioassay and the mosquito bpoas-
say to detect cigvatoxicity in fish. Pp. 491-49%. In
Gabrie, C. & Salvat, B. (eds), ‘Proceedings of the
Fifth International Coral Reef Congress, Tahiti,
Vol 4'. (Antenne Museum-Ephe: Moorea).

BANNER, A.H., SCHEUER, PJ., SASAKI, S.,
HELFRICH, P. & ALENDER, C.B. 1960. Obser-
vations on ciguatera-type toxin in fish. Annals
New York Academy of Science 90; 770-787.

BOSCATO, L.M., EGAN, G.M. & STUART, MC.
1989, Specificity of two-site immunoassays. Jour-
nal of Immunological Methods 117; 221-229,

DICKEY, R.W., BENCSATH, F.A., GRANADE, H.R.
& LEWIS, R.J, 1992, Liquid chromatographic-
mass Spectrometric methods for the determination
of marine polyether toxins. Bulletin de la Société
de Pathologie Bxotique £5: 514-515.

DORNER, J.W, & COLE, R.J. 1989. Comparison of
two ELISA screening tests with liquid chromatog-
raphy for determination of aflatoxins in raw
peanuts, Journal Association Official Analytical
Chemists 72: 962-964,

GAZZAZ, S.S., RASCO, B.A, & DONG, FM, 1992.
Application of immunochemical assays to food
analysis, Critical Reviews in Food Science and
Nutrition 32: 197-229.

HOKAMA, Y. 1985. A rapid, simplified enzyme im-
Mmunoassay stick test for the detection of
ciguatoxin and related polyethers fram lish tis-
sues. Toxicon 23: 939-946.

HOKAMA, Y. 1991. Immunological analysis of low
molecular weight marine toxins, Journal Toxicol-
ogy - Toxin Reviews 10; 1-35.

HOKAMA, Y_, OSUGI, A.M.. HONDA, 5.4.4,
MATSUO, MLK. 1985. Monoclonal antibodies in
the detection of ciguatoxin and other toxic
polyethers in fish tissues by a rapid poke stack test.

SCREEN FOR CIGUATERIC FISH

Pp. 449-455. In Gabrie, C. & Salvat, B., (eds),
‘Proceedings of the Fifth International Coral Reef
Congress, Tahiti, Vol 4°, (Antenne Museum-
Ephe: Moorea).

HOLMES, M.J., LEWIS, R. J., POLI, M.A. & GIL-
LESPIE, N.C. 1991. Strain dependent production
of ciguatoxin precursors (gambiertoxins) by Gam-
bierdiscus toxicus (Dynophyceae) in culture.
Toxicon 29: 761-775.

HUNGERFORD, J.M. 1993. Seafood toxins and
seafood products, Journal AOAC Intemational
76: 120-130.

KIMURA, L.H., HOKAMA, Y., ABAD, M.A.,
OYAMA, M. & MIYAHARA, J.T. 1982. Com-
parison of three different assays for the assessment
of ciguatoxin in fish tissues: radioimmunoassay,
mouse bioassay and in vitro guinea pig atrium
assay. Toxicon 20: 907-912.

LEGRAND, A-M., FUKUI, M., CRUCHET, P.,
ISHIBASHI, Y. & YASUMOTO, T. 1992. Char-
acterization of ciguatoxins from different fish
species and wild Gambierdiscus toxicus. Pp. 25—
32. In Tosteson, T.R., (ed.), ‘ Proceedings Third
International Conference on Ciguatera Fish
Poisoning, Puerto Rico’. (Polyscience Publica-
tions: Québec).

LEWIS, R.J. 1992. Ciguatoxins are potent ich-
thyotoxins. Toxicon 30: 207-211.

LEWIS, R.J. 1992. Socioeconomic impacts and
management of ciguatera in the Pacific. Bulletin
de la Société de Pathologie Exotique 85; 427-434,

553

LEWIS, R.J. & ENDEAN, R. 1984, Ciguatoxins from
the flesh and viscera of the barracuda, Sphyraena
jello, Toxicon 22: 805-810.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fishes. Toxicon 30; 915-919.
LEWIS, R.J. & SELLIN, M. in press. Recovery of

ciguatoxin from fish flesh. Toxicon.

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M, 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycedontis javanicus,
Muraenidae). Toxicon 29: 1115-1127.

MURATA, M., LEGRAND, A.M., ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386.

TRUCKNESS, M.W., STACK, M.E., NESHEIM, S.,
PARK, D.L., & POHLAND, A_E. 1989. Enzyme-
linked immunosorbent assay of aflatoxins B}, Ba,
and G) in corn, cottonseed, peanuts, peanut butter,
and poultry feed: a collaborative study. Journal o
f the Association Official Analytical Chemists 72:
957-962.

VERNOUX, J.P., LAHLOU, N., MAGRAS, L.Ph. &
GREAUX, J.B. 1985. Chick feeding test: a simple
system to detect ciguatoxin. Acta Tropica 42:
235-240.

554

THE CHANGING FACE OF CIGUATERA
PREVALENCE. Memoirs of the Queensland Museum 34(3)
554. 1994:- Ciguatera cases in Queensland (recorded mostly
by the Queensland Departmentof Health) between 1965-1992
are compiled into a database of 920 cases attributable to 343
outbreaks, Pelagic fish, mainly mackeral, account for 65% of
all recorded cases while reef fish account for 35% of cases.
Pelagic fish were found to have a significantly higher
prevalence of 8 of the 27 surveyed symptoms than reef fish,
these being temperature reversal, diarrhoea, nausea, vomiting,
abdominal pain, joint pain, dental pain and ataxia, Northern
fish (=24°S catch location) accounted for 33% of recorded
cases while southern fish accounted for 67% of cases. North-
ern fish were more likely than southern fish to be associated
with a neurological symptom profile (odds ratio=2.0; 95% CI
11.38, 2.82]. Neurological profiles (neurological symptoms
only) accounted for 18.2% of recorded cases. This symptom
profile has become more common over the last decade,

ORAL AND INTRAPERITONEAL ADMINISTRA-
TION STUDIES OF TOXINS DERIVED FROM FISH
TISSUES AND EXTRACTS OF CULTURED G.
TOXICUS IN THE HUMBUG (D. ARUANUS), DAM-
SEL-FISH (P. WARDI) AND THE STRIPEY (L. CAR-
PONOTATUS). Memotrs of the Queensland Museum 34(3):
554, 1994:— Toxin admimstration expenments were designed
to compare effects of ciguatoxin(s) (CTX) and toxin(s) in
extracts of G. toxicus (GDT) between teleost fish, and be-
tween species of teleosts; to quantify bioaccumulation of
toxins in fish skeletal muscle; and to obtain evidence of
bioconversion of GDT to CTX in treated fish.

Based on interpretation of signs and death-times, CTX and
GDT administered i.p. are potent teleost neurotoxins. A com-
parison of dose effect of G. toxicus extract in D. aruanus and
P. wardi shows variable susceptibility to G. foxicus-related
toxins in fish that may be related to trophic niche,

MEMOIRS OF THE QUEENSLAND MUSEUM

reflecting a significant shift in toxic fish consumption from
southern pelagic to both northern and reef fish.

A subset of the 920 cases (N=657) were used to model
temporal and geographical shifts from 1976-1992 in major
responses such as lime to onset of first symptom (ONSET)
and prevalence of a neurological profile. Statistical modelling
included robust regression modelling (generalised additive
modelling) and staustical graphics. Significant and complex
shifts in temporal and spatial prevalence were found. Results
and implications of this modelling are discussed.

M.Y. Chaloupka, Queensland Department of Environment
and Heritage, 160 Anne Street, Brisbane, Queensland 4001;
Richard J. Lewis & M, Sellin, Southern Fisheries Centre,
Queensland Department of Primary Industries, P.O. Box 76,
Deception Bay, Queensland 4508; 1 May 1994,

Feeding and subsequent extraction and quantification of
CTX tn L. carponotatus defined approximate oral effective
dosages and rates of incorporation in skeletal muscle. Feeding
expenments in L. carponotatus indicated that the potency of
GDT is at least half thatof CTX, L. carponatatus, D. aruanus
and P. wardi were unable to bioaccumulate or bioconyert
GDT to CTX under these experimental conditions in quan-
lilies sufficient for detection in the mouse bioassay of residues
derived from the skeletal muscle of experimental fish.

Scott T. Hahn, Michael F. Capra, Centre for Bivlogical
Population Management, Queensland University of Technol-
ogy, GPO Box 2434, Brisbane 4001. Australia & Donald M.
Miller, Southern Mlinois University Schoal of Medicine, IL,
USA; 12 April, 1993.

INVERSE-DETECTED NMR OF CIGUATOXIN: QUATERNARY CARBON
LOCATIONS CONFIRMED IN CTX-1]

RICHARD J, LEWIS AND IAN M. BRERETON

Lewis, R.J. & Brereton, 1.M. 1994 08 01: Inverse-detected NMR of ciguatoxin: quaternary
carbon locations confirmed in CTX-1. Memoirs of the Queensland Museum 34(3), 555-559.

Brisbane, ISSN 0079-8835.

Short-range (HMQC, !Jcu) and long-range (HMBC, 2:4Jcy) 2-dimensional inverse-detected
heteronuclear nuclear magnetic resonance spectra of 0.45mg of ciguatoxin-] are shown.
These spectra provide independent support for the structure proposed for ciguatoxin-1 and
confirm the !3C assignments and the location of the two quaternary carbons in ciguatoxin-1.
The presence of four ether linkages was also confirmed from the HMBC experiment.

Richard J. Lewis, Southern Fisheries Centre, Queensland Department of Primary Industries,
PO Box 76, Deception Bay, Queensland 4508; lan M. Brereton, Centre for Magnetic
Resonance, University of Queensland, St Lucia, Queensland 4072; 22 November, 1993.

A major advance in ciguatera research was the
determination of the structure of ciguatoxin
(Murata et al.,1989,1990). Ciguatoxin-1 (=CTX-
1, Fig.1) is the most toxic ciguatoxin isolated
to-date (Lewis et al.,1991) and is dominant in
ciguateric fish flesh (Lewis & Sellin,1992).CTX-
1 is probably responsible for the clinical
syndrome that follows consumption of ciguateric
fish; especially carnivorous species,

The structure of CTX-1 was proposed on the
basis of one-dimensional 'H NMR and nOe
spectroscopy, two-dimensional homonuclear
scalar coupled spectroscopy (7/1) and mass
spectroscopy (Murata et al.,1989,1990). Short-
range (one bond) inverse-detected heteronuclear
experiments (HMQC) supported the structure
proposed for CTX-1 (Murata et al.,1992). The
absolute stereochemistry of CTX-1 (Fig.1) has
been proposed (Suzuki et al.,1991), Short-range
(HMQC) and long-range (HMBC) inverse-

detected spectra of CTX-1, determined using the
method of Martin & Crouch (1991), provide in-
dependent support for this structure, including
confirmation of the location of the two quaternary
carbons in the molecule.

METHODS

CIGUATOXIN-1 (CTX-1)

CTX-1 was purified in 1991 (Lewis et al.,
1991). NMR experiments were performed on a
0.45mg sample of CTX-1 in 0.45ml of pyridine-
ds (99.96%, Cambridge Isotope Laboratories),

NUCLEAR MAGNETIC RESONANCE (NMR)
SPECTROSCOPY

The formation of heteronuclear multiple quan-
tum coherence between protons and '"C nuclei
provides a powerful tool for molecular structure
determination. Two dimensional experiments

a7
ie

FIG.1.Structure of ciguatoxin-1 (Ri = OH) proposed by Murata etal. (1990). Connectivities confirmed from the

HMBC spectrum are shown bolded.

556

10 4

15 7

20 5

Carbon chemical shift (ppm)

25.54

30 7

T T . T
1.2 1.0
Proton chemical shift (ppm)

70 4 r
75 Ol

80 4

Carbon chemical shift (ppm)

85 4

90 5 5

4.0
Proton chemical shift (ppm)

MEMOIRS OF THE QUEENSLAND MUSEUM

Carbon chemical shift (ppm)

i = + i? a *
3.0 25 2.0 15
Proton chemical shift (ppm)
i pol. 1
120 4 E
D |
125 4 L
z
= | x CQ:
a 130 i:
g tos r
E @ 0 18 t
Ss () Qa
7 f ¥
3 ‘| 7 0 0
5 135 4 rF
© 4 22 0 © 6 6 6
|
140 7 =

6.0
Proton chemical shift (ppm)

5.8

FIG.2. HMQC spectrum of ciguatoxin-1 (CTX-1) at 400 MHz in pyridine-ds (30°C). cu clustered into four
regions shown in detail in panels A to D. Carbon number of the carbons giving rise to cy are labelled according
to the structure of CTX-1 (Fig.1). Data were zero filled to give a matrix consisting of 4 K x 1 K points.

provide extensive proton to carbon connec-
tivities, both direct via one bond scalar coupling
(‘Jcu) with the HMQC experiment and long
range via 2 and 3 bond scalar couplings (7"Jcu)
with the HMBC. The long-range experiment al-
lows correlations across and to quaternary car-
bons (i.e. carbons without attached protons) and
across heteroatoms (e.g. oxygens). Whether

short- or long-range correlations are determined
depends on the evolutionary manipulation of
heteronuclear coherences according to the value
of the respective heteronuclear coupling con-
stants (~140Hz for one bond couplings and 5—10
Hz for 2 and 3 bond couplings). The sensitivity
of these experiments can be maximised if the
generated heteronuclear multiple quantum

NMR OF CIGUATOXIN-L

Chemical shit (ppt)

Chemical shift (ppm)

FIG,3.HOHAHA spectrum of ciguatoxin-1 (CTX-1) at S0OMHz in pyndine-
ds (30°C), Chemical exchange between hydroxyl protons of CTX-1 and

357

scans averaged) over a spectral
width of 4310Hz) and was op-
timised for 8.3Hz couplings.
The two- dimensional
homonuclear Hartman Hahn
experiment (HOHAHA) was
performed according to Bax &
Davis (1985) at SOOMHz on an
AMX-500 at 30°C (246 t
values, each with 2048 points
(88 scans averaged) over a
spectral width of 4505Hz). No
water suppression was used for
these experiments. The NMR
data were processed usin
FTTOOL and spectra apalyied
on a SUN SPARCstation-2
also using this software. For
each 2-D experiment, a gaus-
sian deconvolution with line
broadening was applied to t
data (data zero filled to give 4K
points in P2) and a Hamming
filter applied to t) data (zero-
filled to give 1K points in Fl)
to obtain optimal signal to
noise in each dimension.
Chemical shifts are given in
ppm downfield of the pyridine
resonance ('H at 7.2] ppm and
'3C at 123,5ppm),

H20 was detected as direct and relayed cross-peaks from each hydroxyl

proton to the water resonance at 4.94ppm (mixing time 55msec). Data were

RESULTS

zero filled to give a matrix consisting of 4 K x | K points,

coherences are converted to observable proton
signals, rather than using magnetisation from the
less sensitive heteronucleus, in the so-called “in-
verse” or proton detected correlation experiment.
The one-bond correlation variant of this experi-
ment is known as the inverse-detected HMQC
and the somewhat less sensitive long-range
variant as the inverse-detected heteronuclear
multiple bond correlation (HMBC) experiment.
These procedures were applied to the ciguatoxins
as described below.

The short-range ('Jcu) inverse-detected
(HMQC experiment) two-dimensional NMR
spectrum of CTX-1 was obtained at 4OOMHz on
a Bruker AMX-400 at 30°C (206 t values, each
with 2048 points (512 scans averaged) over a
spectral width of 4000Hz). The long-range
(Jeu) inverse-detected (HMBC) two-dimen-
sional NMR experiment on CTX-] was per-
formed at SOOMHz on a Bruker AMX-500 at
30°C (495 t; values, each with 2048 points (96

The HMQC spectrum of
CTX-1 (Fig.2) detected all ‘cu except those
associated with the proton resonances at 4.86
ppm which were obscured by the H20 resonance
which was not suppressed in this experiment.
This spectrum allowed us to inde-
pendentlydetermine the °C chemical shifts for
each carbon on CTX-1! and the 'H chemical shifts
of the attached proton (Table 1), These data cor-
respond to the assignment of Murataet al. (1992),
with the minor exception of carbon 36 to which
we assign a °C chemical shift of 81.0ppm, as
opposed to 83.69ppm. These data confirm all
methines in CTX-{ and the assignment of the
double bond in the flexible portion of CTX-!
(ring F). In addition, all carbons assigned as at-
tached to oxygen had °C chemical shifts charac-
teristic of such a chemical environment. The “C
assignments were confirmed by careful com-
parison of the Soaps pone of protons ob-
served in the HMQC, HOHAHA and reference

558

20 © CSRS

Oo

aia |
aa/Had

. Cie
eChimbs GEN

CAML
» CML

Carbon chemical shift (ppm)

80 Capitan

100

+ CSYMS3

4.0 a5 3,0 2.5 2.0
Proton chemical shift (ppm)

FIG.4.HMEC spectrum of ciguatoxin- | (CTX~1) at 500 MHz in pyridine-ds

MEMOIRS OF THE QUEENSLAND MUSEUM

ona 1] mM solution). However,
we were able to confirm the
location of four ether linkages
by this experiment (Figs 1,4).

DISCUSSION
CSq/1 @ @ CSUN EN
ests gacsimn| = We obtained HMQC and
eis) 8. «HMBC spectra of a 1mM solu-

tion of CTX-1 in pyridine-ds at
30°C. The HMQC spectrum in-
dependently confirmed the '3C
- assignments of CTX-1 given
by Murata et al. (1992). The
HMBC spectra confirmed the
position of carbons 33 and 52,
the two quaternary carbons
present in CTX-1 (Fig.1). This
spectrum also confirmed the
location of four of the 13 ether
linked rings (Fig. 1). The loca-
tion of these quaternary car-
bons and 12 of the 13 ether
linkages was previously in-
ferred from °C chemical shifts
and one-dimensional nOe ex-
periments (Murata et al,,1990),

The HMBC experiment
detected all carbons two or

CAMNSHO »
q CU/58
Cunisé OF Caa/HSo

cain &

masiisy ©

SUSI 6 C5z/itAo

1.5 1.0

(30°C). 74Jcx are labelled accarding to the proposed structure of CTX-1 three bonds from methyl
(Fig. 1), Data were zero filled to give a matrix consisting of 4 K x 1 K protons but owing to the rela-

points.

1-D 'H NMR spectra. All 'H chemical shifts
assigned from the HMBC experiment overlapped
the chemical shifts assigned from the HOHAHA
experiment also obtained at 30°C (Fig.3). In ad-
dition to scalar coupled connectivities, the
HOHAHA (S5msec mixing time) also detected
the chemical exchange between hydroxy] protons
of CTX-1 and residual H20 in the solvent. Similar
exchange was detected previously for CTX-2
(Lewis et al.,1993). Such signals provide a useful
means of identifying the exchangeable hydroxyl
protons in molecules.

The HMBC spectra of CTX-~1 (Fig.4) includes
2.3 cy for all methyl protons. Fortunately, the two
quaternary carbons in CTX-1 (carbons 33 and 52)
were within three bonds of a methyl, allowing us
to confirm unambiguously the °C chemical shift
(Table 1) and jocation (Fig.1) of these carbons.
Most other long-range couplings expected for the
proposed structure of CTX-] were not detected,
at least in part owing to the small quantity of
CTX-1 available (experiments were performed

tively poor signal to noise of

this experiment (compared to
the HMQC experiment) it was not sufficiently
sensitive to detect many of the *4Jcy couplings.
These couplings may also be either larger or
smaller than 8.3Hz, the coupling size for which
the experiment was optimised. In conclusion, our
data support the structure proposed originally for
CTX-1 by Murata et al. (1989).

ACKNOWLEDGEMENTS

We thank Peter Barron (Bruker, Australia) for
running the HMQC experiment and Ray Norton
and David Doddrell for encouragement and sup-
port.

LITERATURE CITED

BAX A. & DAVIS, D.G, 1985. MLEV-17-based two-
dimensional homonuclear magnetisation transfer
spectroscopy. Journal of Magnetic Resonance 65:
355-360,

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S,, MACLEOD, J,.K. & SHEIL, M.M. 1991.

NMR OF CIGUATOXIN-1

TABLE 1. @C'and 'H NMR chemical shifts for CTX-1
at 30°C (400MHz, pyridine-ds) from the HMQC ex-

dic 2
ey ors m) ecw ru ate oom
aa | 40.0 | 2.52,2.52
Se ee 14.5 | 4.16
6.36 PE Be 2
| 4 | 1310 | 636 | 34 | 808 3.31
7.00 | 4.86 | 35 | 364 |1.92,2.26|
| 6 | 1360 | 5.91 — 81.0 334
7 1268 | 578 [3 73.1 3.50
| a | 34.6 [2.54,2.73| 38 | 468 [1541.84]
eae ae bent a
jo {as 375 [90 | a1] 1.71, 2.03
Pu | 74a [| aro | at | ais | 3.21
fiz [ s20 | 343 | a2 | s40 | 335
1 13-| 736 | 3.34 | 43 | 414 [1.78,2.59
14 | 37.5 |1.85,2.56| 44 | 74.8 | 4.47
15 | 79.7 3.55 | 45 | 87.6 3,20
i6 | sos | 403 | 46 | 440 | 2:59
7 | 1336 | 574 | 47 | 770 | 421
13 | 131] 539 | 48 | 726 4.06 |
1 | 333 407 | 49 | 779 3.96
20 | 836 | 421 | so | 390 | 201
2 | 1323 | 567 | 51 | 422 1.68
22 | 1359 | 604 | 52 | 1097 | -
as7_| 402 | 53 | 459 [234,240
24 | 85.0 3.64 | 54 | 7079 | 4.86
25 | 325 |~22,2961 55 | 75.1 [4.18,4.18
26 | 128.2 1.37
27 | 128.2
32.7 ; 132
30 | 954 | 3.63 1,23

2 1H = chemical shifts at 4. an (2 protons) were
obscured by the water resonance (!°C values from
Murata et al.,1992).

* 43C chemical shift values for quaternary carbons from
the HMBC experiment at 30°C (500 MHz, pyridine-
ds).

559

Punfication and characterization of ciguatoxins
from moray eel (Lycodontis javanicus, Murae-
nidae). Toxicon 29: 1115-1127.

LEWIS, R. J. & SELLIN, M. 1992. Multiple
ciguatoxins in the flesh of fish. Toxicon 30: 915—
919

LEWIS, R.J.. NORTON, R.S., BRERETON, LM. &
ECCLES, C,D, 1993, Ciguatoxin-2 is a
diastereomer of ciguatoxin-3, Toxicon 31: 637—
643.

MARTIN, G.E. & CROUCH, R.C. 1991. Inverse-
detected two-dimensional NMR methods: ap-
plications in natural products chemistry, Journal
of Natural Products 54: 1-70,

MURATA, M., LEGRAND, A.M., ISHIBASHI, Y. &
YASUMOTO, T. 1989, Structures of ciguatoxin
and its congener. Journal American Chemical
Society 111: 8929-8931.

MURATA, M., LEGRAND, A.M. ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambterdiscus
toxicus, Journal American Chemical Society 112;
4380-4386,

MURATA, M., LEGRAND, A. M., SCHEURR, P.J. &
YASUMOTO, T, 1992. "C NMR assignments of
ciguatoxin by inverse-detected 2D spectroscopy
and an explanation of NMR signal broadening.
Tetrahedron Letters 33; 525-526.

SUZUKI, T., SATO, O., HIRAMA, M.,
YAMAMOTO, Y.. MURATA, M.,
YASUMOTO, T. & HARADA, N, 1991. Enan-
tioselective synthesis of the AB ring fragment of
gambiertoxin 4B. Implication for the absolute
configuration of gambiertoxin 4B and ciguatoxin,
Tetrahedron Letters 35: 4505-4508.

560

ON THE GLOBAL INCREASE OF HARMFUL ALGAL
BLOOMS. Memoirs of the Queensland Museum 343(3): 560.
1994:— Harmful algal blooms have occurred throughout
recorded history but during the past two decades they have
increased in frequency, intensity, and geographic distribution;
their effects on human health and economics have increased
accordingly. To some extent, this reflects our increased
awareness of toxic species and the enormous expansion in
aquaculture efforts. Evidence is accumulating, however, that
human activities contribute significantly to this increase
through the stimulation of exceptional blooms by cultural
eutrophication (e.g. from domestic, industrial and agricultural

CIGUATOXIN-1 INDUCES SPONTANEOUS SYNAP-
TIC ACTIVITY IN ISOLATED SYMPATHETIC
GANGLIA OF GUINEA PIGS Memoirs of the Queensland
Museum 34(3): 560. 1994:— Anelectrophysiological study has
been undertaken of the actions of purified ciguatoxin-1 (CTX-
1) on the neurones of guinea pig sympathetic ganglia isolated
in vitro, using conventional intracellular microelectrode tech-
niques. Low concentrations of CTX-1 (0.2-0.8nM) applied
even briefly (<15min) via the perfusing solution induced a
dramatic increase in the spontaneous occurrence of excitatory
synaptic potentials (ESPs) which persisted for many hours.
The amount and pattern of activity varied between neurones
and occurred in the absence of any change in passive or active
electrical properties of the neurones themselves. Single
supramaximal preganglionic stimuli evoked a summed
response which was unaltered after exposure to CTX-1, but

MEMOIRS OF THE QUEENSLAND MUSEUM

wastes; acid precipitation, deforestation and increased runoff
from cleared land) and by the spreading of nuisance organisms
in ships’ ballast water. The global distribution of these
phenomena is illustrated with examples from Japan, North
America, Europe, South-East Asia and Australia, and involv-
ing dinoflagellates, diatoms, prymnesiophytes, raphidophytes
and cyanobacteria.

Gustaaf M. Hallegraeff, Department of Plant Science,
University of Tasmania, GPO Box 252 C, Hobart, Tasmania
7001, Australia; 12 April, 1993.

was followed by a variable duration high frequency burst of
ESPs. These bursts resembled those occurring spontaneously
in the same cell, and apparently arose from individual
preganglionic axons. The effects were abolished by reduced
Ca**, w-conotoxin, low doses of TTX or raised divalent cation
concentrations. The results indicate that some preganglionic
axons have CTX-binding sites that open Na* channels causing
spontaneous depolarization and initiating repetitive dischar-
ges.

Paul Hamblin, Department of Physiology & Pharmacology,
University of Queensland, St Lucia, Queensland 4067,
Elspeth M. McLachlan & Richard J. Lewis, Southern
Fisheries Centre, Department of Primary Industries, P.O. Box
76, Deception Bay, Queensland 4508; 12 April, 1993.

INVERTEBRATES IMPLICATED IN THE TRANSFER OF GAMBIERTONINS TO THE

BENTHIC CARNIVORE POMADASYS MACULATUS
RICHARD J. LEWIS, MICHAEL J. HOLMES AND MICHELLE SELLIN

Lewis, R.J., Holmes, M.J. & Sellin, M. 1994 08 01: Invertebrates implicated in the transfer
of gambiertoxins to the benthic carnivore Pomadasys maculatus. Memoirs of the Queensland
Mucewm 34(3): 561-564. Brisbane, ISSN 0079-8835.

The food chain hypothesis for the transfer of ciguatoxins (CTX) to carnivorous fish has
gained widespread acceptance. ‘This study was undertaken to determine the vector(s)
transferring gambiertoxins to the often ciguateric blotched javelin fish (Pemadasys
maculatus) in Platypus Bay, Queensland. P. maculatus is a benthic camivore which in
Platypus Bay was found to feed predominantly on small shrimps and crabs that live amongst
Cladophora sp. that also harbours Gambjierdiscus toxicus. Of the potential prey of P,
maculatus in Platypus Bay, only the shrimps (mostly A/phens sp.) contained detectable levels
of ciguatoxin-like toxins, implicating shrimps as an important vector in the transfer of
gambiertoxins to carnivorous fish, Any toxic effects of G. toxicus on shrimps may facilitate
the selective feeding of fish on shrimps containing the highest toxin levels, Sich selective
feeding provides a mechanism for the funnelling of toxins from G. toxicus to P. maculatus.
It remains to be established if shrimps are capable of biotransforming the gambiertoxins to
ciguatoxins or whether biotransformation of the gambiertoxins is accomplished exclusively
hy fish, Given that P. maculatus is at mes highly toxic, and within a year can be non-toxic,
it is likely that the gambiertoxins enter the food chain as intense bursts that perhaps last for
only several weeks. Depuration and/or detoxification are likely to account for the apparent
rapid loss of gambiertoxins and ciguatoxins from shrimps, crabs and P, maculatus,

Richard J. Lewis, Michael J, Holmes, and Michelle Sellin, Southern Fisheries Centre,
Queensland Department of Primary Industries, PO Box 76, Deception Bay, Queensland

4508; 22 November, 1993,

The food chain hypothesis. for the wansfer of
ciguatoxins (CTXs) to carnivorous fish has
gained widespread acceptance through the results
of numerous studies (Randall,1958; Yasumolo et
al.,1971,1977a,b,1979; Banner,1974; Murata et
al,,1990; Holmes et al.,1991; Lewis et al,,1991;
1992), Key steps in the food chain hypothesis
include (i) the uptake by herbivorous fish of gam-
biertoxins (GTXs) produced by Gambierdiscus
foxicus and (ii) the transfer of the loxins from
herbivorous to carnivorous fish. Grazing mol-
luscs (Yasumoto & Kanno,!976) and fish that
feed on invertebrates (Banner,!974) have also
been implicated in ciguatera. The involvement of
invertebrates in the ciguatera food chain has been
speculated upon (Kelly et al..1992) following
laboratory observations that brine shrimp were
capable of feeding on G. toxicus. However,
evidence that invertebrates play an important role
in the transfer of CTX or their precursors remains
circumstantial.

Platypus Bay, Queensland, regularly produces
ciguateric fish including the piscivorous Spanish
mackerel (Scomberamorus commersont) and bar-
racuda (Sphyraena jello) (Lewis & Endean,
1983,1984), To reduce the adverse impacts of
ciguatera, a ban has been imposed on capture of

these species in Platypus Bay. Another common
fish, Pomadasys maculatus (blotched jayelin
fish, Fig. [), can also be toxic in this area (Lewis
et al., 1988) and may be a link in the transfer of
CTXs to Spanish mackerel and barracuda (Lewis
& Sellin, 1992), PF. meeulatus are often more toxic
than Spanish mackerel and toxic individuals of
both species are contaminated with CTX-1, -2
and -3 at similar relative levels (Lewis & Scl-
lin, 1992), In this paper we report that P,
maculatus probably accumulates CTX through
feeding on invertebrates (especially the shrimps}
living in the macroscopic algae (Cladophora sp.)
that harbours G. texicus in Platypus Bay.

METHODS

POMADASYS MACULATUS IN PLATYPUS BAY

P. maculatus were captured in Platypus Bay
(24 58'S, 153° 10°E) by line or trawl net in ~15
m of water (Holmes et al,, this memoir). Feeding
preferences were determined from visual assess-
ment of the stomach and intestinal content of P-
maculatus collected intermittently over a yeur (n
= 40). Small beothic fish that could be potential
prey of P. maeulates could not be confirmed by
scuba diver observations in Platypus Bay. The

562

MEMOIRS OF THE QUEENSLAND MUSEUM

FIG.1, Examples of Pomadasys maculatus captured from Platypus Bay, Note the downward deflection of the
mouth that indicates this species is specialised for bottom foraging.

relative intestine length and pH (after dilution
with water) of pooled stomach and intestinal! con-
tents (n=5) were measured to assess the digestive
Strategy.

INVERTEBRATES FROM PLATYPUS BAY

Invertebrates (alpheid shrimps, crabs,
nematodes, polychaetes, gastropods) living in as-
sociation with the green macroalga, Cladophora
sp., carpeting the sandy substrate of the study site
were collected by small dredge, beam trawl or
diver. From 8-10 May 1991 a small dredge was
used to collect several species of benthic
“worms’, in addition to shrimp and crab samples,
The diver-collected (9 May 1988) Cladophora
(9.2kg) was processed exclusively for the small
gastropods present. A beam trawl was used to
obtain (22 October 1991 and 20 March 1992) two
further Cladophora samples of 191 and 154kg
from which additional shrimps and crabs were
collected and extracted for toxins, The visible
invertebrates in each of the above samples were
sorted by hand.

ASSESSMENT OF TOXIN LEVELS IN
INVERTEBRATES IN PLATYPUS BAY
Invertebrates were extracted for ciguatoxin-

like toxins with acetone and the acetone-soluble
material partitioned as previously described
(Lewis et al.,1992). The ether-soluble and
selected butanol-soluble fractions (up to 30mg)
were dried, suspended in Tween 60/saline and
assayed in 20422 mice (Quackenbush strain,
either sex). Signs in mice (n=2) following in-
traperitoneal (i.p.) injection of these fractions
were used to characterise the toxicity of each
fraction (Holmes et al.,1991; Lewis et al.,1991).

RESULTS

Toxicity was detected in the ether-soluble frac-
tion of shrimps but not in the ether-soluble
material of other invertebrates (Table 1). Mouse
bioassay signs induced by the ether-soluble toxin
in the shrimps included severe diarrhoea and
laboured respiration, signs consistent with an in-
jection of a sub-lethal dose of ciguatoxin-like
toxins. Two additional samples of shrimps (178
and 163g) and crabs (78 and 11g), collected by
beam trawl, had levels of gambiertoxins below
the limit of detection of the mouse bioassay. The
butanol-soluble material from the gastropods as
well as from the shrimps and crabs from these
latter two collections were also assayed by mouse

INVERTEBRATE TRANSFER OF GAMBIERTOXINS

TABLE 1. Yield (mg) and toxicity (MU) of ether
extracts of invertebrates collected by dredge from
Platypus Bay, Queensland.

Invertebrate Average
Weight Size
(g) g (range)
Shrimps (mostly “| 425 0.15 (0.02- 30(0.5)
Alpheus sp.) 1,2)
Crabs (mostly 21.9 0.2 (0.04- 36 (0)
Thalamita sp.) 1,7)
Small nematodes 38.1 0.05
Large tube-dwelling | 7.9 0.2
polychaetes

|__Gastropods_| - | ~0.1 |

‘Ciguatoxin-like activity quantified in mouse units
(MU). One MU = 1 LDs0 dose for a 20 g mouse,

bioassay. Toxins resembling the maitotoxins
were detected in the butanol fraction but these
were not characterised further.

P. maculatus in Platypus Bay which ranged
from 30-300g (fork length 13.5-24.0cm) has a
thin walled stomach (pH=7.0) and a relatively
short intestine (10-21lcm) (pH=6.3). The
downward pointing mouth (Fig.1) indicates that
this species is a specialised benthic forager. Iden-
tifiable stomach contents of P. maculatus com-
prised mostly shrimps and crabs with
occasionally some Cladophora and small uniden-
tified fish.

DISCUSSION

P. maculatus has a near neutral pH digestive
system that would be unlikely to provide the
conditions for acid- catalysed spiroisomerisation
of CTX-2 (or the putative 52-epi CTX-1 named
CTX-4) to CTX-2 and CTX-1, respectively. Thus
the CTX-1 and CTX-3 detected in P. maculatus
flesh (Lewis & Sellin,1992) may arise as an ar-
tefact that results from the purification of CTX-2
and -4 on silicic acid supports eluted with acid
solvents such as chloroform.

Of the invertebrates inhabiting Cladophora
beds in Platypus Bay, only shrimps contained
detectable levels ciguatoxin-like toxins (Table 1).
The toxin levels in shrimps declined to levels
below those detectable by mouse bioassay for
two subsequent collections. Shrimps (and per-
haps crabs) may be a vector in the transfer of
gambiertoxins to carnivorous fish. Another pos-
sibility is that P. maculatus accumulates gambier-
toxins from the G. toxicus that is ingested along
with the small amounts of Cladophora ingested

563

incidentally with the invertebrates. This pos-
sibility is considered remote since the prominent
herbivore in the area (Siganus spinus, a species
of similar size to P. maculatus) consumes almost
entirely Cladophora and is seldom toxic (unpubl.
data). Another possibility is that P. maculatus
feeds on the dead remains of ciguateric fish.

For shrimps to accumulate gambiertoxins they
must be capable of ingesting G. toxicus. This
appears likely, since brine shrimp have been
shown to feed on G. toxicus in the laboratory
(Kelly et al.,1992). The detection of ciguatoxin-
like toxins in shrimps and maitotoxin-like toxins
in shrimps and crabs from Platypus Bay indicates
that both groups of invertebrates consume G.
toxicus. Analysis of intestinal content of shrimps
(n = 2) revealed a range of detritus similar to or
larger than G. foxicus, but no G. toxicus, This
analysis was conducted on shrimps collected at a
time when no detectable gambiertoxin could be
extracted from these shrimps.

G. toxicus cells have been shown to be toxic to
brine shrimps (Kelly et al.,1992). Toxic effects of
G. toxicus on shrimps may facilitate the selective
feeding of fish on shrimps containing the highest
toxin levels. Such selective feeding provides a
mechanism for funnelling G. toxicus toxins, espe-
cially gambiertoxins, to P. maculatus (Fig.2). It
remains to be established if shrimps are capable
of biotransforming the gambiertoxins to
ciguatoxins or if this capacity is exclusive to fish.
The piscivorous fish likely to prey on P.
maculatus include Scomberomorus commersoni,
Sphyraena jello and Seriola lalandi (yellow-tail
kingfish).

Environmental and/or genetic factors leading
to a proliferation of gambiertoxins and conse-
quent outbreaks of ciguatera remain to be

environ— . 3
mental Proliferation of
genetic gambiertoxins
Invertebrates
(shrimps)

P. maculatus
(30 - 300 gl
FIG.2, A model for the food chain transfer of
ciguatoxins and/or gambiertoxins to P. maculatus and
piscivorous fish in Platypus Bay. This study has im-
plicated shrimps as a key vector. The size of the fish
(kg) involved in this transfer are indicated.

Gambierdiscus
toxicus

Piscivorous fish
(1.5 -— 20 kg)

564

elucidated. Given that P, maculatus is at times
highly toxic (Lewis & Sellin,1992) and within a
year can be non-toxic, it is likely that the gam-
biertoxins enter the food chain as intense pulses
that perhaps last for only several weeks. At these
times the shrimps would presumably be highly
toxic. Depuration and/or detoxification may ac-
count for the apparent rapid loss of gambiertoxins
and ciguatoxins from shrimps, crabs and P.
maculatus.

ACKNOWLEDGEMENTS

We thank B, Davidson, D. Trama, I. Halliday,
W. Young, A. Wong Hoy, R. Street and D.
Cameron for assistance collecting samples in
Platypus Bay and the Queensland Museum for
assistance identifying the invertebrates.

LITERATURE CITED

BANNER, A.H. 1974. The biological origin and trans-
mission of ciguatoxin, Pp. 15-36. In Humm, H.J.
& Lane, C.E. (eds), ‘Bioactive compounds from
the sea’. (Marcel Dekker: New York).

HOLMES, M.J., LEWIS, R.J., POLI, M.A. & GIL-
LESPIE, N.C. 1991. Strain dependent production
of ciguatoxin precursors (gambiertoxins) by Gam-
bierdiscus toxicus (Dinophyceae) in culture.
Toxicon 29; 761-775,

KELLY, A.M., KOHLER, C.C. & TINDALL, D.R-
1992. Are crustaceans linked to the ciguatera food
chain? Environmental Biology of Fishes 33: 275—
286.

LEWIS, R.J. & ENDEAN, R. 1983. Occurrence of a
ciguatoxin-like substance in the Spanish mackerel
(Scomberomorus comimersoni). Toxicon 21; 19-
24.

LEWIS, RJ. & ENDEAN, R. 1984. Ciguatoxin from
the flesh and viscera of the barracuda, Sphyraena
jello. Toxicon 22: 805-810.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fishes. Toxicon 30: 915-919.
LEWIS, R.J,, CHALOUPKA, M.Y., GILLESPIE, N.C.
& HOLMES, M.J. 1988. An analysis of the human
response to ciguatera in Australia, Pp. 67-72. In
Choat et al. (eds), ‘Proceedings of the Sixth Inter-
national Coral Reef Symposium, Townsville, vol,

MEMOIRS OF THE QUEENSLAND MUSEUM

3’. (6th International Coral Reef Symposium Ex-
ecutive Committee: Townsville).

LEWIS, R.J,, SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae). Toxicon 29: 1115-1127.

LEWIS, RJ., SELLIN, M., STREET, R., HOLMES,
M.J, & GILLESPIE, N.C, 1992, Excretion of
ciguatoxin from moray eels (Muraenidae) of the
central Pacific. Pp. 131-143, In Tosteson, T.R.
(ed.), “Proceedings of the Third Intemational Con-
ference on Ciguatera Fish Poisoning, Puerto
Rico’. (Polyscience Publications; Québec).

MURATA, M,, LEGRAND, A.M. ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386,

RANDALL, J.E. 1958. A review of ciguatera, tropical
fish poisoning, with a tentative explanation of its
cause, Bulletin of Marine Science 8: 236-267,

YASUMOTO, T. & KANNO, K. 1976. Occurrence of
toxins resembling ciguatoxin, scaritoxin, and
maitotoxin in a turban shell, Bulletin of the
Japanese Society of Scientific Fisheries 42: 1399-
1404.

YASUMOTO, T., HASHIMOTO, Y.. BAGNIS, R.,
RANDALL, J.E. & BANNER, A.H, 1971.
Toxicity of the surgeonfishes. Bulletin of the
Japanese Society of Scientific Fisheries 37: 724—
734.

YASUMOTO, T., BAGNIS, R., THEVENIN, S. &
GARCON, M, 1977a, A survey of comparative
toxicity in the food chain of ciguatera, Bulletin of
the Japanese Sociely of Scientific Fisheries 43:
1015-1019.

YASUMOTO, T.,. NAKAJIMA, LL, BAGNIS, R. &
ADACHI, R. 1977b. Finding of a dinoflagellate
as a hkely culpnt of ciguatera. Bullen of the
Japanese Society of Scientific Fisheries 43; 1021-
1026,

YASUMOTO, T., INOUE, A., BAGNIS, R. & GAR-
CON._M. 1979. Ecological survey on adinoflagel-
late possibly responsible for the induction of
ciguatera. Bulletin of the Japanese Society of
Scientific Fisheries 45: 393-399.

CIGUATERA AND HERBIVORES: UPTAKE AND ACCUMULATION OF
CIGUATOXINS IN CTENOCHAETUS STRIATUS ON THE GREAT BARRIER REEF

RICHARD J. LEWIS, MICHELLE SELLIN, {NOEL C. GILLESPIE, MICHAEL J. HOLMES, ANNIE
KEYS, RAEWYN STREET, HEATHER SMYTHE, HAZRA THAGGARD AND SARAH BRYCE

Lewis, R.J., Sellin, M,, 7Gillespie, N.C., Holmes, MJ, Keys. A., Street, R., Smythe, H.,
Thaggard, H., and Bryce, S. 1994 08 01: Ciguatera and herbivores: uptake and accumulation
of ciguatoxins in Ctenochaetus striatus on the Great Barrier Reef. Memolrs of the Queensland
Museum 34(3); 565-570. Brisbane. ISSN 0U79-8835.

Ctenochaetus striatus is a common detritivorous grazer likely to be a key species transfering
ciguatoxin precursors (gambiertoxins) to carnivorous reef fish, Toxins in tissues from ©,
Striatus collected in the Great Barrier Reef were characterised by mouse bioassay and
chromatography. The biodetritus on which it feeds were collected with an airlift suction
apparatus und the toxins present compared with those in C. striatus. Toxins resembling
gambiertoxins and ciguatoxins predominated in all samples. Lesser amounts of {ast acting
and unidentified toxins were also detected. Mailotoxin was not detected, Similar concentra-
tions of the ciguatoxins and gambiertoxins were detected in C. striatus from John Brewer or
Davies Reefs, despite the former having major crown of thorns starfish damage. Toxins
detected in C. striatus from these reefs were below levels thal would result in prey species
becoming ciguateric. This assessment is consistent with the historically low risk of contract-
ing ciguatera from carnivorous fish captured at these reefs, We were unable to detect any of
the ay gambiertoxins in the liver of C. striatus, suggesting that these toxins were
biotransformed to the more polar ciguatoxins (ciguatoxin-1, -2 and/or -3) in the liver of
herbivorous fish, The concentrations of ciguatoxin in the visceral contents of C. striatus were
3- to 6-fold Jower than the levels of such toxins in the biodetritus. perhaps as a result of
bacterial degradation associated with the active fermentation employed as part of the
digestive strategy of this species, Alternatively, the gambienoxins may have been rapidly
assimilated from the intestinal contents of C. siriatus, a feature that may explain the important
role this species apparently plays as a vector for transfer of gambiertoxins (and ciguatoxins)
to carnivorous fish.

Richard J, Lewis, Michelle Sellin, Noel C, Gillespie, Michoe! J. Holmes, Annie Keys, Raewyn
Street, Heather Smythe, Hazra Thaggard & Sarah Bryce, Southern Fisheries Centre,
Queensland Department of Primary Industries, PO Bax 76, Deception Bay, Queensiand

4508: 22 Novertber 1993.

Surgeonfish (Acanthundae) are common ben-
thic herbivores on coral reefs world-wide. Espe-
cially common is Ctenochaetus striatus, a
detritivorous grazer often suggested as a key
species involved in the uptake and transfer of
toxins involved in ciguatera (Randall,1958;
Yasumoto et al., 1971; Banner,1984). C. striatus
feeds by combing biodetritus from turf algae and
coincidentally ingests a variety of toxin produc-
ing benthic dinoflagellates in the process. Of
these dinoflagellates only Gambierdiscus toxicus
has been confirmed to produce toxins which ac-
cumulate in fish and cause ciguatera (Yasumoto
et al., 1977; Murata et al.,1990; Holmes et al,
1991; Lewis et al.,1991; Lewis & Sellin, 1992),

Early studies suggested that G, toxicus
produces one lipid-soluble toxin of similar
poraity to ciguatoxin (Yasumoto et al.,1977);
1wever, recent studies (Murata et al_.1990; Hol-
mes et al.,1991; Holmes & Lewis. 1992; Legrand

et al..1992) determined that G. fexicus produces
several less-polar ciguatoxin precursors named
gambiertoxins (a class of sodium channel ac-
tivator toxins). These gambiertoxins apparently
undergo oxidative metabolism, at some un-
defined point(s) in the food chain, giving rise to
the vanous ciguatoxins including CTX-1, CTX-2
and CTX-3, the principal toxins found in the flesh
and viscera of ciguateri¢ carnivorous fish
(Murata et al.,1990; Lewis et al., 1991; Lewis &
Sellin, 1992; Lewis ef al.,1993). With this new
understanding of the toxins involved in ciguatera,
we have examined reef biodetritus and the vis-
ceral contents, viscera and liver of C. stricates.
Comparison of the toxicity of C. striatus col-
lected from the crown of thors starfish damaged
John Brewer reef and the lightly damaged Davies
Reef allowed an assessment of the impact of such
damage on toxin levels in fish

Si

METHODS

SAMPLE COLLECTION

Adult C. striatus were collected by spear from
back reef areas of John Brewer Reef (18° 38'S,
147° O4'E) and Dayies Reef (188 50°S, 147°
39°E). Specimens were collected in 2-5m dunng
9-10 December (987 from two sites at John
Brewer Reef (n=22 fish ateach site) and from one
site at Davies Reef (n=30 fish). Viscera and liver
were removed soon after capture and visceral
contents (including stomach contents) were
separated from the viscera by carefully stripping
them out. The remaining viscera was sub-
sequently rinsed in seawater to remave any rem-
nants of visceral contents. The visceral contents,
viscera. and liver were separately pooled for each
site, and wet weights determined for each pooled
sample. The samples were initially stored (4
days) in an equal volume of methanol (preserv-
ative) at OC. On return to the laboratory
samples were stored at —20°C prior to extraction.

Biodetritus samples were taken at approx-
imately 2m from John Brewer (site 2) and Davies
Reefs. To mimic the feeding of C. striatus, we
used an airlilt suction apparatus fitted with a
plastic bristled brush and powered by compressed
air from a SCUBA tank. This allowed removal of
biodetritus from the turf algae covering dead
coral surfaces by a combination of scrubbing and
suction actions. The turf algal areps sampled were
typical of areas subject to the major feeding ac-
tivity of C. striaius. Matenal from 0.8m? of turf
covered dead coral was collected into a floating
plankton mesh sock (50m mesh}, The particu-
laie material remaining in the sock was con-
centrated to a small yolume, diluted 1:1 v/v with
methanol and stored as for the C_ striatus
samples,

ISOLATION OF TOXINS

Samples were first freed of the methanol pre-
servative before homogenisation in acetone (3x,
3:1 y/w). The dried extracts were then suspended
in 90% methanol-water and the hexane- soluble
material removed (3x, I:1 v/v) by liquid-liquid
partitioning, The 90% methanol-solubie material
remaining was dried, suspended in water and
extracted with diethyl ether (3x, 1:3 v/v). The
ether-fraction was further separated into cold
acetone-soluble and msoluble matenul following
precipitation at -20°C. The acetone-solubics
were further fractionated on silicic acid columns
(100 mesh, Mallinkrodt using a mumimum of 30
g silica/g extract) eluted with chloroform-

MEMOIRS OF THE QUEENSLAND MUSEUM

methanwd (c:m) mixtures of increasing polarily as
descnbed previously (Lewis et al.,1991). The
water-soluble material from each site was ex-
tracted with n- butanol and pooled before further
fractionation on a silicic acid column (Biosil A,
Biorad) eluted with c:m mixtures as described by
Holmes et al. (1990).

Mouse Bidassay

Fractions were suspended in a 1% Tween
60/0.9% saline solution and bioassayed by in-
traperitonea!l (i.p.) tnjection of Quackenbush
strain mice (20+ 2g, either sex, n=2—5). Signs of
intoxication following injection were recorded la
allow characterisation of the type of toxin present
(ie. ciguatoxin, gambiertoxin, fast acting or un-
determined). To avoid non-specific toxic effects
<30mg of each fraction was injected per mouse,
Fractions were designated as containing CTX-1
if bioassay signs of severe laboured respiration
and loss of activity as well as at least diarrhoea,
hypersalivation or lachrymation. Fractions were
designated as containing GTX or less-polar
ciguatoxins (e.g. CTX-2, CTX-3) if signs of hind-
limb paralysis were observed in addition to the
sign for CTX-1. Fast acting toxins were identified
by injecting doses varying by 2- to 10-fold and
recording time to death. Fast acting toxins typi-
cally caused deaths within an hour at doses ap-
proximating the minimum lethal dose. This
protecol was sufficient to indicate such toxins
had dose vs. time to death relationships clearly
different from those for ciguatoxins, gambier-
toxins or maitotoxins (Holmes et al.,1990.1991:
Lewis et al.,1991,1992). Fractions designated as
containing ciguatoxin or gambiertoxin on the
basis of bioassay signs were quantified from the
time to death relationship: log (dose) = 2.3 log (1
+ T'), where dose is in mouse units (MU) and
time to death (T) is. in hr (Lewis et al.,1992). This
approach allowed quantification and toxin char-
acterisation with a minimum number of mice.
Animal experiments were conducted in accord-
ance with NHMRC animal ethics guidelines.

RESULTS

DIVER OBSERVATIONS ON C, STRIATUS FEEDING
C. striatus is the predominant grazing species
at John Brewer and Davies reefs with most C.
striatus feeding in shallower waters (1—5m), This
herbivore was observed to feed throughout the
day by combing biodetritus from turf algae cover-
ing the exposed dead coral surfaces. C. striatus
was not observed to remove the turf algae in the

CIGUATOXINS IN CYENOCHAETUS STRIATUS

process of removing the biedetritus adhering to
these algae. This was confirmed by a visual as-
sessment of the gut contents.

G. TOXICUS COLLECTIONS AND TOXIN
ANALYSIS

G. foxicus were found attached to numerous
species of macroscopic algae sampled at John
Brewer Reef (Table 1). G, toxicus were also a
conspicuous component of the vacuumed
hiodetritus but the G. toxicus in these samples
were not quantified. The biodetritus upon which
C. striatus feeds and the visceral contents, viscera
and liver of C. striatus were extracted and bioas-
sayed for the presence of toxins. Table 2 indicates
the toxins detected after liquid-liquid partitioning
into hexane, ether-(acetone-soluble and -in-
soluble fractions) and butanol-soluble fractions.
The less polar (hexane-soluble) material did not
contain detectable toxicity, whereas the more
polar fractions were often found to be toxic. The
mouse bioassay detected gambiertoxin- and
ciguatoxin-like toxins and several fast acting
toxins in these more polar fractions.

The toxins in the acetone-soluble ether-fraction
were further characterised by mouse bioassay
after silica gel chromatography (Table 3). This
allowed separation of (i) less polar toxins (GTX-
4b-like) which elute with 97:3 e:m, (ii) toxins of
similar polarity to CTX-1,-2 or -3 or the more
polar gambiertoxins which elute with 9:1 c:m,
and (ii) toxins with polarity similar to the
Maitotoxins which elute with 0:1 c:m (Murata et
al.,1990; Holmes et al.,1990,1991; Lewis et al.,

567

TABLE }. Population densities of Gambierdizcus
fexicus on macroalgae on John Brewer Reef (Seplem-
ber, 1986)

Macroalgae G.toxicus/ 100g | Depth
rs ae ™)

| Spvridiafilamentosa _| pyridia fllamentosa

3 [3]
Padina australis ae 1-5 | 3 |

Balwets opuntia 40 - ae 2-10

Al. turia

1991; Holmes & Lewis,1992), Toxins eluting
with 97:3 or 9:1 c:m induced signs of toxicity
characteristic of the ciguatoxins or gambier-
toxins. The most polar toxins induced either
CTX-1-like signs, signs of undetermined origin
or were fast acting, with the toxicity varying
between the sites and the samples investigated.
The relative concentration of toxins (MU/g) in
the 97:3 and 9:1 c:m fractions at each reef are
compared in Fig. 1.

Toxins in the butanol extracts were also char-
actensed after silicic acid chromatography (Table
4). After silica gel chromatography, toxins were
found only in the c:m 9:1 and O:1 eluates. The fast
acting toxin from the visceral contents appeared
to be less polar than the fast acting toxin in the
detritus. Carry-over of ciguatoxins into the
butanol fraction may explain the toxicity in the
c:m 9:1 eluates. Toxins inducing signs charac-
teristic of the maitotoxins were not detected in the

TABLE 2. Yield (g) and characterisation* of lipid-soluble extracts from Ctenochaetus striatizs and detritus, Great
Barrier Reef.

“John Bre Brewer Reef | John Brewer Reef (1) _| HT John Brewer R Brewer Reel eof} Davies Reet
= Visceral} Viscera| Liver | Detritus | Visceral] Viscera
Content Content
377 pea (200¢ pee dies | (100z) ve (4342 rom oa ) ) | (150g) | (331g)
| Hexane | 529 | 959 | 2.22 | | 814 | 13.26 | 3.98 | 20.97 | 5.40
Acelone- 0.625 | O23h | 3.114 —_ 0.78» 0.426 0. ots | L Ea O.41> 1.164
soluble ether
Acetone- 0.63 0.03> | 0.444 cota 0.12b
insoluble ether
pea | eer pert eta heer 0.92d ¢ | 2.308 |

* Fractions were vaphrat on the basis of solubility and those lethal to mice at = Og/kg were characterised as
indicated by a-c below (n=2).

* Sign(s) in mice typical of ciguatoxin-1 (CTX-1) including lachrymation, hypersalivation and/or diarrhoea.

® Signs in mice typical of less polar ciguatoxins (CTX-2, CTX-3, GTX-4b), include those sign(s) induced by
CTX-1 plus hind-limb paralysis.

* Fast acting toxin.

* Sign(s) of CTX-1 in mice for fractions non-lethal at | .Og/kg.

568 MEMOIRS OF THE QUEENSLAND MUSEUM

o

0.5

o4

LC] Detritus

| Visceral content

[J Detritus

VA Visceral content
ka Viscera
BB Liver

KS] Viscera

= Liver

KKK
Se

o>
we

=
S¢

ae

MU/g (97:3 eluate)
MU/g (9:1 eluate)

as

Ee
was

x
ee

Pe
Ss

coed
aa

Cae

Brewer Brewer Davies Brewer Brewer Davies
(1) {2} (1) (2)

FIG, 1, Concentration of gambiertoxims and ciguatoxins in Ctenochaetus striatus and associated reef biodetritus.
(A) Toxin levels in the 97:3 chloroform-methanol (c:m) eluates (low polarity gambiertoxins). (B) Toxin levels
in the 9:1 c:m eluates (ciguatoxins and high polarity gambiertoxins). Toxin levels are given as total mouse units
(MU)/g of sample. Cold acetone-soluble ether extracts were applied to silicic acid columns in cach case.

butanol fractions, even after silica gel chromatog- indicating that turf algae is a niche in which G.
raphy. toxicus might proliferate.

Several toxins were detected in extracts of the

DISCUSSION turf biodeiritus and C. striatus visceral contents,

viscera and liver. Ciguatoxin(s) and gambier-

C. striatus feed by combing biodetritus from toxin(s) predominated in these samples which

turf algae covering dead coral surfaces of these also contained several fast acting and several

reefs (Purcell & Bellwood,1993), The turf algae unidentified toxins. Interestingly, no maitotoxin

(and other macroalgae) on these reefs were was detected. Concentration of ciguatoxins and

covered with moderate numbers of G, toxicus, gambiertoxins in the tissues of C. striatus from

TABLE 3, Yield [g (MU)] and bioassay signs* of cold acetone-soluble ether-extracts separated by silicic acid

chromatography
(1.12

| | John Brewer Reef (b John Brewer Reef (2)

Fract | Visceral | Viscera| Liver Visceral | Viscera| Liver | Detritus] Visceral | Viscera

ion® | Content Content

cim (0.52) (0.192) | G.072) | (0.069) p 38g) | (1.372) | (0.112) 1.29g) 0.382)
11)8 (0 0) 0) 6)c 0) 0) (14h (oy 14)b
0.04 | oo | 015 | *0.001| 0.07 0.08 | 0.06

(0) (20) (37) (0) (0) (0) (4° (0)

33)¢ (29)b.c (22h (13)8 (32) (18)b¢ (16)

(0) (0) (0) (0) (0) (0) (74

(4e_ | ye | (0) | oy | yt |)

0.21
(0)

estimated to contain 0.SMU),
*© Signs in mice as defined in Table 2.
Signs in mice not clear.
© Fractions eluted from 100 mesh silicic acid with chloroform-methanol (c:m) mixtures of increasing polarity.

CIGUATOXINS IN CTENOCHAETUS STRIATUS

TABLE 4. Yield (g) and bioassay signs* of butanol
extracts® separated by silicic acid chromatography

Fraction | Detritus | Visceral | Viscera Liver
Content
(c:m) (0.8g) (2.5g) (2.0g) | (2.2g)
1:0 0.001 0.01 0.006 0.006
97:3 0.002 0.05 0.02 0.001»
9:1 0,0064 0.74¢ 0.07% 0.074
0:1 0.11¢ 1.30 1.00 1.80

* Fractions lethal to mice at <1.5g/kg were charac-
terised (n=2),

2-4 Signs in mice as defined in Table 3.

* Butanol extracts (g) were pooled for the three sites,
except for the detritus fraction which was from the
Davies Reef collection only.

Fractions eluted from Biosil A silica gel with
chloroform-methanol (c:m) mixtures of increasing
polarity.

John Brewer and Davies Reefs were similar,
despite evidence of major damage at the former
reef as a result of a crown of thorns starfish
infestation. Our sample areas were surveyed in
1985 and had live:dead coral ratios of 1:3 and 2:0
for John Brewer and Davies Reefs, respectively
(The Crown-of-Thorns Study,1985). If the area
of turf algae is a factor that limits C. striatus
density in reef areas, increased areas of dead coral
would allow larger areas of turf algae which could
support higher densities of C. striatus. These
higher herbivore densities might in turn increase
the proportion of such herbivores in the diet of
carnivorous fish. This scenario could result in an
increase in the rate carnivorous fish accumulate
ciguatoxins. Increased areas of turf algae would
also reduce the feeding pressure on turf algae
which could possibly favour higher densities of
G. toxicus on turf alae. Environmental factors, as
yet unidentified, may also increase the levels of
gambiertoxins per unit area of turf algae and such
factors are perhaps more important in increasing
the rate at which gambiertoxins and ciguatoxins
enter the food chain of fish in coral reef areas.
The ciguatoxins in fish are believed to arise
through the biotransformation (oxidative meta-
bolism) of gambiertoxins produced by G. toxicus
(Murata et al.,1990; Holmes et al.,1991; Holmes
& Lewis, 1992). Since low polarity gambiertoxins
were not detected in C. striatus liver, we propose
that the liver is a site, perhaps the major site, for
biotransformation of the gambiertoxins in this
species. Concentrations of ciguatoxin-like toxins
in the biodetritus were considerably less than (3-
to 7-fold) the concentrations found in the visceral
contents of C. striatus. C. striatus does not

569

employ acid digestion (gut was found to have a
pH = 6.4) but fermentation is an important step in
the digestion and assimilation of biodetritus by C.
striatus (H. Choat pers. comm.). Such a reduction
in toxin levels between the biodetritus and the
visceral contents may stem from microbial
degradation of the gambiertoxins to less potent
forms. Alternatively, the uptake of gambiertoxins
from the visceral contents of C. striatus may be
rapid. A rapid uptake of gambiertoxins by fish
(and man) may be a feature common to this class
of polyether toxins.

This study found that levels of gambiertoxins
entering C. striatus were typically higher than
levels in the liver of this species. Consequently
the gambiertoxins and their biotransformed
products (ciguatoxins) do not appear to be ac-
cumulated in a simple, additive manner, suggest-
ing that depuration of ciguatoxins and/or
gambiertoxins may be significant in C. striatus.
Such depuration by herbivores could, at least in
part, contribute to the rapid decline in the
ciguatoxin levels in a population of moray eels
(Lewis et al.,1992) and the rapid decline in
ciguatera incidence in some Pacific Island
countries (Lewis, 1992). A similar conclusion can
be drawn from the study of Bagnis et al. (1985)
who showed that a decline in G. toxicus numbers
paralleled the decline in C. striatus toxicity.

Fish sampled in this study had relatively low
levels of ciguatoxins compared with C. striatus
from French Polynesia (Yasumoto et al.,1971;
Bagnis et al.,1985). To initiate a ciguatera out-
break at these locations on John Brewer and
Davies Reefs, we suggest that orders of mag-
nitude higher gambiertoxin production per unit
area of turf algae are required. Assays with higher
sensitivity and specificity for toxins involved
(e.g. antibody-based assays selective for the dif-
ferent gambiertoxins and their metabolites) or
sites harbouring more toxic fish are likely prereq-
uisites to the further study of toxins in C. striatus.

LITERATURE CITED

BANNER, A.H. 1984. The biological origin and trans-
mission of ciguatoxin. Pp. 15-36. In Humm, H.J.
& Lane, C.E. (eds), ‘Bioactive compounds from
the sea’. (Marcel Dekker: New York).

BAGNIS, R., BENNETT, J., PRIEUR, C. &
LEGRAND, A.M. 1985. The dynamics of three
toxic benthic dinoflagellates and the toxicity of
ciguateric surgeonfish in French Polynesia. Pp.
177-182. In Anderson, D.M., White, A.W. &
Baden, D.G., (eds), ‘Toxic dinoflagellates.’ (El-
sevier: Oxford).

570

HOLMES, M.J. & LEWIS, R.J. 1992. Multiple gam-
biertoxins (ciguatoxin precursors) from an
Australian strain of Gambierdiscus toxicus in cul-
ture. Pp. 520-529. In Gopalakrishnakone, P. &
Tan, C.K. (eds), “Recent advances in toxinology
research, vol.2’. (National University of Sin-
gapore: Singapore),

HOLMES, M.J., LEWIS, RJ. & GILLESPIE, N.C.
1990, Toxicity of Australian and French
Polynesian strains of Gambierdiscus toxicus
(Dinophyceae) grown in culture: characterization
of a new type of maitotoxin. Toxicon 28; 1159-
1172,

HOLMES, M.J., LEWIS, R.J., POLI, M.A. & GIL-
LESPIE, N.C. 1991. Strain dependent production
of ciguatoxin precursors (gambiertoxins) by Gam-
bierdiscus toxicus (Dinophyceae) in culture,
Toxicon 29: 761-775.

LEGRAND, A.-M., FUKUI, M., CRUCHET, P.,
ISHIBASHL, Y. & YASUMOTO, T. 1992. Char-
actenzation of ciguatoxins from different fish
species and wild Gambierdiscus toxicus. Pp_ 25—
32. In Tosteson, T.P. (ed.), ‘Proceedings of the
Third International Conference on Ciguatera Fish
Poisoning, Puerto Rico’. (Polyscience Publica-
tions: Québec).

LEWIS, R.J. 1992. Socioeconomic impacts and
management of ciguatera in the Pacific.Bulletin
de la Société de Pathologie Exotique 85:427—-434.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fishes. Toxicon 30; 915-919,

LEWIS, R.J., SELLIN, M., POLI, M.A., NORTON,
R.S,, MACLEOD, J.K., & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycedontis javanicus,
Muraenidae). Toxicon 29: 1115-1127,

LEWIS, R.J., SELLIN, M., STREET, R., HOLMES,
M.J. & GILLESPIE, N.C. 1992. Excretion of
ciguatoxin from moray eels (Muraenidae) of the

MEMOIRS OF THE QUEENSLAND MUSEUM

central Pacific. Pp. 131-143. In Tosteson, T.R.
({ed.), ‘Proceedings of the Third International Con-
ference on Ciguatera Fish Poisoning, Puerto
Rico’. (Polyscience Publications: Québec),

LEWIS, R.J., NORTON, R.S., BRERETON, IM. &
ECCLES, C.D. 1993. Ciguatoxin-2 is a
diastereomer of ciguatoxin-3. Toxicon 31: 637-
643.

MURATA, M., LEGRAND, A.M. ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin from the moray
ee] Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus, Journal of the American Chemical
Society 112: 4380-4386,

PURCELL, S.W. & BELLWOOD, D.R. in press. A
functional analysis of food procurement in two
surgeofish species Acanthurus nigrofuscus and
Ctenochaetus striatus (Aanthuridae). Environ-
mental Fish Biology.

RANDALL, J.E. 1958. A review of ciguatera, tropical
fish poisoning, with a tentative explanation of its
cause. Bulletin of Marine Science 8: 236-267.

THE CROWN-OF THORNS STUDY 1985. ‘An as-
sessment of the distribution and effects of the
starfish Acanthaster planci (L) on the Great Bar-
rier Reef. 8. Townsville Sector’. (Australian In-
stitute of Marine Science: Townsville),

YASUMOTO, T., HASHIMOTO, Y., BAGNIS, R.,
RANDALL, J.E. & BANNER, A.H. 1971,
Toxicity of the surgeonfishes. Bulletin of the
Japanese Society of Scientific Fisheries 37; 724—
734,

YASUMOTO, T,, BAGNIS, R., THEVENIN, S. &
GARCON, M. 1977. A survey of comparative
toxicity in the food chain of ciguatera. Bulletin of
the Japanese Society of Scientific Fisheries 43:
1015-1019.

CELL BIOASSAY FOR THE DETECTION OF CIGUATOXINS, BREVETOXINS.
AND SAXITOXINS

RON L. MANGER, LINDA S. LEJA, SUE Y. LEE, JAMES M. HUNGERFORD
AND MARLEEN M. WEKELL

Manger, R.L., Leja, L.S., Lee, 5.¥., Hungerford, LM. & Wekell, M.M. 1994 08 (1: Cell
bioassay for the detection of ciguatoxins, brevetoxins, and saxitoxins. Mentoirs of the
Queensland Museum 34(3); 571-575. Brisbane. ISSN 0079-8835.

We have developed an assay using neuroblastoma cells for detection of sodium channel-
specific marine toxins based on an end point determination of mitochondrial dehydrogenase
activity in the presence of veratridine and ouabain. This cell bioassay allows detection of
either sodium channel blocking agents, such as saxitoxins, or sodium channel enhancers such
as brevetoxins and ciguatoxins. The assay responds in a dose dependent manner, differen-
tiates the toxic activity as either sodium channel blocking or enhancing, and is highly
Sensitive, Assay response to brevetoxins and to ciguatoxic extracts is rapid, allowing dose
dependent detection within 4-6hr. The method is simple, utilizes readily available reagents,
uses Substantially less sample than required for mouse bioassay, and is well within the scope
of even modest tissue culture facilities. This cell-based protocol has the potential to serve as
an altemate and complementary method to the standard mouse bioassay,

Ron L. Manger, Linda S§. Leja, Sue Y. Lee, James M. Hungerford, and Marleen M. Wekell,
Seafood Products Research Center, U.S. Food and Drag Administration, P.O. Box 3012,

Bothell, Washingion 98041-3012; 28 March, 1994

Monitoring programs for manne toxins have
depended in large part upon mouse bioassays.
Although mouse assays have for many years
provided a fairly reliable assessment of risk, there
is Mounting pressure to develop alternative
methods to reduce the reliance on animal testing.
Kogure et al. (1988) and Jellett et al. (1992)
developed tissue culture assays in which mouse
neuroblastoma cells are treated with veratridine
and ouabain resulting in altered cell morphology
and decrease in viability. Toxins which block
sodium channels antagonize this effect, rescuing
the cells in a dose dependent manner. Evaluation
is cither through the visual scoring of 200 or more
cells per sample or well, a potentially time con-
suming and operator dependent task, or is de-
pendent upon the physical removal of affected
cells through the cumulative steps of rinsing.
fixing, and staining.

We have deyeloped a cell bioassay for detec-
tion and quantitation of sodium channel activat-
ing toxins such as the brevetoxins and cigua-
toxins. We have modified and simplified the
aboye assays for determination of sodium chan-
nel blockade. Assessment of cytotoxicity in the
present method uses a colorimetric index of me-
tabolic activity. Those cells which are metabo-
lically active reduce a tetrazolium compound,
MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-
diphenyltetrazolium), to a blue-colored formazan
product (Mosmann,1983)). This method requires

minimal processing and the results can be read on
a standard multiwell scanning spectrophotometer
(ELISA plate reader).

MATERIALS AND METHODS

TOXINS

Purified saxitoxin (Calbiochem) was diluted to
the appropriate concentration with complete tis-
sue cuJiure medium prior to cell assay.
Brevetoxins PbTx-] and PbTx-3 (Calbiochem)
were dissolved in methanol and diluted 1;100 in
complete tissue culture medium, from which
serial dilutions in complete medium were made,
A ciguatoxic fish extract (methanol fraction),
prepared from wrasse (Chetlinus rhodechrous)
and mouse broassay data were generously
provided by Dr. Yoshitsugi Hokama, University
of Hawaii. A stock solution of this material was
prepared jn the same manner as for the
brevetoxins. Extracts of crab viscera and mouse
bioassay data were generously made available by
Chery] Eklund and James Bryant, FDA, Bothell,
WA.

MTT BIOASSAY

Cultures were prepared for bioassay as
described by Jellett et al. (1992) with modifica-
tions described below. Neuro-2a cells (ATCC,
CCL131) were grown in RPML 1640 (Sigma)
containing 10% fetal bovine serum (Gibco),

572

>

0.8

0.6

0.4

0.2

ABSORBANCE (570 nm)

0.0
0.10 0.20 0.30 0.40 0,50 0.60
STX (ng)

wo

1.0

8

0.6

0.4

0.2

ABSORBANCE (570 nm)

0.0
0.00 0.10 0.20 0,30 0.40

PSP EXTRACT (dilution)

0.50

FIG.1.Effect of increasing concentration of pure
saxitoxin upon MTT development in the neuroblas-
toma assay (la). Standard MTT development time of
approximately 15min allowed detection in the range
of 0.1ng/10p1 to 0.6ng/10u] saxitoxin addition per
culture well. Aqueous extracts of Dungeness crab
viscera examined for PSP activity in the MTT
neuroblastoma assay (1b). Extracts that had tested at
122ug/100g (@) and no detectable activity (©) using
the mouse bioassay were tested at various dilutions in
the cell bioassay, Circle without error bar is the
average of 2 replicates. The error bar indicates
+SD.Values represent mean of 3-4 replicates.

glutamine (2mM) (Sigma), sodium pyruvate
(1mM) (Sigma), streptomycin SOug/ml (Sigma),
and penicillin 50units/ml (Sigma) (complete
growth medium). Cells were seeded into 96-well
plates at a density of 5 x 10° cells/ml in 200p1
complete growth medium per well. Cultures were
incubated at 37°C/5% CO2 for approximately
24hr.

Culture wells received 101 of sample and 10u1
additions of aqueous stocks of 10mM ouabain
(Sigma) and 1mM veratridine (Sigma) pH 2.
Each sample concentration was tested in replicate
(3 to 5 wells). A minimum of 15 wells per plate
were processed as ouabain/veratridine treated

MEMOIRS OF THE QUEENSLAND MUSEUM

controls (no sample addition), and a minimum of
5 wells served as untreated controls (without
ouabain/veratridine and without sample). In the
case of sodium channel activators, such as the
brevetoxins and ciguatoxic extract, 10u1 samples
were added to replicate culture wells in both the
presence and absence of ouabain and veratridine.
Control wells received added culture medium to
make up for volume differences.

Following incubation with samples, the over-
laying medium was removed from cultures, and
without a wash step, 60ul of a 1:6 dilution of
MTT stock (5mg/ml in PBS, pH 7.4) in complete
growth medium was added to each well. Cultures
were incubated for approximately 15min at 37°C,
medium was then removed, and 100] of DMSO
was added to each well. The plates were immedi-
ately read with an automated multiwell scanning
spectrophotometer using a test wavelength of
570nm and a reference wavelength of 630nm.

RESULTS

Saxitoxin dependent inhibition of the ouabain/
veratridine induced cytotoxicity was measured
directly by alterations in MTT metabolism
(Fig.la). Purified saxitoxin was detected at a
level of 0.1ng/10u1 addition using an approx-
imate MTT development time of 15min. Assay
sensitivity could occasionally be enhanced by
increasing MTT development time to c.45min,
with a resultant detection limit of c.0.02ng/10u1
addition (data not shown). Assay sensitivities
were comparable to that reported by Jellett et al.
(1992). In the absence of ouabain/veratridine
treatment saxitoxin at the concentrations tested
had no measurable effect. For the purpose of
comparison, 0.1Ing/10ul and 0.02ng/10ul
saxitoxin are equivalent to shellfish extracts of
2ug/100¢ tissue and 0.441g/100g respectively.

As a preliminary test of detection of naturally
incurred PSP in samples, acid extracts of Dunge-
ness crab viscera exhibiting positive and negative
PSP activity by the AOAC mouse bioassay
(122yg/100g and none detected/100g respective-
ly) were examined by the MTT neuroblastoma
assay (Fig.1b). The cell bioassay detected mean
values (with standard deviations) of
124+44119/100g in the positive extract (mouse
bioassay) and 33+2,1g/100g in the extract nega-
tive by mouse bioassay. The dose response curves
tended to plateau with increasing concentrations
of extract (dilutions of <1:4), suggesting a com-
peting or potentially interfering cytotoxic com-
ponent.

CELL BIOASSAY FOR CIGUATOXINS, BREVETOXINS AND SAXITOXINS

ABSORBANCE (570 nm) &

0.0 20 4.0 6.0 8.0
PbTx-1 (ma)

10.0

ABSORBANCE (570 nm)

10.0

90 2.0 4.0 6.0 3.0
PbTx-1 (ng)

FIG.2.Brevetoxin PbTx-1 cytotoxicity as measured by
the MTT neuroblastoma assay. (a, b) PbTx-1
cytotoxicity was assayed at 2hr (©), 4hr (#), 6hr (®),
and 1 $hr (A). (b) Cytotoxicity was insignificant in the
absence of ouabain and veratridine at the maximum
incubation time of 18hr (4Q). Values represent mean
of 4 replicates. The error bar indicates +SD.

Brevetoxins and ciguatoxins significantly en-
hance veratridine induced sodium influx in
neuroblastoma cells (Catterall & Risk,1980;
Bidard et al.,1984). We reasoned that this effect
would accelerate the rate of ouabain/veratridine
induced cytotoxicity and could therefore be the
basis of a detection method for sodium channel
activators such as brevetoxins and ciguatoxins. In
the dose range explored, titratable cytotoxicity
was observed as early as 4hr (Fig.2a) and was
total at 18hr. Brevetoxin im the absence of
ouabain/veratridine was not cytotoxic even at the
highest concentration and incubation time tested
(10ng/10ul, 18hr exposure) (Fig.2b). PbTx-3
produced similar results as observed for PbTx-1
in the cell bioassay (data not shown).

A ciguatoxic extract was examined with the
same MTT cell assay format as utilized for
brevetoxins. The extract was diluted and applied

575

to neuroblastoma cells in the presence or absence
of ouabain/veratridine. Within 6hr the sample
produced significant dose dependent cytotoxicity
only in cells treated with ouabain/veratridine
(Fig.3). Even after prolonged exposures of up to
22hr the ciguatoxic extract was not cytotoxic in
the MTT cell bioassay in the absence of
ouabain/veratndine treatment (data not shown).

DISCUSSION

In the present study we sought to develop a
diagnostic cell-based assay for determining either
sodmm channel blocking or enhancing activity.
Furthermore, we explored the possibilities of im-
proving previous methods by simplifying the
end-point assessment of cells treated with sodium
channel blocking agents in the presence of
ouabain and veratridine.

Simplifying the assay was met by incorporating
a sensitive colorimetne test of cellular metabo-
lism, MTT, originally described by Mossman
(1983). The method is a rapid, versatile, quantita-
tive, and simple technique based upon the meta-
bolism of a tetrazolium salt, MTT, by mito-
chondrial dehydrogenase activity in viable cells.
This assay does not require washing or fixation
steps. MTT metabolism has established itself as
an accepted in vitro method in such diverse areas
as the assessment of growth factor activity (Kot-
nik & Fleischmann, 1990), radiosensitivity of cul-
tured cells (Wasserman & Twentyman,1988),
and the evaluation of chemotherapeutic agents
upon target cell lines in culture (Carmichael etal.,

1.00

0.50

ABSORBANCE (570 nm)

2

0.0 10 2,0 3.0 4.0 5.0
CIGUATOXIC EXTRACT (ug)

FIG.3. Ciguatoxic extract from wrasse analyzed by the
MTT neuroblastoma assay. Ciguatoxic extract was
diluted and applied ta cells in the presence of ouabain
and veratridine, Ghr (@) and 22hr (0), or without
ovabain and veratridine, 22hr (A). Values represent
mean of 4 replicates. The error bars indicate +SD.

374

1987; Alley et al.,1988; Mangerct al., 1989). Due
to its many advantages and wide acceptance it
seemed reasonable to explore the utility of MTT
in a modified cell bioassay for the detection of
marine toxins active at the sodium channel, and
had been earlier suggested by us as a potentially
useful approach (Manger, 1993).

An additional goal of our studies was to incor-
porate modifications in the cell-bioassay to allow
detection and quantitation of marine toxins that
activate sodium channels. Insight as to how this
might be accomplished came from Catterall &
Risk (1980) and Bidard ct al. (1984). Their re-
search demonstrated that these toxins enhanced
the **Na influx effect produced by veratridine
treatment in neuroblastoma cells. We reasoned
that this brevetoxin or ciguatoxin activity would
also relate directly to an observable enhancement
of ouabain/veratridine induced cytotoxicity in
our assay. This was observed in our modified
MTT neuroblastoma assay as exhibited by a dose
dependent enhancement of cytotoxicity follow-
ing treatment with etther of these toxins. The lack
of nouceable cytotoxicity in the absence of
ouabain/veratridine ts in agreement with specific
toxin activity Via interaction with sodium chan-
nels.

The MTT cell bioassay was significantly more
sensitive than the mouse bioassay. The animal
assay can detect saxitoxin to a lower limit of
40ug/ 100g tissue (Hungerford & Wekell,1992).
In contrast, the cell bioassay can routinely detect
purified saxitoxin at a Jevel of 0.1ng/10ul, which
is the equivalent of 2ug/ 100g tissue. Occasional-
ly, with extended MTT development time, the
observed limit of detection was 0.02ug/10p)
(0.4ug/100g).

Examination of crab viscera samples with the
MTT cell bioassay demonstrated good correla-
tion with mouse bioassay results. A sample deter-
mined to have 122\1g/100g tissue of PSP by
mouse bioassay resulted in a mean value of
124+44119/100g tissue using the MTT cell biaas-
say. Interestingly, a crab viscera sample that was
PSP negative by mouse bioassay had a mean
value of 33+2n1g2/ 100g tissue in the cell hinassay,
however, this level of saxitoxin js below the
standard detection limit of the animal test
(40p2/100¢). Examination of additional crab vis-
cera extracts by MTT cell bioassay have
produced results in agreement with the mouse
bioassay (data not shown).

The modified MTT cell bioassay is also more
sensitive to the brevetoxins than the mouse bioas-
say, The LDso for mice is 0.01 mg/20g animal, i.p.

MEMOIRS OF THE QUEENSLAND MUSEUM

injection (Hungerford & Wekell,1992), This
would correlate to 0.1 mg/ 100g tissue extract and
would be the equivalent of a Ing/ lO) sample in
the MTT cell bioassay. In the present study the
MTT cell bioassay detected brevetoxins at levels
of 0.25ng/JOpl,

The ciguatoxic extract tested in our studies
produced death in 20g mice following injection
of SOmg in Iml within 2.Shr (Amra et al.,1990).
This represents about 1.8 mouse units of the
equivalent of [5ng CTX-] as estimated by the
method of Legrand et al. (1989). Thus, the
sodium channel activity of this extract was readi-
ly detected in the MTT cell bioassay at levels of
less than 10% mouse units, corresponding to low
or sub pg concentrations of CTX-1.

Mouse bioassays for brevetoxins and cigua-
toxins involye long observation periods, ranging
from several hours to 48hr (Hungerford,1992),
whereas, the MTT cell bioassay can be processed
within 4-6he, The MTT cell bioassay is also well
suited to aulomation, providing a convenient
biological assay that can accommodate a large
number of samples and which can be ac-
complished within one day. Although i vitro
methods cannot presently substitute entirely for
the data derived from animal studies, these
methods do offer the potential to reduce the
reliance upon animal testing and to Facilitate the
rapid screening of test samples. In the event that
mouse bioassays are prohibited or limited by law,
cel] bioassays for these marine toxins may pro-
vide an altemative screening method.

Subsequent to these preliminary studies we
have confirmed the estimated assay sensitivily to
ciguatoxins using purified CTX-] and CTX3C
(Manger et al., in press).

LITERATURE CITED

ALLEY, M.C..SCUDTERO, D.A,, MONKS, A., HUR-
SEY, M.L., CZERWINSKI, M.J., FINE, D.L.,
ABBOTT, B. J., MAYO, J. G. SHOEMAKER,
R.H. & BOYD, MLR. 1988. Feasibility of dig
screening with panels of human tumor cell lines
using 4 microculture tetrazolium assay. Cancer
Research 48; 589-601.

AMRA, H., HOKAMA, Y., ASAHINA, A.Y.,
SHANG, B.S. & MIYAHARA, ).T, 1990,
Ciguatera toxin in Cheilinws rhodechrous (po ou
wrasse), Food and Agncu)tural Immunology. 2:
119-124.

BIDARD, J.N., VIVERBERG, P.M., FRELIN, C,,
CHUNGUE, E., LEGRAND, A.M, BAGNIS, R,
& LAZDUNSKI, M. 1984. Ciguatoxin is a novel
type of Na* channel toxin. Journal of Biological
Chemistry 259: 8353-8357.

CELL BIOASSAY FOR CIGUATOXINS, BREVETOXINS AND SAXITOXINS

CARMICHAEL, J., DEGRAFF, W.G., GAZDAR, A.
F., MINNA, J. D. & MITCHELL, J. B. 1987.
Evaluation of a tetrazolium-based semiautomated
colorimetric assay: assessment of chemosen-
sitivity testing. Cancer Research 47: 936-942.

CATTERALL, W. A. & RISK, M. 1980, Toxins T4¢
from Prychodiscus brevis (formally Gym-
nodinium breve) enhances activation of voltage-
sensitive channels by veratridine. Molecular
Pharmacology 19: 345-348.

HUNGERFORD, J. M. & WEKELL, M. M. 1992.
Analytical methods for marine toxins. Pp.416—
473. In A. T. Tu (ed.), ‘Handbook of natural
toxins, yol.7, Food poisoning’. (Marcel Dekker:
New York).

JELLETT, J.F., MARKS, L.J., STEWART, J.E.,
DOREY, M.L.,WATSON-WRIGHT, W. &
LAWRENCE, J.F. 1992. Paralytic shellfish
poison (saxitoxin family) bioassays: automated
endpoint determination and standardization of the
in vitro tissue culture bioassay, and comparison
with the standard mouse bioassay. Toxicon 30:
1143-1156.

KOGURE, K., TAMPLIN, M.L., SIMIDU, U, & COL-
WELL, R.R. 1988. A tissue culture assay for
tetrodotoxin, saxitoxin and related toxins.
Toxicon 26:191-197.

KOTNIK, K. & FLEISCHMANN, W. R. JR. 1990. A
simple and rapid method to determine
hematopoietic growth factor activity. Journal of
Immunological Methods 129: 23-30.

LEGRAND, A.M., LITAUDON, M., GENTHON,

575

J.N., BAGNIS, R. & YASUMOTO, T. 1989.
Isolation and some properties of ciguatoxin. Jour-
nal of Applied Phycology 1: 183-188.

MANGER, R., COMEZOGLU F. T., WOODLE, D.,
JACKSON, T., PRIEST, J., SINKULE J., MOR-
GAN, A. C. JR. & SIVAM G. 1989. Immunocon-
jugates of ribosomal inhibiting drugs:
comparative potency of trichothecenes and stand-
ard chemotherapeutic agents. Proceedings of the
American Association for Cancer Research 30,
415a.

MANGER, R. 1993. Cell bioassays for seafood toxins.
Journal of the Association of Official Analytical
Chemistry 76, 120-128.

MANGER, R.L., LEJA, L.S., LEE, 8.Y., HUNGER-
FORD, J,M., HOKAMA, Y., DICKEY, B.W.,
GRANADE, H.R., LEWIS, R., YASUMOTO, T.
& WEKELL, M.M. in press. Detection of sodium
channel toxins: directed cytotoxicity assays of
purified ciguatoxins, brevetoxins, saxitoxin, and
seafood extracts. Journal of the Association of
Official Analytical Chemistry.

MOSMANN, T. 1983. Rapid colorimetric assay for
cellular growth and survival: application to
proliferation and cytotoxicity assays. Journal of
Immunological Methods 65; 55-63.

WASSERMAN, T.T. & TWENTYMAN, P. 1988. Use
of a colorimetric microtiter (MTT) assay in deter-
mining the radiosensitivity of cells from murine
solid tumors. International Journal of Radiation
Oncology Biology and Physics 15: 699-702.

576

DETECTION OF CIGUATOXIC FISH BY USING THE
BINDING PROPERTY OF CIGUATOXINS TO VOLT-
AGE-DEPENDANT SODIUM CHANNELS. Memoirs of
the Queensland Museum 34(3) 576. 1994:- Binding studies
indicate that CTX (coded -1B), the principal toxin isolated
from moray eel viscera and CTX-4B (or GT-4B) isolated from
wild dinoflagellate, Gambierdiscua toxicus, competitively
inhibit binding of the brevetoxin (3H)-PbTx-3 to rat brain
membranes. Affinity of CTX-1B is around 30 times higher
than that of PbTx-3 while CTX-4B has around the same
affinity as the brevetoxin. Results confirm that the two toxins
act at the voltage dependant sodium channel of rat brain

EVALUATION OF INTRAVENOUS MANNITOL FOR
TREATMENT OF ACUTE CIGUATERA FISH
POISONING. Memoirs of the Queensland Museum 34(3):
576. 1994:— The Ciguatera Double Blind Study is an inves-
tigator-initiated, grant supported, multicenter, randomized,
controlled trial which is designed to: 1) investigate the ef-
ficacy of intravenous 20% mannitol in comparison to a
placebo (intravenous 5% dextrose in water) for treatment of
acute ciguatera fish poisoning; 2) determine the response time
to treatment; and 3) determine relapse rate 48hrs post treat-

MEMOIRS OF THE QUEENSLAND MUSEUM

membranes. Experiments on minor toxins isolated from
ciguatoxic material are underway. Preliminary results indicate
a common property of the compounds to inhibit the binding
of PbTx- 3. This property is used to evaluate the ciguatoxicity
of hazardous fish. A rapid extraction procedure and a routine
binding assay have been established.

Anne-Marie F. Legrand and Catherine J. Lotte, Institut Ter-
ritorial de Reserches Médicales Louis Malardé, PO Box 30
Papeete, Tahiti, French Polynesia; 1 May 1994.

ment. Mannitol and the 5% dextrose were randomly assigned
to patients who presented with ciguatera fish poisoning to 1
of 4 hospitals. Medical treatment was provided through a
protocol. Patients response was monitored at 10min, 30min
and 2.5hr after therapy was begun. Patient followup was done
for 48hr after the treatment was given.

Neal Palafox, Box 686, John Hopkins School of Public Health,
Maryland U.S.A.; 12 April, 1993.

CONFOCAL LASER SCANNING MICROSCOPY: A NEW TOOL FOR STUDYING
THE EFFECTS OF CIGUATOXIN (CTX-1B) AND D-MANNITOL AT MOTOR NERVE

TERMINALS OF THE NEUROMUSCULAR JUNCTION IN SITU
JORDI MOLGO, PASCAL JUZANS AND ANNE MARIE LEGRAND

Molg6, J., Juzans, P. & Legrand, A.M. 1994 08 01: Confocal laser scanning microscopy; a
new tool for studying the effects of ciguatoxin (CTX-1B) and D-mannitol at motor nerve
terminals of the neuromuscular junction in situ, Memoirs of the Queensland Museum 34(3);
577-585. Brisbane. ISSN 0079-8835,

The confocal laser scanning microscope was used in conjunction with the fluorescent probe
FM1-43 to study the effects of ciguatoxin (CTX-1B) and D-mannitol at motornerve terminals
of the neuromuscular junction in situ, CTX-1B caused time-dependent changes in the surface
urea of motor nerve terminals and perisynaptic Schwann cell at living neuromuscular
junctions. These changes were completely prevented by tetrodotoxin indicating that they are
related to both entry of Nat and increased quantal acetylcholine release, D-mannitol at
concentrations reported to exert an effective hydroxy! radical scavenger action neither
prevented the action of CTX-1B nor antagonized its effects. However, at higher concentra-
tions D-mannitol exerted osmotic effects that caused shrinkage of both motor nerve terminals
and Schwann cell somata previously swollen by the acbon of CTX-1B probably by shifting
water from the intracellular to the extracellular compartment.

Jordi Molgé and Pascal Juzans, Laboratoire de Netirobiologie Cellulaire et Moléculaire,
Centre Natiorul de la Recherche Scientifique, 91198-Gif sur Yvette Cedex, France; Anne
Marie Legrand, Institut Territerial de Recherches Médicales Louis Malardé, Associ€é &@

lPsstiner Pasteur, B. P. 30 Papeete, Tahiti, Polynésie Francaise; 2 February, 1994,

Ciguatoxins (CTX) are a family of potent
lipid-soluble cychc polyethers (Scheuer et al.,
1967; Tachibana et al.,1987: Legrand et al., 1989;
Murata et al.,1989,1990; Lewis et al., 1992;
Lewis & Sellin,]991) involved in ciguatera fish
poisoning (Anderson & Lobel, 1987; Russell &
Egen,]991-; Swift & Swift, 1993). The poisoning
is characterized by severe gastrointestinal and
neurological disturbances (Bagnis et al.,1979:
Withers,1982; Gillespie et al,,1986) which
develop after consumption of coral reef fish.

The chemical strictures (Murata ct al., 1989,
1990; Lewis et al..1991) of structurally related
ciguatoxins (CTX-1B or CTX-1 or CTX, which
is probably the major toxin involved in ciguatera,
CTX-2 and CTX-3) are reminiscent of breve-
toxins, (Baden,1989; Murata et al., 1989, 1990;
Lewis et al.,1991; Gawley et al., 1992) and they
share a common binding site with the brevetoxins
on the neuronal voltage-sensitive sodium channel
proteins (Lombet ¢4 al. 1987; Lewis et al.,1991).

CTX selectively acts on Nat channels in the
node of Ranvier of single myelinated nerve fibers
in such a way that voltage-clamped Na* channel
currents are activated al potentials about 30mV
more negative than unmodified channels and fail
to inactivate during long-lasting depolarizations
{Benoit et al.,1986). It ts likely that a persistent
activation of Na* channels by CTX at the resting

membrane potential leads to a membrane
depolarization and the spontaneous action poten-
tials reported on neuronal, axonal and muscle
membranes (Benoit 1 al.,1986; Bidard et al.,
1984: Molgé et al.,1990), Tetrodotoxin (TTX)
which blocks voltage-gated sodium channels in
thase membranes, prevents such actions (Benoit
ei al,.1986; Bidard et al., 1984; Molgé et al.,
1990), CTX has also been reported to increase
intracellular Ca*+ concentration in NG1O8-15
hybrid cells bathed in standard medium or in a
Ca**+-free medium sopplemented with EGTA
(Molgé et al.,1992a,1993b), CTX-induced Ca**
mobilization prevents further effect of bradykinin
(1j.M) suggesting that CTX also stimulates the
inositol 1,4,5-trisphosphate-releasable Ca** store
(Molga et al.,1993b), Since TTX prevents the
CTX-tnduced increase in intracellular Ca®* con-
centration it would appear that Na? influx through
voltage-gated Na* channels somehow leads to
release of intracellular Ca**. Such a direct
relationship of Na*- dependent Ca** modilization
in neuronal cells is unknown.

CTX increases the rate of release of [*H]-
aminobutyric acid and FH] dopamine from rat
brain synaptosomes. These actions are sensinve
to blockaje by TTX but are unaffected by Ca’*
channe) antagonists like nitrendipine and D-600
(Bidard et al..1984). Since CTX has no action on

578

FIG.1, Low magnification view of motor nerve ter-
minals (NT), Schwann cells (SC) and intramuscular
axons (a) in a living frog cutaneous pectoris neuro-

muscular preparation. Notice the nonmyelinating
Schwann cells covering the branches of the nerve
terminals. Tridimensional reconstitution by a projec-
tion of 30 horizontal section series. The structures
have been stained with the fluorescent membrane dye
FM1-43 for 60min and then washed free of FM1-43.

Na*-K*-ATPase activity, it was suggested that
the enhancement of neurotransmitter release may
be due to a depolarization-induced Ca* influx
(Bidard et al.,1984). CTX was reported also to
enhance Ca?*-dependent ACh release from pure
cholinergic synaptosomes (Molg6 et al.,1992b).
If CTX depolarized synaptosomal membranes to
levels above that needed to activate voltage-gated
Ca** channels, then it would be expected that
membrane depolarization, via Ca’* influx,
would contribute to this Ca**-dependent ACh
release caused by CTX. However, blockade of
Ca”* channel subtypes in Torpedo synaptosomes
(Moulian et al.,1993) by simultaneous applica-
tion of FTX, a toxin purified from Agelenopsis
aperta venom, synthetic omega-conotoxin and
Gd** (Molg6 et al.,1991b) did not prevent ACh
release caused by CTX in the presence of Ca?*
(Molg6 et al.,1993a). These results may suggest
that CTX exerts its effects on ACh release from
Torpedo synaptosomes by increasing synap-
tosomal Na* levels sufficiently to reverse the
Na*/Ca** exchange system which normally uses

MEMOIRS OF THE QUEENSLAND MUSEUM

the Nat gradient to extrude Ca”*. In the reversed
mode the exchanger will import Ca”*.

CTX also increases spontaneous quantal
acetylcholine release at frog neuromuscular junc-
tions even in a nominally Ca’*-free medium sup-
plemented with EGTA (Molg6 et al.,1990).
TTX completely prevented activation of the
release process by CTX suggesting that the CTX
effect depends on Na* entry into the terminal
(Molg6 et al.,1991a). Furthermore, ultrastruc-
tural studies performed at neuromuscular junc-
tions in which quantal transmitter was exhausted
irreversibly by CTX, after 3—4hr of toxin action,
revealed a marked depletion of synaptic vesicles
per nerve terminal cross-section. The depletion
of synaptic vesicles was accompanied by enlar-
gement of the presynaptic membrane coupled to
the swelling of the terminal (Molg6 et al.,1991a;
Comella, Molg6 & Legrand unpubl. results) sug-
gesting that CTX impairs the recycling process
that, under normal conditions, maintains the
synaptic vesicle population during quantal
release.

Experiments described here aim to characterize
some of the basic changes occurring at the
neuromuscular junction in situ during the action
of CTX. For this purpose we have used a
lipophilic dye, that becomes fluorescent only
after incorporation into the outer leaflet of surface
membranes, in conjunction with the recently
evolved confocal laser scanning microscope
which allows optical sectioning of the neuromus-
cular junction at a desired thickness and a sub-
sequent 3-dimensional reconstitution of the
imaged motor nerve terminals.

Confocal laser microscopy appears as one of
the most exciting and valuable techniques for
optical sectioning, high resolution three dimen-
sional imaging and reconstitution of fluores-
cence-labelled or reflecting cellular structures.
This kind of analysis can be done on living nerve-
muscle preparations without the need of physical
sectioning and enables the investigation of
processes, like the time course of action of CTX,
which is not readily studied in fixed preparations.

MATERIAL AND METHODS

Experiments were performed using isolated
cutaneous pectoris nerve-muscle preparations
from adult male frogs, Rana esculenta (20-25g)
between October and April. The excised nerve-
muscle preparation was pinned to the bottom of
a rhodorsil-lined plexiglass chamber (2ml), ex-
posed for 5-60min to the dye (FM1-43,

CONFOCAL LASER SCANNING MICROSCOPY

579

240

160

80

Intensity (arbitrary units)

10 15 20 pm

Intensity (arbitrary unlts)

D

10

15 ye

FIG.2. Images of a neuromuscular junction (A) and of a perisynaptic Schwann cell (C) stained with the dye
FM1-43, In B and D, the intensity of the fluorescence between the lines shown in A and C is indicated. The
peaks of the histograms in B and D (a,b,c,d) correspond to the zones labelled in the images A and C. The images
A and C represent the 3-D reconstitution by a projection of 30 serial sections (0.5j.m steps).

Molecular Probes, Eugene, Or., U.S.A) [N-(3-
triethy! ammonium) propyl]-4-(dibutylamino-
styry] pyridinium, dibromide (241M) dissolved in
standard physiological solution of the following
composition (mM): NaCl, 115.0; KCI, 2.1;
CaCl, 1.8; and N-2-hydroxyethylpiperazine-N’ -
2-ethanesulphonic acid (HEPES), 5 (pH=7.25);
and then washed with the standard physiological
solution. The experiments were carried out at
20°C. Only neuromuscular junctions of surface
fibers were studied. In some experiments excila-
tion contraction of cutaneous pectoris muscles
was uncoupled by treating the preparations with
2M formamide (Sigma, St. Louis, U.S.A) as
previously described (del Castillo & Escalona de
Motta, 1978). In other experiments, D-mannitol
(Sigma, St. Louis, U.S.A) was added to the
standard solution and osmolality was determined

using a Knauer (Berlin, Germany) freezing-point
osmometer, Ciguatoxin (CTX-1B) was extracted
from Gymnothorax javanicus (moray eel) liver
and viscera (Legrand et al.,1989; Murata et al.,
1990). Tetrodotoxin was from Sigma (St. Louis,
US.A.).

Neuromuscular junctions were imaged with a
Sarastro-2000 confocal laser scanning micro-
scope (Molecular Dynamics, California, U.S.A.)
composed of an upwright NIKON optiphot-2
microscope equipped with a single argon-ion
laser beam emmitting light at 488nm (high
power, maximum output 25mW), with a 3%
neutral density transmission filter to prevent dye
bleaching. A510nm dichroic mirror and a 510nm
long pass emission filler were used, The aperture
setting was 50m. The photomultiplier was set at
a constant level ina given expernment (between

580)

CHANGES AFTER CTX ACTION

GM cont ine) 2h ane

NERVE TERMINAL |"

SCHWANN GELL fp —

0 25 50 75 100 125 150 175 200
% CHANGE IN SURFACE AREA

FIG,3. Relative changes caused by 10nM CTX-1b on
the surface area of motor nerve terminals and
Schwann cell somata with respect to controls at dif-
ferent times: of toxin action. The black columns
denote respective controls.

600-900V). Images were acquired with single
scans or after averaging. Neuromuscular junc-
tions were routinely visualized with a 40x water
immersion objective (0,55 numerical aperture),

Control of the scanner module and image
analysis of the data files was achieved with a
Silicon Graphics workstation (Mountain View,
Ca, U.S:A) integrated into the Sarastro system.
Images were analyzed with a Silicon Graphics
Personal Iris 4D/35G workstation using a
UNIX™ operating system and the software
Image Space from Molecular Dynamics. A
series of optical sections were taken at 0.)3-
0.5msteps. Images from each expenment were
processed identically and stored on rewritable
magneto-optical disks. [n all experiments
heuromuscularjunctions were imaged hefore and
after the yarious treatments.

RESULTS AND DISCUSSION.

STAINING OF THE NEUKOMUSCULAR JUNCTION
IN SITU

The fluorescent Staining appears on motor
nerve terminals, on myelinated nerve fibres, and
in perisynaptic Schwann cells somata (Fig,]).
This staining is difficult to wash out after such a
long exposure (G0min) to the dye. The mechan-
jsm of staining scems Lo be due to the high affinity
of the dye for lipid membranes coupled with an
inability to penetrate, so that the dye seems to
partition only into the outer leaflet of surface
membravies (Betz et al,,1992), In contrast to pre-
views Work by Betz et al. (1992), we have found

MEMOIRS OF THE QUEENSLAND MUSEUM

that the FM1-43 dye also stains Jiving motor
nerve terminals in an activity-independent
fashion. Staining in the various membrane struc-
tures was detected on resting preparations ex-
posed for only Smin to the dye and then washed
out, with dye-free medium. This staining lasted
for more than 12hrs.

When the nerve terminal and the Schwann cell
somata were imaged at higher magnification by
a stack of horizontal scans, the image of the 3-D
volume described by the section series (look
through projection) revealed both surface and
internal structures (Fig.2). The intensity of
fluorescence was more marked at the contours
and edges than in the interior of both structures,
Pixel intensity plots of line scans (Fig.2b,d)
showed peaks corresponding to the limits of the
nerve terminal membrane and Schwann cell
somata membrane. The axoplasma of the ter-
minal and the cytoplasma of the Schwann cell
exhibited lower intensity. Having characterized
the dye staining in motor terminals and Schwann
cells, we performed further experiments in order
to determine whether CTX-!B was still active in
enhancing quantal transmitter release after ap-
plication and washout of the dye. Under these
conditions, as in control junctions (see Molgo et
al.,1990), CTX-1B (2.5nM) increased the fre-
quency of miniature endplate potentials (data not
shown), These results indicated first that the
FM 1-43 dye did not perturb the effect of CTX-1B
and second that the dye was suitable for following
eventual changes in the nerve terminal surface
area during the action of CTX-1B

Errect oF CTX-Jb ON Moor NERVE
TERMINALS LN SiTu

In junctions in which muscle contraction was
prevented by pricr weatment with formamide,
stained with the PM143 dye and then washed
out, one of the nerve terminals was selected and,
imaged before and after different mes af CTX-
IB (10nM) addition to the standard medium.
Usually 10 horizontal section images (0.5j2m
step) were made for complete reconstructed
view of the nerve terminals at each time period
investigated. Increases in the nerVe temminal sur-
face area were evident within 15-1 7min of CTX-
1B (100M) addition to the medium, this increase
m surtace area continued for 3hrs. Relative chan-
ees in surface area at 1, 2 and 3hrs of CTX-1B
acuion (Fig.3) were greatest during the first hour
(50+2.0%; n=3) compared with the second and
third hour of CTX-1B exposure. After the second
and third hour nerve terminals only increased

CONFOCAL LASER SCANNING MICROSCOPY

3.420.15 and 5.740.29% respectively wath
respect to the first hour (Fig.3). When junctions
were pretreated with TTX (1M) no such chan-
ges were observed during 3hrs of CTX-1B action.
Thus, the increase in nerve terminal surface area
of motor nerve terminals might be related to both
increase of intraterminal Na~ concentration and
to the enhanced quantal release, None of the 6
nerve terminals imaged during 3hrs with CTX-
1B showed fluorescent spots inside the terminals,
as Observed with high K* medium (Betz. et al.,
1992)_ This result supports the previous view that
CTX-1B blocks the recycling of clear synaptic
vesicles (Molgé et al.,1991a).

EFFECT OF CTX-LB ON PERISYNAPTIC
SCHWANN CELLS IN SITU

Satellite cells of the nervous system, oligo-
dendrocytes and astrocytes in the central neryous
system and Schwann cells in the peripheral nerv-
ous system, have a regulatory role in synaptic
transmission. Thus, glial cells can be depolarized
by K* accumulation near active neurons in situ
(Orkand et al., 1966), can respond to many chemi-
cal transmilters in vitro (Orkand,19§2; Dave et
al., 1991) and express a diversity of ion channels
(Ritchie,]992; Sontheimer,1992). At the frog
neuromuscular junction non-myelinating
Schwann cells cover the motor nerve terminal
and send fine processes around it (Birks et al.,
1960: Dreyer & Peper.1974). The Schwann cell
processes are generally located between the ac-
tive zones at irregular intervals (Couteaux &
Pécot Dechavassine,1970). We tested whether
perisynaptic non-myelinating Schwann cells are
affected, like motor nerve terminals. by CTX-1B
(10nM). Schwaan cells also markedly increased
their surface area during 1, 2 and 3hrs of CTX-1B
action (Fig.3). After 3 hours of CTX-1B action
the increase was more marked (642.9%) than
in motor nerve terminals. The changes of the
Schwann cells lying directly over the nerve ter-
minal were similar to the changes observed in
cells located lateral to the nerve terminals. TTX
(1M) completely prevented such changes when
applied before CTX-1B, Therefore, in addition to
acting on motor nerve terminals CTX-1B also
acts on Schwann cells, Is this related to the action
of the CTX-1B on sodium channels of Schwann
cells orisitthe result of the changes in quantal
acetylcholine release caused by CTX-1B? The
tole of the nonmyelinating perisynaptic Schwann
cells at the frog neuromuscular junction dunng
synaptic activity and particularly during transmit-
ter release remains unknown. Recent studics

581

have shown that motor nerve stimulation in-
duces an increase jn intracellular Ca** concentra-
tion in Schwann cells (Jahromi et al.,1992),
Since Schwann cells undergo profound changes
during the action of CTX-1B it is likely thi they
may play a role in the maintenance of the
neuromuscular junction.

EFFECT OF CTX-LB ON MUSCLE FIBRES IN
SITU

When skeletal muscle fibres in which muscle
contraction was prevented (by formamide treat-
ment) Were imaged at junctional sites before and
after Shr of CTX-1 Baction, the changes observed
in muscle fibre surface area were of the order of
1—1.5% (n=4). Attempts to investigate the ef-
fects of CTX-1B in muscle fibres with functional
excitation-contraction coupling failed due uy the
spontaneous contractions induced by the toxin
which prevented imaging during the first hr of
toxin action.

EFFECTS OF D-MANNITOL AFTER CTX-1B
ACTION

Mannitol was reported to markedly improve
neurologic and muscular dysfunction in patients
with acute ciguatera (Palafox et al.,1988). Al-
ihough these observations were uncontrolled, the
dramatic clinical improvement suggested that
mannitol may have a valuable therapeutic effect
on this intoxication, The mechanism of action of
mannitol is obscure (Russell & Egen,1991). One
possibility that was suggested is that D-mannitol
may neutralize the toxin by some covalent cou-
pling or complexation. Another possibility is that
mannitol may exert osmotic effects by increas-
ing extracellular osmolality. Finally, one should
take into account that mannitol reacts with free
radicals and is considered as an effective
hydroxyl radical scavenger (Halliwell & Gut-
teridge,1985). The possibility that hyperosmotic
D-mannitol may exert its action on ciguatera duc
to its hydroxyl radical-scavenging properties or
its water-draining effect has been suggested by
Peam et al., (1989).

A free radical is an atom or molecule that
contains one Or more unpaired electrons so that
to altain stability either donates its electron to
other molecules or acquires an extra electron
from adjacent molecules, Indeed, free radicals
are highly reactive and, becanse of their in-
stability, damage cells and tissues. D-mannitol
reacts with the very reactive and short-lived
hydroxy! rudical in a way that its concentration
can be limited. Hydroxy! radicals also stimulate

La
oOo
re

RINGER (A)
R+MANNITOL 44% |

A+MANNITOL 2% |

R+MANNITOL 8% |

Osmolality (m Osmol/Kg)

FIG.4. Osmolality of the standard Ringer's solution (R)
and of the various standard solutions (R) to which
D-mannitol was added.

phospholipase Az leading to release of arachi-
donic acid. One interesting property of hydroxy!
radicals ts their ability to initiate lipid peroxida-
lion by extracting a hydrogen atom from polyun-
saturated fatty acids such as arachidonic acid.
This leads to cell membrane damage and fre-
quently to cell death,

We performed experiments on isolated frog
neuromuscular junctions in order to determine
whether D-mannitol could modify the actions of
CTX-1B. Por this purpose, we used a dose of
D-mannito! which has no osmotic effects per se
but which protects kainate-induced death of
cerebellar neurons in culture by scavenging
hydroxy! radicals (Dykens et al.,1987). When
preparalions were pre-treated with 202M D-
mannitol, subsequent addition of CTX-1B
(2.5nM) caused similar effects to those observed
in the absence of D-mannitol i.e. there was an
increase of spontaneous quantal release and spon-
taneous asynchronous contractions (data not
shown) and depolarization of the muscle
membrane (Molg6 et al..1990). Thus, it appears
that 204M D-mannitol does not prevent the ac-
tions of CTX-1B at the neuromuscular junction.
When nerve terminals and Schwann cells were
imaged during 3hrs of CTX-1B action with the
confocal laser scanning micrascope the typical
changes above reported, i.e. changes in surface
area of nerve terminals and Schwann tells, were
observed.

We conclude that D-mannitol concentrations
which has been ed to exert an effective
hydroxyl radical-scavenger action neither
prevented binding of CTX-IB nor antagonized
effects of the toxin.

Further experiments were performed with

higher concentrations of D-mannito] (54.9mM
=1%; 109. 8mM=2% and 439.1mM =8%) added

MEMOIRS OF THE QUEENSLAND MUSELIM

to the standard Ringer’s solution to determine
osmotic effects of this agent. Since solutions con-
taining 1, 2 and 8% D-mamnnitol added to the
standard Ringer solution (Fig.4) had asmolalities
that are 24% (1% D-mannitol), 48.8% (2% D-
mannitel) and 203.7% (8% D-mannitol) higher
than the standard Ringer's solution and, in-
creases In Osmotic pressure causes dramatic in-
creases in spontaneous quantal release. At the
frog neuromuscular junction a 50% increase in
osmotic pressure by addition of sucrose causes a
reversible 45-fold increase in miniature endplate
potential frequency, as previously reported (Fatt
& Katz,1952). We did not attempt to study the
effects of CTX-1B in the presence of such high
coneentrations of mannitol. Instead we tried D-
mannitol after CTX-1B action at the neuromus-
cular junction to determine whether this agent at
different osmolaties could modify the changes in
the nerve terminal and Schwann cells previously
described with the toxin. D-mannitol effectively
caused a shrinkage of nerve terminals and
Schwann cells previously swollen by the action
of CTX-1B (Fig.5). When the effects of mannitol
wete quantified in nerve terminals it was evident
that 2% mannitol applied for 30min decreased the
nerve terminal surface area by 21+1.0% and the
Schwann cell surface area by 15.2%, while the
muscle fibre was decreased by 11.4+0.2%
(n=4). In control junctions D-mannitol reduced
the nerve terminal surface area by only 131.2%,
the Schwann cell surface area by 10=0.5% and
the muscle surface area by 11=0.6% indicating
that mannitol was more effective in reducing
nerve terminal surface area in nerve terminals
treated with CTX-1B than in control nerve ter-
minals. However, muscle changes caused by
mannitol were no different in CTX-1B treated
junctions as compared to controls. D-mannitol, at
concentrations that increased the osmolality of
the standard solution by c,50%, reversed swell-
ing of motor nerve terminals and Schwann cells
observed during long-term effects of CTX-1B,
These findings are important since the clinical
improvement observed in acute ciguatera after
mannitol treatment may be ascribed to the as-
motic action exerted by this agent in the
peripheral nervous system and skeletal muscle
fibres, which would result in a shift of water from
the intracellular to the extracellular compartment.

DISCUSSION

Cell swelling of motor nerve terminals and
perisynaptic Schwann cells was a common

CONFOCAL LASER SCANNING MICROSCOPY

FIG.5, Nerve terminal (NT) and pensynaptic Schwann cell somata (SC) from a junction treated for 3hr with
2.5nM CTX- 1B (a) and after 30min of D-mannitol (2%) added to the standard solution (b). Note the shrinkage

of structures after mannitol action.

response to CTX-1B application at living
neuromuscular junctions, Imaging methods are
the only way in which the shape changes accom-
panying cell volume changes can be determined.
However, determinations of cell volume are not
easy, even in a relatively simple synapse as the
neuromuscular junction, The term cell yolume is
4 complex concept because nerther motor endings
nor perisynaptic Schwann cells have simple in-
dividual geometric shapes and relationships.
Furthermore, the mechanisms of volume
homeostasis in motor endings have not been ex-
plored.

Changes in nerve terminal volume caused by
CTX-1B may result from the fusion of synaptic
vesicles to the presynaptic membrane and the
influx of Na* across the presynaptic membrane.
Previous electron microscopic studies of motor
endings in fixed specimens revealed time-de-
pendent increase in the nerve terminal perimeter,
alterations in nerve terminal mitochondria and
profound depletion of synaptic vesicles after
CTX-1B action (Molg6 et al., 1991: Molgo.
Comella & Legrand, unpubl.).

The Na* content of the nerve terminals is ex-
pected to be increased by CTX-1B. Under normal
conditions, water is in thermodynamic equi-
librium across the terminal membrane. However,
any change in the intracellular Na* concentration
will result in a rapid water flow from the ex-
tracellular to the intracellular compartment. Be-
cause the nerve terminal is readily distensible,
transmembrane water movements will result in
nerve terminal swelling. Schwann cell somata
swelling in situ is probably also related to the
increase in Na* concentration through activation

of sodium channels sensitive to the action of both
CTX-1B and TTX. The contribution of enhanced
quantal transmitter release to the swelling of
Schwann cells remains to be established,

D-Mannitol at concentrations reported to exert
an effective hydroxyl radical scavenger action
neither prevented the action of CTX-1B nor an-
tagonized its effects. However, at higher con-
centrations mannito] exerted osmotic effects that
caused shrinkage of both motor nerve terminals
and Schwann cel] somata previously swollen by
the action of CTX-1B probably by shifting water
from the intracellular to the extracellular com-
partment.

This report demonstrates that CTX-1B causes
time dependent changes in the surface area of
motor nerve terminals and perisynaptic Schwann
cells in living neuromuscular junctions. We have
shown that confocal laser microscopy is a new
tool for research on the effects of ciguatoxins on
living tissues. While the extent of its future ap-
plications in the field of the ciguatoxins is hard to
predict, its potential for neurobiological] research
appears enormous.

ACKNOWLEDGEMENTS

We are grateful to Dr Spencer Brown, Institut
des Sciences Végétales C.N.R.S. at Gif sur
Yvette for cogent advice conceming the use of
the confocal microscope and for stimulating dis-
cussions conceming the analysis of data. PJ. is
supported by a fellowship from Direction des
Recherches Etudes et Techniques (D.R.E.T.).
This work has been supported by grant 91/090
from D.R.E.T

584

LITERATURE CITED

ANDERSON, D.M. & LOBEL, P.S. 1987. The continu-
ing enigma of ciguatera. Biological Bulletin 172:
89-107.

BADEN,D.G. 1989. Brevetoxins: unique polyether
dinoflagellate toxins. FASEB Journal 3:
1807-1817.

BAGNIS, R., KUBERSKI, T. & LAUGIER, S. 1979.
Clinical observations on 3,009 cases of ciguatera
(fish poisoning) in the South Pacific. American
Journal of Tropical Medicine and Hygiene 28:
1067 — 1073.

BENOIT, E., LEGRAND, A.M. & DUBOIS, J.M.
1986. Effects of ciguatoxin on current and voltage
clamped frog myelinated nerve fibre. Toxicon 24;
357-364.

BETZ, W.J., MAO, F. & BEWICK, G.S. 1992. Ac-
tivity- dependent fluorescent staining and des-
taining of living vertebrate motor nerve terminals.
Journal of Neuroscience 12: 363-375.

BIDARD, J.N., VITVERBERG, H.P.M., FRELIN, C.,
CHUNGUE, E., LEGRAND, A.M., BAGNIS, R.
& LAZDUNSKI, M. 1984. Ciguatoxin is a novel
type of Na* channel toxin. Journal of Biological
Chemistry 259: 8353-8357.

BIRKS, R., HUXLEY, H.E. & KATZ, B. 1960. The fine
structure of the neuromuscular junction of the
frog. Journal of Physiology (London) 150:
134-144,

COUTEAUX, R. & PECOT-DECHAVASSINE, M.
1970. Vésicules synaptiques et poches au niveau
des zones actives’ de la jonction neuromus-
culaire. Comptes Rendus de |’Académie des
Sciences, Paris, Série D. 271: 2346—2349.

DAVE, V., GORDON, G.W. & MCCARTHY, K.D.
1991. Cerebral type 2 astroglia are heterogeneous
with respect to their ability to respond to
neuroligands linked to calcium mobilization, Glia
4:440-447.

DEL CASTILLO, J. & ESCALONA DE MOTTA, G.
1978. A new method for excitation-contraction
uncoupling in frog skeletal muscle. Journal of Cell
Biology 78:782—784.

DREYER, F. & PEPER, K. 1974. A monolayer prepara-
tion of innervated skeletal muscle fibres of the m.
cutaneous pectoris of the frog. Pfliigers Archiv
348:257 —262.

DYKENS, J.A., STERN, A. & TRENKNER, E. 1987.
Mechanism of kainate toxicity to cerebellar
neurons in vitro is analogous to reperfusion tissue
injury. Journal of Neurochemistry 49: 1222-1228.

FATT, P. & KATZ, B. 1952. Spontaneous subthreshold
activity at motor nerve endings. Journal of
Physiology (London) 117: 109-128.

GAWLEY, R.E., REIN, K.S., KINOSHITA, M. &
BADEN, D.G. 1992. Binding of brevetoxins and
ciguatoxin to the voltage-sensitive sodium chan-
nel and conformational analysis of brevetoxin B.
Toxicon 30: 780-785,

GILLESPIE, N.C., LEWIS, R.J., PEARN, J. et al.

MEMOIRS OF THE QUEENSLAND MUSEUM

1986. Ciguatera in Australia: occurrence, clinical
features, pathophysiology and management,
Medical Journal of Australia 145: 584—590.

HALLIWELL, B. & GUTTERIDGE, J.M.C. 1985. The
role of transition metals in superoxide-mediated
toxicology.Pp.45-82. In L.W. Oberley, ed., ‘Su-
peroxide Dismutase, Vol. III’. (CRC Press: Boca
Raton, Florida).

JAHROMI, B.S., ROBITAILLE, R. & CHARLTON,
M.P. 1992, Transmitter release increases in-
tracellular calcium in perisynaptic Schwann cells
in situ, Neuron 8: 1069-1077.

LEGRAND, A.M., LITAUDON, M., GENTHON,
J.N., BAGNIS, R. & YASUMOTO, T. 1989.
Isolation and some properties of ciguatoxin. Jour-
nal of Applied Phycology 1: 183—188.

LEWIS, R.J. & SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fish. Toxicon 30: 915—919,

LEWIS, R.J,. SELLIN, M., POLIN M.A., NORTON,
R.S., MACLEOD, J.K. & SHEIL, M.M. 1991.
Purification and characterization of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae), Toxicon 29: 1115—1127.

LOMBET, A., BIDARD, J.N. & LAZDUNSKI, M.
1987. Ciguatoxin and brevetoxins share a com-
mon receptor site on the neuronal voltage-depend-
ent Na’ channel. FEBS Letters 219: 355-359.

MOLGO, J., COMELLA, J.X. & LEGRAND, A.M.
1990. Ciguatoxin enhances quantal transmitter
release from frog motor nerve terminals. British
Journal of Pharmacology 99:695—700.

MOLGO, J., COMELLA, J.X., SHIMAHARA, T. &
LEGRAND, A.M. 1991a. Tetrodotoxin-sensitive
ciguatoxin effects on quantal release, synaptic
vesicle depletion, and calcium mobilization. An-
nals of the New York Academy of Sciences 635:
485—489. 7

MOLGO, J., DEL POZO, E., BANOS, J.E. & AN-
GAUT-PETIT, D. 1991b. Changes of quantal
transmitter release caused by gadolinium ions at
the frog neuromuscular junction. British Journal
of Pharmacology 104: 133-138.

MOLGO, J., BENOIT, E., COMELLA, J.X. &
LEGRAND, A.M. 1992a. Ciguatoxin: a tool for
research on sodium-dependent mechanisms.
Pp.149-164. In P.M. Conn, (ed.), “Methods in
neuroscience, vol.8, Neurotoxins’. (Academic
Press: New york).

MOLGO, J., SHIMAHARA, T., MOROT GAUDRY-
TALARMAIN Y., COMELLA, J.X. &
LEGRAND, A.M. 1992b. Ciguatoxin-induced
changes in acetylcholine release and in cytosolic
calcium levels. Bulletin de la Société de
Pathologie Exotique 85: 486-488.

MOLGO, J.. MOROT GAUDRY-TALARMAIN, Y.,
LEGRAND, A.M. & MOULIAN, N. 1993a.
Ciguatoxin extracted from poisonous moray eels
(Gymnothorax javanicus) triggers acetylcholine
release from Torpedo cholinergic synaptosomes
via reversed Na*-Ca2* exchange. Neuroscience
Letters 160: 65—68.

CONFOCAL LASER SCANNING MICROSCOPY

MOLGO, J., SHIMAHARA, T. & LEGRAND, A.M.
1993b. Ciguatoxin, extracted from poisonous
morays eels, causes sodium-dependent calcium
mobilization in NG108-15 neuroblastoma x
glioma hybrid cells. Neuroscience Letters 158:
147-150.

MOULIAN, N. & MOROT-GAUDRY-TALAR-
MAIN, P. 1993. Agelenopsis aperta venom and
FTX, a toxin purified from it, inhibit acetylcholine
release in Torpedo synaptosomes. Neuroscience
54: 1035-1041.

MURATA, M., LEGRAND A.M., ISHIBASHI, Y. &
YASUMOTO, T. 1989. Stnictures and configura-
tions of ciguatoxin and its congener. Journal of the
American Chemical Society 111: 8929—8931.

MURATA, M., LEGRAND A.M., ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations of ciguatoxin from the Moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386.

ORKAND, R.K. , NICHOLLS, J.G. & KUFFLER,
S.W. 1966. Effect of nerve impulses on the
membrane potential of glial cells in the central
nervous system of amphibia. Journal of
Neurobiology 29: 788—806.

ORKAND, R.K. 1982. Signalling between neuronal
and glial cells. Pp.147—157. In T. A. Sears, (ed.),
‘Neuronal-glial cell interrrelationships’.
(Springer-Verlag: New York).

585

PALAFOX, N.A., JAIN, L.G., PINANO, A.Z.,
GULICK, T.M., WILLIAMS, R.K. & SCHATZ,
LJ. 1988. Successful treatment of ciguatera fish
poisoning with intravenous mannitol. Journal of
the American Medical Association 259: 2740—
2742.

PEARN, J.H., LEWIS, R.J. & RUFF, T. et al. 1989,
Ciguatera and Mannitol: experience with a new
treatment regimen. Medical Journal of Australia
151:77-80.

RITCHIE, J.M. 1992. Voltage-gated ion channels in
Schwann cells and glia. Trends in Neurosciences
15: 345-350.

RUSSELL, F.E. & EGEN, N.B. 1991. Ciguateric fishes,
ciguatoxin (CTX) and Ciguatera poisoning. Jour-
nal of Toxicolology-Toxin Reviews 10, 37—62.

SCHEUER, P.J., TAKAHASHI, W., TSUTSUMI, J. &
YOSHIDA, T. 1967. Ciguatoxin isolation and
chemical nature. Science 155: 1267—1268.

SONTHEIMER, H. 1992. Astrocytes, as well as
neurons, express a diversity of ion channels.
Canadian Journal of Physiology and Pharmacol-
ogy 70: $223-S238.

TACHIBANA, K., NUKINA, M., JOH, Y. &
SCHEUER, P. 1987. Recent developments in the
molecular structure of ciguatoxin. Biological Bul-
letin 172; 122-127.

SWIFT A.E.B. & SWIFT, T.R. 1993. Ciguatera. Jour-
nal of Toxicology-Clinical Toxicology 31: 1—29,

WITHERS, N.W. 1982. Ciguatera fish poisoning. An-
nual Review of Medicine 33: 97—111.

586

MODIFICATION OF NERVE CONDUCTION IN THE
RAT BY BREVETOXIN (PBTX-3). Memoirs of the
Queensland Museum 34(3): 586. 1994:— Brevetoxins are
lipid-soluble polycyclic ether toxins isolated from the marine
dinoflagellate Ptychodiscus brevis. The toxins PbTx-2 and
PbTx-3 bind to a specific receptor site (site 5) on the voltage-
dependent sodium channel, a site shared with ciguatoxin. This
study set out to examine the effects of PbTx-3 and a possible
antagonist on the parameters of nerve conduction.
Electrophysiological studies were carried out on the
ventral coccygeal nerve of male Wistar rats. Prior to ex-
perimentation each animal was anaesthetised with intramus-
cular Leptan (420pI/kg). A Medelec MS92a electro-
myography unit was used for recordings. PbTx-3 (15pg/kg)
was administered intravenously over 15 minutes. In an-

CIGUATERA RESEARCH - AN HISTORICAL
PERSPECTIVE. Memoirs of the Queensland Museum
34(3): 586. 1994:— Ciguatera research at the University of
Hawaii was initiated in the mid-1950’s by the late Professor
A.H. Banner, who formulated four principal objectives: 1,
What is the molecular structure of the toxin? 2, What is the
origin of the toxin? 3, Can a diagnostic test be devised that
distinguishes toxic from nontoxic fish? 4, Can an effective
human therapy be found?

Elucidation of the molecular structure was of central con-
cern since success with the other three goals would be greatly
enhanced, or depend on, a knowledge of structural features.

Inadequate supplies of toxic fish, establishment of a
suitable bioassay, and technology of the 60’s made for slow
progress.Even after the discovery of a dinoflagellate as the
primary toxin producer in 1977, moray eels had to remain the
sole source of toxin, since G. toxicus cultures yielded only the
water-soluble maitotoxin, which was distinctly different from
the lipid-soluble ciguatoxin extracted from, carnivorous fish.

MEMOIRS OF THE QUEENSLAND MUSEUM

tagonist experiments lignocaine (5OOwg/kg) was delivered
intravenously, over 30 minutes.

PbTx-3 produced a significant increase in both the mag-
nitude and duration of supernormality to that of control ner-
ves. This toxin also increased the absolute and relative
refractory periods and decreased the conduction velocity.
Lignocaine returned these parameters towards control values.

These results demonstrate that PbTx-3 alters nerve con-
duction parameters of rats in a similar way to ciguatoxin. It is
suggested that brevetoxin may provide a suitable model in
further studies pertaining to possible therapeutic agents for
ciguatera poisoning.

Christine E. Purcell, John Cameron & Michael F. Capra,
School of Life Science, Queensland University of Technology,
GPO Box 2434, Brisbane 4001, Queensland; 12 April, 1994.

While an extensive search for an algal food source or a
toxin precursor produced no useful leads, it gave rise to
significant discoveries, most prominent among them of
palytoxin, which in time became a benchmark in marine
natural product chemistry, and indeed, in all of organic
chemistry.

The first clue that ciguatoxin belonged to the structural
type of polyethers did not come from sophisticated instrumen-
tation, but from its behaviour in chromatography which paral-
leled that of okadaic acid, a compound first isolated and
characterized as a constituent of a sponge and subsequently
identified as a metabolite of the dinoflagellate Prorocentrum
lima. Interestingly, okadaic acid has low mammalian toxicity
and has become an important probe in the study of cellular
regulation.

Paul J. Scheuer, Department of Chemistry, University of
Hawaii at Manoa, Honolulu, HI 96822, U.S.A.; 12 April,
1993.

REEF MANAGEMENT AND SEAFOOD MONITORING PROGRAMS
FOR CIGUATERA

DOUGLAS L_ PARK

Park, DL. 1994 08 01: Reef management and seafood monitoring programs for ciguatera,
Memoirs of the Queensland Museum 34(3); 587-594. Brisbane. ISSN 0079-8835.

Ciguatera (CTX) toxins in fishery products are odorless, tasteless, and generally undetectable
by simple chemical test, bioassays traditionally monitor suspect fish. Assurance that suscep-
tible foods are safe to eat will come from marketplace screening, separation of adulterated!
product to less risk uses, and, where feasible, prediction of potentially hazardous food
production/harvesting areas, An effective screening method for use in the marketplace must
be: (a) easy to use and interpret; (b) able to test a large number of samples in a short period;
(c) accurately differentiate between toxic and non-toxic product; (d) low cost; (c) available
in sufficient quantity to meet private, industrial, and regulatory agency demands; and (1)
where possible identify toxins involved. The solid-phase immunobead assay (S-PIA,
Ciguatect™) has the highest potential for this purpose. The kit can be used on fishing vessels,
ai receiving docks, processing plants, distribution organizations, retail outlets, consumers,
and regulatory agencies and is designed for non-laboratory use by untrained personnel,

Douglas L. Park, Department of Nutritional Sciences, University of Arizana, Tucson, AZ

85721 USA; 2 February 1994,

Ciguatera fish poisoning is a centuries old ill-
ness, endemic to tropical and subtropical areas
(WHO,1984), sometimes shipped to nontropical
population centres (North America). Humans are
exposed through consumption of fish which have
accumulated toxins produced by dinoflagellates.
An estimated 50,000—500,000 cases of ciguatera
occur each year (Ragelis, 1984). Symptoms are
gastrointestinal, neurological and cardiovascular
and can persist for weeks and even years
(Juranovic & Park,1991). U.S. public health
agencies (Food and Drug Administration and Na-
onal Marine Fisheries Service) have been striy-
ing 10 implement a combative seafood safety
program for years.

Public health research on ciguatera has focused
on protection of human health and enhancement
of commerce in subtropical reef fish. To set up an
effective seafood monitoring program, itis neces-
sary to understand how products become toxic
and so develop an analytical technique for detec-
tion. Historically, methods of analysis for
ciguatoxins (CTX) have been labor-intensive,
me-consuming, and not able to identify in-
dividual toxins (Juranovic & Park,1991).

TOXIN PRODUCING ALGAE AND PRIN-
CIPAL TOXINS INVOLVED IN
CIGUATERA

CTX accumulates in benthic feeding her-
bivorous fish and then up the food chain to man.
Benthic toxigenic dinoflagellates suspected in

ciguatera poisoning include Gambierdiscus
toxicus, Prorocentrim lima, P. concavum, P.
emarginalum, P. mexicanum, P. rathynum, Am-
phidinium carterae, Ostreopis ovata, O. siamen-
sis, O. lenticula, Ceolia monotis, Scrippsieila
subsalsa, and Thecadinium sp,

Of several toxins which may be responsible for
ciguatera, ciguatoxin has been isolated as the
mapor toxin from large carnivores while smaller
amounts have been detected in herbivores. An
explanation for this could be that CTX accumu-
lates preferentially in large carnivores due to its
greater lipid solubility. Murata et al. (1990)
reported the structures of ciguatoxin from the
moray cel (Gymnothorax javanicus) and its likely
precursor lrom G. toxicus. The congener was
shown to be a less oxygenated analog of cigua-
toxin, However, ithas not been demonstrated that
the toxin produced by the dinoflagellate is the
precursor to ciguatoxin(s) accumulating in fish.
Until sufficient quantities of individual toxins
become available and suitable detection methods
for these toxins are developed, it will be difficult
to determine toxin properties. At least five toxins
ure implicated in ciguatera: they are cigualoximn
(CTX), maitotoxin (MTX), scaritoxin (STX),
okadate acid (OA), and a recently named toxin,
proroventrolid (Bagnis et al..1974; Chungue et
al.,1977; Maulin et al., 1992, Tachibana, 1980;
Tindal! et a)..1984; Yasumoto et al_,1971; Yasu-
moto et al.,1984; Yasumoto & Murata,1988a;
Yasumotoa & Murata,|9&8b; Yasumoto &
Scheuer, 1969). Recent studics suggest that in ex-

SkH

THR ISLAND OF HAWALL

Hamakiun Coast

Fatayutasetiew

] ‘Mustiatn Booed

Watknton

fetiacdow's boy ae
Kati

BNon-loxic Sites
§ Oicie Sites

RHP Tou ° ® Reportud Cass

FIG. 1. Reports of cases of ciguatera fish poisoning
outbreaks recorded by Hawaii State Department of
Health (1981-1990), Toxic (Kona Coast) and non-
toxic (Hamakua Coast) sites for collecting biomarker
model are noted.

cess of 20 toxins may be inyolyed in the ciguatera
phenomenon (Juranovic et al.,in press;
Legrand,1991; Lewis et al.,1991; Lewis & Sel-
lin, 1992: Lewis,1992). Relative concentrations
and toxin profiles for each toxie fish vary greatly
and are unknown due to the lack of individual
toxin reference standards and specific analytical
methods.

Okadaic acid is available commercially as a
standard reference material. For CTX, however,
ciguatoxin and possibly maitotoxin and their
analogs are the toxins with the highest toxic
potentials, The toxic potential of okadaic acid js
several orders of magnitude lower than
clguatoxin,

ANALYTICAL METHODOLOGY

Analytical methods for phycotoxins vary ac-
cording to the application, i.e., screening, iden-
ufying, etc. (Park,1994). For ciguatera biossays
have been used in the laboratory but are un-
suitable as a marketplace test. Most earlier
methods were based on biological endpoints
which had major limitations on levels of detec-

MEMOIRS OF THE QUEENSLAND MUSEUM

tion and specificity. Many native tests for fish
toxicity have been examined, including dis-
coloration of silver coins, or copper wire, the
repulsion of flies or ants, and rubbing the liver on
the gums to ascertain if it causes a tingling feeling
(Juranovic & Park,1991). With the possible ex-
ception of rubbing the liver on the sensitive tis-
sues of the mouth, all have proven invalid. As
more reference material and standards became
available, chemical and immunochemical
methods have emerged.

Bioassays have one common disadvantage: the
lack of specificity Fur individual toxins. Alterna-
tive methods based on immunochemistry
(Hokama et al,,1977) are applicable to screening
fish in the marketplace. The original assay, a
radioimmunoassay (RIA) for ciguatoxin, was
developed using antibodies produced against a
conjugate of human serum albumin and
ciguatoxin (isolated from toxic moray eel) in-
jected into sheep and rabbit. This assay was used
successfully to test ciguatera and to screen for
toxic amberjacks (Seriola dumerilt) where 15%
of the fish Were rejected during a 2-yr study on
the Hawaiian market (Kimura et al.,1982).
Despite this success, the assay was not suitable
for routine use due to high cost, instrumentation
requirements, and time involyement.

In 1983, a competitive enzyme immunoassay
(ETA) commonly called the ‘sticktest’ was
developed using the polyclonal antibody used in
the RJA, and evaluated on Hawaiian reef fishes
(Hokama et al.,1983; Hokama et al.,1984; Hok-
ama,1985). As with its predecessor, this antibody
demonstrated close structural similanty of CTX,
MTX, brevetoxin, and OA. EIA used liquid-
paper applied to bamboo sticks to isolate and bind
the toxins (Hokama, 1985). This assay was able to
distinguish between toxic and nontoxic fish. Test
results revealed a high number of false-positives,
although no false-negatives Were observed (Hok-
ama et al.,1987; Hokama & Miyahara,1986).

The stick test Was modified further using
monoclonal antibodies specific for CTX, OA,
and a synthetic fragment of OA, that are more
specific than the sheep antibody (Hokama ct
al,,1990; Hokama et al., 1992; Hokama et
al.,1986). This antibody gave peak tilers of | Sng,
10ng and SOng, respectively, for CTX, the frag-
mentof OA, and OA (Hokamaet aJ., 1992). Com-
petitive inhibition analyses showed that 4ng
purified CTX blocked completely the antibody
reaction with crude CTX, OA and the fragment
of OA at similar concentrations (approximately
SOng), This assay was used to test fish specimens

SEAPOOD MONITORING FOR CIGUATERA

from documented cases of ciguatera (Hawaii
Department of Health) with 98% agreement
(Hokama et al.,1989). A preliminary collahora-
tive evaluation study of the rapid enzyme im-
munoassay stick test was conducted (Ragelis,
1987,1988). Eight of the nine laboratones in-
volved obtained results within acceptable limits
for each of 3 fish cake samples homogenized with
ciguatoxin, The relative standard deviation of
reproducibility (RSDx) was 23-30%, Due to the
Jack of a chemically identifiable standard, the ful)
collaborative study was not conducted.

This assay was modiffed to a solid-phase im-
munobead assay format (Hokama,1990) com-
monly known as the “paddle test’, using bamboo
paddles coated with liquid paper. This format was
used to test 26 cases of ciguatera with [00%
agreement. In a study comparing the stick and
paddle tests. 436 specimens with varied levels of
toxicity showed 80% agreement (Hokama, 1990).

Patents covering the stick and paddle tests were
purchased by HawaiiChemtect International. The
original format was modified to an innovative
rapid solid-phase immunobead assay {S-PLA,
Ciguatect’™) for ciguateric toxins (CTX) and
diartheic shellfish poisoning (DSP) outbreaks
(Park & Goldsmith,I991]; Park et al. [992b)-
Toxins are determined by binding them to a
membrane attached to a plastic strip and exposing
the toxin-ladened membrane to a monoclonal
antibody-colored latex bead complex which has
a high specificity for the toxins of interest. The
intensity of the color on the membrane denotes
the toxins, CTX toxicity potential can be deter-
inined directly on edible tissue or following
specific extraction procedures. The method has
been used to evaluate CTX potential in fish ob-
tained from Hawaii, Australia, and the Caribbean
(Park et al.,1992b). The Ciguatect™ test kit has
heen compared to the mouse assay for the detec-
tion of toxic fish, i.e. fish of tropical origin for
sale in Canada (Todd ef al,,1992) and fish col-
lected from St. Thomas, U.S. Virgin Islands
(Dickey etal..this memoir). Both studics reporied
a high percentage of fish toxic to the mouse and
positive for CTX-related toxins by the
Ciguatect™ test kit or following a rapid extrac-
tion and purification procedure. CTX-related
LOXiNs are present in a significant number of fish;
however, the toxicological or public health sig-
nificance is unknown, i.e., would the toxin(s)
present (profile, potency and concentration) pose
a Significant risk for acute poisoning and/or
chronic toxicity? Todd and co-workers used 135—
250g equivalent fish flesh for injection into the

589

mouse where 85% of the mice died within
24hours. Dickey et al. (this memoir) used 45—
180g equivalent fish flesh and 67% of the mice
died within 48 hours. Interpretation of mouse
assay resulis must be made with caution, how-
ever. Mouse toxicity results have been useful in
confirming toxins in ciguatera outbreaks, al-
though the mouse is relatively resistant to CTX.
Because of the lack of specificity, the mouse
bioassay should not be used to predict cigua-
toxicity. This was particularly apparent with the
Dickey et al. study where for 22% of the
specimens one animal died within a short time
frame and the duplicate animal survived 48 hours.
43%, of the animals that died, died within JOmin.
Shon death times (<S0hnin) with the mouse ane
not considered ciguatoxic. Additionally, the
statistical procedure used by the authors can only
be applied when thé method used for companson
by definition is 100% accurate for sensitivity and
specificity (Riegelman & Hirsch,1989). The
mouse assay is unlikely to be 100% accurate in
determining ciguatera toxicity. As was pointed
out by Hoffman etal. (1983) and Vernoux (1993).
the mouse can have utility when symptomatology
is used as well (symptomology was used as a
criteria in the Dickey et al. study). A preferred
procedure for the evaluation of a test method is
outlined below.

For those products where additional testing rs
desired, possibly for samples testing positive, the
University of Arizona and HawaiiChemtect In-
ternational have developed a rapid extracion
method (REM™) capable of extraction and piar-
tia) purification of ciguateri¢ toxins in <30 mings
(Park et wl..1992a). The REM procedure isolates
and purifies CTX-related toxins on the same
chemical basis as used in more exhaustive CTX
extraction procedures. For the REM™ , toxins are
extracted with a chloroform:water:methinol
mixture and pariuioned into selected phases by
varying polarity. When the REM™ is used in
combination with the Ciguatect™ test kit, the
limit of detection for ciguatoxin, okadaic acid and
related toxins 1s <0.03ng/¢ fish flesh. Also, at this
point chemical methods based on thin layer
(TLC) or high performance liquid chromatog-
raphy (HPLC) technology can be used to confirm
individual toxins.

Methods based on thin layer (TLC) and high
performance liquid chromatography (HPLC)
have been developed for selected individual
toxins associated with CTX (Lee et al., 1987;
Legrand,1991, Dickey er al., 1990). These
methods can be applied as a regulatory tool where

590

TABLE 1. Precision parameters of collaborative data
for solid-phase immunobead assay (Ciguatect™)
determination of ciguatoxins and related polyether
compounds in parrot fish, surgeon fish and amber-
jacks from the Hawaii Island.

Sr

Ny

Parrot Fish (Scarus sp.)

Surgeon Fish
Ctenochaetus sp.

REM Extracts

Parrot Fish (Scarus sp.) | _3.1_ | 0.18] 0.37| 5.8 | 11,9

Surgeon Fish 3.8 |0.18}0,38] 4.8 | 9.9
(Ctenochaetus sp.)

| Amberjack (Caranx sp.) | 4.9 |0.18|0.37| 3.7 | 7.6

S;= Standard deviation of repeatability

S r= Standard deviation of reproducibility

RSD,= Relative standard deviation of repeatability
RSD r= Relative standard deviation of reproducibility

sophisticated laboratory facilities are available.
HPLC techniques have been applied to analysis
of okadaic acid in fish and shellfish (Gamboa et
al., 1992; Yasumoto, 1985; Lee et al.,1989; Dick-
ey et al.,1990), Park and co-workers (unpubl.
data) have developed a TLC method for okadaic
acid in fish tissue and dinoflagellate cultures.
Specificity of this methodology is enhanced by
exhaustive purification of toxins extracted from
fish tissue. Unfortunately, CTX and okadaic acid
develop similarly on TLC. These methods, al-
though not suitable for routine screening
programs, could play an important role in con-
firming the presence of individual toxins in fish
products.

HPLC methodology have been reported for
ciguatoxin and several analogues (Murata et
al.,1990; Lewis et al.,1991; Lewis & Sel-
lin,1992), These studies reported four major
ciguatoxins. Legrand (1991) and co-workers
(Legrand et al.,1990) used HPLC methodology to
isolate multiple ciguatera toxins from wild Gam-
bierdiscus toxicus and toxic herbivorous and car-
nivorous fish.

METHOD VALIDATION

Any method intended to be used in a seafood
safety monitoring program must pass stringent
in-house evaluation and be validated through an
inter-laboratory study to determine precision

MEMOIRS OF THE QUEENSLAND MUSEUM

(reproducibility, repeatability) and accuracy
(recovery) parameters of the method. Method
validation programs are administered by AOAC
International (AOAC) and International Union
for Pure and Applied Chemistry TUPAC). These
validation programs include two phases. The first
phase is a ruggedness test or feasibility (mini-col-
laborative) study. Acceptable results in this phase
lead to a collaborative study (Phase II). These
studies (mini- and full-collaborative) involve the
distribution of coded samples (in duplicate) of
fish, preferably authentic ciguatera fish poison-
ing specimens, to participating laboratories.
Known amounts of specific standards are also
added to some of the samples. The samples are
analyzed following exact method protocols and
results returned to the study organizer. Phase I of
the validation process, i.e., mini-collaborative
study, was carried out for the S-PIA using fish
fillets and REM™ extracts (Park et al.,1992a,b).
The AOAC/IUPAC inter-laboratory mechanism
was used. The precision of the S-PIA
(Ciguatect™) to detect CTX has been evaluated
through analysis of toxic and non-toxic fish fillets
(amberjack, surgeon, and parrot fish) and
REM" extracts of the same fish obtained from
fishing areas around the Hawaiian Islands.
Toxicity potentials of purified extracts were
determined using the mouse and brine shrimp
(Artemia sp.) assays. The analysis showed ac-
ceptable repeatability and reproducibility
parameters (Table 1). University of Arizona,
FDA and NMFS laboratories participated in the
study. The study confirmed excellent perfor-
mance and interpretation of results, and
demonstrated acceptable precision parameters
(Park et al.,1992a,b). A full-collaborative study
of the test kit is recommended. The validation
study will include test portions of naturally and
artificially contaminated fish with ciguatoxin and
okadaic acid. Toxicity potentials will be deter-
mined using the mouse and brine shrimp assays.

SEAFOOD SAFETY MONITORING
PROGRAMS

An effective food safety monitoring program
comprises: 1, monitoring fish harvesting areas for
CTX; 2, establishment of regulatory limits, and;
3, screening commercial fish products. Unaccep-
table product can be further tested to identify the
toxin(s).

MONITORING FISH HARVESTING AREAS
Since seafoods commonly associated with

SEAFOOD MONITORING FOR CIGUATERA

ciguatera poisoning outbreaks are associated with
highly mobile fish, collecting and testing such
fish alone could provide misleading information
so testing less mobile species (e.g. invertebrates)
should be included.

The Ciguatect™ S-PIA screened 36 species of
nearshore invertebrates off the Island of Hawaii
for ciguatoxin and related polyethers (Fig.1)
(R.G. Kvitek, Moss Landing Research Lab-
oratories, and D.L. Park, University of Arizona,
unpubl. data). Specimens included snails, sea ur-
chins, sea cucumbers, crabs, brittle stars, bival-
ves, and zoanthids. Invertebrates were collected
at 6 ‘toxic’ locations along the Kona coast, and at
3 ‘non-toxic’ sites along the Hamakua coast
where there had been only one reported case of
ciguatera since 1980.

A significant positive correlation between
assay results and site-specific ciguatera history
was found for the cowry Cypraea maculifera
(Fig.2). While assay results for most other species
indicated low or no ciguatoxin, cone snails
(Conus), ophiuroids (Ophiocoma) and sea
cucumbers (Holothuria) tested positive frequent-
ly. There was no correlation, however, for these
three genera between assay results and site his-
tory. These results suggest that invertebrates, par-
ticularly grazers and deposit feeders, and
especially cowries, accumulate ciguatoxins and
related polyether compounds at sites known for
ciguatera fish poisoning outbreaks and have the
potential utility of being bio-indicators of reef
toxicity. This marine specimen, or other inver-
tebrates native to other areas, could be an integral
part of the ciguatera monitoring program.

ESTABLISHMENT OF REGULATORY LIMITS
Based on mouse and mosquito bioassay data,
several levels of concern have been proposed.
Since multiple toxins of varied toxin potential are
involved with CTX poisoning, it is not practical
to use a single compound for this regulatory limit.
Historically, a seafood safety monitoring pro-
gram for ciguatera has been hampered by the lack
of reference standards, particularly ciguatoxin.
Okadaic acid is the only toxin associated with
CTX poisoning in sufficient quantities to serve as
a reference standard. Toxicity of okadaic acid,
however, is significantly lower than ciguatoxin
where acute mouse toxicity potentials for
ciguatoxins and okadaic acid are 0.45 and
210yug/kg, respectively. The term okadaic acid
equivalents (OAE) could be used, however, to
standardize analytical methods and in the estab-
lishment of regulatory limits, provided the rela-

591

FE) Hamakua Coast

3
fo}
1S)
3
S
C}
MM

Cypraea maculifera

SN
oe8eo0@ we
o> ge » 82 3 vo »w
#88 §€ #332 4
Ss fF 6 8S Os & 5S
=| 6p oo 8 @&
a s ©@ U ~»~ & FD
9 ga 8 & £3
a «as
“ 2g 49 Ae
42 5 =
ai a

FIG. 2. Ciguatoxin immunoassay scores for Cypraea
maculifera from Kona and Hamakua Coasts of
Hawaii. Color intensity on the test strip assigned a
value between 0-6 where 0 = nondetectable and 5 =
color intensity equal to 5ng okadaic acid. Numbers in
() indicate number of individual specimens tested and
values pooled.

tive potencies of all toxins involved with the CTX
phenomenon are used in calculation of action
levels. For this to be feasible, the test employed
must recognize all toxins involved in the poison-
ing in a similar manner and the action level focus
on the toxin of highest potency. Again, the term
OAE would be used because multiple toxins are
involved in the poisoning.

SCREEN FISH IN THE MARKETPLACE/COMMER-
CIAL CHANNELS

Any method for screening marketplace
seafoods must: a, be easy to use and interpret; b,
be rapid, i.e., able to test a large number of
samples in a short time; c, accurately differentiate
between toxic and non-toxic samples; d, have low
cost; e, be available in quantities to meet private,
industrial, and regulatory agency testing

demands: and f, where feasible, confirm toxin
identity.

The S-PIA method (Ciguatect™) has high
potential for screening market place fish. When
fully validated, the kitmay be used at the harvest-
ing, processing, distribution, retail] or other point
through the marketing route. Testing fish early
after capture is recommended, since this will
minimize cost expended for the product and
potential economic loss to the industry. The kit
can be used on-board fishing vessels, at receiving
docks, processing plants, distnbuting organizit-
tions, retail outlets. consumers, and regulatory
agencies. The self-contained assay is available as
a single analysis kit designed for non-laboratory
use by untrained personne], Organizations con-
ducting large numbers of analyses would be more
inclined to use the Jaboratory kat which contains
sufficient material for >50 tests.

The Seafood Safety Monitoring Program
would involve large-scale testing of fish accord-
ing to an acceptable sampling plan. Fish or lots
testing negative to the screening procedure would
be allowed to proceed normally in commercial
channels. Each point identified above would be a
quality control point. Product testing positive for
toxic potential would be diverted ta lower risk
uses oF retested to confirm toxic potential. This
can be done by using the REM procedure which
isolates, purifies and concentrates the toxins
before retesting or by using alternative les
methods for specific toxins.

LITERATURE CITED

BAGNIS, R., LUOSSAN, M.E. & THEVENIN, 8.
1974. Les tntoxiesbons par poisons perrxyuets
me Iles Gambier. Medicine Tropicale 34:323—

CHUNGUE, E., BAGNIS, R.. FUSETANAI, N. &
HASHIMOTO, Y- 1977. [solation of two toxins
from pesrotfish Scarus gibus. Toxicon 15:89-93.

DICKEY, R.W.. BOBZIN, S.C,, FAULKNER, D.J..
BENCSATH, F.A. & ANDRZEJEWSKI, D,
1990. Identification of okadaic acid froma Carib-
bean dinoflagellate, Prorocentnem concquum,
Toxicon 28: 371-377.

DICKEY, R.W., GRANADE, H.R. & MCCLURE,

FD. this memoir. ‘Evaluation of the Ciguatect ™
immunoassay for the detection of ciguatera-re-
Jated biotoxins in Canbbean fish’.

GAMBOA, P.M., PARK, D.L. & FREMY. J.M. 1992.
Extraction and punfication of toxic fractions from
barracuda (Spiiyraena barracuda) implicated in
Ciguatera poisoning, Pp, 13-24, In T.R. Tosteson
(ed.), ‘Proceedings of the 3rd Internauional Con-
ference on Ciguatera Fish Poisoning’ (Polys-
cience: Morin Heights, Quebec).

MEMOIRS OF THE QUEENSLAND MUSEUM

HOFFMAN, P.A., GRANADE, H.R. & MUMILLAN,
J.P. 1983. The monse ciguatoxin bioassiy: a dose
response curve and symptomatology analysis.
Toxicon 21; 363-369,

HOKAMA, Y. 1985. A rapid simplified enzyme im-
Munoassay stick test for the detection of
clguatoxin and related polyethers from fish tls-
sues, Toxicon 23; 939.

HOKAMA, Y. 1990, Simplified solid-phase im-
munobead assay for detection of ciguatoxin and
related polyethers. Journal of Clinical Laboratory
Analysis 4: 213-217,

HOKAMA, Y., ABAD, M.A, & KIMURA, L.H. 1983,
A rapid enzyme immunoassay (EIA) forthe detec-
Uon of ciguatoxin in contaminated fish tissues.
Toxicon 21; 817-824.

HOKAMA, Y., ASAHINA, A.Y., HONG, T.W.P.,
SHANG, E.S. & MIYAHARA, J,T. 1990
Evaluation of the sick enzyme immunoassay in
Carnax sp, and Serivla dyumerilf associated with
ciguatera, Joumal of Clinical Laboratory Analysis
4: 363-366.

HOKAMA, Y., BANNER, A,H, & BOYLAN, D.A.
1977. A radioimmunoassay for the detection of
ciguatoxin, Toxicon 15; 317-325.

HOKAMA, Y,, HONDA, S.A.A., KOBAYASHI,
M.N., NAKAGAWA, L.K., ASHINA, ALY. &
HIVAHARA, J.T, 1989, Monoclonal antibody
(MAD) in detection of ciguatoxin (CTX) and re-
juiced poyethers by stick-enzyme immunoassay
(5-BA) in fish tissues associated with ciguatera
poisoning, Pp.303-309, In 5. Natori, K,
Hashimoto & Y. Ueno (eds), ‘Mycotoxins and
phycotoxins "88". (Elsevier: Netherlands)

HOKAMA, Y., HONDA, 8.A.A., UYEHARA, K.,
SHIRAI, L.K. & KOBAYASHI, M.N. 1986.
Monoclonal antibodies ta low dalton natural
marine toxins, (Abstract), Jounal of Toxicology:
Toxin Review 5(2): 194.

HOKAMA, Y,, HONG, T.W,.P,, [ISOBE, M.,
ICHTIKAWA, Y. & YASUMOTO, T, 1992, Cross
reactivity of highly purified okadaic ucid (QA),
synthetic, spiroketal east sphere of OA and
ciguatoxin. Journal of Clinical Laboratory
Analysis 6: 54-58.

HOKAMA, Y., KIMURA, L.H., ABAD, M.A.,,
YORKOCHI, L., SCHEUER, PJ., NUKINA, M.,
YASUMOTO, T., BADEN, D.G,, & SHIMIZU,
Y. 1984. An enzyme immunoassay for the detec-
bon of ciguatoxin and competitive inhibitions of
related natural polyethers toxins. American
Chemical Society Symposium Serics 262: 307—
320,

HOKAMA, Y. & MIYAHARA, J.T, 1986, Ciguatera
Poisoning: Clinical and immunological aspects.
Journal of Toxicology; Toxin Reviews 5; 25-31.

HOKAMA, Y,, SHIRAT, L.K., IWAMOTO, LM.,
KOBAYASHI, M.N., GOTO, CS. &
NAKAGAWA, L.K. 1987, Assessment of a rapid
enzyme immunoassay stick test for the detection

SEAFOOD MONITORING FOR CIGUATERA

of ciguatoxin and related polyether toxins in fish
issues. Biological Bulletin 172: 144-153.

JURANOVIC, L.R..& PARK, D.L. 1991, Food bome
toxins of marine origin: Ciguatera. Revue En-
vironmental Contamination and Toxicology 117:
51-94.

JURANOVIC, L.R., PARK D.L. & FREMY, J.M. in
press. Isolation/separation of toxins produced by
Gambierdiscus toxicus and Prorocentrum con-
cavum. Journal of Aquatic Food Product Technol-

ogy.

KIMURA, L.H., ABAD M.A., HOKAMA, Y. 1982.
Evaluation of the radioimmunoassay for detection
of ciguatoxin in fish tissues, Journal Fisheries
Biology 21: 671-680.

LEE, J.S.. YANAGI, T., KENMA, R, YASUMOTO,
T. 1987. Fluorometric determination of diarrhetic
shellfish loxins by high performance liquid
chromatography. Agricultural and Biological
Chemistry 51: 877-881.

LEE, J.S.. MURATA, M. & YASUMOTO, T. 1989.
Analytical methods for ihe determination of diar-
rhetic shellfish toxin. Pp. 327-334. In S. Natori,
K. Hasimoto & Y. Ueno, (eds), ‘Mycotoxins and
phycotoxins’, (Elsevier; Netherlands).

LEGRAND, A.M, 1991, Les toxines de [a ciguatera, In
‘Proceedings of Symposium on Marine Biotoxins,
30-31 January 1991, Paris, France’.

LEGRAND, A.M., FUKUI, M.. CRUCHET, P.,
ISHIBASHI, Y. & YASUMOTO, T, 1990), Char-
acterization of toxins from differcnt fish species
and wild G. roxicus. Pp. 25-32, In T.R. Tosteson,
(ed.), “Proceedings 3rd International Conference
on Ciguatera’ (Polyscience Publishers: Morn
Heights; Quebec).

LEWIS, R.J. 1992. Ciguatoxins are potent ich-
thyotoxins. Toxicon 30: 207-211.

LEWIS, R.J.& SELLIN, M. 1992. Multiple ciguatoxins
in the flesh of fish, Toxicon 30: 915-919,

LEWIS, R.J.. SELLIN, M., POLI, M.A., NORTON,
R.S., MACLEOD, I.K. & SHEIL, M.M, 1991.
Punfication and characterzation of ciguatoxins
from moray ee! (Lycedantis javanicus, Muraco-
idae), Toxicon 29; 1115-1127.

MOULIN, F.. VERNOUX, J.P... FREMY, JM, &
LEDOUX, M. 1992 ‘Dinoflagellate toxins |i
volved in marine fondborne intoxicatinn’,
(CNEVA, Laboraipire Central d’' Hygieuc
Alimentaire; Paris),

MURATA, M., LEGRAND, A.M. ISHIBASHI, ¥.,
FUKUI, & M, YASUMOTO, T. 1990. Structures
of ciguatoxin and ils congener. Journal ol the
Amenean Chemical Society 112: 4380-4386.

PARK, D.L. 1994.. Evolution of methods for assessing
ciguatera toxins in fish. Revue Environmental
Contamination and Toxicology 136:1-20.

PARK, D.L. & GOLDSMITH, C.H, 1991, Inter-
jaboratory validation of the solid-phase im-
munobead assay for the detection of toxins
associated with ciguatera poisoning, (Presented al
the Sth International Conference on Toxic Marine

Phytoplankton, 28 October-! November, 1991
Newport, Rhode Island),

PARK, DL... GAMBOA, P.M. & GOLDSMITH, CH,
1992s. Validation of the solid-phase immunobead
assay (Ciguatect’”’) for toxins associated with
ciguatera poisoning. (Presented at the 106th Intes-
national AOAC Annual Meeting, 31 August-3
September 1992, Cincinnati, OH).

PARK, D.L.. GAMBOA. P.M. & GOLDSMITH, C.H.
1992b, Rapid facile solid-phase immunobead
assay for screening ciguatoxic fish in the marke!
place. Bulletin de la Société de Pathologie Exati-
gue 85; 504-507,

PARK, D,L,, JURANOVIC, L.R, & MANTEIGA, R,
in press. Toxic/mutagenic potential of toxins
produced by Gambierdiscus toxicus and
Prorocentrum concavuen. Journal of Aqnatic
Food Product Technology,

RAGELIS, E,P, 1984. Ciguatera seafood poisoning
overview. Pp, 22-36. In Ragelis, E.P. (ed.),
‘Seafuexl Toxins’ (American Chemical Society;
Washington D.C).

RAGELIS, E,P. 1987. Seafood Toxins. Journal of the
Association of Analytical Chemistry 70: 285—
287.

RAGELIS, E.P. 1988. Seafood Toxins. Journal of the
Association of Analytical Chemistry 71: 81-83

RIEGELMAN, R.K, & HIRSCH, R-P. 1989. Diagnos-
lic diserimination of tests. Pp. 151-163. In R.K.
Riggelman & R.P. Hirsch, (eds) ‘Studying a study
and testing a test’. (Little, Brown, & Co.; Boston),

SAWYER, P., JALLOW, D., SCHEUER, P., YORK,
R.. MCMILLAN, J., WITHERS, N., FUDEN-
BERG, H. & HIGERD, T. 1984. Effect of
ciguolers-associated toxins on body tempernture
in mice, Pp.321-329, In E. Ragelis, (ed.) “Seafood
taxins’. (American Chemical Society:
Washington, D.C.),

TACHIBANA, K. 1980. Structural studies on manne
toxins. Unpubl, Ph.D, Thesis, University of
Hawait

TINDALL, D.R., DICKEY, R.W,, CARLSON, R.D.&
MOREY-GAINES, G. 1984. Ciguatoxic dino-
flagellates from the Canbbean Sea, Pp.225—240,
In E. Ragelis, (ed.) ‘Seafood toxins’. (American
Chemical Society: Washington D.C.).

TODD, E.C.D,, MACKENZIE, J.M., HOLMES,
C.F.B.. KLIX, H. & PARK, D.L. 1992. Com-
Parison between the mouse bioassay. the proscin
phosphatase inhibition bioassay and the solid:
phase immunobead assay for detection of
ciguatoxic potential in tropical fish. Presented al
4th Intemational Conference on Ciguatera Fish
Poisoning, May 4-8, Papecte, Tahiti, French
Polynesia.

VERVOUX, JP. this memoir. The mouse ciguateyxin
hivassay: Directions for ase,

WORLD HEALTH ORGANIZATION, 1984. Aquatic
(manne and freshwater) biotoxins, Environmental
Health Critetia 37,

YASUMOTO. T. 1985. Recent progress in the

594 MEMOIRS OF THE QUEENSLAND MUSEUM

chemistry of dinoflagellates. P.259. In D.M. YASUMOTO, T. & MURATA, M. 1988b. ‘Polyether

Anderson, A.W. White, & D.G. Baden, (eds), toxins implicated in ciguatera and seafood
‘Toxic dinoflagellates’. (Elsevier: New York). poisoning’. (Faculty of Agriculture, Tohoku
YASUMOTO, T., HASHIMOTO, Y., BAGNIS, R., University: Tsumidori).
RANDALL, J.E. & BANNER, A.H. 1971. YASUMOTO, T., RAJ, U. & BAGNIS, R. 1984.
Toxicity of the surgeonfishes. Bulletin of the ‘Seafood poisoning in tropical regions’.
Japanese Society of Scientific Fisheries 37: 724— (Laboratory of Food Hygiene, Faculty of Agricul-
734. ture: Tohoku University).
YASUMOTO, T. & MURATA, M. 1988a. ‘Polyether YASUMOTO, T. & SCHEUER, P.J. 1969. Marine
toxins produced by dinoflagellates’. (Faculty of toxins from the Pacific-VIII ciguatoxin from

Agriculture, Tohoku University: Tsumidori). moray eel livers. Toxicon 7: 273-276.

CIGUATERA POISONING: CURRENT ISSUES IN LAW
JOHN PAYNE

Payne, J. 1994.08 01: Ciguatera poisoning: current issues in law. Memoirs of the Queensland
Museum 34(3); 595-599, Brisbane. ISSN 0079-8835.

The current situation with regard to liability under Queensland law relevant to ciguatera
poisoning is reviewed. [tis argued thatall sectors of the fishing industry should be acquainted
with their responsibilities under common law and under the statutes of Workplace Health &
Safety Act, Trade Practices Act, and Sale uf Goods Act to prevent litigation in the event of

un incident.

John Payne, Peter Channel & Associates, GPO. Box 3114, Brisbaite, Queertsland 400];

22 Navember, 1993,

Areas of Queensland law relevant to ciguatera
poisoning are (i) Liability pursuant to Common
Law and (ii) Statutes (Workplace Health &
Fite Act, Trade Practices Act, Sale of Goods
Act).

LIABILITY AT COMMON LAW

Liability at common law can be based on
breach of contract, Tort or Statute. Breach of
Statute will be dealt with below. Breach of con-
tract, usually in the form of breaching implied
duties of care, gives rise, to the extent that privity
of contract allows, to similar duties to that which
urise in tort liability. Tort, or civil wrong. is based
ona concept of a duty of care. For an action to lie
in tort, three elements are required to be proven:
|) damage, 2) a relationship of proximity, and 3)
want of reasonable care, in circumstances of
foreseeable risk.

The starting point is the case of Donaghue v.
Stevens', which involved purchase by Donaghue
of a bottle of ginger beer, in circumstances where,
due to the bottle being opaque, the contents of the
bottle could not be seen, The bottle in fact con-
tained the remains of a decomposed snail which
fact was not ascertained by Ms Donaghue until
ufter she had consumed the contents of the bottle.
She was not the original purchaser of the bottle,
which had been bought by a friend and arguably
no contractual relationship existed as between her
and the maker. In the leading decision Lord Atkin
held

‘you must take reasonable care to avoid acts or
omisstons which you can reasonably foresee
would be likely to injure your neighbour - who,
then, in law is my neighbour. The answer seems
te be - persons who are so closely and direetly
affected by my act that | ought reasonably have
them in contemplation as being so affected when

Tam directing my mind tothe acts or omissions
which are called in question.’

In applying this principle to ciguatera poison-
ing, the first element to consider is the question
of proximity or to whom is the duty owed. Based
on the ‘neighbour’ principle of Lord Atkin, it
would be any person whom the provider of fish
ought reasonably have in contemplation as likely
to be affected. This would include the eventual
consumer, Whether or not that person be the pair-
chaser of the fish. The relationship vis-a-vis the
consumer would be: commercial catcher, mark-
eter, vendor (fresh), and provider (prepared).
Each of these (individual or corporate) would
owe a duty of care to the consumer of the fish.

The duty is to protect from foreseeable risk of
harm. To determine whether or not that duty has
been breached, consider:(1) whether there wias it
foreseeable risk of injury; (ii) whether the
foreseeable risk gave rise to the injury - causation;
Gii) whether the foreseeable risk could be
prevented; and (iv) whether the foreseeable risk
should, in all the circumstances, be reasonably
prevented.

The standard by which the test of breach is
measured is that of the reasonably prudent per-
son’, which in respect of ciguatera would be ‘the
reasonably prudent commercial catcher.
marketer, vendor or provider’.

Whether a risk of injury is, or is not, foresee-
able, depends on the circumstances of an ine
dent. In the Wagon Mound No. 25, it was held
"a person must be regarded as negligent if he
does not take stejs to eliminate a risk which he
knaws and ought to know ts a real risk and not a
mere passibility which would never influence the
mind of a reasonable man.’

In respect of causation, it must be the foresee-
able nisk which gives rise Lo the injury, This does
not mean that the ‘precise’ injury must be

aN

foreseen but rather the general nature or calegory
of the injury, i.e. strain, break, poisoning’.

The case of McLean v. Tedman? deals with the
issue of prevention. That case dealt with the sys-
tem of work adopted by garbage collectors and
provided that once the collector had raised an
alternative system of work (which could have
been adopted and so avoided risk of injury), that
it was up to the employer to establish that such
system would not work in the circumstances of
the case.

Finally, the Court will need to decide whether
or not, in all the circumstances of a mutter and
where the three elements of forseeability, causa-
tion and prevention have been made oul, as to
Whether or not there has been a failure to provide
reasonable care. Jn reaching ils conclusion the
Court will, inter alia, consider matters such as:
senousness of the nsk, i.e. ils potential to harm:
effect on the person upon whom the duty 1s cast,
i.e. whether or not it will unwarrantedly impede
the process of industry; and cost of implementa-
tion or effect of implementation or alternatives

Further, in reaching its decision and consider-
ing the elements of breach, the Cour will have
regard to practical matters such as: prior com-
plaint. state of general knowledge and/or
specialised knowledge on the issue of risk, what
steps have been taken to investigate and eliminate
risk, whether risk of a similar nalure or magnitude
has been removed or otherwise deale with,
whether or not subsequent to injury an alteration
has been made. and custom and practice within
the industry.

From the layman’s perspective of ciguatera, the
following elements are discernible: species of
fish, location of breeding ground, range of
symploms - from mild to serious, prohibition on
species/breeding ground, incubation period for
onset of symptoms, sensitisation to the ciguatera
toxin, increased toxicily im respect of certain parts
of the fish, size of fish, weatment. inability to
detect.

These factors need to be considered 1p the light
of an individual experience to determine whether
ar not liability will be incurred. For example. if a
commercial catcher of fish sold Red Bass, caught
anywhere in Queensland, or narrow burred
Spanish mackere!, caught off Platypus Bay, then
there is no doubt he would he liable in Tort to any
person who consumed the fish and became
symptomatic. Equally. it may be the case thatauny
provider of prepared fish wha provided as part of
a seafood restaurant menu, barracuda liver taken
only from large barracuda, would be liable. This

MEMOIRS OP THE QUEENSLAND MUSEUM

case would, of course, be dependent on the state
of knowledge. However, it should be remem-
bered that if is not the individual's state of
knowledge that is relevant, but rather the state of
knowledge of the reasonably prudent provider
etc. It is arguable that a seafood provider in
Queensland should be aware of the existence of
ciguatera poisoning and its likely causation, This
is especially so given the regular press coverage
given to the subject and the existence of ap-
propnate Departmental information.

At the other end of the scale, it is probably not
arguable that liability would accrue to a provider
who sold barramundi which in turn lead to
symptoms of ciguatera poisoning. This is par-
ticularly so, noting the low incidence of ciguatera
poisoning linked to the species where only one
case 1s ascribed during the period 1965-1984".

Somewhere between these extremes will, of
course, be the grey area of concern to the industry.
Forexample. the selling of narrow barred Spanish
mackerel which has. been linked with 226 cases
in Queensland between 1965-1984, This is of
more concern when Gillespie, Lewis et al. (1986)
statements are considered: “a large number af
cases of Ciguatera are not reparted to health
uuthorilies, so the trae incidence of ciguatera is
difficult te assess’ and ‘Whether these figures
reflect a trend towards an increasing incidence
of ciguatera or increased public awareness is not
known, hur it is certain that the abovementioned
Reports represent only a@ proportion of the out-
breaks that have occurred’.

Therefore given what appears to be a relatively
high incidence of ciguatera cases/outbreaks as-
sociated with narrow barred Spanish mackerel,
given the potential of ciguatera poisoning to
cause severe health problems and given that
species such as Red Bass which are known to
cause risk in other Pacifie countries (but which
have been involved in few reported cases local-
ty)!" are prohibited, then it is arguable that if the
consumption of commercially caught narrow
barred Spanish mackerel gave nse to ciguateri
poisoning, that liability would accrue. This may
nat be that clear as, for example, it may be the
case that a professional fisherman could argue
that narrow barred Spanish mackerel were only
of concern if caught, for example, off Fraser
Island, This is a matter Which depends on its own
facts und will be clanfied as research continues.

There may well be arguments as to why itis not
Trexsonable to remove narrow barred Spanish
mackerel from the catch in areas other than the
zones ef concem, i.e. Cairns/Townsyille, Rack-

CIGUATER A - CURRENT ISSUES IN LAW

hwmpton and Fraser, orto remove certain sizes of
catch. Such arguments are plausible, but it is
emphasised are dependent on the relevant facis
and level of knowledge.

In summary, to assess whether or not liability
accrues in any given circumstance, it is necessary
to: (i) show that a relationship of proximity exists.
{ti) show that there is a failure or want of
reasonable care, by demonstrating that there was
a foreseeable nsk of injury, which gave rise to the
type of damage which was foreseen, which could
reasonably have been prevented, and (iii) the
consideration of whether or not such breach has
occurred will be dependent on the facts and cir-
cumstances of the poisoning and the events that
precede it.

The author considers itineyitable that there will
be successful Ltigation in respect of ciguatera
poisoning. It is simply a matter of time.

WORKPLACE HEALTH & SAFETY ACT-
DUTIES OF CARE

The most important changes to Occupational
Health & Safety in Australia, are the introduction
of Robens-style legislation. Robens’ legislation
is based on the self regulation of Occupational
Health and Safety in workplaces as opposed to
regulation by way of sanction imposed from out-
side the workplace.

The Queensland Workplace Health and Safety
Act, which is the embodiment of the Robens
model, was assented to on 12th May, 1989, with
Section 6, 36 and $7 commencing on | 0th June,
1989 and the remaining provisions commencing
on 31st July, 1989'', Regulations were enacted
and commenced on 31st July, 1989, excepting
regulations dealing with diving, which com-
menced on 30th October, 1989'*. This Act
amended, or repealed the Construction Safety
Act, the Inspection of Machinery Act, the Health
Act and the Shops and Factories Act. It is now the
Act dealing with Occupational Health & Safety
for the great majority of Queensland workers.

Since enactment there have been substantial
amendments to both the Act and the Regulations.
The most significant amendment being the in-
clusion of the rural industry within the parameters
of the Act by amendment in 1990.

Central to the Robens’ model are: duties of
care, internal nha te assessment, and broad
based prescriptive alternatives, i.e. codes of prac-
tice,

The duty of care is expressed as a legislative
formula in the Act: ‘an employer wha fails to

397

ensure (he heulth and safety al work of all the
employers, employees, save where it is net prac-
ticable for the employer to dea so, convnits an
offence against this Act..'*_

With the definition of practicable in the Act
being, praciecable, means practicable having
regard io:-

(a) the nature of the employment of, as the case
may be, the particular aspect ef the employment
concerned; and

(b) the severity of any potential injury or herm
to health or safety that may be invalved, and the
degree of risk that exists Mm relation to sueh poten-
Nal injury er harm, and

(c) the state of knowledge about the injury or
harm to health or safery that may be involved,
about! the risk of that injury or harm to health: or
safety eccurring and about any ways of preveni-
ing, remieving ar mitigating that injury, harm or
risk; and

(ad) the availability and suitability of ways te
prevent, remave or mitigate that injury or hearre
to health ur safety or risk; and

(¢) whether the cost of preventing, removing or
mitigating that injury or hana te health or safety
ar that risk is prohibitive in the circumstances."

This duty reflects broadly the common law
principle of the duty of care which has evolved
through personal injuries case law and Was enun-
ciated by Lord Atkin in the case of Domaghue v.
Stevens,

Section 9 provides for the duty of care and
impases the relationship in respect of employers
and employees. The Act, however, docs not sole-
ly relate to workplace health and safety, but ex-
tends well beyond what is perceived to be the
employment relationship. This is the result of the
origins of the Workplace Health & Safety Act!®.

Of particular relevance to the commercial fish-
ing industry is Section LO of the Workplace
Health & Safety Act; ‘(2) An employer who fails
to conduct his or her undertaking in such a man-
ner as to ensure that his ar her own health and
safety and the health and safety of persons nor in
the employer's employment and members of the
public who may be affected are not exposed to
risks arising from the conduet of the employer's
undertaking, except where {tts not practicable far
the employer te do se, commits.an offence against
this Act.’

It is clear thal this would include commercial
catcher, marketers, vendors (fresh), and the
provider ( prepared).

The definition of practicability applies to Seo-
tion 10 and the terms of the definition should be

598

considered. Basically, practicability provides for
a similar test as is used for breach of duty of care
under tort. The pnncipal difference is that there
is no requirement for causation. Thal is, no injury
needs to occur for there to be a breach of the
Workplace Health & Safety Act. That means that
if a risk exists which could be reasonably
removed and ought to be reasonably removed and
is oot so removed, then an offence occurs.

Matters discussed above in respect of the
breach of duly of care under tort, i.e, forseeahility,
preventability and reasonableness are equally ap-
plicable toa consideration of practicability. There
are however, in the writer's view, some essental
differences between the duty owed pursuant to
tort and the duty under the Workplace Health and
Safety Act. These duties arise from the fact that
the Workplace Health and Safety Act is a quasi
criminal act!’, which means ils provisions must
be sinctly construed to the benefit of the in-
dividual against whom the sanction is imposed.
This 1s of parucular relevance to the last elenient
of practicability which, as noted above, is: “(e)
Whether the cost of preventing, removing or
mitigating thar injury or harm to health or safety
of that risk is prohibitive in the circumstances,

This on a strict constriction should be con-
sidered in the light of the abilities of the in-
dividual commercial catcher, marketer, vendor
and provider to meet that cost. In other words,
rather than an application of the test of the
reasonable person, the individual should be con-
sidered.

Nonetheless, given the industry's knowledge of
ciguatera poisoning and potential nsk of harm to
members of the public the Act may well have
becn breached.

Breach of the Workplace Health & Safety Aci
will also support an action at common law for
damages. Further, the penalties range from a fine
of $3,000 or 6 months imprisonment for a person
{other than a body corporate) where the Act is
contravened to a fine of $30,000 or 6 months
imprisonment where a death or serious hodily
injury occurs (again this ts for a person other than
a4 body corporate), Offences for bodies corporate
range from $12,000 to $120,000"". By Section
124 of the Act, a person who is a managing
director or other governing officer of who at any
lime acts or takes part in the management, ad-
ministration of government of the business in
Queensland of a body corporate can be liable to
punishment by imprisonment.

In my view the industry must comply with
provisions of the Workplace Health & Safety Act

MEMOIRS OF THE QUEENSLAND MUSEUM

and should consider its position vis-a-vis whether
or not it is practicable for the risk to be removed,

TRADE PRACTICES ACT &
SALE OF GOODS ACT

When a consumer purchases an item from a
retailer there is an oral contract and into this oral
contract certain terms are implied by law, These
implied terms are measured in law to balance the
relationship between retailer and consumer to

ect the consumer from the ‘caveat emptor’
(i.e, buyer beware) principle.

The implied terms of the contract are either
conditions or warranties. Basically, a condition js
a vital or fundamental term whereas a warranty 1s
a collateral or subsidiary term.

Where there is a breach of a condition the
consumer can return the goods, get a refund and
sue for compensation for any loss suffered as a
consequence of the breach. Where there is a
breach of warranty on the other hand the con-
sumer cannot retum the goods and get a refund,
but the consumer can sue for compensation for
any loss suffered as a consequence of the breach.

Therefore, if a retailer breaches a condition or
a Warranty, he/she is exposing himself/herself to
an action by the consumer for compensation for
any loss suffered as a consequence of the breach.

Certain conditions and warranties are implied
by the Commonwealth's Trade Practices Act and
the States’ Sale of Goods Act. Most significany in
the current context is the implied condition that
the goods be of merchantable quality.

Under Section 66(2) of the Trade Practices Act,
goods are of merchantable quality *...if they are
Jit for the purpose ar purposes for which geods of
that kind are commonly bought as it is reasonable
to expect, having regard to any description ap-
plied te them, the price (if relevant) in all other
circumstances.’ A similar definition is provided
by 817(2) of the Queensland Sale of Goods Act,

Basically, to be of merchantable quality the
goods must:-(a) pass without objection in the
trade and description given to them in the con-
tract, and

(b) be of fuir and average quality within the
description, and

(c) be fic for the usual purpose for which such
goods are used, and

{d) one with variations allowed by the agree-
ment of even kind, quality or quantity within each
unit and among all units, and

{e) be adequately contained, packaged and
labelled, and

CIGUATERA — CURRENT ISSUES IN LAW

(f) conform to the promises or affirmations of
fact made on the label or container.

The implied condition or merchantable quality
does not apply where the defects are brought to
the consumer’s attention prior to sale or where
reasonable examination of the product occurs and
the defects ought to have been revealed by this
examination. This situation is of course unlikely
to occur in relation to ciguatera affected fish.

It could be argued that where ciguatera affected
fish is sold it may not be of merchantable quality
as it would not be of fair and average quality
within the description and would not be fit for its
usual] purpose, i.e. human consumption. A person
would then arguably sue for the damage that has
been suffered.

Where the goods are not of merchantable
quality, the retailer may be exposing him/herself
to a suit by the consumer for compensation for
loss suffered as a consequence of such breach.
These rights of redress are of course only avail-
able to the purchasers of the affected fish.

CONCLUSION

Itis arguable that action could be taken against
industry members in relation to ciguatera poison-
ing, either by suit under common law, by prosecu-
tion under the Workplace Health & Safety Act,
or action pursuant to the Trade Practices or Sale
of Goods Act. I suggest the industry should be
pro-active and consider what steps can be taken
to address potential liability.

599

LITERATURE CITED

1.[1932] AppealCase 562
2.Glasgow Corperation v. Muir [1943] 1 AC
447

3.[1967] AC 617.

4.See generally Hamilton v. Nuroof (1956) 56
Commonwealth Law Reports 18.

5.[1984] 155 CLR 306.

6.See generally General Cleaning Contractors
vy. Christmas [1953] AC 180, Read v. J. Lyons &
Co Limited [1947] AC 156.

7.Table 3, Ciguatera in Australia, Vol. 145
Medical Journal of Australia, December, 1986,
p. 584, Gillespie, Lewis et al.

8.ibid,

9.ibid, p. 586.

10.ibid, p. 587.

11.Ss 1 & 2 commenced on the day of Assent.

12.Regulations 1 & 2 commenced on 29th July,
1989,

13.See Section 34 of the Workplace Health &
Safety Act.

14.See Section 9 of the Workplace Health &
Safety Act.

15.See Section 6 of the Workplace Health &
Safety Act.

16.The Queensland Workplace Health and
Safety Act, M. Quinlan, T. Farr, J. Payne, Journal
of Occupational Health & Safety - Australia, New
Zealand, 1989, 5(3), p. 265 - 274.

17.See Section 118, 119 and 124.

18.See Section 118 & 119 and for definition of
serious bodily injury see Section 6.

600

STRUCTURES OF MAITOTOXIN AND
CIGUATOXIN CONGENERS ISOLATED FROM CUL-
TURED GAMBIERDISCUS TOXICUS. Memoirs of the
Queensland Museum 34(3): 600. 1994:— Maitotoxin (MTX)
was isolated from cultured cells of Gambierdiscus toxicus
from the Gambier Islands (GII1 strain). To determine the
structure, the toxin was cleaved into 3 fragments (A, B, C) by
sodium periodate oxidation, followed by sodium borohydride
reduction. Structures of fragments A and B were determined
by 2D NMR experiments, The structure of fragment B, the
largest fragment of 2306 Dalton, was negative FAB MS/MS
experiments, Comparison of the spectra between the frag-
ments and intact MTX allowed us to assemble the whole
structure of MTX. MTX has molecular weight 3422 (nominal,
as disodium salt) and is constructed from 142 carbon chain,
comprising 32 ether rings, 21 methyls, one exomethylene, 28
hydroxyl groups, and two sulfate esters,

MEMOIRS OF THE QUEENSLAND MUSEUM

Two ciguatoxin (CTX) congeners, CTX3C and CTX4A,
and a new polyether toxin named gambierol were isolated
from the culture of G. toxicus from Rangiroa Atoll (GRI1
strain). CTX4A is 52-epiCTX4B and CTX3C is 1,2,3,4-nor-
E-homo-CTX4B. The ladder-shaped polyether skeleton of
gambierol differs from the other two. Production of CTX4A
and CTX3C, by cultured G. toxicus unambiguously confirmed
the generic origin of ciguatera toxins.

Takeshi Yasumoto, Masayuki Satake, Michio Murata, Faculty
of Agriculture, Tohoku University, Tsutsumi-dori
Amamiyamachi, Aoba-ku, Sendai, 981 Japan & Hideo Naoki,
Suntory Institute for Bioorganic Research, Wakayamadai,
Shimamotocho, Osaka 618, Japan; 12 April, 1993.

CIGUATERA: DILEMMAS IN CLINICAL RECOGNITION,
PRESENTATION AND MANAGEMENT

JOHN PEARN

Pearn, J. 1994 08 01; Ciguatera: dilemmas in clinical recogmbon, presentation and manage-
ment. Memoirs of the Queensland Museum 34(3): 601-604. Brisbane, ISSN 0079-8835.

Both the clinician and consulting scientist are confronted with several key problems in the
recognition and management of the ciguatoxic victim. Failure to cansider the possibility of
Ciguatera in a patient presenting with any one or more of the pleomorphic constellation of
symptoms and signs which are the hallmark of the discase, remains the most important
ongoing dilemma of management, A differential diagnosis involves ‘formulation of a list of
diseases, commensurate with the clicited history and the observed signs, arranged in
decreasing order of likelihood’, In mild single cases the difficulty of raising a differential
diagnosis is compounded by lack of some symptoms. Another dilemma is interpretation of
the chronicity of symptoms and this remains a clinical research challenge. A further dilemma
is use of Mannitol and timing its introduction. Clinical research shows that Mannitol is. not
effective if administered more than 48 hrs after symptoms appear.

John Pearn, Department of Child Health, University af Queensland, Royal Children's
Hospital, Brisbane, Queensland 4029; 10th May, 1994.

The pleomorphic nature of ciguatera, the sub-
jectivity of many ofits symptoms and the absence
of any definitive laboratory diagnosis for clinical
cases make this condition one of the most chal-
lenging in clinical medicine.

The research dilemmas of ciguatoxin clas-
sification, source, assay and lesion pathogenesis
ure paralleled by clinical dilemmas of diagnosis,
symptom interpretation and management. In the
evolution of the understanding of any human
disease there exists a ‘wmdow of time’ in which
one has to make the best of all available clinical
experience, however anecdotal and however im-
perfect, in the practical management of an in-
dividual victim, In the case of ciguatera we are
hopefully nearing the end of this era of clinical
empincism. Recent identification of the mole-
cular structure of several] of the ciguatoxins
(Murata et al,1990), advances in understanding
of sodium channel pathophysiology (Benoit et
al.,1986: Lombet et al..1987) and unequivocal
histological evidence of nerve and muscle chan-
ges all contribute to the better interpretation of
the miscellany of symptoms and signs which is
the hallmark of this ‘treacherous and increasing-
ly-occurring marine fish public health hazard’
(Russel & Egan,1991),

Many clinical dilemmas remain. These uncer-
lainlies are perplexing for the physician but like
all dilemmas their resolution will advance the
understanding of this enigmatic, common and
important disease.

DIFFERENTIAL DIAGNOSIS

Diagnosis of ciguatera is essentially clinical.
Currently. it ts the failure to consider the pos-
sibility of ciguatera, in a patient presenting with
any one or more of the pleomorphic constellation
of symptoms and signs which are the hallmark of
this disease, which remains the most important
ongoing, dilemma of management. This fact, the
overlooking of the possibility of ciguatera rather
than any omission of documenting the symptoms
and signs remains the major problem in the
management particularly of sporadic cases.

One dilemma is that there is no published work
on the proportion of sporadic versus multiple
cases in any published case series, Although the
clinical syndrome is now very well defined (Gil-
lespie et al.,1986) the syndrome boundaries for
subacute and chronic cases still remains uncer-
tain. The ‘gold standard’ of the chronic ciguatera
syndrome must include case studies of multiplex
{or epidemic) cases, followed prospectively.

The concept of differential diagnosis is ‘the
formulation of a list of diseases, consistent with
the elicited history and the observed signs, ar-
ranged in decreasing order of likelihood’. All
familiar with ciguatera are aware of the mulu-
plicity of other different diagnoses which are
included in the list of possibilities generated by
the perplexed victim and his or her family, by the
attending first aider, or by the admitting doctor in
ihe emergency room of the referral hospita),

Differential diagnosis, in sporadic cases, in-
cludes such conditions as viral and bacterial
enterocolitis, viraemias of diverse types, some
types of hypersensitivity reaction, poisoning with

© organic and inorganic agents and various
types of neuroses,

Bacterial and viral gastroenteritis can be ac-
companied by prostration, rash, arthralgia and
myalgia and bradycardia. Viral infections can
calse puzzling constellations of symptoms and
signs including rashes, arthralgia and myalgia,
pastro-intestinal disturbances and neurogenic
paraesthesiae. In the pasi, patients with un-
doubted ciguatera have been labelled as suffering
from chronic Viral diseases, auto-immune dis-
ease, possible insecticide and heavy metal
poisoning, psychosis and neurosis, hysteria and
malingenneg.

Subacute and chronic cases, or cases presenting
for the first time after several days of symptoms,
are always difficult to diagnose. A partecular dif-
ficulty is the fact that loss of energy. loss of
uppetite and subjective feelings of weakness are
very common indeed in the general population.

The differential diagnosis of ciguatera is al-
Ways a two-stage process. The first stage ts lo
deduce that one of the icthyosarcotoxacinias is
present; the second is to run through the other
possibilities of puffer fish poisoning (fugu),
maitotoxaemia, clupeotoxism and histamine
poisoning, This latter condition, due to histamine
poisoning, occurs especially after the ingestion of
spoiled Pomatomus, of common ‘tailor’ fish of
eastem Australia; and occasionally after the in-
gestion of Arripisor ‘Wester Australian salmon’
(Smart,1992). The differential diagnosis of the
icthyasarcatoxaemias also includes the vanious
forms of diarrhoeal and paralytic shellfish
poisoning, especially after the ingestion of mixed
seafood meals which include both potentially
loxic species such as coral trout (Plectropomus
maculata) and mackerel (Scoamberamorus com-
mersant) together with oysters and scallops.

The author has encountered several cases
presenting first following rechallenge with
ciguatoxic food one case involving the ingestion
of battery-fed chicken, in which the poultry was
probably fed on fish meal. In this type of case the
diagnosis of the putative original cigualoxic in-
toxication can only be made jn retrospect.

CHRONICITY OF CIGUATERA

One of the main clinical dilemmas ts interpret
ing the true significant of chronic symptoms.

MEMOIRS OF THE QUEENSLAND MUSEUM

How long can ciguatera last? Most experienced
workers have followed cases prospectively and
know that objective signs of poisoning usually
persist for a few days or several weeks only; yet
all know that the subjective, often distressing
symptoms such a5 prostration, arthralgia and
myalgia and disordered cutaneous sensation can
persist in an unbroken continuum of such subjec-
tive symptoms for many months. Can ciguatera
produce symptoms, say, after two or three years?
At tts stage of scientific knowledge there are
numerous anecdotal case reports, but doubt per-
sists ubout the true persistence of symptoms for
more than one year or so. At this point of scien-
tific endeavour. no cumulative frequency his-
tograms have been generated, by symptoms, for
proven cases followed prospectively. Thus, the
chronicity of ciguatera remains an tmportant
clinical research issue for the future.

Recent neurophysiological experiments have
indicated that the toxin is acting at its affector
sites, in organ-bath preparations in fractions of
nanomolar concentrations. This fact. combined
with its fastness in some neurophysiological ex-
periments lends plausible support to the concept
that true symptoms may persist for years rather
than months, The principal target of ciguatoxin is
on unmyelinated fibres. [Lis not implausible that
one of the most toxic substances known to science
{ciguatoxin), and one of such demonstrated
strong attachment to its receptor site in the
sodium channel, might produce bizarre auto-
nomic-related symptoms for very long periods
after the initial insult, Permanent damage to ner-
ves, or residual binding of the toxin fo its target
receptors, may help explain the often observed
phenomenon of recrudescence of symptoms,
even in the face of an otherwise subclinical dose
of Toxin,

INDIVIDUAL SUSCEPTIBILITY

There is considerable individual clinical sus-
ceplibility to ciguatoxin, Not infrequently, dif-
ferent family members eating the same toxic fish,
and often apparently in similar amounts, are af-
fected to different degrees. The mass of toxic fish
eaten is obviously important; and in the case ef
very toxic fish even small differences in plate
portions may be reflected in large differences in
the mass of toxin which is ingested. Experience
in Japan with fugu fish poisoning is that the eating
of very large portions of otherwise relatively safe
fish has resulted in fatalities (Matsubara,198 1).

CIGUATERA — CLINICAL RECOGNITION, MANAGEMENT

Such cases highlight particularly the importance
of portion size - and conversely, the need for
prudence in the face of potentially risky meals.

Personal clinical experience with managing
multiple affected victims who have eaten from
the one ciguatoxic fish suggests that individual
clinical variation is the rule, rather than the excep-
tion, All experienced workers have encountered
situations where some members may be totally
unaffected following the ingestion of a ciguatoxic
fish meal, whilst others cating portions of similar
size may be severely affected. Research
biologists undertaking the mouse assay for
ciguatoxin, also encountered this in a situation
where pairs of mice were being used in the
hiological assay_ Not infrequently one member of
the pair will be dead within 1-3 hours and the
other (although usually affected) will survive.
These ‘mouise-splits’ so often parallel the clinical
discordance one sees among the human victims
of mini-epidemics.

The basis for this vartable susceptibility
remains unknown. A significant genetic com-
ponent is likely although, even within affected
families (in family outbreaks), there is not infre-
quently widespread variation in the severity of
symptoms and objective signs, Different species
react differently to the toxin, both in terms of
quantilative response as crude evidence of
poisoning on the one hand, and in qualitative
syndromuc variation on the other. The ‘straub tail”
seen in poisoned mice is quite different from the
syndrome seen in the (more sensitive) afflicted
cat, often used as the practical test animal in real
life domestic situations where a family is wishing
lo consume a risky species of fish.

Some believe that children are particularly sus~
ceptible and certainly in various LDso assays for
other toxins, neonatal mice are more sensitive
than the standard 19-21 gram adults which are
more traditionally used in the specific ciguatoxic
mouse assay, | have encountered clusters of fami-
ly poisonings where children appear to be more
severely affected. The dilemma remains however
that children se often ingest more of the fish, and
in a particularly toxic fish meal a relatively small
increase in ingested mass (in relation to a child’s
body weight) may result im a supra-threshold
level of ingested toxin. Similarly, sex differences
in responses to the toxin are often hinted at,
anecdotally in the case of women whom it is
thought may be particularly susceptible to the
long term effects, No formal atterpts at initial
dose quantification, with long term follow up by
sex, have been reported,

65

GEOGRAPHIC VARIATION
SYMPTOMATOLOGY

Confusion exists abow the relative incidence of
different symptoms in different parts of the
world. Whilst all case series report such things as
circum-oral tingling, diarrhoea and vomiting,
other symptoms such as dysuria (Gillespie et al.,
1986), dental pain, pruritus and piloerection are
reporled much more frequently in certain
geographic regions than in others. Some differen-
ces are undoubtedly due to sampling errors, dif-
ferences in case descriptions and different
standards of history taking and of reporting,
However, there are obviously different toxins and
different toxin subtypes in different areas. In-
deed, it seems inescapable that the human clinical
syndrome of ciguatera is the result of ingestion of
a cocktail of different ciguatoxins. A priori, it
would be unrealistic not to expect different clini-
cal syndromes under these circumstances, in dif-
ferent parts of the world, There is some evidence
that antibody profiles to toxims from fish taken
from different parts of the world differ in their
cross-reactivily, This gives further credence to
the belief that there are subtle differences in
ciguatera syndromes in different parts of the
world. Nevertheless, in all reported series, a
profile of core symptoms is seen and includes
gastrointestinal symptoms, neurological com-
plications such as paraesthesiae and temperature
dysaesthesiae, myalgia and arthralgia. This also
reflects different case definitions which are used.

The role of mannitol therapy (Palafox et al.,
1988; Pearn e2 al.,1989) remains indeterminate,
although the necessary double-blind study (from
the Marshall Islands) is in progress. Inthe writer's
experience, administration of intravenous man-
nitol in a dose of lg/kg body weight, given as an
oedema-reducing regimen over a maximum ad-
ministration time of 45 minutes, produces dra-
matic alleviation of symptoms within 2-3 hours
in some patients. The role of mannitol therapy in
cases presenting to medical altention afler this
time remains contreversial and this dilemma will
not be resolved until treated cases are followed
prospectively, L give mannitol, in cases present-
ing acutely even although the symptoms may be
milder, in the anticipated belief that the risk of
long term sequelae will be reduced, What has
been established is that mannitol given to the
Ciguatera patients js safe, and that no synergism
between toxin and mannitol has been observed.

To the clinician practising in high-nsk endemic
regions of the tropical littoral, multiple-case out-

604

breaks pose no problem in diagnosis and with the
advent of mannitol therapy management is much
more straightforward. The major problem in the
clinical management of ciguatera remains in the
need for more widespread awareness of the pos-
sibility of the disease, and earlier diagnosis. To
the first aider, nurse or physician encountering
(particularly sporadic) cases, often distant in
place and sometimes distant in time from the fish
source, missed diagnosis still remains the biggest
challenge in the management of this important
disease.

LITERATURE CITED

BENOIT, E., LEGRAND, A.M. & DU BOIS, J.M.
1986. Effects of ciguatoxin on current and voltage
clamped frog myelinated nerve fibre. Toxicon 24:
356-362.

GILLESPIE, N.C., LEWIS, R.J., PEARN, J.H.,
BOURKE, A.T., HOLMES, M.J., BURKE, J.B.
& SHIELDS, W.J. 1986. Ciguatera in Australia.
Occurrence, clinical features, pathophysiology
and management. Medical Journal of Australia
145: 584-590.

LOMBET, A., BIDARD, J.-N. & LAZDUNSKI, M.
1987. Ciguatoxin and brevetoxins share a com-
mon receptor site on the neuronal voltage-depend-

MEMOIRS OF THE QUEENSLAND MUSEUM

ent Na* channel. Federation of European
Biochemistry 219: 355-360.

MATSUBARA, I. 1981. Puffer-fish, a dangerous
delicacy from the Pacific. Pp. 16-19. In Pearn, J.H.
(ed.), ‘Animal toxins and man’. (Qld Department
of Health: Brisbane).

MURATA, M., LEGRAND, A.M., ISHIBA, S.Y.,
FUKUI, M. & YASUMOTO, T. 1990. Structures
and configurations and ciguatoxin from the moray
eel Gymnothorax javanicus and its likely precur-
sor from the dinoflagellate Gambierdiscus
toxicus. Journal of the American Chemical
Society 112: 4380-4386.

PALAFOX N.A., JAIN, L.G., PINANO, A.Z.,
GULICK, J.M., WILLIAMS, R.K. & SCHATZ,
IJ. 1988. Successful treatment of ciguatera fish
poisoning with intravenous mannitol. Journal of
the American Medical Association 259: 2740—
2743.

PEARN, J.H., LEWIS, R.J., RUFF, T., TAIT, M.,
QUINN, J., MURTHA, W., KING, G., MAL-
LETT, A. & GILLESPIE, N.C. 1989. Ciguatera
and mannitol: experience with a new treatment
regimen. Medical Journal of Australia 151: 77-80.

RUSSELL, F.E. & EGAN, N.B.. 1991. Ciguateric
fishes, ciguatoxin (CTX) and ciguatera poisoning.
Journal of Toxicology - Toxin Reviews 10: 37-62.

SMART, D.R.. 1992. Scombroid poisoning. Medical
Journal of Australia 157: 748-751.

CIGUATERA: RISK PERCEPTION AND FISH INGESTION
JOHN PEARN AND RICHARD LEWIS

Pearn, J. & Lewis, R.J. 1994 08 O01: Ciguatera: risk perception and fish ingestion. Memairs
of the Queensland Museum 34(3):605-608. Brisbane. ISSN 0079-8835,

Avsurvey of 37 attendees at the Clinical Ciguatera session of the Ciguatera Management
Workshop, Bribie Island, Apri] 1993, completed a questionaire to assess nsk perception
relative to fish ingestion among @ group acutcly aware of the ciguatera threat. The perceived
tisk difference between ingestion of fish personally purchased in the marketplace and fish
served in a restaurant was assessed with responses from different groups (males/females,
clinicians/biologists) within the sample being compared. No one would accept a risk of 10%
in purchasing fish personally but one clinician would accept a risk of 20% in a seafood
restaurant and 25% of respondees would accept a higher nsk in a restaurant than in their
purchasing wnprepared fish.

Joha Pearn, Department ef Child Health, University of Queensland, Royal Children's
Hospital, Brisbane Queensland 4029 & Richard Lewis, Southern Fisheries Centre, Depart-
ment of Primary Industry, PO Bex 76, Deception Bay, Queensland 4058; 10 May, 1994 ,

Onc of the most practical questions relating to
human ciguatera poisoning is the question “What
msk will laccept before eating a fish meal?”. The
answer to this question governs fishing policy,
marketing regulations, species selection for gour-
met dining, and individual choice of menu.

Many factors influence the statistical sk of
contracting ciguatera. Fish species, size of an
individual fish, fish habitat (Lee,1980), season of
catch and size of portion all modify the intrinsic
risk of developing ciguatera (Bagnis et al,]979;
Lawrence et al,1980; Russell & Egan,1991)-

The statistical risk of contracting ciguatera
varies with country and species (Hakama et al.
1993). The risk of contracting ciguatera on
Niutao Island m Tuvalu is | in 10 (Dalzcll this
memoir). In Micronesia, the risk of ciguatera
from eating Moral eel viscera may be > | in 20.
With Hawaiian jackfish (Curanx sp., or ‘papio’)
the risk is 1 in 100 (Hokama et al,1993). In
Queensland, the risk of ciguatera is <] in 3,000
(Gillespie et al,1986) and from a meal of coral
trout (Plectropomus maculata) itis <1 in 5,000.

Individuals modify their behaviour not on the
basis of these objective figures, but on the per-
ceived subjective risk (Pearn,1973,1977). Sub-
jective risk is determined by such factors as sex
(women usually being more conservative in the
face of a gambling situation), personality (op-
timists being less conservative), past experience,
concepts of probability and the percetved out-
come including fatality (Peamn,1973). In the
specific risk of ciguatera following fish consump-
on, itis known that the objective risk of fatality
is <] in 1,000 of clinical cases in Australian

(Tonge et a], 1967) and French Polynesian reports
(Bagnis et al,1979).

An objective risk of contracting ciguatera of <L
in 1,000, with a risk of fatality of 1 in 1,000,000
for arandom fish meal, is for many alow or trivial
tisk. On the other hand, many deny themselves
the pleasures of gourmet fish meais because o!
the subjective (or perceived) risk which is seen to
be more threatening than these low figures imply.
We were interested to obtain data on this
phenomenon of perceived or subjective risk in the
context of human fish-consumption behaviour.
Besides perceptions of subjective risk, risk-
taking behaviour is known to be influenced by the
social setting in which fish is consumed, and of
course by peer influence. This has significant
implications for consumption of risk species in
seafood restaurants and in other social settings.

Dunng a Workshop on Ciguatera Management
a gathering of world experts on ciguatera par-
ticipated in a risk-assessment study of professed
personal deciston-making in the face of a
hypothetical ciguatera risk. We report here the
results of this study of professed risk-taking be-
haviour in the context of a ciguatcra threat,

METHODS

SUBJECTS

The subjects included all 37 individuals (31
males and 6 females) who attended the ‘Clinical
Ciguatera’ Session of the International Workshop
on Ciguatera Management, held at Bribie Ishiend,
Queensland. on 15th Apri] 1993, under the
auspices of the (Australian) Fisheries Research
and Development Corporation and the

MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE L. Maximum acceptance risks for buying a marketplace fish which might be ciguatoxic. Thirty-seven
world experts in ciguatera, at the International Ciguatera Management Conference, Queensland, Australia. 1993-

Queensland Department of Primary Industries.
No subject refused to take part in this study. All
were university graduates and (comprising as
they did the world leaders in this subject) all were
fully informed of the implications of ciguatera
poisoning, Of the 37 subjects, nine were practis-
ing clinicians and 25 were biological research
scientists working in this field. Subjects came
from Australia, USA, France, Japan and the
United Kingdom. All were proficient in English,

QUESTIONNAIRE

Each subject completed a persona] questian-
naire, anonymously, giving details of sex, dis-
cipline {research scientist, clinician etc) and
professed acceptance risks in the instance of two
separate question-scenarios. A brief (S$ min) ver-
hal exposition about the nature of the study was
presented by one of us (JP) prior to the comple-
tion of the questionnaire.

Each subject was asked to respond to two
specific questions:-

].Imagine you are staying In a country where
ciguatera occurs. You are buying fish in a shop or
market, to take home for yourself or your family.
What (maximum) level of risk of ciguatera would
you accept. before buying the fish?

2. Imagine you are in a seafood resiaurant, in a
country where ciguatera occurs. Fish is served. It
is a species known occasionally to be ciguatoxic.
At what (maximum) nsk level for cigzuatera,
would you eat the fish meal?

RISK SCALES

Subjects recorded their personal (subjective)
acceptable risks in 3 ways: a) by a linear (Likert)
scule marked from 0 to 100%, on which the
subject draws a line at their personal msk accep-
tance level, b} by a vulgar fraction, and c) by a
percentage figure. In each of these systems, arisk
of ‘0° means that an individual will not accept any
risk whatscever. In the context of the specific
questions we asked this implies that the in-

ACCEPTANCE RISKS
| Allfemale subjects [| 0%
[Clinicians Peto

; EMAXIMUM —|~—s MEDIAN ACCEPTANCE RISK

dividual would not eat a msk-species of fish under
any circumstances. A risk of 100% or 1.0 implies
that an individual would go ahead and consume
fish even if it was certain that the subject would
contract ciguatera from such ingestion. The
specific risk of 50% (with the appellation ‘1 in 2’
risk) was marked on the Likert scale.

RESULTS

Of the 37 subjects, all but one recorded their
professed risk-acceptance levels in each of the 3
modalities. For Question 1, about acceptance
nsks for ciguatera when buying fishin the market
place (Table 1), 7 subjects (19%) said they would
not accept any risk, and would not buy risk
species of fish for which there was any chance
whatsoever of contracting ciguatera; 13 (35%)
said they would accept a risk of 1% or greater,
that ts a risk of } in 100 or greater; 4 (3 research
scientists and | male clinician) said they would
accept a risk of 10% (1 in 10) of contracting
ciguatera.

For question 2, concerning restaurant con-
sumption of potentially toxic fish (Table 2), 8
subjects (229) said they would not eat any fish
species in a seafood restaurant, where there was
any risk whatsoever of contracting ciguatera; 8
(22%) said they would accept risks of 10% {1 in
10) or higher. This widespread difference in
professed behaviour, in the face of a medical risk,
parallels the widespread attitude to risk seen in
other medical situations (Pearn, 1973).

Analysing the individual responses to each of
the two questions revealed that 21 subjects (589%)
did not change their professed risk-acceptance
Jevel when confronted by the different social and
peer pressures. inherent in eating in a seafood
restaurant. Nine subjects (25%) professed to ac-
cept higher risks in the seafood restaurant
scenario, with a median increase in risk, in this
group, by a factor of five. In the open-ended
section of the questionnaire marked “comments’,

CIGUATERA — RISK PERCEPTION

v7

TABLE 2.Maximum acceptance risks for eating a potentially ciguatoxic fish meal in a seafood restaurant.
Thirty-seven world experts on cignatera, the International Ciguatera Management Conference, Queensland,

Australia,1993.

|___Allfemale subjects =: subjects
All Ali male subjects subjects
Clinicians
Research scientists
_All subjects

6 wrote that they would accept higher potential
risks in a seafood restaurant scenario, because of
social and peer pressures, and because of such
themes as “being an honoured guest’, or ‘good
manners in a group situation’.

DISCUSSION

This study shows that the majority of subjects,
themselves expert in ciguatera, accepted risks for
contracting the disease which were greater than
the real life objective risks around the Pacific rim
and in the Caribbean. No worker professed to
accept a planned fish-buying risk greater than
10% (1 in 10), although | individual would be
prepared to accept risks of 20% (1 in 5) of con-
tracting ciguatera from eating in a seafood res-
taurant. [t is the objective (mathematical) risk of
ciguatera which concems questions about fishing
industry policy, species prohibition and the fund-
ing for management, monitoring and research. By
contrast, subjective msk determines the choice of
fish for personal and family consumption, menu
selection and such diverse themes as legal and
compensation issucs. Decision-making in the
face of a threat always involves a balance be-
tween perceived or subjective risk on the one
hand, and the outcome (or utility) of a won
gamble on the other. In the study reported here,
the ‘utility’ - the joy of enjoying a gourmet fish
meal together with the risk of escaping clinical
ciguatera - this ‘utility’ is as consistent for a
within-group pattern as it is possible to imagine.
The collective ‘utility’ - good health after risk-
fish ingestion, or its inverse, clinical ciguatera -
was fully understood by all participants, all of
whom were giving papers on the subject at .an
international conference.

Many subjects think of personal risk in quite
specific and individual ways. Optimists tend to
regard themselves as invulnerable and will take
quite high (objective) risks. Pessimists on the

0 - ee ee
0.01 - 5%
0 -20%
G - 20%

es

01s ;

other hand and those with obsessive traits will
reject risks and not enter a gambling situation
where the nsks are mathematically very low (e.g.
<1 in 1,000 or even <1 in 10,000). Almost every-
one behaves inconsistently in their life's be-
haviour when it comes to risks, Some will accept
quite high risks in some areas of human activity
(speeding in the car, for example; or driving after
drinking) but will not accept very low tisks in
other areas. For example almost all home owners
will not leave their home uninsured against fire,
even though the objective risks are <1 in 60,000
and the outcome often not us severe as the conse-
quences of a motor vehicle accident.

The relationship between ciguatera and public
and commercial lability is a topical theme, There
is an undoubted duty of care to reduce the risk of
ciguatera to individual subjects. This applies both
to legal liability in common law and to statutes in
various Workplace Safety Acts and in Fair Trad-
ing Acts, The courts of vanous countries try to set
what is a ‘reasonable’ or ‘practicable’ nsk, with
penalties potentially imposed on those who ex-
pose individuals to risks greater than these ar-
bitrary levels. The current study reported here
shows that experienced, informed ciguatera
scientists and clinicians collectively take greater
risks than are currently accepted as ‘safe’ in the
fishing industry and in restaurant commerce.
What the implications of this are, in the evolution
of regulations and for case law, is for the future
to determine. Certainly, the law always demancls
public health regulations and commercial ‘duty
ofcare’ to be set at much safer levels (thatis lower
risks of exposure) than thal pragmatically ace
cepted by individuals functioning in their own
personal lives. This research confirms this
general observation in the specific context of
ciguatera. Large individual differences exist in
risk-taking behaviour generally (Pearn,!1973;
Pearn,1977), differences which are shown here to
apply to ciguatera specifically.

608

Health regulations and case-law practice (the
latter set by precedent) strive to protect all in
society - not only those whose personal behaviour
tends to be risky. In the developed world, the
perspective of the fishing industry is thus to see a
majority of informed individuals who will accept
a risk of upto 1 in 1,000 (0.1%) of contracting
ciguatera from dining in a seafood restaurant. Of
the world’s ciguatera experts 69% professed that
they would be happy to accept such a risk under
these circumstances. Whether or not this is ‘risky’
behaviour is also a subjective judgement. Current
regulations and some legal opinion indicates that
even those who profess to accept such risks must
be protected. Health regulations and local custom
(such as the banning from sale of the red bass,
Lutjanus bohar) operate to ensure that the objec-
tive risk to diners is significantly less than the
subjectively- acceptable risk level.

The phenomenon of subjective risk is culture-
dependent. Many individuals and indeed many
communities in the developing Pacific countries
accept the risk of ciguatera as a fact of life. In
some such communities individuals accept risks
>1 in 10. Where to set public health risk accep-
tance levels is thus difficult. If one is too conser-
vative, education and local community policies
will tend to reduce the impact of a highly
nutritious, high quality delicious food source with
consequent greater dependence on tinned fish and
tinned meat - the so-called dietary colonialism. In
Western countries of the Caribbean and the
Pacific rim, objective risk rates also vary from
society to society.

The fact that a significant proportion (25%) of
subjects recorded that they would, in a restaurant
setting, accept a higher risk than their own per-
sonal food-buying ‘baseline’ risk, imposes spe-
cial responsibilities and duties of care on
commercial restaurateurs. This implies that the
special vulnerability of patrons, a proportion of
whom are caught against their will and feel that
they have to take higher risks than they would in
other circumstances, need special protection. At
the very least, it implies that the objective math-
ematical risk of a random fish meal producing
ciguatera should be reduced as much as possible,
and suggests that restaurateurs should be aware

MEMOIRS OF THE QUEENSLAND MUSEUM

of the geographical source of risk-species which
they serve.

Attitudes to subjective risk are never static.
They change as scientific knowledge of ciguatera
increases; and will change further as practical test
systems for detecting individual ciguatoxic fish
become available. When they do, risk-acceptance
habits of the fish-eating public will change again,
as new community baselines are set for the risk
of ciguatera.

LITERATURE CITED

BAGNIS, R., KUBERSKI, T. & LAUGIER, S. 1979.
Clinical observations on 3,009 cases of ciguatera
(fish poisoning) in the South Pacific. American
Journal of Tropical Medicine and Hygiene 28:
1067-1073.

DALZELL, P. this memoir. Management of ciguatera
fish poisoning in the South Pacific.

GILLESPIE, N.C., LEWIS, R.J., PEARN, J.H.,
BOURKE, A.T., HOLMES, M.J., BURKE, J.B.
& SHIELDS, W.J. 1986. Ciguatera in Australia.
Occurrence, clinical features, pathophysiology
and management. Medical Journal of Australia
145: 584-590.

HOKAMA, Y., ASAHINA, A.Y., SHANG, E.S.,
HONG, T.W. & SHIRAI, J.L. 1993. Evaluation
of the Hawaiian Reef Fishes with the solid phase
immunload assay. Journal of Clinical Laboratory
Analysis 7: 26-30.

LAWRENCE, D.N., ENRIQUES, M.B., LUMISH,
R.M. & MACEO, A. 1980. Ciguatera fish poison-
ing in Miami. Journal of the American Medical
Association 244: 254-258.

LEE, C. 1980. Fish poisoning with particular reference
to ciguatera. Journal of Tropical Medicine and
Hygiene 83: 93-97.

PEARN, J.H. 1973. Patient’s subjective interpretation
of risks offered in genetic counselling. Journal of
Medical Genetics 10: 129-134.

PEARN, J.H. 1977. The subjective interpretation of
medical risks. Medikon 5: 5-8.

RUSSELL, F.E. & EGAN, N.B. 1991. Ciguateric
fishes, ciguatoxin (CTX) and ciguatera poisoning.
Journal of Toxicology - Toxin Reviews 10; 37-62.

TONGE, J.L, BATTEY, Y., FORBES, J.J. & GRANT,
E.M. 1967. Ciguatera poisoning: a report of two
outbreaks and a probable fatal case in Queensland.
Medical Journal of Australia 2: 1088-1090.

CLINICAL ASPECTS OF CIGUATERA: AN OVERVIEW
TILMAN A. RUFF AND RICHARD J. LEWIS

Ruff, T.A. & Lewis, R.J. 1994 08 01: Clinical aspects of ciguatera: an overview. Memoirs
of the Queensland Museum 34(3): 609-619. Brisbane. ISSN0079-8835.

Ciguatera is a polymorphous disease posing important health, nutritional, economic and
social problems for inhabitants of endemic areas, and occasionally for those in non-endemic
areas. Limited progress has been made in understanding the pathophysiology of the disease
and in developing effective treatment.

Clinical features of the disease are reviewed, and incidence, morbidity and mortality data are
outlined. Methods to prevent ciguatera and progress in treatment of ciguatera are discussed,
and key issues and needs for future research are described. These include: 1, consistent
epidemiologic data, using a consistent case definition; 2, the human immune response to
ciguatoxins; 3, the pathophysiological mechanisms underlying human disease, potentiation
of disease by alcohol, and the phenomenon of sensitisation; 4, better tests for ciguatoxins;
and 5, effective and safe treatment for affected patients.

Tilman A. Ruff, Department of Social & Preventive Medicine, Monash Medical School,
Alfred Hospital, Commercial Road, Prahran 3181, Victoria; Richard J. Lewis, Southern
Fisheries Centre, Department of Primary Industries, PO Box 76, Deception Bay, Queensland

4508; 22 November, 1993.

Ciguatera is the disease caused by the con-
sumption of fishes contaminated with
ciguatoxins, which originate from Gambierdis-
cus toxicus (Adachi & Fukuyo), a unicellular
dinoflagellate alga associated with coral reefs
(Adachi & Fukuyo,1979). Most toxic fish are
captured during inshore fishing near coral reefs.
Ciguatera is a circumtropical disease, likely to
affect >25,000 persons annually. Its greatest im-
pact is in Pacific island countries (Lewis, 1992a).
Although rarely fatal, possibly because fish suc-
cumb before concentrations lethal for humans
can be accumulated (Lewis,1992b), its morbid-
ity is considerable. Ciguatera has been reviewed
several times (Gillespie et al.,1986; Lewis,1986;
Hokama,1988; Vernoux,1988; Hokama,1991;
Juranovic & Park,1991; Russell & Egan,1991;
Lewis,1992a; Lewis & Ruff 1993).

Often regarded as an interesting tropical medi-
cal curiosity rather than a subject for serious
medical study, a good deal of the clinical litera-
ture on ciguatera is rather repetitive and anecdotal
and does relatively little to advance our under-
standing of the disease and its management.

Clinical manifestations of ciguatera are
protean. In areas where the disease is not en-
demic, the diagnosis is often not considered by
physicians unfamiliar with ciguatera, and a wide
variety of erroneous diagnoses may be made,
including neurosis. In many parts of the world,
increasing international travel, and increasingly
widespread transport and consumption of warm
water fish, especially coral reef fish, make it more

likely that cases will be seen outside endemic
areas. The possible severity, chronicity and pos-
sibility of effective treatment make it important
to consider the diagnosis in those presenting with
a compatible illness soon after eating fish.

CLINICAL FEATURES

CLINICAL MANIFESTATIONS

Ciguatera results in variable combinations of
gastrointestinal, neurological, general and car-
diovascular manifestations. Symptoms usually
develop 1-6 hours after ingestion of toxic fish - in
about 90% of cases within 12 hours (Gillespie et
al,1986, Gillespie, 1987), but in a few after more
than 24 hours (Bagnis et al.,1979; Bagnis &
Legrand, 1983; Narayan,1980). Gut involvement
usually consists of an acute self-limiting
syndrome akin to gastroenteritis, which may be
severe, but generally lasts less than 24-36 hours
(Gillespie et al 1986, Gillespie, 1987; Frenette et
al.,1988; Engleberg et al.,1983). Symptoms may
include abdominal pain, nausea, vomiting, diar-
rhea and tenesmus (rectal pain). Resulting in-
travascular volume depletion (‘dehydration’) and
electrolyte disturbances may be severe, par-
ticularly in young children. Volume depletion
and hypotension may be compounded by
myocardial depression and disturbed vasomotor
regulation (including deranged blood pressure
control). Neuromuscular disturbances are most
commonly sensory, but may also be motor. Al-
though neurological dysfunction is typically sug-

610

gestive of predominant involvement of peripheral
nerves, effects may occur at any level of the
nervous system from cerebral cortex to muscle.
Neurological manifestations are usually bilateral,
but may be asymmetrical (Hamburger, 1986) or
unilateral (Hashmi et al.,1989). Manifestations
may include coma, seizures, ataxia (disordered
co-ordination and balance), cranial neuropathies
including ophthalmoplegia (paralysis of eye
movement), myelopathy (spinal cord dysfunc-
tion), peripheral sensory, motor and autonomic
neuropathy and myositis (muscle inflammation).

Typical sensory symptoms are distal limb,
perioral and lingual paraesthesia and dysesthesia
(disordered sensation) - with prominent numb-
ness and tingling - and often a very unpleasant
form of hyperesthesia (abnormal, heightened
sensation) particularly associated with cold ob-
jects producing a distressing burning sensation
(Gillespie et al.,1986). Sometimes a reversal of
temperature sensation occurs, such that cold ob-
jects feel hot and vice versa. Reduced distal sen-
sation and reduced or absent tendon jerks are the
commonest neurological signs. A sensation of
fizzy, metallic taste may occur, Muscle weakness
- most commonly distal or generalised, oc-
casionally asymmetrical - sometimes involves
bulbar and respiratory muscle groups. Airway
protection and ventilatory support may be re-
quired in severe cases. Diffuse muscle pain is
common, and may be associated with elevated
blood levels of muscle enzymes and biopsy
evidence of myositis (Nakano, 1983).

General (non-localising) symptoms include
malaise, lassitude, irritability, depressed mood,
pruritus (itching), sleep disturbance and unusual-
ly vivid dreams. Headaches, arthralgia (joint
pain, particularly involving shoulders, elbows,
knees and ankles), pruritis (localised or
generalised), dental pain, a sensation of looseness
of the teeth and dysuria (painful urination) may
also occur, A variety of skin rashes, most com-
monly maculopapular, are sometimes present and
may be associated with desquamation (peeling)
during the healing phase.

Bradydysrhythmias (slow cardiac rhythm dis-
turbances), atrio-ventricular heart block, myocar-
dial depression and loss of vasomotor regulation
with hypotension, often postural, may occur
during the early phase and tend to resolve more
quickly than general and neurological symptoms.
Autonomic dysfunction may be manifested by
sweating, lacrimation (excessive tears), saliva-
tion and internal ophthalmoplegia (paralysis of
ocular accommodation and pupillary responses).

MEMOIRS OF THE QUEENSLAND MUSEUM

Symptoms often fluctuate from day to day and
at different times of day. The time course is
generally one of improvement over days to
weeks, but symptoms not uncommonly persist for
months, or rarely years. Consumption of alcohol
commonly exacerbates symptoms (Gillespie et
al.,1986; Gillespie, 1987). Death is rare (0.1% of
recorded cases) (Gillespie et al., 1986; Juravnovic
& Park,1991; Gillespie, 1987; Bagnis et al.,1979;
Bagnis & Legrand, 1987). Clinical manifestations
and severity may vary considerably, even among
individuals poisoned by the same fish. In the
absence of a specific human diagnostic test for
ciguatera, this wide variation in clinical manifes-
tations and the clinical nature of the diagnosis
make reliability difficult. Diagnosis is especially
difficult when only one person presents with less
than a full hand of symptoms. Nerve conduction
studies may be helpful, and demonstration of
toxin in any remaining fish samples, while very
useful, is often not possible. Commonly used
clinical criteria for diagnosis of ciguatera are
gastrointestinal and neurological symptoms fol-
lowing ingestion of potentially toxic fish. This
combination, however, occurred in only 25/53
(55%) of patients in one common source outbreak
(Engleberg, 1983), and 52/57 (91%) of patients in
another (Frenette, 1988).

PERSON-TO-PERSON TRANSMISSION

Although the vast majority of ciguatera cases
are caused by ingestion of toxic fish, various
forms of person-to-person transmission have
been described, and are indicative of the potent,
persistent and lipid-soluble nature of ciguatoxins.
These include: transmission via milk to breastfed
infants (Bagnis & Legrand,1987; Thoman,1989;
Blythe & De Sylva,1990), though hyperaesthesia
of the nipples of a lactating mother may interfere
with breast-feeding (Pearn et al.,1982); trans-
placental transmission, resulting in transient
neurological manifestations in the newborn fol-
lowing maternal illness near term (Pearn et
al.,1982); and apparent sexual transmission from
female to male (penile pain after intercourse in
the male partner of an affected woman) (Geller et
al.,1991) and vice versa (pelvic and vaginal pain
after intercourse in the female partners of affected
men)(Lange et al., 1989).

SENSITISATION AND RECURRENT ATTACKS
These are two of the most enigmatic aspects of
ciguatera, and increase its morbidity as well as its
social and economic consequences. Not only
does immunity not follow an attack of ciguatera,

CLINICAL ASPECTS OF CIGUATERA

but there is evidence from a variety of locations
that second and subsequent attacks tend to be
more severe than first attacks (Bagnis et al.,
1979). Also well documented is the phenomenon
of sensitisation. Persons who have previously had
ciguatera may suffer a recurrence of typical
ciguatera symptoms after eating fish which do not
cause symptoms in other persons (Narayan,
1980). Consumption of alcohol or chicken may
have the same efféct (Gillespie et al.,1986; Gil-
lespie, 1987). Such sensitisation can occur many
months or even years after an attack of ciguatera.
Both these factors are most troublesome in areas
where people depend heavily on fish as their
major dietary source of protein.

The basis for sensitisation and recurrent attacks
tending to increase in severity is not known, but
has been generally presumed to be immunologi-
cal, although the symptoms are not typically al-
lergic. A serum bank is being established at CSL
Limited in Melbourne, Australia, as a basis for
exploring the nature of sensitisation following
ciguatera (Sutherland & Lewis, 1992).

PATHOLOGY AND PATHOPHYSIOLOGY

Human pathological studies of ciguatera are
few. Nakano (1983) reported high blood levels of
creatine phosphokinase (CPK, a muscle enzyme)
in 7 men affected with ciguatera on Midway
Island, Central Pacific. The CPK level, initially
>1000 IU/L (normal <200 IU/L) in each, returned
to normal within 10 days. While their motor and
sensory nerve conduction velocities remained
normal, electromyography revealed changes con-
sistent with an acute myopathic process. Inser-
tional and spontaneous activity were normal.
Mild recruitment (minimal effort) produced
small motor units of short duration; maximal
recruitment (maximal effort) revealed enhanced
motor units of low amplitude. Repetitive nerve
stimulation suggested possible neuromuscular
junction fatigue in 2 patients. Muscle biopsies
from 3 patients showed muscle fibre splitting,
degeneration and necrosis, with subsarcolemmal
tubular aggregates and small lipid vacuoles.

A near-fatal case in Hawaii was associated with
prominent generalised muscle spasms and high
blood levels of CPK (41,000 IU/L, reference
range 45-35) and other muscle enzymes (Kod-
ama et al.,1989). Palytoxin present in smoked
mackerel from the Philippines was thought to be
responsible. Similar cases have also been
described following parrot fish ingestion in Japan
(Noguchi et al.,1987). A possible association be-

611

tween polymyositis (a chronic inflammatory dis-
ease of muscle) and ciguatera occurring some
years previously has been suggested (Stommel et
al.,1991) but remains speculative.

The major morbidity of ciguatera, however, is
probably attributable to its effects on peripheral
nerves, Ayyar & Mullaly (1978) reported slowed
sensory conduction velocities without decrease in
sensory nerve action potential amplitude in af-
fected patients. Other studies (Allsop et al., 1986;
Cameron et al.,1991; Cameron & Capra,1991)
documented increased distal motor and sensory
latencies, reduced motor and sensory conduction
velocities, prolongation of the absolute refrac-
tory, relative refractory and supernormal periods,
reduced sensory amplitudes and F wave latencies.
These findings are consistent with a neuropathic
process which in traditional neurological terms is
predominantly demyelinating rather than axonal
in type (primarily damaging the myelin sheaths
of nerves, which are part of Schwann cells, rather
than the nerve fibres themselves).

The report of human nerve biopsy in ciguatera
(Allsop et al.,1986) found striking edema of vac-
uoles in Schwann cell cytoplasm adaxonally (im-
mediately abutting axons), with axonal com-
pression and vesicular degeneration of myelin.
Nakano (1983) described diffuse slowing of brain
electrical activity, elevated cerebrospinal fluid
pressures and abnormal brainstem auditory-
evoked responses in ciguatera patients, although
these are not common findings.

One interesting finding by Cameron & Capra
(1991), in the rat tail nerve in vivo, is that a blood
ethanol (alcohol) level of 0.05% was found to
significantly increase the magnitude and duration
of the abnormal supernormal response observed
in ciguatoxin-treated rats. The mechanism of this
potentiation, which is consistent with common
clinical experience in humans, is yet to be
elucidated. The nature of the human immune
response to ciguatera is essentially unknown.

TREATMENT

Despite advances in understanding the nature
and pharmacology of ciguatoxins, this has yet to
translate into major specific therapeutic advan-
ces. No specific antidote is known for any of the
many marine dinoflagellate toxins, including
those causing ciguatera. Therapy remains
primarily symptomatic and supportive. Many
types of treatment have been tried and although
some important uncontrolled observations have
been reported, particularly in relation to man-

612

nitol, no double-blind controlled clinical trial
results are available for any treatment modality.

Supportive and symptomatic therapy may in-
clude bed rest, analgesia, fluid and electrolyte
replacement, airway protection and ventilatory
support, circulatory support (including positive
inotropic agents), management of dysrhythmias
(most commonly bradycardias and atrio-
ventricular block, occasionally necessitating
temporary cardiac pacing), general care of the
unconscious patient, antihistamines and cool
showers for pruritis, hypnotics, etc. In French
Polynesia standard (Bagnis et al.,1992) but un-
proven (Calvert,1991), therapy for hospitalised
patients has consisted of intravenous infusions of
vitamins C and B6 (pyridoxine) and calcium
gluconate. A wide variety of traditional remedies,
including a considerable number of plants, are
used in various areas (Cooper,1964; Narayan,
1980; Amade & Laurent, 1992; Dufva et al., 1976;
Bourdy et al.,1992) Screening of traditional plant
remedies with a novel mouse bioassay has found
that an extract from leaves of Argusia argenta can
reduce the effects of ciguatoxin (Amade &
Laurent, 1992). Efficacy or safety in humans of
traditional remedies are unknown,

Occasional success has been reported with low
dose amitriptyline, a tricyclic antidepressant, par-
ticularly for chronic paraesthesia and other
neurological symptoms (Bowman, 1987; Davis &
Villar, 1986; Calvert et al.,1987). Fluoxetine (an
antidepressant drug which is a relatively specific
serotonin-uptake inhibitor) was reported to
reduce chronic fatigue in two patients with
ciguatera in whom symptoms had persisted for
over nine months (Berlin et al.,1992). Nifedipine
(a calcium channel blocker) (Calvert et al.,1987)
and tocainide (a lignocaine-like local anaesthetic
agent) (Lange et al.,1988; Lange & Kreider, 1988)
have some theoretical appeal but experience with
their use is very limited.

The most dramatic reported experience of suc-
cessful treatment of ciguatera has been that of
Palafox et al. (1988) in the Marshall Islands, who
treated 24 patients with acute ciguatera with in-
travenous infusions of mannitol, an osmotic
diuretic agent most commonly used in the treat-
ment of cerebral edema. Mannitol is inexpensive
and readily available, but must be given by in-
travenous infusion and accompanied by careful
patient monitoring. Two patients in coma and one
in shock are reported to have responded within
minutes, with full and rapid recovery, hitherto
virtually unknown in severe ciguatera (recovery
typically takes at least one, and more usually two

MEMOIRS OF THE QUEENSLAND MUSEUM

weeks). Neurological and muscular manifesta-
tions improved dramatically; gastrointestinal
symptoms resolved more slowly. A variety of
case reports and uncontrolled observations in-
volving small numbers of patients (Pearn et al.,
1989; Williamson,1990; Stewart,1991) docu-
mented a clear clinical impression that in some
patients (including young children) (Williams &
Palafox,1990), mannitol is dramatically effica-
cious, notwithstanding the highly variable natural
history of the disease. Patients at the more severe
end of the disease spectrum and who are treated
early (within 24 hours of symptom onset) would
appear most likely to benefit from mannitol. The
mechanism of action of mannitol in ciguatera is
unclear - possibilities suggested (Pearn et al.,
1989) include a direct anti-ciguatoxin effect via
a scavenger mechanism, or an osmotic effect
reducing Schwann cell edema, thereby ameliorat-
ing neurological dysfunction. Experimental
studies on interactions between ciguatoxin and
mannitol indicate that mannitol does not act to
reduce the affinity of the sodium channel for
ciguatoxin, nor does mannitol act as a scavenger
for ciguatoxin (Lewis unpubl. data), suggesting
that the osmotic effect is the most likely mode of
action.

In the first controlled trial of mannitol (Bagnis
et al.,1992) 34 patients were treated, compared
with 29 patients treated with vitamins B6 and C
and calcium. Patients were well matched, and a
clinical score based predominantly on the number
and severity of subjective symptoms showed sig-
nificant benefit 1 and 24 hours after onset of
treatment, particularly for paraesthesiae and
gastrointestinal symptoms. The study suffers
from a number of weaknesses: it is unclear
whether the patients or the observers were
blinded, the clinical score was based excessively
on subjective criteria, no follow-up beyond 24
hours is reported, and the differences between
treatment groups, while statistically significant,
would appear not to be of major clinical sig-
nificance. The clinical condition of some patients
deteriorated in the first 24 hours despite mannitol
infusion. Further studies of mannitol treatment
are underway, at least in the Marshall Islands,
Kiribati, Fiji and Florida. A rigorously con-
ducted, double-blind controlled clinical trial, in-
cluding as many objectively determined
parameters as possible and with adequate follow-
up is needed. At present, given the safety of
mannitol and the rapidity with which benefit may
be evident, the administration of mannitol would
seem justified in patients whose illness is

CLINICAL ASPECTS OF CIGUATERA

20005 FJ) (727,000)

1abe
of ain lee

1500 FRENCH POLYNESIA (173,000)

o-
1000 VANUATU (156,000)
—!
0 a ,

800) NEW CALEDONIA (160,000)

s_..

WESTERN SAMOA (168,000)

1975

1980

1985 1990

613

15005 kIRIBAT! (67,000)

MARSHALL ISLANDS [32,000)

} TUVALU (8,000)
200 +
100+
of 44
TOKELAU (1,600)
O24 67 a
100-4 AMERICAN SAMOA (38,000)

50 L j
ol a,

rua
1975 1980 1985 1990

FIG. 1. Annual cases of ciguatera for selected Pacific countries, 1973-1992. Data from SPEHIS,1973-1992. The
1988 population estimates for each country are indicated in parenthesis (FAO,1990), Prior to 1982, data for the
Marshall Islands also included data from the Federated States of Micronesia, the Northern Marianas and Palau.

moderate or severe, and particularly those who
present during the acute phase of the illness,
typically within 24 hours of the onset of symp-
toms. A dose of 1g mannitol per kg body weight,
as a 20% solution, infused over about 30 minutes,
has been most commonly used (Palafox et
al.,1988; Pearn et al.,1989). The clinical impres-
sion is that half this dose, infused over 60
minutes, appears to be less effective (Pearn etal.,
1989). No adverse experiences have been
reported with use of mannitol in patients with
ciguatera, but it is prudent to ensure that patients
are replete in intravascular volume prior to com-
mencement of mannitol infusion,

The remoteness of small and widely scattered
island communities from health care services,
particularly in the Pacific, imposes limitations on
availability of medical treatments, particularly

one requiring careful supervision and intravenous
infusion. A safe orally-active therapy requiring
minimal supervision is desirable.

All patients suffering from ciguatera should be
advised to avoid fish and alcohol for at least 3
months, and to reintroduce them into their diet
cauliously, recognising that ingestion of either
may precipitate a relapse. Many sufferers of
cigualera, particularly in Western cultures and
where fish are not a crucial foodstuff, lose all
inclination to again eat reef fish.

INCIDENCE OF CIGUATERA

The most comprehensive regional database on
ciguatera (SPEHIS,1973-1992) also includes
other forms of marine food poisoning (scombroid
poisoning. clupeotoxism, mullet poisoning, puf-

614
Cases/10,000 pop.
0 50 100
————__._________
RA)
TOK AL
TVA
FRENCH POLYNESIA [x
VANUATU [a
MARSHALL ISLANDS [
cook ISLANDS [a
Ful
NORTHERN MARIANAS [i
NEW CALEDONIA [i
WALLIS AND FORTUNAS ff (average cases
AMERICAN SAMOA ff per annum)
NIUE |
WESTERN SAMOA |
GUAM |
NAURU |
MICRONESIA |
PALAU |
TONGA |
PAPUA NEW GUINEA *|
PITCAIRN

SOLOMON ISLANDS *

FIG. 2. Incidence of ciguatera in Pacific Island
countries. Cases per 10,000 population are indicated
Data are given per annum (p.a.) and were averaged
from SPEHIS data (1973-1991) covering the period
1985-1990. Asterisks indicate incomplete reporting
to SPEHIS from these countries.

fer fish poisoning and invertebrate intoxications).
However, ciguatera typically dominates as a
cause of fish poisoning in the Pacific region
(Lewis, 1992a). Ciguatera is reportedly prevalent
throughout Pacific island countries with the ex-
ception of the Solomon Islands and Pitcairn Is-
land (Fig. 1). Ciguatera is invariably substantially
underreported. In Australia it is estimated that as
few as 20% of cases are reported and <10% of
ciguatera cases in Western Samoa are reported to
SPEHIS (Lewis, 1992a). Similar levels of under-
reporting are likely in other countries. Under-
reporting may vary within and between countries,
and over time.

For countries of the South Pacific, the highest
average incidence of reported ciguatera for the
period 1985-1990 was c.100/10,000 population
per annum(p.a.) in Kiribati, Tokelau and Tuvalu
(Fig.2). The average reported incidence of

MEMOIRS OF THE QUEENSLAND MUSEUM

ciguatera was less than half these levels in French
Polynesia, Vanuatu, the Marshall Islands and the
Cook Islands. The remaining 13 countries
reported <15 cases/10,000 people p.a. Over the
same period, the average reported incidence of
ciguatera in Queensland (population 2.9 million)
was 0.16 cases per 10,000 p.a., a level similar to
that reported for Tonga. By way of comparison,
in the Iles Saintes, Guadeloupe, in the Caribbean,
annual ciguatera incidence has been estimated to
be 30 (Czernichow et al.,1987), in the US Virgin
Islands (Caribbean) to be 73 (Morris et al.,1982)
and in Miami to be 5/10,000 population
(Lawrence et al.,1980).

INDIRECT EFFECTS OF CIGUATERA

Throughout Pacific island countries there is a
heavy dependence on the inshore fishery resource
of reefs for dietary protein and animal fats. Johan-
nes (1990) suggested that the inshore fisheries
resource is of greater importance per capita to
Pacific island countries than in any other region
of the world. Nowhere is the impact of ciguatera
greater than in atoll countries of the Pacific where
intake of reef fish is often above 100g/per person
per day (Lewis, 1992a). Ciguatera is also impor-
tant in relative terms, being one of the more
commonly reported diseases (SPEHIS,1973-
1991).

Ciguatera may have indirect effects on health
by predisposing victims to poor nutrition and
other diseases, and via its social and economic
effects. The ability of subsistence communities to
provide food, especially difficult on the poorer
Pacific atolls, may be impaired due to the neces-
sity of reducing fish consumption to reduce the
risk of ciguatera (Lewis,1986). Ciguatera may
have direct economic effects, reducing trade op-
portunities in potentially ciguateric fishes and
damaging tourism (Lewis,1986). The effect of
ciguatera on fish consumption is likely to be least
in countries where alternative dietary protein
sources to locally caught fish are costly and few,
and where a system of traditional beliefs acts to
reduce perceptions of the adverse effects of
ciguatera (Lewis,1992a). People in larger and
developed countries (e.g. Australia) with more
diverse food sources and a less traditional orien-
tation to the sea may be less accepting of
ciguatera than are people in many Pacific Island
countries.

The need to avoid fish after an outbreak of
ciguatera may exacerbate undernutrition, espe-
cially among children (Eason & Harding,1987).

CLINICAL ASPECTS OF CIGUATERA

Fear of poisoning may accentuate dependence on
imported food. In many Pacific locations, as
much as 90% of fish eaten comes out of a can
(Lewis, 1986). Increased intake of imported food
is often associated with a higher salt, fat and
refined carbohydrate diet that contributes to an
increase in chronic degenerative diseases such as
diabetes (Zimmet et al.,1981), gout (Prior et al.,
1987), hypertension (Zimmet et al., 1980) and
atherosclerotic yascular disease (Taylor &
Thoma,1985) in indigenous Pacific populations.

PREVENTION

INDIVIDUAL LEVEL

Individuals can reduce their risk of contracting
ciguatera by: 1, avoidance of warm water reef
fish, particularly those with a known propensity
to be toxic, and avoidance of certain pelagic fish
which feed on them (e.g. barracuda and mack-
erel), especially in areas with a history of
cigualera; 2, avoidance of all fish at locations
which are 2 known recent or current source of
toxic fish. 3, complete avoidance of moray eels,
Which are commonly highly toxic (Murata et
al..1990: Lewis et al.,1991; Lewis et al.,1992),
except when captured in areas with no history of
ciguatera; 4, avoidance of carnivorous fish may
reduce, but does not eliminate, the risk of con-
iracting Ciguatera. Ciguatoxins tend to be con-
centrated as they pass up the food chain, and
larger fish (particularly 2.5kg) are more likely to
be toxic (Hessel et al.,1960); 5, avoidance of the
head, roe and viscera of potentially toxic fish.
Concentrations of ciguatoxins in fish liver may
be up to 50 times higher than in muscle (Banner,
1976); 6, eating a small portion (20-100g) from
any one fish at the first sitting (Lewis, 1992a); 7,
feeding a large fish flesh meal to a cat which is
observed for at least 6 hours prior to human
consumption of portions of the same fish
(Lewis, 1992a; Cooper,1964); 8, washing the
flesh of herbivorous fish (such as patrol and sur-
geon fish), in several changes of water prior to
consumption has been recommended on the basis
that this may remove some of the water-soluble
maitotoxin (Juranovic & Park,1991), This hus
not, however, been demonstrated to be useful

PUBLIC HEALTH MEASURES

These include: |, education of fisherpeople and
the public in affected areas about the risk of
ciguatera and how this risk can be reduced
(Ahmed,1991); 2, closure of known highly toxtc
arcas to fishing (Ahmed,1991); 3. dans on the sale

615

of high risk fish from known toxic locations, Such
bans have been ermployed in American Samoa
(Dawson, 1977), Queensland (Lewis et al.,1988),
French Polynesia (Lewis,1986), Fiji (Sorokin,
1975), Hawaii (Ahmed,1991; Gallop & Pon,
1992) and Miami (Craig,]1980); apparently with
some success, but with allendant economic Joss;
4, detection of ciguatoxic fish prior to consump-
linn. Such tests should be specific and sensitive
for the toxins implicated in human disease. They
should be sufficiently sensitive to detect 0.1 nM
ciguatoxin-l per kg of fish flesh (Lewis, 1992h),
To be used effectively at the community level,
they should be robust, temperature-insensitive,
reliable, inexpensive and simple to use.

Hokama pioneered development of such a test
to detect ciguateric fish (Hokama.1991). A radio-
immunoassay (RIA), subsequently modified to a
simpler enzyme immunoassay (ELA) (Hokama,
1985) has been further simplified to a ‘stick lest’
which has been used to screen fish caught in
Hawaii (Hokama et al..1990). All of 57 fish im-
plicated in causes of ciguatera, and provided by the
Hawaiian Department of Health in 1987-89
lested pesilive on a stick chnzyme immunoassay
(S-EJA) using @ monoclonal antibedy against
pee (MAhb-CTX) (Hokama et al..1990),
All 86 Carany sp Quick) and Seriola dumerili
{ambenack) provided by sports fisherpersons and
found to be negative on the S-ELA test. were
consumed without incident (Hokamactal.,1990).
However a high proportion, 1195/2190 (S59), of
randomly tested fish of 19 different, potentially
ciguatoxic species tested borderline or positive
(hokama et al..1990), suggesting a high rate of
false posinive tests. The false negative rate how-
ever, which ts of greater importance, would ap-
pear to be acceptably low.

Although the test has problems of specificity,
cross-reacting with a vanety of polyether toxins,
such as okadaic acid, which play an uncertain role
in ciguatera, and is not sufficiently robust to be
used in the field (Hokama et al.,1990), it holds
promise as a practical measure in ciguatera con-
trol. particularly for large fish processed commer-
cially. Several groups are in the process of
developing such antibody-based tests, including
Hawaii Chemtec Inc, which plans to commercial-
ize a modified version of the Hokama test. Re-
search on detection of ciguatoxins using fluor
escence high pressure liquid chromatography
(HPLC) is also in progress. HPLC-based assitys,
perhaps linked to fluorescence or mass spectral
detectors, have the potential to confirm cigua-
toxins in small samples of fish Hesh.

616

A rapid inexpensive test may eventually sup-
plant the riskier process in use in some Pacific
island areas, whereby an adult human eats or a cat
is fed fish from an area, several times a year, to
reassess the toxicity present in locally- caught
reef fish. Such testing may be used particularly to
protect children from ciguatera (Cooper, 1964).

Long-term monitoring of populations of
dinoflagellate(s) associated with ciguatera, their
toxicity and toxicity of fish at various levels of
the food chain at a range of sentinel sites may be
of benefit in predicting ciguatera in an area. This
may enable timely action, such as closing an area
to fishing, or restricting types or sizes of fish
caught, before an outbreak occurs (Ahmed,
1991). Such monitoring, particularly in areas of
human impact on coral reefs (particularly through
construction activities, other forms of coral
damage, terrestrial and marine pollution, includ-
ing sewage and agricultural runoff), could also
make an important contribution to our under-
standing of the genesis of ciguatera. Such
monitoring should be initiated with baseline
studies prior to major developments likely to
damage or alter a coral reef. There is widespread
concer, particularly in the Pacific, that coral reef
damage and pollution associated with population
increase and economic development may in-
crease the incidence of ciguatera (Lewis, 1992a,
Lewis,1986). The possible effects of global
warming, stratospheric ozone depletion and other
global environmental changes on ciguatera are
unknown and provide additional justification for
long-term environmental monitoring.

Restriction of human activities likely to be as-
sociated with coral reef damage. In some Pacific
islands, such as the Line islands (Ross,1947),
Gilbert Islands (Cooper,1964), and Hao,
Moruroa and Mangareva in French Polynesia
(Ruff,1989a,b) military dumping of material on
reefs, construction activities and nuclear test ex-
plosions have been associated with outbreaks of
ciguatera. Similarly, outbreaks have followed
shipwrecks, shore modification and other con-
struction activities in the Marquesas (Lewis,
1984a) and Hawaii (Gallop & Pon,1992; Lewis,
1984b). Although not supported by firm data,
local Aboriginal people in East Arnhem Land
ascribe the occurrence of ciguatera near the Gove
peninsula to the construction of a township and
alumina plant at Nhulunbuy in the early 1970s.

FUTURE DEVELOPMENTS

Key issues and areas for research include: 1, the

MEMOIRS OF THE QUEENSLAND MUSEUM

need for a consistent case definition of ciguatera,
a crucial basis for comparable epidemiologic and
clinical data; 2, better tests for ciguatoxins, in-
cluding ones which can be applied to human
clinical samples. Antibodies which are more
selective and have higher affinity for the
ciguatoxins than those currently available are
needed; 3, understanding of the pathophysiologi-
cal mechanisms underlying human disease,
potentiation of the disease by alcohol, and the
phenomenon of sensitisation; 4, understanding of
the human immune response to ciguatera may
provide a basis for more effective control, par-
ticularly through immunisation; and 5, treatment
for ciguatera which is simple to administer
(preferably orally), inexpensive, and which is
demon-strated to be effective and safe. In the
short term, a well-conducted randomised control-
led double-blind trial of mannitol therapy is
needed.

ACKNOWLEDGEMENTS

We thank Karen Bokser for secretarial assis-
tance.

LITERATURE CITED

ADACHI, R. & FUKUYO, Y. 1979, The thecal struc-
ture of a marine toxic dinoflagellate Gambierdis-
cus toxicus gen, et sp. nov. collected in a
ciguatera-endemic area. Bulletin of the Japanese
Society of Scientific Fisheries 45: 67.

AHMED, FE. (ed.) 1991. ‘Seafood Safety’. (National
Academy Press: Washington D.C.) 87-110,
ALLSOP, J.L., MARTINI, L., LEBRIS, H., POL-
LARD, J., WALSH, J. & HODGKINSON, S.
1986. Les manifestations neurologiques de la
ciguatera. Revue Neurology (Paris) 142: 590—

597.

AMADE, P. & LAURENT, D. 1992. Screening of
traditional remedies used in ciguatera fish poison-
ing treatment. P, 503. In Gopalakrishnakone, P. &
Tan, C.K. (eds), ‘Recent advances in toxicology
research, vol. 2’. (National University of Sin-
gapore: Singapore).

AYYAR, D.R. & MULLABY. N.J. 1978. Ciguatera:
clinical and electrophysiological observations.
Neurology 28: 354.

BAGNIS, R, KUBERSKI, T. & LANGIER, S. 1979.
Clinical observations on 3009 cases of ciguatera
(fish poisoning) in the South Pacific. American
Journal of Tropical Medicine and Hygiene 286:
1067-1073.

BAGNIS, R.A. & LEGRAND, A.M. 1987. Clinical
features on 12,890 cases of ciguatera (fish poison-
ing) in French Polynesia. Pp. 372-393. In
Gopalakrishnakone P. & Tan, C.K. (eds).

CLINICAL ASPECTS OF CIGUATERA

‘Progress in venom and toxin research". (National
University of Singapore: Singapore).

BAGNIS, R., SPIEGEL, A.. BOUTIN, J.P.,
BURUCOA, C., NGUYEN, L., CARTEL, LL.,
CAPDEVIELLE, P.. IMBERT, P., PRIGENT.D.,
GRAS, C. & ROUX. J. 1992. Evaluation de
L'efficacite du mannitod dans Je traitement de Je
Ciguatera on Polynesie Francaise. Medicine
Tropicale 52:67—73,

BANNER, A.H, 1976. Ciguatera: a disease from coral
teef fish. P.177. In Jones, O.A,& Endean, R. (eds).
“Biology and geology of coral reefs vol.3’.
{Academic Press: London).

BERLIN, R.M., KING, S.L. & BLYTHE, D.G. 1992.
Symptomatic improvement of chronic fatigue
with Nuoxetine in ciguatera fish poisoning. Metli-
cal Joumal of Australia 157; 567,

BLYTHE, D.G. & DE SYLVA, D.P. 1990, Mothers
milk turns toxic following feast. Journal of the
American Medical Association 264: 2074.

BOURDY. G., CABALION, P., AMADE, P. &
LAURENT, D. 1992. Traditional remedies used
in the Western Pacific for the treatment of
ciguatera poisoning. Journal of Ethnopharmacol-
ogy 36; 163.

BOWMAN, P.B. 1984. Amitriptyline and ciguatera
(letter), Medical Journal of Australia 140; 802.

CALVERT, G.M. 1991. The recognition and manage~
ment of ciguatera fish poisoning. Pp.l—-1]. In
Miller, D.M, (ed.), “Ciguatera seafood toxins*
(CRC Press: Boca Raton).

CALVERT, G.M., HRYHORCZUK, D.O. & LEIKIN,
J.B, 1987 Treatment of ciguatera fish poisoning
with amitsiptyline and nifedipine, Clinical
Toxicology 25: 423428.

CAMERON, J,. FLOWERS, A.E. & CAPRA, MP.
199]. Electrophysiological studies on ciguatera
poisoning in man (Part Il). Jounal of Neurological
Sciences 11; 93-97.

CAMERON, J. & CAPRA, MF, 1991. Neurological
studies oa the effects of cigualoxin on mammalian
nerve, Pp, 21-32. In Miller, DM. (ed.) "Ciguatera
seafood toxins”. (CRC Press: Boca Raton).

COOPER, M.J. 1964, Ciguatera and other marine
poisonimg m the Gilbert Islands, Pacific Science
18: 41 1-40.

CRAIG, C.P_ 1980. It’s always the big ones that should
get away. Journal of the American Medical As-
sociation 244: 272.

CZERNICHOW, P., DROY, J.M.. EZELIN, F. &
LEROY, J. 1984. Epidemiology of ciguatera in
the Nes Saintes (Guadeloupe). Revue Epidemioli-
que Sante Publique 32: 315-321.

DAVIS, R.T. & VILLAR, L.A. 1986, Symptomatic
mmprovement with amilriplyline in ciguatera lish
poisoning. New England Joormal of Medicine
315: 65,

DAWSON, J.M. 1977. Fish porate in American
Samoa. Hawaii Medical Journal 36:239-243.

DUFYA, E.. LOISON, G. & HOLMSTEDT, B, 1976.

617

Duboisia myoporoides: native antidote against
Ciguatera poisoning. Toxicon 14; 55-64.

EASON, B.J., HARDING, E. 1987. Neurotoxic fish
poisoning in the Solomon Islands. Papua New
Guinea Medical Joumal 30; 49-52,

ENGLEBERG, N.C., MORRIS, J.G,, LEWIS, J.,
MCMILLAN, J.P., POLLARD, R.A. & BLAKE,
P.A. 1983, Ciguatera fish poisoning: 8 major com-
mon-source outbreak in the US Virgin Islands.
Annals of Intemal Medicine 98: 336-337,

FAQ.1990. ‘Fisheries statistics yearbook: catches and
landings 1988". (FAQ: Rome). 102.

FRENETTE, C.. MACLEAN, D. & GYORKOS T.W,
1988. A large common-source outbreak of
ciguatera fish poisoning. Joumal of Infectious
Diseases 158: 1128-1131

GALLOP.J.H, & PON, EW. 1992, Ciguatera: areview.
Hawaii Medical Joumai 51: 91-99,

GELLER, R.J., OLSON, K.R.& SENECAL, PE. 1991.
Ciguatera fish poisoning in San Franciscu,
California, caused by imported barracuda,
Western Journal of Medicine 155; 639-042.

GILLESPIE, N.C., LEWIS, R.I., PEARN, J.H,,
BOURKE, A-T.C., HOLMES M.J., BOURKE,
J.B, & SHIELDS, W,J, 1986, Ciguatera in
Australia: occurrence, clinical features,
pathophysadlozy anc management. Medical Jaur-
nal of Australia 145; 584-390,

GILLESPIE, N. 1987. Ciguatera poisoning. P.160. In

“ovacevich, J., Davis, P, & Pear, J. (eds), “Toxic
plants and animals. A guide for Australia’.
(Queensland Museum: Brisbane).

HAMBURGER, H.A. 1986. The neuro-ophthal-
mologic signs of ciguatera poisoning: a case
report. Annals of Ophthalmology 18: 287-288

HASHMI, M.A., SOROKIN, J.J. & LEVINE, $.M.
1949. Ciguatera fish poisoning. New Jersey
Medicine 86: 469-471.

HESSEL, D.W., HALSTEAD B.W, & PECKHAM,
N.H. 1960. Marine biotoxins. I. Ciguatera poison:
some hiological and chemical aspects, Annals of
the New York Academy of Sciences 90; 783.

HOKAMA., Y. 1988, Ciguatera fish poisoning. Journal
of Clinical Laboratory Analysis 2: 44,

HOKAMA, Y. 1991. Immunological analysis of low
molecular weight marine toxins. Journal of
Toxicology and Toxin Reviews 10:1.

HOKAMA, Y_ 1985, A repad simplified enzyme im-
munoassay stick test for the detection of
ciguatoxin and related polyethers from fish tissue.
Toxicon 23: 939.

HOKAMA, Y., ASAHINA, A.Y., HONG, T.W-P.,
SHANG, E.S. & MIYAHARA, J.T. 1990.
Evaluation of the stick enzyne immunoassay in
Corarsz sp. and Seriofa dumerili associated with
cigualera, J, Clinical Laboratory Analysis 4: 362
306

JOHANNES, R.E. 1990, Managing small-scale
fisheries in Oceania: unusual constraints and op-
portunities. P. 5, In Campbell, H,, Menz, K. &
Waugh, G. (eds), ‘Economics of fishery manage -

618

ment in the Pacific Island Region’. (ACIAR: Can-
berra). Proceedings no. 25.

JURANOVIC, L.R. & PARK, D.L. 1991. Foodborne
toxins of marine origin: ciguatera. Reviews of
Environmental Contamination and Toxicology
117: 51-94,

KODAMA, A.M., HOKAMA, Y., YASUMOTO, T.,
FUKUI, M., MANEA, S.J. & SUTHERLAND,
N. 1989. Clinical and laboratory findings im-
plicating palytoxin as cause of ciguatera poisoning
due to Decapterus macrosoma (mackerel).
Toxicon 27: 1051-1053.

LANGE, W.R, LIPKIN, K.M. & YANG, G.C, 1989,
Can ciguatera be a sexually transmitted disease?
Clinical Toxicology 27: 193-197.

LANGE, W.R., KREIDER, S.D., HATTWICK, M. &
HOBBS, J. 1988. Potential benefit of tocainide in
the treatment of ciguatera: report of three cases.
American Journal of Medicine 84: 1087.

LANGE, W.R. & KREIDER, S.D. 1988. ‘A pilot study
of the potential benefit of tocainide in the manage-
ment of ciguatera toxicity’. Abstract WPA4.1.
First Conference on International Travel
Medicine, Zurich.

LAWRENCE, D.N., ENRIQUEZ, M.B., LUMISH,
R.M. & MACEDO, A. 1980. Ciguatera fish poison-
ing in Miami. J. American Medical Association
244: 254-258.

LEWIS, N.D. 1984a. Ciguatera - parameters of a tropi-
cal health problem. Human ecology 12; 253.

LEWIS, N.D. 1984b. Ciguatera in the Pacific: incidence
and implications for marine resources develop-
ment. P.9. In Ragelis, E.P. (ed.), ‘Seafood Toxins’.
(American Chemical Society: Washington DC)
(ACS Symposium Series 262).

LEWIS, N.D. 1986. Epidemiology and impact of
ciguatera in the Pacific: areview. Marine Fisheries
Review 48: 6-13.

LEWIS, R.J. 1992a. Socioeconomic impacts and
management of ciguatera in the Pacific. Bulletin
de la Société de Pathologie Exotique 85: 427-434.

LEWIS R.J. 1992b. Ciguatoxins are potent ich-
thyotoxins, Toxicon 30: 207.

LEWIS, R.J., CHALOUPKA, M.Y., GILLESPIE, N.C.
& HOLMES, M.J. 1988. An analysis of the human
response to ciguatera in Australia. Pp.67—72. In
Choat J.H. et al., (eds), ‘Proceedings of the 6th
International Coral Reef Symposium, Australia’.
(6th International Coral Reef Symposium Execu-
tive Committee: Townsville). Vol. 2.

LEWIS, R.J. & RUFF, T.A. 1993. Ciguatera: ecologi-
cal, clinical and socioeconomic perspectives.
Critical Reviews in Environmental Science and
Technology 23: 137-156.

LEWIS, R.J., SELLIN, M., STREET, R., HOLMES,
M.J. & GILLESPIE, N.C. 1992. Excretion of
ciguatoxin from moray eels (Muraenidae) of the
central Pacific. P.131. In Tosteson, T. (ed.),
‘Proceedings Third International Conference on
Ciguatera Fish Poisoning’. (Polyscience Publica-
tions: Quebec).

MEMOIRS OF THE QUEENSLAND MUSEUM

LEWIS R.J., SELLIN, M., POLI M.A., NORTON,
R.S., MACLEOD J.K. & SHEIL, M.M. 1991.
Purification and characterisation of ciguatoxins
from moray eel (Lycodontis javanicus,
Muraenidae), Toxicon 29: 1115.

MORRIS, J.G., LEWIN, P., SMITH, C.W., BALKE,
P.A. & SCHNEIDER, R. 1982. Ciguatera fish
poisoning: epidemiology of the disease on St.
Thomas, US Virgin Islands. American Journal of
Tropical Medicine and Hygiene 31; 574-578.

MURATA, M., LEGRAND, A.M., ISHIBASHI, Y.,
FUKUI, M. & YASUMOTO, T. 1990, Structures
and configurations of ciguatoxin from moray eel
Gymnothorax javanicus and its likely precursor
from the dinoflagellate Gambierdiscus toxicus.
Journal of the American Chemical Society 112:

4380.
NAKANO, K.K. 1983. Ciguatera poisoning: an out-
break on Midway Island. Clinical,

electrophysiological and muscle biopsy findings.
The Journal of Neurological and Orthopaedic
Surgery 4: 1-16.

NARAYAN Y. 1980. Fish poisoning in Fiji. Fiji Medi-
cal Journal 8: 67-574.

NOGUCHI, T., HWANG. D., ARAKANA, O.,
DAIGO, K., SATO, S., OZAKI, H., KAWAI, N.,
ITO, M. & HASHIMOTO, K. 1987. Palytoxin as
the causative agent in the parrot fish poisoning.
Pp. 325-335. In Gopalaknishnakone P. & Tan,
C.K. (eds), ‘Progress in venom and toxin re-
search’, (National University of Singapore: Sin-

gapore).

PALAFOX, N.A., JAIN, L.G., PINANO, A.Z.,
GULICK, T.M., WILLIAMS, R.K. & SCHATZ,
LJ. 1988. Successful treatment of ciguatera fish
poisoning with intravenous mannitol. Journal of
the American Medical Association 259: 2740—
2742.

PEARN, J., HARVEY, P., DE AMBROSIS, W.,
LEWIS, R. & MCKAY, R. 1982. Ciguatera and
pregnancy. Medical Journal of Australia 1: 57-58.

PEARN, J.H., LEWIS, R.J., RUFF, T., TAIT, M.,
QUINN J., MURTHA, W., KING, G., MAL-
LETT, A. & GILLESPIE, N.C. 1989. Ciguatera
and mannitol: experience with a new treatment
regimen. Medical Journal of Australia 151:77-80.

PRIOR, LA.M.,WELBY, T.J., OSTBYE, T., SAL-
MOND, C.E. & STOKES, Y.M. 1987. Migration
and gout: the Tokelau Island migrant Study.
Bnitish Medical Journal 295: 457.

ROSS, S.G. 1947. Preliminary report on fish poisoning
at Fanning Island (Central Pacific). Medical Jour-
nal of Australia 11:617.

RUFF, T.A. 1989a. Ciguatera in the Pacific: a link with
military activities. Lancet 1: 201-205.

RUFF, T.A. 1989b. Fish poisoning in the Pacific: a link
with military activities. Canberra; Peace Research
Centre, Research School of Pacific Studies, Aust-
ralian National University. Working Paper 63.

RUSSELL, F.E. & EGAN, N.B. 1991. Ciguateric

CLINICAL ASPECTS OF CIGUATERA

fishes, ciguatoxin (CTX) and ciguatera poisoning,
Journal of Toxicology and. Toxin Reviews 10: 37.

SOROKIN, M. 1975. Ciguatera poisoning in north-west
Viti Levu, Fiji Islands. Hawaii Medical Journal
34: 207.

South Pacific Epidemiological and Health Information
Service, SPEHIS annual reports, South Pacific
Commission, New Caledonia. 1973-1992.

STEWART, M.P.M. 1991. Ciguatera fish poisoning:
treatment with intravenous mannitol. Tropical
Doctor 21: 54-55.

STOMMEL, E.W., PARSONNET, J. & JENKYN, L.R.
1991. Polymyositis after ciguatera toxin exposure
(abstract). Archives of Neurology 48: 874-877.

SUTHERLAND, S.K. & LEWIS, R. 1992. Patients
with ciguatera: request for convalescent sera.
Medical Journal of Australia 157: 140-141.

TAYLOR R. & THOMA, K. 1985. Mortality patterns
in the modernized pacific island nation of Nauru.
American Journal of Public Health 75: 149-155.

619

THOMAN, M. 1989. Ciguatera in a breastfed baby.
Veterinary and Human Toxicology 31: 71.
VERNOUX, J.P. 1988. La ciguatera dans I’ile de Saint-
Barthélémy: aspects épidémologiques, toxicol-
ogiques et préventifs. Oceanologica Acta 1: 37.

WILLIAMS, R.K., PALAFOX, N.A. 1990. Treatment
of pediatric ciguatera fish poisoning. American
Journal of Diseases of Children 144: 747-748.

WILLIAMSON, J. 1990. Ciguatera and mannitol: a
successful treatment (letter). Medical Journal of
Australia 153: 306-307.

ZIMMET, P., FAAIUSO, S., AINUU,
J.,. WHITEHOUSE, S., MILNE, B. & DE BOER,
W. 1981. The prevalence of diabetes in the rural
and urban Polynesian population of Western
Samoa. Diabetes 30: 45-51.

ZIMMET, P., TAYLOR, R., JACKSON, L.,
WHITEHOUSES.L., FAAIVASO, S. & AINUU,
J. 1980. Blood pressure studies in rural and urban
Western Samoa. Medical Journal of Australia 2:
202-205.

PATHOLOGICAL CHANGES IN MURINE HEARTS INDUCED BY INTERMITTENT

ADMINISTRATION OF CIGUATOXIN
KIYOSH! TERAO, EMIKO ITO, MISAKO OHKUSU AND TAKESHI YASUMOTO

Terao, K., Ito, E., Ohkusu, M. & Yasumoto, T. 1994 08 01: Pathological changes in murine
hearts induced by intermittent administration of ciguatoxin, Memairs of the Queensland
Memoirs 34(3): 621-623. Brisbane. ISSN 0079-8835.

Ciguatoxin (CTX) at doses of 0,1 or 0.05pg/kg were given orally by intubation into male
ICR mice once a week for 25 weeks (0.1,1g/kg group) or 40 weeks (0.05 j.g/kg group), Until
about 10 weeks after the beginning of the experiments the mice in both groups showed no
abnormal clinical signs, After about 18 weeks, mice treated with 0.1j.2/kg showed marked
hypertrophy of the hearts; no pathological changes were seen in the hearts of the other mice.
There was swelling or rupture of the endothelium of the capillaries and widening caused by
exudation or collagen fibers in the interstitial space. Occasionally, degenerated or swollen
mitochondria were prominent in the myocardium. Accumulations of platelets in the capil-
laries were frequently observed, Mice treated with low CTX dose showed no pathological
changes even at the ultrastructural level until 40 weeks. 1HUS CTX has a potent cumulative
effect on the cardiac tissue.

Kiyoshi Teruo, Emiko Ite & Misaka Ohkusu, Research Center for Pathogenic Fungi and
Microbial Toxicoses, Chiba University, 1-8-1, Inohana, Chuo-ku, Chiba 260, Japan; Takeshi
Yasumoto, Faculty of Agriculture, Tohoku University, 1-1 Tsutsumidari Amemiya, Aoba-ku,

980 Sendai, Japan; 2 February 1994.

Ciguatera is one of the most serious tropical
food poisonings of fish in the vicinity of coral
Tecfs; its symptoms are neurological, gastrointes-
tinal and cardiac (Baden,19&3}. Several toxins
were isolated from contaminated fishes or cul-
tured dinoflagellates and the chemical structures
determined (Murata et al.,1989). Among them,
CTX is the most poient.A feature of ciguatera is
its obstinacy and recurrence of attacks (Bag-
nis, 1968, Bagnis et al., 1979). Our study of short-
term, successive administration of CTX at a low
dose, confirmed that CTX has accumulative effect
on the cardiac tissue. In experiments reported
here we re-examined the potency of the cumula-
tive effects of the toxin on the mouse heart tissue.

MATERIALS AND METHODS

TOXIN

CTX used herein was isolated from reef snap-
pers (Lutjanus bohlar) from Micronesia and
Okinawa. Phycotoxin was purified as described
by Legrand et al. (1989). One mouse unit is the
minimal lethal dose of CTX 24hrs after i.p, ad-
ministration into 20g mouse and is equivalent to
0.35 g/kg body weight (Yasumoto pers, comm.).
Oral LDs50 was not determined because the avail-
able dose of phycotoxin was limited,

EXPERIMENTAL ANIMALS
Male ICR mice (4 weeks of age weighing 2U-

23g) were obtained from Charles River Japan
Inc., Tokyo. 40 mice were divided into 3 groups,
Group 1; Five mice were given physiological
saline wilh a stomach tube once a week for 40
weeks and served as control. Group 2: Twenty
mice were given CTX (0.1 2e/kg of body weight).
Group 3: Fifleen mice were given low CTX
(0.05g/kg). Mice in groups 2 and 3 were given
the phycotoxin in a similar manner to group 1.
Two mice each trom group 2 were sacrificed by
cervical dislocation 5 hours after the intubation at
the 12th, 14th, 18th and 23rd week from the
beginning. After the treatment all surviving mice
were fed a standard diet (CE-2, Nihon Clea Inc.,
Tokyo). Two mice from the group 3 were also
killed and controlled in a similar manner to mice
in group 2 al the 18th and 40th week.

MORPHOLOGICAL EXAMINATION

After necropsy all internal organs were fixed in
10% neutral formalin and embedded in paraftin-
The slides for light microscopy were stained HE
and PAS, For TEM, pieces of the heart, kidney,
and liver were fixed by cold paraformaldehyde-
glutaraldehyde solution (final concentration: 2%)
and inmimersed in buffered 1% OsO4 at room
temperature. Then dehydrated in a series of
graded ethanol, embedded in Epon $12, and cat
with a diamond knife on a Porter I ultratome. The
ulirathin sections were siained with arany]

622 MEMOIRS OF THE QUEENSLAND MUSEUM

ul " =~ i ', = ‘ i —_ =
FIG. 1. A, Electron micrograph of heart from mouse given 0.1jxg/kg CTX orally by intubations once a week for
25 weeks. Between cardiac muscles (M) are thick bundles of collagen (Co). Blood capillaries (C) are embedded
in the collagen fibers, V: vein. Bar: 10m. B, Electron micrograph of the heart from mouse given 0. lug/kg
CTX orally by intubations once a week for 25 weeks, In. asmall vein (V) a thrombus (arrow heads) is developing,
M: cardiac muscle. Bar: 10pm, C, Electron micrograph of the heart from a mouse given 0.05 pg/kg CTX orally

by intubation for 40 weeks. No discernible changes are seen. M: cardiac muscles, Bar: 10jim,

PATHOLOGICAL CHANGES IN MURINE HEARTS BY CIGUATOXIN

acetate and Jead citrate and examined with a
Hitachi H700H TEM.

RESULTS

Oral administration of CTX at a dose of
0.1j.g/kg resulted in no abnormality, whereas i_p.
injection of the same dose caused severe watery
diarrhoea within 10 minutes. In contrast, mice
injected i.p. with the phycotoxin at 3 dose of
0.05.2/kg induced no diarrhoea at all.

Even at the ultrastructural level a single oral
dose of 0.1j.g/kg of CTX produced no abnormal
changes in the heart muscle. Long-term, intermit-
tenladministration of CTX ata dose of 0.05pi2/kg
induced no pathological changes in the heart tis-
sue until 40 weeks. In contrast, intermittent ad-
ministration of 0.lj:g/kg CTX once a week for
over 12 weeks resulted in severe morphological
changes in heart tissue.

Mice given 0.0Sg/kg CTX showed no abnor-
mal behaviour through the experiment. Mice
treated with CTX at an intermittent oral dose of
O.1g/kg also showed no abnormality during the
first 12 weeks. After that, however, shock often
occurred shortly after administration of CTX.
Usually the animals recovered spontaneously
within 10 minutes, After 18 doses or at the 18th
week two mice were sacrificed. Both ventricles
of the animals were dilated al necropsy. His-
topathologically, muluple single cell necroses
were offen seen in the mural or papillary muscles
of the left ventricle. TEM examination showed
swelling of myocardiac cells and edema between
bundles of muscle fibres. Mitochondria in these
cells became rounded and the matrix was
electron-dense. Occasional dissociation of inter-
ealated discs was noted. Blood capillanes were
embedded by bundles of collagen fibers and
electron-dense flocculent materials (Fig,1 A).
Capillanes and small veins were often occluded
by accumulation of blood platelets (Fig.1B).

Mice given 0,05.g/kg CTX produced no chan-
ges in heart tissue until 40 weeks (Fig.1C).

DISCUSSION

The most prominent morphological changes
after administration of CTX oceurred in the heart
muscles (Terao et al.,1990,1991,1992). Almost
all cardiac muscle cells and the endothelium of
blood capillaries in the cardiac interstitium were

623

markedly swollen and ruptured with severity of
change dependent on dose. These edema may be
caused by the increased influx of Na* channels of
carthac muscle cells (Ohizumi,1990). In our pre-
vious report, short-term successive administra-
tion of CTX at low dose resulted in a cumulative
effect of CTX on the mouse heart (Terao et al.,
1992). In the present study, an intermittent dose
resulted in similar severe morphological changes
in the heart tissue. CTX may bind very tightly to
Na* channels on the cardiac muscle cells or to
those on the endothelium of the capillaries in the
intershtium.

LITERATURE CITED

BADEN, D.G. 1983, Marine food-bome dinoflagellate
toxins, Pp, 99-150. In G.H. Bourne & LF. Danielli
(eds) “International Review of Cytology 82°
(Academic Press: New York).

BAGNIS, R. 1968. Clinical aspects of ciguatera (fish
poisoning) in French Polynesia. Hawaii Medical
Journal 28; 25-28,

BAGNIS, R., KUBERSKL, T. & LAUGIER, 8. 1979.
Clinical observations on 3,009 cases of cigusteris
(fish poisoning) in the South Pacific, American
Joumal of Tropical Medicine and Hygine 28:
1067-1073,

LEGRAND, A.M., LITAUDON, M., GENTHON,
J.N., BAGNIS, R. & YASUMOTO, 'T, 1989,
Isolation and some properties of CTX. Joumal of
Apphed Physiology |: 183-188.

MURATA, M., LEGRAND, A.M,, ISHIBASHI, Y. &
YASUMOTD, T, 1989, Structures of CTX and its
congener. Journal of Amencan Chemical Society
111; 8929-8931,

OHIZUMI, YY. SHIBATA, S. & TACHIBANA, K.
1981. Mode of a excitatory and inhibitory actions
of CTX in the guinea pig yas deferens, Joumal of
Pharmacology and Expenmental Therapeutics
221: 748-752.

TERAO, K., [TO, E. & YASUMOTO, T. 1990.
Pathomorphological studies on experimental
Mulotoxicosis and ciguatoxicosis in mice. Pp.
55-70. In TJ. Tosteson (ed.), ‘Ciguatera Puerto
Rico 1990) (Polyscience Publication: Quebec),

TERAO, K., ITO, E. OARADA, M., ISHIBASHI, Y.,
LEGRAND, A.M. & YASUMOTO, T. 1991.
Light and electron microscopic studies of
pathologic changes induced in mice hy CTX
poisoning. Toxicon 29; 633-643.

TERAQ, K., ITO, E. & YASUMOTO, T. 1992. Light
and electron microscopic studies of the manne
heart after tepeated administrations of CTX ar
CTX-4c, Natural Toxins 1: 19-26.

THE MOUSE CIGUATOXIN BIOASSAY: DIRECTIONS FOR USE TO CONTROL FISH
FOR CONSUMPTION

JEAN-PAUL VERNOUX

Vemoux, J. P. 1994 08 01: The mouse ciguatoxin bioassay: directions for use to control fish
for consumption. Memoirs of the Queensland Museum 34(3): 625-629. Brisbane. ISSN

(79-8835,

Ciguatera fish poisoning causes serious health problems around the world but the diversity
and heterogeneity of ciguatoxins are delaying chemical and immunological remedies.
Realistic methods for ciguatoxin screening of fish are needed for public health studies. The
mouse bioassay may prove useful since it is simple and relatively cheap. Qualitative and
semt-quantitative analyses for ciguatoxins from 50, 100 or 200g of fish tissue using
respectively 2, 6 or 12 animals are precisely described.

Jean-Paul Vernoux, Laboratory of Cellular and Molecular Physlolagy, University of Caen,
Esplanase de la Paix - 14032 CAEN Cedex, France, 2 February 1994.

Ciguatera is a human illness, sporadic in nature,
caused by the ingestion of a wide variety of fish
typically associated with coral reefs. These fish
are transvectors for muluple ciguatera toxins
(mainly ciguatoxins) acquired through their diet
(Vernoux & Abbad el Andaloussi,1986; Legrand
etal.,1990; Lewis & Sellin, 1992). Fish poisoning
is present in many coral reef areas of the world
(Bagnis,1981} with 10,000-50,000 persons af-
fected each year. [ts incidence has been estimated
at 1-4/000 population in the Pacific area (Bag-
nis,1979; Lewis,1984: Yasumoto et al..1984),
4.2/000 in the Virgin Islands (Olsen et al..1984),
and 3-10/000. at Saint Barthelemy Island (Ver-
noux, this memoir) and in the Saintes Islands
(Czernichow et al.,1984). With the increasing use
of air travel, tropical reef fish or consumers (in-
habitants or tourists) are constantly moving and
ciguatera is being documented in temperate
regions (Lange et al..1992). Thus, ciguatera is a
world health problem and there is a need for a
practical screening method for ciguatera toxins.
Amounts of ciguatoxin in fish that pose a public
health problem are very low. One approach to the
estimation of dangerous levels of toxin in fish
tissue is to dose it in those portions that had
elicited human ciguatera poisoning. Some
authors did this using mouse or mosquito biaas-
says (Yasumoto et al.,1984; Chungue et al. 1984;
Bagnis et al.,1987: Vernoux, this memoir). As-
suming that 500ng CTX can kill c.1000g of
mouse (3-p. LDso of CTX into mice=0.45p:g/kg),
these studies show that the minimum threshold
for the human pathogenic dose is c.50-100ng.
This level corresponds to 100-200g of mouse
killed by a 200¢ fish portion i.e. the flesh sample
has a specific toxicity of 0.5—lg of mouse killed

per ¢ of flesh). So only trace quantities of
ciguatoxins are needed to elicit human poisoning
(0.25ppb). Therefore only those detection
methods capable of detecting as little as 250—
500pg of CTX per gram of fish flesh need be
considered, HPLC and immunological methods
could be appropriate methods, Unfortunately,
until now the detection of these extremely low
levels of multiple ciguatoxins (>20, Legrand et
al.,1990) has delayed the use of HPLC methods,
Inimunoelogical methods that possess. the desired
sensitivity and specificity have already been
developed (Hokama,1990; Park et ai.,1992).
However, the potential of these methods for large
scale usc remains to be demonstrated, at least in
part duc to the possibilities of multiple
ciguatoxins in fish contamination. A biological
assay which encompasses all the different toxins
and gives total toxicity could be used as a robust
method for public health control of fish. Two
bicassays having the desired sensitivity are the
mosquito injection bioassay (Chungue ef al,,
1984) and the mouse injection bioassay (Hof-
fman et al.,1983), The latter is preferred since it
is More convenient, simple, specific and has been
widely used. After having used it for 20 years, we
present some recent developments in its use as
screening method.

METHODS

FLESH EXTRACTION

Raw or cooked minced flesh can be used since
ciguatoxins are heat resistant. If cooking, use a
boilable cooking pouch filled with raw minced
sample and boil in water for 30 minutes,

A typical procedure to extract 50g of flesh is:

626 MEMOIRS OF THE QUEENSLAND MUSEUM

TABLE 1. Testing of fish for consumption by the mouse bioassay

1, prepare LR for 50g of flesh and dilute with 2ml of 1% Tween 60 saline at 37°C and homogenize thoroughly.

2, inject i.p. into 2 male mice (18-24g) at dosage d=0.04ml/g of mice (i.e. 1g equivalent of flesh per gram of
mouse).

3, Observe symptoms during 4hr and conclude for ciguatoxin presence (=penile erection) for neurotoxin presence
(=respiratory distress) or for okadaic acid or fatty acid presence (crawling gait, slow breathing and general

cyanosis).

4, note death after 24hrs and weigh the survivors. Use the following table to indicate edibility.

Observed mortality after a
1g.eq. of flesh injection/g
of mouse

2/2

<0.5 MUg/g.e.F

Interpretation

not edible
not edible

bordeline
edible

1, homogenise 3 min in a Waring blender with
150ml of acetone, 2, filter onto a buckner funnel
and wash the remaining cake with 30ml of 80%
acetone; discard the cake, 3, remove acetone on
a rotary evaporator under reduced pressure and
reduce the volume of the remaining aqueous solu-
tion to 30ml (add water if below), 4, add 10ml of
ethanol, shake and extract twice with 40ml of
diethy] ether, 5, remove diethyl ether and residual
water under reduced pressure (addition of ethanol
allows to remove quickly residual water), 6, dis-
solve the diethyl ether residue in 25ml of 80%
methanol and wash twice with 50ml of hexane;
discard the hexane solubles, 7, remove methanol
and water under reduced pressure; resulting dry
residue is called lipid-soluble residue (LR), 8, if
weight of LR is >75mg dissolve in 10ml of 80%
methanol and wash again twice with 20ml of
hexane, 9, emulsify LR in 2ml of 1% Tween 60
saline and keep at —20°C until use.

DETERMINATION OF TOXIN CONCENTRATION
LR emulsified in 1% Tween 60 saline was
heated at 37°C and i.p. injected into male mice
weighing 18—24g (2 mice per dose). A series of
doses that vary in a geometric progression by a
factor of 1.1938 are assayed. Doses are expressed
in gram equivalents of flesh per gram of animal
(g.e.f/g). They were chosen in succession in the
following series of numbers running from 10”!
to 10° which increase successively by a constant
power of 10 (since 1.1938! x 10! = 10"). Four-
teen numbers in such a series are 0.10, 0.12, 0.14,
0.17, 0.20, 0.24, 0.29, 0.35, 0.41, 0.49, 0.59, 0.70,
0.84, 1.0. One gram equivalent corresponds here
to 0.04 ml of LR solution and the volumes to be
injected are respectively: 0.04 ml/g of mouse
multiplied by the numbers indicated in this series
i.e. 0.004 ml/g; 0.0048 ml; 0.0056 ml; 0.0068 ml,

etc. Note that the total weight of mouse to be
injected requires a total volume below that avail-
able in the experiment and the use of mice <20g
for the highest dosage may be necessary.

Doses <1g.e.f/g are injected in 0.8ml per 20g
mouse by carrying out dilutions directly in the
syringe with 1% Tween 60 saline at 37°C. The
approach described above allows determination
of the LDso and the minimum lethal dose (MLD)
which is the lowest dose capable of killing two
mice in the two mice group (or one mouse in the
one-mouse group) after 24 hr. The toxin content
is expressed in terms of Mouse Units gram (MUg)
where 1MUg is 1g of mouse killed by the MLD
(or the LDso) expressed in g.e.f/g. The toxin con-
centration is expressed in MUg per gram of flesh
(MUg/g.e.f) which is the reciprocal of MLD (or
LDsv).

The suitability of fish for consumption may be
controlled by mouse bioassay using 50g of flesh
(Table 1). Quantitative and semi-quantitative
conclusions are presented. Two hours are needed
to prepare LR. If dissolution of LR is difficult, 0.1
ml of ethanol can be added per 1.9ml of Tween
solution. A negative control is run using two mice
injected with a blank solution (without LR). Non-
toxic fish extracts give negative results (they
elicit no symptoms).

Acute toxicity may be determined in two steps
(Table 2) with a minimum of animals (Lorke et
al.,1983). For this method 200g of flesh should
be extracted and four hours are needed to prepare
the LR. In the initial investigation, which requires
an amount of extract corresponding to about 100g
of flesh and 6 mice, an approximate range of
doses producing the toxic effects is established.
Normally this initial investigation would include
doses used in the control of fish for consumption
method i.e. injection of |g.e.f/g as a first step; in

MOUSE CIGUATOXIN BIOASSAY

627

TABLE 2. Determination of the acute toxicity of LR, presiaredtts from n 2008 of flesh, in two steps.

Ist s ste EXPERIMENTAL
Doses d) in g.e.f/g of mouse Approximate
sea 3 deduced toxin

in MUg/g.e.f.

Doses d) in g.e.f./g Corresponding toxin
chosen for the second test concentration in MUg/g.e.f.
ee cewataneey el concentration (2 mice per dose) according to the (MLD)

_ RESULTS

* Number of animals died/number of animals used
One mouse per dose
© This dose allows control of fish for consumption,

a Oor 1/2 >? and <4 0.29
és 4) é si

[Lap 0.59 | 0.
2.0 :

[0.84] 0.71 |
(1.4) | (1.2)

0.59 1,00

oe

1.19
ot

1.43
(0.70

1.70
(0.59

(0.35) | (0.29)
4.17 5.88 | 7.14
it |__| | | (0.20) | | (0.17)

[ ] Possible only when control of fish for consumption is used instead of the first test.

that case according to the observed lethality 0/2
or 1/2 it may be possible to bypass the 0.49 and
0.24 g.e.f/g injections (but not if a 2/2 lethality is
observed). Based on these results, further specific
doses are administered to a group of one mouse
or two per dose depending on the predicted toxin
concentration (<1 or >1MUg/g.e.f). From the
results of these two tests, 6 or 7 successive levels
of dosage (di + d2) are then assayed. For a two-
mice group an LDso can be calculated by the
method of Weil (1952) using the equation
log LDso = log Da + log R (f + 1)

Da being the lower dosage level, R the
geometric factor and f a value given in Tables.
This permits a simple and rapid estimation of the
LDso with a corresponding confidence interval.

For the one animal group per dose, LDso is
estimated as the geometric mean of the doses for
which 0/1 and 1/1 are found (Lorke et al.,1983).
The minimum lethal dose can be used instead of
LDso.

RESULTS AND DISCUSSION

CONTROL OF FISH FOR CONSUMPTION

The proposed acetone method is much more
economic and is as rapid as the method of Lee et
al. (1987) who used methanol as the extracting
agent instead of acetone for okadaic acid, a toxin
chromatographically related to ciguatoxin. The
methanol method is convenient only if the tissue
portions to be extracted are <10g. So we prefer

the procedure with acetone (Vernoux,1981) and
we have been using this method since 1981.

In our method the LR yield must be below
0.15% of the flesh since the less impurities
present the more marked the symptoms in mice
for a given dose. Doses received by mice with this
method do not exceed 1.5 mg of LR/g of mouse.
Our proposed interpretation of symptomatology
and lethality includes:

- the unique propensity of ciguatoxin to induce
penile symptoms i.e penile cyanosis and/or tran-
sitory and incomplete erection (sometimes even
reaching priapism i.e. complete and permanent
erection seen following sub-lethal doses (Ver-
noux & Bagnis,1976; Vernoux et al.,1985). This
symptom was recently confirmed by Terao et al.,
(1991) who pointed out the penis as a target organ
for ciguatoxin.

- the symptoms in mice after i.p injection of
okadaic acid (Vernoux & Moulin,1989) or fatty
acids (Vernoux,1981) are different from that
elicited by CTX but resemble the effects of
maitotoxin, a toxin never detected in fish flesh
(Yasumoto et al.,1984).

- the existence of a narrow range of doses (d -
2d) between 0% and 100% lethality (Hoffman et
al.,1983; Lewis & Endean 1984; Vernoux &
Moulin, 1989).

- the general observation of a minimum
pathogenic dosage only 1/4 to 1/3 the LDso
dosage and the link between the pathogenic
dosage and loss of weight (Chungue et al.,1984;
Vernoux, 1988).

628

- additional observations suggest a 5 MUg or
10 MUg dosage/g of mice if survival time is
respectively about | hour and half an hour but it
may vary considerably with fish species (Ver-
noux and Tahla,1989). The relationship between
the mouse response and the quantity of toxin
present is given in Table | and 2.

ACUTE TOXICITY DETERMINATION

Extracting 200g is convenient for investigating
fish for consumption and for quantifying the toxin
concentration in the 0.5-7.14MUg/g.e.f range.
This range is sufficiently wide to include toxin
concentrations found in fish in the Australia,
Pacific area or the Caribbean. There is no upper
limit for the determination of toxin concentration,
since the more toxic the flesh the less RL con-
sumed in the test. Unlike the method for the
control of fish for consumption, acute toxicity
determination takes more than one day to con-
duct. Fortunately, stability of toxins in 1% Tween
60 saline is complete when samples are stored at
—20°C for up to 6 months.

The geometric factor R = 1.1938 was chosen to
provide a closely spaced series of dosage levels.
This geometrical series of numbers increases suc-
cessively by a constant power of 10 as already
mentioned above. Furthermore, as (1.1938)* =
2.0 another geometric factor R = 2 can be used
and numbers of the corresponding series are
therefore included in the first one. The two-step
method shown in Table 2 was developed using
these series. Since the first one is a closely spaced
series, this enhances the precision of the MLD
determination. In this case we observed that the
MLD values obtained with two animal groups
were equivalent to the LDso values obtained with
four animal groups (Vernoux and Tahla,1989).
This experimental correlation can be easily ex-
plained since the slope of the dose response curve
for ciguatoxin is high with a narrow dose range
(d—2d) between 0% and 100% lethality, thus in-
cluding MLD and LDso values (Vernoux and
Moulin, 1989). Nevertheless the two-mice group
or even one-mouse group also give reliable LDso
values (Weil,1952; Lorke,1983). Here with the
two-mice group, to calculate LDso we use the
method of Weil (1952) since it is easier and more
rapid than the method of Litchfield & Wilcoxon
(1949) and the former approach allows the con-
fidence interval to be estimated. However, we
prefer MLD determination to LDso calculations
since a greater accuracy is not necessary in view
of the range of variation from one dose to another.

It might be thought that MLD or LDso could be

MEMOIRS OF THE QUEENSLAND MUSEUM

determined from the curve of dose (d) versus
survival time (t) particularly since the test can be
conducted in one day. Nevertheless this method
is convenient only if the toxin concentration in
flesh is 2 2MUg, thus limiting its application. The
relationship is d = LD50 (1 + 1/t)’or d/LD50 = (1
+ If)’ i.e. number of MUg/g.e.f = (1 + 1/t)°.
Unfortunately b is fish-species dependent and it
varies from 2 to 3 (Vernoux,1991) thus com-
plicating the situation. So this method is of
limited interest in controlling fish for consump-
tion.

CONCLUSIONS

Presence of multiple ciguatoxins in fish flesh
led us to propose here a simplified mouse
ciguatoxin bioassay. Exhaustive description of
the method should allow it to be use to control
fish by any unspecialised hygiene laboratory. The
limited quantity of flesh used (200g) is con-
venient for investigating fish for consumption
and it allows to determine toxin concentration in
all situations. The fixed method should replace
the multiple mouse ciguatoxin bioassay methods
for which toxicological bases are not very clear.
Standardisation of the mouse strain could be
realised with a known toxin having a similar
physiological effect, brevetoxin for example. We
hope that our proposals will gain wide accep-
tance, since the mouse ciguatoxin bioassay
proposed here provides both a qualitative and
semi-quantitative bioassay for ciguatoxins in
fish.

ACKNOWLEDGEMENT

This study was realised within the scope of the
French National Program on Marine Phycotoxins
and supported by an Ifremer contract.

LITERATURE CITED

BAGNIS, R. 1979. L’ichtyosarcotoxisme de type
ciguatera ; Phénoméne complexe de biologie
marine et humaine. Oceanologica Acta 4: 375—
387.

BAGNIS, R. 1981. Etude morphologique, biologique,
toxicologique et écologique de!’ agent causal prin-
ceps de la ciguatera, le Peridinien Gambierdiscus
toxicus. Thése d’Etat en Biologie Humaine, 180p.
Université de Bordeaux II, France.

BAGNIS, R., BARSINAS, M., PRIEUR, C., POM-
PON, A., CHUNGUE, E. & LEGRAND, A.M.
1987. The use of mosquito bioassay for determin-
ing the toxicity to man of ciguateric fish. Biologi-
cal Bulletin 172: 137-143.

MOUSE CIGUATOXIN BIOASSAY

CHUNGUE, E., BAGNIS, R. & PARC, F. 1984. The
use of mosquitoes (Aedes aegypti) to detect
ciguatoxin in surgeonfish (Crenochaetus striatus)
Toxicon 22: 161-164.

CZERNICHOW, P., DROY, J.M., EZELIN, F, &
LEROY, J. 1984, La ciguatera aux Tles Saintes
(Guadeloupe) . maladie transmise par les pois-
sons, La presse Médicale 13: 222.

HOFFMAN, P.A., GRANADE, H.R, & MCMILLAN,
J.P. 1983. The mouse ciguatoxin bioassay: a dose
response curve and symptomatology analysis,
Toxicon 21: 363-369,

HOKAMA, Y. 1990. Simplified solid-phase im-
munobead assay for detection of ciguatoxin and
related polyethers. Journal of Clinical Laboratory
Analysis 4: 23-217.

LANGE, W.R., SNYDER, E.R. & FUDALA, P.J. 1992,
Travel and ciguatera fish poisoning. Archives of
Intemal Medicine 152: 2049-2053.

LEE, J.S,, YANAGI, T.,. KENMA, R. & YASUMOTO,.
T. 1987. Fluorimetric determination of diarrhetic
shellfish toxins by high performance liquid
chromatography. Agnecultural and Biological
Chemistry 51: 877-881.

LEGRAND, A.M,, CRUCHET, Ph, BAGNIS, R.,
MURATA, M., ISHIBASHI, Y. &
YASUMOTO, T. 1990. Chromalogrmphic and
spectral evidence for the presence of muluple
ciguatera Loxins. Pp. 374-378. In E. Graneli ctal.,
(eds) ‘Toxic marine phytoplankton’ (Elsevier:
Netherlands).

LEWIS, N.D. 1984, Ciguatera in the the Pacific: mn-
cidence and implications for marine resource
development. American Chemical Society, Sym-
posium Series 262: 289-316.

LEWIS, R.J. & ENDEAN, R. 1984. Ciguatoxin from
the flesh and viscera of the barracuda Sphyraena
Jello, Toxicon 22: 805-816,

LEWIS, R.J. & SELLIN, M. 1992. Muluple ciguatoxins
in the flesh of fish. Toxicon 30; 915-919.

LITCHFIELD, J.T.lr & WILCOXON, F. 1949. A
simplified method of evaluating dose effect ex-
periments. Journal of Pharmacology and Ex-
perimental Therapeutics 92: 99-113,

LORKE, D. 1983. A new approach to practical acute
toxicity testing, Archives of Toxicology 54: 275—
287,

OLSEN, D.A., NELLIS, D.W. & WOOD, R.S. 1984.
Ciguatera in the Eastern Caribbean, Manne
Fisheries Review 46: 13-18.

PARK, D.L.. FREMY, J.M., GAMBOA, P.M, &
GLEIZES, E. 1992. Innovative solid-phase im-

629

Munobead assay for the detection of okadaic acid
and related DSP toxins in shellfish. Presented at
2nd Intemational conference on shellfish purifica-
tion. Rennes (France). April 1992.

TERAO, K,, ITO, E.. ORADA, M., ISHIBASHI, Y.,
LEGRAND, A.M. & YASUMOTO, T. 1991,
Light and electron microscopic studies of
pathologic changes induced in mice by ciguatoxin
poisoning. Toxicon 29; 633-643,

VERNOUX, J.P. & BAGNIS, R, 1976, Fractonnement
d'extraits Jipidiques ciguatoxiques en milieu al-
calin. Biochimie 58: 479-484.

VERNOUX, J.P. 1981. L’ichtyosarcotoxisme de type
ciguatera aux Antilles et en Polynésie Francaise:
tests de ciguatoxicité et chaine trophique
ciguatérigéne. PhD Thesis, University of Bor-
deaux IJ,

VERNOUX, J.P, LAHLOU, N.. ABBAD EL AN-
DALOUSSI, §., RIYECHE, N, & MAGRAS, L,
Ph. L985, A study of the ciguatoxin in individual
Caribbean fish. Acta Tropica 42; 225-233.

VERNOUX, J.P. & ABBAD EL ANDALOUSSL, S.
1986, Heterogeneity of ciguatoxins extracted
from fish caught al the coast of the French Antilles.
Biochimie 68; 287-291.

VERNOUX, J.P. 1988. La ciguatera dans I'Tle de Saint
Barthelemy: aspects épidémiologiques,
toxicologiques et préventifs. Oceanologica Acta
11: 37-46.

VERNOUX, J.P. & MOULIN, F, 1989. Intoxications
alimentaires a dinoflagellés et dosage des toxines
associées aux intoxications de type ciguatérique et
diarthéique dues @ la consommation d’ animaux
marins. Science des Aliments, hors série 10, 9:
68-83.

VERNOUX, J.P. & TAHLA, F- 1989, Fractionation and
purification of some muscular and visceral
ciguatoxins extracted from Caribbean fish. Com-
parative Biochemistry and Physiology 946; 499
504.

VERNOLX, J.P. 1991. Moyens d'investigation pour
confirmer une intoxication de type ciguaténque.
In Fremy, J. (ed.). ‘Proceedings of Symposium on
manne biotoxins, (CNEYA).

WEIL, CS. 1952. Tables for convenient calculation of
median-effective dose (CDso) and instructions in
their use, Biometrics 8: 249-263.

YASUMOTO, T., RAJ, U., & BAGNIS R. 1984.
“Seafood poisoning in tropical regions’.
(Laboratory af Food Hygiene, Faculty of Agricul-
tare, Tohoku University: Sendai), 74p.

CIGUATERA IN THE FRENCH WEST INDIES
J.P. VERNOUX AND J. LEJEUNE

Vernoux, J. P. & Lejeune, J. 1994 08 01: Ciguatera in the French West Indies. Memoirs of
the Queensland Museum 34(3): 631-638. Brisbane. ISSN 0079-8835.

Ciguatera fish poisoning was investigated on the Island of Saint-Barthelemy, Leeward
Islands, Caribbean Sea, from 1979 to 1989. Clinical features include gastrointestinal and
neurological disorders. 440 fish caught in fish-pots or by hook and line were checked by
mouse and chicken bioassays. Jacks (Caranx spp.) and barracudas were highly ciguatoxic.
Weight and toxicity were not correlated except for the most toxic species Caranx latus. Small
carnivorous fishes classified as invertebrate feeders are likely involved in the transfer of
ciguatoxin in the food chain since they contained significant levels of toxin. Herbivores (e.g.
surgeonfishes or parrotfishes) which are not locally implicated in ciguatera, contained
sometimes low levels of ciguatoxin. Gambierdiscus toxicus occurred in coastal waters of
Saint Barthelemy but this species may not directly produce ciguatoxin.

J.P. Vernoux, Laboratory of Cellular and Molecular Physiology, University of Caen,
Esplanade de la Paix, 14032 Caen Cedex, France; J. Lejeune, Department of Mathematics,
University of Caen, Esplanade de la Paix, 14032 Caen Cedex, France; 2 February, 1994.

In the Caribbean, ciguatera poisoning is known
from the Bahamas (Pesty,1975) to Martinique
(Menger,1979) including Florida (Lawrence et
al.,1980) Cuba (Bagnis,1979a) and Puerto Rico
(Gilman, 1942; Payne & Payne, 1977) witha high-
er incidence in small Islands such as the Saintes,
the Virgin Islands (Czernichow et al.,1984;
Hanno, 1981; Morris et al.,1982) and the Leeward
Islands (Morice,1965). One of the Leeward Is-
lands, Saint Barthelemy, is a small (c.25km), arid
(without river or spring), tropical Island at
17°55’N, 62°50’ W (Fig.1). The 3 adjacent islands
of Saint Barthelemy, Saint Martin and Anguilla
are surrounded by a wide shelf (max. depth 60m),
with fringing reefs which continue across the
shelf. These submerged reefs often form scattered
coral-shoals 10—20m underwater. At the edge of
the open sea the shelf falls abruptly to >200m.
This coral reef ecosystem shelters a great variety
of reef fishes (Vernoux et al.,1988).

Laboratory research on ciguatera was initiated
in this area in the 1980's. It was shown that in fish
the poison is lipid-soluble and quite similar to
ciguatoxin isolated in the Pacific (Vernoux et
al.,1982; Hoffman et al.,1983). Heterogeneity of
ciguatoxins extracted from viscera or flesh of
Caribbean fish was also demonstrated (Vernoux
& Abbad el Andaloussi,1986) and confirmed
(Vernoux & Tahla,1989; Gamboa et al.,1990).
Gambierdiscus toxicus (Bergman & Alam,1981;
Besada et al.,1982) was suspected as the cigua-
toxin elaborator and as the producer of maito-
toxin (Miller et al.,1984). A study of ciguatera
poisoning and occurrence of ciguatoxins in fish
was carried out on Saint Barthelemy. Results are

presented here, together with results from experi-
ments with G. toxicus sampled at Tahiti in 1976
(in Dr. Bagnis’s laboratory).

MATERIAL AND METHODS

Fish collected at numerous locations on Saint
Barthelemy between 1979 and 1989 were iden-
tified from Stokes (1980). Most were caught in
fish traps on the sea bottom (depth: 30-50m)
within the reef habitat. King mackerel (Scom-
beromorus cavalla) and cero (Scomberomorus
regalis) were collected by trolling, and greater
amberjack (Seriola dumerili), black jack (Caranx
lugubris) and african pompano (Alectis crinitus)
by hook and line. Whole animals or separated
tissues were frozen at —20°C for transport to the
laboratory and stored until processed.

Lipid-soluble residues (LR) were prepared
from flesh or viscera by a routine acetone or
methanol method, respectively (Vernoux et
al.,1985a). For the qualitative testing, fish liver or
LR were fed to chicks as reported elsewhere
(Vernoux et al.,1985b; Vernoux & Lahlou,1986).
For quantitative testing a mouse bioassay was
used (Vernoux etal.,1985a). The toxin concentra-
tion was expressed in Mouse Units gram per gram
of tissue (MUg/g), where 1 MUg is the weight-
specific minimum lethal dose (IMUg=1MU/20).
Fish was assumed to be ciguatoxic when typical
symptoms of ciguatera (Vernoux, this memoir)
were observed in both chick and mouse.

For experiments in Tahiti in 1975 and 1976, a
mixture of algae and detritus was scraped from
the surface of dead corals collected at the Gam-

2. MEMOIRS OF THE QUBENSLAND MUSEUM

BD ate
ES Se ET vided
=

N =
Usa, ;
co ee |
oS “ “ \
re \
in |
Ne Tria oe
‘ ?
ow.
+
Fn : Ma atin
_—— =
[rip na *
= “ os) ey, ia
c~} 3 phe
Kare Co eee,
4 SS | creer ey
; Beater “Sy Sayer" <
ay =
af Antpt les
ivy Lesser
i
Ava ribs HE Caribieon Spa
I rubacas Angilpas
iy ey HP
= _
oM a
f } 4
~ SS
{
1 sy =

In some cases, poisoning was m-
duced after a second meal of the
same fish, Patients complained of
sensitivity disturbances (painful tin-
gling about the mouth and throat, hot
and cold reversal) nausea and vomit-
ing, abdominal pain, diarrhoea,
myalgia, weakness and hypotension.
Visual disturbances (blurred vision)
and persistent itching were also
noted. After symptomatic therapy,
complete recovery occurred after a
Baki few days, weeks or months. After a
beer? primary poisoning some individuals
| develop a fish feeding allergy and
cannot eat any fresh or canned fish.
a Toxin analysis of fishes implicated
FA, in these ciguatera outbreaks revealed
oe ciguatoxins at >1MUg/g of flesh.

FIG.1, The Caribbean region,

bier Islands. This mixture was fractionated by
hand according to size and, then by sisccessive
passes through sieves of different mesh sizes
(Blutex Nylon; 85m, 36m and !0j.m). Vortex
shaking helped to separate associated organisms.
Our routine acetone method was used to extract
ciguatoxin in cach biodetritus fraction. Remain-
ing dry organic matter was treated with boiling
methanol ( Yasumoto & Endo, 1973) to extract the
maitotoxin. This allowed a separation of
ciguatoxin (CTX) and maitotoxin (MTX) with
httle reciprocal contamination, since maitotoxin
is poorly soluble in acetone (Yasumoto et al.,
1976), These toxins were identified by symptom-
atology induced in mice and by their chromato-
graphic behaviour on a silicic acid column. For
analysing toxins in the gut contents of her-
bivorous fish the method of Yasumoto et al.
(1977a) was used.

Analysis of variance (ANOYV A) was done using
SPSS-PC software. A non-parametric ANOVA
K-W test was also used in parallel},

RESULTS

EPIDEMIOLOGY

With the help of some local physicians, fish
poisoning was. surveyed at Saint Barthelemy
(3000 inhabitants), 10-30 cases were identified
per year either from consultauions or from
patients treated at the hospital. Generally they
were fishermen or tourists having consumed
small jacks, big mackerels, snappers or seabass.

DISTRIBUTION OF TOXICITY

During the course of this study we observed
that fish only contained toxin in their flesh if their
livers were also toxic. Furthermore, ciguatoxin
concentration was always higher in viscera than
in flesh (at least twice as toxic).

The most toxic species (1-10 in Table 1) are
large piscivorous species except for A. afer which
is a small (<300g) invertebrate feeder.

Intermediate toxicity was detected in species
{1-15 (Table 1). Lower toxicity was present in
other species (16-30 in Table 1). The above clas-
ses include small species (<500g) such as M.
martinicus, P. arenatus, B. rufus and H- radiatus
and M. plumieri (<ikg) which feed mainly on
benthic invertebrates; these species can be more
toxic than some larger piscivorous species. Only
one species of herbivorous fish, the ocean sur-
geonfish (29 in Table 1), had low toxicity. No
toxicity was found in other herbivorous fish (31—
35 in Table 1). The longjaw squirrelfish (36 in
Table 1), a specific crustacean feeder, was not
toxic.

Statistical analysis of variance of individual
toxicity for species 1-11 (Table 1) (independent
of weight) indicated that the means were sig-
nificantly different (p<0.01) for the 11 species
group unless C. latus and C. bartholomaet were
removed from this group. Without C. Jatys, the F
test was not significant (p=0, 13) but the K-W test
was significant (p=0.04). This discrepancy can be
explained by the difference in size between the C,
bartholomaei sample (n=45) and the other

CIGUATERA IN THE FRENCH WEST INDIES

633

TABLE 1. Flesh ciguatoxicity in various fish species from Saint Barthelemy Island: results presented in
decreasing order according to toxicity.

12

No. Species Specimen TOXIN CONCENTRATION IN MUg/G OF FLESH |
weight | <0.05 1-1.49 |1.5-1.99 >10
1_|Seriola dumerili 6-29
| 2 | Caranx latus 1.3-6 a
3 | Caranx ruber 0.75-2.3
| 4 |Caranx bartholomaei 0.754.8
5 |Alectis crinitus 3.3-12
6_| Sphyraena barracuda 3-10
|_ 7 | Epinephelus morio 6.5-8
8 |Alphestes afer 0.1-0.25 (16)0.1)12)(2) (6)
9 | Scomberomorus cavalla 15-20 1
10 | Scomberomorus regalis 34 1 1
11_|Gymnothorax funebris 3.5-14.5 6 2

Malacanthus plumieri

Lutjanus jocu
Lutjanus griseus
Lutjanus buccanella
Priacanthus arenatus
Bodianus rufus
Halichoeres radiatus
Gymnothorax moringa
Mulloidichtys martinicus

Mycteroperca venenosa

Caranx lugubris

0.3-0.6

(5)(20)(5)(4)(6) | (5)(15)
(4)(2)02)

1 2
1.9-2 2 1
0.3-1.5 312) |)
0.4-0.6 3)(2

0.2-0.4 (2)(15)(12)

0.5-1 (4)G3)

15-2

| Lutjanus analis 45 2 == ap
24 | Seriola rivoliana 2.745 2
25 | Mycteroperca tigris 0.7 (2)
26 | Epinephalus guttatus 0.7-0.9 (3)
| 27 | Epinephalus_adscension 0.50.7 (4)
28 | Calamus calamus 0,2-0.4 (6)
29 |Acanthurus_ bahianus 0.05-0.15 (37)(6) |
30 | Balistes vetula 1.5-2 1 1
31_|Acanthurus chirurgus 0.2-0.4 | BOM
| 32 |Acanthurus coeruleus | 02-04 | (10) |
33 _|Scarus coeruleus 2 1 |
34 | Scarus vetula 06 | (2) | Lae
35 | Sparisoma viride 1 1 | ai
36 | Holocentrus ascensionis 0.1-0.2 (11) |

Specimens were tested individually except where indicated by brackets in which case the number of pooled
specimens is given.

samples (n<10). Caranx latus had the highest
mean toxicity of species examined, suggesting
that it is the most dangerous species at Saint

Barthelemy.

The relationship between individual toxicity
and weight was investigated within each of nine
species for which n>2; the smooth curves (Fig.3)

were fitted with the linear regression model:

634

oO 5 10 15 20
L L naa

MEMOIRS OF THE QUEENSLAND MUSEUM

year. From 1979-1985 the

n

<4 G FUNEBRIS S BARRACUDA

S CAVALLA| median concentration per
| year of ciguatoxin in their
t flesh (n=62) was stable
around 1MUg/g of flesh.

in MUg/g of flesh
2

During the same period

Cc. LATUS Cc, RUBER

E,. MORIO/~

bea
| -

other bioindicator species
such as M. plumieri and A.
afer (benthic fishes) which
were thought to feed direct-

#4 A AFER A CRINITUS

TOXIN CONCENTRATION

ly on the ciguatoxin
producers were also
studied. Their toxicity level
per year was constantly
0.2-0.6MUg/g of flesh. So

_C. BARTHOLOMAEI-

T T T TT T T T TT

Weight in Kg

FIG.2. Relation between weight and toxicity for some toxic fish species.

toxicity = Bo + 8: weight +E,

where toxicity is in MUg/g and weight is in kg.
A significant linear correlation between toxicity
and weight was found (81/0) only for C. latus
(p<001; Bo=-0.85502; Bi= 0.79673) and A.
crinitus (p=.014; Bo=-1.31314; 81=0.29066).

USE OF BIO-INDICATORS

Since jacks occupy the upper part of the
ciguatera food chain, we chose C. bartholomaei
and C. latus adults (>1kg) to investigate the
degree of bioaccumulation of ciguatoxin per

=

gut contents
. ) — :
Scibbus (418 2) ee

gut contents

yt

0 5 10 15 20 25 30 35
MUg per gram of gut contents

Bictx | MTX

FIG.3. Distribution of toxins in gut contents of surgeonfish C.
striatus and parrotfish S. gibbus caught at the same place in the

Gambier Islands.

the amount of toxins in the
food chain at Saint Bar-
thelemy appeared stable
over this period. Neverthe-
less, in the last years of this
study toxicity of C. latus
was 3 times higher (n=10), but no change was
found in the toxicity of C. bartholomaei (n=15).

DISTRIBUTION OF TOXINS AND G. TOXICUS IN
GuT CONTENTS OF HERBIVOROUS FISHES AND
IN CORAL SAMPLES
Extracts of coral samples collected at Saint
Barthelemy blanketed with algae did not contain
detectable ciguatoxin, even though G. toxicus
was present in low numbers on these samples.
The maitotoxin and ciguatoxin analysis of gut
contents of herbivorous fish, obtained in 1975-—
1976 from French Polynesia (Fig.3) shows that
materials ingested by parrotfish S. gibbus
or surgeonfish C. striatus contained
ciguatoxin (fat soluble toxin) and
maitotoxin (acetone precipitated toxin).
Furthermore, the concentration of
ciguatoxin in gut contents of parrotfish
was always higher than that of maitotoxin,
whereas the ratio of the two toxins was
reversed in the gut contents of the sur-
geonfish. Thus ciguatoxin level appears
unrelated to maitotoxin level.
CTX content (in MUg) is quantitatively
dominant in scraped coral substrate and in
4s the fraction >85ym (=algae+detritus),
four times as much as in the G. toxicus
fraction (Fig.4A). Attempts to remove
CTX from coral substrate by scraping in
the presence of a low pressure spray of
water were unsuccessful: 30-60% of total
CTX content remained attached to the
coral substrate (3 experiments). With
crude material (not frozen) results were

CIGUATERA IN THE FRENCH WEST INDIES

similar. The MTX content was well correlated
with G. toxicus (which accumulated in 36-85.m
fraction) and the ratio of MTX to CTX was >1
as it was for the gut toxin contents of C. striatus.

Specific CTX content (in MUg/g of dry matter)
seems to reside not only with the G. toxicus
fraction but also on the other size fractions, par-
ticularly the fraction <36y.m (particles + washing
water) (Fig.4b).

G. toxicus thecal structure can be extremely
resistant to different conditions: freezing, water
dilution, acetone extraction and ultra-turrax
sonication.

DISCUSSION

The involvement of fat soluble ciguatoxins in
ciguatera in the Caribbean has been confirmed by
chemical studies on the toxins in C. bartholomaei
(Vernoux et al.,1982), B. rufus, M. martinicus,
M. plumieri, E. morio, G. funebris, S. barracuda,
S. cavalla, S. dumerili (Vernoux & Abbad,1986),
and A. crinitus, C. latus (Vernoux & Talha, 1989).
In the Virgin Islands, similar species were clini-
cally found to cause ciguatera poisoning (Brody,
1971; Morris et al.,1982; Engleberg et al.,1983)
and ciguatoxin from Lutjanus buccanella has
been well documented (Hoffman et al.,1983).

At Saint Barthelemy the same species of fish
were ciguatoxic as in the Pacific (Randall,1958;
Halstead,1978). Nevertheless, jacks are much
more toxic in the Caribbean (Arcisz,1950) than
in the Pacific Ocean (Caranx ignobilis was the
only species suspected in the Pacific by Bagnis
(1981)). The importance of feeders on small ben-
thic invertebrate feeders in the ciguatera food
chain at Saint Barthelemy is worth pointing out
since this suggests that ciguatera transmission
begins primarily at the invertebrate level. Inver-
tebrates appear to be less important in the Pacific
area, though some feeders on invertebrates, such
as Lethrinus kallopterus at the Marshall Islands
(Randall,1980), Cheilinus undulatus at Tahiti
(Bagnis,1968) and other Lethrinidae at New
Caledonia (Bagnis,1979a) are ciguatoxic.

In the French West Indies all ciguateric species
were shore fish associated with reefs. With the
exception of two semi-pelagic open water
species, the king mackerel and cero, they were
bottom dwelling species generally found at a
depth of <50m (>100m for black jack and greater
amberjack). Ciguatoxins are therefore well corre-
lated with the benthic fish. Further illustration is
provided by Caranx ruber which is dangerous at
Saint Barthelemy when caught in fish traps

635
Toxin content in MUg
6,000 93.88 %
€
5,000 a
4,000 ra
3,000
2,000 |s2.82 % 29.58 %
4,000 Li | —— ty +ha% 14.71%
a 326% ae
CTX Concentration in MUg/g MTX
400 * 1,800
1,600
b ¢
/ MTX\ 1,400
300 | \ <=36 ym
1,200
1,000
200 a
cx 800
600
100 : 400
a
200
os — t 0

washed
coral algae and
substrate —_ detritus
(7.4 Kg)

washed small washing
particules water
(4 liters)

0-36 pum
(dry matter 3g) (dry matter2.59) <10um

>85 pm

(dry hair 209) 1.2 10'cells

FIG. 4. Distribution of toxins MTX and CTX in frac-
tionated material from the Gambier Islands either
considering total toxin content (a) or toxin concentra-
tion (b).

(sedentary individuals) but not in nets (migrating
shoals of fish).

Unlike the Pacific, at Saint Barthelemy her-
bivorous fish are regularly consumed without
suspicion. Similar observations have been
reported in other areas of the Caribbean (Bagnis
1979a; Czernichow et al.,1984) in New
Caledonia (Bagnis,1979b) and in the Indian
Ocean (Lebeau & Telmar,1978). Nevertheless
detectable ciguatoxicity in the surgeonfish A.
bahianus indicates that low (subsymptomatic)
ciguatoxin levels may be present in some Carib-
bean herbivorous fish species, illustrating that
fish edibility depends on toxin level as already
described (Bagnis & Vernoux,1975). Here we
cannot exclude the possibility that ciguatoxin is
present in the other herbivorous fishes, though
ciguatoxin levels are probably (extremely) low.
The higher toxicity of A. bahianus suggests a
different diet. We studied 3 Caribbean surgeon-
fishes (Stokes, 1980; Randall, 1983) that have dif-
ferent feeding habits (Randall,1967): A.

AM MEMOIRS OF THE QUEENSLAND MUSEUM
aimaenea'te Tying fnatlageiiaes —al-glgns', tents of herbivorous fish at Tabiti
organisins ‘, have not previously been considered.
\ pre y ‘
‘, Chanteau (1978) studied scraped
an, WMeUS coral as well as other fractions for
(norganic! cTK wonucer toxicity. She obtained toxin par-
organic nutrient blooming tg: . ~_ +
—~ Colontzution titioning results similar to ours, even
ia wal agen | though scraping was practised with a
Stress produced Seleciidin at E P & P
Lieing oral @ vorieue - MEtallic brush, i.e. MTX content was
é Se, highly correlated with the G, toxicus
frending ~S

Desin ot coral
_

—
Oy Physieal tacors ~~~ Spreading
| ‘a ey
|
| att

—

Flimsy =" CTX prasiucer
eid plaliteration

Benihic weirs
with attached living
organisms

FIG, 5, A proposed mechanism for the initiation of CTX producer

proltleration,

chirurgus and A. hahianus has a thick walled
gizzard-like stomach and ingests morganic sedi-
ment (sand, small shells etc.) with the algae they
crop from solid substrata (especially A. bahianus)
(in Randall,1967;, and personal observations);
and A. coeruleus has a thin walled stomach in
which we have observed a higher content of gveen
algae and it does not usually ingest sand. The
former are therefore grazers whereas the latter is
a browser following the definition of Lobel
(1981). In the West Indies, parrotfishes are also
grazers that feed mainly on algae attached to dead
coral (Randall,1967), except for large Sparisoma
viride and occasionally large Scarus vetula which
graze on live coral (Prydl,1979). Taking into ac-
count these feeding habits and the corresponding
toxicity results, we deduce that the source of the
toxin in the food could be an organism linked to
bottom detritus, The ciguatoxin producer G,
toxicus, found on dead coral (Bagnis et al..1980)
and closely associated with reef sediments or
with macroalgae (Taylor,1979: Yasumoeto ct
al., 1979) may be the ciguatoxin-producer at Saint
barthelemy. This conclusion is consistent with
the presence of G. toxicus (Bagnis, 1981; Besada
et al.,1982; Bourdeau & Durand-Clement,1991)
and with previous studies (Yasumoto et al.,
1977b; Bagnis etal.,1980) which have implicated
G, toxicus as the ciguatera produced in the
Pacific. Nevertheless, the somewhat conflicting
resulls obtained in the study of scraped coral and
discussion conceming ingested toxins in gut con-

Dead. corul
with

containing fraction and <10% of total

Was present in scraped coral
substrate, and in this substrate the
proportion of MTX to CTX were in-
verted (<1), These observations and
ours demonstrated that; MTX is a G.
texicus marker, especially since it is
never excreted out of the dinoflagel-
late (Yasumoto et al.,1979b). How-
ever, CTX contamination level was
independent of the amount of G.
toxicus in these studies, This latter
statement was corroborated by
analysis of toxins in the gut contents
of herbivorous fish: our results and
those uf others (Yasumoto et al.,1975; Yasumoto
el al. 1977a) showed that the relative proportions
of MTX and CTX are inverted in surgeonftsh (C.
striatms) compared with parrotfish (5. gibbus)
caught in the same area, The difference in propor-
tion corresponds to the difference in their feeding
habits: the former is a browser which ingests
chlorophyll-bearing material (10%) as well as
detritus and numerous G. toxicus were found in
its stomach and gut contents, while the parrotfish
is a coral feeder exclusively (103-260pg of
algae/100g of ingested sample according to
Yasumoto et al. (1977a) and no G. texicus was
visible im its stomach or gut contents (Wer-
noux,!981). Thus ciguatoxin producer could be
associated also with living coral (perhaps the
zooxanthellae’), Bagnis et al. (1980) stated that
they could not find any indication that CTX {or
MTX) was excreted into the culture medium,
while Shimizu et al., (1982) and Campbell et al.
(1987) using fluorescence labelled sheep anti-
ciguatoxin antibody both found that the outer
wall of a certain percentage of G. toxicus con-
tained ciguatoxin and/or ciguatoxin-like com-
pounds. Thus the ciguatoxin producer could be a
very small organism dependent (ornot) on certain
epiphytic dinoflagellates such as G. toxicus
and perhaps zooxanthellae. The findings that (i)
ciguatoxin inhibits cellular multiplication of
unicellular marine algae (Durand et al.,1985), (ii)
coral death often seems necessary to induce

roducer

CIGUATERA IN THE FRENCH WEST INDIES

ciguatera production (Bagnis,1¥41), (ii) dead
coral provides new surfaces for dinoflagellates
implicated in ciguatera fish poisoning (Kohler &
Kohler, 1992) and (iv) the presence of CTX in
dead corals with G. toxtcus logether suggest thal
the ciguatoxin producer could contribute to the
death of living coral which could in turn enhance
G, toxicus proliferation (Fig.5}.

LITERATURE CITED

ARCISZ. W. 1950. Ciguatera; tropical fish poisoning.
U.S. Fish and Wild Spectal Scsence Reprint 27:
3.

BAGNIS, R. 1968. Clinical aspects of ciguatera (fish
poisoning) in French Polynesia. Hawai Medical
Joumal 28: 25-28.

BAGNIS, R. & VERNOUX, J.P. 1975, Ciguatoxine et
poissons de récifs comestibles, Bulletin de Ja
Société de Pathologie exotique 68: 320-325,

BAGNIS, R. 1979a. Données récentes concernant
quelques aspects biologiques de Ja ciguatera aux
Antilles, Caratbes Médical 2: 30-36,

BAGNIS, R. 1979b. Ciguatera fish poisoning in New
Caledonia. Clinical and epidemiological aspects.
Revue, Epidémiologique. Santé Publique 27; 17-
29,

BAGNIS, R., CHANTEAU, S., CHUNGUE, E., HUR-
TEL, J.M.. YASUMOTO, T. & INOUE, A. [980,
Ongins of ciguatera fish poisoning; a new
dinoflagellte, Gambierdiscus toxicus Adachi and
Fukuyo, definitively involved ay a casual agent.
Toxicon 18: 199-208,

BAGNIS, R. 1981. Etude morphologigue, biologique,
toxicologique et écologique de ]'agent causal prin-
ceps de la ciguatera, le P éridinien Gambierdiveus
toxicus, Thése Sciences, | 80p, Université de Bor-
deaux II,

BERGMANN, J.S. & ALAM, M. 1981, On the toxicity
of the ciguatera producing dinoflagellte Gambier-
discus toxicus Adachi and Fukuyo, isolated from
the Florida Keys. Journal of Environmental
Health Science 16; 493—500.

BESADA, E.G,, LOEBLICH, L.A, & LOEBLICH,
A.R. 1982, Observations on tropical benthic
dinofagellatcs from cigualera endemic areas:
Coolia, Gambierdiscus and Ostreopsis. Bulletin of
the Marine Science 32; 723-735.

BOURDEAU, P. & DURAND-CLEMENT, M. 1991.
Ichtyvosareotoxisme de type ciguatera: étude
Ecoloxicologique et morphalogique de Gambier-
discus toxicus aux Antilles Francaises (St Bar-
thelemy), In J.M. Premy, (ed.}, “Proceedings of
symposium on marine bioloxins, 30-31 Janvier
1991". (CNEVA).

BRODY, R.W, 1971. Fish paisoning in the Eastern
Caribbean. Proceeding of Gulf and Caribbean
Fisheries Institute 24: }O0-1 16

CAMPBELL. B., NAKAGAWA, L.K_,
KOBAYASHI, MLN, & HOKAMA, Y, 1987,

Garnbierdiscus toxicus in gut content on the Sur-
poonfish (Ctenochaetus strigosus (Herbivore } and
its relation Lo toxicity. Toxicon 25; 1125-1127.

CHANTEAU, S. 1978, Réle d'un dinoflagellé benthi-
que dams la biogentse de la ciguatera, Ph.D Thesis,
Unversity of Clermont Ferrand I, 63p.

CZERNICHOW, P., DROY, I.M., EZELIN, F, &
LEROY, J. (984. La ciguatera aux les Saintes
(Guadeloupe); maladie tansmise par les poissons,
La Presse Médicale 13: 222.

ENGLEBERG, N.C,. MORRIS, J.G. Jr, LEWIS, J.
McMILLAN, J.P., POLLARD, R.A. & BLAKE,
P.A. 1983, Ciguatera fish poisoning: amajorcom-
Ton-seurce Outbreak in the U.S. Virgin Islands.
Annals of Intemal Medecine 98: 336-337.

FRYDL, P. 1979, The effect of parrotiish (Scaridac) on
coral in Barbados (W.1,), International Review of
Ges, and Hydrobiology, 64: 737-748.

GAMBOA, P.M, PARK, DL, FREMY, J.M. 19)
Extraction and purfieation of toxic fractions from
barracuda (Sphyreena barrecwda) implicated in
ciguatera poisoning Presented at 3rd conference
on Ciguatera, Apnl 30) May 4. Lajac, Puesto
Rico,

GILMAN, R.L. 1942. A review of fish poisoning in the
Puerto-Rico, Virgin Islands area. U.S, Nivil
Medical Bulletin 40, 19-27,

HANNO, W.A. 1981. Ciguatera fish poisoning in the
Vargin Islands. Journal of American Medical As-
sociation 2453/5): 464.

HALSTEAD, B.W. 1978. “Poisonous and venomuus
manne animals of the World’ 1043p. (Princeton
University Press; Darwin, N.J.).

HOFFMAN, P.A., GRANADE, H.R. & McMILLAN,
JP, 1983, The mouse ciguatoxin bioassay; a dese
response curve and symptomatology analysis.
Toxicon 21: 363-369.

KOHLER, S. & KOHLER, C, 1992, Dead bleached
cori! provides new surfaces for dinoflagellates,
implicated in ciguatera fish poisonings. Environ-
mental Biology of Fishes 35: 413-416.

LAWRENCE, D.N.. ENRIQUEZ, M.B., LUMISH,
R.M, & MACEO, A. 1980. Ciguatera fish poason-
ing in Miami. Journal of American Medical As-
sociation 244: 254-258.

LEBEAU, A. & TELMAR, J.M. 1978, La ciguatern
dans |'Océan Indien. Revue Travaux Institue
Paches Maritimes 42: 325-345,

LOBEL, P.S. 1979, Trophic biology of herbivorous reel
fishes; alimentary pH und digestive capabilites.
Journul of Fish Biology 19; 365-397.

MENGER, D. 1979. Ichtyosurcotoxisme: a propos de S
casobservés dans tes Antilles Frangaises (Marine
que). Thése Médecine, 86p., Faculté de Médecine
de Grenoble [,

MILLER, D.M,, DICKEY, R,W. & TINDALL, D.R
1984. Lipid extracted from a dinoflagellate Gam-
bierdisows (exicus. Amencan Chemical Soctety
Symposium Series 262; 241~255.

MORICE, J. 1965, Catalogue descriptif des poissons

638

venéneua du bane de Saint-Barthélémy. Revue
Travaux Institut Péches Maritimes 29: 4—1 30.

MORRIS, J.G.Jr., LEWIN, P.. SMITH, C.\W., BLAKE,
P.A, & SCHNEIDER, R. 1982, Ciguatera fish
poisoning: epidemiology of the disease on Saint
Thomas, U.S, Virgin Islands, Amencan Journal of
Tropical Medicine and Hygiene 31: 5374— 578,

PAYNE, C.A. & PAYNE, S.N. 1977. Ciguatera in
Puerto Rico and the Virgin Islands (letter). New
England Journal of Medicine 296; 949-950,

PESTY,H. 1975. Mer des Caraibes, alerte 4 la ciguatcra
~ Semaine des Hopitaux, Supplément aux intor-
mations du 26 Mai, pp. 10-12.

RANDALL, J.E. 1958. A review of ciguatera, tropical
fish poisoning, with a tentative explanation of its
cause, Bulletin of Marine Science Gulf Carrib-
bean 8: 236-267.

RANDALL, J.E. 1967. Food habits of reef fishes of the
West Indies. Studies in Tropical Oceanography 35:
665-847,

RANDALL, J.E. 1980. A survey of ciguatera at
Enewetak and Bikini, Marshall Islands with noles
on the systematics and food habits of ciguatoxic
fishes. Fish Bulletin 78; 201— 249,

RANDALL, J.E£. 1983, “Camibbean ree! fishes” (revised
ed.), 3350p. (T.P.H. Publications. New Jersey).

SHIMIZU, Y., SHIMIZU, H., SCHEUER, P-L,
HOKAMA, Y., OYAMA, M. & MIYAHARA,
J.T. 1982, Gambierdiscus toxicus, a ciguatera-
causing dinoflagellate from Hawaii, Bulletin of
the Japanese Socicty of Scyentific Fisheries 48/6):
811-813,

STOKES, F.J. 1980, ‘Handguide to the coral reef fishes
of the Cambbean’, 160p. (Wm. Collins Sons:
Melbourne),

TAYLOR, F.J.R. 1979. A description of the benthic
dinoflagellate associated wilh matioloxin and
ciguatoxin including observations on Hawaiian
material, Pp.71-76, In Taylor, DL. & Seliger,
H.H. (cds) ‘Toxic dinoflagellate blooms’ (El-
sevier: Amsterdam).

VERNOURX, J.P, 1981, L’ichtyosarcotoxisme de type
Cteustera aux Antilles et en Polynésic Franguise:
tests de ciguatoxicité et chaine trophique
cituntérigéne. Ph.D. Thesis, 140 p,, Université de
Bordeaux.

VERNOURX, J.P., GAIGN, M., RIYECHE, N., TAG-
MOUTI, R., MAGRAS, L.Ph, & NOLEN, J,
1982, Demonstration of a liposoluble ciguateric
toxin in Caranx bartholomaer caught in the
French West Indies. Biochimie 64: 933-939,

VERNOUX, J.P., LAHLOU, N.. ABBAD EL AN-
DALOUSSI, S., RIYECHE, N. & MAGRAS,
L.Ph. 1985a. A study of the distribution uf
cigualoxin in individual caribbean fish Acta
Tropica 42; 225-233.

MEMOIRS OF THE QUEENSLAND MUSEUM

VERNOUX, J.P., LAHLOU, N., MAGRAS, L.Ph, &
GREAUX, J.B. 1985b. Chick feeding test: a
simple system to detect ciguatoxin, Acta Tropica
42: 235-240,

VERNOUX, |.P. & ABBAD EL ANDALOUSSI, S.
1986, Heterogeneity of ciguatoxins extracted
from fish caught at the coast ofthe French Antilles.
Biochimie 68: 287-291,

VERNOUX, J.P, & LAHLOU, N. 1986. Contréle
biologique de la ciguatoxine chez le poussin: ana-
lyse des sympt@mes induits et de Ja towicité
d'extraits de poissons cignatoxiques de Vile de
Saint-Barthélémy. Bulletin de la Société de
Pathologie exotique 79: 140-147.

VERNOUX, J.P., MAGRAS, M. & MAGRAS, Ph.
1988. ‘Coral fishes of the West Indies’. 128p.
(Lalanier: Pans).

VERNOUX, J.P. & TALHA, F, 1989. Fractionation and
purification of some muscular and visceral
Cigualoxins extracted from canbbean fish. Com-
parative biochemistry and Physiology 94B(3):
499-504,

YASUMOTO, T., BAGNIS, R., VERNOUX, J.P. &
CHUNGUE, E, 1975_ Toxins in the viseera and
ingested materials of herbivorous fish and mel-
Tuses linked to coral reets. Annual report for the
South Pacific Commission 1975; 407,

YASUMOTO, T., NAKAJIMA, L., CHUNGUE, E, &
BAGNIS, R. 1977a, Toxins in the gut contents of
a parrotfish. Bulletin of the Japanese Socicty of
Scientific Fisheries 43: 69-74,

YASUMOTO, T., BAGNIS, R,, THEVENIN, 8, &
CARGON, M. 1977b. A survey of comparative
loxicily in the food chain of ciguatera. Bulletin of
the Japanese Society of Scientific Fisheries 43;
1015-1019.

YASUMOTO, T., INOUE, A. & BAGNIS, R. 1979.
Ecological survey of a toxic dinoflagellate as-
sociated with ciguatera. Pp.221-224. In Taylor,
DL, & Seliger, H.H, (eds), ‘Toxic dinoflagellate
blooms”. (Elsevier: Amsterdam).

YASMOTO, T., NAKAJIMA, I, OSHIMA, Y. &
BAGNIS, R. 1979b, A new toxic dinoflagellate
found in association with cigualera. Pp, 65-70. In
Taylor, D.L. & Seliger, H.H. (eds), “Toxic
dinoflagellate blooms’. (Elsevier: Amsterdam).

YASUMOTO, T. & ENDO, M. 1973. Toxicity study
on a marine snail Turbo argyrostoma.{, Presence
of two sulfur containing amines in the acetone
soluble traction. Bulletin of Japanese Society of
Sejentific Fisheries 39: 1055-1061,

YASUMOTO, T., BAGNIS, R., VERNOUX, J.P.
1976, Toxicity of the surgeonfishes {l: Properties
of the principal water soluble toxin. Bulletin of the
senass Society of Scientific Fisheries 43: 359-
365,

LIST OF PARTICIPANTS

LIST OF PARTICIPANTS IN THE WORKSHOP

Dr tng BABINCHAK, National Marine Fisheries Service, PO Box 12607, Charleston, SOUTH CAROLINA, USA

29422-2607

Dr Raymond BAGNIS, Université Frangaise du Pacifique, BP 4635, Papeete, Tahiti, POLYNESIE FRANCAISE

Mr Paul BIRD, 4 Spicely Crescent, HEATLEY, QLD 4814

Dr Donna Glad BLYTHE, Rosenstie! Medical Marine Biology, 4950 Le Jeune Road, Suite A, Coral Gables,

FLORIDA, USA 33146

Mr Alan BREMNER, Intemational Food Institute of Queensland, 19 Hercules Street, HAMILTON, QLD 4007

Dr James BROCK, Medical Faculty, UNIVERSITY OF NEWCASTLE, NSW 2308

Mr John BURKE, Southem Fishenes Centre, PO Box 76, DECEPTION BAY, QLD 4508

Dr Bruce CAMPBELL, Sullivan and Nicolaides, 134 Whitmore Street, TARINGA, OLD 4068

Mr iy CAMPBELL, Queensland Commercial Fishermen's Orgamsation, 17 Hume Parade, PARADISE POINT,

QLD 4216

Dr Mike CAPRA, Centre for Biological Population Management, Queensland University of Technology. BRIS-

BANE, QLD 4000

Mr Milani CHALOUPKA, Department of Environment & Heritage, 160 Ann Street, BRISBANE, QLD 4001

Mr Richard CORMICK, Department of Chemistry, Monash University, CLAYTON, VIC 3168

Mr Paul DALZELL, South Pacific Commission, BD DS, NOUMEA, NEW CALEDONIA

Dr David DAVIES, Sth Pacific Underwater Medicine Society, Suite 6, Killowen House, St Anne's Hosprial, MT

LAWLEY, WA 6050

Mr Richard DEWIS, School of Life Science (Biochemistry), Qld University of Technology, George Street,

BRISBANE, OLD 4000

Dr Robert DICKEY, Fishery Research Branch. FDA, PO Box 158, DAUPHIN ISLAND, ALABAMA, USA 36528

Mr Mike DREDGE, Southern Fisheries Cenire, PO Box 76, DECEPTION BAY, OLD 4508

a Andrew FLOWERS. Centre for Biological Population Management, Qld University of Technology, BRISBANE,
LD 4000

Tr Noel GILLESPIE, Bribie Island Aquaculture Centre, DPI, PO Box 191, Bellara, BRIBIE ISLAND, QLD 4507

Mr Daniel GRZEBYK, Centre d'Océanologie de Marseille, (CNRS - URA 41), 13007 MARSEILLE, FRANCE

Ms Sharon GUY, Department of Chemistry, Monash University, CLAYTON, VIC 3168

Dr Scott HAHN, School of Life Science (Biochemistry), Qld University of Technology, George Street. BRISBANE,

QLD 4000

Dr Gustaaf HALLEGRAEFF, Department of Botany. University of Tasmania, GPO Box 252C, HOBART, TAS

7001

Mr Paul HAMBLIN, C/- Department of Physiology & Pharmacology, UNIVERSITY OF QLD, QLD 4072

aga Yoshitsugi HOKAMA, Dept of Pathology. University of Hawaji, 1960 East-West Rd, HONOLULU, HL.

USA 96822

Dr Mike HOLMES, Southem Fisheries Centre, PO Box 76, DECEPTION BAY, QLD 4508

Mr Dana ICHINOTSUBO, Department of Pathology, University of Hawaii, 1960 Easi-Wesi Rd, HONOLULU. HL

USA 96822

Mr Phillip JOBLING, Department of Physiology & Pharmacology, University of Queensland, ST LUCIA, QLD

4067

Dr Ursula KALY, AIMS, PMB No 3, TOWNSVILLE QLD 4810,

Dr Geoff KING, Medical Superintendent, Mossman Hospitals Board, PO Box 332, MOSSMAN QLD 4873

Dr Richard LANG, Department of Physiology, Monash University, CLAYTON, VIC 3168

Mr Dan LEE, Department of Pathology, University of Hawaii, 1960 East-West Rd, HONOLULU, HI, USA 96822

Dr Anne-Marie LEGRAND, Institut Louis Malardé, PO Box 30, PAPEETE, TAHITI, FRENCH POLYNESIA

Dr Richard LEW1S, Southem Fisheries Centre, PO Box 76, DECEPTION BAY, QLD 4508

Dr Ronald MANGER, US Food and Drug Administration, 22201 23rd Dr, 8.E,, PO Box 3012, BOTHELL, WA,

USA 98041-3012

Professor Elspeth MCLACHLAN, Department of Physiology & Pharmacology, University of Queensland, ST

LUCIA, QLD 4067

Dr Dietrich MEBS, Zentrum der Rechtsmedizin, University of Frankfurt, Kennedyallee 104, D-6000 Frankfurt/Main

70, RFA

Dr Jordi MOLGO, CNRS, Laboratoire de Neurobilogic, Cellulaire et Moléculaire,,91190 Gif-sur Yvelie, FRANCK

Dr Neal PALAFOX, 5711 Oakshire Road, Baltimore, MARYLAND, USA 21209

Dr Douglas PARK, University of Arnizon, Department of Nutrition & Food Science, 309 Shantz, TUCSON,

ARIZONA, USA 85718

639

MEMOIRS OF THE QUEENSLAND MUSEUM

LIST OF PARTICIPANTS (continued)

Mr John PAYNE, Solicitor, PO Box 286, TOOWONG, QLD 4066

Professor John PEARN, Department of Child Health, University of Queensland, Royal Children’s
Hospital, BRISBANE, QLD 4029

Dr Patrick PERLMUTTER, Department of Chemistry, Monash University, CLAYTON, VIC 3168
Dr Barry POLLOCK, Fisheries Services, Department of Primary Industries, GPO Box 46, BRIS-
BANE, QLD 4001

Ms Christine PURCELL, Centre for Biological Population Management, Queensland University of
Technology, BRISBANE, QLD 4000

Dr Tilman RUFF, Social & Preventive Medicine, Monash Medical School, Alfred Hospital, Com-
mercial Road, PRAHRAN, VIC 3181

Professor Paul SCHEUER, University of Hawaii, Department of Chemistry, 2545 The Mall, HONO-
LULU, HI, USA 96822

Ms Michelle SELLIN, Souther Fisheries Centre, PO Box 76, DECEPTION BAY, QLD 4508

Dr John SHERIDAN, Health and Welfare Building, 63-79 George Street, BRISBANE, QLD 4000
Mrs Raewyn STREET, Southern Fisheries Centre, PO Box 76, DECEPTION BAY, QLD 4508