INFLUENCE OF COLD
ON HOST- PARASITE INTERACTIONS

PART III

o=

_jj Editor

f^ ELEANOR G. VIERECK

QP
82
A 75

ARCTIC AEROMEDICAL LABORATORY
FORT WAINWRIGHT
ALASKA

1963

COLD AND COLDS

Sir Christopher Andrewes

Overchalke Coombe Bissett
Salisbury, England

ABSTHACT

Two problems arise: (1) Does cold really precipitate colds in individuals? (2) Why
are colds commoner in summer? Are these two problems or only one problem? (1)
The popular belief that cold or damp can precipitate a cold is very strong. Experi-
ments at the Salisbury Common Cold Unit failed to confirm the notion. Can the belief
be due to popular confusionbetween cause (fall in temperature) and the earliest symptom
(a feeling of chilliness); or does this effect of cold only act in certain infrequent in-
dividuals or at certain infrequent moments, when virus in a latent state can be acti-
vated? Chilling does not seem to activate colds in small isolated communities in the
Arctic or elsewhere. (2) Colds are certainly commoner in temperate zones in winter
than in summer. Many observers report three waves - one in September- October,
one in January, and often a third, lesser one in March. Very many things can be cor-
related with change in season; but which are the important ones? There is little evi-
dence that temperature is itself the direct cause. Rather better, but still unconvincing,
evidence might implicate humidity, perhaps because cold viruses could survive better
in the air at low relative humidities. A "winter factor" could operate in such a way,
favouring the virus' survival, or indirectly through change in people's habits, again
permitting more cross- infections to occur. There may be "conditioned epidemics",
a vinis such as Influenza being seeded into a population but not manifesting itself un-
til conditions, perhaps meteorological ones, are favourable. Possibly the physiology
of the respiratory mucosa is all important. A homoeostatic mechanism may not re-
spond promptly enough to environmental changes and virus attack may be favoured
at this "unguarded hour". Recent work at Salisbury and elsewhere, permitting identi-
fication and titration of "Rhinovi ruses" offers hope of a direct attack on some of
these questions. Understanding of the natural history of colds may offer better hope
of controlling such infections than a programme directed towards specific viruses.

I have not come all the way to Alaska to tell you what the relation
between cold and colds is; rather, I have come in hope that in the
discussion following my remarks some light may be shed on the
obscurity which veils this subject. Why are colds called "colds"?
Is it because the subject feels cold, or is it because chilling is
thought to precipitate an attack? Or perhaps because colds come at
the cold season of the year ? We have really to consider two prob-

301

ANDREWES

lems which may or may not turn out to be one and the same prob-
lem: 1) Does chilling really set off a cold infection and if so how?
2) Why are colds commoner in winter than in summer? In other
words, is the seasonal prevalence of colds simply due to a sum-
mation of the effect of chilling on individuals ? At the moment there
is good scientific evidence that colds are in fact commoner in winter.
Evidence as to the effect on the individual is more doubtful.

Cold and Colds in Individuals

Many people are quite confident that chilling will bring on a cold
in them, and they nearly all say that this can happen within an hour
or two. A friend of mine has to take time off when she wants her
hair washed, as it must be dried immediately and properly, other-
wise she "invariably gets a cold". We have offered to test this ex-
perimentally but have met with no co-operation.

Experiments were conducted at the Common Cold Research Unit
at Salisbury to test the effects of chilling. In this unit, volunteers
come for ten days at a time, having offered their services as "human
guinea-pigs". They are kept isolated, usually in pairs, and appropri-
ate precautions are taken so that there is no subjective bias in de-
ciding whether or not our experimental procedures have given them
a cold. In an experiment several years ago, we took three groups of
six volunteers (Andrewes, 1950). The three pairs in one group were
given diluted virus, which was expected to produce only a few colds.
Another six were soaked in hot baths and then made to stand, un-
dried, in their bathing dresses in a draughty corridor for 30 min-
utes. After that, they wore wet socks for some hours. A third group
of six received the dilute virus plus the chilling treatment. Chilling
alone produced no colds. The "Virus alone" group got two colds,
while the group with virus plus chilling developed four colds. Inter-
esting, but not statistically significant. We repeated the experiment
and again chilling albne did nothir^. But in the other two groups the
result was reversed; there were fewer colds with chilling than
without it. In a third experiment our volunteers went for a walk in
the rain; on returning rather cold and tired, they found that we had
turned off the heat in their quarters. They were treated in three
groups as before, and, as in the second test, chilling was not seen to

302

COLD AND COLDS

have any effect in predisposing to colds. Dowling et al. (1958) re-
ported similar findings.

There is quite an extensive literature, reviewed byThomson and
Thomson (1932), on the effect on the nasal mucosa of chilling
applied locally to the nose, to the whole body, or to the feet. The ex-
periments have mostly been carried out in hope of determining why
chilling causes colds. No evidence is adduced on the preliminary
question of whether chilling causes colds. Mudd and Grant (1919),
amongst others, found that cooling the body surface causes blanching
of the upper respiratory mucosa, and a fall in temperature there of
rather less than 1° C. Cooling of the feet alone was rather less
effective. Different observers are not in agreement as to whether
local draughts playing upon the nose are more or less effective than
cooling of the body generally. Schmidt and Kairies (1931) confirmed
other observers that chilling caused mucosal ischaemia. This hap-
pened to everyone they tested, but the rate of return to normal was
very variable, a fact having possible bearing on varying suscepti-
bility to colds. There is one report that chilling causes increased
acidity in the saliva with a suggestion that nasal secretion may be
similarly affected; this, if true, could perhaps be related to the
rather acid conditions which common cold viruses seem to like in
tissue- culture. Draughts could, of course, operate by causing local
desiccation and hence temporary stagnation in the sheet of mucus
which is normally flowing continuously backwards over the mucous
membranes... But I am in danger of becoming fascinated by these
reports of thirty years ago and am being entrapped into discussing
"how" and avoiding the question of "whether".

1 offer for discussion three possible explanations of the conflict
between popular belief in this matter and our own experimental
findings.

The first is that the popular belief is a fairy-tale, having no real
basis in fact. The second is that people confuse early symptoms
with cause. Assuming, as I do, that a common cold is essentially a
virus infection, it is hard to explain, on any hypothesis, how this
could be "full-blown" within an hour or two of the chilling episode,
as is usually reported. Is it not more likely that an early symptom
of the virus infection is an undue sensitivity to the effects of the

303

ANDREWES

physiological adjustments which follow a fall in temperature? Most
people not infrequently get wet feet or sit in draughts without feeling
chilly and without a subsequent cold; these incidents they forget,

I must digress before putting forward my third explanation. There
is evidence as regards several respiratory virus infections, and
more particularly influenza, that virus may be widely seeded into a
population and yet not give rise to an immediate epidemic. Something
has to be right before the outbreak can get started. Particularly re-
markable was the way in which the A2, commonly called Asian in-
fluenza, spread rapidly in tropical countries, yet was seeded fairly
freely into North America and Europe some months before anything
very much happened there. With onset of cooler weather, the epi-
demic broke out. Common cold viruses have occasionallybeen iso-
lated from normal noses and it is not unlikely that in some people
they, like some other viruses, may be in a state of unstable equi-
librium with their host, awaiting activation by an appropriate stim-
ulus. We may not have induced in any of our comparatively few
volunteers the right kind of equilibrium with the virus we gave them;
it may be only a small percentage of the population in which chilling
would upset a balance and unleash a cold.

One thing seems fairly certain. In small isolated communities in
the Arctic, Antarctic, or elsewhere, cold viruses are lost or else
the members of the small community soon become immune to the
viruses circulating amongst them (Paul and Freese, 1933), Chilling
does not induce colds in them unless there is some contact with the
outside world which could possibly introduce fresh viruses. Many of
you will be able to tell me if I am misinformed in this matter.

Season and Outbreaks of Colds

Let us now turn to cold and its effects on the incidence of colds
in large populations. Every chart I have seen tells the same story:
colds are much more frequent in winter than in summer. One obser-
ver (Lederer, 1928) asserts that summer colds are more often
sporadic and not associated with other cases in the family. In the
Northern Hemisphere there is commonly a peak in the incidence of
colds with the onset of cooler weather in September and October, a

304

COLD AND COLDS

second peak early in the New Year, and, less regularly, a third
lower peak about March. The New Year peak corresponds to the
favourite period for Influenza A epidemics. There are very many
differences between summer and winter, not only in temperature
and humidity, but in resulting changes in our habits, our dress, and
our diet; and it is extremely difficult to pin-point any one of these as
responsible for the winter increase in colds. Undoubtedly, a fall in
temperature precedes an outbreak of colds, but low temperature as
such does not do so; it is the change that matters. Milam and Smillie
(1931) found that onthe tropical island of St. John,the daily variation-
over a range of 6.5° C- was the same as the difference between the
summer and winter maxima (6.5° C to 8° C) ; yet colds were virtu-
ally absent from late May to late October. They thought that with
smaller temperature changes, colds were both scarcer and milder
than in colder climates. Van Loghem in Holland (1928) and also ob-
servers in North America have recorded that outbreaks of colds
occur simultaneously over wide areas of thecountry.lt is very dif-
ficult to explain this on any theory of simple per son-to- person
spread. The outbreaks seem rather tobe precipitated by temperature
changes.

Several writers consider thatrelativehumidity is more important
than temperature. Hemmes et al. (1960) reported that the virus of
influenza, a winter infection, survived better in the air under con-
ditions of low than of high humidity, whereas the virus of poliomj^e-
litis, a summer disease, behaved in an opposite manner. Hope
Simpson (1958) has pointed out that with the onset of colder weather,
people light fires indoors, or turn on the central heating, thus
causing a considerable drop in relativehumidity. At just such times
he sees in his practice a sudden increase in upper respiratory in-
fections. He does not venture to suggest whether this could be due
to better survival of viruses in the air, as Hemmes' results would
indicate, or to some effect onthe host's resistance. The effect on
virus survival seems unlikely, for rhinoviruses or common cold
viruses are closely related in their fundamental properties to the
enteroviruses, which include poliovirus; and their stability at various
relative humidities resembles that of the poliomyelitis rather than
of the influenza virus. So on Hemmes' line of reasoning, colds
should be a summer disease. Further, the fall in relative humidity
which Hope-Simpson records as happening in the autumn, is only

30 1

ANDREWES

noteworthy in unoccupied rooms. In crowded rooms where cross-
infection might be expected to occur, relative humidity is never
very low.

What of temperature changes as causing changes in people's
habits ? Undoubtedly people tend to spend their spare time together
indoors in the winter and much more out of doors in summer. This
would tend to encourage spread of infection in winter - or so one
might at first sight suppose. But if one considers workers in offices,
factories, and shops (a considerable part of the population) , their
lives differ in summer and winter for only a small part of the day.
Many of them use crowded public transport all through the year and
they are perennially cheek-by-jowl in their shops, factories or of-
fices. Probably there is better ventilation in these surroundings in
the summer tims,butnothingachievedby students of air hygiene has
yet encouraged us to believe that respiratory infections are likely
to be greatly reduced by such means; nor yet by U-V irradiation,
chemical aerosols or other methods designed to give an equivalent
result. The fact is that as regards colds, the experimental evidence
available suggests that cross- infection takes place mainly through
direct hits with infectious particles at close range rather than
through minute droplet nuclei (Lovelock et al., 1952). So ventilation
could hardly be expected to play a major role.

There are, of course, some circumstances in which particular
kinds of habit changes leadingto close aggregations certainly favour
spread of virus infections. Quite a number of viruses, som2 types of
adenoviruses, Coxsackie A 21 (or Coe virus), and Influenza B, all
cause outbreaks of respiratory infections mainly in recently col-
lected service recruits or in children re-assembling at boarding-
schools after holidays. Even here season plays a role, for the adeno-
virus outbreaks amongst recruits are not important during summer
months.

We seem to be frustrated at every turn. Every promising clue
seems to peter out. Are recent advances in knowledge about colds
and other viruses likely to be able to help us? I think they are. First
of all, Tyrrell and his colleagues at Salisbury (Tyrrell et al., 19 59)
have found out how to cultivate viruses from a high proportion of
common colds in adults. These we are calling Rhinoviruses (nose-

306

COLD AND COLDS

viruses). They can be grown in cultures of human embryonic kidney,
and in diploid cell lines from human embryonic lung; a minority of
the strains grow also in monkey kidney and other primate cells .The
trick is to cultivate at a lower temperature than is conventional
(33'-' C), at a lower pH, and in rolled tubes to give good oxygenation.
New developments make their study easier. They grow in a wider
range of cells ; they can be studied quantitatively by counting the
tiny foci of degeneration produced in cell sheets (Parsons and Tyr-
rell, 1961) ; even macroscopically visible plaques can now be pro-
duced (Porterfield, 1962). These Rhinoviruses resemble entero-
viruses in being very small, ether- resistant viruses. They differ in
their cultural requirements, greater lability towards acid, habitat,
and pathogenicity. They are of many different serological types, "and
although we don't yet know how many, 30 is probably a low estimate.
We are now passing the preliminary stage of establishing that these
viruses do, in fact, cause many colds all over the world. The tech-
niques developed can now be applied to studies of epidemiology.
Quantitative studies are particularly needed. For many years in our
work at Salisbury we could only detect virus by seeing whether or
not material under study would produce colds when dropped up the
noses of volunteers. Our subjects reacted very variously, and at
best we could only infect 50 per cent of them. Quantitative studies
were almost impossible. Now that one can count rhinovirus plaques,
things are very different. It should not be difficult to discover the
whereabouts of virus in the environment, and just when, how, and in
what quantity it is shed from an infected person. I should not be
wholly surprised to discover that virus shed by a cold-sufferer in
summer wasquantitatively less than in winter, so that there was less
danger of infecting others. All sorts of other quantitative studies
should be applicable from now on, including those concerned with
seasonal variation in resistance to colds.

Resistance to Colds

Jackson and Dowling (19 59) in Chicago produced evidence that
resistance to five strains of colds was specific, directed against
a particular strainofvirus.Discovery of the serological multiplicity
of cold viruses fits in with this. Antibodies to particular strains
seem to be well correlated with resistance to those strains. Yet other

307

ANDREWES

evidence suggests that there is a non-specific element in immunity.
It is generally agreed that in isolated communities immunity to colds
falls, so that contact with civilization after along interval is quickly
followed by colds, and often they are severe ones. Why? The usual
explanation is that people in a large community are constantly sub-
jected to little doses of virus which reinforce their immunity, often
without giving rise to any symptoms. Yet if there are 150 or 57 or
some other large number of cold viruses, it is unlikely that we are
all being regularly exposed to all of them. If not, why are we so
much more resistant than these lately- isolated people? And why
don't they get 57 colds when they beginto mix with society? It seems
that if there is a nonspecific as well as a specific immunity, the
facts couldbereasonablyexplained. This could be mediated by inter-
feron, a virus- inhibiting protein produced by cells under the stimu-
lus particularly of dead or damaged virus. Its production seems to
supply a quickly acting method of halting a virus infection until anti-
bodies canbe made and mobilized. Its action is local and nonspecific,
being directed not merely against the virus which evoked it. In tis-
sue-culture, rhinoviruses are amongst those most susceptible to its
action, and the Salisbury workers are at present engaged on studies
in volunteers of a possible role of interferon in cold virus infections.

If I had to offer a working hypothesis as to the effect of cold in
favouring colds, it would be along the following lines. The human
body, especially the respiratory tract, is exposed to large environ-
mental changes. The complicated anatomy of the nose is part of a
homeostatic mechanism designed to protect the lungs from sudden
changes. But it may not operate completely and instantaneously.
There may be a lag, and during an "unguarded hour", a virus reach-
ing or already inthe mucosa,in equilibrium with the host,may snatch
its chance. I doubt whether a local fall in temperature alone suf-
fices; it may well be an indirect consequence of temperature fall.
Some evidence suggests that interferon production is not so good at
lower temperatures, and that is the sort of thing about which I am
speaking.

I regard an attack on this difficult matter as the most important
phase of our war against respiratory infection. Injecting vaccines
against dozens of sero- types of virus seems to be a rather unprom-
ising business. Discovery of the relation between cold and colds

308

COLD AND COLDS

could lead to measures effective not only against rhinoviruses but
against all the other respiratory plagues as well. I feel that the bal-
ance may be tipped by something quite small but devilishly elusive.
Perhaps our discussions at this symposium will bring forth a clue.

LITERATXJRE CITED

1. Andrewes, C. H, 1950. Adventures amongviruses. III. The puzzle

of the common cold. New England J. Med. 242: 235.

2. Dowling, H. F,, G. G. Jackson, I. G. Spiesman, and T. Inouye.

1958. Transmission of the common cold to volunteers under
controlled conditions. IV. The effect of chilling of the subjects
upon susceptibility. Am, J. Hyg, 68: 59.

3. Hemmes, J. H., K. C. Winkler, andS.M. Kool. 1960. Virus sur-

vival as a seasonal factor in influenza and poliomyelitis.
Nature 188: 430.

4. Jackson, G, G,, and H. F. Dowling. 1959, Transmission of the

common cold to volunteers under controlled conditions. IV.
Specific immunity to the common cold. J. Clin. Invest. 38: 762.

5. Lederer, R. 1928. Die "Wintergipfel" der Atmungserkrankungen

Wien. Klin. Wchschr, 41: 257.

6. Lovelock, J. E., J. S. Porterfield, A.T.Roden, T. Sommerville,

and C. H. Andrewes. 1952. Further studies on the natural
transmission of the common cold. Lancet 2: 657.

7. Milam, D. F., and W. G.Smillie.l931. A bacteriological study of

"colds" on an isolated tropical island (St. John). J. Exp, Med.
53: 733.

8. Mudd, S., and S. B, Grant, 1919, Reactions to chilling of the body

surface. J. Med. Res. (Boston) 40: 53.

309

ANDREWES

9. Parsons, R., and D. A. J. Tyrrell. 1961. A plaque method for

assaying some viruses isolated from common colds. Nature
189: 640.

10. Paul, J. H., and H.L.Freese. 1933. An epidemiological and bac-

teriological study of the common cold in an isolated arctic
community (Spitzbergen). Am. J. Hyg. 13: 517.

11. Porterfield, J, S, 1962. Titration of some common cold viruses

(Rhinoviruses) and their antisera by a plaque method. Nature
194: 1044.

12. Schmidt. R., and A. Kairies. 1931. Experimentelle Studien zur

Genese den "Erkaltungs-Katarrhe". Dent. med. Wchnschr.
2: 1361.

13. Simpson, R. E. Hope, 1958. Common cold: fact and fance.

Brit. Med. J. 1: 214.

14. Thomson, D., and R. Thomson. 1932. The common cold. Ann.

Pickett-Thomson Research Lab. vol. VIII.

15. Tyrrell, D. A. J., M. L. Bynce, G. Hitchcock, H. G. Pereira,

C. H. Andrewes, and R. Parsons. 1960. Some virus isolations
from common colds. Lancet 1: 235.

16. Logham, J. J. van. 1928. An epidemiological contribution to the

knowledge of the respiratory diseases. J.Hyg. (Camb.) 28:33.

DISCUSSION

BLAIR: Sir Christopher, I was particularly interested in your
viewpoint with regard to the effect of chilling. I never had oc-
casion to do a study of this particular matter with regard to our
patients whom we have cooled down to hypothermic levels, and
in the last 100 patients I can recall for whom I have been re-

310

COLD AND COLDS

sponsible, there has not been one single instance of a cold, and
furthermore, no pneumonia, so I think that this would support
your idea that chilling is probably not of any importance in the
development of a cold.

CAMPBELL: Of course, it's been a long time since I have had
bacteriology; we used to think that bacteria may be involved
here. You didn't mention bacteria at all. I just wondered whether
in your studies you had found any correlation between bacterial
flora and virus infection?

ANDREWES: Well, 1 think everybody agrees that in the late
stages of the cold, secondary infection with bacteria comes and
causes the sinusitis and the yellow , muco-purelent stuff. This
is generally about in the early stages of the cold; it is very
difficult to find anything abnormal about the bacterial flora then,
but there is no doubt that in the later stages they do come in
and confuse the picture.

CAMPBELL: But no particular type, I mean.

ANDREWES: No, it varies, I think.

WALKER: Have you done any studies on immunity to the com-
mon cold? Jackson and Dowling, 1 believe, report a surprisingly
slow development of the immunity.

ANDREWES: Well, yes, we are studying that, and we agree
with them that the immunity seems to be specific. We have one
of the people working at Salisbury, Dr. Periera, who is the origin
of the HGP strain of virus, and he has been bled at intervals.
His serum, before he had a certain cold in 1957, had no anti-
bodies to his strain. Then he had this attack of cold and his
antibodies came up pretty quickly. Since then, he has been fol-
lowed along and he has had several colds and we have gotten
different viruses from him. The original HGP virus is one of
the ones that grew in monkey tissue. One of the other strains
was one that didn't grow in monkey, but would grow in human
tissues and another was a strain which would transmit colds to
volunteers, but we couldn't grow it at all. All this time he has

311

ANDREWES

had only slightly falling antibody to his original strain and this
has been quite unaffected by his attacks of these other colds.

WALKER: And his antibody came up in a matter of a couple
of weeks?

ANDREWES: I can't recall, but fairly soon.

WALKER: As I recall, the reports of Jackson and Bowling
show that this was much slower in their experience.

ANDREWES: I don't remember their doing it with antibodies.

WALKER: It seems to me they got a peak only after six months
or a very prolonged period of time, which is strange,

SCHMIDT: I wonder if you might have a comment on the work
that I think has been done by Kruger and others concerning
the concentration of the negative or positive charged ions in
the atmosphere and its effect on the mucosal lining.

ANDREWES: 1 am afraid I don't know about that.

TUNEVALL: There is one thing that relates to that season
of the year, and that is when our children return from their
holiday to school. Couldn't the incidence of colds perhaps be
correlated to these instances when the children get together?

ANDREWES: Well, there is no doubt that children are very
much more effective spreaders of infection than adults. We car-
ried out some epidemiological studies in a rural valley near
Salisbury, and it was found that colds in adults in families in
which there were school-age children were two and a half times
as common as in families when there were no school- age children.
There is no doubt that the little darlings spread the virus very
efficiently.

1 Krueger et al. 1959. Proc. Soc. Exp. Biol. Med, 102: 355-357.

312

COLD AND COLDS

SULKIN: You mentioned that to limit the occurrence of the
common cold through classical immunization might present a
difficult problem because of the growing number of serologic
types. I wonder, in your opinion, if it is conceivable that a new
type of immunization procedure might be evolved through the
mechanism of interference. That is to say, since inactive viral
particles will produce interferon, is it conceivable ?hat by in-
troduction of say, influenza virus into the external nares, that
one might manufacture sufficient interferon to cope with any
one of these common cold viruses?

ANDREWES: We have actually tried that and it didn't come
off, but we don't despair of an approach on that line. Issacs
found that the production of interferon is not only mediated by
virus approach; he thinks it is fundamentally a reaction to any
foreign nucleo- protein approach, and it is possible to produce
some interferon with nucleo- protein approaches of non-viral
origin. One man made an interesting suggestion which we haven't
followed up; that is, that the low incidence of colds in the sum-
mer might be due to the fact that in the summer, people are
constantly being stimulated with nucleo- protein produced from
pollen.

NUNGESTER: I hate to agree with Dr. Blair and sell Joe Berry's
long underwear short here, and I wonder if you believe that ex-
periments with humans where there is more or less uniform
cooling are characteristically comparable to the sort of things
that you have done at Salisbury, In keeping with the work of
Mudgrant and Goldman,^ it isn't the uniform cooling of the whole
body that is important, but the irregular coolir^ of parts of the
body that is important.

WALKER: And I would add, to the right temperature.

NUNGESTER: Yes.

ANDREWES: Our feeling about this effect of cold is that the

2 Annals of Otology, Rhinology, and Laryngology. 30: 1. 1921.

313

ANDREWES

way we did the experiment it didn't show anything, but we are
not willing to say that there isn't a relation. We failed to achieve
it, and we hope other people will find a way they can do it; then
we will believe them. We are not biased about the whole thing.

NUNGESTER: A few years ago, we got to playing with this
sort of thing. In the process of taking nasal washings, we took
just a little bit of a history of the people from whom we took
the nasal washings, if they had a cold. These are students, of
course; they like to brag on how hard they have been working,
so you have to discount this a bit, but it seems that there was
almost a pretty good correlation between loss of sleep and the
incidence of a common cold. There seems to be more of a cor-
relation with this sort of fatigue and stress than there was with
exposure to cold. Have you had any experience with fatigue?
Of course, you have told us that you brought the fatigue ele-
ment in.

ANDREWES: Yes, also the opposite effect; if people go away
for a summer holiday and come back feeling absolutely on top
of the world, they're likely to get a cold almost at once,

SCHMIDT: In connection with humidity and its possible effect,
it occurred to us, during our studies here in Alaska, that it might
be involved. Humidity is very low during the winter months in
Central Alaska, to the extent that you waken with a very dry and
crusty throat. It is hard to imagine that one wouldn't be most
susceptible to any sort of respiratory illness under these con-
ditions. When Dr. Beard, of the Armed Forces Epidemiological
Board (at that time, at least) was here, I discussed this as-
pect with him and he indicated that he had considered this to be
a factor some years ago. He investigated this by working in a
desert area where it was warm but very dry, and he didn't ob-
serve a higher incidence of infection in his subjects. Although
our attempts were rather clumsy and humble, no relationship
could be established between the relative humidity and upper
respiratory infection. Even though you are extremely uncomfor-
table, you seem to survive very nicely.

ANDREWES: Of course, under those desert conditions is the

314

COLD AND COLDS

time when you get epidemics of cerebral spinal fever. I think
that seems to go in the Sudan and other regions around the Sahara,

CAMPBELL: I would like to ask a fundamental question. It
has nothing to do with cold exposure, but in immunization to
viruses that seem like polio, you have to have the specific strain
to produce immunity. We are so conditioned with the pneumo-
coccus work in which the polysaccharide plays a major role
in infection. How do you envision the structure of a virus that
is so specific? There must be something in common with all
these strains of cold viruses, and if we had enough, you pro-
bably could show immunologically or serologically that there
were cross reactions, like the C substance in pneumococcus.
Do you envision a capsule or something around the virus ?

ANDRE WES: Most of these viruses consist of nucleo- protein
core and the other protein outside, and in the case of the smaller
viruses, that is about all there is to it. We have found evidence
of a very slight amount of cross-reaction between some of these
colds. It is only very trivial, but there are small amounts, as
in the case of the pox group of viruses covering not only all the
animal poxes, but also myxoma and a number of others. It has
quite recently been shown by Japanese workers, and by Fenner
in Australia, that there is a common nucleo- protein antigen which
was overlooked for many years. I wouldn't at all think it im-
possible that a similar thing would be found with some of these
smaller viruses. Whether it would be of any use in inducing im-
munity would be anybody's guess.

WALKER: Most commonly, this nucleo- protein represents the
common antigen of the group. It is not a protective antibody,

CAMPBELL: Well, you could modify it some way. They must
have something in common. They like to live in the nose; diph-
theria likes to live a little further down, but both bacteria and
viruses, of course, are local; that's where you get the term
neurotropic and dermatropic.

ANDREWES: We need to collaborate with an immunochemist.

315

ANDREWES

CAMPBELL: Next year, I will be over.

NORTHEY: I'd like to go one step further. How do you feel
that antibody really acts in these infections ?

ANDREWES: Well, it seems to play more of a part than we
expected. With some of these other respiratory infections, you
seem to be able to get repeated infections in spite of the presence
of some antibody in the serum, but in the case of the common
cold viruses, it appears that if you have got an antibody of that
particular strain, you are likely to resist challenge by that par-
ticular strain. We were really surprised to find that there did
seem to be considerable relation between antibody and immunity.

MET CALF: Is there a difference between the antibody titer
that you find in the serum and the nasal secretions, and if so,
is this likely to be of any significance ?

ANDREWES: Well, we have looked for neutralizing things in
the nasal secretion, but that was some time ago before we really
got on to the way to grow these viruses, so I don't think we know
the answer to that question. In the case of influenza, of course,
you do get the same kind of antibody in the nasal secretions as
you do in the serum, but much less of it.

MET CALF: I was thinking of the obvious relationship of a
fluid bathing a vulnerable point of cells subject to attack.

ANDREWES: Of course, all the earlier work in that is muddled
up by the fact that people didn't appreciate the possible presence
of interferon.

PREVITE: Just one naive question, not being very familiar with
virology. We often hear that there are many different viruses
capable of causing what we call the "common" cold. Has any-
one made any effort to ascertain how great and variable are
the number of viruses in any given area?

ANDREWES: There have been a number of studies of that.
Dr. Harare in Chicago has published some work on that, and it

316

COLD AND COLDS

appears that it depends on what age group you take. If you take
small children, you will find .that the rare influenza viruses
produce quite a lot of infection. In any age group, the respira-
tory syncytial virus may be prevalent in one year, and in another
year it may be completely absent. Rhinoviruses seem to be the
hard core which cause more colds than anything else, but I should
make it clear that there are still a great many colds from which
we havn't been able to cultivate any viruses, although these
things will still produce colds in volunteers. So there is still
quite a lot to learn, but I think the proportion of the colds caused
by different agents probably varies very much from time to
time and from place to place.

Well, I don't know that anybody has actually solved all my
more difficult problems for me, but anyway, I am very grate-
ful for all the suggestions that have been made in the matter.

317

EFFECT OF ENVIRONMENTAL
TEMPERATURE ON VIRAL INFECTION^

Duard L. Walker, M. D.

University of Wisconsin

Medical School
Madison 6, Wisconsin

ABSTRACT

In considering the effects of cold on viral infections, four questions seem to be of
particular importance. These are: (1) Can exposure to cold cause an acute but mild
and inapparent infection to become an apparent and severe disease? (2) Can it seriously
worsen an apparent viral infection? (3) Can it activate a latent viral infection? (4) What
are the mechanisms by which cold exerts an effect on viral infections? Indication
that the answers to the first three questions can be "yes" is available from studies of
infections in animals. Studies in this laboratory on Coxsackie infections in mice are
pertinent to the first question. In infant mice the Conn.- 5 strain of Coxsackie B-1
virus causes a generalized, lethal infection, but in adult mice the infection is limited
to a mild, inapparent pancreatitis. Exposure of adult mice to a 4° C environment, how-
ever, results in illness with essentially 100 per cent mortality. Pertinent to the second
question are studies on the myxoma- fibroma viruses. Marshall has shown that exposure
to cold increases the severity of disease in rabbits infected with attenuated strains of
myxoma virus. Relative to the third question, Shope has found that exposure to cold
weather appears to activate latent infections of swine influenza virus in swine. Although
the mechanisms by which cold exerts an effect on viral intections have not been studied
extensively, there is growing evidence that it may be by simply lowering tissue tempera-
ture to one more favorable for multiplication of the infecting virus. At normal tempera-
tures Conn.- 5 Coxsackie virus multiplies only in the pancreas of the adult mouse. Ex-
posure to 4° C results in reduction of body temperature by 1.0° C to 1.5° C and in multi-
plication of virus in many organs. Exposure of mice to 36° C raises their body tempera-
ture 2° C and inhibits multiplication in all tissues, including the pancreas. Study of
Conn. -5 virus multiplication in in vitro cultures reveals that the virus does not multiply
well in adult mouse tissues at 37° C or at higher temperatures, but does multiply well
at 35° C.

1 Studies by the author on this subject were supported by grants from the National
Institute of Allergy and Infectious Diseases and the National Cancer Institute.

319

WALKER

Studies concerned with the effect of environmental temperature on
viral infections have sometimes been confusing and seemingly con-
tradictory, but cold has most often been found to aggravate viral in-
fections. For the purposes of this symposium, then, it seems to me
that there are four questions that should be considered: 1) Can ex-
posure to cold cause a mild, inapparent viral disease? 2) Can this
exposure seriously worsen an apparent viral infection? 3) Can it
activate a latent viral infection? 4) What are the mechanisms by
which cold exerts its effect on viral infections?

It is difficult to find direct experimental data bearing on these
questions that have been obtained in studies of infections in man, but
some evidence is available from studies in animals, and some in-
formation is available from study of viral infection of cells in
culture. The data are not extensive, but I think they provide indi-
cation that the answers to the first three questions can be yes, at
least with selected virus-host systems and under laboratory con-
ditions, perhaps even in man, and that we may find such effects if
we make the proper search.

I should like, then, to disclose this evidence. Some of it is from
others, but I shall limit consideration to that which I think has a
quite direct bearing on the questions that I have posed.

EXAMPLES OF THE EFFECT OF COLD ON VIRAL INFECTION

Can Exposure to Cold Cause a Mild, Inapparent Viral Infection to
Become an Apparent and Serious Disease?

Boring and I have studied a model viral infection in mice caused
by the Conn.- 5 strain of type Bl Coxsackie virus (Boring, ZuRhein,
and Walker, 19 56; Walker and Boring, 1958). Although this virus
produces a generalized and lethal infection in infant mice, it causes
in adult Swiss mice only a pancreatitis. The mice seldom show out-
ward signs of illness even after very large inocula. However, when

320

TEMPERATURE AND VIRAL INFECTION

Inoculum
(i. p. )

Environmental
temperature

Deaths**/no, inoc.

Virus
Suspension*

25° C
»+o C

0/20
19/20

Normal mouse
tissue suspension

25° C

4° C

0/20
0/20

Figure 1. Lethal effect of Coxs^ckie virus infection in adult mice at 4° C. *4000
LD-^ for infant mice. ••No. of mice dying during 10 days of observation.

Ambient temperature
C
Before After

inoculation* inoculation

Deaths /no.
inoculated

4° for 2

days

25°

0/12**

25°

1°, 1 day, then 25°

0/12

25°

4°, 2 days, then 25°

0/12

25°

«+o

12/12

25°

25°, 1 day, then 4°

12/12

25°

25°, 2 days, then 4°

9/12

25°

25°, 1 days, then 4°

9/12

25°

25°, 6 days, then 4°

0/12

Figure 2. Relationship of time of exposure to coUl ami lethal effect of Coxs;ickie
virus in adult mice. 'ISOO infant LIJ50 given 1. p. ••Deaths in 10 uay observation period
after mice placed at 4'^ C.

321

WALKER

inoculated mice are placed in a room at a temperature of 4 C, the
infection then becomes quite uniformly lethal. Figure 1 shows a
t)T)ical experiment. Acute, limited exposure to cold is not sufficient
to change this infection in adult mice from a restricted, asympto-
matic one into a fatal infection. Continued exposure through several
days is necessary (Fig. 2),

That the deaths of inoculated mice at 4° C are related to viral
infection can be shown by neutralizing the virus with specific anti-
serum prior to injection or by passively immunizing the mice prior
to inoculation. This prevents the deaths at 4° C. In addition, it can
be shown that this phenomenon is not caused simply by inability of
mice with pancreatitis to survive in the cold, but is due to a real
enhancement of the infection. Measurementofvirus levels and study
of tissue histology indicate that at ordinary temperatures viral
multiplication is limited to the pancreas, but that in mice at 4 C,
viral multiplication and tissue damage takes place in many tissues.
Data on this point will be presented in a later section.

An investigation of Briody and associates (Briody et al., 1953)
concerned with what could well be called inapparent infection is also
pertinent here. These workers examined the effect of cold on the
process of adaptation of influenza A' virus to multiplication in the
lungs of mice. Unadapted influenza virus usually multiplies to some
extent in mouse lungs, but it causes little pneumonia or mortality
until after a series of serial passages has resulted in a selection of
virus capable of rapid and abundant multiplication in the lungs of
mice. When inoculated mice were maintained at 5*^ C, however, the
virus grew to 100-fold higher levels, the extent of pneumonia was in-
creased and mice began to die of influenzal pneumonia after only a
few passages.

Can Exposure to ColdSeriously Worsen an Apparent Viral Infection?

This question, of course, is not very different from the previous
one, because in many instances the difference between inapparpit and
apparent infection is only one of degree. Nevertheless, ihe question
serves well to introduce work on myxoma virus infections.

322

TEMPERATURE AND VIRAL INFECTION

The effect of environmental temperature on viral infection has
appeared to be of some practical importance in Australia in the
evolution of myxomatosis in wild rabbits. Marshall (1959) noted that
there was repeated suggestion that myxomatosis spreading naturally
through wild rabbits was more lethal in winter than in summer, and
Mykytowycz (19 56) observed that rabbits experimentally infected with
an attenuated strain and housed in unheated quarters had a higher
mortality rate in winter than in summer. To test the possibility that
these observations were related to ambient temperature, Marshall
exposed inoculated rabbits to fluctuating cold (-1° C to +1° C for 16
hours and 15° C for 8 hours each day) and compared the results with
those at normal room temperatures (20° C to 22° C) and at elevated
temperatures (37°C to 39^0 for 16 hours and 26° C for 8 hours each
day). These fluctuating temperatures were chosen to simulate day
and night fluctuations of winter and summer. He found that ambient
temperature had little effect on infections with a highly virulent
strain of myxoma virus or with the quite virulent rabbit pox. But if
rabbits were inoculated with an attenuated strain of myxoma virus
that at 22° C caused death of about 60 per cent of rabbits, then ex-
posure to cold increased the mortality to over 90 per cent and ex-
posure to heat reduced mortality to about 30 per cent (Fig. 3), Myxo-
matosis in a rabbit is a verydistinctivedisease, and comparison of
the signs of disease in the rabbits at the various temperatures, as
well as differences in the levels of virus in the blood, were fairly
convincing indications that the differences in mortality rate were
due to alterations of the extent and severity of infection instead of to
some other effect of the temperatures.

Somewhat similarly, Sulkin (1945) has shown that after inoculation
wath influenza A virus, mice maintained at 15,5° C have significantly
more pulmonary consolidation and mice at 35° C less consolidation
than do mice held at 21° C to 25° C, And I judge from the abstracts
of Drs. Metcalf and Marcus that they will be providing additional
data pertinent to this question.

Can Exposure to Cold Activate a Latent Infection?

Latent here means an inapparent infection that exhibits chronicity
and some degree of host-virus equilibrium. Although this is a par-

323

WALKER

Ui

^

x:

>i

(T3

(X>

X)

<-{

^

u

0)

o

a

Mh

(0

o

u

A

o

<7»

<X>

CO

U

o

o

+-•

Wh

o

o

o

o

r»

l£>

CO

CM

d)

(-,

3

4-*
03

O

U

<U

O

a,

CM

s

CM

(U

1

+-<

o

+->

0

c

o

a)

CN

W

5-1

^

>,

fO

tD

Xl

nH

t^

;-•

0)

O

a,

4-1

w

U

u

x:

o

<-{

00

+

^

o

o

■p

Mh

CO

CM

w

w

■H

■!-•

•H

•H

X>

^

-%

u

■H U

to

13

r-\ ^

0)

D 0

^->

O U

u

O +->

0)

c c

Mh

•H O

c

c o

n

:d

b ^

324

TEMPERATURE AND VIRAL INFECTION

ticularly interesting question, there seems to be relatively little
data that concerns such infections and involves hosts and viruses
that lend themselves to detailed study. Dr. Andrewes has already
dealt with the common cold in man. There is one study, however,
that should be mentioned here,

Shope has pointed out that circumstantial evidence concerned with
epizootics of swine influenza has long indicated that the stimulus
responsible for precipitating attacks of the disease in swine latently
infected with the virus is in some way associated with sudden
changes in weather and especially with the onset of cold,wet weather.
He performed an experiment (Shope, 1955) in which he prepared 25
swine by feeding them earthworms containing lungworm larvae
carrying swine influenza virus. After about 30 days, during which
the animals remained well, he then exposed them for from 4 to 24
hours to adverse weather conditions, which always included rain or
snow and low temperatures. No data were obtained on the effect of
this on the body temperatures of the animals. Eight uninoculated
controls and 15 inoculated animals remained well, but 4 swine did
develop the typical illness of swine influenza and 6 others developed
serological evidence of infection.

POSSIBLE MECHANISMS FOR THE
EFFECT OF COLD ON VIRAL INFECTIONS

The studies that I have outlined indicate that with some hosts and
viruses under proper circumstances cold can aggravate and activate
viral infections. I have limited discussion here to those studies that
do point to these possibilities, because I think failures to find an
aggravating effect of cold, or even the occasional report of the op-
posite effect, are only to be expected with some host- virus combina-
tions and under certain circumstances. But the fact that there are a
good many reports of no effect or a protective effect of cold must be
kept in mind in any evaluation of possible mechanisms. I shall dis-
cuss this further in a later section.

325

WALKER

In any consideration of mechanisms that might account for the
effects of cold on viral infection two possibilities quickly come to
mind. One is that host defenses are modified by exposure to cold,
and the second is that cold acts as a stressing agent and modifies
the host through alteration of hormone balance.

The host defense most frequently considered and seems particu-
larly pertinent in viral infection is antibody production. The effect
of cold on antibody production has been discussed by others in this
meetir^, but I want to point out that Marshall (19 59) considered the
possibility of inhibition of antibody production in his study of rabbit
myxomatosis andshowedthatunder the conditions of his experiments
antibody against sheep erythrocytes developed as rapidly and to as
high levels in rabbits exposed to cold as in control rabbits.

In our study of Coxsackie virus infections in mice. Boring and I
have never actually measured the antibody developed against Cox-
sackie virus in mice at4°C, because we found that cold still exerted
its effect even if we delayed exposure of infected mice until the anti-
body producing process was well under way (Boring et al,, 1956),
Specific neutralizing antibody appears in the blood of mice on the
third day after inoculation with small quantities of Coxsackie Bl vi-
rus and on the fourth day the antibody is at substantial levels (Fig, 4;
Boring and Walker, unpublished data). Even though exposure to cold
is delayed until the fourth day, its effect is not nullified (Fig, 2),
This suggests to us that even if cold were found to have an in-
hibiting effect on antibody production, this would still not explain
the influence of cold on the infection. In addition to this, current
work using mice thymectomized at birth indicates that an adult
mouse does not develop a generalized lethal infection with Coxsackie
Bl virus even though the mouse is incapable of producing antibody.

Other host defenses such as non-specific viral inhibitors and
interferon may conceivably be altered by cold, but there is little
positive evidence for this, as yet.

The role of stress is difficult to assess. There is little question
that the cold exposure used inthe studies that I have discussed were
of a degree sufficient to cause some of the physiological changes
identified with stress. And treatment of animals with large doses of

326

TEMPERATURE AND VIRAL INFECTION

2048

-

^"^^

1024

-

f

....«r' >»

512

-

1

256

-

If
If

128

-

•f
If

If

If

64

.

1 1
1 /
* f
1 /

32

-

1 /
1 /
1 1

1 1

UJ 16
H
1- 8

-

1 1
1 /

f

0 I

1 1

/ /

CD

/ f

1 ^^

_

2 3

DAYS

Figure 4. Antibody response to Coxsackie virus infection in mice at 25" C ambient
temperature. Adult Swiss mice were given 280 infant mouse LD5Q 1. p. The antibody
titers from two experiments are plotted as separate curves. Antibody assays were
made in infant mice and antibody titers are expressed as the reciprocal of the serum
dilution that neutralized LD^q of virus.

cortisone causes an aggravation of a number of viral infections
similar to that seen with exposure to cold. Administration of 2.5 mg
of cortisone to adult mice infected with Coxsackie Bl virus results
in a generalized and lethal infection quite similar to that produced
by exposure to cold (Boring, Angevine, and Walker, 1955). But
Boring and I (1958) were not able to produce similar effects by
treating mice with physiologically active dosesof ACTH. In addition
to this, the fact that short exposures to cold were not effective, that
adaptation of mice to cold did not nullify the effect, and that ex-
posure to the stress of heat caused an entirely different response
has led us to think that just the stressing effect of cold cannot ac-
count for its effect on viral infection,

I have indicated some skepticism that inhibition of antibody pro-

327

WALKER

Day of Virus titers

Tissue infection 4° c 25° C 36° C

Blood

2

10-^-3

10-^-^

<

10-1-0

U

10-3.7

< 10-1-0

<

10-1.0

Brain

2

10-1-5

10-1.5

<

10-1.0

U

10-2.1

< 10-1.0

<

10-1.0

Pancreas

2

10-^-9

10-7.7

<

10-1-0

4

10-6.7

10-5.7

<

10-1.0

Heart

2

10-3.75

10-3.5

<

10-1.0

»t

10-'^-'^

< 10-1-0

<

10-1-0

Liver

2

10-4.5

10-'+. 0

<

10-1-0

!♦

10-5.0

< 10-1.0

<

10-1.0

Figure 5. Virus in tissues of adult mice infected with Coxsackie virus* and held
at 4° G, 25° C, or 36° C. •Inoculated with 140 infant mouse LD50 i. p.

duction or stress can account for the effect of cold on viral in-
fections. For discussion of another possible mechanism I want to
return to the model of Coxsackie virus infections in mice. In study-
ing the effect of cold on this infection, Boring and 1 (19 58) followed
the fate of virus in various tissues of infected mice. It was evident
that in mice at normal temperatures an initial viremia was followed
by significant viral multiplication only in the pancreas. But in mice
at 4° C, the virus multiplied to relatively high levels and produced
marked damage in several tissues. It was also found that if mice
were maintained at an elevated ambient temperature, viral multi-
plication was inhibited in all tissues, including the pancreas (Fig. 5).
In fact, even ifthe pancreatitis was allowed to progress for 36 hours
after inoculation, exposure to 36° C still brought about a prompt
drop in virus titer in the pancreas and rapid elimination of the
virus. An important point is that when rectal temperatures of mice
are measured during exposure to such temperatures, it can be

328

TEMPERATURE AND VIRAL INFECTION

Figure 6. Effect of environmental temperature on the body temperature of mice.
Eich. point represents the mean of the rectal temperatures of 10 mice. (Walker, D. L.,
and Boring, W. U. 1958. J. Immunol. 80: 39-44.)

demonstrated that at 4° C the mouse's internal temperature is
lowered about 1° C to 2° C and exposure to 36° C raises the rectal
temperature about 2° C to 3° C (Fig. 6).

These phenomena suggest that in the host cell there are tempera-
ture-sensitive reactions that can control Coxsackie virus multi-
plication. Additional support for this has been provided by Boring
and Levy (1962) who demonstrated that in HeLa cells in vitro the
optimum temperature for multiplication of the Conn.- 5 strain of
Coxsackie virus is 36° C and that multiplication is markedly in-
hibited at 38° C to 39° C, and is also reduced at temperatures below

329

WALKER

Temperature
incubation

of

Fibroma

Virus

Myxoma

290 C

5.5*

5.5

33° C

7.5

6.5

350 C

7.5

7.5

37° C

6.5

7.5

40° C

"4.5

7.5

Figure 7. Multiplication of myxoma and fibroma viruses in primary rabbit kidney
cells at various temperatures. *Log rabbit skin infectious units per ml produced in
48 hours.

36^ C. We have recently been looking at this more directly by com-
paring the multiplication of Bl Coxsackie virus in primary cultures
of adult and infant mouse tissues at various temperatures in order to
determine the optimum temperature for multiplication in the tissues
usually affected in the mouse. Our preliminary results suggest that
the virus can multiply in adult tissues if the temperature of incuba-
tion is reduced to 35° C and that it is inhibited at higher tempera-
tures .whereas in infant tissues it multiplies to high titer at 37° C to
38° C as well as at lower temperatures. Our experiments have not
proceeded far enough, however, to provide really reliable data.

I have already indicated that in his study of myxoma virus in-
fections in rabbits, Marshall (19 59) found a pattern similar to the one
I have described for Coxsackie virus; that is, an ameliorating effect
of high temperature as well as enhancement of the infection by cold.
Similar protective effects of elevated ambienttemperature had pre-
viously been reported by Thompson (1938). Both Thompson and Mar-
shall found that rabbits at testtemperatures had shifts of their rec-
tal temperatures of only about 0.5° C downward in the cold and
and 0.5° C to 1° C upward in the hot room, but their skin tempera-
tures changed 4° C to 5° C. The possibility that these temperature
changes may have been important in controlling the course of the in-
fection is supported by other evidence indicating that multiplication
of myxoma and fibroma viruses is easily affected by change of tem-
perature. The myxoma and fibroma viruses and their various vari-

330

TEMPERATURE AND VIRAL INFECTION

ants present a spectrum of viruses that are very closely related in
many physical and biological characteristics, including antigenic
makeup. However, after intradermal inoculation virulent myxoma
virus invades, causes generalizeddisease, and is almost 100 per cent
lethal for domestic rabbits, while fibroma virus causes only local
benign tumors that eventually regress without any apparent harm to
the rabbit. It is noteworthy that the only tissues in which fibroma
will multiply and cause lesions in the adult rabbit, even if injected
intraperitoneally or intravenously, are the surface tissues of skin
and testes. It can be demonstrated quite easily that myxoma and
fibroma viruses differ markedly in their capacity to multiply at
temperatures above 35° C. Thompson (1938) demonstrated this in
vivo when he raised the skin temperature of rabbits by exposure to
heat and showed that fibroma lesions were quite easily inhibited
while the disease caused by virulent myxoma virus required higher
temperatures to bring about amelioration. This can be shown in
primary cultures of rabbit tissues in vitro. Kilham (1959) has pro-
vided some data onthis,andsomeof our own data are shown in Fig-
ure 7. Fully virulent myxoma virus multiplies to high titer even at
temperatures of 40° C to 41 C, while fibroma virus reaches peak
titers at 32° C to 35° C and is inhibited to some extent at tempera-
tures as low as 36° C to 37° C and is severly inhibited at higher
temperatures. Variants of myxoma virus exist that are reduced in
their virulence and are intermediate between myxoma and fibroma
viruses in the characteristics of the disease that they produce in
rabbits. It was one of these that Marshall used in his study. We are
currently comparing the capacity of myxoma virus strains to multi-
ply at various temperatures and their relative invasiveness and viru-
lence in rabbits.This work has not progressed far, but it appears that
for many attenuated strains reduction in virulence is accompanied by
decreased capacity to multiply at temperatures comparable to the
internal temperature of the rabbit.

Other studies indicating a relationship between virulence and
capacity to multiply at temperatures above 37° C can be cited. Bed-
son and Dumbell (1961) have shown this relationship with several
poxviruses and their virulence for chicken embryos. Most detailed,
however, have been the extensive studies of Lwoff and associates on
poliovirus (Lwoff, 1959; Lwoff and Lwoff, 1960, 1961). Lwoff has
demonstrated in cells in culture that poliovirus is able to multiply

331

WALKER

Figure 8. Carrier cultures of mumps virus in human conjunctiva cells (C-M cultures)
at different temperatures. Cultures were grown at a?'^ C and then move., to test tempera-

332

TEMPERATURE AND VIRAL INFECTION

tures for 24 hours,
serum. A. 37° C.

Cells fixed and stained with fluorescein- conjugated anti-mumps
B. 40° C. C. 35° C. D. 33° C. Magnification -X 1000.

333

WALKER

only within certain limits of temperature. For many strains a tem-
perature of 40° C or above inhibits multiplication and viral pro-
duction is depressed below 33° C. It appears that elevated tempera-
ture produces a block in the second half of the virus multiplication
cycle and low temperatures block the first part of the cycle. But he
points out (also Dubes and Wenner, 1957) that strains vary in the
temperature that is optimum for their multiplication and that, in
general, neuro virulence in monkeys is associated with those strains
that are capable of good multiplication at 40° C, and lack of viru-
lence with those incapable of multiplication at 37° C or above.

There is additional evidence from quite a number of studies using
tissue cultures or chicken embryos (Enders and Pearson, 1941;
Thompson and Coates, 1942; Sharpe, 19 58; Hoggan and Roizman,
1959; Wheeler and Canby, 1959) that the temperature range within
which viruses can multiply vigorously is often quite limited, and
that the upper limit in particular may be quite sharp and abrupt.
For several of the viruses studied the optimum temperature for
multiplication is a degree or two below that of the internal body
temperature of some of the common mammalian hosts, I want to men-
tion briefly some data concerned with mumps virus, because it
could conceivably provide some insight into how cold could affect
latent infections,

Hinze and I have been studying a carrier system of mumps virus
in human conjunctiva cells. This system has been maintained for
several years (Walker and Hinze, in press). The cells multiply at a
rate comparable to control cultures, and show little evidence of
deleterious effect even though use of fluorescent antibody demon-
strates that 90 per cent or more of the cells contain viral antigen.
As routine, these cultures are grown at 37° C, at which tempera-
ture about 1 of every 100 cells appears to be excreting virus as
judged by the fact that erythrocytes will adsorb to the cell surface.
There is a low level of virus in the medium. At 37° C practically all
of the cells contain antigen, but it is restricted to a few sharply out-
lined, discrete, massesinthecytoplasm(Fig. 8a). At 40° C the anti-
gen is perhaps even more restricted in its distribution (Fig. 8b).
But at 35° C or 33° C the antigen becomes widely distributed in the
cytoplasm in small granules (Fig. 8cand8d) and the cells tend to be-
come rounded and ragged in appearance. At 35° C or 33° C erythro-

334

TEMPERATURE AND VIRAL INFECTION

C3^es will adsorb to 50 per cent or more of cells and the virus con-
centration in the medium increases by 10 to 100-fold indicating that
at the lower temperatures the equilibrium between cells and virus
is upset and there is much more virus production and release in the
cultures.

Lwoff (19 59) has emphasized the narrow zone of optimum tempera-
ture for viral multiplication and the inhibitory effect of elevated tem-
perature to argue that fever may be an important host defense
mechanism in choking off viral infection, and he has pointed to the
variation in optimum temperature among virus strains as a partial
explanation of variation in virulence. It seems to me that these phe-
nomena can also explain some of the observed effects of cold on viral
infection and, indeed, may even accountfor the variability of obser-
vations, I suggest that one can expect to see an effect of cold on viral
infection under certain circumstances. The appropriate circum-
stances would be when an animal (or man) is infected with a virus,
or a particular strain of a virus, that has an optimum multiplication
temperature a degree or two lower than the body temperature of the
host. Under these circumstances the body temperature of the host
could be an important controlling factor in limiting viral multiplica-
tion and in keeping the infection a mild, or inapparent, or latent one.
And lowering the temperature in the right tissues (surface or inter-
nal, depending on the virus) only a degree or two could result in a
markedly increased pace ofviral multiplication, and thus lead to ob-
vious aggravation or activationof the infection. Under these circum-
stances we would not expect to see an effect of cold unless the ex-
posure were sufficient to bring about an appropriate drop in temper-
ature in the right tissues. Nor need we expect to see an enhancing
effect on a highly virulent virus with a temperature optimum near
that of the host tissue.With certain host-virus combinations even an
ameliorating effect of cold might be expected if the tissue tempera-
ture could be dropped below and maintained below the optimum tem-
perature range of the virus. This concept of an effect of cold due to
direct influence onviral multiplication processes within the host cell
does not, of course,in itself exclude the possibility that stress, or
other physiologic changes due to cold, may contribute to the reaction.
But it seems to me that many of the demonstrations of an effect of
cold on viral infection can be accounted for on the basis of a direct
effect of tissue temperature on intracellular reactions.

335

WALKER
LITERATURE CITED

1. Bedson, H. S., and K. R. Dumbell, 1961. The effect of temperature

on the growth of pox viruses in the chick embryo. J, Hyg. 59 :
457-469.

2. Boring, W. D., D, M. Angevine, andD.L. Walker. 1955. Factors

influencing host-virus interactions, I. A comparison of viral
multiplication and histopathology in infant,adult, and cortisone-
treated adult mice infected with the Conn.- 5 strain of Cox-
sackie virus. J. Exp. Med. 102: 7 53-766.

3. Boring, W. D., and R. S. Levy. 1962. Studies on the production of

B-1 Coxsackie virus by HeLa cells. J. Immunol. 88: 394-400.

4. Boring, W. D., G. M. ZuRhein, and D. L. Walker. 1956. Factors

influencing host- virus interactions. II, Alteration of Coxsackie
virus infection in adult mice by cold. Proc. Soc. Exp. Biol.
Med. 93: 273-277.

5. Briody, B, A., W. A. Cassel, J. Lytle, and M. Fearing. 1953.

Adaptation of influenza virus to mice. I. Genetic and environ-
mental factors affecting an A- prime strain of influenza virus.
Yale J. Biol. Med. 25: 391-400.

6. Dubes, G. R., and H. A. Wenner.1957. Virulence of polioviruses

in relation to variant characteristics distinguishable on cells
in vitro. Virology 4: 27 5-296.

7. Enders, J. F., and H. E. Pearson. 1941, Resistance of chicks to

infection with influenza A virus. Proc. Soc. Exp. Biol. Med.
48: 143-146,

8. Hoggan, M. D., and B.Roizman. 19 59. The effect of the tempera-

ture of incubation on the formation and release of herpes sim-
plex virus in infected FL cells. Virology 8: 508-524. -

336

TEMPERATURE AND VIRAL INFECTION

9. Kilham, L. 19 59. Relation of thermoresistance to virulence among

fibroma and myxoma viruses. Virology 9: 486-487.

10. Lwoff, A. 1959. Factors influencing the evolution of viral diseases

at the cellular level and in the organism. Bacterid. Rev. 23;
109-124.

11. Lwoff, A,, and M.Lwoff. 1960.Sur lesfacteurs du developpement

viral et leur role dans I'evolution de 1 'infection. Ann. Inst,
Pasteur 98: 173-203.

12. Lwoff, A., and M. Lwoff, 1961. Les evenements cycliques du

cycle viral. I, Effets de la temperature. Ann, Inst. Pasteur
101: 469-504.

13. Marshall, I, D, 1959. The influence of ambient temperature on the

course of myxomatosis in rabbits. J, Hyg, 57: 484-497.

14. Mykytowycz, R. 1956. The effect of season and mode of trans-

mission on the severity of myxomatosis due to an attenuated
strain of the virus. Austral. J. Exp. Biol. Med. Sci. 34: 121-132.

15. Sharpe, H. S. 1958. Effect of temperature on the multiplication of

foot-and-mouth disease virus in suspensions of kidney cells of
the pig. Nature 182: 1803-1805.

16. Shope, R. E. 1955. The swine lungwor mas a reservoir and inter-

mediate host for swine influenza virus. V. Provocation of
influenza by exposure of prepared swine to adverse weather.
J. Exp. Med. 102: 567-572.

17. Sulkin, E. S. 1945. The effect of environmental temperature on

experimental influenza in mice. J. Immunol. 51: 291-300.

18. Thompson, R. L. 1938. The influence of temperature upon pro-

liferation of infectious fibroma and infectious myxoma viruses
in vivo. J. Infect. Dis. 62: 307-312.

337

WALKER

19. Thompson, R. L,, and M.S.Coates. 1942. The effect of tempera-

ture upon the growth and survival of myxoma, herpes, and vac-
cinia viruses in tissue culture. J. Infect. Dis. 71: 83-85.

20. Walker, D. L., and W.D. Boring. 1958. Factors influencing host-

virus interactions. III. Further studies on the alteration of
Coxsackie virus infection in adult mice by environmental
temperature, J. Immunol. 80: 39-44.

21. Wheeler, C. E., and C, M. Canby. 1959. Effect of temperature on

the growth curves of herpes simplex virus in tissue culture.
J. Immunol. 83: 392-396.

DISCUSSION

CAMPBELL: It seems to me that the enzyme system shown
in Figure 1 is an unusually sensitive system as far as kinetics
go. If you were to plot the effect of temperature on your enzyme
reaction, what would you find?

METCALF: We have done this repeatedly; I would say possibly
30 or 40 times.

CAMPBELL: I am not questioning the results. It's just so
unusual. Earlier this afternoon somebody talked about the sen-
sitivity of viruses to temperature as far as susceptibility goes,
so this is a rather unusual reaction. Didn't you notice that the
effect of temperature between 37*-* C and 20° C is really a tre-
mendous difference in the activity of your enzyme?

METCALF: In other words, Dr. Campbell, you think that this
is more than would seem to be indicated by experience with
other systems ?

CAMPBELL: Well, it's a little different, I would think. It
would be really interesting to do a kinetic study on it.

338

TEMPERATURE AND VIRAL INFECTION

METCALF: All I can say is, while we haven't performed pre-
cise kinetic studies, we have examined the reaction over a period
extending from zero hours through 24 hours. We find that there
is a linear release of neuraminic acid for twenty to thirty minutes,
after which the rate of release slows down greatly, but does con-
tinue for eighteen to twenty- four hours. The initial effect of
virus dilution upon the release of neuraminic acid is overcome
at the end of twenty- four hours. The effect of temperature seems
to consist of a retardation of the reaction with less neuraminic
acid split off.

CAMPBELL: That will be an interesting system to study thermo-
dynamically.

WALKER: I have been trying to relate this to some work that
Dr. Billie Padgett and I have done. Dr. Padgett approached the
question of the role of the enzyme of influenza a little differently,
and actually aimed at selecting a strain of influenza B virus
that was different in its enzymatic characteristics. She was
able to get a line of virus that is quite different in its enzymatic
characteristics from the parent strain in that its peak zone of
activity is at about 35 C, whereas the parent strain has its
peak activity at about 37° C. The enzymatic activity of this virus
is markedly inhibited at temperatures above 35° C, particularly
if calcium ion is removed from the medium. She has made an
effort to use temperature as a means of cutting the multipli-
cation cycle and seeing where the effect of this enzyme may
be. The parent and variant viruses appear to be very similar
in all other respects, antigenic make-up, general character-
istics, and so on, except for their temperature optimum for
enzymatic activity. She finds that the time at which elevation
of temperature will affect the multiplication cycle of the variant
virus is only toward the end of the cycle, suggesting again that
the enzyme is not particularly important in penetration. Ele-
vation of temperature with this variant virus early in the cycle
doesn't have much effect, but late in the cycle it does.

We also have followed the appearance of enzyme and virus in
the cell, but we have interpreted the results a little differently.
I am a little reluctant to think that the enzyme necessarily has an

339

WALKER

effect upon the cell. Our interpretation has been that as multi-
plication goes on and virus is produced possessing enzymatic
activity, you see then the appearance of enzymatic activity in
the cell. This is associated with cell damage, but it would be
difficult to decide which is cause and which is effect.

METCALF: Well, this is perhaps true, but in our case, every
time that we demonstrate enzymic competence and show this by
enzyme fabrication, we experience a corresponding reduction in the
exhibitor content and there is demonstrable damage to the cells.
I agree that it is difficult to separate cause and effect, but en-
zyme fabrication and cell damage are intimately related. Per-
haps this would be a compromise between the two viewpoints.

WALKER: Schlesinger, too, has shown cyclic waves of virus
level and substrate levels in the cells.

ANDRE WES: Dr. Sulkin, you were quoted by Dr. Metcalf.
Have you got any comments to make?

SULKIN: No, not particularly, because when those experiments
were done back in 1941 and 1942, we knew nothing about neura-
minidase, inhibitors, and so forth. It was sort of a naive ex-
periment.

ANDREWES: We were all naive in those days.

SULKIN: The results have been subsequently duplicated by other
workers,

CAMPBELL: Frank Lanni has done quite a bit on this inhibitor.
Didn't he come up with a chemical nature of it and do some kinetic
studies ?

METCALF: Lanni studied the interaction between enzymatically
active swine influenza and egg white inhibitory mucoprotein.
He described the kinetics of inhibitor inactivation, basing his

1 Larmi, F., and Y. T. Lanni. 1955. Virology 1: 40-57.

340

TEMPERATURE AND VIRAL INFECTION

interpretation of the data upon Gottschalk's model of the structure
of urinary mucoprotein. '"^ The inhibitor reduction method of
assay was used in these studies. There was no attempt to define
the chemical structure of inhibitor.

2 Gottschalk, A. 1952. Nature 170: 662-663.

3 Gottschalk, A. 1954. The Blakiston Co., New York,

341

THE INFLUENCE OF COLD ON VIRUS INFECTIVITY
Dr. T. G. Metcalf

Department of Bacteriology

University of New Hampshire

Durham, New Hampshire

ABSTRACT

This study was on the characteristics of influenza A2 strains which might play a
role in facilitating virus invasion of host cells, and the effect of cold upon the invasive
process. Strains of influenza A2 selected for study were isolated from fatal cases of
influenza in humans. The strains were examined for mouse toxicity following intra-
venous injection, cytotoxicity in HeLa and L- cells, mouse and chick embryo IDg^ values,
and neuraminidase activity. Virus enzyme action was singled out for special attention
on the basis of its constant association with A2 strains. The enzyme activity was de-
termined (1) by means of thiobarbituric acid analysis for free neuraminic acid, and
(2) by reduction of the hemagglutination titer of mucoid inhibitor. Substrates used
included neuraminmucoid from edible birds nest; neuraminlactose from bovine colostrum,
and ovomucin from hens eggs. The rate and extent of enzyme action exhibited by virus
showed a progressive decline as the temperature was lowered from 37° C to 4° C.
A comparison of enzyme activity, virus titer, and inhibitor concentration in chick em-
bryos at 37° C and 20° C showed corresponding decreases in enzyme activity and virus
titer while inhibitor remained virtually unchanged at lower temperatures. Infection
of mice was followed by extensive lung damage and death at 4° C. Considerably less
damage and fewer fatalities within the test period were found at 20° C. Contrary to
the chick embryo experience, the enzyme activity and virus titer showed increases at
the lower temperature. Cell monolayers of monkey or hamster kidney were used in
conjunction with fluorochrome analysis to follow the course of cell invasion by virus
at 37° C and 20° C. Virus was first demonstrated in the cytoplsism around the nuclear
membrane at 6 hours following incubation at 37° C. Virus presence in the cytoplasm
was shown for at least 72 hours. The course of invasion at 20° C was different. Virus
was first detected after 10 hours. It appeared in the cytoplasm, but failed to show a
significant increase in numbers. It was possible to show virus accumulating at the per-
iphery of allantoic membranes within 12 hours when embryos were incubated at 37° C.
No virus accumulation was observed in membranes from embryos incubated at 20° C.
Single-caged mice pre-exposed and maintained at 4° C showed massive virus invasion
of lung and bronchi, while group-caged mice showed only minimal virus invasion.

The effect of cold upon the course of influenza virus infections
represents a special problem in host-parasite relationships. Is the
primary influence of temperature a factor altering the biology of a
host, or is it directed against the virus? Which is of greater impor-

343

MET CALF
tance, physiological patterns of the host, or virus infectivity?

This study was concerned with the initial stage of virus-cell
interaction using influenza virus with selected hosts as the experi-
mental model. The influence of low temperature upon the interaction
has been examined from the standpoint of its effect upon virus in-
fectivity.

The proposal that neuraminidase facilitates the penetration of a
host cell by influenza virus presumes that infectious virus should
possess enzyme activity. This viewpoint has been presented by
Gottschalk (1957) who believes enzyme action renders cell wall
mucoproteins more permeable to virus and thereby facilitates
virus penetration of a host cell.

This study began with the knowledge that experimental influenza
infections have been represented as occurring inthe absence of en-
zyme action (Fazekas, 1948; Fazekas and Graham, 1949), Without
adopting a position on the essentiality of enzyme to the infectious
process, it was reasoned that clues to the influence of cold upon vi-
rus infectivity might be gained by a consideration of the effect of
cold upon neuraminidase production by influenza virus.

THE EFFECT OF COLD UPON THE NEURAMINIDASE ACTIVITY
OF INFLUENZA VIRUS

Enzyme action was measured by two methods. The thiobarbituric
acid (TEA) method of Warren (19 59) was used to determine the free
neuraminic acid resulting from combination of virus and a neura-
minic acid containing substrate. A neuraminmucoid substrate was
prepared from oriental "edible birds nest" by a modification of the
method of Lawton et al. (1956), Edible birds nest is a salivary mu-
coid produced by the CoUocalia species of swift. The second method
included the use of ovomucin obtained from hens eggs (Gottschalk
and Lind, 1949) in an inhibitor reduction (IR) titration following com-
bination of virus and ovomucin (Isaacs and Edney, 1950).

344

COLD AND VIRUS INFECTIVITY

" E
o

E .S

^ E

^ c

X) y.

5) "s o

E ^

Ego

o < ^
^ r

3 o.

"^ CO

y\jM ^uj aad v|si aaa j StI

2 E

345

METCALF

.25 .5 1.0

3.0 4.0

TIME (Houri)

Figure 2. The effect of neuraminidase concentration upon the release of neuraminic
acid from mixtures of influenza virus (A2/Jap 30 5/57) and neuramin mucoid at 37^ C.

Using neuraminmucoid and enzyme mixtures incubated at 37° C,
free neuraminic acid was linearly split off during the first 20 min-
utes. When the same mixtures were combined and incubated at 20° C
a slow linear increase in free neuraminic acid was observed during
the 30 minute test interval. No reaction took place within 30 minutes
at 4° C. The use of the IR method for measurement of enzyme action
confirmed the findings of the TB A analyses. Differences observed in
the results obtained by the two methods were indicative either of a
different order of sensitivity, or a difference in the substance being
measured. For example the IR titration gave a yield of approxi-
mately 90 per cent split product compared to about 53 per cent for
the TBA method. Enzymic action at 20° C demonstrated by TEA
analysis was not shown by the IR method.

The concentration of enzyme affected the release of neuraminic
acid as shown in Figure 2, Not only was the rate of release propor-
tional to enzymic concentration, but also to the total amount re-
leased. The same result using other substrate materials at 37° C

346

COLD AND VIRUS INFECTIVITY

>> E

3 O

(1) S

^1

I-

y^^4 X^ jad VN a»J J ^tI

5 ^

fc ^

Jo u
P

fO (M

1^

S" E

5 c

a "^

»> 3 "3

H ii

)3npoJei mds %

347

METCALF

was reported recently by Noll, Aeyagi and Orlando (1961) and Mayren
et al. (1961).

When ovomucin was used in place of neuraminmucoid, different
results were obtained, Enzymic destruction of substrate occurred
rapidly at all temperatures when measured by the IR method. Only
slight differences in the rate of of substrate hydrolysis and the time
required to reach the same endpoint could be shown. The results of
the TBA analyses showed the rate of enzymic attack on ovomucin to
be similar to the values found with neuraminmucoid. The efficiency
of the hydrolysis was considerably improved, however, and indicated
a more labile, more available enzyme-vulnerable structure (Fig. 3),

The inhibitory titer of the two inhibitor preparations for indicator
virus varied markedly. Based on the bound neuraminic acid content,
ovomucin containing 3 x 10"'^ micrograms was inhibitory. Using the
same basis neuraminmucoid containing 7 x 10"^ micrograms was
inhibitory. The ovomucin inhibitory value of 3 x lO"'^ micrograms
agreed well with the value 4-8x10"'^ micrograms reported by Gott-
schalk and Lind (1949) for their preparation of ovomucin, Tamm and
Horsfall (1952) reported an inhibitor titer of 3 x 10" '^ for urinary
mucoprotein at equilibrium with influenza virus. There are no
values published for neuraminmucoid.

Ovomucin contained 0,18 per cent total protein, determined by the
Biuret test ; 0.0 5 per cent hexose, deter mined by the modified Orcinol
method of Rosevear and Smith (1961); and 1.5 micrograms per ml of
total bound neuraminic acid. The neuraminmucoid preparation con-
tained 4.86 per cent total solids, 1.68 per cent total protein, 0.97
per centhexose,andllmicrogramsper mlof bound neuraminic acid.

For the purposes of the study, the results showed that neuramini-
dase continued to attack substrate at 20° C, but with decreased
efficiency. Both the rate and extent of enzyme action were affected.
The marked difference in the values obtained with ovomucin and
neuraminmucoid substrates was interpreted as the result of chem-
ically different compounds possessing enzyme- labile structures of
varying accessibility and possibly different chemical composition.

348

COLD AND VIRUS INFECTIVITY

SIMULTANEOUS MEASUREMENTS OF VIRUS, INHIBITOR, AND
ENZYME LEVELS DURING INFLUENZA INFECTIONS

In order to examine the effect of low temperature upon virus in-
fectivity, studies were directed toward the relationships existing be-
tween virus multiplication, enzyme production, and inhibitor levels
in the infected host. The relationships were determined first at the
normal environmental temperature of the host and secondly at sub-
normal temperatures. Both the chick embryo and white mouse were
used as experimental hosts .Special attention was given to A2 strains,
several of which had been isolated from fatal human cases of influ-
enza.^ In addition, A and Al strains were used. The enzyme activity
of the different strains varied from the Al which possessed low
reactivity, to the A strains which varied from low to intermediate
reactivity, to the A2 strains which had the greatest activity.

A typical experiment using the chick embryo was conducted as
follows. Ten-twelve day embryos were injected via the allantoic
fluid with 0.1 ml volumes of virus of known concentration. A similar
number of control embryos were injected in the same way with the
same volume of diluent. Control and test embryos were incubated at
36° C to 37° C and 20° C. Six to ten embryos from each group were
removed at regular intervals following injection, the allantoic mem-
branes were harvested, washed, and weighed, and 50 per cent sus-
pensions were then prepared, Aliquots of the suspensions were
rapidly frozen and stored at -70° C. The membrane suspensions
were examined for their virus content by means of chick embryo
LD5Q titrations. The enzyme content was determined by TBA or IR
assay, and the inhibitor level was measured using indicator virus.

The result of a typical experiment conducted at 37° C using A2
strains was as follows. Virus multiplication was indicated by a
steady increase indetectable virus after 1 hour. Beginning at 4 hours
the inhibitor content of the membranes began to decline and con-

1 Obtained through the courtesy of Dr.R.Q. Robinson, International Influenza Center for
the Americas, Communicable Disease Center, US Public HealthService, Atlanta, Georgia,
and Dr. A. F. Rasmuss»en, Jr., University of California Medical Center, Los Angeles,
California.

349

MET CALF

r. 'a

— -3 £

■7, o

E :i 3

au'Ejquiaui uivj9 jsd yisj 93J j 8tI

1 s :5

BTXJTA

in ■* M ^4

0

000

0

0

000

..\

.... .^.

y

y

/

\^ ■••••

/

,

y

. y

y

■ •••.

y

'•••..

^
^

-\

—

r'
1

1

^^>^

•3 i

.~ >> LO

c ^

O ^H

P H.

o —
5 c

Zl

Z ^ o z:

3ujXzu;3

jo;iqn^ui

^ 2 g

II

350

COLD AND VIRUS INFECTIVITY

1 ._

: I

1 \
1 \

1

. 1

'••' "T

/

i

l\ /-

if- 1 -

— Jl 1^ E "5

S 2 2

0

- W

E t; o -•

sucjqiuaui iu«j8 jad y^j aaj j 8i1

•^TA

O Ol O Ol 0| Ol Ol 0| Ol o

I I I I I I I I I I -I

5 i

.2 -o
t? E

^ c

IS
^ u

C 0)

\ x: 2

— ^ 4^ O

Q. N O -

ii m o u ^

•g 3 o 3 ♦J

E '^ tn 1 I

=" - S ^

3 c - S^ t-

auiXzu^

joj^qpr^i

BC o

J3 M js -a

351

MET CALF

s^JTA

r<\ IM

O OI O j o o o

o

o

o

o

~^ ^

^

/

/

y

/

y

/

/

/

::>

1 , — ^^

I 1

.2 a

3 Si

> *^

3UlXzU3

aojiqujui

So o 0
r *• o

sruiA

auiXzu^

rt

^

1

4)

c

F

1

fc

O

01

, ,

>>

7?

E

Q

b

o

U

a

0)

n

0)

^2

h(l

V

0

•^

n*

"^^

^^^

g

^

c

n

CO

1

n

0

a

b

f

■*-»

■^

c

ni

J.

o

c

."^j

c

o

Q.

?^

o
o

o

3
E

tn

3

%

tn

3
>

3

L.

c

>

,

4-*

r.

1

to

2

(4

0

o

d

o
o

In

O

0

t>

CO

u
o

5#S

joitqiqui

x; ^ £ -a

352

COLD AND VIRUS INFECTIVITY

'^ (D ■-

-i o 33 s-

^ 5

3 > ^

PPV onrau^jnsfj aaj J 6^

3

"" ,R

i

•2 d

o _

:i5

c «

S .2
1 ^

P

in r)

~ V,

o^

1 ;.

1 °"

-C rt

1 OJ

c t

' h

. >>

E E

rg

^1

O •

^ 4,

4} —

DC m

o ..

c B

O '2

9

2 1

etUTA

O 4)

.2 cu

n aj

<u

suiAzu^

O

-J u

he o '^
■-' *j c~

Ck ■- CO

jo^iqiqui

353

METCALF

tinued to fall throughout the 24 hour test period. Sometime after 12
hours the production of enzyme withinthe membranes was detected,
with a steady increase shownduringthe remainder of the test period.
Confirmation of the fabrication of enzyme within the infected allan-
toic membrane was obtained in separate experiments by detection of
a rise of free neuraminic acid as measured by the TBA assay. The
level of free neuraminic acid in infected membranes began to in-
crease at 12 hours after infection and continued to increase through
24 hours. No increase could be shown for the non- infected mem-
branes (Fig, 4).

When experiments of this kind were conducted withA2 strains in
chick embryos incubated at 20° C, the following results were found.
There was no significant increase in the virus concentration of mem-
branes, no decrease in inhibitor level, and no enzyme formation was
detected. Again there was agreement between inhibitor reduction
titrations and TBA assays (Fig. 5).

The findings obtained in chick embryos with A2 strains possessing
a high enzyme activity were next compared to the results found with
a PR-8 strain of low enzyme activity. Virus multiplication and en-
zyme production occurred, and the inhibitor level decreased accord-
ingly when the experiment was conducted at 37*^ C. At 20 C no virus
growth or enzyme production was detected. The inhibitor level
fluctuated but did not show a steady decline (Fig. 6).

Throughout the study a virus inoculum of 1000 LD50 doses was
used. It was decided to determine the effect of increasing the doses
given in the inoculum upon the relationships of enzyme production
and inhibitor content. Accordingly, an inoculum of 10 ' LD^q doses
per embryo was used. The net result was an accelerated appearance
of enzyme and earlier decline of inhibitor. The virus content of
membrane reached an early peak accompanied by a rapid rise in en-
zyme production between 6 and 9 hours. Membrane inhibitor fell
rapidly between 2 and 6 hours (Fig, 7),

The pattern of events emerging from these experiments confirmed
the existence of a direct relationship between virus multiplication
and the appearance of enzyme activity in the allontoic membrane.

354

COLD AND VIRUS INFECT I VITY

Nell et al., (1961) reported similar findings in an earlier publica-
tion which dealt with infected chick embryos. Accompanying the rise
in virus titer and enzyme content was a corresponding drop in the
concentration of membrane inhibitor. This finding tended to support
the contention of Isaacs and Edney (19 50) who indicated a role for in-
tracellular receptors in influenza infections on the basis of a corre-
lation between the effectiveness of indicator viruses as interfering
agents and their position in the "inhibitor gradient" of allantoic mem-
brane extracts. The time of appearance of enzyme depended on the
speed ofvirus growth. When 10 LD5Q doses were given, enzyme for-
mation was usually detected after 12 hours. If the number of LDgQ
doses administered was increased to 10 ''', enzyme formation was ob-
served between 6 and 9 hours. These results were in agreement with
previously reported findings on virus multiplication by Henle(1953)
and Acker mann and Francis (19 54). When the environmental tempera-
ture was decreased to 20° C, virus growth failed to materialize, no
enzyme was fabricated and the inhibitor level remained unchanged.

Once the relationships of enzyme fabrication, virus growth, and

inhibitor decline had been shown in embryos, attention was directed

to these same considerations in influenza infected mice. Albino mice

o
were infected intranasally with mouse adapted A2 and Al strains.

Two groups of mice were established, one group- caged at the normal
animal room temperature of 20° C, and the other at 4° C. The mice
placed at 4° C were individually caged in plastic containers without
bedding material. Water was added to the containers in amounts ade-
quate to give a thindiscontinuous film of moisture covering the floor.
The mice were maintained at4°C until signs of distress were noted.
These included shivering, ruffled fur, apathy to stimuli, and blanched
ears. Mice from the two groups received an inoculum previously de-
termined to give 50 per cent lung consolidation at 48 hours in mice
maintained at 20° C. An equal number of control mice from the two
groups received an inoculum of sterile saline. Further control mice
were inoculated to determine the 50 per cent lung consolidation end-
point. Mice from test and control groups at 20° C and 4° C were
sacrificed at 0, 8, 24 and 48 hours. Lungs and trachea were exam-
ined for their virus titer, enzyme, and inhibitor content,

2 Obtained through the courtesy of Dr. T. Francis, Jr., School of Public Health, Uni-
versity of Michigan, Ann Arbor, Michigan.

355

MET CALF

BtUU

M a> . CO , r» . >o . «n ^ f^ <^^ —

ol o o o o o o o o o

l-J-^'T

3UlXzU3

jojTqn{ui

is

o *

" 8

Q .

u •
u •

■s =.

■F ^

- 1

3

cf

<

Oi

2

c

o

2

c
o

c

CM

0

o
en

3

O

3

P

■*^

tn

a

.T

>

^

0)

1

■s

E

3

'a

1
1

c

0

3

■^

o

Ih

c

>

c

u

u
0

J3
00

0

x:

4)

0

o

73

^

^

bi

o

iS

3

6^

snatA

c

ooloooooo o

= E
o 1

■8 1

E S

5 3

•_i^ 00 ,^ 'i

> w

aui^z"3

aojiqiqui

c»< 5 SP

356

COLD AND VIRUS INFECTIVITY

snjTA

q ol oi ol oi o oi o| o o

I \

I -^ \ .

I \

I \

I \ •

I \

I \

o

i

tn r

^2

i

■2 H5

r.

1) ^

E ^ =

1? S

a 1)

o

11

1

O

?.s

«ii.

(U 00

!

rg

2

c

0

P

1 :S

s

4) w tn
6f5

u 4)

t i

•a S

« U

o ,s

jo)Tqn{ui

8TUTA

.^^

•B5

n 4)

c 'f

00

.2 §

S E

11

a ^

0, 'a

^^

E Z

«

>. 1-

^ -

JS

•* ,,

0 u.

rg U

2

3 <

P

Q. "^

5 S

■3 Si

^?

00

3 -

u c

> -

. c

A Si

Oi C

0. 0 .

b " u

aojTqyxpii

357

METCALF

Gross macroscopic examination of the lungs showed complete con-
solidation to have occurred in the mice kept at 4^ C while 50 per
cent consolidation was found in lungs removed from mice kept at
20° C. Enzyme production occurred in both groups, but inhibitor
levels declined only in the cold exposed mice. A cold environment,
therefore, led to a more rapid growth of virus which in turn led to a
greater yield of enzyme and a decline in inhibitor content (Fig, 8).

The importance of neuraminidase activity to cell damage was
shown by the results obtained following inoculation of mice with Al
and A strains having no detectable enzyme action. Virus multipli-
cation was limited, no enzyme was produced and there was no loss
of inhibitor. The lungs removed at sacrifice were normal in appear-
ance and no cell damage could be demonstrated (Fig. 9).

FLUOROCHROME STUDIES OF EXPERIMENTAL
INFLUENZA INFECTIONS

Host response to virus infection at the cellular level was studied
with the help of the fluorochrome acridine orange. Monkey kidney
monolayers were infected with A2 strains of high and low enzyme
activity at environmental temperatures of 20° C and 37° C, and ex-
amined at intervals up to 5 days. "Flying coverslips" were fixed in
chilled absolute methyl alcohol, hydrated, and stained 5 to 9 minutes
with a 1 to 10,000 dilution of acridine orange at a pH of 3.8. The
coverslips were examined with a Fluorestar microscope used in
conjunction with an Osram HBO- 200 illuminator. The exciter filter
used was a Corning No. 5860, half- thickness filter. Photomicro-
graphs were made using Ektachrome, type B film.

Non- infected monolayers showed the green fluorescing nuclei
characteristic of DNA material, with yellowish- white nucleoli. The
cytoplasm exhibited a rust color symptomatic of RNA. Six hours
after infection the cytoplasm around the nucleus began to show a
brighter rust fluorescence. By 18 hours the fluorescence became a
bright orange- red color and could easily be distinguished. At 30

358

COLD AND VIRUS INFECTIVITY

hours cjdioplasmic inclusions were observed. From 72 to 96 hours
the infected cell showed extensive cytopathogenic change. The spe-
cificity of the stain was controlled by the use of RN'ase which made
it possible to distinguish the presence of nonspecific staining.

The sequence of events in the invasion of cell monolayers did not
differ noticeably at 37° C or 20° C. The presence of virus, its intra-
cytoplasmic growth, cytopathology, and cell destruction, could not
be differentiated with certainty at the two temperatures.

Examination of allantoic membranes was accomplished by means
of paraffin sections prepared after the method of Coons and Sainte-
Marie (1960). Three to five micron sections were cut on a rotary
microtome and stained with acridine orange. Virus growth at 37° C
was shown to occur predominantly in the entodermal cells lining the
allantoic cavity, although some virus staining material was found in
the ectoderm. Infected membranes kept at 20° C showed practically
no virus staining material in the entodermal cells.

When sections of mouse lung were examined, the extent of virus
invasion was shown. A section of non- infected mouse lung failed to
show virus- staining material. Sections of lungs from mice infected
at 20° C showed virus-staining material gathered around the bronchi
as well as scattered throughout the lung. Sections from infected
mice of the 4° C group showed a more massive accumulation of
virus- staining material. Greater concentrations of virus appeared
immediately adjacent to the bronchi.

The results of the fluorochrome examinations furnished visual
evidence that the neuraminidase- active strains of influenza used in
the study found their way into host cells of chick embryos, monkey
kidney monolayers and mice. The initial stage of cell invasion was
not affected by the degree of enzyme activity associated with the A2
strains. Thus, strains of high enzyme activity could not be shown to
possess a penetration advantage over strains of low enzyme activity.

359

METCALF
DISCUSSION

The measure of infectious virus used in the study was the ability
to multiply sufficiently within a host cell to damage it by means of
enzymic attack. The direct relationship shown between cell damage
and neuraminidase production led to this conclusion. Proceeding on
this basis, the results indicated that cold had no effect per se upon
the infectious quality of virus. For example, the penetration of a host
cell was not affected, and the in vitro effect upon neuraminidase ac-
tivity was minimal. Instead, cold was shown to exert an influence on
the host- parasite relationship subsequent to virus penetration. The
findings suggested that cold acted to alter some host structure im-
portant for defense and accordingly render the cell more vulnerable
to enzyme action. It was postulated that the structure affected by cold
was mucoprotein in nature. This viewpoint would represent a modi-
fication of Gottschalk's enzyme penetration theory (19 57), The impor-
tant difference lies in the relationship of enzyme activity to cell
damage in the present study, rather than cell penetration.

The results were an extension of the findings of Sulkin (1945) who
reported a more serious infection in cold exposed mice with little
change in mortalities. The fatalities occurring in the present study
were attributed to a combination of enzymic competence of the virus
strains used and physiologic insult sustained by the mice.

The findings of the present study constituted an interesting con-
trast to the results reported by Sarracino and Soule (1941), These
authors suggested that infection depends on the amount and virulence
of virus rather than the general resistance of the host. My study
indicated that virus possessir^ enzyme activity was infectious and
thus virulent. It also showed that the severity of influenza infections
was not the exclusive result of virus virulence, but could be con-
ditioned by cold-induced physiologic alterations within a host.

An attempt at translation of the results of the present study in
terms of the "winter factor" of Andrewes (1958) would be presump-
tuous. The results did suggest, however, that a study of physiologic
alterations of mucoprotein. structures of a cell exposed to virus

360

COLD AND VIRUS INFECTIVITY

possessing enzymic competence might provide clues to the nature of
the winter factor.

SUMMARY

Infectious influenza virus was considered to be characterized by
an enzymic potential capable of causing host cell damage. No effect
of cold upon infectious virus per se could be shown. Any influence
exerted by cold upon experimental influenza infections was shown to
be closely related to a host- induced effect. Host cell damage was be-
lieved to follow enzymic action upon muco- protein structures of the
cell.

LITERATURE CITED

1. Ackermann, W, Wilbur, and Thomas Francis, Jr. 1954. Charac-

teristics of viral development in isolated animal tissues. Ad-
vances in Virus Research 2: 81-108.

2. Andrewes, C. H. 1958. Symposium on the aetiology, spread and

control of epidemic influenza. A, The epidemiology of epidemic
influenza. J. Royal Soc. Health 78: 533-536.

3. Coons, A. H., and Guy Sainte-Marie. 1960. Personal Corre-

spondence.

4. Fazekas de St. Groth.S. 1948. Viropexis, mechanism of influenza

virus infection. Nature 162: 294-295.

5. Fazekas de St. Groth, S.,and D. Graham. 1949. Modification of vi-

rus receptors by metaperiodate. Infection through modified re-
ceptors. Australian J. Exp. Biol. Med. Sci, 27: 83-98.

361

METCALF

6, Gottschalk, A. 1957. Neuraminidase: The specific enzyme of in-

fluenza virus and vibrio cholerae, Biochim, et Biophys. Acta
23: 645-646.

7, Gottschalk, A., and P. E. Lind. 1949. Ovomucin, a substrate for

the enzyme of influenza virus. I. Ovomucin as an inhibitor of
hemagglutination by heated lee virus. Brit. J. Exp. Pathol, 30 :
85-92.

8, Henle, W. 1953. Multiplication of influenza virus inthe entoder-

mal cells of the allantois of the chick embryo. Advances in
Virus Research 1: 141-227.

9, Isaacs, A., and M. Edney. 1950. Interference between inactive and

active influenza viruses in the chick embryo. I. Quantitative
aspects of interference. Australian J. Exp, Biol, Med, Sci.
28: 219-230.

10, Lawton, V,, J. V. McLoughlin, and W. T.J. Morgan. 1956. Solu-

bilization of water- insoluble mucopolysaccharides and new
method for preparation of blood group substances. Nature 178:
740-741,

11, Mayron, L. W., B. Robert, R, J. Winzler, and M, E. Rafelson, Jr.

1961. Studies onthe neuraminidase of influenza virus. I. Separa-
tion and some properties of the enzyme from Asian and PR-8
strains. Arch. Biochem. and Biophys, 92: 475-483.

12, Noll, H,, T. Aoyagi, and J. Orlando. 1961. Intracellular synthesis

of neuraminidase following infection of chorioallantoic mem-
branes with influenza virus. Virology 14: 141-143,

13, Rosevear, J, W,, and E, L. Smith. 1961. Glycopeptides. I. Iso-

lation and properties of glycopeptides from a fraction of human
gamma- globulin. J. Biol. Chem. 236: 425-435.

14, Sarracino, J, B., and M, H. Soule. 1941. Effect of heat, cold,

fatigue and alcohol on resistance of mice to human influ-
enza virus. Proc. Soc. Exp. Biol, Med, 48: 183-186,

362

COLD AND VIRUS INFECTIVITY

15. Sulkin, S. E. 1945. The effect of environmental temperature on

experimental influenza in mice. J, Immunol. 51; 291-300.

16. Tamm, I,, and F. L. Horsfall, Jr. 1952. A mucoprotein de-

rived from human urine which reacts with influenza, mumps,
and Newcastle disease viruses. J. Exp. Med. 95: 71-97.

17. Warren, L, 1959. The thiobarbituric acid assay of sialic

acids. J. Biol. Chem. 234: 1971-1975.

DISCUSSION

REINHARD: Dr. Walker, what is the temperature of a baby
mouse?

WALKER: I don't know, but I would like to. This brings up a
very interesting point. I indicated that we have begun work in-
volving using both tissues of baby mice and adult mice. It is
quite evident that raising the body temperature of a baby mouse
does not produce the same effects that I showed here for the
adult mouse. We have exposed infant mice to incubator tem-
peratures up to about 35° C. Above this temperature, the infant
mice do not survive. I think perhaps the baby mice could tolerate
higher temperatures, but the mother refuses to feed them or
her lactation stops at higher temperatures. She apparently has
enough troubles at that temperature without having these little
furnaces hanging on her, so she retreats to the opposite side
of the cage and the mice are not fed. I don't know what the tem-
perature of baby mice is at ambient temperatures of 35° C,
but I think it is probably higher than normal. I haven't succeeded,
so far, in finding a probe that will allow me to measure the
body temperature of a forty- eight- hour- old infant mouse.

One other thing should be pointed out, too, and that is that the
pancreas appears to be different from other tissues. Coxsackie
virus will multiply in the pancreas of the adult mouse even at

363

METCALF

ordinary temperatures, but one can raise the body temperature
sufficiently so that even this multiplication is choked off.

NUNGESTER: This pancreas thing intrigues me very much.

WALKER: It intrigues me, too, because there are other vi-
ruses that seem to multiply well in the pancreas.

NUNGESTER: But by and large, the pancreas is not suscep-
tible to bacterial infection, at least as far as I know.

WALKER: But I think it is susceptible to a number of viruses.

REINHARD: I might offer the information of some recent gastro-
enterological research which indicates that the pancreas has
different proclivity for amino acid uptake. Certain kinds of amino
acids are readily absorbed by the pancreas, and this might be
one of the differences causing increased tissue susceptibility.

WALKER: So far, we have had great difficulty in trying to
get it to persist or multiply in culture.

METCALF: Dr. Walker, I think you have a very interesting
theory. In view of your results as well as the work of Dubos and
Wenner,^ I wonder if you are forced to describe a polydisperse
virus population with respect to genetic variance. Are you se-
lecting one member of a polydispersed population at a given
temperature?

WALKER: I am not any more worried with that than with any
other explanation of multiplication at any temperature. I think
temperature is going to be a selective force on virus popula-
tions, and I think that, as Lwoff has shown, it is possible to
isolate strains that are capable of multiplying at higher tem-
peratures. This is part of the basis for his contention that tem-
perature is related to virulence. Is this pertinent to your question?

1 Dubes, G. R., and H. A. Wenner. Virology 4: 275-296.

364

COLD AND VIRUS INFECTIVITY

MET CALF: I wondered if you felt this to be a condition which
might prove deleterious to your theory.

WALKER: No, I think that this is reasonable. This is just
another selective pressure, and we will have the same problems
with selection that will apply to any other explanation of the
effect of temperature on infection.

CAMPBELL: I might point out that it is pretty well known
that in neonatal animals there is a question whether they pro-
duce any antibodies at all so that when you expose them to anti-
gen, they may actually exhibit a so-called tolerance or immune
paralysis. Their immune mechanism is not working at this point.
You had an exposure of six days at 25*^ C.

WALKER: And then a shift to cold.

CAMPBELL: Yes, and then in this case, you didn't have any
effect.

WALKER: Right.

CAMPBELL: I was wondering whether you thought that this
exposure, some way or another, actually causes an adaptation.

WALKER: The point was, if they were held at ordinary room
temperature for as long as six days after receiving virus, then
the shift to 4° C had no effect, but after only four days, the cold
still caused high mortality. My explanation or speculation is
this; by six days, the virus has been reduced to very low levels
in practically all of the tissues except the pancreas. By that
time, there is still a lot of antibody in the blood, and although
virus activity might be increased in the pancreas, there would
be no opportunity for it to be disseminated widely to other tis-
sues that would be made more susceptible by the lowered tem-
perature. The virus must be widely spread to cells before anti-
body appears in the blood, and the exposure to cold must occur
while there is still a sufficient quantity of virus in the tissues.
I would expect in most of these tissues that the virus is pretty
well gone by six days.

365

MET CALF

PREVITE: Your findings are parallel in some ways to those
that I have reported on endotoxin. If cold exposure was delayed
too long, the animal had manipulated this poison somehow so
that he was no longer sensitized to it by cold stress.

WALKER: Then your test material is gone.

PREVITE: Correct.

SULKIN: Dr. Campbell called attention to the fact that the
infant mouse does not produce antibody, Coxsackie viruses are
unique among viruses in that they elicit an antibody response
very, very, rapidly. In man, for example, antibodies are demon-
strative very early after first signs of infection. Yet a week-
old mouse will produce antibody, whereas the one-day- old mouse
will not. Overland showed that the maturation of the antibody
forming mechanism in the animal accounts for this rapid acqui-
sition of resistance to infection with Coxsackie viruses.

MITCHELL: Will someone please define for me this infant
mouse? Don't say baby; I know that,

WALKER: In Coxsackie B-1 virus infection, for instance, the
mouse is most sensitive to the virus up to about forty- eight
hours in age. Large inocula of this virus will cause lethal in-
fections in mice up to about eight days of age, but then there
is a very abrupt cut-off with no deaths occurring in mice in-
oculated at an older age. But the mice are most susceptible to
small inocula in their first forty- eight hours after birth.

MITCHELL: I was interested in those extra days of exposure;
after this delay, the animal would be ten, eleven, or twelve days
old.

WALKER: No, you misunderstood, because this exposure to
cold is in adult mice. The infant mouse enters in here only be-
cause the adult mouse in the cold behaves rather like the infant
mouse at normal room temperature.

PREVITE: Dr. Walker, in some of your data you mentioned

366

COLD AND VIRUS INFECTIVITY

that you used ACTH, and this apparently didn't enhance viral
multiplication. What was the regimen that you used and how many
injections did you give?

WALKER: Four mg in gel, as I recall, given every twelve
hours for a period of six days. ACTH has been shown very rarely
to have an effect upon viral infections in the mouse. I don't
understand why it should not. You showed the effect of ACTH
upon toxin, but this may be quite a different thing. ACTH does
produce physiological changes in the mouse that indicate that
the ACTH is active, but the failure of ACTH to aggravate viral
infections seems to be a peculiarity of the mouse. Cortisone
has quite an effect upon viral infection in the mouse, but ACTH
has very little.

McCLAUGHRY: Certainly the thesis that Dr. Watson has pre-
sented appeals to me very much as a person interested in human
physiology because of the fact that the general pattern of inter-
actions of related functions of cells and tissues fits the category.
The modifications of a particular kind of cell reaction or tis-
sue reaction modifies the balance among all functions, and there-
fore, the modification of temperature at which cells are held
might very well favor certain enzymic processes which would
be involved, perhaps, in replication of viral protein as compared
with certain other cellular processes.

MARROW: Does the level of interferon which responds to
the increase in temperature account for part of this decrease
in virulence with increased temperatures?

ANDREWES: I think it may work another way. Is sacs gave
a review on interferon at the Montreal meeting; one of the things
that he mentioned was that he has been in touch with Lwoff about
this, and he feels, concerning the virulence of the strains of
virus which do better at high temperatures, that perhaps they
are able to do so because they become less sensitive to the
action of interferon.

367

SUPPRESSIVE EFFECT OF LOW ENVIRONMENTAL
TEMPERATURE ON VIRAL INFECTION IN BATS^

S. Edward Sulkin and Rae Allen

Department of Microbiology

Southwestern Medical School

Dallas 35, Texas

ABSTKACT

Having established the susceptibility of insectivorous bats held at 24° C to 29° C to
experimental infection with rabies and encephalitis viruses and determined that virus
particles introduced peripherally may invade and multiply in brown fat in addition
to other tissues, studies were extended to establish the influence of low temperature
on the course of these infections in these animals. Viral proliferation in interscapular
brown fat, an organized, bi-lobed structure believed to be present in all hibernating
animals, suggested that this tissue might provide a focus of infection or target organ
from which virus could disseminate to other tissues. In view of the function of brown
fat in maintaining the hibernating animal, it seemed logical to assume that virus par-
ticles present in this tissue would remain viable during periods of hibernation, sus-
tained by the same mechanism which nurtures the whole animal, and, upon arousal of
the animal would multiply and disseminate to other sites. This rationale has formed
the basis for our studies on the influence of low temperature on viral multiplication
in bats and the data to be presented will be drawn from the following major areas under
investigation: (1) The susceptibility of various species of bats maintained at 24° C to
29° C to experimental infection with rabies and encephalitis viruses. (2) The influence
of low temperature on initiation of viral infection in bats and on the course of a pre-
viously established infection in these animals. (3) The influence of low temperature
on antibody formation in bats in response to bacterial and viral antigens (some in-
formation relative to rates of antigen and antibody degradation in the torpid bat will
also be included). (4) Reference will be made to studies on the influence of low tempera-
ture on viral multiplication in monolayer and explant cultures of tissues of the bat
as compared with tissues of warm blooded animals.

1 The original work reported herein was supported by research grants from the Caruth
Foundation, the National Institutes of Health, United States Public Health Service, and the
Commission on Viral Infections, Armed Forces Epidemiological Board, and was sup-
ported in part by the Office of the Surgeon General, Department of the Army.

369

SULKIN AND ALLEN

The isolation of rabies virus from naturally infected insectivorous
bats in the United States inrecent years has focused the attention of
epidemiologists and virologists on the Chiroptera as still another
reservoir host for this virus in nature. Since our laboratory reported
the first human death believed to have been due to the bite of a
naturally infected bat (SulkinandGreve,1954),webecame interested
in the association of bats with rabies and more especially in how
these animals are able to survive infection with this hitherto pre-
sumed fatal viral infection. Investigators in South America and
Trinidad demonstrated that the vampire bat could suffer sustained
infections with rabies virus and pass the virus through infected
saliva to susceptible animals without themselves showing overt
symptoms or succumbing to thedisease(Enright,19 56). Wenow have
experimental evidence that insectivorous species of bats native to
the United States may be infected with rabies virus and never show
symptoms of the disease even though virus can be demonstrated in
brain as well as other tissues (Sulkin, 1962),

RABIES STUDIES IN INSECTIVOROUS BATS

Studies on the course of experimental rabies infection in insecti-
vorous bats were undertaken with the view to determining the sus-
ceptibility of these animals to peripheral inoculation with rabies
virus and to ascertain which tissues were involved in the infection
(Sulkin et al., 1959). We have also attempted to determine the in-
fluence of certain physiological characteristics of the bat on experi-
mental rabies infection in these animals (Sulkin et al., 1960; Sims,
Allen and Sulkin, unpublished data). In this regard, the phenomenon
of hibernation was the first area to receive attention.

Species of bats which inhabittemperate zones are efficient hiber-
nators who remain quiescent for several months of the year. It has
been suggested that this conservation of energy is the reason why
these mammals, despite their small size, are relatively long-lived
(Griffin, 1958). Also, many investigators (Rasmussen, 1923, 1924;
Remillard, 1958; Johansson, 1959, 1960; Kayser, 1961) believe that

370

VIRAL INFECTION IN BATS

the so-called "hibernatii^gland", abilobed, organ- like accumulation
of brown adipose tissue found in the interscapular region of hiber-
nating animals (and in certain non-hibernators as well), plays a role
in sustaining the hibernating animal during its period of winter dor-
mancy, Histochemical analysis of the interscapular brown fat of the
little brown bat (Myotis 1. lucifugus) indicates that this tissue builds
up quantities of lipid during pre-hibernation months and that the
various biochemical constituents of this tissue are depleted as hiber-
nation progresses (Remillard, 1958), Studies comparingthe nature of
brown and white fat of several animal species indicate that brown
adipose tissue is histologically distinct and physiologically more
active than white fat (Fawcett, 1952), Recognition of the role of brown
fat in sustaining the hibernating bat during periods of inactivity, to-
gether with reports dealing with the multiplication of Coxsackie virus
(Pappenheimer et al,, 1950; Grodums and Dempster, 19 59) and polio-
virus (Shwartzman, 1952) in the brown fat of mice and hamsters,
suggested that this tissue of bats might provide a site for rabies
virus sequestration in the asymptomatic host; a latent focus of in-
fection from which virus sustained during the period of hibernation
by the same mechanism which nutures the whole animal could be
activated to invade and multiply in other tissues, including salivary
gland, and thereby be perpetuated in nature by this host. Accordingly,
experiments were undertaken to determine whether rabies virus in-
troduced peripherally would, in fact, invade and multiply in the brown
adipose tissue of insectivorous bats. Data compiled from such a
study (Sulkin et al,,1959) are summarized inTable I, Two species of
bats, the Mexican free- tailed bat (Tadaridab. mexicana) , which is a
quasi- hibernator, and the little brown bat (Myotis 1. lucifugus) , a true
hibernator, were used in these experiments.A strain of canine rabies
virus isolated from the brain of a fatal human case was used. Each
bat received an intramuscular injection into the heavy muscle over
the chest of approximately 8000 mouse i.e. LD^q contained in 0.1 ml.

Although the Mexican free-tailed bat proved to be relatively un-
susceptible to experimental rabies infection, virus was demonstrated
in the brown fat of 22 per cent of those animals shown to be in-
fected by viral assay in white Swiss mice. The infection in this
species was most evident 20 to 40 days after intramuscular inocu-
lation of virus. On the other hand, rabies virus was found to be
widely distributed on thelittlebrownbat9 to 26 days following inoc-

371

SULKIN AND ALLEN

rxJ

O O

<U G
PL, .H

X)

a>
o

c

w

o
a

CO
(0

CO

o

CO

nH CO

• •

CM ^

CM 00

CD J3-

(0

s

S

o

bO

H

3

X

i+-t

v

•H

£

o

3

•

-•

^1

(TJ

Hi

ta

•H

CO

ti

•H

(0

•H

T3

O

(0

>i

tH

Z

3 M o

8 P §

S O 3 -5

E

3

o 35

^ £

0) 00 J

- 2

3

3

t-i

35

>

Ol

tn

in

0)

05

oj 2 Q. g

--; (U

rf

„

Ifl

i)

^

C

2

o

a.

a
*

"o

3

f

♦

>.

E

c

Q
J

£

o

t.

k.

JO

C

a.

u

o
tn

o

X!

3

v

OJ

M

a.

o

E

o

rt

^

0

—

t-i

u

c

D — —

H s -

372

VIRAL INFECTION IN BATS

Illation, and virus concentration in some of the tissues approached
the level of the stock mouse brain virus suspension used in inocu-
lating these animals. Virus was demonstrated in the brown fat of
30 per cent of the experimentally infected Myotis. These data, to-
gether with the recent demonstration of rabies virus in the brown
fat tissue in naturally infected insectivorous bats, would support the
h)T)othesis that this tissue provides nutriment for a latent focus of
infection (Bell and Moore, 1960 ;Sulkin, 1962), Studies are in progress
to determine the frequency with which virus can be demonstrated,
under natural conditions, in the brown adipose tissue of bats netted
in different geographic areas at different times of the year.

The mechanism by which brown fat may actually serve as a depot
for storage of rabies virus is still under study, A working hypothe-
sis concerns the possibility that at least in hibernating species,
rabies virus sequesters in brown adipose tissue in the quiescent
host during the period of hibernation and is subsequently activated
by the physiological alterations which occur prior to awakening and
by emergenceintoawarmer environment. Studies concerned with the
effect of low environmental temperature on the pathogenesis of ra-
bies in insectivorous bats would suggest that this may actually be
the case. The remaining portion of this discussion will be limited
to those areas which relate to the subject of this conference.

It is clear from previous discussions in this conference that en-
vironmental temperature has been shown to have a significant effect
on several experimental virus-host systems, both in vivo and in
vitro. It is also apparent that the outcome of such experiments de-
pends on the particular virus used and on the host or cell system
under study (Sulkin, 1945; Walker and Boring, 1958; Hoggan and
Roizman, 1959; Lwoff, 1959).

During the course of studies on the role of Chiroptera as reser-
voir hosts for viruses innature,itbecame evident that these exper-
iments with bats provided an unusual opportunity for determining the
effect of temperature on viral multiplication in the intact animal.
The bat does not possess a thermoregulatory mechanism for main-
taining a constant normal body temperature, but rather, the body
temperature of a bat is that of his environmnet, except when he is in
the active state of walking or flying (Hock, 1951; Morrison, 1959),

373

SULKIN AND ALLEN

The body temperatvire of resting bats has been shown to parallel
closely that of their environment over a range of 2° C to 30° C, and
the metabolic rate of these animals varies directly with their
body temperature (Hock, 1958).

Data have been accumulated comparingthe course of experimental
viral infections in bats held at 5° C or 10° C with groups held at
24° C or 29° C and with groups held at low temperatures for a
period of time and then transferred to room temperature (24° C) or
29° C (Sulkin et al., 1960), Animals held in the cold become torpid
within hours of being placed at the low temperature, and we have
on occasion determined the rectal temperature of a number of cold-
adapted bats and found it to be quite close to ambient. On the other
hand, we cannot be positive that animals held at 24° C or 29° C
maintained constantly body temperatures equivalent to their environ-
ment. For a period of time following the handling associated with
virus inoculation and transfer to experimental cages, the bats
appear excited and are quite active in moving about their cages. It
is likely that during this period the animals have body temperatures
well above ambient. Within a day or so, however, they become quiet
and hang in groups in their cages; they are seldom seen moving
about except in the evening when they come down to the floor of the
cage to receive the food and water offered daily. There is doubt-
lessly considerable variation among body temperatures of individual
bats at these higher temperatures, depending on their degree of
activity. The difficulties inherent in recording individual body tem-
peratures on large numbers of infected bats precludes the possi-
bility of obtaining such data. In this presentation we will be com-
paring the course of viral infection in animals whose body tempera-
tures we know to be low (ambient temperature 5° C or 10 C) with
the course of the infection in animals whose body temperatures we
feel must have fluctuated between ambient (24° C or 29 C) and
possibly 36° C or higher (Morrison, 1959).

Early in these studies it became apparent that animals collected
during the fall or early winter months survive for longer periods of
time in the cold than animals obtained in the spring or summer
months. Subsequently, these experiments were initiated only in the
fall, even though they could onlybe repeated or extended on a yearly
basis. Furthermore, a recent report by Menaker (1962) indicates
that the state of h3TDothermia achieved with summer bats is not

374

VIRAL INFECTION IN BATS

identical with their winter hibernation, and serves to emphasize the
necessity for carrjdng out low temperature studies in bats gathered
during the fall months.

Table II summarizes a study of the suppressive effect of simu-
lated hibernation on the susceptibility of little brown bats to two
strains of rabies virus. Of those receiving the canine rabies virus
(Thompson) and placed in the cold room,infection was demonstrated
in only 7.5 per cent of 40 bats studied over 40 days. The infection
rate amor^ animals kept at 29° C was 36 per cent, and virus was
demonstrated between the ninth and thirtieth day following virus in-
oculation. It is interesting that virus was demonstrated more fre-
quently in the brown adipose tissue (25.7 per cent) than in the sali-
vary gland (11.4 per cent). The infection rate was even higher among
animals held in the cold for two weeks and then transferred to the
warm room. Again virus was demonstrated more frequently in the
brown fat (30.4 per cent) than in the salivary gland (17.1 per cent).
Quantitative data not shown in this tabulation indicate that virus ti-
ters in the few animals which developed evidence of infection while
in simulated hibernation were 10"^^ or less, while titers in various
tissues of animals in the warm room ranged from one log unit to as
high as 10"'^^. In the experiment with the rabies virus (59V13B) re-
covered from the pooled brown adipose tissue of naturally infected
little brown bats,virus wasagaindemonstratedinfequentlyintissues
of animals inoculated and placed directly in the cold room, but re-
mained viable and was activated when animals were transferred to
the warm room.Although the infection rates were similar with both
rabies virus strains, a careful study of the results revealed differ-
ences in the patterns of infection produced by these viruses in active
bats and in those awakened foUowir^ a period in simulated hiberna-
tion.The bat rabies virus occurred more frequently in the brown adi-
pose tissue than in either salivary gland or brain,It can be seen that
virus was detectable in the interscapular brown fat in 92 per cent
of animals shown to develop infection, and inthe salivary gland and
brain in only 28 per cent and 50 per cent respectively. Rabies in-
f6ction was demonstrated in 42 per cent of the animals held in the
cold for 17 days before transfer to the warm room, and the infection
pattern seemed to be affected by the period in simulated hibernation.
In these animals virus was demonstrated in the brown fat, salivary
gland, and brain of infected animals with about equal frequency.

375

SULKIN AND ALLEN

as > w

•H

^— \

^

CO

/-s

o

l£>

ro

CN

•

•

r-i

•

•

^

H

rH

^— '

o

ID

CQ

o>

CD

in

ID

•K

C

-o

•H

S

-o

.H

0)

bO

^

•-\

in

^

4->

.— s

•

•

•

•

(0

^

rH

H

C^

o

GO

CN

(4

N^'

H

iH

CN

m

4-»

(0

(0

>

C

•H

o

rH

e

m

0)

CO

-a

V)

•M

3

ITJ

t:*

^^

•H

C^

.a-

O

H

>

g

<->

•

•

/^

•

•

rH

in

o

rH

CN

t>«

o

s.^

CN

CO

>■— '

CD

in

P4

OQ

+j ^3

C «

Q) +J

O O

in

rH

CD

CO

O

o

0)

•

•

•

•

•

•

(4 >4h

c^

(D

CO

If

o

CN

0) c

CO

Jt

m

^

CU 'H

■a

0)

■a

4-'

0)

o

■M

<u

W

l+H

<U

o

r^

r^

t^

00

O

c

+J

^

CT>

zt

^

CN

in

'f-i

'^

'^

•^

^^

*^

^^

•

en

in

CO

CN

J-

<-i

•

o

CO

CSI

r-i

CN

o

2

2:

rH

ID 0)

•P U

*

C 3

4C

«

0) 4->

«

*

6 «

o

o

o

O

O

o

C ^

o

in

O)

Oi

in

(D

CD

O (1)

CN

CN

CN

CN

•H e

C^

C^

> 0)

m

m

C -P

u

o
in

Q

o

o

r-i 0)

J

o

o

3 W

o

o

O 3

•

»

•

o o

O

00

(O

c

E

•

6
0) o

B (0

S i

m

CO

■M M
«» >

in

a

§ >; « «

2 2 5 E

• 2 4) 5

— I (U »H <

^ p a ♦

"^ 4) C

15 «

;2 ^'

^2
•5 "3

- tfl '
^ 9. V.

IM rt) uj
■^ ^ **H

?• E

■a o -^ o

m c o <i;

T3 "S S tfl

rt £ o >.

^ jc =«

2 ►^ to -o

Q "1 J2 2:

. tn 2 t,

§ « £ «S

•0 0 0°*-

ii S > E ^

2 !« 0) 4) ^

§ 3 a -2 ■=

::; o ;: ♦^ .

H-. ca to ^

■5 s

^ >

0) •<

* ct)

to g

to t-^ a

c <a '^

bc tA 7?

2 J=

376

VIRAL INFECTION IN BATS

suggesting that after a period of latency in a dormant animal, acti-
vated virus may reach the salivary gland more rapidly and with
greater frequency. In addition, quantitative studies have indicated
that the virus concentration in this tissue is greater than in infected
animals which have not experienced a period of hibernation. These
data are presented graphically in Figures 1 and 2,

The results of these experiments suggest that seasonal fluctu-
ations'in environmental temperature may provide not only a mech-
anism for virus storage in this reservoir host, but may also increase
the chances of virus transmission by thishost.lt is recognized that
this concept may only apply for strains of rabies virus which
circulate in bat populations.

ARTHROPOD- BORNE VIRUS INFECTIONS IN BATS

During the course of studies on experimental rabies infection in
insectivorous bats, we became interested in determining how these
animals would react to experimental infection with other viral
agents, A survey of the literature (Sulkin, 1962) indicated that bats
have been associated in various ways with a number of disease-
producing agents, suggesting that these animals might act as reser-
voir hosts for many viruses in nature. An area of particular interest
to us, from an epidemiological standpoint, concerns the mechanisms
involved in the overwintering of the arthropod-borne viruses. Al-
though investigators have long sought to determine the whereabouts
of these viruses during the winter months intemperate zones where
mosquito vectors do not carry out year-round transmission cycles,
this void stillremains in our under standing of the complex biological
life cycles of these disease- producing agents. Reports of the sus-
ceptibility of bats to experimental infection with various arthropod-
borne viruses (Ito and Saito, 1952; Corristan, LaMotte and Smith,
1956; LaMotte, 1958) suggested this animal as an additional reser-
voir host, but provided only limited information as to the course of
the infection or the tissues involved. We were prompted, therefore,
to pursue studies on experimental arbovirus infection inbats to de-

377

SULKIN AND ALLEN

tA 1A «^

-M. ««.

>-

CO

2 I

2 -I
|.|

4) ^

4) £

s -

o o

■= 1

- « 2

« in

E J-

■? a.

O a;

1) c -? *^

a E

£ E

3 i

b E ^

378

VIRAL INFECTION IN BATS

S

■? i -

<— ^

S 5

o —

o • , fc, 2

w

vTl

s

?r

1

t

0)

o

3

O ;;;

i: 3 2

2 = i:

5 E

? i-s

s ^

° — "S

S 5 a;

0) S E o

•— i7> n "3

>

c 2

^ H ^ B

379

SULKIN AND ALLEN

determine the degree of susceptibility of various species to certain
virus strains and to locate the tissues involved in the infection which
provided sites for viral proliferation and resultant viremia. Again,
emphasis was placed on determining if virus invaded and multiplied
in interscapular brown adipose tissue, thereby providing a mecha-
nism for survival of arbovirus particles in the hibernating animal
in a manner similar to that described for rabies virus.

In the initial experiments (SulMn, Allen and Sims, 1960), we were
able to confirm certain aspects of the observations reported by
Corristan et al. (1956) and LaMotte (1958), Viremia was demon-
strated in bats following peripheral inoculation of Japanese B or
St. Louis encephalitis viruses and, in addition, the lipotropic char-
acteristic of these viruses was demonstrated. These experiments
have now been extended to include various bat species and several
virus strains. Bats were inoculated subcutaneous ly and placed in
specially designed cages which allowed them to receive fresh food
and water daily with nimimum danger to the caretakers. The virus
inoculum consisted of approximately 150 weanling mouse i.e. LD-^
of mouse brain suspension, or infected tissue culture fluid, or bat
blood. To simulate natural circumstances, bats were kept in near
darkness throughout the day and were offered food at the close of
each working day. Unless otherwise indicated, animals were main-
tained at 24° C * 2° C (relative humidity 65 per cent) and were ob-
served several times daily throughoutthe course of the experiments
for possible signs and symptoms of encephalitis. When tissues were
obtained for virus assay, extreme care was taken to rid specimens
of as much blood as possible so that resulting virus titers would re-
flect virus multiplication in a specific tissue and not virus present
in blood circulating through that tissue. These experiments, which
will be published in detail elsewhere (Sulkin, Allen and Sims, un-
published data), indicate that the Mexican free- tailed bat is only
slightly susceptible to the high mouse passage Nakayama strain of
Japanese B encephalitis virus, but is significantly more susceptible
to a mosquito isolate (OCT- 541) which had been through two hamster
kidney tissue culture passages and one passage in little brown
Myotis. In the latter case, the infection was widely disseminated in
animals sacrificed over a period of 3 weeks, with virus concentra-
tions reaching 3 logs or more inblood, brown fat, and kidney. In one
instance virus was demonstrated in the brown fat and not in the

380

VIRAL INFECTION IN BATS

blood of an animal sacrificed 21 days after virus inoculation. The
passage history of the virus strain seems to deter mine its degree of
infectivity for the Mexican free- tailed bat, where the recently iso-
lated strain is much more infective than the laboratory- adapted
strain. Similar results were obtained with two strains of St. Louis
encephalitis virus. This bat species is quite resistant to infection
with the high mouse passage Hubbard strain, but is quite susceptible
to infection with a strain recovered from a flicker bird and in its
seventh mouse brain passage. Virus was first demonstrated in blood
one to two days after virus inoculation, increased to levels of 4.0
log units by the seventh day, and was still detectable in animals
sacrificed about one month post- inoculation. In addition, virus was
demonstrated with much greater frequency in brown fat than in
brain or kidney. In a few instances virus was demonstrated in tissue
preparations from animals that were not viremic at the time of
sacrifice.

The little brown bat is highly susceptible to the OCT-541 strain of
Japanese B encephalitis virus. Virus was widely distributed in the
various tissues assayed during an observation period of about one
month; virus was demonstrated in the blood and in brown fat with
about equal frequency, and in several instances virus was demon-
strated in the brown adipose tissue and not in blood or other tissues
tested. This bat species is only slightly susceptible to infection with
the flicker bird strain of St. Louis encephalitis virus. The big brown
bat (Eptesicus f, fuscus) is highly susceptible to infection with the
OCT-541 strain of Japanese B encephalitis virus. All animals devel-
oped widespread infection during the first 12 days following virus
inoculation, titers in the blood and brown fat reaching more than
3.0 log units in many instances. Subsequent to the twelfth day, the
infection seemed gradually to subside; only 2 of 10 animals tested
had a detectable viremia by the twenty- fifth day. The susceptibility
of this bat species to the other virus strains has not yet been
determined. •

381

SULKIN AND ALLEN

INFLUENCE OF TEMPERATURE ON
ANTIBODY PRODUCTION IN BATS

In planning studies on the influence of hibernation on the progress
of these experimental viral infections inbats, it seemed desirable to
determine the pattern of antibody production in these animals fol-
lowing peripheral inoculation of various arthropod-borne viruses.As
pointed out earlier, the unique thermoregulatory mechanism of bats
sets them apart from other animals and suggests this animal as an
ideal experimental host whose temperature and metabolism maybe
altered at will, permitting the influence of temperature on infection
and antibody production to be studied over extended periods. The use
of bats also permits the achievement of hypothermia without the use
of drugs or surgery which might have an independent effect on the
physiochemical balance of the host.

A bacterial antigen was selected for a pilot experiment on the
effect of temperature on antibody production because of the ease
with which one candemonstrate agglutinating antibodies. The typhoid-
H antigen- antibody system was chosen as representing a relatively
simple immunization procedure and antibody assay technique re-
quiring very small serum samples, and the big brown bat was used
because it can withstand repeated bleeding by cardiac puncture.
One group of bats was maintained at room temperature (26° C) and
another in simulated hibernation (10° C), Rectal temperatures were
taken periodically. All bats received 3 intraperitoneal doses of 0.25
ml each of an 8- hour formalized broth culture of Salmonella typhosa
on days 0, 15 and 24 of the experiment. Several animals were bled
periodically and all were bled on the 35th day following the initial
immunizing dose. The antibody level of each serum sample was de-
termined by means of the tube agglutination test. The results indi-
cate that the ability of the big brown bat to produce antibody in
response to injections of a bacterial antigen is dependent upon its
body temperature, which parallels that ofthe environment (Fletcher
et al., 1962). After 35 days and 3 doses of antigen, over 70 per cent
of the animals had agglutinin titers ranging from 1:8 to 1:672, In
striking contrast, the bats held at 10° C failed to produce any de-
tectable antibody during this period. Some animals which had been

382

VIRAL INFECTION IN BATS

maintained at room temperature were transferred to the cold room
on the 42nd day, and blood specimens were obtained one month later.
Persistence of antibody was observed in 5 bats, but in each instance
agglutinin titer dropped. Bats kept in simulated hibernation for 42
days were transferred to room temperature and one month later (or
72 days after initiation of the experiment) blood samples were ob-
tained. Although no additional antigen was administered upon trans-
fer to room temperature, 6 of 10 animals uniformly negative after 30
days in the cold possessed agglutinating antibody one month after
transfer to the warm room. Presumably, antigen persisted during
hibernation.

Several studies have been made to determine the immune response
of big brown bats experimentally infected with Japanese B encepha-
litis virus and held at room temperature or in simulated hibernation
for extended periods. Results indicate an extremely close associa-
tion between the cessation of viremia and the appearance of detect-
able neutralizing antibody. Even virus may be detected in low titer in
blood samples containing neutralizing antibody. Also, preliminary
studies indicate that bats experimentally infected with this virus
do not regularly produce antibodies detectable by the complement
fixation (CF) and hemagglutination- inhibition (HI) techniques cur-
rently in use. Experiments designed to study the antibody response
to the hibernating bat and the fate of any neutralizing antibody pre-
sent in the animals entering hibernation indicate that antibody is
produced by bats kept at room temperature, while none is detectable
in those held in the cold. As in the case with the bacterial antigen,
there is evidence that transfer to room temperature stimulates anti-
body syntheses. In general, it has been found that bats which had
equivocal titers(l,0tol.7LNI)^atl7to 21 days were usually strongly
positive about 2 months post- inoculation. Although viremia develops
in all experimentally infected bats, some fail to develop even equi-
vocal titers. As was the case in experiments with the bacterial anti-
gen, none of the bats aroduced neutralizing antibody in response to
Japanese B encephalitis virus during 43 days at 10° C. Similar re-
sults were Obtained with some bats which were inoculated with ra-
bies virus and bled after 43 days in the cold (Sulkin et al., 1960),

2 Log neutralization Index (LNI). AnLNIof < UO was considered negative, 1.0 to 1.7 equi-
vocal, and 1.7 or above positive.

383

SULKIN AND ALLEN

INFLUENCE OF HIBERNATION ON INFECTION AND ANTIBODY

RESPONSE OF BIG BROWN BATS INOCULATED WITH

JAPANESE B ENCEPHALITIS VIRUS

Having established the susceptibility of the big brown bat (Epte-
sicus f. fuscus) to experimental infection with Japanese B encepha-
litis virus and determined the duration of the viremic phase, the
degree of involvement of brown adipose tissue, brain, and kidney,
and learned something of the antibody response of infected animals
held at room temperature, the influence of low temperature on this
infectious process could then be studied. In a study on Japanese B
encephalitis infection in bats during simulated hibernation, LaMotte
(1958) reported that viremia is suppressed in animals placed at
10*^ C immediately after inoculation;upontransfertoroom tempera-
ture after as long as 3 months in the cold, demonstrable levels of
virus appeared in the blood within 2 to 5 days.

In designing our studies, we have attempted to reproduce in the
laboratory situations as they might occur in nature. Whenever
possible, experiments are planned so that the periods in which ani-
mals are held at low temperature in the laboratory correspond to
the season of natural hibernation in an effort to simulate hibernation
rather than merely attain a state of hypothermia (Menaker, 1962).
Groups of bats are placed in cold rooms at varying times following
virus inoculation to represent (1) animals enterir^ hibernation im-
mediately after becoming infected and before infection develops to
a demonstrable level; (2) animals entering hibernation at the peak
of the infectious cycle when virus is present in blood and other tis-
sues; and (3) animals entering hibernation while in the immune phase
of the infection when virus is no longer regularly demonstrable in
blood or other tissues and significant levels of neutralizing antibody
are present. Although these studies are still in progress, we have
analyzed sufficient data to enable us to construct a schematic dia-
gram which we believe to be representative of experimental infection
with this virus in big brown bats under the various environmental
conditions described (Fig. 3),

The uppermost section in Figure 3 depicts infection and antibody

384

VIRAL INFECTION IN BATS

20

30 40 50 60 7C 80 90

DAYS AFTER. /NOCULATlON

Figure 3. Schematic representation indicating suppressive effect of simulated hiber-
nation on progress of infection and immunologic response of bats (Eptesicus f. fuscus)
following intramuscular inoculation of Japanese B encephalitis virus (150 weanling
mouse i. c. LD^q). Shaded areas refer to periods in simulated hibernation (10° C). Solid
bars indicate virus demonstrated by intracerebral inoculation of weanling mice. Clear
bar indicates no virus recovered. Log neutralization index (LNI) indicated by curves.

response in animals held at room temperature from time of inocu-
lation. Some animals become viremic as early as 24 to 48 hours
after receiving 150 LD5Q doses of virus, whereas virus is seldom
isolated from brown adipose tissue before the 3rd or 4th day post-
inoculation. We indicate demonstration of virus in blood and brown
fat over a period of approximately 25 days. It is difficult to define
precisely the duration of active infection, since there appears to be
much individual variation. We do know that virus can be demon-
strated in brown fat and/or blood of virtually all animals tested 7 to
12 days after receiving virus and that subsequently, the number of
infected animals and the concentrations of virus in their tissues
gradually diminish. This decrease in viral proliferation coincides
with the time when neutralizing antibodies first become detectable,
and it is at this point that we have demonstrated on occasion low
concentrations of virus inthebloodof animals with positive neutral-
ization indices. A significant number of animals have positive LNI
30 to 50 days post- inoculation, although some infected animals never
develop antibody detectable by the methods used. Studies on the

385

SULKIN AND ALLEN

persistence of antibody and the occurrence of cyclic episodes of in-
fection in animals held at room temperature are incomplete at this
time.

In the remaining sections of the figure, the shaded areas indicate
the periods following injection of virus when bats are maintained at
low temperature (10 *~^ C) and the influence of these periods of simu-
lated hibernation on the progress of infection has been plotted. When
animals are placed at 10° C immediately after inoculation, viral
multiplication is suppressed. Occasionally, one can demonstrate a
trace of virus in blood or brown fat, but it is difficult to determine
by assay of specimens in mice the fate of the virus inoculum during
this period in the cold. When the animals are transferred to room
temperature, virus reaches demonstrable levels in blood and brown
fat 2 to 3 days later. It appears that the infection which occurs
following a period of suppression in the cold is of shorter duration
than in room temperature groups, and antibody reaches significant
levels more rapidly. On the other hand, bats placed at 10° C one week
after virus inoculation, while at the peak of the infection cycle,
seem to suffer prolonged infection. Virus can be isolated regularly
from brown fat and blood for at least 30 days after the animals are
placed in the cold. Although subsequent data are incomplete, we
have included an instance in which virus was isolated from the brown
adipose tissue of a bat more than 2 months after the animal was
placed in simulated hibernation. The blood of this animal did not
contain demonstrable virus, but companion animals circulated virus
3 days after transfer to room temperature. None of the infected
bats developed antibody while in the cold, but plasma samples with
positive LNI were obtained from bats in this group approximately 2
weeks after transfer to room temperature. In still another series of
experiments mapped in the bottom section of Figure 3 we are
attempting to determine if an animal entering hibernation in an im-
mune state could, upon arousal, circulate virus again spontaneously.
Data along these lines accumulate slowly, since infected bats must
be held at room temperature 30 to 50 days for antibody develop-
ment, then withstand an additional month or more at 10° C and sur-
vive transfer back to room temperature before the end results can
be obtained. At the present time we have some data indicating that
at least some bats with positive LNI, when placed in the cold, lose
their antibody during a period of a month or more in hibernation

386

VIRAL INFECTION IN BATS

and upon transfer to room temperature 3 months after the initiation
of the original infection may experience a second cycle of infection
and antibody response.

DISCUSSION AND SUMMARY

Our primary interest in studyingthe influence of low temperature
on experimental viral infections in bats has been to accumulate in-
formation which would help in understanding how periods of winter
hibernation could affect natural virus infections in these animals.
As might be expected, the depressed metabolic state and the low
body temperature of bats held at 5° C or 10° C are not conducive to
active viral multiplication. Of far greater interest is the observation
that bats in simulated hibernation are able to sustain viral infections
during this period of inactivity. Although virus does not appear to
replicate at a readily measurable rate during this period, viability
is maintained. Upon transfer to a warmer environment simulating
spring arousal, virus lying dormant in brown adipose and/or other
tissues begins to multiply and active infection, demonstrable by con-
ventional assay methods, ensues. This sustenance of viral particles
in the hibernating bat seems more remarkable when accomplished
in animals placed at lowtemperature immediately after virus inocu-
lation. In our studies with rabies virus, animals were inoculated
intramuscularly and placed immediately at 5° C or 10 C, allowing no
time for virus to attach to or penetrate the cells of the animals at
room temperature or the higher temperatures believed to be opti-
mum for this animal virus; yet some bats transferred to a warm
environment even a month or more later were subsequently shown
to develop rabies infection. During the periods inoculated animals
were held at low temperature, only an occasional animal could be
proved infected upon sacrifice, and in all cases rabies virus was
demonstrated in low titer, indicating that virus was not multiplying
very rapidly in these bats. Evidence was obtained, however, in an
experiment with a bat rabies virus strain in little brown Myotis,
that some t)Ape of viral activity occur red during the period the inocu-
lated animals were held at 5° C which resulted in what appears to

387

SULKIN AND ALLEN

be a modification of the virus strain, Sadler and Enright (1959) have
recorded the suppressive effect of low temperature on rabies in-
fection in bats and have correlated it with the lowered metabolic
rate of the animals. These investigators found no evidence of rabies
virus multiplication in animals placed at low temperature immedi-
ately after virus inoculation, but if infection were allowed to develop
for 6 days at 22° C prior to transfer of animals to 4 C, some evi-
dence was obtained of virus multiplication in the brains of animals
which died after 30 to 90 days in the cold.

The experimental data recorded to date on the influence of low
temperature on rabies virus infection in bats suggest that winter
hibernation may have a profound effect on natural rabies infections
in these animals. There is no doubt that the period of winter dor-
mancy could exert a sparing effect perhaps sufficient to maintain a
constant state of latency in a large population. In addition, there is
the suggestion that rabies virus particles which overwinter in the
tissues of the bat and are subject to either complete dormancy or a
shift to a much decreased rate of multiplication may be altered by
this period at below optimum temperature.

Because the big brown bat is highly susceptible to Japanese B en-
cephalitis virus and suffers an infection which can be characterized
as to incubation time, duration of the initial infectious phase, and
antibody response, the format of our low temperature studies with
this virus-host system is more inclusive than has been possible in
work with rabies virus. Assuming that in a given population animals
would vary as to the stage of their natural arbovirus infection, we
are studying the effects of simulated hibernation on groups of ani-
mals which are pre-viremic, viremic or post-viremic, and immune.
The results of these studies allow us to speculate along the following
lines: (1) bats which enter hibernation soon after receiving an in-
fective dose of Japanese B encephalitis virus, but before infection
has developed to demonstrable levels, would be capable of sustaining
infection through months of winter hibernation and upon arousal in
the spring provide infective blood for feeding vectors; (2) bats which
enter hibernation with hightitersof Japanese B encephalitis virus in
blood and tissues apparently suffer a prolonged infection, possibly
due to the suppression of antibody production in the cold; (These
animals could provide infective blood for at least a month for mos-

388

VIRAL INFECTION IN BATS

quitoes present at the hibernating site. In this regard, it is of in-
terest that LaMotte (1958) has shown that mosquitoes will feed on
bats at 10° C. Thus the viremic, hibernating bat could provide a mid-
winter feeding of infected blood for mosquitoes, making possible a
two-step winter transmission chain as suggested by Bellamy, Reeves
and Scrivani (1958) for western equine encephalitis virus in birds
and mosquitoes. In addition, the bat which enters hibernation in a
viremic state, although his blood titer gradually decreases during
the winter, would be capable of circulating virus again upon arousal
in the spring) and (3) bats which enter hibernation at a time when
they possess neutralizing antibodies against Japanese B encephalitis
virus may show negative neutralization indices by the termination
of their winter sleep and the warmer environmental temperature,
together with the physiological alterations which occur at time of
arousal and emergence, could serve to activate latent Japanese
encephalitis virus infection and produce viremic hosts without the
necessity of re- infection.

We feel that these studies demonstrating experimentally the var-
ious conditions under which the hibernating bat can serve to over-
winter an arbovirus add strong supportive evidence to the growing
concept of the role of hibernating animals inthe maintenance of these
agents in nature during periods when vectors are not active. Another
hibernating mammal, the hedgehog, has been shown experimentally
to circulate Russian spring- summer encephalitis virus during
periods in the cold (van Tongeren, 1958) and experimental studies
with western equine encephalitis virus in snakes (Thomas and Ek-
lund, 1960) and Japanese B encephalitis in frogs (Chang, 1958) sug-
gest that these poikLlothermic animals may be capable of overwin-
tering these viruses.

In addition to the broad epidemiological aspects of the studies
presented here, the compiled data also provide information of a
more fundamental nature relative to the influence of temperature on
the virus-cell-host relationship per se. There is no doubt that the
multiplication rate of both rabies and Japanese B encephalitis
viruses is suppressed in the bat by maintaining the animal at low
temperature. The manner in which infection with both these agents
is sustained for long periods of time in the dormant animal, however,
suggests that some type of viral activity occurs in this host at low

389

SULKIN AND ALLEN

temperature. The apparent alteration in the bat rabies virus strain
which occurred as a result of a period of "incubation" in bats at
5° C further supports this concept. There is also indication that the
Japanese B encephalitis virus may be altered by passage through
bats at 10° C. The virus demonstrable inthe blood and other tissues
of these animals upon transfer to 24° C often produces a disease in
mice characterized by increased incubation time, bizarre paralytic
symptoms and frequently, recovery. In experiments now inprogress
we are studying strains of viruses following single and multiple
passages through bats held at low temperature. These in vivo studies
are being paralleled with in vitro experiments using monolayer and
explant cultures of bat brown fat, kidney and embryonic tissue, as
well as preparations of tissues from warm-blooded animals in an
effort to produce virus strains altered by passage at low tempera-
ture in different host systems. A study of virus strains altered by
the pressures of low temperature exerted inthe intact bat, cultures
of bat tissues, and cultures of tissues from warm-blooded animals
should provide devinitive information concerningthe mechanisms of
temperature-induced variations in animal viruses.

LITERATURE CITED

1. Bell, J. F., and G. J. Moore. 1960. Rabies virus isolated from

brown fat of naturally infected bats. Proc. Soc. Exp. Biol. Med.
103: 140-142.

2. Bellamy, R. E.,W. C.Reeves, and R. P. Scrivani. 19 58. Relation-

ships of mosquito vectors to winter survival of encephalitis
viruses. II. Under experimental conditions. Am. J. Hyg. 67:
90-100.

3, Chang, I-C. 1958. Studies on Japanese B encephalitis in cold-

blooded animals. Pediatrics 4: 27-49.

4, Corristan, E., L. LaMotte, Jr., and D. G. Smith. 1956, Suscep-

tibility ofbats to certain encephalitis viruses. Fed.Proc. 15:584.

390

VIRAL INFECTION IN BATS

5. Enright, J. B. 1956. Bats and their relation to rabies, Ann, Rev,

Microbiol. 10: 369-392.

6. Fawcett, D, W. 1952, A comparison of the histological organiza-

tion and cytochemical reactions of brown and white adipose
tissues. J. Morphol. 90: 363-405.

7. Fletcher, Mary Ann, Ruth Sims, Rae Allen, and S. Edward Sulkin.

1962, Influence of temperature on antibody production in the
bat, Texas Rep. BioU Med, 20: 142,

8. Griffin, Donald R. 1958, Listening in the Dark. Yale Univ, Press,

9. Grodums, Irene, and George Dempster. 1959. The age factor in

experimental Coxsackie B-3 infection, Canad. J. Microbiol. 5:
595-604.

10. Hock, R. J. 1951. The metabolic rates and body temperatures of

bats. Biol. Bull. 101: 289-299,

11. Hock, R, J. 1958, Hibernation, p. 61-133. In: Cold injury. Trans,

5th Conf, Josiah Macy Found,

12. Hoggan, M, D,, and B.Roizman. 1959. The effect of the tempera-

ture of incubation on the formation and release of herpes
simplex virus in infected FL cells. Virology 8: 508,

13. Ito, T., and S. Saito. 1952. Susceptibility of bats to Japanese B

encephalitis virus. Jap. J, Bact, 7: 617-622,

14. Johansson, B, 1959, Brown fat: a review. Metabolism 8: 221-240.

15. Johansson, Bengt, 1960, Brown fat and its possible significance

for hibernation, p, 233-248, In mammalian hibernation, Lyman
and Dawe, eds, Cambridge, Mass.

16. Kayser, C, 1961,Thephysiology of natural hibernation, Pergamon

Press, New York, v 8.,

391

SULKIN AND ALLEN

17. LaMotte, L. C, Jr. 1958. Japanese B encephalitis in bats during

simulated hibernation. Am. J. Hyg. 67: 101-108.

18. Lwoff, A. 1959. Factors influencing the evoluation of viral dis-

eases at the cellualr level and in the organism.Bact.Rev. 23:
109-124.

19. Menaker, Michael. 1962. Hibernation- Hypothermia: An annual

cycle of response to low temperature in the bat Myotis lucifu-
gus. J. Cell. Comp. Physiol. 59: 163-173.

20. Morrison, P. 19 59. Body temperatures in some Australian mam-

mals. I. Chiroptera. Biol. Bull. 116: 484-497.

21. Pappenheimer, A. M., J. B. Daniels, F. S. Cheever, and T. H.

Weller. 1950, Lesions caused in suckling mice by certain vi-
ruses isolated from cases of so-called non- paralytic polio-
myelitis and of pleurodynia. J. Exp. Med. 92: 169-190.

22. Rasmussen, A. T. 1923. The so-called hibernating gland. J.

Morphol. 38: 147-205.

23. Remillard, G. L. 1958. Histochemical and microchemical obser-

vations on the lipids of the interscapular brown fat of the fe-
male vespirtilionid bat Myotis lucifugus.Ann.N.Y.Acad.Sci.
72: 1-68.

24. Sadler, W. W., and J. B. Enright. 1959. Effect of metabolic level

of the host upon the pathogenesis of rabies in the bat. J. Infect.
Dis. 105: 267-273.

25. Shwartzman, G. 1952. Poliomyelitis infection in cortison-

treated hamsters induced by the intraperitoneal route. Proc,
Soc. Exp. Biol. Med. 79: 573-576.

26. Sims, Ruth, Rae Allen, and S. E. Sulkin. 1963. Studies on the

pathogenesis of rabies in insectivorous bats. III. Influence of
the gravid state. Unpublished data.

392

VIRAL INFECTION IN BATS

27. Sulkin, S. E. 1945. The effect of environmental temperature on

experimental influenza in mice. J. Immunol. 51: 291-300.

28. Sulkin, S. E., and M. J. Gr eve. 19 54. Human rabies caused by bat

bits. Texas State J. Med. 50: 620-621.

29. Sulkin, S. E., P. H. Krutzsch, R. Allen, and C. Wallis. 1959.

Studies on the pathogenesis of rabies in insectivorous bats. I.
Role of brown adipose tissue. J. Exp. Med. 110:369-388.

30. Sulkin, S. E., Rae Allen, Ruth A. Sims, P. H. Krutzsch, and C.

Kim. 1960. Studies on the pathogenesis of rabies in insectivor-
ous bats. II. Influence of environmental temperature. J. Exp,
Med. 112: 595-617.

31. Sulkin, S. E., R. Allen, R. Sims. 1960, Lipotropism in pathogen-

esis of encephalitis viruses in insectivorous bats. Virology 11:
302-306.

32. Sulkin, S. E, 1962. The bat as a reservoir of viruses in nature.

v.4.In Progress in Medical Virology.Karger/Basel, New York,

33. Sulkin, S. E., Rae Allen, and Ruth Sims. Studies of arthropod-

borne virus infections in Chiroptera. I, Susceptibility of in-
sectivorous species to experimental infection with Japanese B
and St. Louis encephalitis viruses. Unpublished data,

34. ThomaSj Leo A., and Carl M, Ecklund. 1960. Overwintering of

western equine encephalomyelitis virus in experimentally in-
fected garter snakes and transmission to mosquitoes. Proc,
Soc. Exp, Biol. Med. 105: 52-55.

35. Tongeren, H. A, E. van. 19 58, Sixth International Congress on

Tropical Medicine and Malaria. (Abstr.) Lisbon, 5-13 Sep-
tember p. 166.

36. Walker, D. L., and W, D.Boring.l958. Factors influencing host-

virus interactions. Ill, Further studies on the alteration of Cox-
sackie virus infection inadult mice by environmental tempera-
ture. J. Immunol, 80: 39-44.

393

SULKIN.AND ALLEN
DISCUSSION

CAMPBELL: Did you do electrophoretic patterns of the hiber-
nating bats ?

SULKIN: We are doing them now, Dr. Campbell, and we are
finding things that I haven't yet had time to sit down and figure
out, but we are seeing some very strange things in these elec-
trophoretic patterns. I really don't know what to make of it.

ANDREWES: I understand that viremia is absent in hiber-
nating snakes and then comes back again, as in your bats. When
they come out of hibernation, is there any evidence in your bats
as to whether this is a function of temperature of not; or hasn't
that been done ?

SULKIN: I don't think we have enough data. Perhaps Dr. Marcus
knows. I don't think so.

SCHMIDT: Many times you used the term "simulated hiber-
nation". Would you care to define it?

SULKIN: I suppose I could delete the word "simulated", but
I think that when you take a bat away from his natural environ-
ment and then try to simulate these circumstances in the lab-
oratory, there is a difference. I think that when we net the bats
in the fall at the time when they ordinarily would be going into
hibernation, we are dealing with true hibernation. This is not
the case when such experiments are carried out with bats netted
in the summer. Such animals, when placed at low temperatures,
are hypothermic.^

WALKER: You don't see any difference in your experiments,
though, in these two seasons ?

1 Menaker, J. 1962. Cell and Comp. Physiol. 59: 163-173.

394

VIRAL INFECTION IN BATS

SULKIN: Well, there are many differences in the same bat
species, Eptesicus fuscus. You can do an experiment of this
sort. You can collect bats in the fall of the year and use thermo-
couples and record body temperature of these bats. Now, if you
take a bat just prior to the time he is goint into hibernation,
and then place him at about 8° C to 10° C in the laboratory,
then record temperatures over a period of time, rectal tempera-
ture will march along at a very steady state. If you stimulate
this bat with a pair of forceps, the temperature goes up, and
goes up very promptly. If you take the same species, Eptesicus
fuscus, the big brown bat, and you do the same experiment,
but do it in July instead using the same probe and the same
amount of insertion into the rectum and so on, everything iden-
tically the same, then you can stimulate this bat and nothing
happens. We have done this with many bats. So this is one sig-
nificant difference between the true hibernating bat and the simu-
lated hibernating bat.

TRAPANI: Don't they wake up during the winter time for
feeding ?

SULKIN: No, they stay in the cage,

TRAPANI: In their natural habitat?

SULKIN: In their natural habitat they don't feed; they store
food in their brown fat, presumably, and this is enough to keep
them going. One can do this with hamsters, too, but it's a trick.

BERRY: In Texas, doesn't the temperature get high enough
to bring them out of the hibernating state?

SULKIN: Most of the bats in Texas are not true hibernating
species, but migratory species. During mid-winter they migrate
south to an area where the temperature is optimum.

BERRY: But bats are in the caves in Texas in the winter
time.

SULKIN: Yes, many Mexican free- tailed bats can be found in

39 5

SULKIN AND ALLEN

certain of the caves during the winter time.

BERRY: But they can't be hibernating all the time, can they?

SULKIN: Most of the bats in Texas, and there are millions
and millions of Mexican free-tails there, are not of a true hi-
bernating species. They go out for night time feeding. Con-
sistent with that is the fact that they have very little brown fat
in the interscapular region, so it's very difficult to do experi-
ments with this particular species,

BERRY: Just out of curiosity, how do you feed them?

SULKIN: The big brown bats (Eptesicus U fuscus) are hand-
fed meal worms the first week we get them into the laboratory.
Most of them learn to feed in about 2 to 3 days. Then we put
them on a concoction that we have prepared which consists
of cottage cheese, banana, and meal worms, to which is added
liver extract with iron and multivitamins.

BERRY: They won't breed?

SULKIN: No. They don't breed under laboratory conditions,
but many of the gravid bats have delivered their young in the
laboratory.

SCHMIDT: Concerning true hibernation or simulated hiber-
nation in the same species, I believe you indicated that if you
induce this simulated hibernation during the summer months,
this animal is then under stress and not truly hibernatir^. I
am wondering if he is burning excessive amounts of body tis-
sue of some sort to accommodate himself. Have you made any
measurements of weight loss comparing these two periods of
time?

SULKIN: No, we haven't done any good experiments along this
line. We are so convinced that this is a different host when
induced to hibernate during the summertime that we do these
experiments on a yearly basis in the late fall.

396

VIRAL INFECTION IN BATS
CAMPBELL: Have you ever injected the brown fat?

SULKIN: Yes, into a non- hibernating animal. The experiments
don't work too well. Two sets of experiments have been re-
ported previously, and conflicting results were obtained.2»3

BLAIR: Bigelow has done this in Toronto.

SULKIN: These are extracts, and they are not very pure.

BLAIR: He transplanted the brown fat and the results were
entirely negative.

SULKIN: Nobody has as yet been able to define the active
component in brown fat tissue that is likely to be related to
hibernation. And I think that there is a growing interest in this
tissue as the true mechanism of hibernation.

TRAPANI: One thing we tried which seemed to indicate some
activity of brown fat was this: We made a saline extract out of
brown fat obtained from the Guinea pigs exposed to 2° G for
about 10 weeks. Ordinarily, you put Guinea pigs into -15° C
and they don't do very well; they die quickly. However, when
saline extract of brown fat was injected into animals kept at
-15° C, they survived another day or so. We never tried it again
because of the lack of time.

SULKIN: I think that if you tried it again, it might not work.
When saline extracts of brown fat were used by previous in-
vestigators, they yielded inconsistant results.

ANDRE WES: We have been hearing quite a lot in the last two
days on the subject of cold and other stressing factors. There
is one form of stress that has never been mentioned. There
were some experiments done a good many years ago on pneu-
mococcal infections in partly- immunized mice when it was shown

2 Zirm. 1956. Zschr. f Naturforsch. 11: 530-535.

3 Kross. 1933. Zschr. f. d. ges. Neurol. Psych. 146: 208-218.

397

SULKIN AND ALLEN

that these infections could be activated by the stress of alco-
hol. This is a subject which needs further investigation.

398

MICROBIOLOGICAL ASPECTS OF HIBERNATION
IN GROUND SQUIRRELS

J. Schmidt

Department of Bacteriology

University of New Hampshire

Durham, New Hampshire

ABSTKAGT

Studies were designed to determine the effect of reduced temperatures on the normal
bacterial and viral flora of the experimental animals and on the fate of other micro-
organisms artificially introduced. Determinations were made of the number of coliform
bacilli, fecal streptococci, and psychrophilic organisms present in the fecal material
before and immediately after hibernation. In addition, the total viable aerobic cell
count was determined. All counts were based on the number of orgamsms per gram
(dry weight) of fecal materal. The data indicate a gradual 3 log increase in the number
of psychrophiles and a simultaneous decrease of similar magnitude in the number of
coliform bacilli during periods of hibernation. No change which could be associated with
hibernation was noted in either the total cell count or the number of fecal streptococci.
The animals were shown to be free of viruses capable of inducing a cytopathogenic
effect in monolayers of either HeLa, human amnion, or monkey kidney cells. Fecal
samples from each of 32 animals were repeatedly tested over a period of four months.
Bacterial viruses (Escherichia coli phage) have been demonstrated in only one of the
40 ground squirrels examined. Several strains of E. coli isolated from the animals were
used as the host cell in the test systems. The effect of hibernation on the presence of
artificially introduced ECHO-6 and Coxsackie B-3 viruses and E. coli phage in the in-
testinal tract of ground squirrels is being evaluated.

Hibernating mammals provide an excellent means for studying
the effect of cold on the biology of experimental infection from the
standpoint of boththehostand the parasite. These animals represent
a unique situation in which a normal system is available in the same
species, at widely separated temperatures. For instance, in the ac-
tive state a ground squirrel has a normal body temperature of ap-
proximately 37° C, whereas in hibernation its normal temperature
may approach 0° C.

The studies reported here deal with the effect of hibernation on

399

SCHMIDT

certain components of the normal bacterial and viral flora and on the
retention of specific viral agents in the intestinal tract of ground
squirrels.

Since hibernation is such an important feature of these investi-
gations, a few words aboutsomeof its salient characteristics are in
order. First of all, it is important to recognize that hibernation is a
natural phenomenon for certain animals and that it is quite different
from experimental hypothermia. Hibernation is a physiologically
controlled and regulated process, while experimental hypothermia is
a result of weakened or overcome normal mechanisms of tempera-
ture regulation. Chilling occurs in enforced hypothermia in spite of
a concerted effort on the part of the animal to maintain its normal
temperature. On the contrary, hibernation is a passive, yet deliber-
ate process in which declines in metabolic rate, respiratory rate,
and heart rate precede the drop in body temperature.

Within a limited range, the body temperature of the hibernating
ground squirrel parallels that of the environment, usually remaining
0,5° C to 3° C above the ambient temperature (Johnson, 1931), Mayer
(1960) reported an average rectal temperature of 4.2*^ C for ground
squirrels hibernating in an environmental temperature of 1.6° C,
The heart rate and respiratory rate are slowed to about 3 per min,
and the metabolic rate is between 1/30 and l/lOO of that found in the
active "resting" animal. Vascular changes include a decrease in
blood pressure as the animal enters hibernation followed by vaso-
constriction once the torpid state has been achieved (Lyman and
O'Brien, 1960), There appears to be a pronounced leucopenia during
hibernation (Svihla, 1953) which has notbeen explained. Although the
rate is very slow. Brock (1960) reports that there is actually some
manufacture of red blood cells during deep hibernation.

There are indications that some of the endocrines play a role in
hibernation, although the specific data are conflicting. Most workers
agree that an involution of the endocrines takes place before the
animal enters hibernation. Indeed, Kayser (1955) insists that hiber-
nation will not take place without this involution.

Although hibernation is partof a yearly cycle of events for ground
squirrels, it is not clearly understood what specific factors are im-

400

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

portant in determining when they will enter or emerge from the
hibernating state. It would appear that the animal receives few, if
any, clues from its environment, such as photoperiodicity or tem-
perature (Pengelley and Fischer, 1957).

Ground squirrels and other mammalian hibernators undergo
natural awakenings from time to time during the hibernating period
(Lyman and Chatfield, 19 55), Kays er (1960) reported that the average
period of continuous hibernation in ground squirrels is about 21 days.
In our work we have been somewhat less successful. The cause of
the periodic arousals inundisturbed animals has yet to be explained.

For a more complete survey of the physiology of mammalian hi-
bernation, the reader is referred to the recent publication edited by
Lyman and Dawe (1960).

MATERIALS AND METHODS

Arctic ground squirrels (Spermophilus undulatus), adult animals
of both sexes, were captured in the Paxson Lake area of central
Alaska during the month of August. They were housed individually in
metal cages equipped with removable wire mesh floors and catch
pans. Their diet consisted of apples, carrots, and Labena pellets
(Ralston Purina Co., St. Louis, Mo.). They were observed in the
colony for 2 months before the studies were initiated. Hibernation
was induced by placing the animals in a cold room at an ambient
temperature of 3° C ± 1° C for the bacterial work and 5° C ± 1° C
for the viral studies. These temperatures were maintained through-
out each ofthe respective investigations. The animals were observed
daily but not disturbed unnecessarily, because various stimuli may
trigger an arousal from the hibernating state.

The primary explants of monkey kidney cells were obtained from
Shamrock Farms (Middletown, New York) and maintained on Eagles
basal medium with Hank's balanced salt solution (BSS), HeLa mono-
layers were prepared from the S-3 clone, kindly supplied by

401

SCHMIDT

Dr. John L. Riggs of the University of Michigan, using 90 per cent
Eagles BM and 10 per cent calf serum as the growth medium. For
maintenance, the calf serum was reduced to 5 per cent. The same
growth and maintenance media were used for the cultures of the
RA strain of human amnion cells.

The specific viral agents used were the D'Amori strain of ECHO-
6 virus and the Nancy nicKo strain of Coxsackie B-3,

Bacterial Studies

Investigations were made of the normal intestinal flora of 15 adult
ground squirrels both in the active state and foUcwing periods of
deep hibernaticn. Fresh fecal droppings were collected on filter pa-
per, and those contaminated with voided urine were excluded. Speci-
mens were kept at 4° C until processing, which took place within 2
hours after defecation. The droppings from an individual animal
were weighed and combined with an equal weight of sterile distilled
water. From this a uniform suspension was made with the aid of a
sterile glass rod. Part of the material was used for the determina-
tion of the per cent moisture and another portion for the preparation
of cultures. Triplicate cultures were inoculated from serial 10-fold
dilutions (10" to 10~ ) of the specimens, and the results obtained
were calculated as the number of organisms per gram dry weight of
fecal material. Quantitative determinations were made of the con-
form bacilli, fecal streptococci, and psychrophilic organisms, in
addition to a total viable aerobic cell count. The coliform bacilli
were cultured on (Difco) eosin methylene blue agar at 37° C. The
plates were seeded with 0.1 ml of the appropriate dilution which was
then spread evenly over the surface of the agar with a bent glass
rod. The inoculum was dispensed with a sterile 0.1 ml pipette. Using
this technique, well-isolated colonies were consistently obtained.
All gram negative rods yielding lactose- positive colonies within 48
hours were arbitrarily considered to be coliform bacilli.

For the enumeration of the psychrophiles, (Difco) tryptone glucose
extract (TGE) agar was employed. Pour plates were prepared using
an inoculum of 1 ml and the incubation was carried out at 2 C to
3° C for 14 days. Organisms producing a visible colony within 14

402

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

days were termed psychrophiles. In addition, pour plates were made
in TGE agar as above, except that these cultures were incubated at
37° C for 48 hours. The results from these latter cultures were re-
ferred to as the total count at 37° C. The number of fecal strepto-
cocci was determined using (Difco) azide dextrose agar with an in-
oculum of 1 ml and 37° C incubation temperature. All plate counts
were made using a Quebec colony counter and only those plates
having between 30 and 300 colonies were included.

Bacterial Virus Study

In order to gain some information concerning the normal occur-
rence of bacterial viruses in the animals, 44 adult ground squirrels
were examined repeatedly (154 individual samples) for the presence
of bacteriophage in their intestinal contents. The bacterial host cells
used were strains of Escherichia coli previously isolated from the
animals in the group.

The method used for the detection of phage was as follows. Fresh
fecal material was collected in a sterile tarred 50 ml centrifuge
tube and the weight of the sample determined, A one to 10 dilution
(weight in volume) was prepared in sterile nutrientbroth and a uni-
form suspension made with the aid of a sterile glass rod. Following
centrifugation at 17,000 X g for 30 minat 5° C the clear supernatant
fluid was withdrawn and tested for the presence of virus by two
methods. In the first, 1 ml of the fluid and 1 ml of a 3 hr broth cul-
ture of the strain of E, coli being tested was added to 5 ml of tryp-
ticase soy broth. This suspension was incubated overnight at 37 C
and the supernatant fluid therefrom was assayed for the presence of
virus according to the following procedure. Serial 10-fold dilutions
(10"^ to 10~ ) of the fluid were prepared in sterile nutrient broth and
0.1 ml added to a tube containing 2 ml of melted (43° C) trypticase
soy agar (0.7 per cent) and 0.1 ml of an 18 hour broth culture of the
test strain of E, coli. This was carefully mixed and poured over the
surface of a layer of nutrient agar in a petri plate. These cultures
were incubated at 37° C and examined for the presence of virus as
indicated by the formation of plaques. In the second method the pre-
liminary overnight incubation was eliminated, and the original super-
natant fluid was tested for bacterial virus by the plaque method.

403

SCHMIDT

These rather serious attempts to isolate phage disclosed that only
one ground squirrel out of the 44 tested was excreting a bacterial
virus reactive against any of the host strains used. This virus (E.
coli phage 42) would effectively lyse our strain 14 of E. coli. There-
fore, we had available a working system which had been isolated
from the colony, and 43 animals which were naturally free of the
agents involved. Accordingly, a study was made of the effect of hi-
bernation on the retention of this bacterial virus following artificial
administration in the intestinal tract of the experimental animals.

Ten squirrels were used in this investigation. Of these, 5 were in
deep hibernation and 5 were active but had been in the cold room for
the same length of time as those in the hibernating state. The tem-
perature in the hibernaculum was 5^ C, as mentioned earlier. Each
animal was given a dose of 1 ml containing 29 x 10^ plaque forming
units of virus. The virus suspension was cooled to 5° C and intro-
duced into the stomach by means of a 2 ml syringe fitted with a 2 in,
length of polyethylene tubing. The tubing was easily directed down
the throat of the animals while they were hibernating. Ether anes-
thesia was used in the case of the active animals.

Following administration of the virus, quantitative determinations
were made for excreted virus by collecting fresh fecal material and
examining it by the direct plaque method described above. After each
specimen was obtained, the animal was transferred to a clean cage,
in an effort to preclude contamination of subsequent specimens.

Enteric Virus Study

Following these investigations, attempts were made to determine
the natural occurrence of enteric viruses in the animals. Thirty-
two squirrels were examined repeatedly over a period of 4 months
for the presence of enteric viruses in their intestinal contents. It
was surprising (justified or not) to find that all were free of viral
agents capable of inducing a cytopathogenic effect in monolayers of
either HeLa, human amnion, or monkey kidney cells. Therefore, we
initiated studies to determine the effect of hibernation on the reten-
tion of specific enteric viruses in the intestinal tract of these ani-
mals following artificial administration. Preliminary tests had indi-

404

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

cated that monkey kidney cells were susceptible to ECHO- 6 virus
and that Coxsackie B-3 virus could be isolated and propagated in
monolayers of HeLa cells. In addition, both viruses could be suc-
cessfully recovered from the fecal material of animals which had
been fed a suspension of the specific agents.

Accordingly, 9 squirrels were given a suspension of 1 x 10^ ECHO-
6 virus particles and 8 others were fed 1 x lo''' particles of Cox-
sackie B-3 virus, using the method of administration described
above. Approximately half of the animals in each group were hiber-
nating and half were in the active state; all had been in the cold
room for the same period of time. Thereafter, fresh fecal samples
were collected at appropriate intervals and tested for the presence
of virus using monolayers of monkey kidney and HeLa cells. Serial
10-fold dilutions (10~^ to 10"^) of the fecal material were prepared
in Earles BSS and a duplicate series of tubes was inoculated with
0.1 ml of the proper dilutions. Incubation was carried out at 37° C,
and the cultures were examined daily for evidence of a cytopatho-
genic effect.

RESULTS AND DISCUSSION

The data indicated a significant increase in the number of psych-
rophiles and a simultaneous decrease in the number of coliform
bacilli following periods of hibernation. On the other hand, very
little fluctuation was observed in either the total counts at 37° C
or the numbers of fecal streptococci, and these did not appear to
be influenced in any way by hibernation. The values shown repre-
sent the averages obtained with the 15 ground squirrels studies. Al-
though there was some variation in the hibernating habits, activity,
food consumption, and normal intestinal flora of the individual ani-
mals, each of the squirrels tested demonstrated these character-
istic trends.

A representative number of the coliform bacilli and psychro-
philes were isolated and their temperature-growth relations deter-

405

SCHMIDT

p 10

X

o

UJ

^ 9

>-

(T
O

UJ

o

7-

? 6-

E
o

<

UJ

i

3

5 -

.

• TOTAL COUNT 37 'C

. , • •

"

0«iiiMm»""W ^ ""•••••fni,,^

-

^^^K^^ PSYCHROPHILIC BACTERIA

s

f

/

^^ COLIFORM BACILLI

1 1 1 1 1 1 ^ ,

4-8 9-13 14-18 19-23 24-28 29-33 34-38 39-43

DAYS AFTER ONSET OF HIBERNATION

Fitnare 1. Tenipei.ilure-grovsth relations Ijelween a reprL'senlative number (j1 psychro-
philic baclena ami culifurni IjaciUi.

mined. None of the coliform bacilli tested grew at temperatures
below 5° C. This would imply that little or no multiplication of these
organisms occurred in the intestinal tract of the animals while they
were in the hibernating state and it would be consistent with the
decrease in the counts as shown in Figure 1. The psychrophiles, on
the other had, grew well when incubated at temperatures down to
0° C. Although Ingraham and Stokes (19 59) report that some psychro-
philic bacteria have a maximum growth temperature above 37° C,
none of the psychrophiles isolated in our study gave evidence of
growth at temperatures above 35° C, During this study the average
duration of continuous hibernation was 6 days with extremes of 2 and
19 days. Between periods of hibernation the animals were active for
an average of 1.3 days, indicating that they were in hibernation ap-
proximately 82 per cent of the time. This figure must be considered
conservative, because it takes an active animal from 8 hours to
several days to become completely dormant, and the waking process
requires about 3 hours (Lyman and Chatfield, 1955).

406

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

o —

14 19 26 30 47 2 8 17 18

ARCTIC GROUND SQUIRRELS GIVEN PHAGE 42

Figure 2. A/clic ground squirrels given phage 42.

The results of the bacterial virus study are presented in Fig-
ure 2. In can be seen that the virus was rapidly eliminated from
the intestinal tract of the active animals, but was retained for
a considerable period of time by the animals which hibernated.
It should be noted that each of the hibernating individuals was in
an aroused, or active state for periods of time more than that
required for complete elimination of the virus in the non-hiber-
nating squirrels. This might be explained by the fact that when
the animals come out of hibernation for these short periods,
they usually do not eat, even though food may be available. This
would, of course, result in a less rapid turn- over of the intestinal
contents and thus permit a retention of the virus. In any case,
it is not suggested that the extended period of virus retention
observed in the hibernating animals was the result of any spe-
cific effect, but rather a reflection of the fact that under these
conditions it simply takes longer for material to pass through
the gastrointestinal tract of the animal. It is likely that the phage
is handled as an inert particle, so far as the animal is concerned.

407

SCHMIDT

20

^ Hibernating

■♦- Specimen - Positive
for VIRUS

- Specimen - Negotive
for VIRUS

I 10 22 5! 13 17 23 24 45

ARCTIC GROUND SQUIRRELS GIVEN ECHO-6 VIRUS

Figure 3. Arctic ground squirrels given Echo-6 virus.

The results of the ECHO-6 studies are shown in Figure 3.
As in the case of the bacteriophage, it can be seen that the virus
soon disappeared from the intestinal tract of the active animals
but could be recovered from the hibernators up to 18 days after
administration. Two of the squirrels were still excreting virus
when the last sample was collected.

Essentially the same effect was observed in the case of the
Coxsackie B-3 virus, as shown in Figure 4. The period of re-
tention in the active animals was somewhat longer than that ob-
served for the bacteriophage and ECHO-6 virus, but these dif-
ferences may be more apparent than real.

As to the fate of the virus introduced, none of our data gave
any evidence that viral replication had occurred. Indeed, we re-
covered considerably less virus than we administered.

Dempster et al. (1961) report that Coxsackie B-3 virus is in-

408

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

12 14 18 21 9 16 26 30

ARCTIC GROUND SQUIRRELS GIVEN COXSACKIE B-3 VIRUS

Figure •*. Arctic ground squirrels given Coxsackie B-3 virus.

fectious for ground squirrels in both the hibernating and non-
hibernating state. They were able to recover virus up to 50 days,
after subcutaneous inoculation, from the brown fat, brain, and
heart tissue of animals which had hibernated. On the other hand,
their active animals were positive for virus at 4 but not at 14
days after a similar inoculation. They make no mention of the
squirrels arousing from hibernation following the injections nor
subsequently during the study period and would seem to infer that
they observed continuous hibernation for as many as 50 days.
Based on our own experience and the published reports of others
(Lyman and Chatfieli, 1955; Kayser, 1960), periods of continuous
hibernation beyond 30 days are unusual; the record, to my knowledge,
has been 40 days (Kayser, 1960), We have not as yet examined
any tissues from our animals for the presence of virus but expect
to do so in some future studies. It is my belief that active virus
could remain in the animal so long as hibernation continued. Such
findings would, of course, have implications with respect to the
possibility of overwintering of enteric and other viral agents.

409

SCHMIDT

The resistance of hibernating animals to infection has been in-
vestigated by Kalabukhov (1958) and others. The available evidence
indicates that resistance to infection is increased duringthe period
of preparation for hibernation as well as during hibernation. There
is general agreement that the enhanced resistance is simply a re-
flection of the physiological state of the animal. Since the infectious
agents tested are not capable of reproduction at temperatures found
in the hibernating host, one would not expect to find evidence of an
active infection under such conditions. At the same time, one might
expect very little phagocytic activity at such reduced temperatures,
and it is known that a leucopenia occurs during hibernation (Svihla,
19 53). These observations would suggest that increased resistance
in the hibernating state is probably not related to the cellular mech-
anisms of resistance.

Adequate data are not available concerning the humoral resistance
of hibernators, Jaroslow and Smith (1961) have studied the disappear-
ance of antigen from the circulatory system of hibernating ground
squirrels. Their data indicate that there is little or no disappearance
of homologus or heterologus proteins from the circulation during
periods of hibernation. Studies concerning the actual production of
antibody during hibernation are lacking.

In our bacterial studies we have shown that the psychrophiles
grow well at temperatures downtoO°C, and that they increase con-
siderably in number in the intestinal tract during hibernation. Most
of the psychrophilic organisms we have encountered have been gram
negative rods and producers of endotoxin. It is conceivable that with
the animal's normal mechanisms of resistance either inoperative or
functioning with reduced efficiency,these organisms might find their
way out of the intestinal tract and into other parts of the body such
as the blood stream. It is tempting for me to speciilate that one of
Nature's reasons for programming those unexplained spontaneous
arousals, which are accomplished at great expense of energy to the
animal, might be to clear the circulation of these and other invading
toxic elements. One of the things we plan to do in the future is
to check the blood of the hibernating squirrels for psychrophiles
and other endogenous microorganisms.

410

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

SUMMARY

The effect of reduced temperatures (associated with hibernation)
on the normal bacterial flora and on the fate of specific viral agents
which were artificially introduced into ground squirrels has been
studied. The data indicate that a gradual increase in the number of
psychrophiles and a simultaneous decrease in the number of con-
form bacilli occurs during hibernation. No change, which could be
associated with hibernation, was noted in either the total cell count
or the number of fecal streptococci.

The viral studies demonstrated that hibernation appears to ex-
tend the period of time during which the specific agents tested could
be recovered from animals following artificial administration.

LITERATURE CITED

1. Brock, Mary Ann. 1960. Production and life span of erythrocytes

during hibernation in the golden hamster. Am. J. Physiol.
198: 1181-1186.

2. Dempster, G., E. Irene Grodums, and W. A. Spencer. 1961. Ex-

perimental Coxsackie B-3 infection in the hibernating ground
squirrel and bat. Canad. J. Microbiol, 7: 587-594,

3. Ingraham, J. L., and J. L. Stokes. 1959. Psychrophilic bacteria.

Bacterid. Rev. 23: 97-108.

4. Jaroslow, B. N., and D. E.Smith. 1961. Antigen disappearance in

hibernating ground squirrels. Science 134: 734-735.

5. Johnson, G. E, 1931. Hibernation in mammals. Quart. Rev. Biol.

6: 439-461.

411

SCHMIDT

6. Kalabukhov, N. I. 1958. Characteristics of heat regulation in

rodents as one of the factors in their sensitivity to plague in-
fection. Zh. Mikrobiol. Epidemiol. Immunobiol. 29:1453-1460.

7. Kayser, C. 1955. Hibernation et hibernation artificielle. Rev.

Path. Gen. Comp. 668: 704-728.

8. Kayser, C. 1960. Mammalian hibernation. I. Hibernation versus

hypothermia. Bull. Museum Comp. Zool. Harvard Univ. 124:
9-30.

9. Lyman, C. P., and P. O. Chatfield. 1955. Physiology of hiber-

nating mammals. PhysioU Rev. 35: 403-425.

10. Lyman, C. P., and A. R. Dawe. 1960. Mammalian hibernation.

Bull. Museum Comp. Zool. Harvard Univ. 124: 549p.

11. Lyman, C. P., and Regina C. O'Brien. 1960. Mammalian hiber-

nation. XVIII. Circulatory changes inthe thirteen- lined ground
squirrel during the hibernating cycle. Bull. Museum Comp.
Zool. Harvard Univ. 124: 353-372.

12. Mayer, W. V. 1960. Mammalian hibernation. VII. Histological

changes during the hibernating cycle in the arctic ground
squirrel. Bull, Museum Comp, Zool, Harvard Univ. 124:131-154,

13. Pengelley, E. T., and K. C. Fisher, 1957, Onset and cessation of

hibernation under constant temperature and light in the golden-
mantled ground squirrel, Citellus lateralis. Nature 180: 1371-
1372.

14. Svihla, A., H. Bowman, and Ruth Ritenour. 1953. Stimuli and

their effects on awakening of dormant ground squirrels. Am.
J. Physiol. 172: 681-683.

412

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

DISCUSSION

REINHARD: How much fecal material and ingesta remains
in the gastrointestinal tract of these hibernating animals?

SCHMIDT: Generally the stomach is empty. This is follow-
ing prolonged periods, and as I mentioned, they do not eat,
usually, during the periods of arousal. This has been investi-
gated by certain workers from the histological standpoint, and
so on. Their stomachs are usually empty, although there will
be fecal material remaining in the intestinal tract, and they
do defecate when they wake up.

REINHARD: Is this a considerable amount throughout the in-
testinal tract?

SCHMIDT: I don't know what you would call considerable. Our
samples are usually about 1 gm,

REINHARD: Has anybody investigated bacterial activity in the
intestine in relation to the possible production of materials
which could serve as sources of energy to the host, when ab-
sorbed?

SCHMIDT: No, as far as I know, but I think the source of energy
is the brown fat.

BERRY: Does the total amount of fecal material discharged
by the control animals that get rid of the virus in three or four
days equal the total amount of fecal material of the hibernating
animals which require eighteen days to eliminate the virus?

SCHMIDT: I am not sure I understand you.

BERRY: This would give some measure of digestive tract
motility.

SCHMIDT: Yes, it is my feeling that this virus is taking a

413

SCHMIDT
ride; in the active animals, it simply gets there quicker.

BERRY: It should be possible to measure the total fecal material
discharged.

SCHMIDT; That is right. We have records on this, but I did
not think that the data that I had suggested anything very un-
usual about this, and as I mentioned, I feel that this retention
is a mechanical sort of thing. Now, perhaps not in the case of
Coxsackie B-3. Dempster^ has reported active infection in his
animals. He's worked with bats and also with ground squirrels
of a different species than mine. There are some things that
we have to get straightened out about Dempster's work, but
he claims that, for instance, the virus actually multiplies during
hibernation.

CAMPBELL: What happens if you remove the brown fat be-
fore or during hibernation?

SULKIN: I think it would be impossible to do this, because
although brown fat is located largely in the interscapular area,
it is distributed elsewhere in the body, so you could never do
a definitive experiment of this sort.

CAMPBELL: But it is fairly localized in mammals; you could
take most of it out.

SULKIN: You could take a pretty good chunk of it out. But
there is a lot of brown fat along the spinal column and else-
where, and it may not take much to do the trick.

WALKER: How important is temperature to this natural hi-
bernation that you speak of? How low must this be in either bats
or squirrels?

SCHMIDT: These animals will become groggy, sleepy, tor-
pid and in a state of hibernation around 19° C, for instance, but

1 Dempster et al. 1961. Canad. J. Microbiol. 7: 587-594.

414

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

they are like a bear; that is, they are easily aroused. Their
body temperature can actually be lower than the ambient tem-
perature. This is supposed to be due to the evaporative cooling
effect, but they will enter what is called a state of hibernation
at around 19° C, and as you lower the temperature, within reaLson
now, they get in an increasingly deeper state of hibernation. We
were trying to get pretty deep hibernation.

TRAPANI: How low is the ambient temperature of the burrow
in the natural habitat of the animal?

SCHMIDT: This has been studied by Dr. Mayer^ here in
Alaska. He has rigged up a very clever harness for the ani-
mal made out of guaze, on which he attaches a thermocouple.
The animal goes down into the burrow and drags the thermo-
couple with him. He has made recordings throughout the winter.
It will actuallv get below 0° C for certain periods of time. I
am sure there is a microclimate involved, too, but the tempera-
ture is quite steady in the burrow.

TRAPANI: Then they will waken, shiver, and then go back
to sleep?

SCHMIDT: The feeling is that they do arouse even in the natural
state, periodically, in their burrows, and they will fool around
a little while and then go back into hibernation.

ANDREWES: Two questions I'd like to ask. Am I correct in
supposing that what you told us implies that these various phages
and other viruses that you put into the hibernating animal do
persist for very long periods of time in contrast with E. coll
which are not only relatively, but actually, disappearing?

SCHMIDT: The E. coli get down to a rather low level. The
studies with bacteria were for longer periods of time; how-
ever, I think there is a difference. I don't know just where these
E. coli go, I can't account for this; there are a number of pe-

2 Mayer, W. V. 1960. Bull. Museum Comp. Zool. Harvard Univ. 124: 131-154.

415

SCHMIDT

ciiliar things here in the bacterial picture. For instance, we see
a rather rapid increase in the psychrophiles beyond what you
can get in a test tube. You know how slowly psychrophilic or-
ganisms grow at, say, 4° C. It takes several days to get much
of an increase, but in the squirrel, after he hibernates for a
couple of days you find approximately a six log increase and,
at the same time, you may observe perhaps a three log de-
crease in the count of E. coli. I have no explanation for this
whatsoever,

WALKER: Could this be a matter of competition in the in-
testinal tract? Could the conditions be such that one has a greater
advantage than the other and E. coli is crowded out?

ANDREWES: I don't see why E. coli should actually die.

SCHMIDT: They won't reproduce at these low temperatures.
They have several things working against them, whereas the
psychrophiles have everything working their way, as it were.
This is the temperature at which the psychorphiles can re-
produce, and at the same time, body defenses are rather im-
mobilized so that they can grow unabated.

WALKER: Yes, the pH may change and there are all sorts
of things to which E. coli may be susceptible.

SCHMIDT: Well, I have tested for antibiosis by the plate method.
It was not an exhaustive study, and although I haven't seen any
evidence of this as yet, it is one of the things that certainly
should be considered. Also the moisture content of the intes-
tinal contents changes somewhat. In addition, there is a lot of
mucus secreted into the intestine, which probably serves to
protect the lining during these periods, and this may have an
effect on the bacteria and also the virus.

ANDREWES: My second question refers to Dempster's story
about the enterovirus multiplying. It would seem from what
you have told us that one might infer that this virus is growing
at a temperature lower than any other viruses do grow, so one
might suspect that any growth there is might occur during these

416

MICROBIOLOGY OF HIBERNATION IN GROUND SQUIRRELS

periods of awakening.

SCHMIDT: That is right. This would seem to be the logical
explanation, although Dempster reports viral replication with
hibernating bats. In connection with this, I asked Dr. Sulkin
about his bats; that is, whether they spontaneously arouse. You
see, they're quite a different animal in that they really don't
have a thermoregulatory mechanism. But ground squirrels do.
You can put ground squirrels in a cold room and if they are
ready to hibernate, they will. If they are not, you can keep them
there for a long period of time and they won't hibernate. Demp-
ster makes the point that these bats which he kept at 2^ C did
hibernate, and there was virus multiplication. Coxsackie B-3
was the agent he was using. In all fairness to Dempster, let
me briefly tell you what he has done. He has made quantitative
estimations of virus in squirrels by inoculating suckling mice
with dilutions of homogenates of various tissues. With the bats,
on the other hand, he reports only "isolation" or "no isolation".
He was able to recover the virus from the blood and certain
other tissues. I think that he considers viremia as being in-
dicative of viral replication. If we would discuss this with him,
which I have not done yet, we might be able to iron this out,
but ' certainly there is this possibility that this minor replic-
cation has occurred during these periods of arousal.

NORTHEY: I am sorry to dig up a dead body here, but some-
thing is bothering me very much. Was it the consensus of this
group that protein metabolism resulting from an increase in
total food consumption directly affected antibody production?
I think this is a very fundamental point here. In other words,
can we say then that the cure for pheumococci pneumonia would
be an increase in food intake?

CAMPBELL: I thought I tried to bring this point out when I
was talking. Very definitely there are two things involved. Mostly,
synthesis of antibody involves a breakdown of the antigen. As
I mentioned a while ago, Dixon has pointed out the possibility
that under stress of hyperimmunization, the body uses gamma
globulin. This is the source of amino acids. Under the normal
state it uses free amino acids, so metabolism must play a very

417

SCHMIDT

important role in all this type of stress mechanism, if it goes
more than three or four days.

NORTHEY: Does it depend upon the antigenicity of the material?
This would depend on how antigenetic the material is, in which
case it would be independent of the protein metabolism.

CAMPBELL: I don't know what you mean by antigenicity.
Some of the things like hemocyanin seem to stimulate more
precipitating antibodies than, say, serum albumin. Now, what
the basis of this is, I don't know. Some people say that anti-
genicity is related to molecular weight because all albumin
is a very good antigen. But more important is seeing how a
material is localized and handled by intracellular enzymes.

418

COLD THERAPY IN BACTEREMIC SHOCK
Emil Blair

University of Maryland

School of Medicine
Baltimore 1, Maryland

ABSTRACT

Approximately 10-12 per cent of bacteremic patients go into shock. Mortality is
60-70 per cent despite aggressive therapy, including bacteria sensitive antibiosis.
The basic problem in all forms of shock appears to be a disparity between MRO2 and
CDO2 (circulatory delivery of oxygen to the cell). The rationale of hypothermia is
directed primarily at MROg and secondarily at CDO2. At 32° C, MRO2 is one-third
of normal, cold pressor effect elevates and sustains arterial blood pressure, heart
rate is slowed, and the renal and the CNS flow and ventilation are augmented. The
patient is brought into a metabolic environment more commensurate with the red'-aced
perfusion. No direct effect on the bacterial organism or on antibiosis was observed.
Similarly, leucocytic response was not altered. Patients were cooled only when they
had become refractory to therapy and were in dire straights. Salvage rate was 50 per
cent in 52 cooled. Excluding deaths from various causes, 14 (27 per cent) died of un-
remitting septic shock. In two of these, opsonin indices were found to reduce markedly.
Current studies in dogs with no therapy appear to confirm the view that hypothermia
(32° C) exerts no direct effect on the organism or on host mechanisms. Death is likely
due in part to crippling of the RES.

The syndrome of bacteremic shock has been long recognized as a
serious clinical problem (Laennec, 1831), While the most common
offending micro-organisms belongtothecoliform group, this usually
fatal malady can be precipitated by gram positive bacteria, by
Rickettsiae and by viruses (Spink, 1960; Ebert and Abernathy, 1961),
The pathophysiology is generally pictured as vascular failure or col-
lapse (Romberg, 1899; Gilbert, 1960) and believed due directly or in-
directly to endo- or exotoxins (Spink, 1960; Thai and Egner, 1956;
Aub et al., 1947). The physiologic dysfunction pattern is similar with
both toxins. It is estimated that about 15 per cent of clinical bacter-
emias develop hypotension and the shocksyndrome (Ebert and Aber-
nathy, 1961). Antibiosis along with supportive therapy reduced a 100

419

BLAIR

per cent mortality to 60 percent — 75 per cent in coliform or staph-
ylococcus shock (Altmeier and Cole, 1958;Smithand Vickers, 1960).
Ps. aeroguinosa continues to claim a 100 per cent death rate (Mon-
crief).

Failure to control satisfactorily the staggering losses prompted
exploration for help in another direction. While the complex pattern
from infection to shock and death is yet to evolve, the underlying
problem appears to be hypoxemia. Cellular metabolic requirements
(MROo) in febrile states accelerate. This calls for a proportionate
increase in circulation to assure adequate delivery of oxygen (CDO2).
At any given metabolic level, the ratio of CDO2 to MRO2 determines
the presence or absence of hypoxemia and its degree. As long as the
circulation meets the requirements, there is no hypoxemia. In the
case of septic shock characterized by circulatory failure, the re-
duced perfusion results in a proportionate decrease in tissue oxygen
tension. The disparity lies with CDO2 (assuming alveolar oxygen
tension is maintained). When the ratio falls below a critical level for
a given metabolic state, the cell dies. Current treatment is aimed at
CDO2 in the form of transfusions, oxygen inhalation and vasopres-
sors. This line of attack is inadequate as indicated by the high rate
of death. The next obvious step is to attempt modification of MRO2
in order to bring it more in line with the reduced CDO2 and restore
the normal ratio. Hypothermia is a means of doing this. With pro-
gressive lowering of core temperature of the homeotherm, there is
an exponential fall in oxygen uptake in the absence of shivering
(Spurr etal., 19 54), At 32° C uptake is reduced by one- third, at 30° C
by one- half, at 25° C two- thirds, and four-fifths at 20° C (Blair,
1960). Reduction in blood flow is roughly proportional to that in oxy-
gen consumption. Thus, in septic shock with an already diminished
flow, it appeared possible that with a lowering of MRO2 by hypo-
thermia, a more favorable metabolic environment may supervene.

In addition to the significant decrease in metabolism at 32° C,
other features at this level provide additional physiological benefits.
Cold pressor effect restores and sustains the arterial blood pres-
sure. The heart rate is slowed and its work allayed; the central ner-
vous system is stimulated and reflex mechanisms are augmented;
ventilation is improved, and renal flow is enhanced. This has been
termed the augmented level (Blair, 1960). The thesis for hypothermia

420

COLD THERAPY IN BACTEREMIC SHOCK

then, is to modify the metabolic environment of the host with likely
no specific action at any given site or enzyme system. In instances
such as bacteremic shock, hypothermia would place the metabolic
community at a physiologic level more commensurate with the ex-
isting attenuated blood flow. Hypothermia as a "therapeutic" tool is
quite ancient (Currie, 1798). In recent years a number of reports
have attested to the value of hypothermia in septic problems (Blair
et al., 1961;Cockett and Goodwin, 1961;Martin, 19 58; Drescher, 1960),

This report consists of a description of clinical experiences and
some aspects of experimental studies. Evaluation of the influence of
hypothermia rested on refractoriness to standard therapy. The pa-
tients continued to receive intensive drug and antibiotic treatment
while under hypothermia. In order to evaluate the effects of hypo-
thermia per se, bacteremic shock was induced in dogs, one series
of which received no treatment other than hypothermia.

CLINICAL STUDY

Criteria for Cooling

The sequence of progressive deterioration from an acute infection
to septicemia, hypotension, and shock does not adhere to a specific
clinical pattern. Too many important factors influence the probabil-
ities of shock in septicemia. Some of these are host resistance, in-
cluding age and nutrition; type and virulence of offending micro-
organism; and type, intensity, and timing of treatment. The criteria
for the development of shock secondary to sepsis are largely based
upon hypotension, tachycardia, hyperpnea, and oliguria. All patients
in septic shock were begun on the accepted regimen as enumerated
previously. Only those who subsequently became unresponsive and
seemed about to die were selected as candidates for hypothermia.

Characteristics of a typical clinical picture are listed in Figure 1.
The patient who appeared tobedoingfairly well was usually elderly.
The temperature dropped slightly,ventilationbecame irregular, and

421

BLAIR

1. Elderly debilitated patient

2. Gram negative coliform bacteria

3. Temperature 39° C - U0° C

4. Sudden collapse:

a. pallor or cyanosis

b. disturbed sensorium

c. blood pressure below 70 mm.Hg,

d. weak, rapid pulse

e. oliguria

f. ventilatory disturbance

5. Blood status:

a. elevated BUN

b. whole blood deficit

c. electrolyte deficit

d. leucopenia often

e. metabolic acidosis

6. Mortality 55-70 per cent

Figure 1. Clinical features of bacteremic shock.

422

COLD THERAPY IN BACTEREMIC SHOCK

Level:

33°

C

- 32° C

35 Patients

37°

C

10 Patients

Duration:

1

3

7 14 30 Days

10

16

12 6 1 Day

Figure 2.

Duration of Hypothermia in bacteremic shock.

Total cases - U5

Died

- 22 Rate - 50%

21-30

21-30

Hl-50

51-60 61-70 71+

5/2

3/0
>50
11/3-27%

3/1

15/7 14/8 5/4
<50
34/19-56%

Contributory

Causes

Hypothermia

Electrolyte

below 30°

C 2

deficit 4

Bleeding

5

Unremitting
septic shock

Aspiration 1

10

Figure 3. Mortality factors.

a pallor engulfed the rapidly deteriorating patient. Urine flow ceased,
leucopenia and metabolic derangements were observed often.Refrac-
toriness to therapy, characterized by an almost total breakdown of
physiologic compensation.was demonstratedby marked disturbances
in sensorium or coma, thready or absent pulse, an arterial blood
pressure below 60 mm Hg or unobtainable, and Cheyne-Stokes or
Kaussmalbreathing.A surface coolingtechnic was applied utilizing a
rubber blanket through which refrigerant fluid (ethyl alcohol) was
circulated. The cooling process was stopped at 34° C to 33° C
and the patient allowed to drift to a 32° C to 31° C level at which

423

BLAIR

Observation

Patients

Per cent

Hypoglycemia

4

9

Pressor effect

36

80

Improved sensorium

41

91

Pulse slow

43

86

Improved ventilation

42

93

Figure 4. Response oi

septic

shock patients to cooling.

Pneumonia

0

Cold Bums

0

Palsies

0

Leucopenia

0

Thrombocytopenia

0

Drift

Often (2 deaths)

Hypoglycemia

4

Figure 5. Complications from hypothermia.

he was then stabilized. By controlling blanket temperature, patients
were maintained at 33° C to 32*-* C for extended periods.

Results of Clinical Study

A total of 45 patients were cooled for periods of one to thirty days
with the majority for a week or less (Fig. 2). Gram negative organ-
isms were responsible for the sepsis in43 of which E. coli was most
commonly predominant. The others were S. aureus, coagulaseposi-
tive.Positive blood cultures were obtainable in less than 50 per cent.
Positive cultures were present during hypothermia and upon re-
warming in those who survived.

424

COLD THERAPY IN BACTEREMIC SHOCK

The salvage rate was 50 per cent (Fig. 3). Only one- half of
the deaths were due to unremitting septic shock. At the time
of all fatalities antibiosis was specifically bacterio- sensitive.
The elderly debilitated patient tolerated septic shock most poorly.
Survival rate was 73 per cent in those under 50 years of age
and only 44 per cent in those over 50. Two deaths were associ-
ated with deeper cooling (28° C and 27° C) and were in individuals
over 50 years of age. Response to cooling was uniformly favor-
able, attesting to the augmenting influence of the level of cooling
employed (Fig. 4). However, this response proved no indicator
of the outcome. Most of the deaths occurred within 72 hours of
cooling. Therefore, survival beyond this time increased prob-
abilities of recovery, unless a complication unrelated to the
sepsis intervened. Complications of hjrpothermia proved to be
of little or no consequence where adherence to 32° C was ob-
served (Fig. 5). Four patients developed hypoglycemia, which
was adequately managed by glucose infusion.

The criteria for rewarming were based entirely on the sub-
jective course of the hj^othermic patient and did not bear neces-
sarily any relationship to the state of the sepsis or host anti-
bacterial response. The primary indication for rewarming was
the patient's vocal and muscular objection to the cold state. The
sensorium was clear, the arterial blood pressure and pulse sta-
bilized, and ventilation was normal. Elevation in temperature
occurred in all patients after rewarming. Hypothermia was re-
instituted only when this was accompanied by evidences of shock.

An example of a survivor cooled for eight days is illustrated
in Figure 6, This was an 18 year old white female, who devel-
oped peritonitis following an appendectomy. Septicemia was evi-
denced by an elevated temperature of 39° C to 40° C, tachy-
cardia, hyperpnea, and leucopenia. Bacteroides was cultured
from the blood stream. Surgical exploration was followed by
shock. Failure of response to therapy highlighted by lethargy
and coma prompted resort to hypothermia. Within 72 hours there
^\as considerable improvement in the patient's status. Blood cul-
tures remained positive during the hypothermic state and be-
came negative 12 days after hypothermia was discontinued. Re-
warming was started on the fourth day, but a convulsive seizure

425

BLAIR

■re

40

38
36
34
32

30-
160
I40

120'
100-
80
60'
40
20

BlX»0D CULTURE

WBC IN THOUSANDS
7 5 7 10 25 21

HEMATOCRIT
26 35 35

TRANSFUSION

31 20 25 33

i- -

0 0 0

? io Iz 13 M is i 26* ii ' 67*' 61

DAYS 01 234 5678

0 12 3 4 5 6

HYPOTHERMIA

Figure 6. Case of bacteroldes septic shock managed with hypothermia.

and coma believed due to brain abscess, necessitated further
cooling. Within 24 hours the patient was again awake. The pa-
tient complained of the cold and was shivering actively. Re-
warming was started again on the eighth day and was followed
by an elevated temperature of 38° C, Since the patient's hemo-
dynamics and neurologic status remained normal, no further cooling
was attempted. Figure 7 shows the blood picture, Durir^ the ini-
tial period of the bacteremia, there was a leucopenia. The white
blood cell count was 5,000 with only 30 per cent mature gran-
ulocytes and 50 per cent immature cells. Just prior to cooling,
the white count increased and continued to be elevated during
hypothermia. Mature polymorphonuclear leucocytes appeared in
the high ratio normally expected in severe sepsis. The alter-
ations in hematocrit coincided with the whole blood deficits, and
returned to normal with blood replacement.

426

COLD THERAPY IN BACTEREMIC SHOCK

DAY 0 4
T«C 40 40

6
36

7
33

12
32

14
38

17
3B

WB£ IN
THOUSAhC

S

RBC IN
MILUONS

30 -
25 ■

•5

•.iB_C y

r

^

^

20

"~*^

•4

15 •

/

^

N^

^

3

10

^^

2

5 ■

•1

0 •

0

DIFFEREN

''''*'- 250.000

280.000

PLATELETS

100

90

o.seGs__

80

/

"""^^

v.... ^

HEMAT

XRIT
50

70
60

I

.*■''

--*

X

40

30

50

\X"

"•*'

20

40

X

10

30'

y ^•-.^

^s^^

0

20

•^ —

^^.^

y

Av

10

LYMPHS^^

4- '~^^

^-■^— ^

^

\.

0

='TJoNo5"^

*■<«»<

^'-'^

:*^

^

^^

40

40

36

33

32

36

38

Figure 7. Blood element changes in a case of bacteroides septic shock with hypo-
thermia.

Discussion of Clinical Observations

Hypothermia proved to be a valuable adjunct in this group of
seriously ill patients. Its effect appeared to be primarily metabolic
and permitted an increase in cellular tolerance to hypoxia. This
"slowing down" allowed time in which to pursue treatment directed
specifically against the micro-organism. It is of special interest
that death occurred even while bacterio- sensitive antibiotics were
being poured in. In part, the failure of antibiosis may be related to

427

BLAIR

reduced perfusion, the failure of the needed carrier mechanism.
The hypoglycemia observed during hypothermia is not in keeping
with the usual observations. Hyperglycemia is fairly characteristic
in normal animals cooled from 31° C to 25° C (Bickford and Mot-
tram, 1960; Blair, unpublished data) andinhumansat 30° C (Henne-
man et aI,,I958),Theelevatedbloodsugar has been attributed to re-
duced metabolism and impaired absorption (Henneman et al., 19 58;
Wynn, 19 54), Endotoxins have been demonstrated to produce hypo-
glycemia after an initial hyperglycemia (Berry et al., 1959), Hypo-
glycemia was not observed in only those patients under hypothermia
in this study. One likely explanation may be that shivering caused
this, but since all survivors shivered at one time or another, other
causes must have existed.

While there is little doubt of the efficacious role of hypothermia
in bacteremia shock, the evidence thereto is essentially imperical.
The continued administration of antibiotics, corticosteroids, and
other drugs may well have obscured the precise effect of hypo-
thermia itself. In order to assess the effect of cooling per se, an
experimental study was undertaken.

EXPERIMENTAL STUDY

Methods

Randomized chlorolosed mongrel dogs of both sexes weighing be-
tween 10-15 kg were induced into septic shockby peritoneal instilla-
tion of 1 to 1.5 gms of feces, suspended in saline. Ventilation was
unsupported. Three groups of dogs were studied: normal (B), septic
shock (S.S.), and septic shock plus hypothermia of 32° C (S.S. + H.).
Arterial blood pressure (ABPm) was monitored, as were heart and
ventilatory rates (HR and Vj^, Arterial and mixed venous samples
and expired air were withdrawn for baseline controls during septic
shock and after hypothermia was induced. Oxygen content was de-
termined by Van Slyke-Neil method, while oxygen uptake (V02) was
determined by Scholander; and from these cardiac output (Q) was

428

COLD THERAPY IN BACTEREMIC SHOCK

Hours

S.S. + H.

S.S.

2

6

8

10

1U+

0

86

100

0
13
38
75
88
100

Figure 8. Per cent mortality rates in experimental gram negative septic shock with
(S.S.) and without hypothermia (S.S. + H.).

HEART RATE

-PER MIN

240'

•oo

MEAN ARTERIAL

•e

•

VENTILATORY

RATE

BLOOD PRESSURE

1

PER MIN

MM Hg

200'

•

•oo

•

•

•

0

:

0

0

160'

•oo

0

•

• 0*

•

r

0

0

ic

•

8

•

oo

120

oo
o ooo

•

e

o

0

s

o

•

oSo

oo

o

•

0

SO

o
•
o

0

•
8

M

•o
•

5

0

•

8

•

0

40

o
•

s

ooo

0

o

0

Figure 9. Hemodynamic and ventilatory rate changes in septic shock (closed circles)
and septic shock plus hypothermia (open circles).

429

BLAIR

Vo,

Vt

200-

CC MIN

L MIN.

160-

• OO o
00 •

M

-16

120-

o

•
o

-12

80-

o

o

-8

oo

• s

o
o

s

0

40-

1*

oo

o
o

1

• 0

0

o

§

-4

ss

SS*H

ss

SS*H

Figure 10. Ventilation in septic shock (closed circles) and septic shock with hypo-
thermia (open circles), Vg, "oxygen consumption; V-p ■= minute ventilation. During septic
shock, an increased degree of breathing was required to yield a given amount of oxygen.
Hypothermia significantly allayed this undue ventilatory work by reducing oxygen re-
quirement (MROg).

computed. Cultures of the blood stream were done 0, 15, 30, and
60 minutes after instillation of feces. White counts and differ-
entials were done also prior to and after septic shock in both
groups. It is emphasized the animals received no fluids or treat-
ment other than hypothermia in the S.S. + H. group, except for
replacement of blood removed for analysis.

Results of Experimental Study

Mortality. Figure 8 shows the mortality rate in both groups.
Eighty- six per cent of the animals died in septic shock in four
hours, and all died within six hours after fecal instillation. In
the hypothermic group 13 per cent died by the four-hour period,
88 per cent by 10 hours, and all by around fourteen hours. Hypo-
thermia did not save any of the animals, but it did prolong survival

430

COLD THERAPY IN BACTEREMIC SHOCK

BLDOO CULTURE

as 13

T«C
A2-

z

4

g

58

l^ _■— -—— ""^^^

36

220
200'

leo

140
120

too

80
60
40^
20'
0

,». •■ • hCART RATE

.A,

X' \ ^ MEAN ARTERIAL

>'■ \ BLOOD PRESSURE

» VENTILATORY RATE

2 3 35 4

HOURS

Figure 11. Example of a septic shock study. Concomittant with massive bacterial in-
vasion of the blood stream, the temperature rose, arterial blood pressure dropped,
ventilation and heart rates accelerated.

significantly.

Physiologic changes. Figure 9 demonstrates alterations in
ABPm, HR, and Vj^. Septic shock was characterized by significant
fall in ABPm, by tachycardia, and by hyperpnea. Hypothermia
resulted in a further fall in ABPm, and a slowing of HR and Vj^ to
or below pre-septic shock values. Non- cooled dogs prior to death
underwent serious reduction in the compensating hj^ierventilation.
Vq and "C^-p (minute ventilation) values are shown in Figure 10.
V02 in septic shock remained essentially similar to pre- infection,
with an increased V-p.Upon cooling, Voo ^^'^ Vrj. declined appreciably.

431

BLAIR

J2

50-

220
200

MEAN ARTERtAL
BLOOD PRESSURE

4 45 6

HOURS

K) 12

Figure 12. Example of a septic shock and hypothermia study. Note again onset of shock
shortly after flooding of blood stream. This likely represents period of failure of clear-
ance mechanisms of the blood stream. Under hypothermia the arterial blood pressure
changed little, but ventilation and heart rates slowed considerably. Upon rewarming,
heart rate re- accelerated and ventilation decreased, but not to previous high pre-
hypothermic level.

Examples of the course of events in an animal from each group are
illustrated in Figures 11 and 12. Hypotension, tachycardia, and hyper-
pnea developed at about the same time the blood stream was over-
whelmed with bacteria. The hypothermic dog lived for 14 hours, and
then succumbed after it had beenrewarmed. Positive blood cultures
persisted during the hypothermic period. Effects on A-Vqo appear
in Figure 13. There was a marked widening (three times normal)
which indicated increased extraction of oxygen. At the time of death
the A-Vq2 widened further, while in the cooled septic dogs there

432

COLD THERAPY IN BACTEREMIC SHOCK

02,

I n

B

I a

I n

SS S&*H
3-4

n

SS*H
6-8

Figure 13, A-Vq differences in septic shock (closed circles) and septic shock with
hypothermia (open circles). Baseline (B) values are similar in both groups. During septic
shock (S.S.), A-Vq rose with further rise later in the S.S. group. Under hypothermia
(S.S, + H.), however, there was significant improvement in A-Vq^.

Organism

Per cent

E. coli
Streptococcus
Psuedomonas
Aerobacter
Al. fae calls

62
46
23
15
8

Figure 14. Incidence ofpredominant micro-organisms cultured from the blood stream.

433

BLAIR

Time Minutes

15 30

Hours
2 3

Per cent
initial

Per cent

TNTC

30 45

25

50 37 13

Figure 15. Percentage rate of appearance time of micro-organisms in blood stream.

WBC IN
THOOSANDS

• SEPTIC SHOCK CONTROLS

0 SEPTIC SHOCKS HYPOTHERMIA

Figure 16. White blood count in septic shock. A marked leucopenia developed with
stabilization and some improvement under hypothermia (32° C). Just before death in
the hypothermic group, there was a further fall in white blood cells.

was a reduction toward, but not to, normal.

Bacteriology. The most common organism cultured from the blood
stream was Escherichia coli (Fig. 14). Cultures included also Strep-
tococcus, Pseudomonas, and Aerobacter. Combinations of E. coli
with these were frequent. Bacteria were detected in the blood stream
as early as 15 minutes after inoculation (Fig. 15). The blood stream
was overwhelmed within two hours, which was about the same time
severe shock developed. Cultures remained positive during the
h3T)othermic period.

Leucoc3^ic response. A severe leucopenia developed in both
groups, reaching its most advanced state at the time of death
(Fig. 16). Induction of hypothermia stabilized the level of white blood
cells. There was also a tendency for some increase in count, but

434

COLD THERAPY IN BACTEREMIC SHOCK

HYPOTHERMIA

SM-2 SH-3 SH-4

Figure 17. Differential counts demonstrate the leucopenia was due primarily to re-
duced granulocytes.

the series is too small to attach significance to this. There was
a further drop just before death, as occurred in the non- cooled
group. The leucopenia was due primarily to disappearance of
granulocytes (Fig. 17). Lymphocytes show a relative rise in
count with no apparent alteration in monocjdes. Cooling resulted
in no further change in the differential; only a temporal extention
of the pattern was noted.

Discussion of Experimental Study

Criticism of experimental preparation. Septic shock in the dog is
not the same as in the human. Avulsion of the cecum may produce a
preparation more similar (Pulaski et al. , 1954). In the study per-
formed here the inoculum was homologous. The temporal relation-
ship also was considerably shorter. Better control of bacteriologic

435

BLAIR

%>N

Figure 18. Relation of minute ventilution (V^) to oxygen consumption. For details
see text.

effects per se may be obtained if a pure strain of E. coli had been
used instead of coliform mixtures, which varied with each inoculum.

In this study, interest was centered primarily on the general
physiologic and metabolic effects of septic shock alone and as modi-
fied by the augmented level of hypothermia.There is less operative
trauma with fecal instillation, and the entire study can be completed
within 6 to 12 hours. Particularly interesting is the fact that the he-
modynamic and hematologic pictures are similar to those caused by
endotoxin (Spink,1960;Gilbert,1960). The mixed flora is more akin to
the clinical insult thanpureendotoxinor E, coli. One significant dif-
ference between endotoxin and coliform induced changes is that Vq^
is not decreased as much, whereas Vrp is greater with coliform sep-
sis (Maxwell etal.,1960).Increase or no change in Vq^ and at least an
initial increase in Vrp are characteristic of the clinical syndrome.

Hypothermic effects. Septic shock in these preparations resulted
in an increased work load on the cardiopulmonary systems. Response
was adequate to a point, but with progressive crippling from the tox-
emia, compensatory mechanisms broke down. The animal descended

436

COLD THERAPY IN BACTEREMIC SHOCK

deeper and deeper into hypoxia and finally succumbed from cardio-
pulmonary failure. The increased ventilatory work required in
septic shock is demonstrated in Figure 18. To provide 80 cc of
oxygen, normally the dog must breathe 2,8 liters/min, (Curve B).
In septic shock (Curve S.S.) an equivalent oxygen availability
required 6.2 liters/min. The dog's breathing apparatus had to
work 120 per cent harder. When cooled (Curve S.S. + H.), 80 cc
of oxygen was obtained with 4.5 liters/ min., reducing work load
in septic shock by one-half, but still 60 per cent above normal.
Failure of circulation with reduced tissue perfusion has been
most often pin-pointed as the main mechanical breakdown leading
to death (Spink, 1960; Altmeier and Cole, 1958), These studies
are in agreement. Figure 19 demonstrates the relationship be-
tween X^Oo ^^^ ^ ^^ the normal. A and A' represent the approx-
imate range of normal flow required for a given Vq-, In Fig-
ure 20 the effects of septic shock with and without hypothermia
are illustrated. Normally, to maintain adequate tissue oxygen-
ation, 100 cc of oxygen is delivered by a minimal flow of 1.1 liters/
min. In septic shock, the equivalent oxygen is supplied by a markedly
reduced flow of 0.4 liters/min., a flow deficit of 65 per cent.
Induction of hypothermia returned the relationship toward nor-
mal. This was achieved primarily by reducing MRO2 with likely
little alteration in perfusion. This represents the single most
important benefit of hypothermia in this condition. Metabolic
environment was brought to a level more commensurate with
the sharply compromised capabilities of the physiologic apparatus.
The pathophysiologic picture of death in the hypothermic dogs
was similar to that in the non-cooled group. The difference is
primarily a temporal one. The shock state is due primarily to
reduced cardiac output, as is also reported to occur in the human
(Gilbert et al., 1955) and in experimental endotoxemia (Maxwell
et al., 1960). This is attributed to reduced venous return, secondary
to hypovolemia and to pooling (Gilbert, 1960; Ebert and Abernathy,
1961), with the latter occurring primarily in the liver and the
splanchnic beds.

The physiologic picture of septic shock in the human and in
the dog are similar. Hj^othermic effects, however, differ principally
in failure to raise ABP in the dog. This may be due to anesthesia
or to species difference response. The criteria of septic shock in

437

BLAIR

6 l/min

Figure 19. Relation of blood flow {(Jj) to oxygen consunnption (Vq^). Curves indicate
normal range of flow for a given amount of oxygen.

0 L/MI

200 229

v^^ce/»..N

Figure 20, Curves A- A' are reproduced from Figure U. In septic shock the curve
falls toward the abscissa demonstrating reduced flow (closed circles). Under hypo-
thermia the Q-Vq2 relationship is improved.

438

COLD THERAPY IN BACTEREMIC SHOCK

the dog are hypotension, hyperpnea, tachycardia, and increased
A-"C^Q2. In the human they are hypotension, tachycardia, and a
reduced and irregular ventilation, and widened A-Vo2»

Leucopenia. The leucopenia is reported due to endotoxic effect
and the hyperpyrexia due to pyrogens from granulocytes. The
mechanism is obscure, but may be in part caused by depression
of bone marrow activity. The breakdown of host resistance (Zwei-
tach et al,, 1957) lies in part to depression of the reticuloendo-
thelial system. Direct endotoxin effect or reduced perfusion or
both may contribute to the RES breakdown. Granulocytes were
affected primarily in this study. However, counts were made
on whole blood, and the increased hematocrit in this type of
shock may introduce an error in the cell counts.

The persistence of positive blood cultures under hypothermia
indicated no significant effect on the bacteria at the temperature
used. Lower temperatures are required for possible influence on
growth (Balch et al., 1955).

Hypothermia presumably does not alter the blood's ability
to clear bacteria (Fedor et al., 1956; Frank et al,, 1956), Pre-
vious experiences on the effect of hypothermia in experimental
infections have produced conflicting reports. Increased mor-
tality rate was observed in rabbits with pneumococcal infection
after cooling to 31° C (Sanders et al., 19 57). A similar study
at 21° C yielded an increased survival rate (Wotykins et al.,
19 58). Death rates in rats from gram negative septicemia were
improved with cooling to 25° C (Balch et al., 19 55). The vari-
ations in results may well be due to species differences, depth
of cooling and virulence of the bacteria used.

SUMMARY

Bacteremia shock is usually of gram negative coliform origi"h.
The mortality persists at 65 to 70 per cent despite appropriate

439

BLAIR

antibiosis and supportive therapy. The underlying physiologic dys-
function appears to be hypoxia secondary to a tissue perfusion
deficit. The rationale of hypothermia lies in the reduction of MRO^
in order to place tissue needs at a level more commensurate with
the reduced blood flow. The action of hypothermia is non-specific.
Its effect lies in the modification of the host metabolic environment.
The augmented level of 32° C has proved effective and safe and
likely does not alter bacterial growth.

ACKNOWLEDGEMENTS

The able assistance of George Henning, Allan Land, Luther
Leibensperger, Dorothy Suter, McRae Williams, and Joseph Wilson
was most valuable. This study was supported by U. S, Public Health
Service Grant No, HE-06154-02 and Research Career Award No.
HE-K3-4232(C1) and by a grant from OTSG, U, S. Army Research
and Development Command,

LITERATURE CITED

1. Altmeier, W. A, and W, B, Cole, 1958, Nature and treatment of

septic shock. Arch. Surg. 77: 498-507.

2. Aub, J. C, P. C. Zemechik, and I. T. Nathanson. 1947. Physio-

logic action of Clostridium oedematieus (Novyi) toxin in dogs.
J. Clin, Invest. 26: 404-411.

3. Balch, H. H., H. E, Noyes, and C, W, Hughes, 1955, The influence

of hypothermia on experimental peritonitis. Surgery 38:1036-
1042.

440

COLD THERAPY IN BACTEREMIC SHOCK

4. Berry, L. J., D. S. Smythe, and L. G. Young. 1959. Effects of

bacterial endotoxin on metabolism. I, Carbohydrate depletion
and the protective role of cortisone. J. Exp. Med. 110: 389-405.

5. Bickford, A. F,, and R. F.Mottram. 1960. Glucose metabolism

during induced hypothermia in rabbits. Clin«Sci.l9: 345-359.

6. Blair, E. 1960. Hypothermia- physiologic rationale. Am. Pharm,

Therap. 1: 758-768.

7. Blair, E., R. W. Buxton, A. R. Mansberger, and R. A. Cowley,

1961. The use of hypothermia in septic shock. JAMA 178: 916-
918.

8. Blair, E. Tobe published. Physiology of immersion hj^iothermia.

U. S. Army R and D Report.

9. Cockett, A. T. K,, and W. E, Goodwin. 1961. Hypothermia in the

management of bacteremic shock. U. S. A. F. Review p. 8-61.

10. Currie, J, 1798, Medical reports on the effects of water, cold

and warm, as a remedy for fever and other diseases. J,
M'Creery, Liverpool.

11. Drescher, C. 1960. Weitere erfahrungen mit der abkohlungsbe-

handlung der allgemeinen peritonitis und anderer mit hyper-
thermic einhergehender erkrankungen. Zbl. Chir, 85: 2342-
2349.

12. Ebert, R. V., and R.S. Abernathy. 1961. Septic shock. Fed. Proc.

Suppl. 9, 20: 179-184.

13. Fedor, E, J., E, R. Fisher, S, H. Lee, W. K. Wertzel, and B.

Fisher. 1956. Effect of hypothermia upon reduced bacteremia.
Proc. Soc. Exp. Biol. Med. 93: 510-512.

14. Frank, E. D., D. Davidoff, E. W. Friedman, and J. Fine. 1956.

Host resistance to bacteria in hemorrhagic shock. IV. Effect
of hypothermia onclearanceof intravenously injected bacteria,
Proc. Soc. Exp. Biol. Med. 91: 188-189.

441

BLAIR

15. Gilbert, R. P. 1960. Mechanisms of the hemodynamic effects of

endotoxin. Physiol. Rev. 40: 245-279.

16. Gilbert, R. P., K. P. Honig, J. A. Griffin, R. J. Becker, and

B. H. Adelson. 1955. Hemodynamics of shock due to infection.
Stanford Med. Bull. 13: 239-246.

17. Henneman, D. H., J. P. Bunker, and W. R. Brewster, Jr. 1958.

Immediate metabolic response to hypothermia in man. J. Appl,
Physiol. 12: 164-168.

18. Herion, J. C, R. I. Walker, and J.G. Palmer. 1960. Relation of

leucocyte and fever responses to bacterial endotoxin. Am. J,
Physiol. 199: 809-813.

19. Hinshaw, L. B.,W. W. Spink, J. A. Vick, E. Mallet, and J. Fin-

stad. 1960. Effect of endotoxin on kidney function and renal
hemodynamics in the dog. Am. J. Physiol. 201: 144-148.

20. Laennec, R. T. H. 1831. Traite de I'auscultation mediate et des

maladies des poumons et ducoeur.p. 138. J. S. Chaude, Paris,

21. Martin, C. 1958. Sur I'hibernation artificielle appliquee au

traitment d'uncastres severe deseptico-pyohemie a staphylo-
coccus. Presse Med. 61: 84-86.

22. Maxwell, G. M., C. H. Castillo, C. W. Crumpton, S. Afonso,

J. E. Clifford, and G. G. Rowe. 1960. The effect of endotoxin
upon the systemic, pulmonary, and coronary hemodynamics and
metabolism of the intact dog. J. Lab. Clin. Med. 56: 38-43.

23. Moncrief, J. Personal communication.

24. Pulaski, E. J., H. E. Noyes, J. R. Evans, and R. H. Bralune.

1954. The influence of antibiotics on experimental endogenous
peritonitis. Surg. Gyn. Obst. 99: 341-349,

442

COLD THERAPY IN BACTEREMIC SHOCK

25. Romberg, E., H. Passler, C. Bruhns, and W. Mueller. 1899.

Untersuchungen (Iber die allgemeine pathologie und therapie
der kreislaufst6rung bei acuten infectionkrankheiten, Deut-
sches Arch. Klin. Med. 64: 652-657.

26. Sanders, F., E. S. Crawford, and M.E.Debakey. 1957. Effect of

hypothermia on experimental intracutaneous pneumococcal in-
fection in rabbits. Surg, Forum 8; 92-97,

27. Smith, I. M., and A. R. Vickers. 1960. Natural history of 338

treated and untreated patients with staphylococcus septicemia.
1936-19 55. Lancet 1: 1318-1322.

28. Spink, W. W, 1960. The pathogenesis and management of shock

due to infection. Arch. Int. Med. 106: 433-442.

29. Thai, A. R., and W. Egner. 1956. The mechanism of shock pro-

duced by means of staphylococcal toxin. Arch, Path. 61: 488-
494,

30. Spurr, G. B,, B, K. Hutt, and S. Horvath. 1954, Responses of

dogs to hypothermia. Am, J. Physiol. 179: 139-145.

31. Wotykins, R. S,, H, Hirose, and B, Eiseman, 1958. Prolonged

hypothermia in experimental pneumococcal peritonitis. Surg.
Gyn. Obst. 107: 363-368.

32. Wynn, V. 1954. Electrolyte disturbances associated with failure

to metabolize glucose during hypothermia. Lancet 2:575-578.

33. Zweifach, B. W., B. Benacerraf, and L. Thomas. 1957, The

relationship between the vascular manifestations of shock
produced by endotoxin, trauma and hemorrhage, 11. The pos-
sible role of the reticulo- endothelial system in resistance to
each type of shock. J, Exp, Med. 106: 403-414,

443

BLAIR
DISCUSSION

BERRY: I would like to take just a few minutes to tell about
some very preliminary findings we have at Bryn Mawr, perhaps
relevant to the problem that Dr. Blair emphasized so clearly;
the problem of cell anoxia in the patient with septic shock. I
want to preface my remarks by recalling my early years as a
scientist, the era of vitamin research. Animals were made
vitamin deficient and had all sorts of symptoms. It seemed
hopeless to find what the vitamin was doing metabolically, be-
cause the symptoms were so diffuse, and this is very much the
problem with endotoxin.

Now, the metabolic effects of endotoxin are so diffuse and so
wide-spread in an animal, it may be impossible to ever pinpoint
the specific enzyme that might be involved. I'm not sure that
I have found the enzyme, but I think there is one enzyme in-
hibited by endotoxins, and this enzyme is important to the whole
well-being of the animal, and may be a valuable clue. The enzyme
is tryptophan pyrrolase. It converts tryptophan oxidatively into
knyurenine which, in turn, is changed into nicotinamide, which
is incorporated in the pyridine nucleotides DPN and TPN. These
are the primary hydrogen acceptors in oxidative metabolism, and
they are, therefore, directly involved in most of the energy re-
lease of the mammalian organism. The reason that I started
looking at this enzyme is because the glycocorticoids of the
adrenal cortex which are the most effective, if not the only ef-
fective compounds capable of protecting animals against endo-
toxin, cause a prompt and large increase in tryptophan pyrrolase
in the livers of experimental animals. One finds, also, that tryp-
tophan will cause an increase in tryptophan pyrrolase, probably
as a kind of an adaptive enzyme response. If the enzyme, tryp-
tophan pyrrolase, really plays an important role in the pro-
tection of an animal against endotoxin, then tryptophan should
be about as effective as cortisone in protecting against endotoxin.
This was not found with tryptophan, much to our surprise, but
it was observed with niacin at a level of 20 jU M, while DPN,
at 0.3 /iM, was also active. This is what one would predict if the

444

COLD THERAPY IN BACTEREMIC SHOCK

protection is dependent upon the pjn^idine nucleotides. An en-
couraging part of this study is the fact that an analogue of niacin,
3-acetyl pyradine, produces in mice some of the symptoms of
endointoxication. The ruffled fur, the apparent drop in blood
pressure, the adrenal insufficiency, etc. are obtained with this
nicotinic acid analogue. If one waits a matter of only four hours
after giving endotoxin and then attempts to protect with any of
these compounds, none is effective. I don't know why. But at
least there is hope that we may have one level of explanation for
the metabolic events that produce cell anoxia, Dr. Blair. We
need to do many more experiments. I hope what I report now
stands up on repetition.

BLAIR: With regard to Dr. Berry's very exciting observations,
there is no question but that before this matter of endotoxic
shock, if that is really what it is, can be handled adequately by
physicians, you have to know a great deal more about the basic,
fundamental changes, and this is certainly an approach to that,
I personally would be most interested in this in relation to oxy-
gen tensions, because there has to be some relationship as to
whether it is the endotoxin itself that produces this deficiency,
or is the hypoxia secondary. In other words, is it simply a matter
of lack of oxygen?

Just in line with this, we at the University of Maryland have
become interested in the use of hyperbaric oxygen and treat-
ment of varied and sundry illnesses. This is going to be the
latest gimmick on the surgical scene now, and I have taken two
of our dogs in septic shock, thrown them into a modified chamber
at three atmospheres of whole oxygen, and brought out two dead
animals, so I don't think that this particular approach is neces-
sary. Maybe we gave them too much oxygen.

With regard to the matter of the hypothermia and the pseud o-
monas infections, on the surface, the matter seems to be an
incongruity from the standpoint of recommendations for cooling
a patient. On the one hand, it seemed that hypothermia is pro-
bably a bad thing to use, and certainly one gets in to very serious
trouble if he allows his patient either spontaneously or under
deliberate cooling conditions, to drop down to too deep a level.

445

BLAIR

In general, we observe patients who are about to go in shock
to show a slight fall in temperature. The extreme hyperpyrexia
does change. He still has a high fever, but there is usually a
drop of 1° C on the average. Apparently the hyperpyrexia is
related to the problem of the protective coverings of the skin.
It's pretty important, apparently. Hypothermia in experimen-
tally burned animals is quite a difficult problem to cope with.
There have been several attempts to use hypothermia in burned
animals, and none of them have been successful. Now, whether
this burn septicemia itself is a separate entity or somewh9.t
different than those from germs that come out of the colon,
I don't know, but certainly there may well be some very sig-
nificant differences in the toxins of these two circumstances.

Concerning cold therapy, there is no such thing as cold therapy
at all. Cold doesn't treat anything. Cold modifies metabolic en-
vironment. Sometimes it is good, sometimes it is bad. The
point of the matter is that use of hypothermia does not dic-
tate the ultimate survival of the patient, or certainly of the
experimental animals. It is only adjunct, and it is a crutch.
Heaven knows we certainly need crutches in treating very sick
people, and hypothermia has been somewhat useful in that respect.

446

SUMMARY OF SYMPOSIUM
W. J. Nungester, M. D.

University of Michigan

Medical School

Ann Arbor, Michigan

Years ago, the great naturalist Agassiz counselled, "Go to Nature,
Take the facts into your own hands. Look and see for yourself."
During the past three days, we have followed this sage advice. But as
we have looked, perhaps we have not seen with complete clarity, for
the problems are complex. Yet those responsible for this symposium
have given us a widely arranged spectrum of subjects for discussion
which have afforded us the opportunity to 1) "go to Nature" and look
at the effects of cold on the incidence of naturally acquired infections
in man; 2) "take the facts in (our) own hands" and see for ourselves
the effects of hypothermy on experimental infection; and finally, 3)
hear reports on the search for mechanisms by which host-parasite
interactions are affected by hypothermy.

In examining the data on the effect of cold on naturally acquired
infections in man, we are confronted by complexities, as was clearly
pointed out by Dr. Berry, Dr. McClaughry, Dr. Babbott, and Sir
Christopher Andrewes, In determining whether or not lower ambient
temperatures affect host- parasite relations, ithas been shown that a
number of factors must be considered, among which may be listed
the following: 1) the temperature gradientbetweennormalbody tem-
perature and the environment; 2) moisture or wind which increase
the transfer of body heat to the environment with a local and general
cooling of the body; 3) non-uniform cooling on parts of the body which
may have peculiar physiological effects, as demonstrated by Mudd
et al, (1921) ; 4) since we know from Pfluger's Law that the physio-
logical response is related more directly to the rate of change of
the stimulus than to the strength of the stimulus, rate of change of
body temperature is significant; 5) physiological conditions of the
host affecting heat transfer from the skin as peripheral dilation (al-
cohol) or the effects of local or general cooling of the body as car-
dio vascular diseases, endocrine disturbances, and so forth; 6) pre-
vious acclimatization of the subject -to cold, whatever this means

447

NUNGESTER

physiologically; and 7) pathogenic microbial flora of the host or his
environment at the time of exposure to low temperature.

To suppliment all this, the question might be asked, does an ex-
cessive heat transfer from the host to his environment affect host
resistance to infectious diseases? As a starting point, we should ask
ourselves what has been decided as to the effect of low ambient tem-
peratures on the resistance of man to infectious diseases. As Dr.
Berry recalled, many of his elders as well as his country doctor all
warned against wet feet (and heat loss). At that time, these people,
being fine Texans, believed that exposure to cold did predispose to
upper and lowerrespiratorytractinfections. Of courseproof of such
widespread beliefs is hard to come by, as Dr. McClaughry and Sir
Christopher have stated, Goldstein (1951) has claimed that extreme
cold is a "potent stress factor in bringing about the common cold".
Troisi et al. (1953) found a higher incidence of upper respiratory in-
fections in 66 men working in cold rooms of meat preservation
plants. Andrewes has shown us that colds are more prevalent in
temperate zones in winter than in summer, and as we know, the rate
of change of heat loss can be quite marked in these zones. The
chances for irregular cooling of body surfaces may be significant.
But I think we would all agree that a definitive answer has not yet
been derived.

It would appear that we need more epidemiological evidence on
the effect of increased heat transfer from the host and on suscepti-
bility to infection. Experiments in the future should attempt to cor-
relate not cold as such, but rather, the various factors associated
with general or local loss of body heat with the incidence of infection.
And what are these various factors ? If only some knowledge as to
what measure to use other than rectal temperature or ambient
temperature were available. Perhaps then we would have a more
direct measure of what exposure to cold does to the host, and by this
many of the variables which now confront us might be reduced con-
siderably.

Turning now to another aspect of the symposium, we might recall
the reports on experiments with animals in which host resistance
was altered by exposure to low ambient temperatures. Conclusive
data was presented by Dr, Previte and Dr. Miraglia which showed

448

SUMMARY

that host resistance to selected strains and doses of salmonella and
staphylococci was decreased by a low environmental temperature.
Dr. Metcalf and Dr. Walker demonstrated that in chilled animals
there is a lowered resistance to influenza and Coxsackie B-1 viruses
respectively. And Dr. Marcus has shown us that chilled mice which
were not acclimatized are more susceptible to Coxsackie B- 5 virus
than are normal animals. On the other hand, Dr. Sulkin has told us
that there is less rabies or encephalitis virus produced in bats at
-2° C than at higher temperatures. These findings clearly indicate
the complexnatureoftheproblemsweface.lt is indeed logical to ex-
pect less virus production in cells whose metabolism is lowered by
cold; but pathogen replication is only one phase of disease produc-
tion, although it is an important one.

Another aspect of these experiments conducted with animals is
that it became clear that in challengingthem while kept at normal or
low ambient temperatures, one must define certain experimental
procedures quite carefully. These include such aspects as ambient
temperature, rectal temperature, cage type, bedding (or lack of
bedding), acclimatization (which should be defined by physiological
measurements, if possible), virulence and dosage of pathogen, and
site of inoculation (S. C, or I. P. being the major methods used).

Dr. Walker's report on virus replication showed that the strain of
Coxsackie used was replicated in the pancreas of animals kept at
normal temperatures, but was not replicated in other tissues. In-
fections of the pancreas are rare, yet this strain of Coxsackie
selects this organ as its site of operations in the normal animal.
When the animals were chilled in Dr. Walker's experiments, the
virus became replicated in many tissues.

Dr. Berry cited Shephard's success in producing growth of M.
leprae in mice injected in the foot pad, and Shephard has surmised
that success was based on the lower temperatures of foot pad tissue.
We all know that T. pallida will not infect rabbit testicles unless the
ambient temperature is lower than 59.5° C. Also, elevated body
temperatures have been used in man to treat s3T3hilis. One wonders
if these effects are related to the metabolism of the pathogen or
host defense mechanisms. It is of some interest to observe that cer-
tain pathogens such as the trypanosomes and leptospira grow much

449

NUNGESTER

better in vitro at 30° C than at 37° C, yet in vivo they do very well
at 37° C.

If we accept the evidence currently available, we must conclude
that cold does lower animal host resistance to many infectious a-
gents.The question with which we must concern ourselves is, "How?"

Host resistance to infectious agents involves anatomical, physio-
logical, biochemical, and immunological factors, and most of these
have been touched upon by the various participants here.

Anatomical. Any break in the skin resulting from frost bite or
cold sensitization obviously increases the chances of infection. Some
of the most serious infections of man result from such losses of
mechanical protection of the skin, as can be seen in the infection
which can complicate burns. Much more subtle anatomical factors
such as the bloodsupply of theskin,the nasal mucosa, and the arch-
itecture of the upper respiratory tract may be concerned with this
problem.

Physiological. The physiological changes in the host resulting
from sudden or prolonged exposure to cold are even more subtle.
Changes of this type have been called to our attention by Dr. Blair
and Dr. Miya. The latter has shown that there is an increased re-
sistance to bacterial endotoxins in the cooled animal, while Dr. Blair
and Col. Moncrief have described their use of hypothermy to treat
bacterial shock in man. Probably the increased resistance of chilled
animals and man to endotoxin is based on some physiological mech-
anism which has not yetbeen defined. As Sir Christopher pointed out,
disease production and mechanisms related to it are important as-
pects of this story. And the report of Washburn (1962) on the changes
in blood circulation in frost bitten skin is pertinent to the lowering
of skin resistance by cold.

My own experience in this area relates to the mucus secretions in
the upper respiratory tract. It is recognized that sterile hog gastric
mucin (Nungester et al., 1936; Olitzki, 1948) and human respiratory
tract mucin (Nungester et al., 1951) lowers host resistance to bac-
terial infections. Experimental pneumonia can be produced readily
by injecting a suspension of bacteria and mucin deep into the res-

450

SUMMARY

piratory tract of dogs or rats (Nungester and Jourdonais , 19 36) . Also,
pneumococci and mucin placed in the nose of rats will produce ex-
perimental pneumonia under certain conditions which alter the nor-
mal physiological defense mechanisms (Nungester and Klepser,
1938), But do such findings have any relation to cold and respiratory
tract infections?

Mudd et al. (1921) have shown that an uneven cooling of skin sur-
faces causes vascular changes in the nose. It is common for nasal
secretions to flow out of the anterior nares when one comes into
contact with cold air. Either the ciliary mechanisms fail or there is
a marked increase in secretions. It seems to me that the latter is
more correct, for it is based on a sudden change in blood supply to
the mucus secreting glands. This hypothesis is based on the early
findings of Mudd and his colleagues, A marked increase in respira-
tory tract secretions based on physiological stimulation (cold air)
or infections will overburden the cilia and cause the mucus secre-
tions to accumulate in the upper respiratory tract. Such an over sup-
ply may drain out the anterior nares, be swallowed, or might pos-
sibly be aspirated. Is it possible, then, that marked cooling of the
body may increase the chances for aspiration of infected mucus
secretions from the upper respiratory tract?

We have some evidence from our laboratory (Nungester and
Klepser, 1938) which showed that the reflex control of the glottis is
decreased by body chilling. Mechanical stimulation of the glottis
area with a smallwireinlightly anesthetized rats resulted in closing
of the glottis in all but 18 per cent of 754 stimulations. With rats
chilled for ten minutes in ice water, the glottis did not close in 46
per cent of 552 stimulations. In 21 other rats, none of them aspir-
ated India ink colored mucin placed in the nose, while 55 per cent of
20 rats previously chilled aspirated the material into the lungs.
Pneumococcus pneumonia developed in 13 per cent of 46 normal rats
inoculated internasally with mucin and pneumococcus. In another
group of 24 rats, chilled in ice water and similarly injected, 42 per
cent developed pneumonia. These results may or may not be related
to severe chilling and pneumonia in man.

Dr. Tunevall has told us that absorption of tetanus toxin is delayed
in the hypothermic animal. This directs our attention to the physi-

451

NUNGESTER

ology of the lymphatics and the peripheral circulation. Possibly the
increased resistance of the hypothermic animal to bacterial endo-
toxin reported by Dr. Miya and the clinical findings of Dr. Blair are
directly related to the findings of Dr. Tunevall. We might call atten-
tion to the findings of Klepinger etal. (1959) that of 58 drugs tested,
all but strychnine, chlorpromazine, and promazine were more ac-
tive in animals kept at 36° C than at 3° C or at 26° C.

Biochemical changes. Does exposure to cold produce any measur-
able effects on the biochemistry of the host? Dr. Trapani has men-
tioned the increased thyroid activity of the hypothermic animal. This
is significant. Dr. Campbell has called our attention to the absence
of the beta anomaly in the descending electrophoretic pattern in the
serum of a man living in an arctic or subarctic climate. He also
mentioned the significance of the brown fat found in the hibernating
squirrel. Dr. Sulkin was particularly interested in this brown fat in
the hibernating bat. The question arises as to what role this peculiar
tissue with its high lipid content plays in hibernation or resistance
to the effects of cold.

Another interesting biochemical find has come to light in the work
of Monier and Weiss (1952). A sharp increase in the excretion of
ascorbic acid (53 per cent) and dehydroascorbic acid (186 per cent)
was noted inhypothermic animals over normal animals. Since ascor-
bic acid is found in large quantities if phagocytic cells and must be
present to a certain level in the phagocytes are to operate properly,
such losses of this essential vitamin must be compensated for, or
host resistance may be lowered through loss of ascorbic acid.

Other biochemical changes in h)T)othermic animals have been
noted which mayor may not be related to changes in host resistance.
For example, Ershoff (1951; 1952) noted a decreased resistance of
pyridoxine on riboflavin deficient rats which had been chilled. Ulti-
mately, we will better understand host-parasite relations when ade-
quate knowledge is available to explain such phenomena on a bio-
chemical basis. Dr. Metcalf 's report on the effects of temperature
on the neuraminidase of the influenza virus in embryonated e^s and
in the mouse lung represents such a biochemical approach to the
problem.

452

SUMMARY

Immunology. The effect of cold on circulating antibody was dis-
cussed by Dr. Campbell. Two factors which he pointed out that should
be kept in mind were 1) the effects of cold on antibody production and
antigen breakdown, and 2) the effects of cold on antibody removal
from the circulating blood.

In general, the participants reported a decrease in circulating an-
tibody in chilled animals. However, Dr. Northey found little difference.

Cold and microbialfloraof host. The effect of maintaining mice at
22° C to 23° C on the invasion of the blood stream and peritoneal
cavity, as reported by Dr. Tunevall, emphasizes how little we know
as to the permeability or impermeability of the gut to the passage of
macromolecules or microorganisms. This permeability might not be
of significance except that as Dr. Tunevall reports, clearance of
bacteria from the blood stream is also impaired in the hypothermic
animal. A basic explanation of both findings is in order.

Another mechanism by which cold may alter host microbial flora
is suggested by Dr .Schmidt's report onthe intestinal flora of squir-
rels as affected by hibernation. He found a sharp increase in psych-
rophilic bacteria.Such a change in intestinal flora, induced by hypo-
th'ermia,could be associated with changes in available vitamin K pro-
duced by the normal intestinal flora and more significantly, with an
increase or decrease of antimicrobial agents produced by intestinal
bacteria.

SUMMARY

In conclusion, it is safe to say that the final blue print showing
the anatomical physiological biochemical and immunological mech-
anisms by which exposure to cold alters host resistance to infection
will not be simple. But it will be a workable paper and will suggest
practical ways of increasing host resistance.

453

NUNGESTER
LITERATURE CITED

1. Culver, E. 19 59. Effects of cold on man. An annotated bibliography

1938-1951. Physiol. Rev. 39: 1-524.

2. Ershoff, B. H. 19 51. Decreased resistance of pyridoxine-deficient

rats to cold exposure. Proc. Soc. Exp, Biol, Med. 78: 385.

3. Ershoff, B. H. 1952. Decreased resistance of riboflavin-deficient

rats to cold stress. Proc. Soc. Exp, Biol. Med. 79: 559,

4. Goldstein, L. S. 1951. Cold weather as a factor in the epidemi-

ology of grippe and the common cold. Arch. Pediat, 68: 577.

5. Keplinger, M. L., G. E, Lanier, and W. B. Deichmann. 1959.

Effects of environmental temperature on the acute toxicity of
a number of compounds in rats. Toxicol, Appl, Pharmacol,
1: 156-161,

6. Monier, M. M., and R.J. Weiss. 1952. Increased excretion of de-

hydroascorbic and diketogulonic acids by rats in the cold.
Proc. Soc. Exp. Biol. Med. 80: 446.

7. Mudd, S., S. B. Grant, and A, Goldman. 1921. Etiology of acute

inflammations of nose, pharynx and tonsils. Ann. Otol. Rhinol,
Laryngol. 30: 1.

8. Nungester, W. J., J, K, Bosch, and D. Alonso. 1951, Resistance

lowering effect of human respiratory tract mucin, Proc. Soc,
Exp, Biol. Med. 76: 777-780.

9. Nungester, W. J., and L. F. Jourdonais. 1936. Mucin as an aid

in the experimental production of labor pneumonia. J. Infect.
Dis. 59: 258-265.

10. Nungester, W. J., L. F. Jourdonais, and A. A. Wolf. 1936. The

effect of mucin on infections by bacteria. J. Infect. Dis. 59:
11-21.

454

SUMMARY

11. Nungester, W. J,,andR.G. Klepser.1938. A possible mechanism

of lowered resistance to pneumonia. J. Infect. Dis. 63: 94-102.

12. Olitzki, L. 1948. Mucin as a resistance- lowering substance. Bac-

teriol. Rev. 12: 149.

13. Sulkin, S. E, 1945. Effect of environmental temperature on ex-

perimental influenza in mice, J, Immunol. 51: 191-305,

14. Troisi, F. M,, A, Amorati, and O. Bonazzi. 1953. Pathological

effects from work in cold environments. Industrial Med. and
Surgery 22: 11.

15. Washburn, Bradford. 1962, Frostbite - what it is - how to pre-

vent it - emergency treatment. New Eng. J. Med. 266: 974,

455 1*997-63

J