CIBA FOUNDATION COLLOQUIA ON AGEING Vol. 2. Ageing in Transient Tissues A leaflet giving details of available earlier volumes in this series, and also of the Ciba Foundation General Symposia and Colloquia on Ageing, is available from the Publishers. CIBA FOUNDATION COLLOQUIA ON AGEING VOLUME 2 Ageing in Transient Tissues Editors for the Ciba Foundation G. E. W. WOLSTENHOLME, O.B.E., M.A., M.B., B.Ch. and ELAINE C. P. MILLAR, A.H.W.C, A.R.I.C. With 96 Illustrations LITTLE, BROWN AND COMPANY BOSTON THE CIBA FOUNDATION for the Promotion of International Co-operation in Medical and Chemical Research 41 Portland Place, London, W.l. Trustees: The Right Hon. Lord Adrian, O.M., F.R.S. The Right Hon. Lord Beveridge, K.C.B., F.B.A. The Hon. Sir George Lloyd-Jacob Mr. Raymond Needham, Q.C. Director, and Secretary to the Executive Council: Dr. G. E. W. Wolstenholme, O.B.E. Assistant Secretary: Miss N. Bland Librarian: Miss Joan Etherington Editorial Assistants: Miss C. M. O'Connor, B.Sc. Miss E. C. P. Millar, A.H.W.C. All Rights Reserved This book may not be reproduced by any means, in whole or in part, with- out the permission of the Publishers Published in London by J. & A. Churchill Ltd. 104 Gloucester Place, W.l First published 1956 Printed in Great Britain PREFACE The Ciba Foundation, London, is an educational ^ah< scientific charity founded by a Trust Deed made in 1947. Its distinguished Trustees, who are wholly responsible for its administration, are The Rt. Hon. Lord Adrian, O.M., F.R.S.; The Rt. Hon. Lord Beveridge, K.C.B., F.B.A.; The Hon. Sir George Lloyd- Jacob; and Mr. Raymond Needham, Q.C. The financial support is provided by the world-wide chemical and pharmaceutical firm which has its headquarters in Basle, Switzerland. The Ciba Foundation forms an international centre where workers active in medical and chemical research are encouraged to meet informally to exchange ideas and information. It was opened by Sir Henry Dale, O.M., F.R.S., in June 1949. In the first six years, in addition to many part-day discus- sions, there have been 37 international symposia, each lasting two to four days, attended by outstanding workers from many countries. Other symposia are planned at the rate of five or six a year. Early in 1954 the Trustees decided on special measures designed to encourage, internationally, basic research relevant to the problems of ageing. As part of this programme the con- ference already arranged in that year on General Aspects of Ageing was taken to be the first in a series of colloquia, to be held annually. A second colloquium was organised in 1955, and at the welcome suggestion of Professor E. C. Amoroso, this one dealt with studies on the ageing of tissues, the normal life of which is shorter than that of the organism as a whole. It was hoped that inferences of more general value might be made from researches on the ageing of the placenta, the repro- ductive system, deer antlers, erythrocytes and so on. This volume contains papers and discussions of this second colloquium, on Ageing in Transient Tissues, at which Professor Amoroso graciously acted as Chairman. vi Preface The informality and intimacy of these meetings have per- mitted discussion of current and incomplete research and stimulated lively speculation and argument. They have also been the occasion for reference to much published and unpub- lished work throughout the world. The proceedings are issued in full, with only the minimum of editing, in order to pass on to a far wider audience the benefits of these meetings. It is hoped that readers will not only gain information and inspiration from this report, but will also feel that they share in these frank and friendly discussions. CONTENTS ^. . . . PAGE Chairman's opening remarks E. C. Amoroso ........ i Organ culture studies of foetal rat reproductive tracts by Dorothy Price and Richard Pannabecker . 3 Discussion: Amoroso, Corner, Jost, Parkes, Price, Villee, ZUCKERMAN ........ 13 The age factor in some prenatal endocrine events by A. Jost ........ 18 Discussion: Amoroso, Dawes, Huggett, Jost, Price, Row- lands, Villee, Wislocki, Zuckerman .... 27 The regenerative capacity of ovarian tissue by S. Zuckerman ....... 31 Discussion: Corner, Dempsey, Huggett, Krohn, Matthews, Parkes, Strauss, Williams, Zuckerman ... 54 The history and fate of redundant follicles by P. C. Williams ....... 59 Discussion: Amoroso, Corner, Dempsey, Harrison, Krohn, Parkes, Rowlands, Strauss, Williams, Zuckerman . 66 The corpus luteum of the guinea pig by I. W. Rowlands ....... 69 Discussion: Amoroso, Harrison, Huggett, Jost, Krohn, Matthews, Rowlands, Tuchmann-Duplessis . . 83 Observations on the cytomorphosis of the germinal and interstitial cells of the human testis by D. W. Fawcett and M. H. Burgos ... 86 Discussion: Fawcett, Montagna, Wislocki, Zuckerman . 96 Mitochondrial changes in different physiological states by E. W. Dempsey 100 Discussion: Dawes, Dempsey, Fawcett, Montagna, Wislocki 103 vii 73500 viii Contents PAGE Morphological aspects of ageing in the placenta by G. B. Wislocki 105 Discussion: Amoroso, Dawes, Hamilton, Harrison, Hug- GETT, JOST, VlLLEE, WlSLOCKI . . . . .114 Chronological changes in placental function by A. St. G. Huggett ...... 118 Discussion: Amoroso, Dawes, Huggett, Jost, Tuchmann- Duplessis, Zuckerman . . . . . .125 Biochemical evidence of ageing in the placenta by C. A. Villee 129 Discussion: Amoroso, Dempsey, Huggett, Jost, Villee, Williams, Wislocki, Yemm . . . ' . . . 144 Uptake of radio-potassium (42K) by the uterus and placenta during the advancement of pregnancy in the rat and the goat by R. J. Harrison and J. L. D'Silva .... 148 Discussion: Amoroso, Hamilton, Harrison, Huggett, Montagna, Strauss, Wislocki . . . . .159 Modifications in the foetal development of the rat after administration of growth hormone or cortisone to the mother by h. tuchmann-duplessis and lucette mercier- Parot 161 Discussion: Amoroso, Huggett, Jost, Montagna, Strauss, TUCHMANN-DUPLESSIS, WILLIAMS . . . . .173 The growth cycle of deer antlers by G. B. Wislocki 176 Discussion: Bourliere, Boyd, Dempsey, Huggett, Mat- thews, Strauss, Wislocki, Zuckerman . . .183 Ageing of the axillary apocrine sweat glands in the human female by W. Montagna ....... 188 Discussion: Amoroso, Harrison, Huggett, Medawar, Montagna, Strauss, Zuckerman . . . . .199 Contents ix PAGE The metabolism of senescent leaves by E. W. Yemm . . . . . t 202 Discussion: Bourliere, Huggett, Krohn, Montagna, Row- lands, Villee, Williams, Yemm ..... 210 The physical instability of human red blood cells and its possible importance in their senescence by J. E. Lovelock ....... 215 Ageing in human red cells by P. L. Mollison ....... 233 Discussion: Amoroso, Dempsey, Huggett, Krohn, Mollison Montagna, Parkes, Rowlands, Tuchmann-Duplessis, Villee, Williams, Wislocki, Yemm .... 239 General Discussion: Amoroso, Bourliere, Corner, Dempsey, Fawcett, Huggett, Jost, Krohn, Montagna, Parkes, Villee, Williams, Wislocki, Yemm .... 246 .V, • • « I I «i •> > >» 0 0 0 4 4 5 6 2 0 4(20 + ) 5(20 + ) 2(21+) with ovaries 15 + 16 + - 17 + 18 + 16+-18 + 19 + 19+-21 + 21+-22 + 4 2 4 3 regressed >> 0 0 0 3 3 2 2 2 0 0 2(20 + ) 2(21+) testo- sterone 15 + 16 + 17 + 16+-18 + 20 + 19+-22 + 9 4 6 persisted >> 1(18 + ) 2(20 + ) 6(19+) 3 4 5 0 4(20 + ) 4(19+) oestra- diol 15 + 16 + 17 + 18 + 17+-19 + 19+-20 + 20+ -21 + 21 +-22 + 3 4 7 4 regressed partial persistance >> 0 2(20 + ) 2(20 + ) 4(21+) 3 2 6 4 0 0 5(20 + ) 4(21+) The urogenital sinus was removed at explantation in some cases. Seminal vesicles were present in explants cultured at 18 + days ; some small prostatic buds occur in a few foetuses at 18 4- days but are normally present by 19 + . The figures in parentheses indicate the age of the youngest explant in which seminal vesicles or prostate glands were found. Organ Culture Studies of Reproductive Tracts 7 results were similar with a replaced testis and one left attached indicating that surgical removal of the testis did not cause Wolffian duct retrogression. In contrast, in tracts with no testes the Wolffian ducts underwent slow retrogression which involved successively, reduction in diameter, loss of the lumen and then of the epi- thelial component so that the ducts were discontinuous, and finally loss of the surrounding sheath of mesenchymal origin. Retrogression was mainly antero-posterior and by 3 to 4 days of culture the ducts were practically gone anteriorly, except for a small segment connected with the efferent ducts; the posterior segment persisted and joined the prostatic urethra. Testosterone in the clots (1 or 2 drops of 4(i.g./drop or 20(j.g./ drop) prevented retrogression of the ducts after removal of the testes. The results of culturing explants at the three older ages, with both testes on the tracts or without testes, were essenti- ally similar to the results with 15 + day-old explants. With both testes present the Wolffian ducts persisted and developed a slight epididymal coiling; without testes the ducts retro- gressed. An important difference however, was observed in tracts with only one testis (left in position or replaced). Although both Wolffian ducts usually persisted, there was a unilateral effect of the one testis if the tracts were spread in an open Y-shape on the clots. It was apparent that the diameter, and in some cases the continuity, of the gonadless Wolffian duct depended upon its distance from the testis. Fig. 1 presents the results diagrammatically for explants cultured at 17+ days. In the gonadless tracts cultured at all four ages, testosterone prevented retrogression of the Wolffian ducts. When ovaries were substituted for testes the ducts retrogressed as in gonad- less explants, and the same result was obtained in tracts cultured at 15+ days with oestradiol. At older explantation ages, oestradiol (1 or 2 drops of 0- 006 (xg. /drop in the clots) caused a partial retention and some cystic enlargement of the ducts in 5 out of 15 explants. 8 Dorothy Price and Richard Pannabecker days post coitum 17.5 -21.5 Fig. 1. Diagrammatic representation of reproductive tracts explanted at 17+ days. A: tract at the time of explantation ; B, C, D, E, F: tracts cultured for 4 days. B: with both testes present the seminal vesicles and prostate glands developed. C : without testes the Wolffian ducts retrogressed, no seminal vesicles appeared but a few prostatic buds developed. D : with one testis (left in place or detached and replaced) the results were as in B in 5 out of 6 explants. E: with one testis, placed at a greater distance from the opposite side of the tract, the Wolffian duct on the gonadless side retrogressed somewhat and the seminal vesicle was lacking or smaller in 8 out of 9 explants. F: with no testes but with testosterone micropellets added to the clot, the Wolffian ducts were retained and the seminal vesicles and prostatic buds developed as in B and D. M : Mullerian duct ; W : Wolffian duct ; S : seminal vesicles ; P : prostate. Seminal Vesicles The development of seminal vesicles proved to be related to the age of the tract at the time of explantation as well as to the presence of hormone. In explants with testes, no seminal vesicles developed in tracts cultured at 15+ days; large dilatations with only a slight indication of a free seminal vesicle tip appeared in 9 out of 24 explants cultured at 16 + days; well-developed glands were found in 18 out of 19 explants of 17+ day-old tracts cultured when dilatations were present in the Wolffian ducts (Fig. 2, A), and within four days of culture they reached a length, in the best cases, equal to A C - ^"''X'V 3*&V3 Fig. 2. A, D and F illustrate normal foetal development of the seminal vesicles: B, C, E and G were explanted at 17+ days and were cultured for 2 days (B) or 4 days (C, E, G). A: normal foetus at 17+ days with Wolffian duct dilatations and medial Mullerian ducts. B: 19+ day-old explant with testes, showing Wolffian duct dilatations and medial utriculus prostaticus. C : 21 + day-old explant without testes ; the Wolffian ducts have retrogressed and no seminal vesicles have developed. D: normal foetus at 19+ days; the large seminal vesicles are near their junction with the corresponding ductus deferens. E: 21+ day-old explant with no testes but with testosterone; ejaculatory duct on the left, ductus deferens and seminal vesicle on the right. F : normal foetus at 21 + days ; length of the seminal vesicles 1 • 1 mm. G : 21 + day-old explant with testes; length of the seminal vesicles 0 16 mm. ( X 90). facing page 8 Fig. 3. A : urogenital sinus region of a 19 + day-old foetus with several ventral prostatic buds and a single dorsal bud. B : same region of a 21 + day-old explant cultured for 4 days with testosterone ; dorsal and dorsolateral prostatic buds are indicated by arrows ( X 120). Organ Culture Studies of Reproductive Tracts 9 that of normal glands at about 19 days (Fig. 2, B, D, G) but no cranial flexures developed. Seminal vesicles w'ere'already developed in tracts explanted at 18+ days. A unilateral effect of one testis on seminal vesicle develop- ment was observed. In 26 such explants in which seminal vesicles developed, the glands were larger or present only on the testis side in 20 but were bilateral and equally developed in 5 explants in which the testis was near the gonadless Wolf- fian duct (Fig. 1). In explants with no testes or with ovaries, no seminal vesicles developed (16+ and 17+ days) or they remained as at explantation (18+ days). Testosterone stimulated the development of seminal vesicles and was more effective with advancing age of the tract at the time of explantation. As in explants with testes, only the tracts explanted at 17+ days were stimulated to produce well-developed seminal vesicles (Fig. 2, E). Oestradiol stimulated seminal vesicle development in 4 out of 11 explants cultured at 16+ and 17+ days. Prostate Glands Part of the urogenital sinus region and the urethra were removed from some reproductive tracts at explantation and when the sinus was explanted, it tended to undergo necrosis particularly in the region of Miiller's tubercle. Table I lists the number of well-preserved sinuses and summarizes the occurrence of prostate glands. The prostatic buds that developed were from the dorsal and dorsolateral regions and probably represented those parts of the prostate gland complex and possibly the coagulating glands (Fig. 3, B). The removal of the bladder at explantation disturbed the ventral part of the sinus and urethra and, although healing occurred, no prostatic buds were formed from that region. Only one prostate gland (of one or two short buds) developed in explants cultured at 15+ days and that with testes on the tract. At older ages, prostate glands de- veloped in all categories of explants. There was variability in 10 Dorothy Price and Richard Pannabecker the number and size of the buds but, in general, the prostatic buds were more numerous and larger in explants with testes or with testosterone. Discussion The results of the culture experiments indicate that a hor- mone from the foetal testes maintained the Wolffian ducts, stimulated the formation of the primordia of the seminal vesicles and their growth, and caused an increase in the num- ber of prostatic buds. These findings extend and corroborate some of the observations of Wells and his co-workers (Wells and Fralick 1951; Wells, Cavanaugh and Maxwell, 1954) in castrated rat foetuses. They concluded that testicular androgen stimulates the prenatal growth of male rat accessory reproductive organs but that they could not determine whether this androgen caused the development of the pri- mordia of these glands. They also observed absence of the flexure of the seminal vesicles and loss of the epithelium of the ductus deferens following castration and attributed this to two factors, the absence of the testes and impairment of deferential circulation due to surgical injury (Wells and Fralick, 1951). Testosterone treatment, however, prevented all of the observed effects of castration. In unilateral castration they found effects on the ductus deferens and the seminal vesicle on the operated side in some foetuses, but they con- sidered that these results were due to circulatory impairment and did not support the idea of local action of a diffusing testis hormone. Explantation of foetal rat reproductive tracts at ages from 15-5 to 18-5 days gave results that suggest that the retention of the Wolffian ducts and the development of seminal vesicle primordia may normally depend upon testis hormone. Under culture conditions, this hormone reached other parts of the tract by diffusion and produced local effects on Wolffian ducts and seminal vesicles in unilaterally castrated explants. Further, the results of detaching and replacing testes on the Organ Culture Studies of Reproductive Tracts 11 tracts indicate that surgical injury can be ruled out as a significant factor in general Wolffian duct retrogression. The question of the relation of testis hormone to the development of the prostate glands cannot be definitely- answered. Prostatic buds developed in gonadless tracts which had been explanted before the appearance of the primordia and this development was reported also for cas- trated foetuses (Wells, Cavanaugh and Maxwell, 1954). In culture, these buds formed in only one tract (with testes) which was explanted at 15+ days, but this result may repre- sent only slowed development. It seems probable that the prostatic region of the urethra is subjected to the influence of testis hormone in the foetus as early as 15+ days, and it may soon be conditioned to produce a few prostatic buds even in the absence of further hormone stimulation. The hormone, as a diffusible agent, would act to bring out an underlying organ- ization of the urethra. However, such an interpretation is complicated by the fact that prostate glands developed in foetal female tracts after explantation (Pannabecker, 1956). The conclusions as to the significance of foetal rat testis hormone in relation to the Wolffian ducts, seminal vesicles and prostate glands and the observation of local action of testis hormone are in agreement with the findings of Jost (1947, 1950, 1953) in his extensive studies on castrated foetal rabbits. His results showed, however, that prostatic buds did not develop when castration was performed at the youngest ages. He and his collaborators (Jost and Bergerard, 1949; Jost and Bozic, 1951) reported the retrogression of the Wolffian ducts in cultured fragments of gonaducts of foetal rats. Raynaud and Frilley (1946, 1947, 1950) reported that following X-ray destruction of the testes in mice, the accessory reproductive glands were smaller or absent and a unilateral effect was observed when only one testis was destroyed. The direct stimulating effects of oestradiol on the Wolffian ducts and on seminal vesicle development in a few gonadless explants is of interest, since it duplicates the results obtained in female rat foetuses when the mothers were injected with 12 Dorothy Price and Richard Pannabecker oestradiol or oestradiol dipropionate (Greene, Burrill and Ivy, 1939, 1940). In the males, the results were variable and the Wolffian ducts were normal in some foetuses but had retro- gressed in others; the seminal vesicles were smaller and pro- static buds were inhibited. These results suggest the possibility of an inhibiting effect of oestrogen on testis hormone pro- duction via the pathway of foetal hypophyseal gonadotropin and, possibly, a direct stimulating action of oestrogen on the male tract. However, hypophysectomy by decapitation in the foetal rat has not shown that the hypophysis produces gonadotrophin, which^is necessary for the secretion of testis hormone (Wells, 1947* 1950). Jost (1953, 1954) has reviewed the problem recently for several species. Summary and Conclusions The results of culture of foetal male reproductive tracts indicate that testis hormone 1) maintains the Wolffian ducts, 2) stimulates the development of the primordia of the seminal vesicles and their further growth and morphogenesis and 3) causes an increase in the number of prostatic buds. The testis hormone diffused through the explanted tracts and produced local effects. It is suggested that these observations may apply to normal sex differentiation in the rat and that foetal testis hormone may be effective by local diffusion as well as by circulatory pathways. Acknowledgement. This investigation was aided in part by Research Grant No. G. 2912 from the National Institutes of Health, Public Health Service, and by the Dr. Wallace C. and Clara A. Abbott Memorial Fund of the Univer- sity of Chicago. REFERENCES Fell, H. B. (1940). J. R. micr. Soc, 60, 95. Greene, R. R. (1942). Biol. Symp., 9, 105. Greene, R. R., Burrill, M. W., and Ivy, A. C. (1939). Anat. Rec.y 74, 429. Organ Culture Studies of Reproductive Tracts 13 Greene, R. R., Burrill, M. W., and Ivy, A. C. (1940). Amer. J. Anat., 67, 305. Jost, A. (1947). Arch. Anat. micr. Morph. exp., 36, 271. Jost, A. (1948). C. R. Soc. Biol., Paris, 142, 196. Jost, A. (1950). Arch. Anat. micr. Morph. exp., 39, 577. Jost, A. (1953). Recent Prog. Hormone Res., 8, 379. Jost, A. (1954). Cold Spr. Harb. Symp. quant. Biol., 19, 167. Jost, A., and Bergerard, Y. (1949). C. R. Soc. Biol., Paris, 143, 608. Jost, A., and Bozic, B. (1951). C. R. Soc. Biol., Paris, 145, 647. Jost, A., and Colonge, R. A. (1949). C. R. Soc. Biol., Paris, 143, 140. Moore, C. R. (1953). J. clin. Endocrin., 13, 330. Moore, C. R., and Price, D. (1942). J. exp. Zool., 90, 229. Pannabecker, R. F. (1956). Manuscript in preparation. Price, D. (1951). Endocrinology, 49, 672. Price, D. (1953). Schweiz. med. Wschr., 36, 835. Raynaud, A. (1942). Actualites sci. industr., Nos. 925 and 926, p. 1-463. Raynaud, A. (1950). Arch. Anat. micr. Morph. exp., 39, 518. Raynaud, A., and Frilley, M. (1946). C. R. Acad. Sci., Paris, 223, 1187. Raynaud, A., and Frilley, M. (1947). Ann. Endocr., Paris, 8, 400. Raynaud, A., and Frilley, M. (1950). Ann. Endocr., Paris, 11, 32. Wells, L. J. (1947). Anat. Rec, 97, 409. Wells, L. J. (1950). Arch. Anat. micr. Morph. exp., 39, 499. Wells, L. J., Cavanaugh, M. W., and Maxwell, E. L. (1954). Anat. Rec, 118, 109. Wells, L. J., and Fralick, R. L. (1951). Amer. J. Anat., 89, 63. DISCUSSION Jost: I was very much interested in Dr. Price's magnificent results and very glad, I must say, to see in what good agreement they are with some personal findings. It seems clear that in the absence of the testes, the Wolffian ducts do not persist and develop. This seems to be true in the rabbit, in the rat and most probably in man also, since it is well known that in cases of gonadal agenesis (so-called "ovarian agenesis") no Wolffian duct persists. Some years ago, we tried in my laboratory to confirm by experiments in vitro the results obtained in vivo in castrated rabbit foetuses. I should like to show you two slides taken from former papers (Jost, 1950; Jost and Bozic, 1951). Small pieces of the genital tract containing both the Wolffian and the Miillerian ducts (rat embryo on day 16£) were culti- vated in serum of castrated or non-castrated adult rats. In the next three days whatever the sex of the embryo the Wolffian ducts retro- gressed and disappeared while the Miillerian ducts developed and pro- liferated; so the ducts behave in vitro as they do in vivo in castrated rabbit foetuses. These results seem to be in good agreement with Dr. Price's more extensive experiments. But, Dr. Price, you did not discuss the point of the Miillerian ducts. 14 Discussion Did you observe any inhibition of the Miillerian ducts in explants with the testes? Price: I did not mention the Miillerian ducts because of lack of time. Our results on these ducts are still being studied and will be reported soon by Pannabecker. We found that explanted Miillerian ducts of both males and females were more difficult to culture at 15 \ and 16^ days than Wolffian ducts. They were actively growing and in a critical period of development at the time of explantation and they tended to become discontinuous under all our culture conditions. As I say, they had rather a hard time but they continued to develop and formed the prostatic utricle in males and the uterovaginal canal in females. When we explanted older tracts, the Miillerian ducts of males re- gressed and the Miillerian ducts of females were retained under all experimental conditions. The Miillerian ducts of females were stimu- lated, not inhibited, when foetal testes were placed against them or male hormone was put in the clot. However, I do not think that this indicates that there is no possibility that foetal testis hormone may inhibit the Miillerian ducts of males. Jost: Did you see any difference in the prostatic development of tracts which were taken on day 16i and those taken on day 17|? You showed that some prostatic buds appear, but the testis begins to work earlier than day 17£ and by that time the testis has been able to act on the urogenital sinus for about one day. I would also like to draw atten- tion to the fact that in some of these cultures the effect of the distance from the testes was very conspicuous ; I observed some years ago such a spatially restricted activity of the testis in vivo under certain circum- stances, a point which has sometimes been questioned, and I am very glad to see that the same thing occurs in vitro. Price: This question brings up the relation of development of pro- static buds to testis hormone. It appears that the testes are already secreting hormone in the foetus between 15| and 16£ days. A few prostatic buds appeared in male tracts explanted at 16^ days without testes, but more numerous buds developed in such tracts explanted at 17^ days. This corroborates some of your findings in foetally-castrated male rabbits. I agree that this indicates longer action of testis-secreted hormone on the urogenital sinus. I may say that in our cultures prostatic buds developed in explanted female tracts with ovaries present and without ovaries. Most of these tracts were from females of our stock — loosely called a strain — that has had a relatively high frequency of female prostate glands. If we accept the conclusions of Mahoney (1942) and Mahoney and Witschi (1947) that the determination of the female prostate in female prostate strains of rats is genetic and not hormonal, we do not need to assume that androgenic hormone is necessary for the development of all prostatic primordia in the rat. Later Witschi (1948) postulated that the deter- mination is hormonal — foetal ovarian androgen in the female and testicular androgen in the male. This may be so but I find it difficult to understand all the variability Discussion 15 in spontaneous occurrence of female prostatic primordia, including unilateral lobes, on the basis of levels of foetal ovarian androgenic hormone. Jost: Did you find any effect of gonadotrophs substance on such cultures or was there any indication of a reduced activity of the testes in cultures without gonadotrophic substance? Did you study the interstitial cells? Price: We did not add gonadotrophic hormone to the clot. We have no idea whether the testes continued to secrete their hormone during culture or whether we were just getting liberation and diffusion of a substance already present in the testes at the time of explanation. The interstitial cell situation is still being studied. Corner: As I see it, the action of the foetal hormone in these cultures was accurately imitated by testosterone, and this suggests, as far as it goes, that the foetal hormone is similar chemically or identical with that of the mature testis. Price: No, I do not believe that they are necessarily similar or identical. Zuckerman: In your reply to Prof. Jost, you suggested that you might be dealing with a "prostatic strain" of female rats. Do most of the females show prostatic rudiments? Price: Some years ago we established a stock of rats with an incidence of approximately 80 per cent of female prostate glands. We are no longer selecting and inbreeding and the incidence has fallen. There is quite a difference between the development of prostatic buds from the sinus in those females which develop the character and those which do not. However, in the latter a few prostatic buds usually appear, remain rudimentary and disappear postnatally, and this is true also in females of the non-prostate strain as reported by Mahoney. It seems possible that the sinus of rats may be capable of producing a certain number of buds even without androgenic stimulation. Our culture experiments have not answered this question. Zuckerman: Could one infer that the presence of prostatic tubules in the females of other species of mammal, for example, the green monkey, C. aethiops sabaeus, in which, as far as I recall, they are almost a con- stant feature, means that these creatures conform more to the explana- tion you have given for your "prostatic strain" than to Prof. Jost's explanation of the development of prostatic primordia? Price: I think that varying levels of androgenic hormone may not necessarily be the whole story in the development of prostatic primordia in females. Jost: Dr. Witschi's (1948) explanation is that, in the prostatic strain of rats, the ovary may produce an androgenic compound. It would be necessary to castrate female foetuses or to cultivate the sinus in very young animals to elucidate the question. Do you believe that the ovary or the adrenal gland may produce some androgenic compound ? Price: Yes. It is certainly true in postnatal rats, and prostate glands of males and females respond to such hormones. Jost: In some cases, for instance in the mole, the ovary contains a very large medullary part which is like a testis, and in such an animal it seems 16 Discussion likely that the ovary may produce some androgenic compound; and indeed all the females of the mole are more or less masculinized. That is a special case. Parkes: Prof. Jost, are you referring to the adult or the embryo in these moles? Jost: Both. The mole embryo was studied by Godet (1949). He ob- served that the female foetus in the mole normally exhibits signs of masculinization such as a prostate. When oestrogens are injected into the pregnant mother, the female foetus becomes more like the normal female in other species: for instance, the prostate fails to appear, and the Wolffian ducts disappear more rapidly. Under normal conditions the new born female mole to a certain extent resembles an experimentally masculinized female of the rat, for instance. Parkes: That is why I asked whether you were referring to the adult or the foetus. Jost: It was done on the foetus. Parkes: I imagine it can be taken as read that the adult ovary pro- duces a certain amount of androgen, but I have never been convinced that the adult mole shows any real signs of masculinization, apart from the queer changes in the anoestrous ovary. Jost: Some signs of masculinization were also reported by Godet in the adult mole at the level of the balano-preputial fold, which undergoes a seasonal cycle, Parkes: You mean masculinization appears when the ovary has gone back to normal? Amoroso: I think it should be emphasized that the morphological changes in the reproductive organs of the mole are so notoriously different from the comparable events in the rat that the greatest caution should be exercised when making comparisons between the two species. But apart from this I would like to ask Dr. Price whether in fact the Wolffian ducts may not themselves exert a profound influence on the development of the Mullerian ducts. It is known, for example, that in elasmobranch fishes the Mullerian ducts arise by longitudinal splitting of the Wolffian ducts and in mammals there is good supporting evidence that the Wolffian ducts and /or Wolffian tubules contribute elements to the growing tips of the Mullerian ducts. Price: Yes, I think there is a real point here. Gruenwald (1941) offers evidence for the chick and the human that the Wolffian duct may con- tribute cells to the Mullerian duct or alternatively that the Mullerian duct may split from the Wolffian duct. In some of our explants we observed that the posterior end of a Mullerian duct entered the sheath of its associated Wolffian duct and in a few cases a Mullerian duct actually joined a Wolffian duct. In the foetus the Mullerian ducts have not yet reached the urogenital sinus at 15^ days. In a few of the cultured tracts that were explanted at that age the ducts continued their growth to the sinus and joined it but growth was less marked in other explants. It is true that some of the variability may be attributable to morphological or physiological changes in the Wolffian ducts in culture. Discussion 17 Villee: Dr. Price, I am interested in your experiments in which the testis is replaced by synthetic hormones added to the culture, and I wonder if you can say anything about the quantity of hormone that is present in the clot? In some cases, the oestradiol you added will have affected organs normally under the control of testosterone and this is, in a sense, the obverse of some of my experiments on citrate metabolism in the placenta. In these I find that oestradiol is the most effective of the steroids I have tried. Testosterone does have a definite effect, but it is only about 1 /1000th as active as oestradiol in stimulating the placental isocitric dehydrogenase. Could one account for the effect of oestradiol, and for the variation when you added oestradiol to the clot, by assuming that testosterone and oestradiol have effects which are qualitatively similar but quantitatively different? Price: We put one or two drops of saline suspensions of micropellets in the clots. The testosterone suspensions were 4 [i.g./drop and 20 \ig. /drop and the oestradiol, 0-006 [i.g. /drop. With regard to the variability in the stimulating effects of oestradiol on Wolffian ducts I doubt that chance local distribution of the hormone is responsible. Greene and his collaborators found much the same variability in foetal females when large doses of oestrogens were in- jected into the pregnant mothers. I do not think that this is a significant finding as far as normal sex differentiation goes. Villee: That would depend on whether these early organ anlage can produce hormones typical of the opposite sex. Price: That is true. REFERENCES Godet, R. (1949). Bull, biol., 83, 25. Gruenwald, P. (1941). Anat. Rec., 81, 1. Jost, A. (1950). Arch. Anat. micr. Morph. exp., 39, 577. Jost, A., and Bozic, B. (1951). C. R. Soc. Biol, Paris, 145, 647. Mahoney, J. J. (1942). J. exp. Zool, 90, 413. Mahoney, J. J., and Witschi, E. (1947). Genetics, 32, 369. Witschi, E. (1948). Ann. Endocr., Paris, 9, 385. THE AGE FACTOR IN SOME PRENATAL ENDOCRINE EVENTS Alfred Jost Laboratoire de Biologie animale, P. C. B. Faculte des Sciences, Universite de Paris It is evident that age is an essential factor in the study of the foetal endocrine correlations, since the endocrines cannot be expected to play any physiological part before they reach an adequate stage of differentiation. However, physiological specialization does not necessarily parallel morphological organogenesis. A priori three points of view about the foetal endocrine glands may be considered and were effectively sus- tained: 1) the endocrines have no physiological importance at all for the developing foetus; such a view is now difficult to maintain, at least for some glands ; 2) they progressively acquire a state of physiological activity which enables them to work immediately at birth; this interpretation often, and more or less implicitly, involves a similarity between prenatal and post- natal endocrine processes; 3) finally, one may assume that foetal endocrinology implies characteristic events occurring at certain stages of foetal life, and which can be distinct from postnatal function, even if they prepare postnatal life. In connection with this last view it should be recalled that some organs pass through a limited phase of sensitivity to hormones. For instance, masculinization of the female uro- genital sinus by testosterone can be produced only during a limited period of time (Turner, 1940; Moore, 1945). Moreover, hormonal treatments may become teratogenic if applied at certain stages: pitressin injected into the rat foetus produces haemorrhages, necrosis and, finally, congenital amputations of the extremities when administered before day nineteen and not afterwards (Jost, 1953c). In recent unpublished observa- tions, I noticed that cortisone injected into the abdominal cavity of rat foetuses (1-3 mg. per foetus) produced cleft 18 Age Factor in Some Prenatal Endocrine Events 19 palates only if given before day sixteen (this effect of cortisone was discovered on mice by Fraser and Fainstat, 1951). In such cases the response of the foetal organism to the same treatment is strictly correlated with its developmental age. In the normal developing foetus, variations in the response of the same target organ to the same endocrine gland may also occur according to the age. The efficiency of the endocrine factor then becomes maximal at certain stages, and it seems not unlikely that the hormonal release might also reach a maximum at the same stages. Some facts presented in this paper will show to what extent such speculations may be suggested by experimental data ; but no effort will be made to give a complete review of the field. Testicular function and sexual structures The testis is responsible for the development of the mas- culine sexual structures, as was observed in castration experi- ments on the rabbit foetus (Jost, 1947, 1953a); the importance of the age factor in these experiments will be recalled in con- nection with two structures. The Wolffian, or mesonephric, ducts exist in both sexes at early stages; in females they persist until day twenty-four when they begin to regress ; in males they develop as deferent ducts and seminal vesicles under the stimulation of the testis. They regress in males also, if castration is performed before day twenty-two ; after castrating on day twenty-three only a small seminal vesicle develops, and after castrating on day twenty-four the Wolffian ducts persist and differentiate along their whole length. They have been "stabilized" for the re- mainder of their life ; from transitory urinary ducts they have developed into definitive male genital ducts. Under the in- fluence of the testis, something happens between about days twenty-two and twenty-four which changes the characteristics of the ducts and which enables them to differentiate as deferent duct and seminal vesicle, even in the absence of the testis. However, at that stage and still for a while, these ducts cannot 20 Alfred Jost be stimulated to glandular activity as in the adult, even by- large doses of testosterone. The urogenital sinus provides another interesting example. This part is identical in young male and female embryos. Prostatic buds appear in the male if the testis has acted on the undifferentiated tissue for a sufficient period of time. Early castration, on day nineteen, completely prevents the forma- tion of the buds; castration performed one or two days later permits the development of two straight unbranched buds; castrating two days later still, on day twenty-three, when only mere anlage have appeared, does not stop prostatic growth. The buds continue to differentiate, although generally in a somewhat reduced manner in comparison with control animals. During the short period of time involved, the testis has completely changed the properties of a special part of the urogenital sinus, which has become the prostatic region. It would be of great importance to know what happened in these cells, what kind of change they underwent. After day twenty-four the male rabbit foetus is definitely marked as a male ; the sexual structures have passed a turning- point after which removal of the testis does not inhibit further differentiation. Even if the testicular function of the normal male were reduced at late stages of pregnancy, the sexual organogenesis would be normal: something similar seems to happen in the human foetus, in which the testicular intestitial cells are very large and rich in cytoplasm from the stage of 3 cm. to a stage of about 15 cm., after which they almost disappear (Gillman, 1948); masculine organogenesis was already complete. The age factor in pituitary function Research into hypophyseal function in the rabbit foetus was initiated in 1947, in order to verify whether the pituitary gland controls testicular activity. The foetuses were deprived of their hypophyses by the decapitation method, the value of which has already been discussed (Jost, 1951). Age Factor in Some Prenatal Endocrine Events 21 Briefly, it was noticed (Jost, 1951) that in the foetuses which were decapitated before the initiation of somatic sexual differentiation, testicular activity remained weak: on day twenty-eight the prostate was as reduced as in foetuses castrated on day twenty-one, and likewise the external geni- talia were feminine. The importance of the age factor in hypophyseal function was studied in two series of foetuses ; some were decapitated early and killed at one-day intervals thereafter, others were decapitated at one-day intervals and killed at the same final stage (day twenty-eight). The analysis of the results indicates that the testis has some activity even in the absence of the pituitary gland. It works in a normal manner until approxi- mately day twenty-two as judged from the first steps of differ- entiation of the genital tract; afterwards, mainly between day twenty-two to twenty-four, the pituitary gland is needed for normal testicular function; after day twenty-four it can be removed without suppressing masculine organogenesis. Relationship between hypophysis and thyroid also indi- cated that, until the twenty-first or twenty-second day, the thyroid differentiation and specialization proceeds normally in the headless foetuses; afterwards the pituitary gland be- comes necessary (Jost, 1953a, 19536). On the adult animal, the McManus technique (periodic acid - Schiff ) has been shown to give indications about those pituitary cells which produce gonadotrophic or thyrotrophic hormone. Therefore this procedure was applied to the hypo- physis of the rabbit foetus (Jost and Gonse, 1953). McManus- positive material accumulates in an increasing number of pituitary cells from day nineteen until day twenty-two to twenty-three and then diminishes in a significant manner; this last point will be illustrated here only by one case con- cerning litter mates, two of which were studied on day twenty-three and two others on day twenty-eight. The average numbers per section of distinctly stained cells, counted in three medial sagittal sections, distant 125jjl, were respectively 232 and 263 on day twenty-three in comparison with 65 and 77 22 Alfred Jost on day twenty-eight. It should be noted that the area of the sections almost doubles in the same time, which reduces the number of positive cells per unit of surface area in the oldest group. Such facts, verified on a rather large series of foetuses, strongly suggest that the activity of the pituitary gland passes through a maximum at a stage at which the testes and the thyroid request hypophyseal stimulation. Then the pituitary gland would not progressively increase its physiological work in parallel with the progress of its gross anatomical differen- tiation, but would release a larger amount of hormone during a limited period of time. The incidence of the age factor on the relationship between hypophysis and adrenal cortex was studied with Miss A. Cohen on a small series of foetuses and some preliminary indications were obtained which require verification. The rabbit adrenal cortex provides a less favourable material than that of the rat, its structure being less clearly denned and more variable. On day twenty-eight the cortex shows an external zone organized in large arcads and rather irregular internal cords, which centrally more or less penetrate into the medulla; other cortical cells are mixed with the medul- lary cells. In foetuses decapitated before day twenty-one or twenty-two and studied on day twenty-seven or twenty- eight the zone of arcads is relatively larger and the internal part markedly reduced. The ratio, width of the arcad zone to width of the internal part, is definitely in favour of the arcads, while in controls the internal part largely exceeds the zone of arcads. On the contrary, in foetuses decapitated on day twenty-six and studied on day twenty-nine, the internal part remained large. These preliminary observations seem to indicate that decapitating the foetus before the cortex is organized terminates in a more marked reduction than decapitating at later stages. In the rat foetus the adrenal cortex grows markedly from day sixteen to day twenty-one; at about the age of nineteen days, modifications in the lipid content of the cortex suggest changes — probably an increase — in the physiological activity Age Factor in Some Prenatal Endocrine Events 23 of the gland (Cohen, 1954). Decapitation introduces a striking underdevelopment of the cortex (see Jost, 1954). A reduction in size of the cortex and some shrinkage of the cells occurs at birth (Josimovich, Ladman and Deane, 1954); alteration in the lipid content was observed by Miss Cohen (1955, unpublished data) in a few foetuses which had not been delivered at 22J days of age. It might be wondered whether such changes are correlated with a reduction in the adrenocorticotrophic activity of the pituitary gland at the time of birth. Glycogen storage in the liver It is almost a hundred years since the discovery by Claude Bernard that the foetal liver stores appreciable amounts t~ 23 tk 25 26 27 28 23 Days Fig. 1. Liver glycogen content expressed in mg./g. of fresh tissue; black points represent control foetuses; crosses represent foetuses decapitated between days 19 and 24 and killed on days 26 to 28; triangle represents foetuses decapitated on day 26 and killed on day 29. of glycogen from a certain stage onwards when suddenly accumulation begins. In the rabbit foetus the turning- point appears on day twenty-five (Lochhead and Cramer, 1908, and Fig. 1). Although the question has not yet been 24 Alfred Jost submitted to biochemical analysis, one may wonder whether at that stage the enzymic assortment of the liver cells under- goes a change and becomes adequate for storing glycogen. Studies were undertaken to determine whether the change in the liver glycogen storage is hormonally controlled (see Jost, 1954). Measurements were made on decapitated rabbit and rat foetuses. Only the former will be considered here. It was first observed that in rabbit foetuses decapitated on day nineteen, the glycogen content was very low on day twenty- eight in comparison with litter mate controls (respectively 1-58 ±0-85 mg./g. and 20 -06 ±3 -33 mg./g. of fresh tissue; Jost and Hatey, 1949). The same result is also observed in all foetuses decapitated before or on day twenty-four and killed on days twenty-six to twenty-eight (Fig. 1 ; Jost and Jacquot, 1954, 1955). In attempting to discover what hormonal factor was responsible for glycogen storage, different hormones were administered at the moment of decapitation. The detailed results will be presented elsewhere (Jost and Jacquot, 1955); briefly, it was observed that ACTH (Organon), in a dose of 2 to 8 mg. inserted under the skin, permits glycogen de- position in the decapitates, either at a normal or at a sub- normal level, in fourteen out of sixteen treated foetuses. Two of these foetuses were thyroidectomized-decapitated, and they also gave positive results. An attempt was then made to obtain glycogen storage with corticoid hormones: no positive result was obtained with cortisone (3*5 mg.), hydro- cortisone (3-7 or 4-1 mg.), desoxycorticosterone acetate (2 mg.), or 9-a-fluorohydrocortisone (25 y.g.; high doses seem to be lethal), nor with the association of cortisone or hydro- cortisone with DOCA or with 1 u. of insulin, or with 50 to 90 (jig. of thyroxine nor with a small amount of growth hormone. Although only a small number of experiments have been carried out in each case, it is curious to notice that no other hormone than ACTH was found to permit glycogen deposition. If the adrenal glands are involved, one cannot discard the possibility that they produce some other com- Age Factor in Some Prenatal Endocrine Events 25 pound than those which were tried, or at other levels. In any case, the fact that giving ACTH can restore the physiological condition in decapitates allows the supposition that the glycogen storage in the liver is hormonally controlled. It then would appear that endocrine factors introduce a turning-point in the carbohydrate metabolism, as they do in the properties of several tissues. In an effort to analyse the significance of this turning-point, it should be established to what extent the glycogen storage system has been changed by the first initial hormonal impulse. Attempts were made to study this point by decapitating after day twenty-five. In four foetuses decapitated on day twenty-six and norm- ally grown up to day twenty-nine, the mean value of the liver glycogen was 6-1 mg./g. (range: 3 to 8-7), while in controls the value was about 30 mg./g. ; in eight 26-day-old control foetuses the mean was 5-1 mg./g. (range 2 to 7-35) (Fig. 1). So it appears that decapitating the foetus on day twenty-six stops the rapid increase in glycogen storage norm- ally occurring after this day, but that the liver continues to store new glycogen at about the level to which it was brought before decapitation, since the values are expressed per g. of fresh tissue and since the liver continues to grow. Further study is necessary to solve the question under dis- cussion exactly. It should also be determined whether a liver, which has not been stimulated at the right time during foetal life, remains able to respond to delayed stimulation. Such problems involve implications for further postnatal life. Conclusions The aim of this informal paper was to draw attention to and perhaps to raise discussion on some suggestions resulting from experiments. Although many of the discussed problems are not yet solved, it seems that a pituitary-controlled hor- monal impulse introduces a turning-point in the further pro- perties of some tissues : the case of the sexual structures must be held as conclusive from this point of view. Other tissues, 26 Alfred Jost such as the liver, also pass through a physiological turning- point, but the importance of this change for later postnatal life is not yet established. The study of the pituitary gland in the rabbit or of the testicular interstitial cells in man, for instance, strongly suggests that such endocrine glands could pass through a period of maximal working at certain prenatal stages, fol- lowed by a period of lesser activity. It would then appear that the normal development of the foetus involves distinct endo- crine events at appropriate stages. Finally, although only one example of an actually transient tissue was considered in this paper, namely the Wolffian duct, the points which were discussed cannot be considered as too far from the main object of this colloquium, if the idea will be accepted that the term " foetus " does not define a particular state of an animal but covers a succession of different transient stages leading up to each other. Acknowledgement. We are indebted and wish to extend our thanks to several Labora- tories which kindly supplied the homones used in these researches: Dr. Choay (Choay Laboratory) for growth hormone and insulin; Dr. Tausk (Organon, Holland) for ACTH; Dr. Velluz (Roussel— UCLAF), for cortisone, hydrocortisone, desoxycorticosterone acetate; Dr. Borman (Squibb) for 9-a-fluorohydrocortisone. REFERENCES Cohen, A. (1954). C. R. Soc. Biol, Paris, 148, 321. Fraser, F. C, and Fainstat, T. D. (1951). Pediatrics, Springfield, 8, 527. Gillman, J. (1948). Carnegie Inst. Contrib. Embryol., 32, 81. Josimovich, J. B., Ladman, A. J., and Deane, H. W. (1954). Endo- crinology, 54, 627. Jost, A. (1947). Arch. Anat. micr. Morph. exp., 36, 271. Jost, A. (1951). Arch. Anat. micr. Morph. exp. 40, 247. Jost, A. (1953a). Recent Prog. Hormone Res., 8, 379. Jost, A. (19536). Arch. Anat. micr. Morph. exp., 42, 168. Jost, A. (1953c). Arch, franc. Pediat., 10, 855. Jost, A. (1954). Cold Spr. Harb. Symp. quant. Biol., 19, 167. Jost, A., and Gonse, P. (1953). Arch. Anat. micr. Morph. exp., 42, 243. Age Factor in Some Prenatal Endocrine Events 27 Jost, A., and Hatey, J. (1949). C. R. Soc. Biol., Paris, 143, 146. Jost, A., and Jacquot, R. (1954). C. R. Acad. Sci., Paris, 239, 98. Jost, A., and Jacquot, R. (1955). Ann. Endocr., Paris, 16, 849. Lochhead, J., and Cramer, W. (1908). Proc. roy. Soc. B, 80, 263. Moore, C. R. (1945). Amer. J. Anat., 76, 1. Turner, C. D. (1940). J. exp. Zool., 83, 1. DISCUSSION Amoroso: It is evident from Professor Jost's account that the effects of decapitation of the foetus are many and varied, and that by utilizing this technique — so familiar in his country — he has contributed much to our understanding of the endocrine climate of the foetus. He has raised a number of interesting points, not the least important of which is the self-perpetuating nature of the urogenital system in the absence of the testes after a certain stage in development has been reached. He has also called attention to the sundry histochemical reactions that he employs as evidence in support of his thesis. It seems reasonable therefore that Professor Wislocki should open the discussion. Wislocki: There is a point I should like to raise with regard to Prof. Jost's report. It concerns the changes he noted in glycogen storage and synthesis in the liver, following decapitation. Claude Bernard, in 1859, first proposed that the placenta is a deputy for the foetal liver until such time as the latter becomes physiologically capable of forming and storing glycogen. How the placenta gives up this function I shall demonstrate tomorrow in several lantern slides which will show the disappearance of glycogen from the rat's placenta in relationship to its appearance in the foetal liver. It would be interesting to know, with respect to the dramatic change which you have demonstrated in the liver following decapitation, whether placental glycogen persists, while its appearance in the liver is postponed? Is the placenta also a target organ of the pituitary? If it is not, does it then maintain its function of glycogen storage following decapitation? Jost: We examined this question with Jacquot, but only a few cases were studied. One difficulty with this study is the high variability of the glycogen content of the maternal placenta from one individual to another at the stage we studied (28th day). But as far as we could see, we observed no clear change from the normal in the placenta of decapi- tates. This means that the placenta of the decapitates did not contain more glycogen than that of controls, despite the low glycogen content of the liver. Likewise the effect on the placenta of high doses of hydrocortisone given to the foetus was not too evident, since the placenta of the treated foetuses did not very clearly exceed the high range of individual variability. Have you any information on that matter, Prof. Huggett? Huggett: With regard to controlling glycogen in rats and rabbits chemically, Prof. Jost is quite correct in saying that there is some factor causing considerable variation. But if you take a number of animals and 28 Discussion if you take the average of a number of any age there is a very definite trend; that is to say, the highest values for glycogen in the placenta are achieved round about the 21st day in the rabbit, and there are very few high values late in pregnancy and in the earliest state, about the 15th day, it begins to appear. Price: I should like to ask Prof. Jost a question relative to the McManus-positive cells which he demonstrated in the pituitary. The material is very abundant at a certain time and disappears relatively rapidly at a critical period when you say testicular hormone is needed. Do you think there is a special release mechanism which is causing the discharge from the pituitary at that time, or do you think that the pituitary reaches a certain stage and then releases automatically? Jost: I cannot explain exactly what happens. I observed changes in the pituitary gland which are concomitant with the need of pituitary hormone for the testis. One answer to your question may be the fact that the same changes occur in the pituitary gland of the female and of the male foetus. The ovary does not seem to need gonadotrophic hormone. The pituitary gland could perhaps function in a self-differen- tiating manner. Zuckerman: Prof. Jost, in elaborating your thesis about turning- points in the "determination" of tissues that are conditioned by pitui- tary hormones, it occurred to me that you might be able to alter their time relations by giving gonadotrophin before the critical days on which the pituitary hormones act (22 to 24). Supposing you injected gonadotrophic hormone into a foetus on day 18 or day 16, would you not accelerate the stage at which masculine organogenesis would proceed normally in the absence of the pituitary? Jost: I believe that it might be difficult to expect some action of hormones given too early to the foetus. As far as I know, experiments done with testosterone, for instance in the opossum or in other animals, showed that the prostate area is able to respond only at the time at which normally the prostatic buds appear in the male. So trying to activate this part of the sinus beforehand was not a success. Zuckerman: Yes, but I was thinking particularly of the pituitary. Why do you not give gonadotrophic hormone on day 20 or 18, and then move these critical two days forwards? This would surely be the test of your hypothesis. Villee: In any such system you have two parts — whatever is producing the hormone, the pituitary for example, and also the cells which are to be stimulated. And Prof. Jost's point, which I think is quite good, is that probably the ability of the cells to respond is the limiting factor here. If you did Prof. Zuckerman's experiment and got a negative result, it still would not disprove Prof. Jost's thesis. There are many examples from the field of amphibian embryology in which there are critical times for response, and you can do what you will before that time and nothing happens. Jost: The phase of sensitivity exists in the receptor and you must respect it. Zuckerman: Accepting all this, I am still wondering why in pursuing Discussion 29 the experimental analysis of the temporal phases of development you do not attempt to derange the whole system in time. In fact, we do not know a priori that the cells of some target tissue mature at a given stage and that they can only react at a given point to some endocrine stimu- lation. This constitutes deduction from experimental results. Huggett: A positive result from Prof. Zuckerman's experiment would be very interesting; a negative result would not disprove anything else Jost: Concerning the physiology of the liver, the point raised by Prof. Zuckerman can be studied on decapitated foetuses given ACTH. Is ACTH able to induce the liver to store glycogen earlier than usual? In a short series of cases we did not observe anything like that. Is ACTH still able to act at late stages on the liver of early decapitated foetuses which did not store glycogen in the right time? We have only a few trial experiments which are not yet conclusive. Such experiments still have to be done. Rowlands: What are the gonadal effects of decapitation at this stage? Are there any changes that would suggest the action of a gonadotrophin ? Jost: In the testes, the interstitial cells do develop but on day 28 they are reduced in decapitates and enlarged when gonadotrophic hormone is given. Moreover, the genital tract is a good indicator of their activity; before day 22 or 24 it develops normally in decapitates. Huggett: Prof. Jost, have you controlled your decapitation by merely transecting the cord leaving the head on ? It seems to me that you have not absolutely proven that this might not be a neural effect. Jost: I tried not exactly that experiment but some other ones. I tried to graft a pituitary gland onto a decapitated foetus and I studied the genital tract. I had only four cases of such grafts on males. In these four cases, I had one positive case in which the decapitated foetus had a normal genital tract, another case which was doubtful, and two cases which were negative. So I should not like to give too much importance to these preliminary observations, but I feel that if it were possible to replace the whole head by a grafted pituitary gland the demonstration would be good enough. Huggett: A series of positive results would be interesting; negative re- sults might be defective surgery. But I would like to see what happened if, instead of actually decapitating, you inserted, say an iridectomy knife, and just transected the cord level with the first cervical, the normal place for spinalizing an animal, sew the wound up and return the total foetus instead of the decapitated trunk. Jost: This is an interesting suggestion. In the past I tried another experiment of this type, by removing the brain and the bulb and leaving the pituitary gland in place. The result was not a success because the pituitary gland was ill-developed, although not completely necrosed. Dawes: I was going to ask about the nervous system from the point of view of the circulation. I would not expect any very great alteration if the head was removed up to half-way through pregnancy. I do not know what happens in rabbits, but in sheep the nervous control of the circulation becomes increasingly important towards the end of preg- nancy, and I wondered whether some of the results you obtain towards 30 Discussion the end of pregnancy in the rabbit might not be a result of alterations of the internal environment in that way. That was one point and the other was this. Ashton, Ward and Serpell (1953)* have shown that, at certain stages of development of the retinal vessels, constriction caused by excess oxygen leads to obliteration from mal-development in newborn kittens. Perhaps if you were to inject pitressin into foetal rabbits at a certain stage of development, their fingers might fall off or fail to develop, simply because at that critical stage there may be con- striction and hence clotting in the peripheral vessels ; that is, simply a vascular effect, and not what we usually think of as a hormonal effect upon cells. Jost: First, concerning the effect of pitressin or other hormones pro- ducing necrosis and congenital amputations of the extremities of the foetus : the final explanation of the haemorrhages involved has still to be found. A careful study of the vessels and measurement of blood pressure is needed. Finally, the technique of decapitation should not be considered as the final ne plus ultra technique in foetal endocrinology. It was useful and still remains of interest since it put on an experimental status questions which were discussed before in a more or less speculative manner. Now we know that correlations exist between the head and several endo- crines. It has not yet been absolutely demonstrated that the nervous system is not involved, but the fact that the appropriate pituitary hormone may replace the head indicates at least that the nervous system is not indispensable. The concomitant changes noted in the pituitary gland and in the other endocrine glands afford another argu- ment in favour of the hormonal interpretation. * Ashton, N., Ward, B., and Serpell, G. (1953). Brit. J. Ophthal, 37, 513. THE REGENERATIVE CAPACITY OF OVARIAN TISSUE S. ZUCKERMAN Department of Anatomy, University of Birmingham If the whole of one ovary and all but a piece of the second are removed from the body, the fragment that remains will hypertrophy, and its volume may ultimately equal that of two normal ovaries. If— to turn to another experiment— a minute quantity of suitable gonadotrophin is injected into a normal female animal, its ovaries will become transformed into a mass of luteal tissue. These two changes are illustrative of several which point to the conclusion that the ovary is not only a very plastic structure, but also one with considerable powers of cellular transformation and regeneration. On the other hand, the conspicuous changes it can undergo give a misleading impression about two central features of ovarian physiology— the capacity of the mature ovary to generate oocytes, and the dependence of the organ's secreting powers on the normal ovarian structure. It is to these two issues, which are critical to ideas about regeneration, that the present paper is addressed. The formation of Oocytes The answer to the question whether neoformation of oocytes occurs, or can be stimulated, in the adult mammalian ovary, depends on histological, cytological, experimental and bio- metric evidence. In an earlier review (Zuckerman, 1951) I considered the merits of the different types of observation which have been brought together to support the conflicting views that oogenesis does occur, or does not occur in the adult ovary, and concluded that the belief that oogenesis continues 31 32 S. ZUCKERMAN throughout reproductive life is very insecurely based. Since then, three sets of papers have appeared, stating the opposite view (see p. 36), and it is necessary to see how the work done in my own laboratory stands in the light of these newer studies. The more important of our own observations, all consistent with the view that the female mammal usually begins her reproductive life with ovaries that are already furnished with a finite stock of oocytes, are the following : 1. The total number of oocytes declines sharply with age (rats: Mandl and Zuckerman, 1951a; monkeys: Green and Zuckerman, 1951, 1954). On this point our observations merely confirm the careful studies of the rat carried out by Arai (1920a), as well as the impressions obtained by a number of other workers in some- what sketchy studies of other species. 2. Oocytes without any cellular cover, or surrounded by a single layer of cells, make up at least 90 per cent of those present in the monkey and rat (Green, Mandl and Zuckerman, 1951). Given that proper care is taken to enumerate all oocytes, and that account is also taken of age and litter- relationships, no evidence can be found that the total number of oocytes fluctuates with the phases of the oestrous or menstrual cycle (rat: Mandl and Zuckerman, 1950; monkey: Green and Zuckerman, 1951, 1954). The contrary view, which has been taken to mean that there is a cyclical generation of oocytes in the mature ovary, has been mainly based on impressions gained from the variations that occur during the course of the menstrual and oestrous cycle, not in the total population of oocytes, but in the num- bers of, and in the incidence of atresia in, large and medium- sized follicles. We have confirmed by numerical study that the latter changes do occur, but have also shown that no value can be attached to views about the neoformation of oocytes in the adult ovary, or about cyclical variations in the intensity of oogenesis, that are based on counts from which the primor- dial germ cells are excluded. The same point is illustrated by Regenerative Capacity of Ovarian Tissue 33 certain observations reported by Simpson and van Wagenen (1953), and in which purified FSH was administered to im- mature rhesus monkeys. Histological preparations of the ovaries gave the impression that this treatment stimulated oogenesis in all its phases, and that "small ovocytes" were formed in the germinal cords. Subsequent counts showed, however, that the total number of oocytes present after treatment was much the same as in the ovaries of normal monkeys of comparable age. In a further set of experiments, in which FSH was administered to unilaterally-ovariecto- mized immature monkeys, it was found that the numbers of oocytes in the hormone-stimulated ovary was of the same order as in the ovary that was removed before treatment began. 3. The number of oocytes does not increase in successful autografts of ovarian tissue, reimplanted after freezing at very low temperatures (— 190° C) or after having been kept at normal room temperature (Green, Smith and Zuckerman, 1956). To the best of my knowledge, this study is the only one so far carried out in which the numbers of oocytes were counted in ovarian grafts. The conclusion to which it points is in line with what is said below about fragments of ovarian tissue left in the body in situ. 4. Oocytes can survive for 2 J months in ovarian homo- grafts in rats, and for a year or more in autografts in monkeys, in the absence of the germinal epithelium (Breward and Zuck- erman, 1949; Mandl and Zuckerman, 1949). Here, our observations merely confirm a number of earlier ones made by other workers (e.g. Herlitzka, 1900; Marshall and Jolly, 1907; Pettinari, 1928). The indication is that oocytes may have a very long life. 5. The total number of oocytes in the one ovary of rats from which all traces of germinal epithelium had been removed by means of the application of corrosive fluids was not significantly different, over a period of nearly 1 J years, from that in the un- treated normal ovaries of the same animals. Correspondingly, AGEING VOL. 2 3 34 S. ZUCKERMAN the total number of follicles in both ovaries of the treated rats was only slightly lower than the expected number in normal rats of the same age (Mandl and Zuckerman, 19516). The results of this experiment extend a corresponding study reported by Moore and Wang (1947), and the conclusion to which both point is that the cellular division which may be observed in the germinal epithelium of adult ovaries bears no necessary relation to the process of oogenesis. 6. X-irradiation of rats and mice leads to the disappear- ance of all oocytes from an ovary, without at the same time causing any definable histological or cytological change in the germinal epithelium (Humphreys and Zuckerman, 1954; Mandl and Zuckerman, 1956a, b). These findings confirm many earlier observations (e.g. Lacassagne, 1913; Brambell and Parkes, 1927; Everett, 1943); Everett believes that the germinal epithelium of the mature mouse, the animal on which he experimented, consists of cells which are somatic in origin, as well as primordial germ cells that originate in the gut entoderm, and which are set aside during early embryonic development. Without provid- ing any cytological evidence to support his view, he suggests that X-irradiation destroys the germinal epithelium without affecting the somatic elements. 7. Compensatory hypertrophy of an ovary is not associated with an increase in the total number of oocytes, but the single hypertrophied ovary contains as many follicles with antra as do the two ovaries of a normal animal (Mandl and Zucker- man, 1951c). These findings confirm Arai's (1920&) earlier study of com- pensatory hypertrophy in the rat. They also accord with Lipschutz's observations (1925, 1928; also Lipschiitz and Voss, 1925) on the cat and rabbit. These indicated that no new oocytes are formed when an ovary or fragment of an ovary undergoes compensatory hypertrophy, and that the number of oocytes present in a small piece of ovarian tissue left in the body is gradually and rapidly exhausted by recurrent phases of follicular maturation. Regenerative Capacity of Ovarian Tissue 35 8. The conclusions stated in the preceding paragraph have been confirmed (Mandl, Zuckerman and Patterson, 1952) in a more extensive series of experiments on litter-mate rats which received the following treatments : A. right ovary removed : more than half of left resected. B. right ovary removed : left ovary untreated. C. same as A, but remaining fragment of left ovary painted with tannic acid in order to destroy the germinal epithelium. D. right ovary removed: surface of whole left ovary painted with tannic acid. Compensatory hypertrophy occurred in the ovarian tissue left in the body, whether or not the germinal epithelium was present, and was not associated with any increase in the expected number of oocytes. At the same time a unilaterally- ovariectomized animal lost fewer oocytes within a given period than a normal animal possessing both ovaries. But the rate of loss was substantially greater than would be expected to occur in a single ovary of a normal animal, and increased inversely with the amount of ovarian tissue left in the body. The latter observation suggests that the smaller an ovarian graft, the more quickly will it become depleted of its oocytes. 9. The total number of oocytes does not increase after hypophysectomy (Ingram, 1953). This experiment was carried out in order to test statements by Swezy (1933) that stimulation by gonadotrophic hormone decreases the rate of oogenesis in the adult rat, and corre- spondingly, that the rate of oogenesis increases after hypophy- sectomy. The hypophysectomized animals which she used were, in fact, younger than her control animals, and would have been expected to possess more oocytes, regardless of other physiological considerations. Her data merely suggest that the number of oocytes decreases more slowly with age in hypophysectomized rats than in normal animals. These are the main observations bearing on the problem of 36 S. ZUCKERMAN oogenesis which have so far emerged from studies in my own laboratory. It is necessary to consider them in the light of the contrary statements that have been made more recently by Aron, Marescaux and Petrovic (1952, 1954a, b); by Van-Eck (1955); by Burkl (1954, 1955); and by Burkl and Kellner (1954a, b, 1955). The new observations reported by Aron, Marescaux and Petrovic (1954&) relate to the apparent occurrence of differen- tiating oocytes in the ovary of the mature guinea pig. They present their findings in the context of a general review, in which they refer to various statements made by other workers in support of the alternative views (a) that oogenesis occurs only during embryonic life; (b) that it continues for a variable period after birth; (c) that all the oocytes formed during embryonic life are destroyed shortly after birth, and are then replaced by a new generation of oocytes; and (d) that new oocytes are formed throughout reproductive life. Aron and his collaborators come down firmly in favour of the last of these hypotheses, and cite a number of papers which they believe prove that oocytes can be formed either from the germinal epithelium or from cord-like invaginations of this epithelium within the substance of the mature ovary. Their selection of the literature is, however, arbitrary; and they make little attempt to analyse critically the papers they cite. Indeed, at least a few are quoted in a sense different from what their authors apparently intended. For example, Nunes (1932) is referred to as one who has helped establish " with certainty" that oocytes differentiate from invaginations of the germinal epithelium within the substance of the ovary. In fact, all his paper records is that in the adult rabbit ovary follicles are often closely related to epithelial invaginations, and his own con- clusion— "on ne peut done nier d'une maniere absolue la neoformation ovigene dans l'ovaire adulte" — is much less categorical than is implied by Aron et aZ.'s references. Lane- Claypon (1905) and Pincus and Enzmann (1937), who are among many others who have studied the rabbit ovary, and both of whom are also cited by Aron et ah, reached quite Regenerative Capacity of Ovarian Tissue 37 different views about oogenesis — Lane-Claypon that oocytes were formed from interstitial cells, and Pincus and Enzmann that not only was this not the case, but that few mitoses occurred in the germinal epithelium, and that no primordial oocytes migrated from this epithelium. In themselves, these differences of view merely indicate that the so-called histo- logical evidence of oogenesis is anything but clear-cut. But the fact that they exist makes it reasonable to expect some justification for giving greater weight to one rather than another interpretation of the histological data. This Aron et al. do not do. Matthews' and Harrison's (1949) observations on the seal are also cited as cast-iron evidence that invaginations of the mature germinal epithelium give rise to oocytes. But on this point, these two authors are completely non-committal. All they say is that "in many instances the smaller crypts, and diverticula of the larger ones, are found to terminate as primary follicles, the epithelium surrounding the oogonia being directly continuous with the germinal epithelium lining the lumina of the crypts. ... It is not yet clear whether the oogonia arise directly from the epithelium of the crypts or whether they have reached their position previously from some other source." Other authors, again, are quoted as stating that oocytes may be present between the cells of the germinal epithelium of the mature ovary, the implication being that they are formed there. The observation is one which must be familiar to all who have studied the ovary. It is, however, another and purely arbitrary matter to infer that the presence of oocytes in this situation indicates that they are derived from germinal epithelium. Furthermore, it is not at all clear that many of the authors cited by Aron et al. themselves regarded, or would now be prepared to regard, their observations as indicative of this particular conclusion. For example, all that Deane (1952), one of the workers referred to, states is that "small ova are occasionally seen within the germinal epithe- lium— more frequently they occur just below it, accompanied 38 S. ZUCKERMAN by a cluster of satellite cells." The further inference about oogenesis, implied by Aron et aL's reference to the paper, is not even mentioned. The only paper cited in connection with this particular issue, and which, in fact, bears critically on the question of oogenesis, is that of Hamlett (1935) on the arma- dillo. It does so, not because it emphasizes the topographical relations of oocytes to the germinal epithelium, but because it describes nuclear changes that may be indicative of oogenesis (see later). Observations about the position of cells in the ovary are far too equivocal to sustain conclusions about the occurrence of oogenesis in adult life. Oogenesis is a process. As I have emphasized before (Zuckerman, 1951), to infer that it occurs from a study of histological sections implies that separate and distinct phases of the process, each necessarily viewed in isolation can, as it were, be set in continuous motion by a dynamic interpretation. This might be a reasonable exercise if there were clear-cut and generally accepted cellular stages which in the mature ovary linked an indisputable oocyte with some other and equally well-defined cellular constituent of the ovary, for instance, a cell of the germinal epithelium. But this is not so. Some writers, for example Allen (1923) and Bullough (1942), have merely assumed that if a cell of the germinal epithelium divides, one or both of the daughter cells will differentiate into an oocyte, and only a very few have attempted to define the cellular stages which, in the adult ovary, might link an oocyte with a cell of the germinal epithelium, or with any other cellular component of the ovary. The majority of histologists have completely failed to con- vince themselves on this point. I have elsewhere suggested an alternative, and far simpler, explanation of the proliferative powers of the germinal epithe- lium (Zuckerman, 1951). It is that the capacity of the cells of the adult germinal epithelium to subdivide reflects the extensive cyclical changes which occur in the shape and size of the ovary as follicles mature and burst, and tear the surface of the ovary, and as corpora lutea develop. These Regenerative Capacity of Ovarian Tissue 39 changes could not possibly run their course in the way they do if the germinal epithelium was not able to proliferate (see also Schmidt and Hoffman, 1941; and Schmidt, 1942). Other kinds of evidence bearing on the question of oogenesis are referred to by Aron et al. in a non-critical, or at best non- committal way, and apart from certain comments which are re- ferred to below, these authors do not resolve the contradictions posed by the work done in my own laboratory to the thesis they support. They proceed to deal with the field of observation in which they themselves have contributed. This concerns the occurrence in the ovaries of the mature lemur, the armadillo and the guinea pig, of nuclear changes which are suggestive of the division of oocytes. The lemurs can be usefully considered first, as they were the first to be described in this connection (Gerard, 1919/20; Rao, 1927; and Gerard and Herlant, 1953). The observations of Gerard (1919/20) and Gerard and Herlant (1953) relate to adult specimens of the African Loris- idae (Galago senegalensis moholi, G. crassicaudatus, G. demidoffi and Perodicticus potto). In all these species "oogenetic cords" containing primordial oocytes are found separated from the germinal epithelium by a tunica albuginea of varying thick- ness. Many of the cells in these cords are actively dividing, and according to Gerard and Herlant can be regarded only as oogonia. Other cells show chromatin configurations charac- teristic of the leptotene, synaptene and pachytene stages of meiosis. Cells corresponding in all cytological respects with oogonia were also recognized in the germinal epithelium itself. Rao (1927) studied the Indian species Loris lydekkerianus, and noted that "tubular" invaginations of the germinal epithelium ramify in certain parts of the ovary of the pregnant, but not non-pregnant, adult. Rao was under the impression that it had been firmly established that oogenesis in mammals continues throughout adult life, and he therefore tried to decide whether oocytes are formed from invaginations of the germinal epithelium, or as Lane-Claypon had suggested in the case of the rabbit, from the interstitial cells of the ovary as 40 S. ZUCKERMAN well. His own view was that cells from the germinal epithelium divide mitotically, and either give rise to granulosa cells or pass into the interior of the ovary as interstitial cells, where they may later become transformed into oocytes. He de- scribes how the nucleus of an interstitial cell may enlarge, with the chromatin filaments of the nucleus separating over the increased area in the form of a loose meshwork interspersed with chromatin nodules. This he regarded as the leptotene stage of meiosis. He then goes on to describe what he con- sidered to be a synaptic condition of the nucleus, indicated by the massing of the filaments into a lump on one side of the nucleus. A pachytene stage is also pictured, in which the synaptic lump is resolved into coarser and more bulky fila- ments. On the other hand, Rao states that the diplotene phase is either fugitive or does not occur, since the dual arrangement of the filaments cannot be made out. All these interpretations are based entirely upon an analysis of separate cells, and Rao is careful to point out that he did not have the opportunity of studying oogenesis, for purposes of comparison, in foetal ovaries. Hamlett's (1935) observations of oogenesis in the mature armadillo relate to a single pregnant specimen which had "a button of tissue projecting abnormally from the otherwise smooth surface" of the ovary. His histological description accords closely with that of Gerard and Herlant for the lemur, but Hamlett, nevertheless, regards his specimen as illustrating not the normal state of affairs, but "what must happen in the rare instances" when the epithelium of a mature ovary in which the tunica albuginea is well developed "undergoes proliferation". "The cells are forced outward, not inward, for they are incapable of penetrating the dense layer of connective tissue and smooth muscle forming the tunica." He goes on to say that "any postnatal replenishment of oocytes" from the surface is normally impossible in the armadillo, as well as in other animals whose ovaries are invested by a dense tunica. The oogenetic phenomena described by Gerard and Herlant and by Hamlett seem, at first reading, to be similar to those Oocytes in Immature and Mature Guinea Pigs Fig. 1. Immature guinea pig. Nest of oocytes showing phases of meiosis, probably pachytene and diplotene. Fig. 2. Pubertal guinea pig. Nest of oocytes showing phases of meiosis, including zygotene. Fig. 3. Mature guinea pig. Nest of oocytes showing phases of meiosis, including apparent zygotens. Fig. 4. Old guinea pig. Nest of oocytes showing phases of meiosis, including apparent zygotene. Fig. 5. Normal oocyte, showing pachytene phase of meiosis. Fig. 6. Three stages of meiosis. facing page 41 Regenerative Capacity of Ovarian Tissue 41 which Aron and his collaborators now report in the guinea pig, and which we can confirm from personal observation One often finds m the ovarian cortex of this animal, cells frequently crowded together in nests, which seem to be either oogonia or oocytes. What appear to be phases of meiosis can be recognized, although it is by no means simple to identify all the stages which Aron et al. have described (cf. Figs. 1-4) On the other hand, they do not describe oogonia undergoing mitosis, their view being that oogonia, having been formed by a straightforward metamorphosis of quiescent germinal elements in the germinal epithelium, become directly trans- formed into oocytes. Aron et al. also believe (a) that this presumed neoformation of oocytes varies considerably from animal to animal; (b) that until the animals are about 2\ years old, when it usually becomes negligible or ceases, the process is independent of age; (c) that the process does not vary cyclically; (d) that its intensity is not affected either by gonadotropin or hypophysectomy; (e) that it does not occur more vigorously in one ovary after its fellow has been removed; and (f ) that it is not affected by the injection of oestrogen. What is significant about these observations is not that Aron et al. were able to observe phases of meiosis in nests of oocytes, but the fact that they failed to find any evidence of mitosis of oogonia, and that they were consequently forced to conclude that the latter are formed from the direct transfor- mation of dormant cells in the germinal epithelium. Others before them, as well as ourselves, have also failed to observe mitosis in presumed germinal elements in the guinea pig ovary (e.g. cf. Myers, Young and Dempsey, 1936; Evans and Swezy, 1931). It is also worth noting that the cellular formations (e.g. nests of "oocytes ") seen in the guinea pig ovary have lent themselves to a variety of interpretations, and that one group of workers, following Loeb (1905, 1932) regard partheno- genetic development of ovarian oocytes as common in this species (e.g. Myers, Young, and Dempsey, 1936), whereas others (e.g. Stockardand Papanicolaou, 1917; Evans and Swezy, 1931; Bacsich and Wyburn, 1945) regard them as a rarity. It is, of 42 S. ZUCKERMAN course, conceivable that the nuclear changes described by Aron et al. in the ovary of the mature guinea pig constitute no more than the meiotic changes which usually occur pre- pubertally in most other mammals. For example, de Wini- warter (1920) has noted that the reduction division of oocytes may occur in the ovaries of cats which have just passed puberty. To the best of my knowledge, however, active meiotic changes are not usually seen, except in degenerating follicles or follicles about to ovulate, in the ovaries of adult rats, mice, rabbits, ferrets and monkeys. None of these reports, indicating or suggesting that meiosis, and in Gerard and Herlant's case mitosis as well, may occur in the germinal cells of the ovaries of mature females of certain species, is furnished with a control account of the histology of the foetal ovary, at those times when oogonia are unquestionably multiplying. This deficiency was recognized by Rao, but is not referred to by the other workers concerned. Bujard (1947) describes the ovary of the immature guinea pig, but nothing he records allows one to determine the extent to which the presumed oogenetic phenomena in the adult ovary, as outlined by Aron et al., and as confirmed by ourselves, correspond with the histological manifestations of actual oogenesis, which presumably is still in progress at birth. Oogenesis in the mature animal must by definition mean the continued formation of several oocytes from some precursor cell, in the same sense that a single spermatogonium is in the end responsible for the production of a very large number of spermatozoa. An essential criterion of oogenesis would, there- fore, have to be the multiplication of oogonia by mitosis, in the same sense that the occurrence of meiosis represents the end, not an intermediate phase, in the process of formation of the ovum. The studies reported by Burkl (1954, 1955) and Burkl and Kellner (1954a, b, 1955) are in general of a more experimental kind than those of Aron et al. In his first paper, Burkl (1954) reports that oestrogen (hexoestrol implantation) increases the rate of neoformation of oocytes in rats, as judged by the Regenerative Capacity of Ovarian Tissue 43 "activity" of the germinal epithelium. This conclusion is hardly borne out by the data that are published, since these, in fact, show that the total number of primordial oocytes was smaller in treated animals than in controls of the same weight- groups. Results of experiments in which rats were fed the highly toxic drug Myleran are also taken by Burkl and Kellner (1954a, b) to support the thesis that oogenesis is a process which continues into adult life. Some of the animals used in this experiment died as a result of the treatment, and only one normal and three experimental rats were left to provide what in effect proved to be the basic thesis of the study, that Myleran acts only on follicles whose oocytes exceed 38 (x in diameter. This figure is used in a series of calculations de- signed to determine the daily maturation rate of follicles, by which is meant the number of oocytes over 38 fx in diameter divided by the period of the experiment. By successive and increasingly unconvincing transformations and comparisons, the conclusion is then reached that some 25 oocytes degenerate in each cycle. Since the stock of oocytes at the onset of sexual maturity is asumed to be 3500, it is concluded that neoforma- tion of oocytes must occur if the ovary is not to be depleted of oocytes too soon. Quite apart from various theoretical and mathematical defects of the argument, the design of the whole experiment is totally inadequate to sustain any such conclu- sion. It is worth noting, too, that the average number of oocytes at puberty in the Birmingham strain of albino rats is 6000, and that a stock of oocytes of this size would be sufficient to sustain the rate of loss of oocytes calculated by Burkl and co-workers without assuming that oogenesis continues after sexual maturity had been reached. In a later paper, Burkl (1955) again concludes, but on this occasion without resorting to oocyte counts or a study of nuclear changes, that oogenesis occurs in the mature ovary of the dog. This he does merely on the basis of the topographical relation of oocytes to cords of epithelial cells which had presumably been derived from the germinal epithelium. And in a paper which follows (Burkl and Kellner, 1955), he 44 S. ZUCKERMAN purports to show that the rate of neoformation of oocytes in albino rats increases after hypophysectomy. Oocyte counts are given for this study, but their statistical analysis, again, does not bear out the conclusion stated. In the first experi- ment the mean number of oocytes for eight prepubertal hypophysectomized animals was 3840 ± 300, and for five controls 3180 ± 170. This difference is not significant statisti- cally. In a second experiment, carried out on mature animals, the corresponding figures for eight hypophysectomized animals are 2650 ± 230 and for five normal controls 2580 ± 230. Again, the difference does not approach statistical significance. Analysis of these figures, in fact, shows that the results of Burkl's experiments agree closely with those of Ingram (1953), carried out in my own laboratory. They fail to demonstrate a significant increase in the number of oocytes after hypo- physectomy, but do indicate that the usual rate of decline in the numbers of oocytes may slow down after the pituitary is removed. Van-Eck (1955) attempts to reach a conclusion about the occurrence of oogenesis in the adult ovary of the rhesus monkey from an experimental estimate of the time it takes an oocyte to become atretic and to disappear. This work is based on oocyte counts of fourteen immature (represent- ing ten animals) and seven mature ovaries (representing four animals). The time taken for a follicle to become atretic was estimated from experiments on three young adult monkeys which were irradiated with 1200 r (600 r on two successive days) between the 19th and 22nd day of the men- strual cycle, and ovariectomized 7, 10 and 14 days later respectively. Oocytes surrounded by more than one layer of cuboidal cells were not present after 7 days, and after 14 days the only oocytes present were those surrounded by one layer of epithelial cells. Van-Eck concludes that "once atresia occurs in a follicle the process is invariably completed within two weeks ; for the small growing follicles the process takes less than one week." On the assumption that the time taken for atresia to occur Regenerative Capacity of Ovarian Tissue 45 is the same "whether the damage is caused physiologically (from unknown factors) or artificially (by X-rays)", and on the basis of estimates of the total number of normal oocytes, and the percentage of oocytes regarded as atretic, the fol- lowing function tn = a(l - r)n in which tn = the number of oocytes after n periods of atresia; a = the original number of oocytes; and r the percentage of atresia, is used to show that the ovaries would be practically exhausted within two years if new oocytes were not formed; consequently, oogenesis must occur in the mature ovary, two years being the theoretical maximum life of the oocyte. Once again both facts and arguments are hardly adequate to support the conclusion. In the first instance, Van-Eck assumes that normally "atresia is of short duration for the primordial and small primary follicles — the follicle is resorbed, leaving no trace, and its place in the ovary replaced with stroma." To the best of my knowledge this contention is arbitrary, in so far as no critical proof has ever been provided that under physiological conditions atresia of young oocytes is a rapid process. Estimates of the rate of atresia, inferred from observations on single ovarian phases, are just as insecure as are inferences of postpubertal oogenesis from a picture of an epithelial cord apparently budded from the germinal epithelium. No-one knows how long an atretic oocyte or follicle in a normal ovary can remain in the same apparent state; nor indeed is there any agreement among students about the occurrence of a high rate of degeneration in prim- ordial oocytes. A second weakness of the work is that it assumes that atretic oocytes can be diagnosed sufficiently well to justify statements such as: "the percentage of atretic ova was constant for each group ". (According to her estimates 4 • 5 per cent of all oocytes are atretic or in process of becoming so.) Elsewhere in the paper, however, she noted that "for the primordial oocytes it is sometimes difficult to determine 46 S. ZUCKERMAN whether the cell is atretic or still normal" — and in such cases she counted the cell as normal. Actually, anyone who has tried to make successive counts of the atretic oocytes in an ovary will know how numerous the marginal cases are, and how difficult it is to obtain estimates which agree closely. A third criticism of Van-Eck's study is the unwarranted assumption that the atresia which occurs as a result of X-ray sterilization is identical in its nature, and in its duration, with what occurs in the normal ovary. The only specific enquiry into this point of which I know (Halberstaedter and Ickowicz, 1947), in fact, indicates that the first histological appearances of atresia in the normal ovary are characteristically different from those which herald atresia after X-irradiation. An ad- ditional difficulty is that Van-Eck implicitly assumes that the time taken for an oocyte to disappear after X-irradiation is independent of the r dosage. The three monkeys which she irradiated were given a total of 1200 r in two applications through the dorsolumbar region of the body wall. It would be extraordinary if the rate at which oocytes disappeared were not increased if the same dosage were applied directly to the ovaries, or alternatively, if a dose five times as great were applied directly (see Lacassagne, 1913). The question whether oogenesis does or does not occur in the mature animal has been so beset with equivocal fact and arbitrary interpretation, that it is essential that new observa- tions on the topic, which clearly has important practical im- plications, should be marked by observational and conceptual precision if they are not to multiply confusion. The first point that needs to be settled is whether the experimental and biometric data summarized on pp. 32-35 are consistent with any thesis other than that oogenesis is in abeyance in the mature rat or monkey. By "settling", I do not, of course, mean the matching of fact with speculation — for example, disposing, as do Aron et al. of the observation that oocytes can persist in an ovary long after it has lost all recognizable signs of its germinal epithelium, by suggesting that it is possible that some unrecognizable or undefined derivative of that epithelium Regenerative Capacity of Ovarian Tissue 47 is still persisting within the substance of the ovary. What we need to know is whether any thesis other than that a finite stock of oocytes is involved can be reconciled with such plain numerical propositions as "the total number of oocytes does not increase in an ovary undergoing compensatory hyper- trophy"; or "the relative rate of loss of oocytes increases inversely with the amount of ovarian tissue left in the body". To some extent Aron et al. appreciate this point, and their way of dealing with it is to suggest, in effect, that the smaller the amount of ovarian tissue left in the body, the more the work it has to do in satisfying the body's need for oocytes, and con- sequently, the more rapid the decline in the oogenetic potency of its germinal epithelium. The difficulty with this hypothesis is not that it is arbitrary; it is un verifiable. The only tangible measure of oogenetic potency which one can conceive of would be the numbers of oocytes produced in the ovary at any given time. As I have tried to show, any realistic analysis of the differences which occur in the numbers of oocytes in different experimental conditions, or as between different normal physiological states, leads inevitably to a conclusion which implies a primary stock of oocytes. The other possi- bility is that oocytes are transitory structures, and that the actual ones that are seen, and which could be counted under the microscope at any given moment, would have been re- placed by another set of oocytes had the ovary been examined on a later occasion (e.g. that the number counted on one occasion was the measure of a given oogenetic potency at that moment ; and a lesser number counted in another set of com- parable ovaries on a physiologically later occasion was a measure of a lesser oogenetic potency). This, in fact, is what is implied by Aron's suggestion. If it were accepted, no more would be demanded of one's credulity than would be from a man who counted the apples on a tree on successive days, and who held that each day he was counting a set of apples which had miraculously replaced those that were there the day before — and just as miraculously disappeared (there being no fallen apples to account for the new ones that had 48 S. ZUCKERMAN suddenly appeared). To some extent it is a matter of taste which view one accepts ; my own is for the simpler proposition that the identities of apples — and equally of oocytes — cannot be so simply switched. What answer should, then, be given to the question whether oocytes continue to be formed in the adult ovary? If we set aside for a moment some of the cytological evidence, the experimental data undoubtedly support the view that as a rule the mature mammalian ovary is incapable of adding to the store of oocytes with which it is furnished by the time of puberty. Most of this evidence has been derived from work on rats, but the more fragmentary observations that have been made on other species agree far more with this thesis than they do with the contrary one. There appear to be certain notable exceptions, and it may well be that oogenesis ceases in some species earlier than in others, in which it may continue for some time after puberty. But in general the ovary appears to be a transient tissue which, so far as its main function of oogenesis is concerned, has little or no powers of regeneration once puberty has been reached. Compared with the testis, the ovary is in this respect a structure which reaches senescence early, and which cannot be rejuvenated by any known hormonal treatment. Ovarian secretion This limitation does not extend, in any corresponding measure, to the ovary's powers of hormonal secretion. These normally manifest themselves first at the time of puberty, and in woman they fade out at the menopause. In the interval between these two temporal events, the ovary's capacity to secrete hormone is under the control of the gonadotrophic secretions of the pars distalis of the pituitary. Its capacity to respond to this stimulation is often held to be independent of the presence of germinal elements, but the evidence bearing on this point is somewhat equivocal. Thus, in rats, an ovarian graft that is apparently devoid of oocytes may some- times secrete sufficient oestrogen to provoke continuous or Regenerative Capacity of Ovarian Tissue 49 recurrent oestrous phases in the host animal for periods up to a year (Parkes, 1956). Similarly, mice which have been steri- lized by X-irradiation, and whose ovaries are devoid of oocytes, may still continue to experience recurrent phases of oestrus for periods up to 69 days (Parkes, 1927). On the other hand, it is a matter of common observation that the postmenopausal human ovary, when depleted of its oocytes, ceases to secrete sex hormone. And it is also believed that it cannot do so, even when subjected to the influence of gonadotropin (see Burrows, 1949). Because of the apparent conflict between these two sets of observations, the whole subject has recently been re-investigated in my own laboratory. The first point which we (Mandl and Zuckerman, 1956a, b) re-investigated was the observation that a mouse ovary that has been depleted by means of X-irradiation of all oocytes and of all recognizable follicular elements may still produce sufficient oestrogen to induce vaginal cornification and even recurrent oestrous cycles (Parkes, 1926, 1927; Westman, 1930). Our technique differed from the earlier experiments in so far as we irradiated the ovaries not through the body wall, but directly. Only ovaries which were found on serial section to be absolutely devoid of oocytes were regarded as sterile. According to our observations on rats, the usual pattern of oestrogenic activity after successful irradiation (as judged by the criterion of a total lack of oocytes) is that short periods of vaginal cornification occur at increasingly irregular intervals, and that after a minimum of five and usually within 40 days, the vaginal smear becomes continuously cornified. The period of cornification lasts from 2 to 14 weeks, and during this time the animals, in our experience, will not mate. The vaginal smear then becomes anoestrous in type. The behaviour of the X-ray sterilized mouse also follows this pattern. Signs of oestrogenic stimulation are not confined to the vaginal epithelium, but can also be observed in the uterus, which while smaller than normal, remains significantly larger than in spayed control rats as long as 26 weeks after treat- ment. On the other hand, changes occur in the weight of 50 S. ZUCKERMAN the body, thymus and spleen during the three months which follow sterilization, which correspond to those that occur after surgical removal of the ovaries. Further experiments showed that these indications of oestrogenic activity continue after bilateral adrenalectomy, but cease after removal of the X-ray sterilized ovaries — which presumably are therefore responsible for the secretion of the hormone. In effect, what all this adds up to is that the capacity of the ovary to secrete sex hormone is neither necessarily, nor absolutely, dependent upon the presence of germinal elements. Earlier histological studies of the ovaries of mice that had been X-irradiated at birth or at the age of three weeks had indicated that the cells which are responsible for the continued secretion of oestrogen after the elimination of the follicular system are those which are derived from what is described as "the first post-irradiation proliferation" of the germinal epithelium (Brambell, Parkes and Fielding, 1927a, b). The tissue produced by this proliferation is "almost indistinguish- able from true luteal tissue . . . the epithelial cells becoming like luteal cells with connective tissue elements of the sheath growing in amongst them like the thecal cells of the corpus luteum." Such proliferations do not occur in the adult ovary. The X-irradiated adult mouse which continues to undergo oestrous cycles was held by Brambell and Parkes (1927) to be under the influence of oestrogen produced by "inter- follicular tissue " and "follicular derivatives ". The latter two elements are, for all practical purposes, indistinguishable from each other, and form the main part of the sterilized ovary. Histologically they closely resemble the cells seen in mice irradiated before puberty, and which are derived from the cords of the first proliferation. We have found that even when no luteal-like cells are present in rats sterilized before puberty, and when no cyclical changes occur in the vaginal smear, the vaginal closure membrane nevertheless breaks down (Mandl and Zuckerman, 1956c). Since treatment with oestrogen makes it do so precociously (Allen and Doisy, 1924), the obvious inference would be that Regenerative Capacity of Ovarian Tissue 51 the sterilized ovary secretes a small amount of oestrogen even when no obvious luteal cells are present, and that the threshold of oestrogenic stimulation responsible for canalization of the vagina is lower than that at which cyclical changes in the vaginal epithelium occur. An alternative possibility is that an X-irradiated ovary which is devoid of apparent luteal tissue does not produce oestrogen but androgen, or some adreno- cortical hormone, and that the latter is responsible for the break- down of the vaginal closure membrane (see Burrows, 1949). Whichever hormone it be, there is little likelihood that it merely accelerates the canalization of the vagina, which would otherwise take place, independently of hormonal action, as development proceeds, since, according to our experience, the vaginal canal remains permanently closed if rats are ovariecto- mized twenty days or so before the usual time of disappearance of the closure membrane. One further fact needs to be considered before returning to the question of the dependence of the secretory function of the ovary on its germinal elements. Is the secretory capacity of the ovary after X-irradiation, whether slight or pronounced, under the control of pituitary gonadotrophin, or is it auto- nomous? The answer is not clear-cut, but inclines to the latter possibility. We have found, in experiments on rabbits and mice, that an ovary that has been effectively sterilized by means of X-rays, and in which all oocytes and follicles have been destroyed, does not undergo compensatory hypertrophy when its fellow is removed (Humphreys and Zuckerman, 1954). This observation has been confirmed in a more exten- sive series of experiments on rats (Mandl and Zuckerman, 1956a). These experiments have also shown that a completely sterilized ovary not only fails to undergo compensatory hyper- trophy, but also does not respond by an increase in size to exogenous gonadotrophin. On the other hand, we also have indications — they are little more than this — that the in- tensity of oestrogenic secretion of the X-irradiated ovary, however slight it may be, can be increased by gonadotrophic stimulation. 52 S. ZUCKERMAN These observations seem to leave little doubt that the mechanism of secretion is different in an X-irradiated as compared with a normal ovary. Not only does it not re- spond to gonadotrophic stimulation like a normal ovary, but at no time (to judge, for example, by the size of the uterus) does it secrete as much oestrogen as a normal ovary. Since neither recurrent periods of vaginal cornification nor per- sistent cornification continues indefinitely after X-irradi- ation, it is also clear that the X-irradiated ovary becomes progressively incapable of secreting at all. This conclusion leaves open the general question whether the decline in secretory capacity is due to the elimination of the germinal elements of the ovary, or to damage inflicted on other ovarian cells by the X-irradiation. At the moment it is difficult to see how this particular problem can be answered. A decision could possibly be based on experiments in which rats are irradiated in a carefully graded series, so that their post- irradiation behaviour could be related to the amount of damage suffered by their ovaries. Conclusion From the fact that reproductive life in higher mammals is of shorter duration than life itself; from the knowledge that fertility in mammals, including man, reaches a peak in early maturity and then gradually declines; from the observation that the senescent ovary cannot be reactivated by any known hormonal treatment ; and from the experimental data reported in this paper, one can only conclude that in general the ovary is a transient tissue with little regenerative capacity, not only with respect to its germinal but also to its secretory functions. The period over which the ovary is a regenerating structure, in the sense that it is able to produce oocytes, may vary from one mammalian species to the other, but, in general, it is an organ which is more senescent, or potentially more senescent, than the testis, both from the germinal and the secretory points of view. Regenerative Capacity of Ovarian Tissue 53 REFERENCES Allen, E. (1923). Amer. J. Anat., 31, 439. Allen, E., and Doisy, E. A. (1924). Amer. J. Physiol, 69, 577. Arai, H. (1920a). Amer. J. Anat., 27, 405. Arai, H. (19206). Amer. J. Anat., 28, 59. Aron, C, Marescaux, J., and Petrovitch, A. (1952). C. R. Ass. Anat., 39, 421. Aron, C., Marescaux, J., and Petrovic, A. (1954a). C. R. Soc. Biol., Paris, 148, 388. Aron, C, Marescaux, J., and Petrovic, A. (19546). Arch. Anat., Strasbourg, 37, 1. Bacsich, P., and Wyburn, G. M. (1945). J. Anat., Lond., 79, 177. Brambell, F. W. R., and Parkes, A. S. (1927). Proc. roy. Soc., B, 101, 316 Brambell, F. W. R., Parkes, A. S., and Fielding, U. (1927a). Proc. roy. Soc, B, 101, 29. Brambell, F. W. R., Parkes, A. S., and Fielding, U. (19276). Proc. roy. Soc, B, 101, 95. Breward, M. M., and Zuckerman, S. (1949). J. Endocrin., 6, 226. Bujard, E. (1947). Acta anat., 4, 68. Bullough, W. S. (1942). J. Endocrin., 3, 141. Burkl, W. (1954). Verh. anat. Ges. Jena, 52, 81. Burkl, W. (1955). Z. Zellforsch., 41, 421. Burkl, W., and Kellner, G. (1954a). Anat. Anz., 100, 322. Burkl, W., and Kellner, G. (19546). Z. Zellforsch., 41, 172. Burkl, W., and Kellner, G. (1955), Acta anal., 23, 49. Burrows, H. (1949). Biological Actions of Sex Hormones. 2nd ed. Cambridge: University Press. Deane, H. W. (1952). Amer. J. Anat., 91, 363. Evans, H. M„ and Swezy, O. (1931). Mem. Univ. Calif., 9, 119. Everett, N. B. (1943). J. exp. Zool., 92, 49. Gerard, P. (1919/20). Arch. Biol., Paris, 32, 357. Gerard, P., and Herlant, M. (1953). Arch. Biol., Paris, 64, 97. Green, S. H., and Zuckerman, S. (1951). J. Endocrin., 7, 194. Green, S. H., and Zuckerman, S. (1954). J. Endocrin., 10, 284. Green, S. H., Mandl, A. M., and Zuckerman, S. (1951). J. Anat., Lond., 85, 325. Green, S. H., Smith, A. U., and Zuckerman, S. (1956). In press. Halberstaedter, L., and Ickowicz, M. (1947). Radiology, 48, 369. Hamlett, G. W. D. (1935). Anat. Rec, 62, 195. Herlitzka, A. (1900). Arch. ital. Biol., 34, 89. Humphreys, E. M., and Zuckerman, S. (1954). J. Endocrin., 10, 155. Ingram, D. L. (1953). J. Endocrin., 9, 307. Lacassagne, A. (1913). Effets Produits sur l'Ovarie. Lyon: Rey. Lane-Claypon, J. E. (1905). Proc. roy. Soc, B, 77, 32. Lipschutz, A. (1925). Brit. J. exp. Biol., 2, 331. Lipschutz, A. (1928). Brit. J. exp. Biol., 5, 283. Lipschutz, A., and Voss, H. E. V. (1925). Brit. J. exp. Biol., 3, 35. Loeb, L. (1904-05). Arch. mikr. Anat., 65, 728. 54 S. ZUCKERMAN Loeb, L. (1932), Anat. Rec, 51, 373. Mandl, A. M., and Zuckerman, S. (1949). J. Anat., Lond., 83, 315. Mandl, A. M., and Zuckerman, S. (1950). J. Endocrin., 6, 426. Mandl, A. M., and Zuckerman, S. (1951a). J. Endocrin., 7, 190. Mandl, A. M., and Zuckerman, S. (19516.). J. Endocrin., 7, 103. Mandl, A. M., and Zuckerman, S. (1951c). J. Endocrin., 7, 112. Mandl, A. M., and Zuckerman, S. (1956a). J. Endocrin. In press. Mandl, A. M., and Zuckerman, S. (19566). J. Endocrin. In press. Mandl, A. M., and Zuckerman, S. (1956c). J. Endocrin. In press. Mandl, A. M., Zuckerman, S., and Patterson, H. D. (1952). J. Endocrin., 8, 347. Marshall, F. H. A., and Jolly, W. A. (1907). Trans, roy. Soc. Edinb., 45, 589. Matthews, L. H., and Harrison, R. J. (1949). Nature, Lond., 164, 587. Moore, C. R., and Wang, H. (1947). Physiol. Zool, 20, 300. Myers, H. L., Young, W. C, and Dempsey, E. W. (1936). Anat. Rec, 65, 381. Nunes, J. P. (1932). C. R. Soc. Biol., Paris, 111, 598. Parkes, A. S. (1926). Proc. roy. Soc, B, 100, 172. Parkes, A. S. (1927). Proc roy. Soc, B, 101, 421. Parkes, A. S. (1956). J. Endocrin., 13, 201 Pettinari, V. (1928). Greffe Ovarienne et Action Endocrine de l'Ovarie. Paris: Doin. Pincus, G., and Enzmann, E. V. (1937). J. Morph., 61, 351. Rao, C. R. N. (1927). Quart. J. micr. ScL, 71 N.S., 57. Schmidt, I. G. (1942). Amer. J. Anat., 71, 245. Schmidt, I. G., and Hoffman, F. G. (1941). Amer. J. Anat., 68, 263. Simpson, M. E., and Wagenen, G. van. (1953). Anat. Rec, 115, 370. Stockard, C. R., and Papanicolaou, G. N. (1917). Amer. J. Anat., 22, 225. Swezy, O. (1933). Ovogenesis and its Relation to the Hypophysis. Lancaster, Pennsylvania: Science Press. Van-Eck, G. J. V. (1955). Anat. Rec, 121, 378. Westman, A. (1930). Acta obstet. gynec. scand., 10, 299. Winiwarter, H. de. (1920). C. R. Soc. Biol., Paris, 83, 1403. Zuckerman, S. (1951). Recent Progr. Hormone Res., 6, 63. DISCUSSION Dempsey: I am not in a hurry to cross swords with Prof. Zuckerman on this issue. I rather agree with him. I haven't any vast experience in this field, and it has been many years now since I was concerned with this issue. I was interested at that time in making some calculations as to the total number of oocytes present in the guinea pig ovary. When these were counted in the pre-pubertal animal's ovary and the number so obtained was divided by my best estimates of the number of oocytes that were lost during each reproductive cycle, I came out with a figure that was perfectly in agreement with the hypothesis that all the oocytes are formed before puberty. There are plenty of them to Discussion 55 last the entire reproductive lifetime of the animal. And this in spite of the fact that is well known that the loss of the ova at each reproductive cycle is very much larger than the number of eggs that come to ovula- tion. My estimate was that somewhere between 100-200 eggs underwent follicular development during each cycle, of which two to six, at most, ovulate, and the remainder become atretic at the period immediately following ovulation. And it was the process of that atresia that led me to wonder as to its mechanism, because it is a very startling thing that almost instantly after ovulation, nuclear changes overtake practically all of the unovulated eggs. I have also found that, if at any stage in the cycle one injected gonadotrophins, abnormal maturation divisions were induced by the treatment, and that the fate of the eggs, so altered, went on through this characteristic form of atresia. The phenomenon of Loeb's that Prof. Zuckerman mentioned is a very common one in the guinea pig. In serial sections through both ovaries of many hundreds of guinea pigs, I have found these masses, which Loeb interpreted as parthenogenetic development of the eggs, to be extremely common. It is a phenomenon which occurs so widely that I quite agree with you it cannot be ignored in attempting to estimate the origin of the various cells that one sees. Comer: I find myself in complete agreement with Dr. Dempsey. My own direct experience with this problem is limited to a conscientious effort, made over more than thirty years of study of serial sections of mature monkey ovaries, to see whether I could discover any sign of change of the cell of the germinal epithelium, or anything else, into an oocyte. This has been a complete failure. I have completely failed to see any convincing production of oocytes in mature monkey ovaries. This is in spite of Swezy's account based in some part on some mis- interpreted specimens from one of my papers. I do not think she ever saw the specimens ; she worked from my figures only. In the recent work of Van-Eck done in part on some of the materials borrowed from me, there seems to me a possibility of a fundamental error. She tried to wipe the slate off, so to speak, by irradiating the ovaries and then used deductions about the growth of follicles immediately thereafter, without apparently realising the possibility that the irradiation had done something to those follicles which grow and degenerate in the next few days or weeks. So there again I find myself fully in doubt as to the neogenesis of ova. Zuckerman: I think the flaw in the Van-Eck papers is really a serious one. Van-Eck has assumed, in the first instance, that the atretic phenomenon which occurs after irradiation is identical with natural atresia. Because you can destroy an oocyte quickly with X-rays, she therefore assumes that you can tell exactly how long it takes for an ordinary oocyte to disappear under natural circumstances. There is no reason for supposing that the two processes are identical. Parkes: There is one question I would like to ask. In this connection, the definitions of the expressions pubertal, postpubertal and pre- pubertal seem to be fundamental. What constitutes puberty from the point of view of regeneration of oocytes ? 56 Discussion Zuckerman: In the context of oogenesis the term varies according to species. If, for example, it is true that the lemurs can produce oocytes when sexually mature, I would be inclined to say that they constitute one extreme in a range of mammals, the other being those species which achieve their entire complement of oocytes before birth. In the case of the rat we know that nuclear configurations of the kind which could be interpreted as oogenesis cease to appear after about day 4. Some people put it even earlier. Parkes: That is long before the first ovulation? Zuckerman: Yes, but even so it would not surprise me to find a few oocytes being formed later. Huggett: Tell me, have you any knowledge as to where these oocytes come from? Zuckerman: I have always assumed that it was true that the gono- cytes migrate into the embryo. Bounoure, Regan, Wolff and Nieuwkoop have studied the problem, and their results show that the gonocytes migrate to the genital ridge. The gonocytes become "fixed" there in some sort of way ; they multiply ; and when they cease multiplying they become differentiated into oocytes. Corner: Hurdig has been able to mark, in human embryos, the gono- cytes with one of the microchemical stains — a lipophosphatase. This seems to be quite clear. Do you not think it is quite believable work and that this corroborates the old story? Strauss: Is it possible for the neoformation of oocytes in the adult mammal to be a question of comparative embryology as well as one of a cytological nature? Zuckerman: In what sense embryological ? Strauss: In the lower archaic mammals, like the Galago and Dasypus, you will find some neoformation. Gerard, as well as Hamlett, claims to have seen neoformation in these primitive mammals, in as distinct a manner as one could wish. Zuckerman: All these creatures started off on their reproductive life as mammals, and as far as we know as far back as the palaeocine. They have continued and evolved quite successfully until now, a few fading out. I do not think we have any reason to regard the Galago as being more archaic than the rest, and that postpuberal oogenesis constitutes a primitive phenomenon. Matthews: In fishes, of course, there is entirely new production of oocytes, but I don't know what happens in the bird. Parkes: Prof. Zuckerman, can you say whether or not endocrine activity on the part of the ovary continues after the destruction of the oocytes and, therefore, of the follicular system, and if so for how long? Zuckerman: I can only give you the facts for experiments on rats which we have done. Including litter-mate controls, 306 rats from 76 litters were used. The ovaries were directly irradiated after being ex- teriorized (the rest of the body being shielded), at a dose about four times higher than the lethal dose when the beam is applied to the body surface. Complete sterilization was judged by the absence of a single recognizable oocyte in serial sections. Most animals entered a phase of Discussion 57 continuous "oestrus" within about forty days after treatment. Phases of oestrus occurred intermittently in the preceding period. Animals which were not in continuous oestrus forty days after irradiation usually turned out to be incompletely sterilized. Having moved into a state of continuous vaginal cornification, the animals remained in that con- dition for from two to fourteen weeks, and then became anoestrous. The weight of the uterus, however, was higher than that in spayed litter- mate controls, even 26 weeks after treatment. A small amount of oestrogen was clearly being produced throughout the period of observa- tion (as judged by the weight of the uterus) even though it was insuffi- cient to cause cornification in smears, except in the earlier part of the experiment. The vaginal mucosa showed some degree of oestrogenic stimulation, and consisted of more than four layers of cells. Approxi- mately the same results have been obtained on the mouse. Parkes: That interests me very much because of the work I did years ago with Brambell on irradiation of the mouse ovary. At that time we were concerned entirely with finding out whether or not cyclic activity would persist after the destruction of the follicular apparatus ; and it did persist. We were not interested then in trying to find out how long endocrine activity would persist, but endocrine activity survived certainly for some months after destruction of the follicular apparatus. It so happened that, talking to Prof. Zuckerman recently, I discovered that he was doing these irradiation experiments and he discovered that I was doing not very dissimilar ones — irradiating rat ovarian tissues in vitro before making a graft. Our work is in an early stage but the results we have obtained so far are at least compatible with his, though we do not know for certain what happens to a graft which has been irradiated in vitro. Zuckerman: Both what Dr. Parkes has just said, and our own findings, suggest that there may be some relation between the secretory capacity of the ovary and the follicular system, in the sense that the elimination of the latter eventually affects the former. But exactly what the relation is, I do not yet know. Parkes: Then there is a further point; that if you put up the dose of the X-rays too high you destroy the whole ovary in the same way as a smaller dose destroys the oocytes only. Krohn: But of course, in the menopausal ovary you have no oocytes left and no hormone production apparently. The two do go hand in hand there. Dempsey: In these experiments in which you produce continuous oestrus after irradiation, will the animals mate ? Zuckerman: No. In spite of a completely cornified smear they will not mate. Dempsey: And have you tried interrupting their continuous oestrus either by giving gonadotrophins or progesterone ? Zuckerman: Neither. Dempsey: There is a phenomenon in the rat in which continuous oestrus is produced by constant illumination and in which as long as the animal is maintained under lights, the vaginal smear will be in a cornified 58 Discussion state. Here, however, the continuous oestrus may be interrupted by a variety of devices, by mating or by administration of gonadotrophin or progesterone, in which case the animal will run a cycle or two and then go back into continuous oestrus again. I was wondering if perhaps part of the phenomenon of your animals ceasing ultimately to be in oestrus was not merely a matter of exhaustion. After all, fourteen weeks of continuous oestrogenic secretion in the life of a rat is a long period. Parkes: In our experience, rats in which cyclic corniflcation is passing into persistent corniflcation will quite often mate, and they will certainly mate under conditions where the ovarian tissue they possess does not have any follicular system. Zuckerman: We specifically tested the animal's readiness to mate, but failed to observe it on any occasion. I do not know why the intermittent corniflcation in the immediate post-irradiation period fades into con- tinuous corniflcation, and then into the anoestrous type of smear. Williams: May it not be something to do with the adrenal? Zuckerman: No. We have removed the adrenals, keeping the animals on saline and that does not interrupt it. Williams: Do I understand that if you adrenalectomize the animals, you do not interfere with the cyclic phase? I was thinking that you might get hypertrophy or overactivity of the adrenal later and this might account for the loss of corniflcation. Zuckerman: The experiments on the removal of the adrenals were done when we knew the animals were so many days in continuous oestrus, and from previous experience, could predict they would con- tinue until corniflcation faded out altogether. Huggett: What is your index of oestrus? Zuckerman: We do not call it "oestrus"; we refer to vaginal corniflcation. I should like to ask Dr. Parkes whether he believes there is any difference between the likelihood of obtaining a successful graft of testicular as compared with ovarian tissue. Parkes: For the gametogenic property or the endocrine one? Zuckerman: Gametogenic, although I do not imply the complete formation of spermatozoa. Parkes: Well, the situation is quite different because unless the testis graft is put in a position where it would normally have a chance of showing spermatogenesis you won't get it, of course. Zuckerman: I do not mean complete gametogenesis ; I mean the multiplication of spermatogonia, and not necessarily spermatocytic differentiation. I was under the impression that it is easier to make a successful graft of testicular than of ovarian tissue, as judged by the continuation of the earlier stages of gametogenesis. THE HISTORY AND FATE OF REDUNDANT FOLLICLES P. C. Williams Imperial Cancer Research Fund, London The fate of redundant follicles, so far as I am capable of dealing with it, is easily summarized. I believe that follicular degeneration or atresia has been observed as a normal happen- ing in the ovary of all mammals that have been studied. It certainly occurs in all laboratory rodents and I am only going to deal with these — primarily with the rat with a side- glance at the guinea pig. Of the many follicles that start on the path of development towards ovulation in these animals, only a small minority reach the goal — most fall by the wayside. It is stated that degeneration may set in at any stage in follicular development, but again I am confining my remarks to atresia of medium-sized and large follicles — broadly, those with an antrum. This is because it is the most striking form of atresia to observe and much the easiest to study. The course of that atresia is familiar. In the guinea pig, atretic follicles show initially (Fig. 1) a pyknosis of the nucleus, and disintegration, of the granulosa cells bordering the antrum. The granulosa layers rapidly degenerate and, as the follicle shrinks, the inner thecal cells proliferate and take on a fibrocytic appearance (Fig. 2). Meanwhile the ovum also degenerates and finally only the theca retains its structure — it remains unaffected by the secondary degeneration of the fibrocytic tissue, so that thecal nodules (Fig. 3) persist giving the interstitial tissue a nodular appearance before merging into general uniform interstitial tissue (Fig. 4). There is much less fibrocytic proliferation from the theca in the rat, but otherwise the process is exactly the same. In both species, the interstitital cells so constituted. 59 60 P. C. Williams although retaining an endocrine function as shown by histo- chemistry (Dempsey and Bassett, 1943; Claesson and Hillarp, 1947; Deane, 1952) and X-irradiation studies (Brambell and Parkes, 1927; Parkes, 1927), are not so prominent as in other species, for example in the rabbit, where they hypertrophy to form an interstitial gland. Thus the fate of the redundant follicle is partly to disappear completely, partly to persist to form part of the endocrine tissue of the ovary. Now to come to the history of atresia. It has been stated that atresia occurs in cycles in the guinea pig (Aron and Aron, 1953) and mouse (Engle, 1927), being most marked after ovulation in the guinea pig, and during the first day of di- oestrus in the mouse, though no such cycle has been proved in the rat (Mandl and Zuckerman, 1950). If there is such a correlation, then it is some evidence that atresia is conditioned by hormonal changes and it is this aspect of the subject I wish to discuss. First of all I should like to recapitulate some old work on the direct effects of oestrogen on the ovary (Williams, 1944). If an immature rat is hypophysectomized the ovaries atrophy. This atrophy can be at least partly prevented by treatment with oestrogen starting at the time of operation. Table I Table I Ovarian Weight in Immature Rats after Hypophysectomy, or Hypo- PHYSECTOMY AND THE IMPLANTATION OF A STILBOESTROL TABLET Weight of ovaries (mg.) Days after operation Hypo. Hypo.+stilb. 1 5-9 7-8 2 50 8-9 3 60 9-2 4 4-6 10-5 5 30 8-4 10 3-3 8-7 15 2-9 9-4 5 rats per group. Ovarian weight in 20 normal immature rats =8 -2 mg. Plate I. The stages of follicular atresia in the ovary of the pregnant guinea pig ( x 45). Fig. 1. Earliest and late stage of follicular atresia. Fig. 2. Intermediate stage of follicular atresia. - Fig. 3. Final stages of folli- cular atresia: thecal nodules. Fig. 4. Merging of the nodules into the general interstitial tissue. The microscopical preparations were kindly lent by Dr. I. W. Rowlands. facing page 60 Plate II. Ovaries of immature rats ( X 22) Fig. 5. Normal animal. Fig. 6. Fifteen days after physectomy. hypo- Fig. 7. Fifteen days after hypo- physectomy and oestrogen treatment. Fig. 8. Twenty-two days after hypo- physectomy ; oestrogen treatment started I? days after operation. History and Fate of Redundant Follicles 61 shows the progressive loss of weight in the ovaries after hypophysectomy and the maintenance of weight if a stilb- oestrol tablet is implanted at the time of operation. The same effects can be produced by daily injections of 50-100 [ig. of oestrone. The histological effects of the treatment are striking. The ovary of the normal immature rat (Fig. 5) contains follicles in various stages of development and atresia, and almost all with a mean diameter greater than 200 \i are vesicular. (Hereafter, it is only follicles of this size that are counted and discussed.) There is little change in the general appearance during the first five days after operation in the untreated animal though more of the vesicular follicles are now atretic, and by the 15th day (Fig. 6) the follicular apparatus is almost non-existent. If oestrogen treatment is started at operation then by the 5th day the ovary is full of medium-sized follicles almost all of which are solid — suggesting that oestrogen stimulates proliferation of the membrana granulosa — and this condition persists until the 15th day at least (Fig. 7), though by this time the follicles are not quite such a pro- minent feature of the ovary as they were at the 5th day. If oestrogen treatment is delayed until seventeen days after operation there is no recovery of ovarian weight, though some remaining rudimentary follicles are stimulated and enlarge (Fig. 8). I think this proves that the stimulation depends upon the presence in the ovary of follicles already possessing a membrana granulosa (it does not effect primordial follicles) but does not depend upon the presence of gonadotrophin in the circulation. Quantitative study of the immediate effects (Table II) shows that the number of follicles per ovary with a diameter of more than 200 \i is actually increased by the oestrogen treatment, that most of these follicles are solid (and in fact those called vesicular mostly show only that localized fragmentation of the membrana granulosa that is the first sign of antrum formation), and that the size of the largest follicle in each ovary remains more or less constant. In the untreated animal the number of follicles of this size 62 P. C. Williams progressively decreases and there is a reduction in the size of the largest follicle. I have recently studied some of this old material from the point of view of follicular degeneration. Some interesting points arose. The normal immature rat ovary contains about 40-50 follicles more than 200 [i in diameter and of these about half are in the early stages of atresia. My impression is that Table II Effect of Stilboestrol Implantation at Operation on the Histology of the Ovaries in Immature, Hypophysectomized Rats Animal Days after operation Follicles > 200 [i in diameter Mean maximum follicle diameter Total Solid Normal — 49 4 399 Hypophysectomized 2 5 10 15 38 25 9 7 5 8 6 7 312 258 218 216 Hypophysectomized and stilboestrol- treated 2 5 10 15 52 85 85 67 16 55 48 38 412 376 378 377 this process is a very rapid one; certainly almost all the follicles in the later stages of atresia are smaller than this, and so are not included in the counts. These is little change in the total count during the first five days after operation but the proportion of atretic follicles is greatly increased. Only 6-10 of the follicles are normal at this time and 20-30 are atretic. The proportion of atretic follicles has increased from 40-50 per cent to 60-80 per cent. At this point we have to reckon with the observation of Aron and his colleagues (Firket, Petrovich, Marescaux and History and Fate of Redundant Follicles 63 Aron, 1953; Aron and Aron, 1953) that a single injection of a crude extract of beef anterior hypophysis produces at a certain dose level atresia of virtually all the follicles in the ovary of the guinea pig. As the dose injected is increased these effects are produced; firstly, follicle stimulation; with a larger dose, atresia; with a still larger dose, luteinization ; and with the highest doses, atresia again. Within the limits producing atresia, the higher the dose the quicker the effect. At first sight, this observation seems to conflict with the increase in atresia produced by hypophysectomy in the immature rat — but I am not sure that there is a conflict. I believe that the operation of hypophysectomy releases gonadotropin into the circulation. The figures in Table I suggest that there is an interruption in the ovarian weight loss on the 3rd day after hypophysectomy. The difference in ovary weight at this point is not statistically significant but I think it is a real one, and Lane and Greep (1935) made a similar observation. There is some supporting evidence for the suggestion that this temporary, slight recovery of ovarian weight is real and a reflection of endogenous gonadotrophin. When a single dose of the gonadotrophin from pregnant mares' serum is given 3-4 days after hypophysectomy to immature rats, the ovarian response is greater than it is if the injection is made before or after this time (Williams, 1945). I think it may well be that gonadotrophin hastens the follicular atresia which certainly occurs, however, in its absence — 21 days after hypophysectomy there are still some follicles in the ovary and some of these are undergoing atresia though the propor- tion (only about 15 per cent) seems to be subnormal. When we study the ovaries of the oestrogen-treated, hypophysectomized rats there is a different picture. During the first four days after operation there are less atretic follicles than there are in the untreated animal — there are 6-12 instead of 20-30. Apparently oestrogen retards atresia. But on the 5th day something new appears. Fragmentation of the membrana granulosa occurs resembling, so far as one can see from a haematoxylin-eosin preparation, the preliminary 64 P. C. Williams stages of normal atresia, except that this occurs peripherally instead of centrally (Figs. 11-12). This is not merely a result of the predominance of solid follicles because it occurs in the same situation in follicles in which antrum formation is beginning. In the untreated, hypophysectomized animal atresia seems to be quite normal ; even when it occurs in almost solid follicles it starts around the oocyte (Fig. 10). The abnormal form of atresia proceeds to completion in the oestrogen-treated animals and eventually outweighs the stock of rudimentary follicles that were present at hypophysectomy, so that the ovary eventually becomes as atrophic as that of the untreated, hypophysectomized rat. This presumably explains a puzzling thing about the original observation: that oestrogen stimu- lated the ovarian follicles, yet long-continued oestrogen treat- ment of intact rats will produce ovarian atrophy. I shall not attempt to explain the peculiarity of this form of atresia, but it may be relevant that Bullough (1942) in his counts of the mitoses in normal follicles in the mouse found that there were more in the central layers of the granulosa than in the peripheral layers, and suggested that it was the oestrogen in the follicular fluid that was responsible for this. In the oestrogen-treated, hypophysectomized rat it is fair to assume that oestrogen is concentrated at the external surface of the follicle. The degenerative process has been stated to be initiated in dividing cells of the membrana granulosa (Salazar, quoted by Corner, 1932). A further point that may be relevant is that this type of atretic fragmentation has been described as being present in almost all follicles of oestrogen-treated, hypophysectomized rats that have been treated at the same time with chorionic gonadotrophin or androgen (Gaarenstroom and de Jongh, 1946). These findings demonstrate that the membrana granulosa is a true transient tissue. Its fate may be to degenerate or to be transformed into another transient tissue — the corpus luteum. The life of the membrana granulosa in the rat is limited to about five days. I think I have proved that oestrogen Plate III. Follicular atresia in immature rats ( x 68) Fig. 9. Normal immature rat. Fig. 10. Hypophysectomized im- mature rat. Figs. 11-12. Hypophysectomized rats treated with oestrogen facing page 64 History and Fate of Redundant Follicles 65 stimulates the proliferation of the membrana granulosa, but even in the continued presence of oestrogen the degenera- tion of this tissue is not prevented — it is only delayed for at most several days. I think that there is evidence too that the degeneration may be hastened by gonadotrophin, perhaps acting by the liberation of ovarian androgen. Apart from this, the ageing process in the membrana granulosa is completely obscure. Histochemistry has revealed some interesting facts about the degeneration process once it has started but no-one can yet tell whether a follicle is proceeding to ovulation or will undergo atresia until the degeneration starts. Postscript The observation recorded by Professor Zuckerman in his communication suggests to me that all the ovarian tissues are transient. He has certainly convinced me that oogenesis is no longer a function of the ovary in adult life. The transience of the membrana granulosa I have just discussed, and that of the corpus luteum is well established. Professor Zuckerman' s findings in rats with X-irradiated ovaries suggest that the thecal tissue and its successor, the interstitial tissue, are transient too. He has confirmed Dr. Parkes's earlier observation that cyclic vaginal cornification may continue in the complete absence of oocytes and fol- licles and has extended this finding by observing that such vaginal cycles do not continue indefinitely. After a number of weeks the vagina reverts to a continuous di-oestrous condition. This suggests that the function of the interstitial tissue is limited in time unless the tissue is continually renewed from the thecal cells of degenerating follicles and corpora lutea. Such a hypothesis would explain the lack of cyclic activity in the postmenopausal woman, in whom, as Professor Krohn pointed out, lack of cyclic ovarian activity is not accompanied by persistence of cyclic vaginal changes. Acknowledgement. I am very grateful to Dr. D. J. Trevan, Mr. E. V. Willmott, F.R.P.S., and Mr. G. D. Leach for preparing the photomicrographs. AGEING VOL. 2 4 66 P. C. Williams REFERENCES Aron, M., and Aron, C. (1953). Arch. Anat., Strasbourg, 36, 69. Brambell, F. W. R., and Parkes, A. S. (1927). Proc. roy. Soc. B, 101, 316. Bullough, W. S. (1942). J. Endocrin., 3, 150. Claesson, L., and Hillarp, N. A. (1947). Acta physiol. scand., 14, 102. Corner, G. W. (1932). Special Cytology, ed. Cowdry, 2nd edn., Sec- tion 39. New York: Hoeber. Deane, H. W. (1952). Amer. J. Anat., 91, 363. Dempsey, E. W., and Bassett, D. L. (1943). Endocrinology, 33, 384. Engle, E. T. (1927). Amer. J. Anat., 39, 187. Firket, H., Petrovich, A., Marescaux, J., and Aron, M. (1953). C. R. Soc. Biol, Paris, 147, 501. Gaarenstroom, J. H., and de Jongh, S. E. (1946). A contribution to the knowledge of the influences of gonadotropic and sex hormones on the gonads of rats. Amsterdam: Elsevier. Lane, C. E., and Greep, R. O. (1935). Anat. Bee, 63, 139. Mandl, A., and Zuckerman, S. (1950). J. Endocrin., 6, 426. Parkes, A. S. (1927). Proc. roy. Soc. B, 101, 421. Williams, P. C. (1944). Proc. roy. Soc. B, 132, 189. Williams, P. C. (1945). J. Endocrin., 4, 127. DISCUSSION Amoroso: Before putting Mr. Williams' paper open to discussion I would like to ask him one question. Are these atretic phenomena equally well expressed in the adult ovary as in the immature ovary? Williams: That I do not know. Zuckerman: Is the first sign of atresia in the normal ovary invariably in the granulosa? A paper published about two years ago discussed the nature of the atretic process in the X-irradiated ovaries as compared with the normal, and the suggestion was made that X-irradiation, like any noxious stimulation, began a process of atresia which started with degeneration of the oocyte. An abortive division first occurs. I under- stood that in the atresia which occurs in the normal ovary, changes in the oocyte began after changes in the granulosa. Williams: I do not know which comes first. Parkes: You seem to have very well-preserved eggs in some of the follicles which have become atretic. Is that usual under those conditions? One expects to find rather a bad looking egg inside an atretic corpus luteum. Rowlands: I can confirm, from my limited experience with the guinea pig ovary, that the egg does seem to be preserved in a good condition for a very long period of time in an atretic follicle and the zona pellucida remains recognizable in a corpus luteum atreticum for a large part of the pregnancy. Strauss: I would like to ask if you think that antrum formation is dependent on an endocrinological process ? Williams: Yes, I think it certainly is. Discussion 67 Strauss: What about those animals which do not develop an antrum, and which have solid ripe follicles like the Centetinae? Williams: I am afraid I know nothing about them. Krohn: Would you attribute changes in the follicle to the vascularity around the follicle, atresia starting, perhaps, at the inside of the follicle because it is furthest away from the blood supply, and the effect of oestrogen being related to a local increase in vascularity? Williams: I don't know whether any work has been done on that. The awkward thing is you cannot distinguish by any means that I know, which follicle is going to become atretic at any time, and which one is going to proceed to ovulation, until the atresia has started. Zuckerman: How can one then decide whether the follicle undergoing atresia which is being looked at at a given moment in a section would or would not have been there if the examination had been made a fort- night later, or a month later? Williams: I do not think you can, except from counts. The follicle population diminishes as the amount of atresia increases. Zuckerman : That is true for the whole ovary. My question applied to a specific group of follicles, which looked as though they were undergoing atresia. I understood that your impression was that the process is rapid. How could you know this by studying the individual atretic follicle ? Williams: You could, I think, only do it by vital staining methods. I have not tried. Krohn: But you are always up against the problem of seeing the same follicle twice. That is the real catch, I think. Amoroso : It is generally known that a fundamental difference exists between the primary follicle in the normal ovaries of young rats up to about 13 days of age and those of older animals following hypophy- sectomy. The follicles of the young rats do not become greatly enlarged in response to pituitary gonadotrophic preparations, while those of the hypophysectomized animal do. From the morphological standpoint there is no apparent difference between the condition of these two types of follicles. The significance of this difference in follicular response is unknown, but it suggests the possibility that follicular ageing may be an important conditioning factor in the response of the follicle to pituitary stimulation. Corner: I have a somewhat unorthodox notion about atresia which I might mention here in order to have it knocked down. In a rhesus monkey, when a very large follicle is about to ovulate, and the contained ovum is in the last stages of maturation, the granulosa layer undergoes a process which is very much like atresia. The nuclei of the granulosa cells become pycnotic; the cells are infiltrated with fluid and become fibroblastic in shape. I have called attention to this in published forms. It is visible in other people's pictures, for instance in some of the pictures Boling published years ago from the rat. Dempsey: It is also true in the rabbit. Corner: Yes, and I think it has been seen in one or two other species also. Now in the rhesus monkey at the time of ovulation and formation of the corpus luteum it is not uncommon to have some small atretic 68 Discussion follicles luteinized, and we have seen a number of cases where only a part of the small follicle was luteinized, forming a little partial accessory corpus luteum one side of a tiny follicle. The other side of that follicle was in full atresia. I have formed the notion that a process of partial atresia, or something very much like it, is an essential step in the conversion of granulosa cells into granulosa lutein cells. Amoroso : That may also be the case in the ovary of the foetal giraffe where atresia is widespread and is accompanied by the conversion of granulosa cells into granulosa lutein cells. Dempsey: Is there any possibility that there are at least two types of atresia ? I was struck by Mr. Williams' pictures of his oestrogen-treated animals which showed an atresia occurring at the base of the follicle as opposed to the generally occurring location near the antrum. Then, there was the phenomenon Dr. Parkes mentioned, that these eggs were extraordinarily well preserved in the atretic follicles, whereas in the atresia that occurs after ovulation in the guinea pig ovary it is my impression that the egg shows changes very rapidly indeed. So, is it possible that we are confusing perhaps two different physiological mechanisms ? Williams: The process of atresia in the oestrogen-treated animal is obscure to me. You don't seem to see the intermediate stages. Un- fortunately my material is very scanty between the fifth and the tenth day after hypophysectomy and oestrogen treatment. You do see what are obviously remnants of atretic follicles, which have obviously under- gone atresia since treatment started. The general appearance is rather different from those in the normal animal. I think this is a different form of atresia. Strauss: Is not atresia the normal fate of an egg follicle, and matura- tion and ovulation only the exception? In humans, for example, the 400 eggs, approximately, which ovulate during the fertile period are unimportant compared with the 400,000 primary follicles. Therefore, there are many more atretic follicles than ripening follicles. What sig- nificance do such atresia and atretic follicles have? Harrison: I was interested to hear you say that you thought granu- losa cells had only a short life. Some years ago I carried out some ex- periments in puncturing mature follicles of the rabbit ovary and with- drawing granulosa cells and then culturing them. They grew, interest- ingly enough, into a series of spheres for two, three or four days; I could never get them to grow longer than that. If, however, minute quantities of a water-soluble oestrogen were added, then the spheres increased in size and would live for several days longer. Williams: That would fit perfectly well with what I think happens. By giving oestrogen you increase the life of the follicle cells for perhaps a few days but no longer. Perhaps there is an analogous situation in the ovarian interstitial cells in Prof. Zuckerman's X- irradiated animals. It may be that these cells too have a limited life, and in fact vaginal cornification may cease because there is no continual renewal of interstitial tissue from atretic follicles in the X-irradiated animals. THE CORPUS LUTEUM OF THE GUINEA PIG I. W. Rowlands Agricultural Research Council Institute of Animal Physiology, Babraham, Cambridge The salient morphological and histological features of the ovarian changes in the guinea pig have been known for half a century. In 1906, Loeb described its early development and drew attention to the early closure of the rupture point of the follicle, the formation of a fluid-filled central cavity and to the occurrence of haemorrhage before the obliteration of the cavity by luteal tissue. In a later paper, Loeb (1911) gave a compre- hensive account of the histology of the ovary during the di- oestrous cycle and pregnancy, taking into account the corpus luteum and follicular activity. He noted the greater size of the former during pregnancy and that it remained a large circumscribed organ until parturition. Nevertheless, cy to- logical evidence of regression was obtained which seemed to be associated with loss of vascularity. A rapid and complete degeneration of all vesicular follicles was noted in unmated and pregnant animals for a short period of about four or five days after ovulation, after which follicular growth was resumed, and by the 10th day large healthy follicles together with others showing signs of recent degeneration accompanied the corpora lutea. A number of the healthy follicles soon undergo very decisive changes which are best described in Loeb's own words — "These follicles are characterized by an increase in cytoplasm of the granulosa cells. The nuclei of the granulosa cells are not so densely packed in these follicles as in the ordi- nary large follicles; this peculiarity being due to the marked development of the cytoplasm. They can easily be recognized in sections stained by haematoxylin and eosin insomuch as they appear stained more reddish in contradistinction to the ordinary large follicles in which the blue colour of the nuclei predominates, while in the mature follicles the red stain of the cytoplasm is a distinguishing feature". 69 70 I. W. Rowlands I have confirmed the presence of these large follicles in the guinea pig ovary during pregnancy and their staining reaction, as noted by Loeb, forms a ready means of their distinction from any other follicles that are present. They are, quite clearly, follicles that are undergoing the changes associated with the pre-ovulatory phase of development. In no single case has any evidence of spontaneous ovulation been found during pregnancy. The object of the present work has been to discover if ovulation can be induced during pregnancy and to study the growth of the resulting corpora lutea in the presence of those associated with gestation. However, before embarking on this experimental procedure it was necessary to make a quantitative study of the development of the corpora lutea of the cycle and of pregnancy for, to my knowledge, no work of this sort has been published. The experiments form part of an investigation into the capacity of the guinea pig to maintain a second concurrent pregnancy. Material and Methods All the guinea pigs were of the Hartley-Dunkin (M.R.C.) strain, of which eighty were used to establish the normal growth of the corpus luteum of the cycle and after mating with a sterile or fertile male. A vasectomized male was used to effect sterile mating. All observations were dated from the post-partum oestrus to ensure that every animal was fertile and that each one contributed to the replenishment of the stock before being put to experimental use. The times of parturition and post-partum mating were accurately recorded for those animals used to obtain early developmental stages of the corpus luteum, but for later stages the occurrence of mating was presumed by the discovery of a vaginal plug or by sperm in a vaginal smear. Chorionic gonadotrophin, used to induce ovulation, was injected into one of the small ear veins and the result observed by examination of tubal washings for eggs or by the serial section of the Fallopian tubes. The ovaries and the repro- ductive tract were fixed in Bouin's fluid overnight and serial The Corpus Luteum of the Guinea Pig 71 sections cut at 7 (x; every 5th section was mounted and stained in haematoxylin and eosin. Measurements of the corpora lutea and the mature follicles were made across three diameters. One of these was got by counting the number of sections in which the organ appeared and the other two by measuring with a micrometer eyepiece the largest section of the organ ; measurements were made at right angles through the centre of the section. The product of the three diameters in mm.3 (D3) has been used throughout to express the dimension of these organs from which true volume may easily be calculated. The Corpus Luteum Quantitative observations At ovulation, which occurs about fifteen hours after parturi- tion, the mature follicle measures about 1 mm. in diameter, by which time changes suggestive of luteinization have already commenced in the membrana granulosa. Fig. 1 gives the mean 8 4 1 "« • • O 5 IO 15 20 No. of days after oestrus Fig. 1. Mean D3 in mm.3 of the corpora lutea of the guinea pig (£) and after sterile mating (0). 72 I. W. Rowlands 6 8 5 «• • •• • • •• • • • • I V T5 30 40 Staga In pp«gnanay idaya) ■fr 68 5 IO IS Fig. 2. Mean D3 in mm.3 of the corpora lutea of pregnant guinea pigs. Fig. 3. Regression lines and formulae of data given in Figs. 1 and 2. The Corpus Luteum of the Guinea Pig 73 D3 of the corpora lutea in 25 unmated guinea pigs and in five which were mated to a sterile male, and Fig. 2 is a similar representation of the same organs in fifty pregnant guinea pigs. The regression lines calculated from the data are given in Fig. 3. The growth rate of the corpora lutea in mated and unmated animals is very similar but the period of growth differs. In the latter it ceases on about the 12th day and the organs then rapidly regress ; the regression is not inhibited as the result of sterile mating. In the pregnant animal the corpus luteum continues to grow until the 18th to 20th day and is maintained at this size for the rest of pregnancy, but after parturition it rapidly becomes disorganized. Its involution occurs when the new corpora lutea are actively growing in the same ovaries. Microscopic observations This investigation is not primarily concerned with detailed cytological changes in the corpus luteum, but a general de- scription of the more gross histological features of the ageing organ is required since it is on these grounds that the distinc- tion between the normal and the induced corpora lutea is made. The mature follicle protrudes only slightly above the surface of the ovary and after its rupture forms only a small crater over which the germinal epithelium is re-established in 48 to 72 hours. Both the granulosa and the theca interna contribute to the luteal tissue which at three to four days consists of a narrow band enclosing a central cavity as de- scribed by Loeb. The cells are small and fairly closely packed ; they have a small dark-staining nucleus surrounded by clear cytoplasm of irregular shape. Mitotic figures are not common at this stage. Possibly because of compression set up in the cavity by the ingrowth of luteal tissue, haemorrhage occurs very commonly and in some of the corpora lutea the cavity becomes packed with red blood cells for as long as the 20th day in some pregnant animals, but in others where bleeding is less severe strands of connective tissue encroach and obliterate the cavity by the 7th day or soon afterwards. The luteal cells 74 I. W. Rowlands meanwhile expand; the cytoplasm enlarges and the nucleus becomes spherical in shape. At this stage the corpus luteum has a good blood supply. Changes up to the 12th day are approximately similar in unmated and mated animals, but afterwards, the corpora lutea in the former gradually undergo regressive changes which become intensified when ovulation recurs on the 16th to the 18th day. Meanwhile in the pregnant animal, the corpus luteum shows little change until after the 20th day when the luteal cells gradually enlarge and tend to a polyhedral shape. Mitoses are seen more commonly and the gland has a very rich blood supply. This condition of apparent high activity persists until about the 35th day when a very gradual ageing process sets in ; the outline of the cells becomes less distinct, the cyto- plasm is more coarsely granular and vacuoles appear, and the nucleus becomes elliptical in shape. Nevertheless, a small number of the luteal cells appear which have all the characters of those in a young, functional corpus luteum. The amount of fibrous connective tissue increases, which probably serves to maintain the size and shape of the gland. The most striking change which occurs over this period is the restriction of the blood supply due to a progressive closure of the vessels. The gland as a whole, however, remains circumscribed until parturition after which it collapses. In so far as a similar change occurs in the corpus luteum of the unmated guinea pig when a new set of follicles rupture, it may well be that the final collapse of the gland at parturition is associated with the recurrence of ovulation at the next oestrus, fifteen to eighteen hours later, rather than from the withdrawal of placental hormones or the evacuation of the uterus. Follicular Activity and the Induction of Ovulation in Pregnancy At the time of post-partum mating, all follicles other than those destined to ovulate twelve hours later are undergoing atresia, and for the four days after rupture no healthy The Corpus Luteum of the Guinea Pig 75 vesicular follicles are to be found. But as Loeb noted, a new wave of follicular growth then takes place and by the 8th to 10th days mature follicles are present, and in the series of ovaries examined at least one of these structures was found throughout pregnancy. No evidence of spontaneous ovulation was obtained. A few initial experiments indicated that a dose of 50 i.u of chorionic gonadotrophin was sufficient to induce ovulation in pregnancy. A summary of the results obtained in 33 guinea pigs, given in Table I, indicates that the follicles in early Table 1 The Frequency of Ovulation induced by an Intravenous Injection of 50 i.u. of Chorionic Gonadotrophin at Different Stages of Pregnancy Injection period (days) No. of animals injected No. of animals ovulating Average no. of follicles ruptured 8-20 21-^0 41-62 10 16 7 5 16 7 20 2-8 3-9 pregnancy (8 to 20 days) are less sensitive to the action of the hormone than are those occurring in late pregnancy, when the average number of follicles that rupture is nearly doubled. The Fate of the Ruptured Follicle Guinea pigs were injected with chorionic gonadotrophin on the 8th, 20th and 35th days of pregnancy; they were killed 8, 14, 20 and 28 days later and the induced and the normal corpora lutea were measured as described above. Classification of the two sets of corpora lutea was based on qualitative rather than quantitative differences shown. By these means the induced corpora lutea at 8, 14 and 20 days of age are readily distinguishable from those of pregnancy, but in some ovaries the differences between the 28-day-old induced corpora lutea 76 I. W. Rowlands and the aged corpora lutea of pregnancy are not great. Although they may be different in size it was felt that this criterion was no certain means of classification, and to enable the writer to be unaware of their size at the time of histological examination, he was shown only a high-power microscopic Table 2 The Growth of Corpora Lutea induced at Different Stages in Pregnancy Code No. IOP. Induction of ovulation in pregnancy (day) Age of in- duced cor- pora lutea (days) Stage in pregnancy (days) No. of corpora lutea Mean D3 (mm3) of corpora lutea Pregnancy Induced Pregnancy Induced 11 13 9 10 12 14 8 8 8 8 8 8 8 8 14 14 20 20 16 16 22 22 28 28 7 4 6 6 2 4 1 2 2 1 2 4 4-96 601 6-72 5-63 6-28 7-55 1-62 305 2 10 2-32 1-94 2-57 21 23 28 24 26 22 25 27 21 21 21 21 21 21 21 21 8 8 14 14 20 20 28 28 29 29 35 35 41 41 49 49 8 9 5 6 6 6 7 6 3 1 2 3 2 4 1 7 5-41 5-89 6-34 4-94 4-95 6-01 7-00 5-21 1-49 2-26 3-57 2-56 3-37 3-53 4-93 2.96 17 15 18 31 16 19 36 20 35 35 35 35 35 35 35 35 8 8 14 14 20 20 28 28 43 43 49 49 55 55 63 63 9 4 5 5 6 6 5 10 3 4 3 5 3 1 4 4 6-98 9-28 8-20 6-69 9-69 5-88 8-39 7-31 3-03 2-64 419 2-39 4-63 4-39 3-89 4-38 The Corpus Luteum of the Guinea Pig 77 field upon which classification was based. A very close agree- ment between classification based on the results of micro- scopic examination and size of the gland was obtained. The most reliable criterion used to separate the two sets of glands was the degree of vascularity but when at a later stage of development closure of the blood vessels had occurred in the Stage in pregnancy (days) Fig. 4. Mean D3 in mm.3 of corpora lutea induced on the 8th (O) ; 21st ( X ) and 35th day (0) of pregnancy compared with the normal structures in unmated and pregnant guinea pigs. induced corpora lutea as well, the distinction was made on cell size, nuclear shape, vacuolation of the cytoplasm and the estimates of the numbers of mitotic figures present. The data obtained are presented in Table 2 and Fig. 4 on which lines fitted by eye, illustrate the trend of growth of the corpora lutea induced at the three different stages of pregnancy. The first wave of follicles to mature, when ruptured artificially on the 8th day of pregnancy, form very small corpora lutea growing at a very slow rate, but although they do not attain the size of those in the unmated animal, they are maintained 78 I. W. Rowlands at their full size for a longer period. It would seem that the later in pregnancy that ovulation is induced, the more rapid is the growth and the greater is the size attained by the cor- pora lutea. The follicles ruptured on the 35th day form corpora lutea of similar size to those produced during the di-oestrous cycle but unlike the latter they remain at their maximum size, as do the corpora lutea of pregnancy. Histologically, the induced corpora lutea resemble those of the unmated animal for they do not undergo the changes characteristic of normal structures during the 3rd to the 4th week of pregnancy. The corpora lutea of pregnancy in the ovaries of the guinea pig in which ovulation was induced on the 35th day (mean D3 = 7-70:1: 0-68 mm.3) were significantly larger than those (mean D3 = 5 • 65 ± 0 • 37 mm.3) in ovaries containing a set of corpora lutea induced on the 21st day. Enlargement does not seem to be related to the number of induced corpora lutea co-inhabiting the ovaries, although in one animal (IOP/19) it did not occur when only one of the latter structures was present. Functional Activity of the Induced Corpora Lutea The inhibition of follicular activity Six pregnant guinea pigs were injected with chorionic gonadotrophin and killed 24 to 96 hours later. All follicles present at the time of injection which failed to ovulate became heavily luteinized and were recognizable as small luteal cysts containing an egg. No normal vesicular follicles were found in ovaries up to 96 hours after injection, but in other guinea pigs killed after eight days they were as plentiful as in a normal animal after mating. Capacity to sensitize the uterus to decidual reaction A few attempts were made to test the capacity of induced corpora lutea to sensitize the uterus in this way. Unilateral pregnancy was established by ligation of the left uterine horn The Corpus Luteum of the Guinea Pig 79 or the left Fallopian tube in eight young guinea pigs, and at various stages after post-partum oestrus they were injected with chorionic gonadotrophs. Six days later silk threads were inserted through the sterile horn and the sites examined histologically six days afterwards. No evidence of a decidual reaction was obtained but in two animals in which ovulation was induced on the 53rd day of pregnancy the infertile horn was enlarged around the site of irritation. It should be realized that the horn sterilized by ligation at its cervical end is a large flaccid structure containing a fluid and bears no resemblance to the uterus in the early luteal stage. The latter condition was more closely obtained in animals that were semi-sterilized by ligation of the Fallopian tube, but in two of the eight animals treated in this way there was no sign of a decidual reaction. Effect on parturition Two guinea pigs were injected in very late pregnancy to observe the effects of the presence of induced corpora lutea on parturition. One, which was treated on the 54th day, gave birth to a litter at the expected time, fourteen days later. No evidence of post-partum mating was obtained but when killed three days afterwards the ovaries contained eleven very young corpora lutea. The second animal, injected on the 62nd day, produced a litter seven days later. She was killed six hours after parturition and the ovaries, which contained six induced corpora lutea, also had numerous mature follicles awaiting ovulation. There is little to suggest, therefore, that the induced corpora lutea affect the time of onset of parturi- tion or the recurrence of post-partum ovulation. Discussion The corpus luteum is an endocrine organ whose main function in unmated animals is the regulation of the di- oestrous cycle by means of its capacity to inhibit ovulation. Of all the structures in the body, it is probably the one which has the shortest life, for in normal circumstances it remains 80 I. W. Rowlands active for periods ranging from a few days to about two weeks according to the species. Its life becomes extended in the process of reproduction and it assumes important functions in the maintenance of gestation. The extent to which it is prolonged and its relative importance to the needs of gestation varies greatly from one species to another. Reynolds (1949), in summarizing the evidence for the role of the corpus luteum in the maintenance of pregnancy in different animals, places the guinea pig in an intermediate position between those species in which luteal tissue is required for foetal survival and those in which it plays no part in the second-half of preg- nancy, for the reason that abortion does not invariably occur after oophorectomy in late pregnancy. The observations reported above have shown quite clearly that the corpora lutea persist as well-organized bodies throughout pregnancy in the guinea pig. But mere existence does not necessarily imply retention of functional activity, and in this connection reference must be made to the work of Loeb and Hesselberg (1917) showing that pregnancy was initiated and maintained in some guinea pigs for as long as thirteen days after oophorec- tomy performed as early as the 3rd to 6th day after mating. In others, treated similarly, abortion occurred. That some fundamental change in the function of the corpus luteum takes place at this very early stage of pregnancy is suggested by the renewal of follicular growth and maturation in the guinea pig ovary on about the 4th to 5th day after ovulation. It is difficult in this species, therefore, to establish the precise relationship between the corpus luteum and the maintenance of gestation. It would seem, in general, that there are two factors closely associated with the prolongation of the life of the corpus luteum. First, in some species, for example the rat, the duration of luteal activity is prolonged by cervical stimulation at mating. In the guinea pig, however, it has been shown above that mating with a vasectomized male fails to prevent the regression of the corpus luteum which normally occurs on the 12th day after ovulation. Secondly, many lines of enquiry The Corpus Luteum of the Guinea Pig 81 suggest that luteal growth is stimulated at the time of implant- ation of the blastocyst. This is shown by many species which exhibit the phenomenon of delayed implantation, for example many species of the Mustelidae (Wright, 1942) and the Pinnipedia (Harrison, Harrison Matthews and Roberts 1952; Rand 1955), and in species in which implantation is delayed by concurrent lactation, for instance the bank vole (Cleth- riomys glareolus) examined by Brambell and Rowlands (1936). In all these there is evidence that the growth rate of the corpus luteum slows down or may even stop when the blastocyst is in the free-living state in the uterus, but when attachment takes place luteal growth is resumed. Experimental evidence of the role of the uterus in prolonging luteal activity has been pro- vided by Nalbandov, Moore and Norton (1955), who have shown that distension of the uterus of sheep by insertion of a bead 8 mm. in diameter delayed the onset of the next oestrus by about eight days. In the sheep, however, no difference in size exists between the corpus luteum of the cycle and that in the pregnant animal, so that it is not possible to show whether the uterine stimulus is able to enlarge as well as to prolong the period of its activity. Clearly, an experiment of this nature is desirable in the guinea pig. Attachment of the blastocyst to the uterine epithelium in the guinea pig occurs on the 6th day after fertilization (Blandau, 1949) at which time the uterus is sensitive to stimuli producing the decidual reaction. There is, however, no suggestion in Fig. 2 of a change in the growth rate of the corpus luteum at this time and if attachment is responsible for the continuation of luteal growth from the 12th to the 20th days of pregnancy (over which period the corpus luteum is regressing in the non-pregnant animal) it becomes clear that the effects of the stimulus are not apparent for some time later. It would seem probable on the basis of the work of Velardo and colleagues (1953) on the quantitative relationship between the decidual response and extension of luteal function in the rat, and from Nalbandov' s experiments on sheep (loc. cit.), that the stimulus for the continued growth of the corpus luteum of pregnancy 82 I. W. Rowlands is provided only when a certain threshold amount of decidual development or uterine distension has occurred. It has been made clear that not only is the life of the corpus luteum prolonged in the pregnant guinea pig but that it is also significantly enlarged. Are these two features different expressions of a response to the same stimulus or are they controlled independently? It is well known that in many species prolongation of luteal activity is not accompanied by enlargement of the gland, which may be an indication that these characters are controlled separately. The observations made on a set of corpora lutea resulting from ovulation induced at different stages during pregnancy confirm this view for, although these structures never attain the size of the normal corpora lutea of pregnancy, they have a longer life than those of the unmated animal. Summary The corpus luteum of the guinea pig reaches maximum size on about the 12th day and regression, which sets in immediately, is accelerated when ovulation recurs four to six days later. Sterile mating does not prolong its life-span. In pregnancy, the corpus luteum grows at the same rate for eighteen to twenty days and is maintained at maximum size until parturition, but histological evidence of ageing occurs in mid-pregnancy. Regression is very rapid after parturition. Follicular growth is resumed on the 4th to 5th day after mating and mature follicles are found at all times after the 8th or 10th days. Ovulation does not occur spontaneously but may be induced by chorionic gonadotrophin (50 i.u.) injected at any time after the 8th day. The number of follicles that rupture and the size of the corpora lutea that are induced are greater in late than in early pregnancy. Those produced on the 53th day grow to the size of the corpora lutea in the unmated animal. They never reach the size of the co-existing corpora lutea of pregnancy but, like the latter, they are The Corpus Luteum of the Guinea Pig 83 maintained at their full size for a long time. The corpora lutea of pregnancy are enlarged when they co-inhabit ovaries con- taining another set induced on the 35th day, but not when ovulation was induced on the 21st day. Follicular growth is inhibited for four to five days follow- ing the induction of ovulation in pregnancy. The induced corpora lutea are incapable of sensitizing the sterile horn of unilaterally pregnant guinea pigs, and their presence in very late pregnancy does not interfere with parturition. A factor associated with decidual growth is considered to be responsible for the additional growth occurring in the corpus luteum of pregnancy and which is different from that main- taining the size of the corpus luteum until parturition. j REFERENCES Blandau, R. J. (1949). Anat. Rec, 103, 19. Brambell, F. W. R., and Rowlands, I. W. (1936). Phil. Trans., 226, 71. Harrison, R. J., Harrison Matthews, L., and Roberts, J. M. (1952). Trans, zool. Soc. Lond., 27, 437. Loeb, L. (1906). J. Amer. med. Ass., 46, 416. Loeb, L. (1911). J. Morph., 22, 37. Loeb, L., and Hesselberg, C. (1917). J. exp. Med., 25, 305. Nalbandov, A. V., Moore, W. W., and Norton, H. W. (1955). Endo- crinology, 56, 225. Rand, R. W. (1955). Proc. zool. Soc. Lond., 124, 717. Reynolds, S. R. M. (1949). Physiology of the uterus. 2nd. Edit. New York: Hoeber. Velardo, J. T., Olsen, A. G., Hisaw, F. L., and Dawson, A. B. (1953). Endocrinology, 53, 216. Wright, P. L. (1942). Anat. Rec, 83, 341. DISCUSSION Harrison: I would like to ask Dr. Rowlands if he has done any histo- logical or histochemical investigations on these follicles which he arti- ficially ruptured during pregnancy; was there any difference in their early development? Rowlands: No histochemical investigations have been made, but with the limited number of ovaries examined by ordinary histological methods there is nothing to suggest that the early development of the induced corpus luteum is abnormal. 84 Discussion Harrison: Do you think they function, do they produce anything? Rowlands: I think it would be most difficult to prove their functional activity. In two animals in which ovulation was induced very late on in pregnancy the additional corpora lutea had no effect on the occurrence of parturition, they regressed immediately afterwards, at the same time as the corpora lutea of pregnancy. The other possibility is to observe their capacity to inhibit follicular growth, but the latter recurs so rapidly after normal ovulation in the guinea pig that little time is available for observations to be made. Certainly, for the first 4-6 days after induced ovulation and in the presence of induced corpora lutea, vesicular follicles are absent. But at eight days after induced ovulation they have reappeared. To this extent, there is some slight evidence that they are functional. Krohn: Do I understand correctly that you cannot do in the guinea pig what you can do in the rabbit — prolong the duration of pregnancy by inducing a new set of corpora lutea ? Rowlands: No. In the two animals in which I induced ovulation very late on in pregnancy, so that at the time of expected parturition the induced corpora lutea would be in a fairly active state, they did not inhibit parturition. Krohn : Is the corpus luteum necessary in the later stages of pregnancy for the pregnancy to continue? Rowlands: The guinea pig is, of course, a rather peculiar animal in this respect for the effect of ovariectomy on pregnancy is not quantal, that is, abortion does not occur in all animals after this treatment. Loeb and Hesselberg, many years ago, claimed that the corpora lutea are required only for about the first 6 days after mating, but that ovari- ectomy on the 3rd to 5th days of pregnancy caused some to abort on about the 12th day. Jost: This has been studied again recently in Prof. Courrier's labora- tory by Artunkal and Colonge in 1949. They found that castrating the guinea pig before day 16 always produces abortion, but this abortion may be avoided by progesterone. Later castration does not induce abortion. So this is good evidence that the corpus luteum is neces- sary during the first 16 days of pregnancy but afterwards it may be suppressed. Huggett: Of course, if the corpus luteum is persistent the placenta may not have to do quite so much. They may be supplementing each other. T.-Duplessis: Do you have any data concerning the growth of the corpora lutea during delayed pregnancy, for example in lactating rats or other animals ? Rowlands: Not in the guinea pig but in the bank vole (Clethriomys glareolus) there is a slow initial growth phase when the egg is free in the tube and in the uterine lumen, which is followed by a second burst of growth following the stimulus provided by implantation. T.-Duplessis: It would be interesting to know what happens in animals like the mink where I believe you have delayed pregnancy of about four or five months. Discussion 85 Rowlands: I have no information to offer about mink, but in various species of seals in which implantation is also delayed for some months after mating, Prof. Harrison has described changes in the corpus luteum coincident with implantation. Harrison: My impression was, in examining that material, that the corpus luteum of ovulation, if you like to call it that, was not properly vascularized, and that at the time of implantation, perhaps some two or three months later, the corpus luteum became revascularized, and that before implantation there was an outbreak of vacuolation of the luteal cells, which went when implantation occurred. Matthews: Something similar seems to occur in the badger also. Amoroso: With regard to the persistence of corpora lutea, the cat resembles the guinea pig more closely than any other mammal. In the cat the corpora lutea of pregnancy persist throughout gestation and involute rapidly only after parturition. Consequently, should a new pregnancy be established early in the post-partum period two genera- tions of corpora lutea are recognizable in the ovaries, one set from the preceding pregnancy and a new set for the concurrent pregnancy. In these circumstances the older set, coming under the dominance of the same endocrine influences as the new set, show signs of resurgence and follow the same growth pattern as the latter; never, however, do they again attain their former maximum size. In the cat the corpora lutea attain their maximum size during the period of implantation, so it may well be that Blandau's figures for the guinea pig covering the period sixth day to twelfth day may actually represent the era when the ovum is actually entering the endometrium. Accessory corpora lutea, such as are found in the mare, are sometimes encountered in the ovary of pregnant cats during the seventh week of gestation. OBSERVATIONS ON THE CYTOMORPHOSIS OF THE GERMINAL AND INTERSTITIAL CELLS OF THE HUMAN TESTIS Don W. Fawcett and Mario H. Burgos Department of Anatomy, Harvard Medical School, Boston, Massachusetts The growth of our knowledge of the histophysiology of the testis has been closely tied to the development of new methods for the study of its structure. The principal events in spermato- genesis were carefully worked out in the late 1800's on tissues prepared by classical cytological staining methods (Lenhossek, 1898 ; Meves, 1899) and further refinements of these procedures, combined with keen observation under the light microscope, have continued over the years to disclose valuable additional details (Duesberg, 1909; Gatenby, Beams and Woodger, 1921; Gresson and Zlotnik, 1945). Of late, the phase contrast microscope has provided a means of confirming, on living cells, many of the earlier observations made on fixed material (Gresson, 1950; Austin and Sapsford, 1951). The use of modern histochemical staining reactions has permitted a partial chemical characterization of certain cytoplasmic components of the germinal and interstitial cells (Montagna, 1952; Mancini, Nolazco and De la Baize, 1952; Leblond and Cler- mont, 1952). Improvements in methods of fixation and micro- tomy have now made it possible to bring the high resolving power of the electron microscope to bear upon the remaining problems in the finer structure of the testis (Watson, 1952; Challice, 1953). The present paper summarizes some results of the application of this powerful new tool to the study of the complex cell transformations in spermatogenesis and the cycle of development and ageing of the human interstitital cells. 86 Cytomorphosis of Human Testicular Cells 87 Observations on Spermatogenesis The Spermatocyte divisions Electron micrographs disclose that, in the division of the secondary spermatocytes, karyokinesis proceeds normally but cytokinesis is delayed. Therefore, binucleate spermatids are fairly common and, in these, differentiation of an acrosome often begins in association with each nucleus before division of the cell body occurs. When constriction does take place, the daughter halves remain connected by a narrow proto- plasmic bridge throughout the early stages of their trans- formation into spermatozoa (Fig. 1). The connection between these cells is somewhat larger than the intermediate body (Zwischenkorper) that is seen in many other cell types during telophase of mitotic division and it persists longer. With the osmic acid fixative used, the bridge contains no visible rem- nants of the spindle apparatus. Although it is true that some of the testicular biopsies examined were obtained from persons being studied clinically for infertile marriage, it is doubtful whether these atypical cell divisions should be con- sidered pathological, for they are seen with about the same frequency in sections of cat testis (Burgos and Fawcett, 1955) and in the monkey, rabbit and toad as well. These persisting intercellular connections seem to have been overlooked in studies on mammalian spermatogenesis with the light microscope. However, McGregor (1899) reported that the secondary spermatogonia, primary and secondary spermatocytes and the spermatids of amphiuma are con- nected in pairs or groups of four by "intermediate bodies which persist unusually long after telophase". It is not known at just what stage in the differentiation of the mammalian spermatids they become completely separated from one another. It is tempting to speculate that some of the double anomalies observed among the spermatozoa of the human ejaculate may have their origin in an exaggeration of this normal tendency for the germ cells to remain connected by in- tercellular bridges through the early stages of spermatogenesis 88 Don W. Fawcett and Mario H. Burgos Formation of the acrosome and head cap The large juxtanuclear body of the spermatid which is commonly referred to as the idiosome-Golgi complex, appears under the electron microscope as an aggregation of minute vacuoles and broad, flattened vesicles (Figs. 2 and 3). The latter are often closely approximated in parallel array and hence, in section, have a lamellar appearance. The small vacuoles seem to arise by being budded off from the edges of the flattened vesicles. With cytological staining methods, the parallel arrays of membranes are apparently impregnated more heavily with osmium or silver than is the rest of the complex and they are identified with the light microscope as rod-like or crescent-shaped Golgi bodies (dictyosomes). The masses of minute vacuoles, staining less heavily, have been designated the idiosomal material (archoplasm). Inasmuch as the Golgi complex in all other cell types examined to date also contains both the minute vacuoles and the parallel arrange- ments of membrane-bounded vesicles, there appears to be no reason for retaining, in the case of the spermatid, the special terms "idiosome", "archoplasm" or "sphere" for a part of this structure. Instead, the whole body is to be regarded as a particularly well-developed Golgi complex, differing in no essential respect from that of other cell types. Early in the differentiation of the spermatid one or two large granules are formed within separate vacuoles of the Golgi complex. These proacrosomal granules are moderately dense to electrons and are generally homogeneous with present resolutions. Coalescence of the vacuoles, and of the granules which they contain, results in the formation of a single sizeable acrosomal granule within a rather large acrosomal vesicle. The latter is bounded by a distinct membrane and, in life, probably has a fluid content but this is represented in the electron micrographs only by a faint, flocculent precipitate seen in the vacuole around the acrosomal granule (Fig. 3). The acrosomal vesicle approaches the nucleus and adheres to its anterior pole and at the same time the granule in its Fig. 1. A pair of spermatids (Spd) joined by an intercellular bridge (at arrow) through which there is free communication from one cell to the other. The spermatids are completely surrounded by Sertoli cells (Se C). facing page 8& jl' ~ GC Ac V Figs. 2 and 3. Corresponding areas of two differentiating numan spermatids. In eaeh, the acrosomal vesicle (Ac V) is closely adherent to the nucleus (N) and its content is slightly more dense than the surrounding cytoplasm. The acrosomal granule (Ac Gr), enclosed within in the vesicle, is somewhat flattened and occupies a shallow depression in the nucleus. There are several sizeable vacuoles and numerous smaller ones in the neighbouring Golgi complex (Go C). These are Jbelieved to coalesce with the acrosomal vesicle and so contribute to its progressive enlargement. GoC Ac Gr i ■n * *. • -^ 1 -fat Ac V i * • JF> N . •* > * •'.» 4 v J** ,. , ■ k.. •• • « < ':'/'*'<*. '•'.'■ .'•' V v 1 '• y ' «*" Fig. 4. The acrosomal vesicle (Ac V) has spread out over the nucleus (N). Its lumen is narrower and the Golgi complex (Go C) is less closely associated with its surface than in the earlier stage (Fig. 3). The hemispherical acrosomal granule (Ac Gr) is situated at the tip of the elongated nucleus. Fig. 5. The acrosomal vesicle here extends down over the anterior half of the nucleus to form the head cap (C). Its lumen is reduced to a narrow cleft and its content is considerably more dense than before. The acrosome granule (Ac Gr) is markedly flattened. Fig. 6. The head of a spermatid in a slightly more advanced stage of development than that depicted in Fig. 5. The cap consists of a homogeneous layer of moderate density bounded by outer and inner membranes which are continuous at the posterior margin of the cap (*). A layer of osmiophilic granular material is visible on the inner aspect of the nuclear membrane where it is in contact with the cap. A cross- section of a tail flagellum (T) is seen at the lower right of the figure. Fig. 7. A late spermatid showing the coarse granular character of its nucleus. The head cap (C) is visible, covering the anterior two-thirds of the head. The post-nuclear cap (Pn C) appears as a clear space within the nuclear membrane at the base of the nucleus. Tail filaments (T) and ring centriole (Ri C) are also included in the section. Cytomorphosis of Human Testicular Cells 89 interior becomes fixed to that portion of the wall of the vesicle that is attached to the nuclear membrane. The Golgi complex thereafter remains closely applied to the surface of the acrosomal vesicle (Fig. 3) and gives rise to numerous vacuoles which are believed to coalesce with the vesicle, contributing in this way to its progressive enlargement. The acrosomal granule also grows, seemingly, by addition of material to its surface from the surrounding fluid. The granule at first is almost spherical in shape and projects well into the vesicle, but later it becomes flattened and frequently occupies a shallow depression in the end of the nucleus (Figs. 2 and 3). As the area of contact between the enlarging acro- somal vesicle and the nuclear membrane increases, a layer of dense granular material is deposited on the inner aspect of that part of the nuclear membrane which is covered by the vesicle (Figs. 4, 5 and 6). The vesicle gradually extends down over the end of the nucleus like a double-layered stocking cap and, as it does so, its lumen is reduced to a narrow cleft (Figs. 4 and 5). Meanwhile, in the elongation of the spermatid, the Golgi complex and the associated cytoplasm migrate to the posterior pole of the nucleus and the plasma membrane at the anterior end of the cell is brought into close contact with the outer membranous wall of the acrosomal vesicle or head cap. The substance of the granule, which originally was con- fined to the anterior pole of the nucleus, now spreads laterally, filling the narrow cleft between the layers of the cap until it extends all the way to the posterior margin. The head cap at this stage, therefore, consists of outer and inner membranes which are continuous at the posterior margin of the cap, and between these is a thin layer of homogeneous material derived from the substance of the acrosomal granule (Figs. 6 and 7). Thus, the study of electron micrographs has clarified the nature of the idiosome and has reaffirmed the origin of the acrosome from the Golgi complex; it has established that the acrosomal vacuole or vesicle is not a fixation artifact but is a real structure which has an important role in the forma- tion of the head cap of the spermatozoon. 90 Don W. Fawcett and Mario H. Burgos Condensation of the karyoplasm After formation of the head cap a sequence of striking changes is observed in the nucleus. Throughout the early stages of differentiation, the karyoplasm of the spermatid consists of closely-packed fine granules (circa 100 a) of rather low density, interrupted here and there by coarser granules of somewhat greater density (Figs. 4 and 5). The nucleolus is inconspicuous and indefinite in outline. During the period of elongation of the nucleus, the fine granules in the karyoplasm are gradually replaced by coarser granules. These probably arise by agglomeration of the pre-existing fine granules. However, more seems to be involved than a change in the state of aggregation of the nuclear material for, associated with the increase in granule size, there is a marked increase in their osmiophilia (Fig. 7). This must reflect an increase in the number of unsaturated linkages. As differentiation continues and the coarse granules become more closely packed, some sperm heads show sizeable clear areas of irregular shape which, no doubt, correspond to the " vacuoles " that have been observed with the light microscope in the heads of human spermatozoa. These have not been seen in the other species studied. Late in the maturation of the spermatids, the coarse dense granules (300-400 A) coalesce into a smooth homo- geneous mass which once again displays only a moderate affinity for osmium. Concurrently with these changes in the fine structure of the karyoplasm there is a steady reduction in the size of the nucleus and a change in the shape of its sagittal section from obtuse to acuminate. The electron microscope finding, that there is a waxing and waning of the osmiophilia of the nucleus, is reminiscent of the earlier reports in the literature to the effect that there is a progressive increase in nuclear staining with iron haematoxylin during the intermediate stages of spermatid differentiation, followed by a diminished uptake of the dye later on when the spermatozoa are maturing. The increase in intensity of nuclear staining in the course of Cytomorphosis of Human Testicular Cells 91 spermatogenesis has generally been looked upon as the result of a concentration or dehydration of the karyoplasm. The electron microscope observations reported here suggest that the process is not as simple as this and that it involves not only profound alterations in the state of dispersion of the nuclear chromatin but, very likely, changes in its chemistry as well. Observations on the interstitial cells Appearance with the light microscope Human interstitial cells occur as groups of elongated, rounded and polyhedral cells of diverse cytological character. The marked differences in their appearance are interpreted by some investigators as representing different phases in the life- cycle of the Leydig cells (Hooker, 1944; Williams, 1950). Since mitotic figures are not observed in mature Leydig cells, it is commonly assumed that they arise by differentiation from spindle cells present in the interstitium, but it remains unsettled as to whether these cells of origin are common connective tissue fibroblasts, or whether they are primitive mesenchymal cells which persist in the adult along the blood vessels and in the lamina propria of the seminiferous tubules. When the young interstitial cells are still relatively undif- ferentiated, they have a plump, fusiform shape with a nucleus of infolded or irregular outline, a finely granular chromatin pattern and a small nucleolus. Although they resemble fibroblasts in some respects, they are generally larger, often contain a few lipid droplets (Montagna, 1952) and, with some fixatives, their cytoplasm has a fibrillar texture (Sniffen, 1950). In the course of their metamorphosis into typical Leydig cells, the nucleus is said to become round in contour, eccentric in position and to develop a very distinct nuclear membrane and a large nucleolus. The cell volume increases and the cytoplasmic filaments gradually give way to fine acidophilic granules. In the mature Leydig cell, the cytoplasm becomes increasingly heterogeneous and contains 92 Don W. Fawcett and Mario H. Burgos coarse, as well as fine, acidophilic granules, golden-brown lipochrome pigment and lipid droplets of various sizes. In addition, many of these cells contain large angular inclusions known as the crystalloids of Reinke (Reinke, 1896). These are visible in the cytoplasm of unfixed interstitial cells as highly refractile, pale yellow bodies which are usually rod-like, rectangular or trapezoidal in shape, but are often rounded at the ends. They are peculiar to the human testis, and their origin, chemical composition and physiological significance are unknown. Appearance with the electron microscope The electron microscope reveals an even greater variation in the fine structure of the interstitial cells than was appre- ciated with the light microscope, and provides additional evidence of a transition from spindle cells to Leydig cells. The spindle cells in the lamina propria of the seminiferous tubules have elongated nuclei and a homogeneous cytoplasmic matrix of low density to electrons. Their cytoplasm is traversed by a few canalicular strands of ergastoplasm which show vesicular dilations along their length. Mitochondria are few in number and simple in their internal structure. The only cytoplasmic inclusions observed are a few small lipid droplets. The larger fusiform cells that are found along the blood vessels and interspersed with the Leydig cells have round or indented nuclei and their cytoplasm is characterized by the presence of large numbers of extremely thin (50 a) filaments of indefinite length (Fig. 10). These are most abundant in the interior of the tapering cell processes where they generally run parallel to the long axis of the cell. The mitochondria and endoplasmic reticulum tend to be near the periphery of the cell. Many small vesicles (300-500 a) are found just beneath the plasma membrane and these sometimes appear to open at the cell surface. Some of these cells contain sizeable agglomera- tions of dense osmiophilic granules clustered around vacuoles of lower density. Fig. 8. A portion of a Leydig cell, including a little of the nucleus (N) and an adjacent area of cytoplasm containing many minute vacuoles, several mitochondria (M) and a few large homogenous lipid droplets (L) believed to consist of neutral fat. Fig. 9. A representative area of Leydig cell cytoplasm containing mito- chondria (M), homogeneous droplets of lipid (L) and intensely osmio- philic aggregations of pigment (P). The latter are comprised of dense granules of varying size, often associated with vacuoles of homogeneous osmiophobic material, possibly mucoprotein. facing page 92 N I i SL CrV 1*1 M WLm: ■ Fig. 10. A small area from one of the fusiform, immature Leydig cells illustrating the fibrillar character of the cytoplasm. Large numbers of very thin filaments (50 a) are generally oriented parallel to the long axis of the cell. Fig. 11. An area from a typical mature Leydig cell. The cytoplasmic filaments have been replaced by enormous numbers of minute mem- brane-limited vacuoles, giving the protoplasm a foamy texture. The sharp angular structure above is the tip of a crystal (Cr), and the edge of the nucleus (N) appears at the lower right. Fig. 12. A portion of an interstitial cell showing both filaments (F) and small vacuoles (V). A fabric-like pattern is visible in the crystal (Cr) at the left. The mitochondria (M) have a matrix of very low density and the internal membranes are sparse and disorderly in their arrangement. Fig. 13. Low power view of parts of four neighbouring Leydig cells revealing the heterogeneous character of their cytoplasm which con- tains simple lipids (L), complex osmiophilic pigment deposits (P), and crystalloids of Reinke (Cr) occurring in a great variety of shapes. A section of an area of a crystal such as that enclosed in the rectangle is shown at higher magnification in Fig. 14. £ r> Crystalloid / Of Reinke Fig. 14. Electron micrograph of a segment of a crystalloid of Reinke (Cr) disclosing a highly ordered internal structure which presents a a fabric-like pattern. Fig. 15. The end of a crystalloid of Reinke at high magnification showing a pattern of densities spaced a uniform distance apart (186 a) in precise rows intersecting at approximately a right angle. This pattern is interpreted in the diagrammatic overlay as the orderly arrangement of protein macromolecules in a crystalline lattice. Cytomorphosis of Human Testicular Cells 93 The mature Ley dig cells are epithelioid in shape but their membranes are not closely coherent as in epithelial tissues. The nuclei are round and have prominent nucleoli. The cytoplasm is not fibrillar, but, instead, has a vesicular character, owing to the presence of very numerous small membrane-limited vacuoles 100 to 200 millimicrons in dia- meter (Fig. 11). Some of these seem to have a content of very low density but the majority appear quite empty. Their origin and significance is by no means clear. Their mem- branous wall is sometimes observed to be continuous with the plasma membrane in a manner which suggests that they either are formed at the cell surface or discharge their con- tents there. In some Leydig cells these minute vesicles are present in moderate numbers, while in others they are so closely packed that little background cytoplasm can be seen between them. It is easy to speculate that they are somehow related to the specific function of the Leydig cells and that the variations in their abundance from cell to cell reflect different states of physiological activity. The mitochondria are not particularly abundant in Leydig cells and are quite variable in shape. Some are elongated, some spherical, and others have a bizarre, swollen appearance. This pleomorphism is consistent with reports, based upon light microscopy, that the mitochondria of interstitial cells are especially labile and very apt to assume enlarged and distorted forms as a result of delayed fixation (Duesberg, 1918). The mitochondrial matrix is of very low density and the folds of the membrane projecting into the interior of the organelle are fewer and more irregular in their dis- tribution than in the mitochondria of most other tissues. In keeping with the acidophilic staining of these cells in ordi- nary histological preparations, the endoplasmic reticulum or ergastoplasm is poorly developed and the small particles of ribonucleoprotein, which are responsible for cytoplasmic basophilia (Palade, 1955), are present in small numbers. Lipid occurs in the Leydig cells in two forms : homogeneous, moderately osmiophilic droplets (Fig. 8), and heterogeneous 94 Don W. Fawcett and Mario H. Burgos aggregates of intensely osmiophilic, granular material (Fig. 9). The former are interpreted as droplets of neutral fat, while the latter are believed to belong to the group of acetone- insoluble lipid complexes often referred to as lipochrome pigments or ceroid. The intensely osmiophilic granular material is often aggregated around the periphery of vacuoles of a homogeneous substance which does not reduce osmium. Sharp angular clefts which sometimes occur in the midst of these conglomerations are interpreted as negative images of crystals which have been dissolved out during specimen preparation. The significance of these pigment masses is not clear, but they are of very common occurrence in the cells of most steroid-producing endocrine glands. The amount of lipid present in the mature Leydig cells is highly variable. Some of them are virtually devoid of droplets of neutral fat but contain considerable pigment. Others have large amounts of both. In electron micrographs, the crystalloids of Reinke have a complex and highly-ordered internal structure which presents a regular pattern bearing a superficial resemblance to that of a woven fabric (Fig. 14). The pattern varies, depending upon how the object is cut, but in crystals which are most favourably oriented with respect to the plane of section, it is possible to resolve a pattern of densities 100-150 A in dia- meter, uniformly spaced about 190 A apart along two axes which are approximately at right angles to one another (Fig. 15). The length of the period along the third axis has not been determined. This pattern is thought to represent the arrange- ment of macromolecules in the lattice of a protein crystal. A few of the crystals are rectilinear, with straight sides and sharp angles, but the majority have rounded contours, parti- cularly at the ends (Fig. 13). There are indications that they are almost as soft as the surrounding cytoplasm for, when sectioned, they cut perfectly smoothly without tearing or setting up vibrations in the block. Protein crystals are apt to be rounded at the ends rather than sharp-angled and, in the "wet" state, they are quite soft. Cytomorphosis of Human Testicular Cells 95 Heretofore, the assumption that the crystalloids of Reinke are protein has been based entirely upon their solubilities and staining reactions as studied with the light microscope. The present study presents additional evidence of their protein nature in their rounded shapes and soft consistency, and it provides the first visual demonstration of their large molecular size and crystalline internal organization. It is apparent that the crystals of Reinke are entirely different from the crystal- loids of Spangaro and Lubarsch which are found in the epi- thelium of the seminiferous tubules. In addition to the spindle-shaped interstitial cells and the typical Leydig cells just described, many cells are observed which appear to be intermediate between these two cell types. These transitional forms have some of the cytological charac- teristics of both. Thus, one may find bundles of delicate filaments, typical of the spindle-shaped interstitial cells, intermingled with numerous sub-microscopic vacuoles of the sort commonly found in mature Leydig cells (Fig. 12). The occurrence of such cells seems to support the contention of earlier workers who postulated a gradual metamorphosis of the spindle cells into mature Leydig cells. Moreover, the striking variability among mature interstitial cells with respect to the degree of vacuolation of their cytoplasm and the abundance of their lipid droplets, pigment granules and protein crystals, strongly suggests that they are of very different ages or that they are in different phases of a cycle of physiological activity. There seems to us to be little to sup- port the contention that the Leydig cells clear themselves of these accumulated inclusions and revert to spindle cells (Rasmussen, 1932; Williams, 1950), but there is some reason to suspect that they age and ultimately degenerate. Certainly many cells are seen which are crowded with lipid and pigment and have grossly swollen, distorted mitochondria, and some of these also show large angular areas which, at high magni- fication, are pale, poorly defined and devoid of visible internal structure, as though they were "ghosts" of crystals which had undergone dissolution. Although dependable criteria for the 96 Don W. Fawcett and Mario H. Burgos recognition of senility in cells at the electron microscope level have not yet been established, it is difficult to escape the conclusion that some of the Leydig cells are maturing while others are no longer physiologically active and are, perhaps, in a state of impending degeneration. Acknowledgement. We are greatly indebted to Dr. Fred A. Simmons of the Department of Gynecology of Harvard Medical School for making available to us the human testicular biopsies upon which this study is based. REFERENCES Austin, C. R., and Sapsford, C. S. (1951). J. R. micr. Soc, 71, 397. Burgos, M. H., and Fawcett, D. W. (1955). J. Biophys. Biochem. CytoL, 1, 287. Challice, C. E. (1953). J. R. micr. Soc, 73, 115. Duesberg, J. (1909). Arch. Zellforsch., 2, 137. Duesberg, J. (1918). Biol. Bull, 35, 175. Gatenby, J. B., Beams, H. W., and Woodger, J. H. (1921). Quart. J. micr. Sci., 65, 265. Gresson, R. A. R. (1950). Quart. J. micr. ScL, 91, 73. Gresson, R. A. R., and Zlotnik, J. (1945). Proc. roy. Soc. Edin. B, 62, 137. Hooker, C. W. (1944). Amer. J. Anat., 74, 1. Leblond, C. P., and Clermont, Y. (1952). Amer. J. Anat., 90, 167. Lenhossek, M. (1898). Arch. mikr. Anat., 51, 215. McGregor, J. H. (1899). J. Morph., 15, (Suppl.), 57. Mancini, R. E., Nolazco, J., and De la Balze, F. A. (1952). Anat. Rec, 114, 127. Meves, F. (1899). Arch. mikr. Anat., 54, 329. Montagna, Wm. (1952). Fcrtil. and Steril., 3, 27. Palade, G. E. (1955). J. Biophys. Biochem. CytoL, 1, 59. Rasmussen, A. T. (1932). In Special Cytology, 2nd edn., ed. by E. V. Cowdry, Vol. 3, p. 1673. New York: Hoeber. Reinke, F. (1896). Arch. mikr. Anat., 47, 34. Sniffen, R. C. (1950). Arch. Path., 50, 259. Watson, M. (1952) A. E. C. Proj. Rep. UR— 185 Univ. of Rochester. Williams, R. G. (1950). Amer. J. Anat., 86, 343. DISCUSSION Montagna: I have scarcely recovered from the breath-taking beauty of these preparations. Before I make any comments (which unfortu- nately have very little to do with Dr. Fawcett's presentation) may I say that I have been accused of dealing only with the testes of feeble-minded individuals. This is largely true because they are the only people we Discussion 97 can convince that castration is a good thing! We have supplemented our studies with specimens of testes obtained from volunteer prisoners. Ageing in the human testis is a very gradual process. There is no period at which ageing processes are really precipitous. The testes from a man, 72 years old — the oldest individual we have — still look pretty good, and still have a certain amount of spermatogenesis going on. Beginning from about 45 years or so, and this is variable with indi- viduals, spermatogenesis gradually slows down and together with this there is an increase in the number of cells which look like Sertoli cells. I call them Sertoli cells because their nuclear morphology and cyto- plasmic inclusions, such as glycogen and lipid granules, are identical with those of Sertoli cells. The oldest testes usually contain numerous Sertoli cells and sper- matogonia and no other cells. The seminiferous tubules do not become aged all at once. That is, whereas one seminiferous tubule might appear entirely normal, one next to it might show strong ageing changes. Some tubules show a banding of ageing changes, a phenomenon which I shall discuss again when I speak about the sweat glands later on. Within the same tubules a segment might show ageing changes, but other segments might still have spermatogenesis. The basement membrane of the seminiferous tubule, which is composed of three layers, undergoes some interesting changes. The middle hyaline or glassy layer of the mem- brane becomes disproportionately large with advancing age. I want to come back to the general misconception of calling the Sertoli cells the Sertoli syncytium. In very young or in very old testes, where Sertoli cells are numerous and not mixed with all the spermatogenic cells, one can see, even with the obsolete methods of the light or phase contrast microscope, a fairly distinct membrane between the individual cells. I am glad to hear Prof. Fawcett refer to certain cells in the inter- stitium as fibroblast-like. In human testes there are, indeed, cells which one cannot tell from fibroblasts. These cells, however, always predict- ably contain some perinuclear sudanophil bodies. Although even the fibroblasts in connective tissue elsewhere may have visible lipid inclu- sions, these are more striking in the fibroblasts of the testes. I fancied some years ago seeing a transition from ordinary fibroblasts to com- pletely differentiated Ley dig cells. I say ' I fancied ' because I have not the vaguest idea if we have a bimodality of action in the fibroblasts : one going towards the fibroblasts and one going towards the Leydig cells. I would like to hear Prof. Fawcett express his opinion on this matter. Fawcett: Up to a year ago I was doubtful of the reported transition from fibroblasts to Leydig cells but it is certainly true that in electron micrographs there are two distinct categories of cells in the interstitium of the testis ; those that have a filamentous cytoplasm and few in- clusions, and those whose cytoplasm is filled with minute vesicles and contains a variety of lipid, pigment and crystal inclusions. One also finds cells with intermediate characteristics having both the little vesicles and the delicate filaments in varying proportions. It is easy to AGEING VOL. 2 5 98 Discussion believe that these cells are transitional forms between fibroblasts and mature Leydig cells. In the course of this study, therefore, I have become more receptive to the suggestion that has often been made in the literature to the effect that fibroblasts or fibroblast-like cells are transformed into Leydig cells during adult life. Zuckerman: Would it be possible to arrange any experimental con- ditions in which the relative frequency of the transition from one to the other type of interstitial cell could be studied ? I recall that when most cells which could possibly be spermatogonia disappear from the semini- ferous tubule, and the cells of Sertoli become denned with considerable clarity, one occasionally finds that the interstitial cells swell. Fawcett: It might be possible to study this, but even if one could set up experimental conditions that would change the relative frequency of the interstitial cell types, it would be difficult to make statistically valid observations on this with the electron microscope, because one's field of view is so restricted and the sampling of the cell population so small that I believe the experimentally produced shifts would have to be very great to be detected at all. Wislocki: With reference to the interstitial tissue, I recall that it varies considerably — histologically — in different mammals. It would be interesting to examine different types of interstitial tissue with the electron microscope. I recall that in rodents, for example in mice and rats, that the cells are more fusiform or spindle-shaped with fewer epitheloid cells of the Leydig type than one sees in man, or, to cite another example, in the testes of deer. Thus there is a range to be explored with reference to various cytological types. I presume that you believe that the ketosteroids of the male gonad are secreted by the Leydig cells. In that event, would you equate the lipid droplets or any other organelles which you have seen in these cells with the production and storage of ketosteroids ? Fawcett: I presume that the ketosteroids are present principally in the Leydig cells, and their solubilities are such that it is most likely that they are in the lipid droplets that are preserved with osmic acid. As you know, lipid is not nearly as abundant in human interstitial tissue as it is in the cat. All of the slides of interstitial cells that I have pre- sented have been from the human testis. The greater amount of lipid in the cat testis made it technically more difficult to study. Adjacent Leydig cells may differ greatly in the abundance of the minute vesicular component of their cytoplasm. We have wondered if this variation reflected a difference in their secretory activity, but what- ever the content of these vesicles may be, it is not preserved by osmium fixation and probably is not a steroid containing secretory material. Wislocki : Pollock, a pupil of Stanley Bennett, working in my depart- ment (Pollock, W. F. (1942). Anat. Rec, 84, 23) described the inter- stitial cells of the cat's testes as containing cytoplasmic material which gave a positive phenylhydrazine reaction and birefringent crystals which were increased by treatment of the tissue with digitonin. It was concluded that these reactions, plus certain solubility properties, indi- cated the presence of steroid compounds in the cells. I wonder whether Discussion 99 the findings with the electron microscope have any bearing on the possible demonstration of ketosteroids in the cat's testis. Fawcett: Yes, that is correct but I believe subsequent work has raised some doubt as to the value of these reactions for the identification of steroids. In human testis the results of the application of these histo- chemical methods are particularly difficult to interpret because some of the same staining reactions are given by both the acetone-insoluble aggregations of pigment and the lipid droplets, and it would be sur- prising if the pigment contained ketosteroid. Montagna: If one assumes that this series of tests reveals the sites of ketosteroids, or hormone, or what have you, in the human testis one must include (1) the primary spermatocytes and spermatogonia, and (2) the Sertoli cells as sites of these substances. These elements give reactions identical with those found in the Leydig cells. Fawcett: By electron microscopy, the Sertoli cells contain an abund- ance of neutral fat droplets but pigment is relatively uncommon. We have very rarely found either lipid or pigment in spermatogonia or spermatocytes. Montagna: Judging only by the osmiophilia? Fawcett: Judging by their appearance in electron micrographs of osmium fixed tissue. Osmiophilic pigment aggregations are easily dis- tinguished from neutral fat in such preparations. MITOCHONDRIAL CHANGES IN DIFFERENT PHYSIOLOGICAL STATES Edward W. Dempsey Department of Anatomy, Washington University Medical School, St. Louis, Missouri Mitochondria, denned by light microscopy as organelles stainable by selective stains and intravitally by Janus green, have been identified in sections examined with the electron microscope. They appear after osmic fixation as ovoid or elongated structures with an outer limiting membrane. This membrane is a double one. Its inner lamina is folded or extended into many inward projecting folds, each of which is composed of two closely apposed membranes. Enclosed by the limiting membrane and between the inner folds is a homogeneous, moderately dense matrix, in which there occurs occasional dark granules. The number and complexity of the internal laminar folds, the density of the matrix and the size and number of the granules all exhibit variations from tissue to tissue, but are reasonably similar from cell to cell in the same tissue. We have examined mitochondria from several tissues in different physiological states, or after various experimental procedures. These organelles prove easily susceptible to alteration. They swell rapidly if fixation is delayed after the death of the animal. This post-mortem swelling can be prevented, however, by subjection to hypertonic solutions. Mitochondria behave, therefore, somewhat like osmometers. After intravital exposure to several foreign or toxic reagents, mitochondria appear to concentrate the foreign substances and to become altered in the process. In rats, to whose drinking water silver nitrate had been added for long periods of time, deposits of dense granules occurred in basement membranes and in macrophages throughout the body and, 100 Changes in Mitochondrial Appearance 101 in addition, in occasional parenchymal cells. In the liver and pancreas, mitochondria, identifiable by their internal folds, were observed to contain silver granules. Further deposi- tion of silver caused disruption of the internal folds and the formation of intramitochondrial vacuoles. Similar changes in mitochondrial appearance have been noted in our laboratory by Dr. Jules M. Weiss after intravital administration of neutral red. We have also seen occasional macrophages from normal animals in which granular aggregates were segregated within mitochondria. The endodermal epithelium lining the yolk-sac cavity is a transient tissue in that it, like the rest of the placenta, lives its life only during the period of gestation. Moreover, in many animals, the yolk-sac is a flourishing organ early in pregnancy but becomes reduced in size and relative importance after the establishment of the chorioallantoic placenta. At mid- gestation in the guinea pig, mitochondria in the yolk-sac epithelium are normal in appearance and have well-defined internal membraneous folds. Near term, however, the mito- chondria are swollen, their internal folds have vanished and many seem to have been converted into pigmented structures. In the mare's yolk-sac, which becomes reduced in size during the middle span of pregnancy, numerous cells become de- tached from the basement membrane. In these effete cells, swollen and disorganized mitochondria are the rule. These changes are not ascribable to poor or delayed fixation, since adjacent cells may contain mitochondria of normal size and structure. The adrenal cortex can be regarded as a transient tissue in that moribund and dead cells are frequent in the inner, juxta- medullary zone, whereas mitoses are ordinarily encountered only in the outer zones. These and other cytological con- siderations lead to the assumption that cells in the outer portion of the zona fasciculata are in an active secretory phase; those in the inner fasciculata are becoming exhausted and those in the zona reticularis are dying. The electron microscopical appearance of these zones has recently been 102 Edward W. Dempsey studied in our laboratory by Dr. Jeffrey D. Lever. In the outer portion of the fasciculata, mitochondria are numerous and closely packed, they frequently have bizarre shapes and are closely associated with a system of membrane-lined vacuoles or tubules. Their internal structure differs from that exhibited elsewhere; the internal membranes here are tubular rather than laminar in configuration. The density of the mitochondrial matrix varies greatly, some mitochondria being excessively osmiophilic. In the zona reticularis the cells are smaller and the mitochondria are less closely packed; thus, it appears that with exhaustion of the cells, there is a loss of mitochondria. Upon stimulation of the gland by exposure to cold or by administration of adrenocorticotrophic hormone, the mitochondria initially become less osmiophilic and their membranous structures become more sharply evident. More prolonged stimulation results in a restoration of the original osmiophilia. In prolonged stimulation, bizarre forms become more common and their structural alterations are more extreme. The limiting membranes are frequently re-dupli- cated and the internal folds become more complex. "Open" forms have been encountered in which the internum of the mitochondrion is continuous with the cytoplasm through pores in its limiting membranes. These observations in various tissues and in different physiological states suggest that mitochondria are labile organelles, easily altered by changes in their physical environ- ments. The experiments on vital staining show that the inclusion within cells of foreign agents is of considerable consequence to the integrity of their mitochondria. Ageing and moribund cells, exemplified by the yolk-sac and the zona reticularis of the adrenal gland, contain fewer and degenerat- ing mitochondria. The experimental changes induced in mitochondria in the secretory zones of the adrenal gland offer hope, with further study, of understanding the manner in which they are formed and their relation to secretory processes. Discussion 103 DISCUSSION Montagna: I would like to ask Prof. Dempsey whether or not he con- siders the mitochondria as fixed rather than labile structures. I find it difficult to reconcile the idea of mitochondria as rigid structures with the observations of Frederic and Chevremont and with those of the Lewises. These investigators all observed living mitochondria undergo numerous vicissitudes around the nucleus. Dempsey: I think there is no question but that the mitochondria are extremely labile structures. Bensley showed many years ago that if one places a rat on a diet of nothing but sugar and water, within a week the mitochondria practically disappear from the cells of the pancreatic acini. Dr. Weiss in my laboratory repeated that experiment a year or two ago and confirmed Bensley's results. He then continued by re- feeding the animal a normal diet, after the exhaustion of the mito- chondria, and was able to show that the organelles regenerate with extraordinary rapidity. Within a matter of hours after administration of protein materials in the diet, the pancreatic cells again have their normal complement of mitochondria. In this kind of experiment it seems to me perfectly clear that they are labile structures. Also, in the adrenal cortex, there are many more mitochondria in the secretory zones than there are in those cells in the degenerating zones. If one permits the assumption that these are two different phases in the life-cycle of the same cell, then it would appear that mitochondria had been exhausted from the cell during its life-span. Wislocki: This paper raises many questions about mitochondria; for example, what are their life-spans, to what degree do they degenerate, to what extent and how do they divide and renew themselves and what changes do they undergo under different physiological and pathological conditions? The electron microscope should, in the next several years, provide answers to many of these questions. Dempsey: Prof. Wislocki's remark reminds me of something I had intended to say and forgot. You, Prof. Wislocki, I believe inadvertently, remarked that the mitochondria perhaps divide and renew themselves. I am not sure that we should assume that the organelles within a cell derive themselves from pre-existing organelles. We are so familiar, I believe, with the cell doctrine and with the dictum that all cells come from previously existing cells, that imperceptibly our minds tend to slide into the similar syllogism that all structures come from pre- existing structures, but they do not necessarily; and it could well be that mitochondria are not produced from previously existing mito- chondria but arise from some other structure in the cell. Wislocki: That is a possibility. Dempsey: I think the question ought to be kept open. Fawcett: The origin of mitochondria from pre-existing mitochondria has been well documented by phase contrast cinematography of living cells in tissue culture. In these films one sees long filamentous mito- chondria dividing into several segments, and these detached segments continue their activity in the cytoplasm and apparently constitute new mitochondria. Under certain experimental conditions, we have seen 104 Discussion mitochondria in electron micrographs of liver that appeared to be under- going division but we must admit that the pictures could equally well be interpreted as two mitochondria coalescing. I believe the electron microscope will ultimately settle the question as to whether mito- chondria divide or are formed anew from other components of the cytoplasm ; but I have been disappointed in our own work to date, in that we seldom obtain pictures that provide clear-cut evidence for one or the other of these alternatives. Dawes: I was puzzled by the use of the word "membrane" ; could you explain that, Prof. Dempsey ? It seemed rather like the story of Frankie and Johnny — that some of the membranes had no beginning and no end. Are they partitions? Dempsey: Well, they are thought to be membranes in that mito- chondria may be shown to be active osmotically. I quoted part of the evidence. It can also be shown that they can be swollen and shrunk experimentally by changing the tonicity of the suspending medium. The structure that we see is a line in a micrograph which has a density greater than the surrounding. The density, we strongly suspect, is due to the fact that it has reduced osmium tetroxide. We have to have some kind of terminology, and "membrane", in view of these two facts, seems to be tentatively reasonable. Wislocki: We shall have to experiment with the effect of fixatives upon the preservation of the mitochondria in their most natural state. Besides artifacts introduced by fixation, mitochondria are subject to functional as well as age changes which must be recognized. It is difficult, with just the one or two fixatives now at our disposal, to say whether an organelle is normal or pathological, or merely showing arti- factual changes as a result of fixation. Dempsey: There is only one technical way that I know in which this difficult problem can even be approached. If in a tissue with a mixed cell population, it is found that one cell type exhibits unaltered mito- chondria whereas the other cell type exhibits altered mitochondria, it seems difficult to implicate the quality of fixation as an explanation for the alterations. This is the criterion we have used. Wislocki: I have observed that identical cells will vary considerably in their appearance in the electron microscope, depending upon whether they were fixed in either osmic acid or a chromic acid fixative. Dempsey: Differing in some details, but I think not in the configura- tion of these so-called membranes in the mitochondria — I believe it is right, Prof. Fawcett, that the density of the matrix of mitochondria may be reduced by over-fixation with Palade's fluid? If one fixes for a long time, this material leaches out of the specimen, so that by removal of the matrix, one obtains an enhanced contrast of the essential skeleton of the mitochondria. Fawcett: Yes, prolonged fixation in Palade's buffered osmium solution will extract some of the mitochondrial matrix, but the membranous structures are not significantly altered. For some purposes it is ad- vantageous to decrease the density of the background somewhat by longer fixation so that the membranes stand out in sharper contrast. MORPHOLOGICAL ASPECTS OF AGEING IN THE PLACENTA George B. Wislocki Department of Anatomy, Harvard Medical School, Boston, Massachusetts By modifying a previous definition devised by Flynn for vertebrate placentas, Mossman (1937) proposed that "the normal mammalian placenta is an apposition or fusion of the fetal membranes with the uterine mucosa for physiological exchange". Thus, the mammalian placenta is not a unified organ composed of homologous units, as the liver and kidney, but consists of several membranous structures which enter variously into relationships with one another in different groups of mammals. In any given animal, these membranous tissues collectively form the placenta, although some differen- tiate and function early in gestation, while others develop and assume placental functions later in pregnancy. Hence, the study of the ageing of the placenta is a matter of investigating not a single structural unit, but a variety of organs which may function either successively or concurrently. Thus, while some of the placental structures of a given animal are degenerating, others are growing and undergoing differentia- tion. One need only mention the successive or concurrent combination in mammals of a bilaminar yolk-sac placenta, a choriovitelline placenta, an inverted yolk-sac placenta (com- plete or incomplete), a chorionic placenta and a chorioallantoic placenta. The combinations and sequences of these structures in various placental types are catalogued and described in masterly papers by Mossman (1937) and Amoroso (1952). Many of these placental structures are provisional or transient and serve only temporarily to support the growing conceptus. Thus, the bilaminar omphalopleure and chorio- vitelline membranes may be eventually completely resorbed 105 106 George B. Wislocki (tree-shrew, vespertilionid bats, lemurs (Loris, Galago), horse, coney (Hyrax)). Following complete inversion of the yolk sac (lagomorphs, many rodents), its parietal wall, as well as the chorion and the decidua capsularis associated with it, degenerate completely. In the pig's placenta the un vas- cularized ends of the chorionic sac wither. In the human placenta, the chorion laeve and decidua capsularis are pro- visional in nature and disappear completely by the fourth month of gestation. The majority of these transient foetal structures are functional and subserve physiological exchange until regression sets in. Moreover, in most mammals, the uterine endometrium undergoes various degrees of destruc- tion and resorption during the course of gestation as a result of being invaded by the trophoblast of the chorioallantoic placenta. The degenerating endometrium forms an important source of nutrient material (histotrophe) for the growing placenta and conceptus. From all of this it is evident that considerable portions of the placenta undergo ageing and ultimate death long before pregnancy is over. These losses are compensated by other parts of the foetal membranes which differentiate into the definitive placental structures which mediate physiological exchange until the end of gestation. The definitive parts of the placental membranes of Euthe- rian mammals, which include a chorioallantoic placenta and sometimes a yolk-sac placenta, undergo progressive histo- logical and cytological changes which have been the subject of numerous investigations (cf. Amoroso, 1952). Histo- chemical changes in the placenta have also been investigated in a variety of mammals including the sow (Wislocki and Dempsey, 1946a), sheep (Wimsatt, 1950, 1951), cat (Wis- locki and Dempsey, 1946&), shrews (Wislocki and Wimsatt, 1947), bat (Wimsatt, 1948, 1949), rat (Wislocki and Padykula, 1953) and man. The human placenta of the second and third months of gestation has been compared by cytological and cytochemical methods with the placenta at full term (Wislocki and Bennett, 1943; Dempsey and Wislocki, 1944, 1945; Wislocki and Demp- Morphological Aspects of Ageing in the Placenta 107 sey, 1948; Wislocki, 1955). It has been compared at the two periods with respect to the localization and amount of nucleo- proteins, enzymes, carbohydrates and lipids which it contains. Electron microscopy has also revealed significant cytological differences between the placenta of the first trimester and at full term (Wislocki and Dempsey, 1955a). The cytological and cytochemical changes in the human placenta involve both the nuclei and cytoplasm of the trophoblast (Wislocki, 1955). Mitotic division decreases and various cytoplasmic structures and substances diminish in size and amount. The lipid droplets and mitochondria of the trophoblastic syncytium decrease both in size and number. The brush border of the syncytium seen with the daylight microscope, and the equivalent microvilli observed with the electron microscope, diminish in length. Glycogen gradually disappears from the foetal placenta. Cytoplasmic ribonucleo- protein, seen in the syncytium as cytoplasmic basophilia in the daylight microscope and as ergastoplasmic vesicles in the electron microscope, decreases. On the other hand, acid and alkaline phosphatases increase. The stroma of the villi becomes less cellular and more fibrous. Biochemical methods afford another means of determining the amount and direction of placental changes during gesta- tion. Thus, in the human placenta, Villee (1955) reports a gradual decline in the amount of placental glycogen, in the degree of anaerobic glycolysis, and in the rate of oxygen consumption. With respect to the decline in glycogen, the histochemical and biochemical observations are in agreement. A recent study of the enzymatic reactions of the rat's placenta from the 13th day of gestation until full term reveals a number of changes (Padykula, 1955, 1956). Adeno- sine triphosphatase and glycerophosphatase (pH9-4), acid phosphatase, esterase (Pearse's method) and succinic dehydro- genase (Seligman and Rutenburg's method) were investigated in frozen, cryostat sections. Homogenized visceral yolk sac was also assayed biochemically for succinic dehydrogenase, adenosine triphosphatase and glycerophosphatase activities. 108 George B. Wislocki The several enzymatic reactions were lowest at 13 days in all parts of the rat's placenta. In the visceral endoderm of the yolk sac, alkaline glycerophosphatase reached a maximum at 17 days and declined to almost none by 21 days. Adenosine triphosphatase rose to a maximum at 19 days but then de- clined somewhat by the end of gestation. Acid phosphatase activity rose steadily until the end of pregnancy. Esterase activity reached a peak at 15 days and diminished slowly thereafter. Succinic dehydrogenase activity was maximal on the 16th day and then declined rapidly. Assays of homo- genates of the yolk sac confirmed the histological findings for adenosine triphosphatase, alkaline glycerophosphatase and succinic dehydrogenase. In the chorioallantoic placenta, adenosine triphosphatase activity increased markedly in the spongy zone between the 17th and 19th days, as did also acid phosphatase between the 15th and 19th days. Various cytological and cytochemical changes have been reported in the placenta of the rat (Wislocki and Padykula, 1953). For example, glycogen begins to accumulate in the visceral endoderm of the yolk sac on the 15th day, reaches a peak on the 18th day, and declines thereafter. A mucopoly- saccharide, demonstrable by the periodic acid-Schiff reagents, increases from the 10th to 16th days and then declines (18th day), but increases once more by the 21st day. In the tropho- blast of the placental labyrinth, where glycogen is never abundant, it increases up to the 18th day and subsequently diminishes. In the spongy zone and the subplacental myo- metrium, glycogen is extremely abundant up to the 18th day but then declines rapidly. The decrease of some substances in the placenta may be related to the assumption of functions by various foetal organs, especially the liver. From study of the rabbit's placenta, Claude Bernard (1895) advanced the idea that the placenta serves as a deputy for the synthesis and storage of glycogen until the foetal liver begins to function. In confirma- tion of this, glycogen first appears in the liver of the rat on the 19th day of gestation when it has begun to disappear from the Morphological Aspects of Ageing in the Placenta 109 placenta (Padykula, 1955). Similarly, it has been suggested (Wislocki, Dempsey and Fawcett, 1948) that the intense cytoplasmic basophilia of the early human placenta, which is attributable to ribonucleoprotein, represents a provision for the synthesis of proteins necessary for foetal growth until such time as this function is taken over by the foetal liver. A progressive increase in permeability of the placental barrier has been reported by various investigators. In rabbits, placental permeability to antibodies (Rudolfo, 1934), phenol- sulphonphthalein (Lell, Liber and Snyder, 1932), neoars- phenamine (Snyder, 1943) and radioactive sodium (Flexner and Pohl, 1941a) increases during the course of gestation. Similarly, in the rat there is a gradual increase in permeability to insulin (Corey, 1932) and radioactive sodium (Flexner and Pohl, 19416). Moreover, Flexner and his associates have observed an increase in the rate of transfer of heavy water and sodium across the placental barrier in other mammals (sow, goat, cat, man). These increases in the rate of placental exchange have been ascribed to a progressive reduction in the number of layers of the placenta accompanied by a diminu- tion in width of the barrier, a combination of changes which has generally been regarded as favouring a progressive in- crease in placental "efficiency". Aside from these gradual morphological and physiological changes, the question arises as to whether, beginning some time late in gestation, the definitive placental structures undergo regressive, terminal changes which have an unfavour- able effect on placental exchange, limit the length of gestation, or influence the time of parturition. Such changes, if present, would comply with Lansing's (1952) definition of ageing as a process of unfavourable progressive change which becomes apparent after maturity, is inversely related to growth and involves a decrease in efficiency of the mechanism for reconstruction. There are relatively few substantial morphological grounds upon which to evaluate to what degree the definitive placental structures undergo unfavourable terminal changes which 110 George B. Wislocki depress physiological transfer or influence the onset of labour. In the human placenta, in which more attention has been de- voted to this problem than in animals, unfavourable, terminal morphological changes have been described. These include the accumulation of fibrin, "fibrinoid" and calcium (Hertig, 1946), the " hyalinization " and loss of the syncytium covering a variable number of the chorionic villi (Tenney, 1936), and haemorrhagic infarcts involving the complete sclerosis of placental villi (Thomsen, 1955). However, to what degree and how these changes actually limit or decrease the overall efficiency or individual functions of the human placenta is mainly a matter of conjecture. With respect to animals, the epithelium of the visceral layer of the yolk-sac placenta of the guinea pig, investigated with the electron microscope (Dempsey, 1953), reveals that the mito- chondria at full term are swollen and some are degenerating. Furthermore, the cell cytoplasm is more granular than earlier in gestation and some of the ergastoplasmic structures are swollen. In the rat, however, which has a much shorter gestation period than the guinea pig, similar degenerative age changes have not been observed (Wislocki and Dempsey, 19556). These differences between the ageing of the yolk sacs of the guinea pig and rat raise the interesting question as to whether the definitive placentas of animals with long gestation periods show more terminal age changes than those with short periods. The shortest gestation periods for Eutherian mammals are 16 days for the golden hamster and 17 to 19 days for the short-tailed shrew, whereas the longest periods are over 600 days for the elephant and between 420 and 500 days for the giraffe and rhinoceros. Perhaps the most rewarding compari- son of the degree of terminal ageing with the length of gesta- tion could be carried out in rodents. There, differing from most orders, the length of gestation varies greatly in different species, whereas the definitive chorioallantoic and yolk-sac placentas are relatively similar in their basic structure. One might profitably compare the terminal placental histology of Morphological Aspects of Ageing in the Placenta 111 species of rodents which have the shortest gestation periods (hamsters, mice, rats) with that of members which have periods six to seven times as long (porcupines, beavers). A sharp decline in the last decile of pregnancy has been observed by Flexner and Gellhorn (1942) in the rate of transfer of radioactive sodium across the placentas of various mammals (rodents, rabbit, cat, goat), excepting the sow in which there is no diminution. A similar decline in sodium transfer has also been reported in the human (Flexner et al., 1948), which Flexner (1955) attributes to the deposition of "fibrin over the villi" and to "thromboses". This unfavourable terminal change in placental transfer of sodium would seem to be valid, although there are no morphological expressions of terminal senescence, presently known in the animals cited, with which to correlate the results. Of the various cytomorphic and biochemical changes which occur in pregnancy, it is difficult to say specifically which are progressive and which regressive. The terminal change in sodium transfer found by Flexner in various mammals and the changes in the mitochondria recorded by Dempsey in the guinea pig's ageing yolk sac might well be regarded as regres- sive. The decline in adenosine triphosphatase and of succinic dehydrogenase noted in the rat's yolk sac by Padykula might also be thought of as signalizing regression, especially since adenosine triphosphatase plays a role in general energy release. But for the majority of changes no definite clues exist as to whether they exert favourable or unfavourable influences upon placental exchange. Premature and pathological age changes have been de- scribed in the human placenta, but they differ only in degree from the normal terminal regressive changes alluded to. As a consequence of the difficulty of measuring or assaying quan- titatively the differences between normal and pathological placental ageing, it has proven impractical to distinguish ordinary terminal changes from premature and pathological ones. The toxaemias of pregnancy and eclampsia have been 112 George B. Wislocki variously ascribed to a variety of placental lesions. Premature and excessive degeneration of the syncytial trophoblast has been held responsible for these conditions (Tenney, 1936; Tenney and Parker, 1940), and, since the syncytium is re- garded as the site of formation of placental ketosteroids (Wislocki and Bennett, 1943; Wislocki, 1955), the observed changes in ketosteroid metabolism in the toxaemias have been ascribed to syncytial degeneration (Smith and Smith, 1948; Sommerville, 1950). Bartholomew and his associates (1932, 1934, 1936) have attributed the toxaemias to spastic constric- tions of the foetal blood vessels of the chorionic villi, resulting in the formation of red infarcts with necrosis of the villi in the infarcted areas. Thomsen (1955) agrees with him, except that he regards the placental lesions as secondary rather than as the primary precipitating cause of eclampsia. Many other speculative explanations of the toxaemias of pregnancy have been advanced (cf. Dieckmann, 1952). Schneider (1950), for example, attributes these conditions to the toxic effects of thromboplastin liberated from pathological haematomas which frequently occur between the decidua and the basal placental plate late in pregnancy. Page (1953) offers a theory involving an unidentified vaso-toxic substance produced in the placenta, combined with a high sodium intake and the presence of large quantities of placental steroids. In two cases of pre-eclampsia, Flexner and his associates found a marked reduction in placental permeability to radio- active sodium. In toxaemia the ability of the placenta to con- centrate amino acids is also decreased (Crumpler, Dent and Lindan, 1950; Villee, 1955). One avenue of investigation of terminal placental ageing remains to be mentioned, namely, the postmature retention of foetuses in utero induced by inhibition of parturition. This was first accomplished experimentally by producing a fresh set of corpora lutea in the rabbit on the 28th day of pregnancy (Snyder, 1934), and later by others as the result of injecting progesterone. Following this procedure, the foetuses remain alive and continue to grow for 4 to 6 days beyond the normal Morphological Aspects of Ageing in the Placenta 113 time of parturition, but eventually die when they become too large to be delivered. It would be interesting to investi- gate both the cytology and the functional capacities of the placentas of such postmature rabbit foetuses. In connection with postmaturity, Masters (1952) cites three cases in women of intra-abdominal pregnancy which were terminated by Caesarean section at or near term, and the placenta allowed to remain in the abdominal cavity; for three weeks thereafter, active chorionic gonadotrophin and sodium pregnanediol excretion was present. REFERENCES Amoroso, E. C. (1952). In Marshall's Physiology of Reproduction, 3rd ed., vol. 2, p. 127. London: Longmans, Green and Co. Ltd. Bartholomew, R. A., and Kracke, R. R. (1932). Amer. J. Obstet. Gynec, 24, 797. Bartholomew, R. A., and Kracke, R. R. (1936). Amer. J. Obstet. Gynec., 31, 549. Bartholomew, R. A., and Parker, R. L. (1934). Amer. J. Obstet. Gynec, 27, 72. Bernard, C. (1859). J. de la physiol. de Vhomme et des animaux, 2, 326. Corey, E. L. (1932). Physiol. Zool, 5, 36. Crumpler, H. R., Dent, C. E., and Lindan, O. (1950). Biochem. J., 47, 223. Dempsey, E. W. (1953). Amer. J. Anat., 93, 331. Dempsey, E. W., and Wislocki, G. B. (1944). Endocrinology, 35, 409. Dempsey, E. W., and Wislocki, G. B. (1945). Amer. J. Anat., 76, 277. Dieckmann, W. J. (1952). The Toxemias of Pregnancy. St. Louis: C. V. Mosby Co. Flexner, L. B. (1955). Josiah Macy Jr. Foundation, Trans. First Conf. on Gestation, p. 11. New York. Flexner, L. B., Cowie, D. B., Hellman, L. M., Wilde, W. S., and Vosburgh, G. J. (1948). Amer. J. Obstet. Gynec, 55, 469. Flexner, L. B., and Gellhorn, A. (1942). Amer. J. Obstet. Gynec, 43, 965. Flexner, L. B., and Pohl, H. A. (1941a). Amer. J. Physiol., 134, 344. Flexner, L. B., and Pohl, H. A. (19416). J. cell. comp. Physiol., 18, 49. Hertig, A. T. (1946). J. Geront., 1, 96. Lansing, A. I. (1952). In Cowdry's Problems of Ageing, p. 3. Baltimore : Williams and Wilkins Co. Lell, W. A., Liber, K. E., and Snyder, F. F. (1932). Amer. J. Physiol., 100, 21. Masters, W. H. (1952). In Cowdry's Problems of Ageing, p. 651. Baltimore: Williams and Wilkins Co. 114 George B. Wislocki Mossman, H. W. (1937). Contrib. to Embryol. No. 158. Carnegie Inst, of Washington, 26, 133. Padykula, H. A. (1955). Anat. Rec, 121, 347. Padykula, H. A. (1956). Manuscript in preparation. Page, E. (1953). The Hypertensive Disorders of Pregnancy. Amer. Lecture Series. Springfield: C. C. Thomas. Rudolfo, A. (1934). J. exp. Zool, 68, 215. Schneider, C. L. (1950). Ciba Foundation Symposium on Toxaemias of Pregnancy. London: J. & A. Churchill, Ltd. Smith, G. V., and Smith, O. C. (1948). Physiol Rev., 28, 1. Snyder, F. F. (1934). Johns Hopk. Hosp. Bull, 54, 1 . Snyder, F. F. (1943). Proceed. Conf. Problems of Human Fertility, p. 144. Nat. Committee Maternal Health. New York. Somerville, I. F. (1950). Ciba Foundation Symposium on Toxaemias of Pregnancy. London: J. & A. Churchill, Ltd. Tenney, B. (1936). Amer. J. Obstet. Gynec, 31, 1024. Tenney, B., and Parker, F. (1940). Amer. J. Obstet. Gynec., 39, 1000. Thomsen, K. (1955). Arch. Gynaek., 185, 476. Villee, C. A. (1955). Josiah Macy Jr. Foundation, Trans. First Conf. on Gestation, p. 104. New York. Wimsatt, W. A. (1948). Amer. J. AnaL, 82, 393. Wimsatt, W. A. (1949). Amer. J. Anat., 84, 63. Wimsatt, W. A. (1950). Amer. J. Anat., 87, 391. Wimsatt, W. A. (1951). Amer. J. Anat., 89, 233. Wislocki, G. B. (1956). In Allen's Sex and Internal Secretion. 3rd ed. Baltimore: Williams and Wilkins Co. Wislocki, G. B., and Bennett, H. S. (1943). Amer. J. Anat., 73, 335. Wislocki, G. B., and Dempsey, E. W. (1946a). Amer. J. Anat., 78, 181. Wislocki, G. B., and Dempsey, E. W. (19466). Amer. J. Anat., 78, 1. Wislocki, G. B., and Dempsey, E. W. (1948). Amer. J. Anat., 83, 1. Wislocki, G. B., and Dempsey, E. W. (1955«). Anat. Rec., 123, 133. Wislocki, G. B., and Dempsey, E. W. (19556). Anat. Rec, 123, 33 Wislocki, G. B., Dempsey, E. W., and Fawcett, D. W. (1948). Obstet. Surv. Baltim., 3, 604. Wislocki, G. B., and Padykula, H. A. (1953). Amer. J. Anat., 92, 117. Wislocki, G. B., and Wimsatt, W. A. (1947). Amer. J. Anat., 81, 269. DISCUSSION Hamilton: I was very interested to hear what Prof. Wislocki had to say on the placenta. Anything he has to say concerning the placenta is always received with very great interest. As you know, I have been interested in the placenta for a great many years and within the last five or six years Prof. Boyd and I have been particularly concerned with the investigation of the human placenta in situ, concerning which we have chiefly considered two aspects. First, the changes that are occurring in the placenta itself and secondly, those in the endometrium, which like the placenta is a transient tissue. To Discussion 115 deal with the endometrium first — we have been investigating the histo- logical changes that are occurring in its blood vessels during pregnancy: we find that there are very extensive changes occurring in the so-called spiral arteries and particularly in the endothelium of those arteries. In the early stages, the endothelium, of course, is quite thin but as gestation proceeds it becomes thickened in many places and indeed many-layered. Why this occurs is not very easily answered but it is a very striking feature indeed. The changes in the placenta are also striking. The villi in the early stages of gestation are oedematous but as pregnancy advances the mesenchyme of these villi becomes fibrotic. I regard this as being in the nature of an ageing phenomenon as we progress from the early stages of the placenta to the later stages. The question of the cause of these quasi pathological changes in the placenta are very difficult to deter- mine. We have not found the fibrinous deposits on the villi that have been described by the pathologists. We also found that syncytium invades the mouths of endometrial vessels, both veins and arteries, that are opening into the intervillous spaces, or what we would rather regard as labyrinthine spaces than purely intervillous spaces. Amoroso: In your allusion to the endometrium, are you making a distinction between the maternal part of the placenta and other parts of the endometrium ? Hamilton: I am. Harrison: One of the points I would like to raise is the question of increase in surface area of the placenta as it ages or as it grows. Prof. Hamilton and I have been working on a series of Fallow Deer which have been obtained throughout pregnancy and we have specimens covering extensively the period of gestation in this animal. We find that the number of villi, if you count them, increases; not only that but their branches also increase very remarkably in number about the end of the first quarter of pregnancy ; and later there is a steady increase in the number of tertiary branches. If you remember, the villi in this animal are long and straight and bifurcate into equally straight branches. I am convinced that some of the so-called changes which have been called increases in efficiency of the placenta are, in fact, reflections of this increase in surface area, which at least I know occurs with one animal, and I would like to ask Prof. Wislocki if he has any observations to make on that in other mammals. Wislocki: That question was extensively discussed at the First Con- ference on Gestation of the Josiah Macy Jr. Foundation, 1955. It was stressed that unit weights do not form a valid basis for comparison of the relative rates of placental exchange in different mammals. The respective surface areas through which metabolic exchange occurs, involving either the effective surface of the placental membranes (trophoblast, vitelline membrane) or the surface area of the placental capillaries, would seem to be the most important factors. Dawes: In the whole animal perhaps what matters is the total amount of exchange. If you define "efficiency" in terms of surface area, you 116 Discussion are going to put the physiologists in a very tight spot (for technical reasons). What we wish to know, ultimately, is the ability of the foetus to survive under particular conditions and there we have to take into account the whole placenta. Is that a reasonable proposition? Wislocki: I agree with you that the whole placenta should be taken into account. Nevertheless, few physiologists are familiar with it in its entirety or have taken cognizance of it. The entire placenta, in various mammals, may be composed of several quite different placental structures including some type of yolk-sac placenta, a chorioallantoic placenta, and the membranous chorion. These may function suc- cessively or concurrently, as the case may be. Some functions and metabolic transfers may occur solely through one of these structures, while others take place through some other one. In the human, which has solely a chorioallantoic placenta, all physiological exchange must take place through it, with the possible exception of some transfer through the membranous chorion. Hamilton: Would Prof. Wislocki agree that both aspects must be considered ; the thickness of the placental barrier and the whole extent of the placental barrier? We have got to bear in mind, a thin layer of very great extent or a thin layer of limited extent and a thicker layer of very great extent. Wislocki: Well, it turns out that as more and more chemistry and histochemistry and enzymic chemistry is done upon the placental barrier that thickness alone, apparently, becomes less and less important because so many of these processes of transfer are influenced by chemical transfer rather than, as was thought previously but still entertained in some quarters, that it is all a consideration of a semi-permeable membrane. Jost: Prof. Wislocki yesterday raised the point of the relation between the foetus and the placenta, and I myself am very interested in such a question. To what extent does the foetal physiology have any importance for placental evolution? Prof. Wislocki, have you any information about those placentas which develop in the absence of the foetus? It is quite easy to obtain a placenta without a foetus either by removal of the foetus or, for instance, by castrating pregnant rats ; then, quite often, the foetus dies and the placenta continues to grow. In such castrated rats, the placentas seemed to me to be even larger than normal. I was very curious to know what happens in such a placenta. Wislocki: No, I have not studied any such placentas. It would be invaluable to investigate, cytologically and histologically, placentas which develop in the absence of the foetus. Placentas accompanying the experimentally induced condition of postmaturity in rabbits would also be particularly interesting to study with reference to senescence, the mechanisms of placental transfer and the synthesis of placental hormones by the trophoblast. Huggett: Some years ago, Dr. Pritchard and I investigated the effects of foetal death in the rat and we found that if the foetus was killed before the 11th day then almost invariably the placenta was absorbed. If, however, it was killed after the 11th day when the foetal mesoderm had Discussion 117 grown into it, then we found it could develop a certain degree of independence and new cell formation. Although it did not grow to full term it had, however, powers of intrinsic development without the foetus. Hamilton: Whether or not the placenta persists will depend, I think, on whether or not it has been vascularized, will it not? Amoroso: That is the point he means. Hamilton: It has not been vascularized in the early stages. Huggett: But if the mesodermal tissues, including vessels, grew within the placenta, then it had power of development. Hamilton: Quite right. Villee: In humans, as you know, the occurrence of a hydatidiform mole is a form of natural experiment of this type, for the foetus presumably dies and only the placenta is left. We have studied several examples of such moles and have shown that this tissue bears a strong resemblance to normal placental tissue in its functional capacities. The little vesicles which are present contain a fluid which is quite comparable to foetal serum. The amino acids are concentrated in the vesicular fluid just as they are in the foetal serum. The tissue itself has metabolic character- istics when incubated in vitro which are like those of normal placenta. CHRONOLOGICAL CHANGES IN PLACENTAL FUNCTION A. St. G. Huggett Physiology Department, St. Mary's Hospital Medical School, University of London Placental functions can be examined in two aspects, permeability effects and endocrinal functions. The first covers transmission of oxygen and nutrients from the mother and of carbon dioxide and waste products from the foetus. The second covers the well-recognized synthesis of reproductive endocrines which affect the mother and possibly the foetus. It also includes production of materials of metabolic import- ance which may be demonstrable in placental cells before disappearing and are to some extent the basis of Claude Bernard's dictum that the placenta is the liver of the foetus. Unfortunately, while many different aspects of placental function have been studied, there are only a limited number of investigations effected at different conceptional ages and it is only possible to mention a few in the time at one's disposal. Gellhorn and Flexner (1942) studied the passage of radio- active sodium across the placenta of the rabbit and other species (Fig. 1). These show increasing permeability until within 10 per cent of full term but a fall off in the last 10 per cent of intra-uterine life. Whether this last 10 per cent is the effect of postmaturity is a point of debate, as is also the standard of reference. Rodolfo (1934) found that the amount of antibodies passing during gestation increased, but the rate of increase was progressively diminished. There was no falling off. Barron (1951) has investigated the oxygen passage across the placenta in the sheep. There is evidence of different degrees of transfer at different placental ages and increasing oxygen tension in the umbilical vein blood in the last few days. My colleague, Dr. Wilfred Widdas, injected into pregnant 118 Chronological Changes in Placental Function 119 £ 90- Ul o 5 80 E % 70 ANIMAL MGS N4 TRANSFERRED TOTAL GESTATION " PER GM PLACENTA PER HR. PERIOD IN DAYS AT MAXIMUM GUINEA PIG 6.26 67 1 0.4- 23 9.25 32 0.79 62 0.43 154 0.028 114 RAT RABPiT CAT GOAT SOW 04 05 06 07 PERIOD OF GESTATION IN TENTHS OF TOTAL Fig. 1. Transfer of sodium across unit weight of placenta in unit time in different species. (After Gellhorn and Flexner, 1942.) rats an amount of 59Fe adequate to produce a constant specific activity in the mother which may possibly also have been the maximum saturation. There was a rise in the total iron in the litters and an increasing percentage of labelled Fe was recovered from the litters (Fig. 2). But when the iron/foetal 26 r 14 Injection Fe59 Days Fig. 2. Percentage of labelled iron recovered from litters at varying ages and the number of rats in the litter. Continuous lines indicate upper and lower limits, and broken line the approximate mean recovery curve. Weaning at 21-28 days after birth. (Huggett and Widdas, 1949.) 120 A. St. G. Huggett weight was calculated it was seen that a relative anaemia was produced (Fig. 3). This, however, was not due either to shortage of iron or inefficiency of the placenta but to the foetal growth being disproportionate. At the risk of overlapping with Professor Wislocki's approach, one would draw attention to the weights of the human placenta given in Adair and Thelander's classical paper in 1925. In no case is there any suggestion that the placenta loses weight as the foetus grows (Fig. 4). In the goat, however, Elliott, Hall and Huggett (1934) found evidence of 100 90 80 70 *o 60 3? 50 I" 40 30 20 10 0 Pregnancy Lactation \. 14 14 Days 21 28 35 42 Fig. 3. Ratio Iron/Weight in foetal and newborn rats. (Huggett and Widdas, 1949.) a slowing up or flattening out of the placental weight curve. On the other hand, umbilical blood flow through the placenta in the foetal sheep has been shown to rise rapidly as term approaches (Fig. 5) (Cooper, Greenfield and Huggett, 1949). When considering the production of materials by the placenta one should draw attention to the production of carbohydrates within it. Mammalian allantoic placentas can be divided into two classes, those whose foetal blood has one sugar, glucose, and those whose foetal blood has two, glucose and fructose, and in these, fructose is always in excess. These latter are all Ungulata or Cetacea. Glucose can flow back to the mother if the gradient is reversed experimentally but fructose never does. This fructose is synthesized by the Chronological Changes in Placental Function 121 PLACENTAL WEIGHT*(*VERAGEf j 300 900 I SOO 2250 3000 3750 -4000 5250 4 FOETAL WEIGHTS IN GRAMS Fig. 4. Relation between placental and foetal weights in the human. (After Adair and Thelander, 1925.) placenta, a function not possessed by the foetus. This synthesis appears to be a steadily increasing function, as shown by Table I. All the other mammals have no foetal fructose and always appear to have much glycogen at some stage in the 20 UO 160 40 60 60 IOO I20 SHEEP FOETAL AGE DAYS Fig. 5. Umbilical blood flow in the foetal sheep at different ages. (After Cooper, Greenfield, and Huggett, 1949.) 122 A. St. G. Huggett Table I Total Fructose Secretion by the Placenta Foetcd age (days) Blood fructose (mg-%) Blood volume (ml.) Fructose in foetal blood (mg-) Total fructose space (ml.) Total fructose in fructose space (mg-) Corrected by Hitchcock (1949) {mg-) 60 80 100 120 140 150 125 100 75 50 15 50 110 250 500 22-5 62-5 110 187 250 130 200 300 430 1800 195 250 300 325 900 225 312 375 406 1125 placenta. All placentas have deposits not only of glycogen but several other materials which all have dates of maximal deposition differing with the material and the species (Table II). It would appear that as the placenta ages it takes on Table II Endoplacental Composition in the Rodent Dates of Maximum Concentrations (Full-term: Rabbits, 30 days; Rats, 21 days) Day of Substance maximum concen- tration mg-% Reference Rabbit Glycogen (foetal) 14 1-54 g./lOOml. Lochhead and Cramer (1908) Glycogen (maternal) 21 5-57 g.,/100 ml. Lochhead and Cramer (1908) Rat Glycogen (whole placenta) 16 830 Huggett (unpublished) Rabbit Phospholipin 30 1500 Boyd (1935) Free cholestrol 30 2T5 55 55 Cholestrol ester 21 400 55 55 Neutral fat 17 970 J J 55 Chronological Changes in Placental Function 123 different functions, not only as a whole but in different parts of the placenta and in different species. The peak of placental glycogen in the rat is at the 16th day, after which it decreases. Some years ago my former colleague, Professor J. J. Pritchard, and I investigated the effect on the placenta of experimental foetal death in the rat (Huggett and Pritchard, 1945; Pritchard and Huggett, 1947). Oestrogens caused decidual necrosis and foetal death but DOSE :- 0«4lmg./. day for 7 days 2 • 8 mg. spread over 7 days PROGESTERONE * IMPLANTED "~ NORMAL CONTROLS IO II 12 i3 14 IS 16 17 18 19 20 21 22 RAT FOETAL AGE DAYS Fig. 6. Daily averages of placental glycogen in the last third of pregnancy in normal pregnant rats and rats im- planted with progesterone during the previous 7 days. Average absorption 0 • 28 per day. progesterone had no such effect; rather the foetus thrived. An extension of these experiments has demonstrated that the implantation of progesterone during a period of 7 days in the pregnant rat with an average absorption of 0 • 41 mg. per day per rat produces an accelerated decrease of the glycogen in the period of placental glycogen decrease (Fig. 6). It also causes a decrease of placental weight accompanied by an increase of foetal weight (Figs. 7 and 8). In summary, therefore, the placenta changes composition not in a regular sequence to old age but spasmodically, suggesting that different functions are exercised at different 124 A. St. G. Huggett E to !- I y Z UJ U < s! ,0°- NORMALS PROGESTERONE —IMPLANTED IC 12 !4 16 18 20 22 RAT FOETAL AGE *. DAYS Fig. 7. Daily averages of placental weights in these same experiments. // NORMAL •'/ CONTROLS 12 14 FOETAL AGE DAYS Fig. 8. Daily averages [of foetal weights in these experiments. Chronological Changes in Placental Function 125 ages, and there is evidence that its "efficiency", certainly for some functions when considered in relation to foetal growth, is increased in its old age; in fact, it might be said to be abruptly expelled in the prime of life. REFERENCES Adair, F., and Thelander, H. (1925). Amer. J. Obstet. Gynec, 10, 172. Barron, D. H. (1951). Yale J. Biol. Med., 24, 169. Boyd, E. M. (1935). Biochem. J., 29, 985. Cooper, K. E., Greenfield, A. D. M., and Huggett, A. St. G. (1949). J. Phijsiol, 108, 160. Elliott, R. H., Hall, F. G., and Huggett, A. St. G. (1934). J. Physiol., 82, 160. Gellhorn, A., and Flexner, L. B. (1942). Amer. J. Physiol., 136, 750. Hitchcock, M. W. S. (1949). J. Physiol, 108, 117. Huggett, A. St. G., and Pritchard, J. J. (1945). Proc. R. Soc. Med., 38, 261. Huggett, A. St. G., and Widdas, W. F. (1949). J. Physiol, 110, 386. Lochhead, J., and Cramer, W. (1908). Proc. roy. Soc. B, 80, 263. Pritchard, J. J., and Huggett, A. St. G. (1947). J. Anal, Lond., 81, 212. Rodolfo, A. (1934). J. exp. ZooL, 68, 215. DISCUSSION Amoroso: It is evident that Professor Huggett does not feel inclined to accept Lansing's definition of the ageing process. As I understand it from what Professor Wislocki has said, Lansing considered the absence of the powers of reconstruction in an ageing tissue to be an important criterion of sensescence. Dawes: I would like to say how much I enjoyed Prof. Huggett's exposition of the physiology of the placenta. I think that in any dis- cussion of measurement of its efficiency, its functional efficiency should be the overriding consideration. There is one rather odd finding that I should like to comment on. Mr. Parry has been measuring the weights of the placentomes in sheep and finds that, after the 120th day, the total weight of the placentomes of a particular variety of sheep falls very considerably from about 600 g. /foetus to 200-300 g. at term. Whether this is a species difference, I do not know; it is just one more of the many facts that are necessary to assess efficiency. Another point ; I was much impressed by the work of Cooper and associates (Cooper, K. E., Greenfield, A. D. M. and Huggett, A. St. G. (1949) J. Physiol, 108, 160) in measuring placental flow, which is tech- nically difficult. They showed that flow through the foetal side of the 126 Discussion placenta rose progressively during pregnancy. Now, blood pressure also rises during the latter half of pregnancy, and it would seem that this rise in blood pressure is in fact due to the development within the foetus of vascular reflexes and the sympathetic nervous system. I wonder, therefore, whether we should not concentrate attention, in future work, on trying to find out what happens to the vascular re- sistance on the foetal side of the placenta. So far as the little evidence available suggests, we may find that the vascular resistance is the same throughout the latter part of pregnancy, and the increase in umbilical blood flow may be due to an increased cardiac output and to changes in the foetus rather than the placenta. In any discussion of the transfer of material across the placenta, there are obviously three variables we have to take into account, first, the foetal placental flow, secondly the diffusion coefficient of the pla- cental tissues, and thirdly uterine blood flow. Now, as to uterine blood flow there is practically no data available. Barcroft made some ob- servations in rabbits but the numbers of rabbits were rather small. I know that there is great technical difficulty in measuring maternal blood flow through the placenta but it looks, from Barcroft's other work, as if it might be the limiting factor in asphyxia of the foetus towards the end of foetal life. This seems one of the most important variables which still have to be measured in determining what has been called the lack of efficiency of the placenta. I am not sure that this lack of efficiency has been firmly established, but in order to investigate it, uterine flow must be measured. Amoroso : Were you comparing the sheep with the rabbit in respect of maternal flow? Dawes: No, the only published work is on the rabbit. Amoroso: I raised that point because we might get into difficulties if we try to relate results derived from the rabbit to sheep and vice versa. I think that point was raised, or at least was hinted at by Professor Wislocki when he referred to the maturation of the foetus. The sheep foetus is, as we all know, a very much more mature organism at term than is the rabbit. Zuckerman: Prof. Huggett's last remarks about progesterone remind me that McKeown and I once did some work, which was concerned with the effect of progesterone on foetal growth, in which the corpora lutea were destroyed one by one with an electrocautery. We found that as the amount of luteal tissue in the ovaries decreased, so the litter size declined. We, therefore, assumed that progesterone had some direct influence on the foetus. It now seems that Prof. Huggett has demon- strated that the influence of progesterone is on the placenta, and the foetus is affected secondarily. Huggett: It is a possible inference. Jost: I am not sure that during pregnancy progesterone is necessary for the foetus itself. Two kinds of observations may be mentioned. First the experiment of Courrier who succeeded in producing ab- dominal pregnancies in castrated rabbits by opening the uterus at the level of a nidation, and castrating the female on the same day. The Discussion 127 intra-uterine foetuses died quickly after castration, the abdominal ones survived and grew. Progesterone is necessary only for the intra- uterine development. Castration in the pregnant rat does not always stop development of all foetuses. Some may survive, but they are often deformed by uterine pressure. In a series of personal experiments, the weight of such foetuses found alive on day 21 was about 4-4 g., with controls of the same strain of 4 • 9 g. Despite the uterine impairment, foetal growth was not very much reduced by the absence of the corpora lutea. Amoroso: Of course, you make the assumption that there is no extra gonadal source of progesterone there. Jost: Yes, but in the rabbit and in rats there is not too much pro- gesterone present; at least it has never been detected. Amoroso: There is enough, nevertheless, to maintain abdominal preg- nancies in ovariectomized rabbits. Moreover, I believe that both Professor Courrier and Professor Klein would dispute your point. Huggett: May I just comment on these two points; first, I entirely agree, I do not think progesterone is necessary for the growth of the foetus ; all that I want to suggest is that it appears to alter the permea- bility and, therefore, nutrients seem to get into the foetus and I take it — I have no knowledge — but I guess that the foetus thus gains a better supply and it can grow a little better. That is as far as one can go. T.-Duplessis: Concerning the relationship between the placental weight and the foetal weight, Prof. Huggett quoted an observation of Adair and Thelander. We have also found that when pregnancy is pro- longed in the rat for 3 or 4 days, both the placental and the foetal weights increase. So this raises the question whether the placenta determines the length of pregnancy ? We think that the placenta is able to function for a longer period of time than it normally does; in fact, when we looked at the placenta in pregnancies prolonged for 3 or 4 days they looked quite normal. It seems that ageing is not a definite process of the placenta : the placenta can reach a greater age than it normally does without showing ageing phenomena. Huggett: Which function were you studying when you said the function ? T.-Duplessis: The foetal-maternal exchange. Since these foetuses are still alive 3 or 4 days after term and are bigger than normal we may assume that the placenta does function correctly. Amoroso : A considerable amount of work has been carried out on the prolongation of pregnancy in these animals by Hill and Parkes, and by Professor Wislocki and his associates which antedates the work of Sir Joseph Barcroft. Huggett: Yes, and they showed that if the foetus does go on too long, then it asphyxiates itself. A?noroso: For my guidance, might I ask Professor Huggett what exactly did he weigh when he worked on the goat's placenta? Huggett: The cotyledons, not the membranes. Amoroso: You stripped the membranes from the caruncles? Huggett: No, we took the caruncles, cut them off leaving the rest of 128 Discussion the chorioallantoic membrane, so what was weighed were the caruncles and placentomes and not the placenta. There are two points; first, Dr. Fahmy and I gave a paper at the Federation in Atlantic City last year (1954) in which we showed that alkaline phosphatase and glycogen never appear in the same cells, but when glycogen disappears alkaline phosphatase comes into the picture. So that we might have a curve showing an uprise of alkaline phos- phatase activity — we have not estimated that. Secondly, as the placenta gets thinner, then it is more efficient; diffusion goes on. As far as one can see from the work being done, the physio-chemical diffusion as such plays practically no part at all. It is only between thinning and diffusion so that it is almost a facilitated, a temporal cytological action involving energy consumption. Amoroso : If we agree that the placenta is a transient tissue, then the structures, variously described as endometrial cups or craters and which occur transiently in the placenta of the equidae, must be regarded as transient tissues within a transient tissue. They have been fully described and are figured in Brit. Med. Bull. (1955) vol. 11. BIOCHEMICAL EVIDENCE OF AGEING IN THE PLACENTA Claude A. Villee Department of Biological Chemistry, Harvard Medical School, and Research Laboratories, Boston Lying-in Hospital, Boston, Massachusetts The concept that the placenta undergoes regressive changes in the latter months of gestation and that pregnancy is termi- nated when the placenta can no longer provide an adequate supply of nutrients to the child was originally stated by Hippocrates (Reynolds, 1949). Careful histological and histo- chemical observations (Wislocki and Bennett, 1943; Wislocki, Dempsey and Fawcett, 1948) have provided evidence of thickening and hyalinization of the walls of the blood vessels and of the presence of thromboses in the last two or three months of gestation. Decreases in the rates of oxygen con- sumption and of anaerobic glycolysis as gestation proceeds have been reported (Loeser, 1932; Wang and Hellman, 1941; Page, 1948; Hellman, Harris and Andrews, 1950). The present paper describes the results of investigations of the intermediary metabolism of slices from placentas ranging in age from six weeks to term. The metabolic activities of the placenta decrease as gestation proceeds, but the decreases are not general and uniform. The decreases in activity are gradual, without sharp change near term, and the placenta at term is still quite active in many respects. Term placentas were obtained directly from the delivery room and were used within five minutes of delivery. Earlier placentas were obtained at Caesarean sections performed for delivery or for therapeutic interruption of pregnancy and were used within five minutes of their removal from the patient's body. Placental slices, about 0-5 mm. thick and weighing about 200 mg., were cut with a Stadie-Riggs micro- tome or with fine scissors. Care was taken in preparing the slices to avoid areas of necrosis. The slices were incubated AGEING VOL. 2 129 6 130 Claude A. Villee in Warburg respirometer vessels at 37° in an oxygen atmo- sphere. The composition of the incubation medium, expressed in millimoles per litre, was K+ 50, Na+ 80, Mg+ + 10, phosphate 40, Cl~ 100, pyruvate 10 and glucose 11-1. In alternate vessels, either the glucose or the pyruvate was labelled with 14C. The initial pH of the medium was 6 • 8 and the pH after incubation was 6*8 ±0-1 (determined by glass electrode). Other slices of the placentas were used for determination of initial glycogen content and of wet weight : dry weight ratios. After a two hour incubation, the slices were removed, weighed, and analysed for glycogen. Aliquots of the incuba- tion medium were analysed for glucose (Nelson, 1944), pyruvate (Friedemann and Haugen, 1943), and lactate (Barker and Summerson, 1941). The respiratory C02 was recovered from the alkali in the centre well and converted to BaC03, then its 14C content was measured with a windowless, proportional flow counter (Robinson, 1950). Glucose was isolated from other aliquots of the incubation medium, after the addition of 10 mg. of carrier glucose, as the glucose phenylosazone. This was washed, recrystallized, washed again, and its radioactivity was measured with the flow counter. Pyruvate was isolated from the other aliquots, after the addition of 10 mg. of carrier pyruvate, as the 2 : 4-dinitro- phenylhyrazone. This was washed seven times with alcohol and water, then its radioactivity was measured with the proportional flow counter. Aliquots of the glucose obtained from the tissue glycogen were converted to glucosazone for radioactivity determinations. From these measurements one can estimate the net amount of glucose utilized or produced, the net amount of pyruvate utilized, the net amount of lactate produced, the amounts of glucose and pyruvate carbons metabolized to C02, the net amount of glycogen produced or utilized, the amounts of glucose and glycogen made from pyruvate, and the amounts of glycogen and pyruvate made from glucose. Preliminary experiments showed that there was no detect- able difference in the metabolism of slices from the chorionic Biochemical Evidence of Ageing in the Placenta 131 and decidual surfaces of the placenta and that slices cleaned of blood by exhaustive washing with cold saline and those cleaned by blotting on filter paper showed similar rates of metabolism. Slices prepared from placentas which had been standing 60 minutes at room temperature showed metabolic rates 20 per cent less than those cut immediately. Oxygen Consumption A gradual decrease in the rate of oxygen consumption of the placenta as gestation proceeds has been reported previously (loc. cit.). The present series includes thirty term placentas 2 - .!••■ 10 15 20 25 50 55 40 WEEKS OF PREGNANCY Fig. 1. The oxygen consumption, in fx\. per mg. of dry weight per hour, of placental slices as a function of the age of the placenta. The arrow at 40 weeks indicates the range of the values obtained for term placenta. 132 Claude A. Villee and thirty-three earlier placentas, ranging in age from 6 to 37 weeks. The decrease in oxygen consumption with age is linear until a gestational age of about 30 weeks, after which the Q0j (microlitres of oxygen utilized per mg. of dry weight of tissue per hour) is relatively constant until term (Fig. 1). The values for oxygen consumption reported here are slightly lower than those reported by Hellman, Harris and Andrews (1950), probably because of the differences in the method of determin- ing the dry weight of the tissue. In their experiments the tissue was removed from the medium at the end of the incuba- tion period, dried and weighed. We have found that the wet weight of the tissue recovered at the end of the experiment is only 60 to 70 per cent of that put in originally. To avoid this source of error, the dry weight : wet weight ratio for each placenta was determined from adjacent slices. The placental slices recovered from the incubation medium at the end of the experiment were weighed and their glycogen content was measured. The ratio of dry to wet weight of the placenta almost doubles during gestation (Fig. 2). 15 12.5 10 * 7.5 5- • • \0 20 25 30 35 40 WEEKS OF PREGNANCY Fig. 2. The ratio of dry to wet weight of the placenta as a function of placental age. The arrow at 40 weeks indicates the range of the values obtained for term placenta. Biochemical Evidence of Ageing in the Placenta 133 Glycogen Content and Metabolism The glycogen content of the placenta was estimated by the method of Walaas and Walaas (1950). The glycogen was isolated and purified, hydrolysed to glucose, and the glucose 50 40 50 20 \. • \ \ 10 IS 20 25 50 WEEKS OF PREGNANCY 55 40 Fig. 3. The glycogen content, expressed as mg. of glycogen per g. of dry weight of tissue, of the pla- centa as a function of age. The arrow at 40 weeks indicates the range of the values for term placenta. was measured by the method of Nelson (1914). There is a marked decrease in the glycogen content of the placenta with age from 6 until 25 or so weeks of age (Fig. 3). The glycogen content remains essentially constant during the last 12 to 15 weeks of pregnancy. From the measurements of glycogen content before and 134 Claude A. Villee after incubation, the net change in glycogen content was cal- culated and expressed as micromoles of glucose units per g. of wet tissue per hour. The placenta early in gestation has a marked ability to synthesize glycogen in vitro (Fig. 4). This ability begins to decrease after 10 to 12 weeks of pregnancy and decreases steadily to term. The average rate of glycogen utilization of the thirty term placentas was —1-92 micro- moles per g. per hour, and only one of the thirty showed a + 4 • • • \ • . .>.^ • * • t ~-^° • • • • • • • • "V- — — ______ ■ — < i • • > t X +2 __D o o E 2 "2 ui O O u i -4 o < 5 10 15 20 25 30 35 40 WEEKS OF PREGNANCY Fig. 4. The metabolism of glycogen by the placenta as a function of age. Positive values indicate net glycogen synthesis; negative values indicate net glycogen utiliza- tion. The values are expressed as /nmoles of glucose units per g. of wet tissue per hour. The arrow at 40 weeks indicates the range of values for term placenta. net production of glycogen (-f-0-22 micromoles per g. per hour). Corroborative evidence was obtained from the radio- activity measurements of the glucosazones derived from the tissue glycogen after incubation. These showed that term placentas have a just barely detectable ability to incorporate [14C]glucose into glycogen in vitro. In contrast, 8 to 14 week placentas have a marked ability to incorporate 14C from both glucose and pyruvate into glycogen in vitro. This ability decreases with gestation and was absent from two of the three placentas aged 22 to 24 weeks. Biochemical Evidence of Ageing in the Placenta 135 The Metabolism of Glucose The ability of the placenta to utilize or produce glucose was calculated from the difference in glucose content of the incuba- tion medium before and after incubation. Some of the earliest placentas tested showed a net production of glucose during the incubation (Fig. 5). From 15 weeks until term, however, + 10 T 5 - 5 -15 % —- 10 15 20 25 30 WEEKS OF PREGNANCY 35 40 Fig. 5. The metabolism of glucose by the placenta as a function of age. Positive values indicate net glucose production; negative values indicate net glucose utiliza- tion. The values are expressed as ^tmoles per g. of wet tissue per hour. The arrow at 40 weeks indicates the range of values for term placenta. the average net rate of glucose utilization proved to be about —7 micromoles of glucose per g. wet tissue per hour. From the difference in the specific activity (c.p.m. per millimole glucose) of the glucose in the incubation medium before and after incubation, one can estimate the amount of glucose produced by the placenta and secreted into the medium and the true amount of glucose utilized. These calculations showed that the placenta at 10 weeks has a glucose production of about -{-12 micromoles per g. per hour and a true glucose utilization of about —17 micromoles per g. per hour. The net 136 Claude A. Villee glucose utilization is, therefore, — 5 micromoles per g. per hour. At term, the true glucose utilization is —7-5 micro- moles per g. per hour, glucose production is zero, and the net glucose utilization is also —7-5 micromoles per g. per hour. Hence, although the average net utilization of glucose is fairly constant from 15 weeks to term, this is in fact due to concomitant decreases in both the production and utilization of glucose. The production of glucose by the liver has been shown to be mediated by the enzyme glucose-6-phosphatase (de Duve, Berthet, Hers and Dupret, 1949; Swanson, 1950), a reaction in which glucose-6-phosphate is converted to glucose plus inorganic phosphate. The glucose-6-phosphate is derived from glycogen via phosphorylase and glucose-1 -phosphate, or by gluconeogenesis from other carbon compounds by a reversal of glycolysis. The human foetal liver gains the ability to secrete glucose only after 12 to 16 weeks of development (Villee, 1953a). Before that time it is unable to regulate the concentration of glucose in the foetal blood stream. The placenta does have the ability to secrete glucose early in pregnancy but loses this ability as gestation proceeds. The placenta thus could regulate the concentration of glucose in foetal blood early in development but not later. The ability of a tissue incubated in vitro to secrete glucose into the incubation medium may be tested in three ways. First, if glucose production exceeds glucose utilization, direct chemical analysis of the medium will reveal the net increase in the amount of glucose present. Second, whether or not there is a net increase in glucose concentration, glucose production can be detected by using 14C -labelled glucose in the incubation medium and measuring the specific activities of glucose isolated from the medium before and after incuba- tion. If the tissue produces glucose, the unlabelled glucose molecules produced will dilute the labelled molecules of glucose present in the medium and the specific activity will therefore decrease as incubation proceeds. Third, the tissue may be incubated in the presence of 14C -labelled pyruvate, or in the Biochemical Evidence of Ageing in the Placenta 137 presence of any substance which is readily metabolized to glueose-6-phosphate. Glucose is isolated from the medium at the end of the incubation period, and its radioactivity is measured. In this way the production of glucose can be demonstrated directly and its rate of formation from the given precursor can be estimated. By all three methods it was shown that the placenta can secrete glucose early in pregnancy. However, it gradually loses this ability, presumably by a loss of the enzyme glucose-6-phosphatase (Table I). Table I The Metabolism of Glucose by Placental Slices Age (weeks) No. of ex p. [imoles jg. tissu e /hour Glucose utilized Glucose produced Glucose produced from pyruvate 8-10 11-12 13-15 19-23 28 32 36 40 4 6 5 5 2 1 2 12 -17-7+0-53* -19-3 + 11 -18-6 + 3-6 -22-3 + 1-9 -15-9 -12-8 -11-2 - 6-6+0-68 + 11-9 + 1-5 + 9-5 + 1-5 + 9-9+2-8 + 14-3 -1-2-1 + 11-7 + 90 0 0 +0-27+0-01 +019+002 +0-16+0-05 0 0 0 0 0 * mean ± standard error of the mean. These experiments demonstrate that the placenta early in gestation has the biochemical mechanism necessary for the storage of glycogen and the secretion of glucose. It could function to regulate the glucose content of the foetal blood stream. Claude Bernard demonstrated the presence of gly- cogen in the sheep placenta and in 1858 made the suggestion that the placenta may act as an "accessory liver". The Metabolism of Pyruvic and Lactic Acids The net utilization of pyruvate, calculated simply from the difference in the content of pyruvate in the medium before and after incubation, by placentas at successive stages of 138 Claude A. Villee gestation is given in Fig. 6. There is a considerable amount of variation between placentas of the same age. A slight de- crease in the net utilization is evident after 10 weeks of development. The amount of pyruvate produced during the incubation period can be calculated from the dilution of the radio-pyruvate in the medium by unlabelled pyruvate produced by the cells. The true utilization, i.e., the net 0 • • • • -10- 0*0 • •_ — - -"•" ' • • • • • • -20 • 10 25 20 25 30 35 40 WEEKS OF PREGNANCY Fig. 6. The utilization of pyruvate by the placenta as a function of age. The values are expressed as /xmoles per g. of wet tissue per hour. The arrow at 40 weeks indicates the range of values for term placenta. utilization corrected for the amount produced, can then be estimated. Both production and true utilization of pyruvate increase from 7 weeks to 24 weeks and then decrease at term (Table II). The net utilization based on these calculations (column 5 of Table II) is maximum at 10 to 12 weeks and then gradually decreases to term. The rate at which pyruvate is produced from glucose remains essentially constant from 7 weeks to term (column 6 of Table II). The fact that labelled carbon from pyruvate can be incorporated into glycogen indicates that the glycolytic cycle can operate in reverse, Biochemical Evidence of Ageing in the Placenta 139 Table II Metabolism of Pyruvate by Placental Slices Age (weeks) No. of exp. [Lmolesjg. tissue (hour Pyruvate produced Pyruvate utilized Net pyruvate utilization Pyruvate produced from glucose 7-9 10-12 13-15 19-24 40 8 15 8 7 22 -fll-8±ll* + 13-0±0-85 + 17-2±l-4 + 190±2-9 + 14-0±0-55 -19-6±l-7 -25-5±l-9 -26-6±2-8 -27-7±2-7 -22-3±ll - 7-8 -12-5 - 9-4 - 8-7 - 8-3 + 5-7±0-53 + 6-9±l-3 + 6-9±l-l + 5-5±0-59 + 50±0-24 * mean ± standard error of the mean. presumably via glucose-6-phosphate. This is evidence that the inability of the term placenta to secrete glucose is due to the absence of the enzyme glucose-6-phosphatase and not to the failure of some part of the glycolytic cycle. The production of lactic acid by placental slices was cal- culated from its rate of appearance in the incubation medium. As with pyruvate utilization, there was considerable variation in the values obtained early in gestation, but there was a gradual decrease in lactate production from 10 weeks to term (Fig- 7). The source of the lactic acid was determined by experi- ments in which slices of term placenta were incubated in media containing different substrates (Table III). When no substrate was present in the incubation medium, lactate was produced at a rate of about 5 (jtmoles/g./hour. This lactate is derived primarily from the breakdown of glycogen, for glycogen disappearance occurred at the rate of 2 to 3 jxmoles/ g./hour. Each micromole of glucose unit in glycogen yields two micromoles of lactate. The presence of 10 jxmoles pyruvate per ml. of incubation medium resulted in the production of an additional 2 to 3 (xmoles of lactate per g. per hour. Glucose in the medium at a level of 1 1 • 1 (jimoles per ml. increased the production of lactate 6 to 7 (jimoles per g. per hour over that 140 Claude A. Villee 30 20 10 .-v. t 10 20 25 30 35 40 WEEKS OF PREGNANCY Fig. 7. The production of lactate by the placenta as a function of age. The values are expressed as /imoles per g. of Avet tissue per hour. The arrow at 40 weeks indicates the range of values for term placenta. made when no substrate was added. Since glucose is meta- bolized to lactate via pyruvate, the greater production of lactate from a glucose substrate than from pyruvate itself can Table III The Production of Lactate by Slices of Term Placentas Gas phase Substrate [imoles of lactate produced per g. tissue per hour None Glucose Pyruvate Acetate Oxalacetate o2 4-94 11-9 613 o2 4-76 10-8 0-84 o2 511 4-78 3-21 o„ 706 6-38 5 03 Na 110 23-3 12-4 N2 905 211 9-58 o2 4-59 13 1 811 o2 4-40 9-7 7-9 Biochemical Evidence of Ageing in the Placenta 141 be ascribed to the greater abundance of reduced diphospho- pyridine nucleotide produced in glycolysis. Under anaerobic conditions lactate production in the absence of added sub- strate was about double that found aerobically. Lactate production from glucose was similarly doubled. The incre- ment in lactate production on the addition of pyruvate under anaerobic conditions was small, again suggesting that the limiting factor was the amount of reduced diphosphopyridine nucleotide present. The presence of acetate or oxalacetate, at levels of 10 [xmoles per ml., decreased the accumulation of lactate. Previous experiments with 14C -label led glycerol had shown that it can be utilized by liver (Tang et al, 1953) but not by muscle (Villee, White and Hastings, 1952). When slices of term placenta were incubated with 14C-labelled glycerol, it was found that the placenta is unable to utilize glycerol in glycolysis (Villee, 1953ft). The evidence favoured the con- clusion that term placenta lacks the enzyme which phos- phorylates glycerol to form a-glycerophosphate, the first step in its metabolism. Effects of Hormones In another series of experiments, tests were made of the ability of slices of term placenta to respond to hormones added in vitro. The presence of insulin (0-5 unit per ml.) increased the utilization of glucose and the synthesis of glycogen ; it had no effect on the utilization of oxygen or the production of lactic acid (Villee, 1953ft). Insulin also increased the rate of the utilization of glucose by slices of decidua from a six- week pregnancy and by slices of hydatid mole (pure chorionic villi). The addition of insulin to slices of early placentas increased the rate of glucose utilization to such an extent that the net production observed in the control was changed to a net utilization of glucose in the vessels to which insulin was added. The addition of cortisone or of an aqueous adrenal extract decreased the utilization of glucose and oxygen. 142 Claude A. Villee Term placenta also retains its ability to respond to oestro- gens added in vitro (Villee and Hagerman, 1953). The effect of oestradiol-17/? or oestrone has been shown to be a stimula- tion of isocitric dehydrogenase, an enzyme which requires diphosphopyridine nucleotide (DPN) as coenzyme (Villee, 1955; Villee and Gordon, 1955; Gordon and Villee, 1955). The Metabolic "Ageing" of the Placenta From the present experiments it may be concluded that in all of the metabolic functions tested, the activity of the placenta decreases as gestation proceeds. If we may equate decreasing metabolic activity with ageing, then the placenta undergoes ageing. The changes observed include : an increase in the solid fraction of the tissue, a decrease in oxygen con- sumption to about half the rate earlier, a marked decrease in the glycogen content of the tissue, a loss of the ability of the placenta to synthesize glycogen in vitro, a decrease in the rate of glucose utilization, a loss of the ability to produce glucose, and less marked decreases in the rates of pyruvate utilization and lactate production. A calculation of the fraction of the respiratory C02 derived from the labelled substrate, obtained by comparing the specific activities of the centre well C02 and the substrate, showed that there was no change in this. The percentage of C02 derived from the glucose of the medium was 3 • 7 ±0 • 6 in the early (10 week) placentas and 4-0 ±0-76 at term. The percentage of C02 derived from the pyruvate of the medium was 20 -8 ±1*34 in the early placentas and 18 -1^1*03 at term. Thus, although the absolute amounts of glucose and pyruvate metabolized by term placentas were decreased, the relative proportion of the C02 derived from each of these substrates was unchanged. Eight experiments have been performed with placentas from women with toxaemia. The ages of the placentas ranged from 32 weeks to term. In all respects tested, the utilization of glucose, oxygen, glycogen, and pyruvate the Biochemical Evidence of Ageing in the Placenta 143 production of lactate and C02, and in initial glycogen and water content, these placentas did not differ significantly from normal placentas of a comparable age. These results confirm and extend those of Heilman, Harris and Andrews (1950). The data presented make it clear that the decreases in metabolic rate are not uniform. Some functions decrease more rapidly, and begin at an earlier time, than others. Glycogen synthesis appears to be one of the first to decrease markedly. The term placenta, as judged by its metabolic activities in vitro, is still quite active. It would seem inaccurate to speak of it as "senile". There is, furthermore, no sharp change at term which could be used to explain the initiation of labour. One could postulate, perhaps, that labour is induced when some critical ratio between the demands of the child for nutrients and the ability of the placenta to supply them is exceeded. The change in the ratio would be brought about, however, more by the rapid increase in the demands of the child as its mass increases than by the decrease in the ability of the placenta to supply nutrients. REFERENCES Barker, S. B., and Summerson, W. H. (1941). J. biol. Chem., 138, 535. De Duve, C, Berthet, J., Hers, H. G., and Dupret, L. (1949). Bull. Soc. Chim. biol., Pans, 31, 1242. Friedemann, T. E., and Haugen, G. E. (1943). J. biol. Chem., 147, 415. Gordon, E. E., and Villee, C. A. (1955). J. biol. Chem., 216, 215. Hellman, L. M., Harris, B. A., and Andrews, M. C. (1950). Johns Hopk. Hosp. Bull, 87, 203. Loeser, H. (1932). Arch. Gynaek., 148, 118. Nelson, N. (1944). J. biol. Chem., 153, 375. Page, E. W. (1948). Obstet. Surv. Baltim., 3, 615. Reynolds, S. M. R. (1949). Physiology of the Uterus, 2nd ed., New York: Hoeber. Robinson, C. V. (1950). U.S. Atomic Energy Commission, Document AECU-719. Swanson, M. A. (1950). J. biol. Chem., 184, 647. Tang, C.-T., Karnovsky, M. L., Landau, B. R., Hastings, A. B., and Nesbett, F. B. (1953). J. biol. Chem., 202, 705. Villee, C. A. (1953a). J. appl. Physiol, 5, 437. Villee, C. A. (19536). J. biol. Chem., 205, 113. 144 Claude A. Villee Villee, C. A. (1955). J. biol. Chem., 215, 171. Villee, C. A., and Gordon, E. E. (1955). J. biol. Chem., 216, 203. Villee, C. A., and Hagerman, D. D. (1953). J. biol. Chem., 205, 873. Villee, C. A., White, V. K., and Hastings, A. B. (1952). J. biol. Chem., 195, 287. Walaas, O., and Walaas, E. (1950). J. biol. Chem., 187, 769. Wang, H. W., and Hellman, L. M. (1941). Johns Hopk. Hosp. Bull., 73, 31. Wislocki, G. B., and Bennett, H. S. (1943). Amer. J. Anat., 73, 335. Wislocki, G. B., Dempsey, E. W., and Fawcett, D. W. (1948). Obstet. Surv. Baltim., 3, 604. DISCUSSION Huggett: The thing that I was interested in is this question of glucose production. It would seem that glucose can be produced metabolically within the placenta, and it would be desirable to distinguish that glucose, if it goes into the foetal blood, from the glucose which is passed from the mother to the foetus. So far as one can see that must be going on right away at full term when the metabolic production of glucose within the placenta ceases. Now the point that interests me there is that in the case of the sheep, where we have been able to study the passage of sugars across the placenta, this passage of glucose from the mother to the foetus can be reversed by reversing the gradient, but it is not in pro- portion to the gradient. The amount that moves is more than can be accounted for by the gradient, and equilibrations do not take place. We have to postulate that there is a cellular activity by the sheep placenta in regard to the passage of that particular monosaccharide. So there are two things — the metabolic production and the trans- mission. I do not know if Dr. Villee would agree with that, but it seems to me a point that we have to keep clear in any analysis. Villee: I would certainly agree with that, and I think there is no necessary correlation between the two; they may be quite different processes. When I am speaking of the metabolic production of glucose I am, of course, referring to the production of new glucose molecules from something else, and its actual secretion into the incubation medium where the aliquots are taken. Now the transport of glucose, either into a cell or across or through a cell, may be an entirely different process, one which does not involve the formation of a phosphorylated glucose intermediate. Of course, in the human being as well as in the sheep, glucose is obviously passing across the membrane, and I think it is in part, at least, by some active process though not necessarily one which involves phosphorylation. Huggett: The only other point I would emphasize is the interesting finding that there is no difference between normal and toxaemic placentas. That is rather unexpected. Villee: You may remember that Hellman was unable to find any Discussion 145 differences between normal and toxaemic placentas either. He had 3 or 4 toxaemic placentas, and as you know they are not too easy to come by. It is only because our laboratory is in a lying-in hospital that we have been able to get as many as we have. Amoroso : What part of the placenta, Professor Huggett, would you say was implicated in maternal glucose production ? Huggett: I have no knowledge at all about the site of production of glucose in any species. Williams: Do I detect in this discussion of glucose production and chemical function a suggestion that these are intrinsically regulated within the placenta, or are they determined by the hormone conditions in the maternal organism ? If glucose production decreases, does that mean anything except that there is no longer any need for the extra amount of glucose that there was earlier and that the greater need of the foetus earlier called in some change in the maternal organism to stimulate an extra production of glucose ? Villee: I do not know if we can speak in terms of need. I think Prof. Wislocki mentioned the histochemical evidence that the foetal liver is unable to store glycogen and secrete glucose early in pregnancy. Williams: It was only a suggestion that there was a need. Villee: Simply, in order to have a developing foetus, we must have some organ capable of regulating the glucose content of the foetal blood. The foetal liver cannot do it but the placenta very fortunately can. When the foetal liver takes over this function there is no longer any need for the placental enzyme and its function declines. I do not really know what is cause and what is effect, but this observation is quite general in enzyme phenomena. It is known that so-called adaptive enzymes occur widely in bacteria and examples of comparable enzyme changes are being found more and more in animal tissue and mammalian tissue. These are changes in enzyme activity and, presumably, in the actual amount of enzyme present, in response to the hour by hour needs for that particular enzymatic activity. The observation that when there is no longer any need for the secretion of glucose by the placenta, the enzyme glucose-6-phosphatase declines and disappears, suggests that this is an example of enzyme adaptation. Williams: Do you think you could recover that activity? Villee: I really do not know. Williams: What happens, for instance, in a diabetic pregnancy? Villee: We have tested quite a few diabetic placentas and they were no different from the normal ones. Williams: But, of course, they are controlled anyway by the insulin given to the mother throughout pregnancy. Villee: Yes, but we have had some diabetics which were in rather poor control and even their placentas showed no differences. Yemm: May I ask Dr. Villee whether during the course of his experi- ments with radioactively labelled glucose, he measured the rate of incorporation into glycogen ? Does the glycogen decline result from higher rate of breakdown or slower rate of synthesis ? Villee: Early in gestation there is a very active incorporation of 146 Discussion radioactive glucose into glycogen, whereas at term this occurs at a rate which is just barely detectable. I think this is primarily the result of a decreased rate of synthesis rather than an increased rate of breakdown. Huggett: I was going to ask whether Prof. Jost could advise us whether the foetal endocrines affect placental metabolism. Do you know whether they actually act on the placental metabolism as apart from the foetal organ metabolism? Jost: I know of no evidence in this field. It has been assumed that the placental glycogen is necessary for the foetus until the time the liver takes on the same function ; such an assumption might suggest that there is some correlation between the changes occurring in the placenta and in the liver of the foetus itself. Some results obtained on decapitated rabbit foetuses do not support this idea since the placental glycogen drops even when the liver does not store glycogen. But the foeto- placental endocrine relationship needs further study. Dempsey: I want to comment on the reference point which Dr. Villee used for the determinations he made. You refer your determinations finally to the dry weight of the slice? Villee: Only for oxygen, as it is generally done that way. The rest of the determinations are referred to the fresh weight of the tissue. Dempsey: In either case it is a reference to a weight determination. But isn't there a considerable change in the solid components of the placenta as it goes on to term ? My impression is that there is consider- ably more collagen, for example, in the placenta at term than there would be in some of the earlier placentas. Villee: The amount of solids goes from about 8 per cent up to about 13 per cent of the total weight. Dempsey: And some of that solid must be metabolically inert. Villee: Yes. Dempsey: So you have as your reference a shifting scale. Would it be possible to have for reference some other determination, which is correlated more sharply with the viable metabolically active tissue, with the living cells, for example, nucleic acid, phosphorus, or something of that sort ? Would you think that this would be as good or a worse reference point than the weight? Villee: It would be at least as good, and perhaps better. Since these changes are gradual and not particularly marked, I have plotted them both as wet weight and as dry weight, and the changes are evident in either case. Dempsey: Yes, but I was thinking that if there is an accumulation of metabolically inert solid material during the course of pregnancy, then the changes in the determinations which you observed would actually be greater than your figures show, would they not ? Villee: That would be so if our slices were representative of the whole placenta including the collagen. Of course, as collagen is a little diffi- cult to slice we try to find a villus as free of this connective tissue as possible. Dempsey: Even in the terminal villus, one of the very thin areas, which I think Prof. Wislocki showed on the screen a little while ago, Discussion 147 there is a considerable number of collagen fibres. So I do not think you can avoid it that way. Villee: No, all we can do is minimize it. Dempsey: Another comment I wanted to make was with reference to the problem of ageing. If I understand correctly the placenta does not show very great signs of ageing because it can still do a great many things, albeit perhaps not quite as rapidly as it did earlier in pregnancy. This somehow does not prove to me that the placenta is not in fact aged, because I think one might also say that an 80-year-old man can still do a great many things. Villee: You may remember, Prof. Dempsey, that I said that if we may equate ageing with these changes, or equate these changes in chemical function with ageing, then the placenta undergoes ageing. Dempsey: Yes. I simply was interested in the manner of ageing. I am sure that there are a great many functions which persist at a perfectly normal rate in an aged organism. That is, there are some functions which are not limiting in an ageing individual. For example, I believe that conduction velocity does not decrease with ageing in the nerve ; it either conducts or it does not conduct. But yet there are age changes in the nervous system in the sense that nerve cells fall out. So a particular biochemical or physiological event may or may not show changes with chronological increase in age. Villee: Quite right. I chose these particular functions since these have to do mostly with the release of energy, and I felt that if we did demonstrate marked changes in these, then I would believe that perhaps the tissue is really ageing in any sense you want to say. Perhaps we should have started the colloquium by defining ageing. Wislocki: In my own presentation, earlier this morning, I discussed the definition of ageing with respect to the peculiar mode of growth and structure of the mammalian placenta. We do, indeed, have to define what we mean by ageing in the case of the placenta. UPTAKE OF RADIO-POTASSIUM (42K) BY THE UTERUS AND PLACENTA DURING THE AD- VANCEMENT OF PREGNANCY IN THE RAT AND THE GOAT R. J. Harrison and J. L. D'Silva Anatomy and Physiology Departments, London Hospital Medical College, London, E.l During recent years increasing use has been made of experiments involving the introduction of radioactive sub- stances into the maternal circulation in order to examine placental uptake of various substances. Such experiments have been carried out in several species at varying stages of gestation, using radioactive sodium (Flexner and Roberts, 1939; Flexner and Pohl, 1941; Flexner and Gellhorn, 1942), iron (Vosburgh and Flexner, 1950), phosphorus (Nielson, 1941 ; Naeslund, 1951), calcium (Shirley, Jeter, Feast er, McCall, Outler and Davis, 1954) and strontium (Finkel, 1947) as tracer substances (for further references see also Reynolds, 1949 and Marshall, 1952). The distribution of injected radioactive potassium ions (as]42KCl) in animal tissues has been investigated by a number of workers (see D'Silva and Neil, 1951, for earlier references; Walker and Wilde, 1952; and Ginsburg, 1952). Their results show that the cells of organs such as the liver and kidney rapidly exchange their potassium with that of the plasma, whereas skeletal muscle cells and red cells exchange their potassium slowly. Since no experiments had been reported on the distribution of radioactive potassium ions injected in tracer amounts into pregnant animals the authors carried out a series of experiments on pregnant and non-pregnant rats (D'Silva and Harrison, 1953). The distribution of potassium was studied at different times after injection and at various 148 42K Uptake by Tissues in Pregnant Animals 149 stages of pregnancy in rats and guinea pigs. A series of goats, pregnant between 35 and 142 days, were subsequently used for further experiments and the relevant results, together with those on the rodents, are given below. The collection of material from a series of goats allowed histological and histo- chemical observations to be made on the changes occurring in the placenta during the advancement of pregnancy. These observations are briefly summarized below. Materials and methods Female albino rats weighing between 170 and 290 g. were used; the animals were killed at times varying from the 11th day of pregnancy until term. Female goats, mainly of the British Saanen breed, were used; the goats were mated at known times and were killed at intervals from the 35th day until just before term. The portions of the uteri and placenta not used for the estimation of radioactivity were fixed in various fluids for subsequent histological and histochemical examination by a variety of methods. Irradiated "Specpure" K2C03 was converted into an aqueous solution (2 per cent or 0-2 per cent w/v) of 42KC1, as indicated by D'Silva and Neil (1951). In rats the amount of K ion injected in each experiment was 0-3 to 1 -2 mg., and in goats 60 to 90 mg., an amount insufficient materially to alter the total amount of K in the extracellular fluid. Descriptions of the technical procedures carried out in order to estimate the radioactivity of the tissues will be found in the paper by D'Silva and Harrison (1953). The potassium content was determined by means of a direct reading flame photometer. Experimental results In a series of rats pregnant for 11 to 18 days the distribution of 42K in the uterus, placenta, kidney and foetuses was studied 5 minutes after intravenous injection of 42KC1. It will 150 R. J. Harrison and J. L. D'Silva be seen from Table I that with the smallest foetus (<9 mm.) the radioactivity in the uterus was less than that in the placenta. When the foetus was 9-18 mm. long the reverse was true. The results of a larger number of experiments, carried out near term and involving the sacrifice of the animals Table I The Activity of Tissues in the Pregnant Rat 5 Minutes after Intravenous Injection of 42KC1 Dura- tion of Foetal Rat No. preg- nancy in days length (mm.) Kidney Uterus Placenta Foetus 55 11 6 5-74 0-825 113 41 13 6 3-41 0-465 0-542 — 35 14 7 6-68 0-406 0-910 0 022* 34 14 8 3 15 0-313 0-391 0 030* 43 14 9 3-33 0-509 0-303 — 44 14 9 2-98 0-501 0-315 — 63 18 9 4-94 0-660 0-545 — 36 14 10 4-32 1140 0-689 0-015* 37 14 10 5-34 0-757 0-601 — 46 15 10 4 03 0-565 0-322 — 38 15 12 5-48 0-822 0-635 — 39 15 12 5-27 0-693 0-509 — 61 17 17 5-51 0-568 0-350 — 60 17 18 401 0-590 0-383 — Mean 4-59 0-630 0-545 S.D. — — ±114 ±0-210 ±0-242 * Includes foetal fluids. Activity of uterus, placenta and foetuses at various times after injection of 42KC1 into rat9. Results are expressed as percentage (x) of total radioactivity, x is defined by the expression : _ counts/min./g. tissue — total injected counts/min. where D is a correction factor allowing for decay. x 100 D at intervals of 5 minutes to 24 hours after injection of 42KC1 are described in the paper by D'Silva and Harrison (1953). The activity of the labyrinthine placenta near term falls from 0 • 53 per cent (per g. of tissue) of the total radioactivity after 5 minutes to 0 • 29 per cent after 24 hours. It is suggested that these results could be explained on the assumption of a slow 42K Uptake by Tissues in Pregnant Animals 151 rate of blood-flow through the uterine vessels (Barcroft and Rothschild, 1932; Barcroft, Herkel and Hill, 1933) together with a rapid rate of exchange of K ions. A further series of experiments was carried out on goats of various breeds, pregnant from 35 to 142 days, into which radioactive 42KC1 was injected 15 minutes before death. In almost every animal portions of the kidney and liver were removed and the radioactivity determined. In all but two animals the activity of the plasma was also determined and this allowed calculation of the relative potassium activity (R.P.A.). The latter is calculated according to the following expression : Activity of tissue mg.K/g. of plasma ~ Activity of plasma mg.K/g. of tissue where the activity is defined by the expression given in the footnote to Table I. Table II shows that the radioactivity of both the kidney and liver varied widely in different animals, but there was no systematic change in relation to the duration of pregnancy. On the whole, the radioactivity of the tissue was directly related to its potassium content which is shown by the rela- tively small variation in the figures for the R.P.A. of kidney (column 7) and of liver (column 10). The membranous chorion was removed from portions of the uterine wall and areas of mucosa carefully stripped from the underlying muscle. The radioactivity and K content of the uterine muscle and mucosa are shown in Table III. There were large changes in the radioactivity of these tissues in different experiments but these were roughly paralleled by changes in K content. Therefore the R.P.A. values for uterine muscle (column 5) and mucosa (column 8) were much less variable than either the radioactivity or the K content. In each animal either a whole placentome, or a portion of one, was removed and examined for its activity and its potassium content. In several animals a number of placen- tomes of varying shape and size were examined individually ; 152 R. J. Harrison and J. L. D'Silva there was no significant difference in their activity. Different slices (about 0-5 g.) of the placentome gave results similar to those obtained from a whole placentome removed from the same animal. The activity of the placentome was 0-0058 per Table II The Activity of Tissues in the Pregnant Goat 15 Minutes after Intravenous Injection of 42KCI Goat no. Days preg- nant Plasma Kidney R.P.A. Liver R.P.A. Activ- ity^ mg.K2 Activ- ity1 mg.K2 Activ- ity1 mg.K2 52 6 7 59 37 9 14 12 46 58 13 32 36 38 1 48 18 20 21 35 39 45 45 59 66 75 87 91 92 96 100 107 116 125 125 136 142 142 0 0012 0 0016 0 0012 0 0014 0 0018 0 0025 0 001 0-0013 00011 0 0011 0 0019 0 0022 0 0047 0 0021 00015 0-008 0 0012 017 0 14 0-40 0 14 016 0-38 0-39 015 0-24 0-22 0 18 0-14 0-22 0-18 0-28 019 016 0018 0 023 0014 0011 0 025 0018 0 022 0010 0 029 0015 0010 0017 0019 0 044 0011 0 020 0 023 0-014 0 024 2-31 2-72 3 19 1-32 2-20 2-68 2-72 2-90 2-50 2-82 2-71 2-40 3 15 206 1-96 305 3-82 311 1-10 0-75 1-45 0-83 0-98 102 1-40 115 1-30 1-20 0-75 0-89 0-26 0-84 1-40 0-84 1-00 0-008 0 013 0-009 0013 0010 0012 0012 0-006 0016 0014 0 004 0010 0012 0 030 0016 0015 0-008 0016 3-30 3-80 413 3 17 3-20 4-45 5-50 4-97 3-87 3-26 3-56 3-41 3-56 3-72 4-10 3-47 5 07 5-81 0-35 0-30 0-72 0-59 0-27 0-41 0-47 0-48 0-94 0-39 0-32 0-52 0-34 0-83 0-37 0-37 Mean S.D. 101 ±0-30 0-48 ±0-20 1 Activity determined as shown in footnote to Table I. 2 Potassium content expressed as mg. of K per g. of wet tissue. cent at the 35th day and 0-0083 per cent of the total injected radioactivity at the 142nd day of pregnancy. The average activity of the placentome throughout pregnancy was 0-0099 per cent ; the lowest figure being 0 • 0023 per cent on the 92nd day and the highest 0-032 per cent on the 91st day. The 42K Uptake by Tissues in Pregnant Animals 153 extreme figures were associated with a correspondingly de- creased or increased potassium content. There is thus no evidence that the uptake of radio-potassium per g. of placen- tome 15 minutes after injection varies during pregnancy. Table III The Activity of Tissues in the Pregnant Goat 15 Minutes after Intravenous Injection of 42KC1 Goat No. Days Preg- nant Uterus R.P.A. Mucosa B.P.A. Activ- ity1 mg.K* Activ- ity1 mg.K2 52 6 7 59 37 9 14 12 46 58 13 32 36 38 1 48 18 20 21 35 39 45 45 59 66 75 87 91 92 96 100 107 116 125 125 136 142 142 0 0063 0-0076 0 006 0 0043 0 0065 0-004 0 004 0-004 0 004 0 0033 0 0024 0-0056 0-0027 0017 0 009 0 0075 0 008 0 0037 0 006 2-26 2-90 2-71 2-26 2 04 2-87 2-87 1-86 105 2-71 3-33 2-88 2-77 216 1-70 2-93 2-52 6-20 3-39 0-39 0-23 0-63 019 0-28 0-21 0-84 0-44 0-26 0-39 0-09 0-50 0-24 0-21 0-59 0 14 0-24 0019 0 008 0 007 0015 0 007 0012 0-0057 0015 0-0082 0 0015 0-0082 0 0036 0-029 0 020 0 021 0 002 0-0058 0010 206 2-69 206 212 211 206 2-20 1-61 2-40 3 10 4-49 213 3 02 1-90 3-80 203 2-84 1-30 0-98 0-34 0-63 0-51 101 1-04 0-74 0-37 0-34 0-61 0-50 0-47 018 0-47 Mean S.D. 0-35 ±019 0-63 ±0-32 1 Activity determined as shown in footnote to Table I. 2 Potassium content expressed as mg. of K per g. of wet tissue. The membranous chorion was rich in potassium, but the quantity present was usually less than that in the placentome (per g. of wet tissue). The radioactivity of the membranous chorion was less than that of the placentome, except in Goat 14, where it was slightly higher. The activity of the mem- branous chorion per g. was 0 • 0004 per cent at the 35th day, 154 R. J. Harrison and J. L. D'Silva and 0-001 per cent of the total radioactivity at the 142nd day of pregnancy. The average of all the estimations of the activity of the chorion was one fifth of that of the placentome 15 minutes after injection. The R.P.A. of the membranous Table IV The Activity of Tissues in the Pregnant Goat 15 Minutes after Intravenous Injection of 42KC1 Goat No. Days preg- nant Placentome R.P.A. Membranous chorion R.P.A. Activ- ity1 mg.K2 Activ- ity1 mg.K2 52 6 7 59 37 9 14 12 46 58 13 32 36 38 1 48 18 20 21 35 39 45 45 59 66 75 87 91 92 96 100 107 116 125 125 136 142 142 0 0058 0-0074 0013 0-0077 0 010 0-006 0-006 0 0044 0 032 0 0023 0 0025 0 0074 0-009 0017 0 023 0013 0 009 0-005 0 0083 1-35 1-47 4 19 2-20 2-53 3-58 2-36 2-78 7-30 1-84 3 15 5-76 2-80 2-50 2-68 2-23 2-30 307 0-61 0-44 102 0-35 0-35 0-33 0-63 0-44 0-27 0-26 0-38 0-48 0-41 0-76 0-48 0-36 0 0004 0 0013 0- 00045 0- 00097 0-0038 0- 00063 0011 0- 00034 0 0013 0 0014 0- 00036 0 0029 0 0004 0 0029 0-0026 0-0014 00011 0- 00049 0-0011 0-96 1-47 1-42 2-39 1-11 102 2-37 2-54 1-45 1-58 1-03 1-90 1-55 1-39 1-36 1-71 1-49 5-40 2-87 006 008 011 0 04 0-32 0-10 0-05 010 0-19 0-31 0 05 0-13 0-09 007 0-14 0 02 005 Mean S.D. 0-45 ±0-22 0-11 ±0 09 1 Activity determined as shown in footnote to Table I. 2 Potassium content expressed as mg. of K per g. of wet tissue chorion was considerably less than that of the placentome, which suggests that the chorion exchanges its K at a slower rate than the placentome. It is possible that this difference is due to the lack of opportunity for exchange between the blood and the cells of the chorion. 42K Uptake by Tissues in Pregnant Animals 155 Morphological observations The experimental observations outlined above made available material from a series of animals killed after a known period of pregnancy. Material from other goats, also of known gestation age, has resulted in the examination of over sixty animals killed at different stages of pregnancy. Recent papers on the developmental changes in the placenta of the Artiodactyla include those by Wimsatt (1950, 1951), Harrison and Hamilton (1952), Bjorkman (1954) and Harrison and Hyett (1954). A full review of earlier papers on the placenta of the cow, sheep and pig has been given by Amoroso (1952). The placenta of the goat is polycotyledonary with from 120 to 180 cup-shaped or concave placentomes (PL I, Fig. 5) arranged more or less in rows. The placentomes on the meso- metrial aspect of the central portion of each uterine horn are, for the first third or more of pregnancy, substantially the larger. As pregnancy advances the cranially placed placentomes increase more rapidly in size than those centrally placed, but do not reach a comparable size. Occasionally flattened, sessile, oval placentomes are encountered together with the concave variety, and in several animals killed during the middle of pregnancy only the flattened, oval variety was present. It should be noted that the uptake of radioactive potassium by the two types of placentome was not significantly different. It is not yet clear if this morphological difference is genetic in origin, or a manifestation of change in the form of the placen- tome during development. Primary villi are present and have penetrated deep into the caruncular tissue even by the 39th day of pregnancy (PL I, Fig. 4). The primary villi are at first straight and simple, but active division at their tips results in the formation of secondary and tertiary villi, each fitting into a corresponding maternal crypt. From the earliest stages studied (35th day) it appears that the trophoblast has considerable powers of attrition and the maternal epithelium is destroyed not only in the crypt walls but also in localized areas in relation to the 156 R. J. Harrison and J. L. D'Silva membranous or intercotyledonary chorion (PL I, Figs. 1 and 3). Near the edges of the developing caruncles the chorion is thrown into folds or intermediate villi, the areas thus resemble the arcade formations described by Assheton (1906) and Bjorkman (1954). The maternal epithelium also shows signs of impending degeneration in the septal walls surrounding the intermediate villi (PL I, Fig. 2). Only in the regions of the mouths of the uterine glands does the maternal epithelium display the characteristics of healthy tissue, and where the chorion comes into contact at the edges of the gland mouth localized areas of cellular attrition are present (PL I, Fig. 3). The powers of attrition apparently possessed by the chorionic epithelium may be reflected by the increase in density of the sub-epithelial stratum compactum to a degree greater than that observed by Hamilton and Harrison (1951) in the non-pregnant uterus of the goat. This increase in density appears to be due not only to changes in the cellular population but also to increase in the quantity of connective tissue and changes in its staining qualities and also in those of the intercellular substance. It is noticeable that the de- struction of maternal tissue attributable to the chorion is capable of causing localized haemorrhages between foetal and maternal tissues. Such haemorrhages can be seen in early PLATE I Fig. 1. Photomicrograph to show destruction of the maternal epithelium during the early stages of pregnancy. The animal had been pregnant for 30 days. Section stained by the P.A.S. method, x 70. Fig. 2. Photomicrograph to show the formation of folds in the chorion at the edge of the developing caruncle. The animal had been pregnant for 30 days. Haematoxylin and eosin. x 70. Fig. 3. Photomicrograph of a membranous chorion in relation to the mouth of a maternal gland. The animal had been pregnant for 30 days. Haematoxylin and eosin. x 75. Fig. 4. Photomicrograph to show the appearance of the young villi in the developing caruncle at the 30th day of pregnancy. Haematoxylin and eosin. X 42. Fig. 5. The appearances of the placentomes on the 50th day of pregnancy, showing haemorrhages on the surfaces of the placentomes. x £. Fig. 6. Photomicrograph of a small accessory placentome on the 66th day of pregnancy, showing an extravasation of blood between the maternal and foetal tissues. Plate I. '-"•kJ* ,- _., * Jjtr-**"-**^;- '-' - <_.' '-''1 facing page 156 Plate II. *^-, 42K Uptake by Tissues in Pregnant Animals 157 pregnancy (PL I, Fig. 6), the middle of pregnancy (PL I, Fig. 5), and in late pregnancy it is common to find evidence of both early and more recent haemorrhages in one area of the edge of the placentome. From about the 45th day of pregnancy numerous laterally directed branches develop all along the extent of each primary villus. These side branches divide frequently and extend into the crypt walls. The latter are soon extensively invaded (PL II, Fig. 2) and examination of the specimens killed during the latter third and terminal stages of pregnancy suggests that there is a gradual attrition of the tissue surrounding each maternal capillary. Near term much of the crypt wall has been lost and only remains as strands of tissue connecting and surrounding the network of maternal vessels (PL II, Fig. 5). There appears to be a change in the characteristics of the septal connective tissue in that it becomes more markedly P.A.S. positive with the advancement of pregnancy (PL II, Figs. 3 and 4). Examination of the membranous chorion during the latter part of pregnancy suggests that the destructive powers of the chorion are not necessarily restricted to the earlier period (PL II, Fig. 1). It is possible that there may be efforts to renew the maternal uterine epithelium by regeneration from the intact epithelium of the gland mouths. In any event there PLATE II Fig. 1. Photomicrograph to show the appearances of the membranous chorion at about the 90th day of pregnancy, and the relationship of the chorion to the mouth of the uterine gland. Haematoxylin and eosin. x 94. Fig. 2. Photomicrograph to show branching of the villi within the placentome of an animal pregnant for 96 days. Haematoxylin and eosin. x 28. Fig. 3. Photomicrograph showing the tips of the villi within the maternal crypts at the 86th day of pregnancy. The section is stained by the P.A.S. method and shows the foetal giant cells, and also the connective tissue of the crypt cells. X 68. Fig. 4. Photomicrograph to show the bases of the foetal villi where they enter the maternal crypts ; also showing the mouths of the crypts. Section stained by the P.A.S. method, x 50. Fig. 5. Photomicrograph of the branches of the villi and the maternal capil- laries in the centre of a placentome at the 142nd day of pregnancy. Haema- toxylin and eosin. x 175. 158 R. J. Harrison and J. L. D'Silva is frequently to be observed a narrow layer of pyknotic, heavily staining nuclei separating the chorionic cells from the maternal stroma, not only in the region of the membranous chorion (PL II, Fig. 1) but also in the placentome (PL II, Fig. 5). This layer may represent the attenuated and dying maternal epithelium, but it is also possible that it is the remnants of foetal cells which have invaded and destroyed the maternal epithelium and have themselves subsequently died (Harrison and Hamilton, 1952). Trophoblastic giant cells, possessing the histochemical characteristics so clearly described by Wimsatt (1951) in the sheep, are observable at the edges of the villi (PL II, Fig. 3). Similar cells can also be observed in situations which strongly suggest that the giant cells pass across and become inter- calated within or even replace the maternal epithelium. It is, however, difficult to decide whether these cells can exist for very long in their close relationship to the maternal stroma. Acknowledgements. The authors express their thanks to Dr. F. L. D. Steel and Mr. C. J. Turner for their help with the experimental work, to Mr. R. Q. Cox and Mr. R. F. Birchenough for technical assistance, and to the Agricultural Research Council and the Yarrow Fund of the London Hospital Medical College for grants to defray expenses incurred in purchasing animals and radioactive isotopes. REFERENCES Amoroso, E. C. (1952). In Marshall's Physiology of Reproduction. 3rd ed. vol. 2, p. 127. London: Longmans, Green and Co. Assheton, R. (1906). Philos. Trans. B., 198, 143. Barcroft, J., Herkel, W., and Hill, S. (1933). J. Physiol, 77, 194. Barcroft, J., and Rothschild, P. (1932). J. Physiol., 76, 447. Bjorkman, N. (1954). Acta Anat., 22, Suppl. 2, 1. D'Silva, J. L., and Harrison, R. J. (1953). J. Embryol. exp. Morph., 1, 357. D'Silva, J. L., and Neil, M. W. (1951). Biochem. J., 49, 222. Finkel, M. P. (1947). Physiol. Zool., 20, 405. Flexner, L. B., and Gellhorn, A. (1942). Amer. J. Obstet. Gynec, 43, 965. Flexner, L. B., and Pohl, H. A. (1941). J. cell. comp. Physiol., 18, 49. Flexner, L. B., and Roberts, R. B. (1939). Amer. J. Physiol., 128, 154. Ginsburg, J. (1952). Fed. Proc, 11, 54. 42K Uptake by Tissues in Pregnant Animals 159 Hamilton, W. J., and Harrison, R. J. (1951). J. Anat., Lond., 85, 316. Harrison, R. J., and Hamilton, W. J. (1952). J. Anat., Lond., 86, 203. Harrison, R. J., and Hyett, A. R. (1954). J. Anat., Lond., 88, 338. Marshall, F. H. A. (1952). Physiology of Reproduction. 3rd ed. vol. 2. London: Longmans, Green and Co. Naeslund, J. (1951). Acta obstet. gynec. scand., 30, 230. Nielson, P. E. (1941). Amer. J. Physiol, 135, 670. Reynolds, S. R. M. (1949). Physiology of the Uterus, 2nd ed. New York: Hoeber. Shirley, R. J., Jeter, M. A., Feaster, J. P., McCall, J. T., Outler, J. C, and Davis, G. K. (1954). J. Nutr., 54, 59. Vosburgh, G. J., and Flexner, L. B. (1950). Amer. J. Physiol., 161, 202. Walker, W. G., and Wilde, W. S. (1952). Amer. J. Physiol., 170, 401. Wimsatt, W. A. (1950). Amer. J. Anat., 87, 391. Wimsatt, W. A. (1951). Amer. J. Anat., 89, 233. DISCUSSION Wislocki: I should like to ask Prof. Harrison whether, besides the chorioallantoic placenta, he also investigated potassium transfer through the yolk-sac placenta, and what differences he found ? Harrison: We made a few estimations of that. The uptake after 15 minutes and one hour were negligible. The uptake was so small that it was not significant in comparison with the uptake of other parts of the placenta. And that frankly did surprise us. I would have expected that there would be considerable uptake, but there was not — but that was not near term and it may be that that is relevant to Prof. Dempsey's comments earlier on. Hnggett: There are two points. First, Cloette observed in the placentomes in sheep and goats the change in cup-shaped to the flattened type, and I would rather associate it with age ; that at full-term or post- partum there is a preponderance of the flattened placentomes, but at mid-pregnancy or earlier you would have these cup -shaped. I have not worried very seriously about the matter. The second point is, Dr. Seoras Morrison has been working with me during the last few months on the uptake of [14C] glucose, given to the mother, by glycogen in the rabbit placenta and there we find quite definite differences between the decidual glycogen and the chorionic glycogen. They do not behave in the same way. I was also very interested in the general finding you had that the placenta compared with the liver has a very latent active uptake — I hope I may use that term. Harrison: Yes — the cup-shaped placentomes have a potassium con- tent and an uptake of radioactive potassium equivalent to that in the flattened placentomes. Both liver and placentome are rich in potassium, but the uptake of radioactive potassium by the placentome is not as great, therefore one thinks of differences in the blood supply, and the rate of flow and thus the opportunity for exchange. 160 Discussion Strauss: I wonder if Prof. Harrison found any potassium in the walls of the foetal blood vessel of the placenta. A couple of years ago we found some potassium in human mature placentas. The potassium is in the syncytiotrophoblast as well as in the walls of the foetal blood vessels. We found about the same distribution of the calcium. Have you any experience of that ? Harrison: We have not done any quantitative estimations on that. We have only made some autoradiographs, and one can detect the potassium in the chorion, but we have not looked carefully or made any comparison with the amount in the walls of the blood vessels. We have estimated the uptake in the umbilical cord and that is insignificant compared with the uptake by the placentome. So it was from that experiment that we thought we could argue that the connective tissue in the placenta did not take up potassium to a degree parallel to the chorionic epithelium. Huggett: You are not distinguishing between Wharton's jelly and the vessels of the cord ? Harrison: We simply took a chunk of the cord. Hamilton: How early do the haemorrhages occur, Prof. Harrison? I have not seen those before in any of the ungulates. Harrison: That picture I showed was, I am almost sure, 41 or 42 days old, but we found such localized haemorrhages right from the start of the formation of the placentomes. It is quite different from the state of affairs that we found in the deer. We did not find any haemorrhages, as far as I can remember, at any stage in any species of deer we looked at. Hamilton: I take it that the trophoblast is eroding maternal blood vessels in some way or another. Amoroso: I have no doubt in my own mind that the trophoblast is endowed with erosive properties. That it may destroy the maternal capillary endothelium is a real possibility. As to the question of the two types of placentomes to which Prof. Harrison has drawn our atten- tion, these have been figured in the uterus of the sheep by Cloette and they are also known to occur in the uterus of many antelopes. They are not indicative of age changes in the placenta. Huggett: It was in the goat that I first saw it. Amoroso : On the other hand, changes in the connective tissues which may often be very widespread, are invariably associated with placental degradation. Montagna: A great deal has been said about the efficiency of the placenta and I believe that we are being as arbitrary with this term, as we have been also with ageing and senescence. I wonder who is the better judge of efficiency, our arbitrary determination of efficiency, that is, whether or not potassium or iron, etc., go through at a certain rate that we decide is efficient or not inefficient, or if the embryo is a better judge of efficiency. Amoroso: Are you suggesting that it is the quality of the embryo at the end of gestation which is the criterion? Montagna: I am saying that the placenta is always efficient if the embryo is going to survive. MODIFICATIONS IN THE FOETAL DEVELOP- MENT OF THE RAT AFTER ADMINISTRATION OF GROWTH HORMONE OR CORTISONE TO THE MOTHER Herbert Tuchmann-Duplessis and lucette mercier-parot Faculty of Medicine and Ecole Normale Superieure, University of Paris "The art of living," said the Greeks, "consists in dying young, but as late as possible." This, surely, is the goal that gerontology has set for itself. When one considers the problem of ageing, which is among the most fascinating for the biologist, one question at once arises: are the development, functioning and evolution of the organs in the course of life unalterably fixed by heredity or do they also depend to a certain extent on internal and external factors over which we may have some influence? The second alternative is the more probable, if not the more desirable. The increase in longevity which we have witnessed constitutes an encouraging argument in this respect. For the past century it has been affirmed that the endocrine glands are capable of delaying or hastening ageing, and it is this idea that has been the basis of hormone therapy and the great advances in endocrinology. If the over-optimistic theories of the early workers in endocrinology have unfortu- nately not proved correct, it is nevertheless probable that the endocrine balance does play an important part in the growth, functioning and ageing of the organs. Numerous authors have carefully analysed the morpho- logical processes in the ageing of the endocrine glands. Loeb (1941), who has made a very detailed study, has de- scribed the progressive modification of the different endocrine glands with age; he has also shown that depending on the AGEING VOL. 2 161 7 162 H. TUCHMANN-DUPLESSIS AND L. MeRCIER-PaROT experimental conditions certain hormones, such as the oestro- gens, are capable of delaying or, on the contrary, of hastening the ageing of the endocrines. It seems, therefore, as if modi- fications in hormonal equilibrium are capable of influencing the general process of development and ageing. During the last two years we have tried to examine the problem of the relationship between the endocrine glands and the somatic development by modifying the endocrine balance of the pregnant female. The results of two series of experiments are given here, one on the action of the somatotrophic hormone (STH), the other on that of cortisone. Somatotrophic Hormone The well-known part played by the anterior pituitary in somatic development has led many authors to consider somatotrophic hormone as a stimulant not only of postnatal growth but also of embryonic development. That, at least, is the conclusion reached by Teel (1926), Hain (1932), Sontag and Munson (1934) and Watts (1935), who observed that the injection of crude extracts of the anterior pituitary into pregnant rats gave rise to foetal gigantism. Using the purified hormone Hultquist and Engfeldt (1949), Engfeldt and Hult- quist (1953), Nixon (1954) and Cotes (1954) also produced in rats enormous foetuses weighing up to 7 and 8 g. However, the injection of STH into pregnant females of other species, such as the dog and the cat, gave inconsistent results (Young, 1946). The question of the foetal gigantism obtained with STH has often been raised in connection with the large children born of diabetic mothers, and highly ingenious interpretations have been put forward to explain the exaggerated foetal growth. However, it seems improbable that STH, due to its high molecular weight, could cross the placental barrier, and the possibility of stimulating embryonic growth with STH seemed astonishing, and therefore we repeated the experi- ments on the rat. Fig. 1. Newborn rats; left, control; Fig. 2. Rat embryos removed on the 20th day right, offspring of an STH-treated of gestation. Left , control ; right , embryo of mother. STH-treated mother. Fig. 3. Above: left to right, embryos of control rats at the 20th, 19th and 18th days of gestation; below: two 20-day embryos removed from STH- treated mothers. facing page 162 B Fig. 4. Sections of rat embryos removed on the 20th day of gestation. Left, controls; right, embryos of STH-treated mothers. A and B— sagittal sections. C — frontal sections. H— hypophysis or pituitary. Ep— Epiphyseal gland. Influence of STH and Cortisone on Foetal Growth 163 Method We injected the somatotrophic hormone (Choay) in doses of 10 i.u. per day into several groups of pregnant rats. One group was injected daily from the 3rd to 16th day, another from the 4th to 16th and others from the 5th, 6th, 7th and 8th to 16th day. In the case of the group treated from the 3rd to 16th day we either allowed the pregnancy to con- tinue until term, or removed the embryos on the 20th day of gestation. Results The administration of STH from the 3rd to 16th day of pregnancy causes a greater increase in the weight of the mother than normal and prolongs the duration of pregnancy by from 3 to 6 days. The foetuses, all stillborn, are large, their average weight being 6-7 g. against the 4 • 5 g. of the controls (Fig. 1). The removal of the foetuses on the 20th day of pregnancy gave apparently contradictory results. The embryos of the STH-treated mothers are, in fact, smaller than the controls. On the 20th day of pregnancy the average weight of 68 foetuses of STH-treated mothers was 1 • 22 g. against the 2 • 18 g. of the controls : that is, a difference in weight of 40 per cent compared with the controls (Fig. 2). These embryos appear normal, but are much more juvenile in appearance (Mercier-Parot and Tuchmann-Duplessis, 1955). On the 20th day the largest of the embryos of STH-treated mothers resemble control embryos of 19 days and the smallest, control embryos of 18 days (Fig. 3). Microscopic study of these embryos confirms this impression of juvenility. Serial section of several embryos did not, in fact, reveal any anomaly, but a very marked difference in the surface of the various organs is regularly found, both in the central nervous system and in the visceral organs. In fact, the organs of the embryos of mothers treated with STH are comparable in appearance on the 20th day to embryos of 18 to 19 days (Fig. 4). The results of these two series of experiments led us to 164 H. Tuchmann-Duplessis and L. Mercier-Parot revise our ideas of "foetal gigantism of pituitary origin". In actual fact, if one compares the weight of the newborn of STH-treated mothers which are born between the 24th and 28th day of pregnancy with that of young rats of the same actual age, that is to say two to six days post-partum, one observes that the latter are always the larger. The greater 10 9-1 8 w 6H -m 5- -C C7 - k S 6 7 Day treatment commenced Fig. 0. Average weight of embryos removed on the 20th day of gestation; the differences in weight are only significant when STH is administered before the 5th day of gestation. were tempted to explain it by an increase of maternal protein anabolism. This could in fact deprive the foetus of part of the nutritive substances intended for it. However, a comple- mentary experiment, in which we commenced STH adminis- tration 4, 5, 6, 7 and 8 days after conception, showed us that foetal growth is inhibited to a significant degree only when STH is administered before the 5th day of gestation (Fig. 6). This fact suggests another mechanism, for example an early enzymatic disturbance, which has a deleterious effect on the development of the embryo, as has been clearly shown by the work of Warkany (1948) and Giroud (1954). 16G H. TucHMANN-DurLESSis and L. Mercier-Parot With regard to the problem of the influence of the maternal pituitary on the development of the embryo, our experiments with hypophysectomy on the 12th day of pregnancy, that is, when the placental secretion is sufficient to make up for the hypophyseal gonadotropins, show that hypophysectomy at this stage is generally well tolerated, and we had only a few postoperative abortions. The progress of pregnancy is in most cases normal. The foetuses removed on the 20th day are normal and their weight identical with that of embryos from control rats which had undergone a spurious operation. The average weight of the offspring of hypophysectomized mothers is 2-12 g. as compared with 2-20 g. for the controls. Our results confirm those of Campbell, Innes and Kosterlitz (1953), and in no case have we observed any diminution in weight like that reported by Knobil and Caton (1953). More- over, when pregnancy is allowed to proceed to term and after, it is observed that foetal growth continues. With the excep- tion of two rats which gave birth on 22nd-23rd day to live foetuses of normal weight, the others showed difficulty in giving birth and we had to have recourse to artificial deliveries between the 24th and 25th day in order to recover the foetuses. It is interesting to note, on this point, that with the prolongation of gestation, the weight of the foetuses increases even in the absence of the pituitary, and that those removed on the 24th day weigh 5 • 6 g. Discussion These observations show that the development of the embryo is relatively independent of the pituitary control. When one compares the results of our experiments with hormone therapy and hypophysectomy, one is led to con- clusions which are directly opposed to frequently expressed opinions on the part played by somatotrophic hormone in embryonic growth. In point of fact, taking as a basis the experiments with STH administration in course of which foetal gigantism was thought to be observed in the rat, Young (1953) admits that the somatotrophic activity of the pituitary Influence of STH and Cortisone on Foetal Growth 167 is increased during gestation. Moreover, the author believes that the foetal gigantism which occurs in certain pathological conditions, notably with diabetic and pre-diabetic mothers, could be brought about by a somatotrophic hypersecretion of the maternal pituitary. However, the somatotrophic hypersecretion in the pregnant female, postulated by Young and often accepted by clinicians, has not been demonstrable. The results of our experiments make the suggestion that the influence of STH is the cause of foetal gigantism, hardly credible. Moreover, the fact that maternal hypophysectomy does not inhibit foetal develop- ment also shows that embyronic growth is not pituitary- controlled. This interpretation is supported by the fact that the destruction or ablation of the foetal pituitary (Raynaud and Friley, 1947; .Tost, 1947) does not appear to slow down foetal growth in the mouse or rabbit. It also links up with clinical observations on the children of acromegalic mothers. Neither Jackson (1954) nor Huant (1955) have observed the stigma of gigantism in the children of acromegalic mothers; in one case, Huant even noted signs of marasmus and under- development. Thus these experimental results and clinical observations confirm each other and that somatotrophic hormone does not enter into the embryonic development of mammals. Nevertheless, the inhibition of the embryonic growth of the rat, which we observed with early administration of STH and which is probably indirectly caused, is worthy of our attention, for the possibility of consistently obtaining embryos of more juvenile appearance may provide an interesting method of studying the metabolic factors conditioning em- bryonic growth. Cortisone The inhibition of embryonic growth by somatotrophic hormone led us to examine the influence of another hormone, cortisone. It is remarkable that whilst somatotrophic hormone 168 H. TUCHMANN-DUPLESSIS AND L. MeRCIER-PaROT increases protein anabolism, cortisone increases protein catabolism. When pregnant rats are treated from the 6th to the 16th day of pregnancy with cortisone in doses of 20 mg. per day, disturbances of gestation and delivery are observed. Courrier and co-workers (1951), like Robson and Sharaf (1952), have similarly observed in the rat and the rabbit disturbances of gestation, abortions and difficulties in delivery under the influence of cortisone and ACTH. With the doses of cortisone which we have used (Tuchmann- Duplessis and Mercier-Parot, 1954), abortions in our breeding rats are relatively rare. The foetuses are, in the majority of cases, live, of normal appearance and their average weight is only 10 per cent lower than that of the controls. We have not observed malformations like those reported by Fraser, Fain- stat and Kalter (1953) in mice. Nevertheless, in spite of their normal appearance, these new- born rats are extremely frail. Although appearing to nurse normally, that is their milk intake is normal, the majority of the young rats die between the 3rd and 4th day. The growth of the rare survivors is, as shown in Fig. 7, considerably slowed down or completely arrested. Sickly, rapidly becoming senile and wrinkled in appearance, these animals do not generally survive for more than 12 to 18 days. Thus, if the administration of cortisone to the pregnant rat only slightly delays embryonic development, it almost com- pletely arrests postnatal growth. Two series of experiments were carried out in order to investigate the possible causes of the inhibition of post- natal growth: (1) the exchange of newborn between cortisone- treated mothers and control mothers, and (2) the treatment of control mothers with cortisone after parturition. In the first series of experiments out of 10 attempts we have only succeeded so far with one exchange of newborn rats. However, the result seemed noteworthy, for, after an initial delay, the few newborn from the cortisone-treated mother and reared by the control mother, developed relatively well. Influence of STH and Cortisone on Foetal Growth 169 Conversely, the offspring of the control mother suckled by the rat treated with cortisone during gestation, are retarded in growth and show dorsal alopecia as well as markedly delayed sexual development. Although in both cases the milk intake of the newborn appears to be normal, the result of the Controls Cortisone Fig. 7. Growth curves of young rats born of mothers treated with cortisone during pregnancy. exchange of mothers indicates that cortisone has an unfavour- able effect on lactation. The results on two sets of animals obtained at a six months interval prove this. The rats in the first set received 20 mg. of cortisone for 5 days following parturition and 10 mg. after this; those in the second set received 20 mg. of cortisone for 15 days following parturition. The cortisone treatment of the 170 H. Tuciimann-Duplessis and L. Mercier-Parot nursing mothers considerably delays the somatic develop- ment of the newborn and increases their mortality (Mercier- Parot 1955). The course of the growth of the young rats Fig. 8. Growth curves of young rats nursed by mothers treated with cortisone after parturition. Influence of STII and Cortisone on Foetal Growth 171 shows two characteristics, as shown in Fig. 8. During the first phase of 4-6 days, the retardation in growth compared with the controls is slight, being of the nature of 15 to 20 per cent. From the 6th day the arrest in development is accent- uated and on the 10th day the difference in weight compared with the controls is from 50 to 60 per cent. The majority of the young die between the 12th and the 15th day. Of 50 new- born, 5 reached the age of 30 days and only two have survived two months. Although these two rats partially made up their delay in somatic development during the second month, the genital organs retained their infantile appearance. Thus at first sight cortisone proves just as deleterious to the growth of young rats when it is administered during lactation as when it is administered during gestation. Conclusions The two series of experiments that we have described here show that, in spite of its autonomy, foetal development can be modified by disturbances in the endocrine balance of the mother. Contrary to the long accepted opinion, it does not seem feasible that foetal gigantism can be accounted for by the maternal somatotrophic secretion. Although hypophysectomy in the mother on the 12th day of pregnancy does not influence the development of the embryos, the early administration of STH unquestionably inhibits the embryonic development. At the same age, the weight of the foetuses of STH-treated mothers is about half that of the controls, and these embryos are more juvenile in appearance. At 20 days the surfaces of the nervous system and of the visceral organs of these embryos have scarcely reached the stage of development of control embryos of 19 days. If the causes of this growth inhibition still remain unexplained, the fact itself is interesting for it opens up new possibilities of research into the metabolic factors capable of influencing somatic growth. The results of cortisone administration show that, in spite 172 H. TUCHMANN-DUPLESSIS AND L. MeRCIEK-PaROT of the slight difference in weight at birth — 10 per cent lower than the controls, the postnatal development of these young rats is greatly inhibited or arrested. Although we have been able to show that this inhibition appears to be connected with disturbances of lactation, the great frailty of the offspring of cortisone-treated mothers seems to indicate the existence of a more or less reversible change in their constitution. Acknowledgement. We wish to express our thanks to M. Memin, for taking the photographs. REFERENCES Campbell, R. M., Innes, L. R., and Kosterlitz, N. (1953). J. Endo- crin., 9, 68. Cotes, P. M. (1954). J. Endocrin., 10, 14. Courrier, R., and Colonges, A. (1951). C. R. Acad. Sci., Paris, 232, 1164. Courrier, R., Colonges, A., and Baclesse, M. (1951). C. R. Acad. Sci., Paris, 233, 333. Engfeldt, B., and Hultquist, G. T. (1953). Acta endocr., 14, 181. Fraser, F. C, Fainstat, T. D., and Kalter, H. (1953). Etudes neo- natales, 2, 43. Giroud, A. (1954). Biol. Rev., 29, 220. Hain, A. M. (1932). Quart. J. exp. Physiol., 22, 71. Huant, E. (1955). Personal communication. Hultquist, G. T., and Engfeldt, B. (1949). Acta endocr., 3, 365. Jackson, W. P. U. (1954). J. din. Endocrin., 14, 177. Jost, A. (1947). C. R. Acad. Sci., Paris, 225, 322. Knobil, E., and Caton, W. L. (1953). Endocrinology, 53, 198. Loeb, L. (1941). Harvey Lectures, 36, 228. Mercier-Parot, L. (1955). C. R. Acad. Sci. Paris, 240, 2259. Mercier-Parot, L., and Tuchmann-Duplessis, H. (1955). C. R. Acad. Sci., Paris, 240, 455. Nixon, W. C. W. (1954). Ann. Endocr., Paris, 15, 20. Raynaud, A., and Friley, M. (1947). C. R. Acad. Sci., Paris, 225, 596. Robson, J. M., and Sharaf, A. A. (1952). J. Physiol., 116, 236. Sontag, L. W., and Munson, P. L. (1934). Amer. J. Physiol., 108, 593. Teel, H. M. (1926). Amer. J. Physiol., 79, 170. Tuchmann-Duplessis, H., and Mercier-Parot, L. (1954). C.R. Acad. Sci., Paris, 239, 1689. Warkany, J., and Roth, C. B. (1948). J. Nutr., 35, 1. Watts, R. M. (1935). Amer. J. Obstet. Gynec, 30, 174. Young, F. G. (1946). Schweiz med. Wschr., 78, 894. Young, F. G. (1953). Expos, ann. Biochem. med., 15, 1. Discussion 173 DISCUSSION Jost: The two kinds of experiments which were shown seem to be rather different. In the first one, pituitary extracts were used and the question immediately arises whether the extracts were completely free of substances stimulating other endocrine glands, such as the adrenals or the ovary. It is well known that maternal hormonal imbalances may interfere with foetal growth. Concerning the direct action of growth hormone on the foetus I ob- served no increase in size of rabbit foetuses which were given directly 1*5 mg. growth hormone (Choay) under the skin. On the other hand, I observed a considerable reduction of the foetal growth as a result of some maternal hormonal treatments. For instance, I compared the average size of rat foetuses after administration of androgenic com- pounds to the mother. In our colony the mean weight of control foetuses on day 21 is about 5 g. In rats injected with testosterone phenyl- propionate, testosterone oenanthate or androstenediol dipropionate (10 mg./day) from day 15 to day 20 the average weight of the foetuses was about 3 g. ; on the contrary, with methylandrostenediol dipropionate the weight was normal. The exact point of attack of these substances is unknown. I wished to use such examples just to draw attention to the fact that maternal hormonal imbalances may alter foetal growth, and to raise the question of the extent to which the injected growth hormone changed the maternal hormonal equilibrium. T.-Duplessis: As far as growth hormone is concerned, I do not think at all that it would be a direct action, as there is good reason to think that such a high molecular weight hormone will never cross the placental barrier. The problem we were studying was the following one: will growth hormone induce foetal gigantism or not ? In the last few months it has been claimed that growth hormone induced foetal gigantism, but we could show that it does not, and we believe that the greater weight of these foetuses is only due to a prolonged pregnancy. How growth hormone prolongs pregnancy, we do not know. It was a very fascinating explanation to say that growth hormone contains some gonadotrophin and, therefore, it will modify the functioning of the ovary. But curiously enough, the results you get with pure growth hormone, that biochemists claim to be devoid of gondadotrophins, are the same as those one gets with crude growth hormone in which there is the high quantity of gonadotrophic hormone. And I may mention also that Prof. Nixon, who utilized a pure growth hormone from a tested Holland preparation, got larger foetuses in the rat. I was puzzled by this problem of growth hormone for other reasons because many years ago it was shown that there was a very good corre- lation between the cytological picture of the pituitary and growth, first in the mother. But surely during the pregnancy, there is a big increase of the eosinophil cells which are supposed to be the source of growth hormone. And then other people have shown that in the embryo first of all eosinophil cells appear and they correlate this very early appearance of eosinophils with the function of the somatic development, while the baso- philic cells will directly determine the sexual development of the embryo. 174 DiscussiOiN Jost: I wish to introduce a remark about some papers claiming that growth hormone injected into pregnant rats produces foetal gigantism. I must say that a careful examination of the published tables often leads to other conclusions than those expressed by the author in the text accompanying the tables. I know of no clear demonstration of foetal gigantism induced by growth hormone. T.-Duplessis: No, they got a very large foetus of about 8 g. and they did tremendous work in measuring the different endocrine glands, and where these glands were larger and so forth, they tried to explain that also by growth hormone secretion or pathological aspects, for instance in a diabetic mother. But I think that they were always dealing with older embryos. Huggett: At the end of the war, Dr. Deryk Fraser and I were interested in foetal growth and the relationship between growth hormone and diabetogenic hormone and the effect on growth gonadally, so we injected growth hormone, which we got from Prof. Frank Young, and we found, as Prof. Tuchmann-Duplessis has found, that the foetuses were smaller. But we also found, as he did, that if you injected not at the beginning of conception but after 5 days, it had no effect, and that gave us the clue and we worked out that in fact what we were dealing with was delayed implantation due to the presence of impurity. We proved this because later Prof. Young produced a very pure growth hormone, completely free of gonadotropin, and we found that it was completely inactive and had no effect whatsoever. Like Prof. Jost, we came to the same conclusions on reading some of the published papers— that our conclusions differed from the authors'. Strauss: I would like to know where the growth hormone came from? Is it a pure hormone or a complex one? T.-Duplessis: We tried two types of growth hormone, a crystalline one prepared by Li in California and a commercial brand (Choay Laboratories) in which there may have been gonadotrophins. I think we have to bear in mind that these hormones only modify the environ- ment of the mother — they do not cross the placental barrier. It was demonstrated by many people that pituitary hormone, which, like ACTH, has a lower molecular weight does not cross the placental barrier. Strauss: But if growth hormone does contain gonadotrophic hormone, then gonadotrophic hormones may be produced by the placenta itself. T.-Duplessis: But the results were the same whether you have a preparation which you are quite sure does not contain gonadotrophin or one which does. It is only the mortality or abortion rate which is higher if you have a mixture of both. This is very significant in the experiments reported by Nixon. Huggett: Swyer and Fouracre Barns also investigated this same problem using pure growth hormone, and got the same result as we got, namely the entire negative effects. Williams: If growth hormone only acts if it is given before the 5th day, you can conceive that this is the period when pituitary luteotrophin is active and that by inhibiting production of this you delay implantation. If you are giving very high doses of growth hormone you may depress Discussion 175 the pituitary even if the preparation contains no gonadotrophin and it then comes down to what sort of dose you are giving. Are you giving the dose that has an effect on an adult or on a growing young rat: is your dose level higher than the physiological level? If so, I think it might be possible to depress pituitary function and delay implantation even if there were no gonadotrophin present. Huggett: These all cause tremendous growth in the mother. Williams: So that you are certainly giving excessive amounts? Huggett: We have given doses that cause the mother to grow but what we did not get was any sign of it passing across the placenta or causing the foetus to grow. Williams: You never got any delayed implantation with the pure compound however much you gave? Huggett: No, that is what Swyer found also. Montagna: How certain is it whether the implantation was, in fact, delayed or whether this was something else ? Huggett: It is not certain beyond the fact that the sizes of the foetuses fitted delayed implantation and that they appear to be entirely normal if you looked on it as being a 15-day foetus when it ought to be 19. Montagna : I asked that question largely because I have been waiting for two days for a chance to bring in skin and finally I am able to do it! If one injects rats with large amounts of ACTH or cortisone there is no effect at all on hair growth if the hair follicles had actually started to grow at the time of treatment. If one starts the injections before the hair follicle itself begins to grow no growth occurs in spite of the applica- tion of stimuli which would normally initiate growth. This is the reason I asked if we are sure that it is due to delayed implantation and not to some other condition. Huggett: That is quite a valid point. Jost: I should like to recall how complicated the problem of delayed implantation is. Canivenc and Mayer recently observed that burning a limb of a pregnant rat will suffice to induce implantation for a long period. Williams: Is it not true that in this burning experiment if you take the adrenal out, there is no delay? Jost: I think so, but I am afraid of making the wrong statement. Amoroso: It has been suggested that if the adrenals are removed you will get the same results. Jost: As far as cortisone is concerned, it is very curious to see to what extent the foetus is protected against cortisone. It is possible to inject into a rat foetus, weighing about 500 mg., 3 mg. of cortisone, and the foetus still develops. If the young newborn is injected with only a small fraction of this dose, it dies very quickly. I wonder whether in Prof. Tuchmann-Duplessis' experiments cortisone did not act chiefly after birth. T.-Duplessis: I do not know why the embryo is protected, though I do not think it is as well protected as it would appear to be, because the normal embryonic development is a very autonomous one, and there is a 50-70 per cent fragility amongst cortisone-treated animals. None of them lived longer than 10 to 20 days. So even if we do not see a morpho- logical modification, they are still very sensitive to cortisone. THE GROWTH CYCLE OF DEER ANTLERS George B. Wislocki Department of Anatomy, Harvard Medical School, Boston, Massachusetts The antlers of deer are deciduous, being annually renewed. In the Virginia deer (Odocoileus virginianus) in its natural habitat, they begin to grow visibly in April or May, reach maturity in August and the velvet is shed in September. The mature antlers, consisting of bare, dead bone, remain in place, firmly attached until mid-winter when they are shed. Antlers grow by the addition of new material at the ex- tremity of the beam and at the tips of the tines as they arise. The growing tip moves away from the antler pedicle, deposit- ing a column of bone which, once laid down, does not increase appreciably in diameter as it matures. All stages of growth and maturation can be seen in a single growing antler; the bone nearest the antler base is the oldest and most mature, whereas that at the tips is most recent and in process of formation. Microscopically, the mode of bone differentiation at the tips of the growing antlers is observed to be intermediate, in a number of important respects, between intramembranous and endochondral ossification (Wislocki, Weatherford and Singer, 1947). In the growing tips beneath the velvet, germinal, preosseous and osseous zones are visible. The germinal zone located immediately beneath the velvet consists of large fusi- form and stellate cells with extremely basophilic cytoplasm and showing numerous mitoses. The cells of the preosseous zone lie in lacunae contained within a matrix. The matrix consists of a dense meshwork of reticular, collagenous fibres and of a ground substance; the latter is metachromatic following staining with toluidine blue and gives an alkaline phosphatase reaction. Each lacuna is enclosed by a capsule which is 176 The Growth Cycle of Deer Antlers 177 strongly basophilic when stained with methylene blue and in- tensely metachromatic with toluidine blue, and gives an intense reaction for alkaline phosphatase. The basophilia and meta- chromasia of the lacunar capsules bespeak the presence of a sulphated acid mucopolysaccharide and a resemblance to cartilage. However, on other grounds, including the presence of a dense meshwork of collagen, the arrangement of the cells and the pattern of the blood supply, the preosseous tissue differs markedly from ordinary hyaline cartilage and endo- chondral bone formation as seen in the growing epiphyseal regions of long bones. Unlike endochondral bone formation where the cartilage cells perish, the preosseous corpuscles of the antlers appear to become directly converted into osteocytes. Throughout their entire period of growth, the antlers are supplied with blood by the arteries of the velvet (Waldo, Wislocki and Fawcett, 1949). Venous drainage occurs at first internally through the pedicle (April, May), but by July this pathway has been completely supplanted by a series of efferent veins which are located in the velvet. Recurrent arterioles at the growing tips of the antlers connect the arteries of the velvet with the capillaries and venous channels of the antler shaft. A ligature tied around the antler base in mid-June produced no permanent interference with either the blood supply or growth, whereas ligation in July caused cessation of growth and death of the antler. These results indicate that in June there are enough internal vascular channels to by-pass the blood around the site of the ligature, whereas in July such internal channels are no longer available. Natural shedding of the velvet appears to be associated with ossification of the peripheral zone of the antler shaft, a process which restricts the venous return from the interior of the antler and obstructs the circulation. With respect to the mode of growth of antlers, in recent interesting experiments two fallow deer were injected with radioactive phosphorus (Bernhard, Brubacher, Hediger and Bruhin, 1953). In a stag with growing antlers, a radiogram showed a gradient of phosphorus deposition which was 178 George B. Wislocki maximal in the tines and declined toward the antler base. Another deer with mature, dead antlers which had shed the velvet, was given radioactive phosphorus. The radiogram showed no activity whatsoever in the antler, but some deposition of phosphorus in the pedicle, which ceased abruptly at the line of union with the antler. Alkaline phosphatase activity in Virginia deer as measured biochemically, shows a similar gradient from the antler tips to base (Aub, 1940). The antlers are well innervated by branches of the trigeminal nerves (Wislocki and Singer, 1946). Numerous small nerves, located in the vascular layer of the velvet, extend outward to the growing antler tips. The dermis of the antlers is supplied with sensory endings for touch or pain. One antler was denervated experimentally in each of two deer just as growth had started. The denervated antlers grew, but became some- what dwarfed and deformed compared with the normal antlers; they lost their velvet in normal fashion and were ultimately shed. We attributed the somewhat irregular and retarded growth of the denervated antlers to traumatic injury attendant upon loss of their sensory receptors rather than to a loss of a direct trophic influence upon growth. The nerves supplying the antlers are probably collaterals which grow out from fibres normally innervating the antler pedicles. The fibres supplying the antlers are destroyed annually when the velvet is lost, but they regenerate seven months later upon renewal of the antlers. In the largest species of deer (elk, caribou), the rate of growth of the nerve fibres must exceed half an inch a day, establishing a record for the rate of growth of nerves. More recent observations indicate that, after the velvet is shed, the proximal portions of the nerve fascicles remain dor- mant over winter in the skin of the pedicles until they regene- rate in the following year (Wislocki and Waldo, 1953). In the dermis, close to the distal edge of the pedicle during the period of dormancy, sizable nerve fascicles are encountered, each of which is enclosed in a heavy sheath of epineurium. The suggestion is offered that these heavy sheaths protect and The Growth Cycle of Deer Antlers 179 shelter the nerve fascicles during the long quiescent period. The protection afforded explains, perhaps, why neuromas are not formed during the extensive period of latency. Besides the annual cycle of the antlers, male deer exhibit seasonal changes in their gonads and accessory reproductive organs. Thus Virginia deer show seasonal differences in the size and histological picture of the testes, epididymides and seminal vesicles (Wislocki, 1943, 1949). Spermatogenesis begins early in July, becomes maximal in October, declines in December and January and is in complete abeyance from then until July. Similarly, the interstitial cells of the testes are more active in the fall than in the spring, as gauged by various cytological reactions for lipids. The seminal vesicles also reach a peak of seasonal activity in the fall as judged by size and secretory activity. The annual growth cycle of the antlers is controlled by endocrine factors. The effects of castration, reported by numerous previous investigators upon various species of deer, have shown quite conclusively that (1) castration in the first eight months of life results in complete suppression of antler growrth, (2) castration after the appear- ance of the first set of antlers, if they are in the velvet, results in the permanent retention of the velvet and failure of the antlers to be shed, and (3) castration in the presence of antlers which have lost their velvet, results in immediate shedding of the antlers and their subsequent renewal and permanent retention in the velvet. These results rest upon the observa- tions and experiments of Caton (1877), Nitsche (1898), Rorig (1899, 1907), Tandler (1910), Tandler and Grosz (1913) and Zawradowsky (1926), to mention a few. The present investigator and his associates have confirmed these previous results and have extended the endocrine ana- lysis of the factors controlling antler growth by the adminis- tration of testosterone to a series of Virginia deer under various experimental conditions (Wislocki, Aub and Waldo, 1947; Aub, Wislocki and Waldo, 1950; Waldo and Wislocki, 1951). An airgun was devised with which slugs of testosterone pro- pionate and other similar materials could be shot into deer. 180 George B. Wislocki The following results were obtained : (1) Administration of testosterone to bucks which had never had antlers, as a result of castration as fawns, induced antler growth. (2) Administration of testosterone to bucks, which had been castrated as yearlings or later and which bore permanent antlers in the velvet, induced prompt shedding of the velvet and subsequent shedding of the antlers (tendency to restore the antler cycle). (3a) Administration of testosterone to normal bucks during the period of antler growth induced shedding of the velvet. (3b) Its administration to normal bucks, subsequent to the shedding of the velvet, inhibited the normal casting off of the antlers and led to their retention for many months. (4) Administration of testosterone to ovariectomized female deer caused the growth of well-developed antlers (masculinization). The information gained about the seasonal changes in the antlers and gonads of Virginia deer, combined with the observed effects of castration and administration of testo- sterone upon antler growth, led to the conclusion that the phases of the seasonal antler cycle in Virginia deer are con- trolled, on the one hand, by testosterone and, on the other hand, by an "antler-growth stimulus", the latter probably of anterior pituitary origin (Wislocki, Aub and Waldo, 1947; Waldo and Wislocki, 1951). The effects of castration of adult deer indicate the existence of a non-gonadal (hypophyseal) factor responsible for antler growth and of a testicular factor mainly responsible for the secondary hardening (internal reorganization) of the antlers and loss of the velvet. Mating takes place in the autumn, during a time when the testosterone level is high and the antlers consist of bare, dead bone. Afterwards the testosterone level declines, whereas the antler- growth stimulus reappears. In the opinion of these investi- gators (Waldo and Wislocki, 1951 ; Wislocki and Waldo, 1953), based upon histological studies and observations on living deer, shedding of the antlers is due to the reappearance of The Growth Cycle of Deer Antlers 181 the growth stimulus. Thus, growth rather than death of the tissues at the antler-pedicle junction represents the major factor causing the antlers to be shed. The observed proliferation of fibrocellular connective tissue at the antler-pedicle junction which accompanies the resorption of the osseous connections appears to be primarily responsible for the separation of the antlers from their pedicles. Furthermore, the hypothesis was advanced that the proliferation of the dermis surrounding the antler base exerts an upward thrust upon the underside of the flange-like osseous burr, causing the live tissues in the pedicle to disengage from the dead antler, whereupon minor trauma and mere weight of the antlers hasten their final separation. In a less well-documented study, Gruber (1937) advanced a somewhat similar concept of the process of antler-shedding in the roe deer (Capreolus capreolus) and the European stag (Cervus elaphus); nevertheless, Bruhin, as late as 1953, still adheres to the older view that shedding occurs during a period of quiescence or latency. The relation of the phases of the antler cycle to seasonal changes in hormone levels in Virginia deer, as elucidated by the studies cited, are illustrated diagrammatically in Fig. 1. In 1943, Wislocki pointed out that the growth phase of the antler cycle is initiated and completed in the spring and summer during a period of increasing and maximal daylight, whereas the shedding of the velvet and rutting (increasing gonadal activity) occur in the autumn during a period of diminishing daylight. Experiments were planned at that time to test the possible role of light in regulating the antler cycle, but these were subsequently never carried out. Meanwhile, a single experiment on a red deer (Cervus elaphus) has been reported by Jaczewski (1952). A stag in which antler growth commenced early in March, was confined daily, beginning on April 1, in a dark stall, from 4:00 p.m. to 8:00 a.m. On June 11, while still confined, the velvet was shed and the stag became aggressive. He was thereupon released and placed out of doors, and on July 12, shed his antlers. New antlers began immediately to grow and by mid- August had developed 182 George B. Wislocki two points. Although Jaczewski placed a different interpre- tation on this sequence of events, the present writer suggests that the reduced daylight from April 1 until June 11, caused both premature shedding of the velvet and rutting, events that would normally have occurred with diminishing daylight SLOW GROWTH tPreporatiom FilSnglnft for > Rounding xShedding / of Pedicle ACTIVE GROWTH NO GROWTH JSLOWGR Bare Antler Composed of J Preparation Dead Bone Firmly ! for Attached ^Shedding/ to its Pedicle \ / L 1 I I JAN. FEB. MAR. APR. MAY JUN JULY AUG. SEPT. OCT. NOV. DEC. JAN. FEB. Fig. 1. Diagrams relating the phases of the antler cycle to seasonal changes in hormone levels in male Virginia deer kept in captivity. (A) divides the cycle into three major phases consisting of periods of slow growth, active growth and inhibition of growth. The phase of slow growth is subdivided to indicate the changes which occur prior to shedding of the antlers (preparation for shedding) and the growth which takes place following shedding and preceding the phase of grossly visible active growth. (B) represents the relative physical size of the antlers during each phase of the cycle. The numbers 1 , 2 and 3 indicate the principal events of the cycle as designated below in (C). The thickness of the black bar indicates relative size of the antlers, the serrated upper border the presence of the velvet, and the shaded portion the occurrence of changes at the antler base in preparation for shedding. (D) illustrates our concept of the seasonal variations of the "antler-growth stimulus" (for normal males and castrates) and of testosterone. (From the original of Fig. 4, Wislocki and Waldo, 1953) in the late summer and autumn. Moreover, upon return of the animal to full daylight, growth was again stimulated, with loss of the antlers (July 12) and the immediate appearance of a new set of antlers. The complete sequence of events in this stag indicates the stimulating effect of withdrawal of light The Growth Cycle of Deer Antlers 183 upon differentiation of the antlers and gonadal function (pre-mature shedding of the velvet and rut) and the growth- stimulating effect of increase of light by the subsequent resumption of antler growth. REFERENCES Aub, J. G. (1940). Unpublished communication. Aub, J. G., Wislocki, G. B., and Waldo, C. M. (1950). New York State Conservationist. Reprint No. 98. Bernhard, K., Brubacher, Z., Hediger, H., and Bruhin, H. (1953). Experientia, 9, 138. Caton, J. D. (1877). The Antelope and Deer of America. Cambridge, Mass.: Hurd and Houghton. Gruber, G. B. (1937). Ges. Wiss. Nachr. Biol., 3, 9. Jaczewski, Z. (1952). Acta physiol. polon., Prace III, 201. Nitsche, H. (1896). Studien iiber Hirsche. Leipzig. Rorig, A. (1899). Arch. EntwMech. Org., 8, 382. Rorig, A. (1907). Arch. EntwMech. Org., 23, 1. Tandler, J. (1910). Anz. Akad. Wiss., Wien, 47, 252. Tandler, J., and Grosz, S. (1913). Die biologischen Grundlagen der sekiindaren Geschlechtscharaktere. Berlin: Springer Waldo, C. M., and Wislocki, G. B. (1951). Amer. J. Anat., 88, 351. Waldo, C. M., Wislocki, G. B., and Fawcett, D. W. (1949). Amer. J. Anat., 84, 27. Wislocki, G. B. (1943). Essays in Biology, p. 631. Univ. of California Press. Wislocki, G. B. (1949). Endocrinology, 44, 167. Wislocki, G. B., Aub, J. G., and Waldo, C. M. (1947). Endocrinology, 40, 202. Wislocki, G. B., and Singer, M. (1946). J. comp. Neurol., 85, 1. Wislocki, G. B., and Waldo, C. M. (1953). Anat. Rec, 117, 353. Wislocki, G. B., Weatherford, H. L., and Singer, M. (1947). Anat. Rec, 99, 265. Zawadowsky, M. M. (1926). Trans, exper. Biol. Zoopark Moscou, 1, 18. DISCUSSION Bourliere: Prof. Wislocki, what is the influence of the age of the animal on the antler cycle? Is the cycle modified in any way in the very old animals? Wislocki: Senescent changes do occur in deer antlers as revealed by observations on the antlers of old deer and, particularly, by several series of shed antlers which have been collected annually from captive deer. These series reveal a progressive annual increase in size and weight of the antlers up to 8 or 9 years of age, after which there is a rapid 184 Discussion diminution in antler size. A picture of such a series is presented by H. E. Anthony (1929). In view of the concept I have outlined of the control of the antler cycle, I attribute the decline of the antlers in senescent deer to a weakening of the hormonal regulation by the gonads and anterior pituitary. Besides becoming extremely small with advanced age, the antlers are frequently deformed and permanently covered by velvet, and are no longer shed (cf. Rorig, A. 1899; 1907). These features also point to a slackening of the hormonal regulation. The permanent retention of the antlers in velvet coincides with the effects upon antlers produced by removal of the gonads of adult deer. On the other hand, the diminution in their size points to decline in the pituitary factor. Zuckerman: Is there another factor involved — the factor of nutrition? Wislocki: Nutrition does play a vital role as a factor in the growth of antlers with respect, particularly, to their size and weight. In a privately kept herd of red deer, in Czechoslovakia, supplemental feeding of the herd with various foods rich in calcium, phosphorus and vitamins, induced the growth of antlers of incredible size (Vogt, 1937). Deer ordinarily consume grasses, leaves, shrubs, lichen, berries and other plants from which they usually obtain sufficient minerals. If their needs are not fully met, they may adopt peculiar habits. Murie (1935) describes the phenomenon in late winter of Alaska-Yukon caribou eating the shed antlers. Individual deer were also reported as eating the velvet and gnawing at one another's antlers just before shedding. Zuckerman: May I just add one further fact to that story? Duck- worth, at the Rowett Research Institute i n Scotland, has shown that major osseous changes occur in the pregnant ewe, including thinning of the bones and complete loosening of the teeth. By the time the animals are ready to lamb under the very hard winter conditions that prevail in Scotland, their jaws do not fit and there may be complete mal- occlusion. When the grass-growing season starts again the tooth again becomes held in the alveolus. Wislocki: With respect to the phenomenon of shedding of the antlers, Waldo and I (Waldo and Wislocki, 1951 ; Wislocki and Waldo, 1953) proposed a new concept of the process of shedding, which recognized growth and proliferation, rather than death, of the tissues at the antler- pedical junction as the major factor responsible for the loosening and shedding of the antlers. We attributed the proliferative changes, accompanied by local decalcification, to a hormonal "antler-growth" stimulus, probably of pituitary origin. Thus, the shedding of the antlers does not appear to be a regressive phenomenon dependent solely on poor nutrition, as would seem to be true of the changes in the teeth of pregnant ewes. However, in connection with pregnancy, it should be recalled that in reindeer, in which the females also have antlers, these are shed just after the termination of pregnancy. This suggests the possible role, at that period, of hormonal changes in the induction of antler-shedding. Nevertheless, it is also probable that metabolic changes in calcium and phosphorus should have occurred during the course of gestation. Discussion 185 In this connection, reference to hermaphrodite deer may be of interest. Hermaphrodites are very prevalent in Capreolus, the European roe deer (Rorig, 1899; 1907) and in Odocoileus, the white-tail and mule deer of North America (Wislocki, 1954; 1956). Besides occasional true hermaphrodites and pseudohermaphrodites, which invariably possess fairly well-developed antlers associated with minute, atrophic, intra- abdominal testes or ovotestes, there is a relatively large number of deer with normal female genital tracts, which have nothing abnormal about them except that they possess small, deformed antlers which are often permanently in velvet. These animals usually appear to have been pregnant and they are often lactating. I should infer from this that, during the course of pregnancy, either the ovaries and /or the adrenals had produced ketosteroids capable of stimulating the formation of rudimentary antlers. The possibilities that such antlers may be induced by the activity of progesterone or an androgen have been discussed (Wislocki, 1954, with literature). Strauss: Prof. Wislocki, you said that the antler bone is softer than the normal bone, does it depend upon the phosphorus content of the antler bone ? Wislocki: It is well known to naturalists that the antlers of deer are much softer than their skeletons. This is probably due to the fact that the bone is rapidly laid down and does not undergo the repeated internal reorganization which occurs in the skeleton. In dead deer, in the open, the antlers weather much faster than the skeleton and they are the first bones to be gnawed away and destroyed by small animals. Strauss: You showed us nearly-closed arteries at the casting off of the antlers, are these "polster" arteries? Wislocki: Well, as to whether the arteries of the velvet of growing antlers resemble what, in German, are called "Polsterarterien", I should reply no. The arteries of the velvet resemble structurally most nearly, I believe, the umbilical arteries. They are thick-walled, cylindrical and capable of constriction so as nearly to obliterate the lumen. The largest arteries appear to have a broad inner, circular coat and an outer, longitudinal one, but in the smaller vessels, the coats are not so sharply defined. There is no inner or outer elastic membrane. Instead, elastic fibres and argyrophilic reticular fibres interweave in a dense network around the smooth muscle cells (cf. Wislocki and Singer, 1946, Plate 2). Boyd: I would like to ask about the method of denervation in these antlers. It seems to me a little curious that it was only sensory fibres that grew into the antlers and I wondered if it was really established that the regenerated nerves included only sensory fibres. Wislocki: I don't know too much about the nerve fibres. I stimulated the velvet of a tame, white-tail deer (Odocoileus virginianus) with a needle and an electric current and learned that the antlers were sensitive to these stimuli. I assumed that these responses were related to the nerve fascicles observed in histological sections of the velvet (Wislocki and Singer, 1946). A recent investigator (Vacek, 1955) supplies more information on the innervation of the stag (Cervus elaphus) and the fallow buck (Dama dama). According to him, sebaceous and apocrine 186 Discussion glands in the velvet were innervated by both myelinated and un- myelinated fibres. Nerve ends were also abundantly demonstrable in the corium, epidermis and on hair follicles. The arteries received medullated and unmedullated fibres with terminations localized in the adventitial and medial coats. Previously, I had assumed that the arteries of the velvet, because of their similarity in structure to umbilical arteries, might be devoid of any innervation. We also denervated an antler in each of two white-tail deer (Wislocki and Singer, 1946), in late spring, and followed the subsequent antler development. The denervated antlers showed no changes attributable to possible loss of trophic neural influences. The denervated antlers grew to practically the same size as the antler on the opposite, normal side. Some deformation present in the denervated antlers seemed to be due entirely to repeated bruises and several fractures which they sustained because of the loss of sensation in the velvet. Boyd: There were no superior cervical sympathectomies — chronic ones? Wislocki: No, we did nothing of that sort. Huggett: You did not test them experimentally by stimulating to see how the vascular plexuses proliferate ? Wislocki: No. We did not do that. Matthews: One or two points: first of all, if the initiating of the growth of the antlers is due to a growth-stimulus substance, how is it that if you have a deer that has been castrated as a fawn and has grown no antlers at all, and then you give it a shot of testosterone, the antlers start growing? And furthermore, there is the other case of the male deer which never grow antlers at all. They are fairly common in Scotland and are known as hummels. They are not castrates ; there is nothing wrong with them except that they just do not grow any antlers. Wislocki: In regard to those questions, I can merely say that it is a strange fact that a fawn, if castrated, grows no antlers at all. However, if it is given later a daily single shot of testosterone propionate, this pro- cedure will trigger a growth-promoting influence which initiates antler growth. Why this is so, I do not know, but the pituitary growth- promoting factor apparently remains dormant until it receives an initial signal, either, normally, from the developing testis, or, in the castrated deer, by administration of a single dose of testosterone propionate. Without this signal, a castrated fawn will remain, forever, without any antlers. With regard to so-called "hummels", that is, male deer sexually active but bearing no antlers, I have no first-hand knowledge. Perhaps the absence of antlers in these deer is a genetic deficit rather than a hormonally- induced condition. It would be interesting to attempt to stimulate antler growth in such animals by the administration of hormones. Dempsey: Could it be the pituitary or could it be the end-organ that is triggered? Wislocki: Well, there is always the possibility of the end-organ. Wxe have considered that but we have so little evidence that they ever got the direct effect on the end-organ. Discussion 187 Zuckerman: Is it possible that you can get other effects in the end- organ? Dr. Harrison Matthews has just cited the hummel, which never grows any antlers at all. Experienced stalkers also believe that, if the sensitive velvet is bruised in one season, and antler growth is deformed, the animal will never grow points again in the corresponding part of his antlers. I do not know whether Dr. Harrison Matthews can confirm that? Matthews: Yes, that is so but is it a very close observation? The fact is that the animal has a deformity in the same place every year, but it is merely guesswork to say that it was caused in the first instance by damaging. Zuckerman: That is certainly true. Matthews: Of course, there is another wonderful yarn that goes around among deerstalkers and that is, if an animal gets accidentally uni- laterally castrated the horn on the opposite side is always deformed or even absent. REFERENCES Anthony, H. E. (1929). Bull. N.Y. Zool. Soc., 32, 3. Murie, Glaus, J. (1935). Alaska-Yukon Caribou, North America Fauna No. 54. United States Department of Agriculture, Bureau of Biological Survey, Washington, D.C. Rorig, A. (1899). Arch. EntwMech. Org., 8, 382. Rorig, A. (1907). Arch. EntwMech. Org., 24, 1. Vacek, Z. (1955). Csl. MorfoL, III, 249. Vogt, Franz. (1937). Neue Wege der Hege. Neudamm: Neumann. Waldo, C. M., and Wislockt, G. B. (1951). Amer. J. Aunt.. 88, 351. Wislocki, G. B. (1954). J. Mammal, 35, 486. Wislocki, G. B. (1956). J. Mammal., 37, in press. Wislocki, G. B., and Singer, M. (1946). J. camp. Neurol., 85, 1. Wislocki, G. B., and Waldo, C. M. (1953). Anat. Rec, 117, 353. AGEING OF THE AXILLARY APOCRINE SWEAT GLANDS IN THE HUMAN FEMALE William Montagna Department of Biology, Brown University, Providence, Rhode Island This work is based on a large number of biopsy specimens collected with a high-speed rotary biopsy punch, 5 or 10 mm. in diameter, from the axillae of normal volunteer subjects. For the study of normal adult glands, specimens were obtained from women 21 to 36 years old; two biopsy specimens were excised at weekly intervals for four weeks from each of the subjects. Fluctuations, if any occurred, in the morphology of the glands during the menstrual cycle could thus be fol- lowed. Specimens were taken at monthly intervals from several pregnant women and for two months after parturition. Ageing changes were studied in a series of specimens from subjects 41 to 78 years of age. Fewer but comparable speci- mens from men were also studied. Apocrine sweat glands occur in most mammals. In the Primates, the lemurs and the Platyrrhines have only apocrine glands, whereas the Catarrhines have both eccrine and apo- crine sweat glands. In the anthropoid apes, the chimpanzee has more eccrine than apocrine glands, but the orang has more apocrine glands (Schiefferdecker, 1922). In the human body apocrine sweat glands are found in the axilla, the mons pubis, the external auditory meatus, the circum-anal area, the areola and nipple of the breast and the labia minora of the female, and in the prepuce and scrotum of the male. Some glands may be found on the face and around the umbilicus. Negroes have more apocrine glands than do Caucasians, and the glands are more numerous in the female than in the male of both races (Homma, 1926). According to Schiefferdecker (1922), the human races can be segregated on the basis of 188 Female Axillary Apocrine Sweat Glands 189 abundance of apocrine glands; Europeans, having the fewest apocrine glands, belong to the highest order, but the Australian Negroes, with the most glands, belong to the lowest. Actually, the frequency and the distribution of apocrine glands have much the same variety in all races (Woollard, 1930). Apocrine axillary glands develop in the fifth foetal month as adventitious buds from hair follicles. At the end of foetal life they open into the funnel-shaped depression of the pilo- sebaceous canal; they are incompletely formed at birth and their characteristic properties develop slowly. Growth, final development and activity are closely related to sexual develo- ment and maturity. The glands attain full development at puberty, and they are said to undergo involution in senescence (Ito, Tsuchiya and Iwashige, 1951). Axillary apocrine glands are compactly coiled, tubular glands (Horn, 1935; Sperling, 1935), with adjacent loops joined by shunts or terminating in blind sacs. The deepest portion of the secretory coils extends into the subcutaneous fat. The single, or, rarely, doubled, duct of the axillary gland, imbedded in the loose connective tissue around the hair follicle, is perfectly straight. It runs near and parallel to the hair follicle and opens inside the pilosebaceous orifice. When ducts open directly onto the surface, the associated hair follicle may have degenerated. The secretory epithelium is composed of irregularly colum- nar cells (Fig. 1), with the terminal portion often elongated and projecting into the lumen. The free border of these cells terminates in very fine cytoplasmic processes which give the surface of the cells the appearance of a cuticular or brush border. Occasionally, segments of individual tubules, or entire tubules, are lined by simple cuboidal epithelium. When the tubules are excessively dilated, the epithelium is reduced to cells so flat that they resemble endothelial cells. The epithelial cells rest upon a mesh of myoepithelial cells aligned roughly parallel to the axis of the tubule. Outside of the myoepithelial cells is a thick, hyalin basement membrane. In order to understand senile changes it is necessary to 190 William Montagna describe first the histological features in normal adult glands. In discussing the ageing processes only a few of the more striking morphological features will be considered. The Glands of Young Adults The axilla contains rows of coiled apocrine glands, each associated with an axillary hair follicle. In histological sections each gland appears as a large nest of tubules cut at various planes (Figs. 1, 3). Connective tissue trabeculae separate adjacent nests; the connective tissue around the gland itself is very delicate. Although every section within each nest is a segment of the same coiled gland, the secretory epithelium of the different segments may be extremely different. In some segments, the epithelial cells are very tall ; in others they may be low cuboidal. Some tubules have a small lumen, others have a wide one. Pigment is abundant in the cells of some segments and nearly absent in those of others. These dif- ferences seem to record the differences in the state of activity of the various segments. Particularly in the axilla of women over 30, one or more segments may be distended and the epithelium reduced to squamous cells (Figs. 2, 3). In contrast to the lumen of normal tubules, which contains a clear colloid that does not stain with either basic or acid dyes, the lumen of dilated tubules contains a flocculent fluid which stains with basic dyes and often stains a strong metachro- matic colour with toluidine blue. The secretory epithelial cells of axillary glands usually contain pigmented granules, but the amount of pigment varies from one individual to another. All of the biopsy specimens removed from the same individual at weekly or monthly intervals show the same characteristic amount of pigment. In different individuals the pigment granules may be nume- rous or scant, barely visible or as large as the nucleus. Pig- mented granules are aggregated around a clear juxtanuclear region; the terminal cytoplasmic extension of each cell is always free of granules. Some cells may be full of pigment. Female Axillary Apocrine Sweat Glands 191 Pigment may be found in the tall cells or in the flat ones, and it may be absent from either of them. Neither type of pigment is affected by lipid solvents and both are relatively intact even in paraffin sections. Neither pigment is argyrophilic. The larger, brown granules stain with basic dyes; the small yellow ones are acidophilic. When stained with Altmann's aniline acid fuchsin-methyl green, the small granules, like the mitochondria, are fuchsinophilic ; the medium-sized granules stain green, and the large ones remain unstained. The larger pigmented granules are autofluorescent and emit a yellow to orange light, even in paraffin sections (Bommer, 1929; Montagna, Chase and Lobitz, 1953). The small yellow pigment granules are either nonfluorescent or emit a pale yellow light of very low intensity. The secretion fluid in the lumen of the tubules emits a very pale autofluorescence. The dark pigment contains some lipid and can usually be coloured with sudan black. The sudanophilia of these granules can be demonstrated in paraffin sections (Fig. 7). The yellow pigment is not sudanophilic. Many of the larger pigment masses, but not the small yellow pigment granules, are acid- fast and can be stained with carbolfuchsin and differentiated with hydrochloric acid. The larger pigment granules are mildly positive with the plasma! reaction, indicating the pres- ence of compounds containing aldehydes, ketones, acetals or compounds containing unsaturated groups (Bunting, Wislocki and Dempsey, 1948). The large pigment granules are reactive to the periodic acid-Schiff method. The reaction is not diminished by pre- vious treatment with diastase or saliva (Fig. 5). The epithelial cells of axillary glands characteristically contain ionic iron (Bunting, 1948; Bunting, Wislocki and Dempsey, 1948; Cavazzana, 1947; Homma, 1925; Homma, 1926; Iwashige, 1951; Montagna, Chase and Lobitz, 1953; Zorzoli, 1952). Homma reports that the glands in 70 per cent of the specimens obtained from cadavers and from surgery contained iron, and Manca (1934) found it in 26 out of 27 speci- mens. Some specimens abound in iron, but others may have 192 William Montagna traces or none at all. Iron is said to be normally more abun- dant in specimens from middle-aged subjects. The distribu- tion of iron is erratic; adjacent tubules in the same specimen or separate coils of the same glandular unit may show dif- ferent amounts of iron. Some cells may abound in iron whereas morphologically identical cells may possess none. The pres- ence or absence of iron is an individual characteristic. Iron is nearly always in the form of small granules ; large iron granules are very rare (Bunting, Wislocki and Dempsey, 1948; Homma, 1925; Homma, 1926). Iron may be present in the tall, apparently secretory cells, as well as in the low cuboidal ones. Regardless of the amount of iron in the epithelium, the residual secretion in the lumen of the tubules contains none (Montagna, Chase and Lobitz, 1953; Zorzoli, 1950), unless the epithelium is collapsing and undergoing obvious degenerative changes. Iron is found as a part of, or together with, the small yellow pigment granules; the large dark-brown granules rarely contain it. When they do, only a thin film at the periphery is reactive to the iron tests, and the brown pigment shows through unchanged. The amount of pigment in the cells is no indication of the presence of iron, but iron is found only in the small yellow granules and only if the large brown ones are also present in the same cells. This has led Manca (1934) to call the pigment which contains iron, haemosiderin, the one which does not, wear-and-tear pigment. However, the yellow pigment cannot be called haemosiderin since it may or may not contain iron. Perhaps an iron-containing substance may or may not be associated with the yellow pigment (Zorzoli, 1950). It is possible that whether or not iron is present, the yellow PLATE I Fig. 1. Segments of a normal apocrine sweat gland from the axilla of a 33- year-old woman. The height of the epithelium varies from one segment to another. On the extreme right are the coils of an eccrine gland. Stained with Heidenhain's haematoxylin. Fig. 2. Dilated segments of an apocrine gland from the axilla of a 41 -year-old woman. The epithelium is reduced to long squamous cells. Most of the colloid in the lumen has fallen ; bits which cling to the slick are stained meta- chromatically. Stained with toluidine blue. facing page 102 PLATE II. Fig. 3. Low power field of the glands from the axilla of a 41 -year-old woman. There are three nests of tubules, representing three glands. The nest in the upper part of the field is normal. The tubules in the nest in the centre of the field are lined by cuboidal cells and some segments show some degenerative changes. The tubules in the nest at the bottom of the field have a very flat epithelium. Stained with toluidine blue buffered to pH 5-0. Fig. 4. Low power field of the dilated tubules of the glands from the axilla of a 74-year-old woman. The epithelium in nearly every tubule was cuboidal or low columnar. Fig. 5. Secretory cells from the glands of a 26-year-old woman. Stained with the periodic acid-Schiff method showing Schiff-positive granules. Fig. 6. Perfectly normal cells from a gland in the axilla of a Gl -year-old woman. Stained with the periodic acid-Schiff method. Compare with Fig. 5. t * •m * . • 7 *• 4 4 I it •1 12 Female Axillary Apocrine Sweat Glands 193 pigment granules could develop into the brown ones. When they contain iron, the yellow pigment granules gradually lose iron, or perhaps the inorganic iron is gradually transformed into organic iron. However, there might be two yellow pig- ments, only one of which contains iron. Ionic iron is never found in the terminal cytoplasm of the cells. Not all axillary apocrine cells contain pigmented granules. In the glands of some individuals the secretory cells possess only traces of pigmented and lipoidal granules. In their places, the supranuclear cytoplasm has usually chromophobic and occasionally slightly basophilic secretion spherules. In some subjects these spherules are the only secretory elements present; in others they may be admixed with pigmented ones. With Heidenhain's haematoxylin the "chromophobic" spherules stain a clear blue-black. They are Schiff-reactive but resist digestion with diastase or saliva. The base of each apocrine cell is usually stippled with nume- rous fine basophilic granules, which often gives these cells a longitudinal striation. The cytoplasm lateral to and above the nucleus is weakly basophilic. In each cell a spherical region above the nucleus, which corresponds to the negative image of the Golgi apparatus, is free of basophilic granules. The apical and terminal cytoplasm of apocrine cells is almost PLATE III Fig. 7. Sudanophilic pigment granules in an axillary gland from a 28-year-old woman. Paraffin section coloured with Sudan black. Fig. 8. Normal distribution of sudanophilic pigment granules in an axillary gland from a 78-year-old woman. Compare with Fig. 7. Paraffin section coloured with Sudan black. Fig. 9. Iron in the axillary gland of a 78-year-old woman. The large pig- mented granules are unreactive to this test. Section treated with the Prussian blue method of Gomori. Fig. 10. Epithelial cells from a dilated gland from the axilla of a 72-year-old woman. The lumen is to the left of the cells. The granules in the cytoplasm are stained metachromatically with toluidine blue. Fig. 11. Flattened epithelial cells from an axillary gland of a 69-year-old woman. The cytoplasm contains glycogen granules. The flocculent content of the lumen is Schiff-reactive. Stained with the periodic acid-Schiff method. Fig. 12. Intensely stained colloid in an axillary gland from a 71 -year-old woman. Section treated with the periodic acid-Schiff method. ageing vol. 2 8 194 William Montagna chromophobic. The basophilic granules, as well as the nucleoli, are no longer stainable with basic dyes after the sections have been incubated in ribonuclease, and they must therefore con- tain ribonucleic acids. The intensity of basophilic staining in the cytoplasm of secretory cells is inversely proportional to the number of recognizable secretion granules present. Thus, nucleic acid may be involved in the synthesis of the secretion granules, pigmented or chromophobic. Normal axillary apocrine glands contain no glycogen (Montagna, Chase and Lobitz, 1953). However, Schiff- reactive substances other than glycogen are numerous in the apocrine cells and have already been described (Fig. 5). Spindle-shaped myoepithelial cells are couched between the single layer of columnar or cuboidal cells and the basement membrane. These cells are 5 to 10 (jl in the widest diameter and 50 to 100 (x in length. In transverse sections of tubules they appear as triangular cells between the bases of the secretory cells. The axes of the cells are oriented roughly parallel to that of the tubule. Myoepithelial cells are best developed in those tubules which are lined with the tallest epithelium. When tubules are dilated and the epithelial cells are flat and elong- ated, myoepithelial cells are very delicate or impossible to demonstrate. The Ageing Glands In the glands of women around 30 years of age, a few seg- ments of tubules become dilated and the secretory epithelium is reduced to cuboidal or squamous cells (Fig. 2). This is actually the onset of ageing changes. Between the ages of 34 and 44 there is a very gradual increase in the number of dilated segments in the entire axilla, and in a group of women 40 to 44 years old, entire glandular units have a flattened epithelium with or without an accompanying dilatation of the tubules (Fig. 3). The dilated tubules are lined with squamous cells. Although these atrophic changes can be found in nearly all specimens from the late thirties to the late forties, or up to Female Axillary Apocrine Sweat Glands 195 the period just before menopause, they are not numerous. The majority of the glands are normal and their secretory cells are tall columnar. In specimens from progressively older women more tubules are lined with a flat epithelium and a larger number of tubules is dilated. The lumen of the tubules lined with the atrophied epithelium contains a flocculent residue, in contrast to the lumen of normal tubules, the content of which is always clear. The most marked ageing changes in the axillary organ occur in the early fifties. At this period many entire glandular systems become atrophied. Atrophy and dilatation of the glands progresses very slowly from here on. We have not found a single specimen, regardless of age, in which there were not at least a few apparently normal and functional glands. This point is subject to great individual variation. In some individuals the glands become ineffective much more rapidly than in others. Many of the tubules in the axilla of a 78-year-old woman appeared normal. The assumption that the epithelium of these glands is normal is drawn from several criteria. The epithelial cells, although subject to variations, are usually tall columnar, with some cytoplasmic extensions projecting from their apices. The cells contain variable amounts of brown and yellow pig- ment granules. The small yellow granules often contain iron (Fig. 9), and the brown granules are usually sudanophilic (Fig. 8), acid fast, autofluorescent, and Schiff-reactive (Fig. 6). In contrast with the glands which remain intact, the epithe- lium of the glands which show atrophic changes has lost most of the characteristic features found in apocrine cells. Morpho- logically, the cells have lost the resemblance to secretory cells. They may be so flat as to resemble endothelial cells (Fig. 10). They rarely contain granules and pigment, lipid or iron. In contrast with normal glands, which never contain it, glycogen may be found in the epithelial cells (Fig. 11) and in the myoepithelial cells of the atrophied tubules. In addition to these granules of glycogen, the cells may also contain saliva- resistant, Schiff-reactive granules. These same granules also stain metachromatically with toluidine blue (Fig. 10). This 196 William Montagna combined staining property is identical with that of mucus. Also, the curdled luminal content of these glands stains like the granules just described. This, too, is like mucus, and is strikingly different from the serous content of normal glands which is clear and nearly achromic. As soon as the epithelial cells become flattened, the myo- epithelial cells become very thin and inconspicuous. In the tubules lined by squamous cells myoepithelial cells cannot be demonstrated. The presence of conspicuous myoepithelial cells, then, is a property of the functional glands. The duct is the most resistant structure of the axillary gland and it is never changed morphologically. All of its features are like those of normal mature glands, even in those subjects in which the glandular tubule is reduced to a series of thin- walled cysts. This point is of some interest. Perhaps in spite of the changes which the glands suffer with ageing, they still retain some secretory activity, although the type of secretion is considerably altered. Mitotic figures are not infrequent in the glandular epithe- lium of aged subjects. It must be recalled that mitosis is not common even in the glands of young women. In the glands of women 50 years of age or older mitotic figures may be found even in the flattened epithelial cells. If these are self -propagat- ing cells, then the changes that they have undergone are not degenerative. Axillary glands are inactive in children, become functional at puberty and remain active throughout the sexual life. In old age axillary odour seems to disappear, indicating a decrease and final cessation of apocrine secretion. In women the glands are believed to undergo periods of greater or lesser activity correlated with the menstrual cycle and with pregnancy. Authors generally agree that the glands are small and in- active during the intermenstruum and become swollen and active during the premenstruum and during the menstruum (Loeschcke, 1925; Schaffer, 1926). Cavazzana (1947) believes that during the premenstruum and menstruum the glands are actually larger than during the intermenstruum; their Female Axillary Apocrine Sweat Glands 197 lumen is dilated, and the secretory cells average taller. The myoepithelial cells are believed to be swollen and farther apart during the menstruum. During pregnancy the changes are like those which occur during the menstruum but they are much more pronounced (Talke, 1903; Waelsch, 1912). Contrary to all others a few authors (Cornbleet, 1952; Loes- chcke, 1925; Richter, 1932) believe that during pregnancy the axillary glands are in a resting state or that their function is depressed. Although these concepts are accepted unquestion- ingly, evidence to the contrary deserves some scrutiny (Klaar, 1926). Unlike the other authors, all of whom studied apocrine glands in autopsy material, Klaar (1926) obtained biopsy specimens from the same women at weekly intervals during the menstrual cycle. He found in normal women no changes in the axillary organ which correspond to the cyclic menstrual changes, and concluded that there is no correlation between functional activity and the diameter of the gland, and that the differences in the height of the secretory epithelial cells and their content of iron in no way corresponds to the menstrual cycle. During pregnancy and in puerperium the glands ex- hibit individual range of variation similar to that found in non-pregnant women. In women after menopause, the glan- dular epithelium exhibited a wide range of variations ; they all had some characteristic dilated or "cystic" tubules lined by flattened epithelial cells and surrounded by a thick connec- tive tissue capsule. Normal glands, however, could be found alongside of the cystic tubules even fifteen years after the onset of the menopause. In castrated women a general slow regres- sion of the apocrine glands is similar to that seen after the menopause. Our observations are in complete agreement with those of Klaar. We have been unable to correlate any changes in the axillary glands with the menstrual cycle. Similarly, no changes have been found in biopsy specimens removed at monthly intervals from pregnant women. Thus, there do not appear to be definite cyclic changes in the axillary apocrine 198 William Montagna glands which parallel those of the gonads. The contradictions between these reports and those of others may be the result of the use of improper and insufficient materials by others. It seems possible that most authors have lacked an appreciation of the full range of variations which occur in the glands even in the same individual. Although it is customary to believe that axillary apocrine glands have a cyclic secretory activity, we have not been able to detect cyclic changes in their morphology. Usually everyone assumes that the glands become func- tional at the onset of puberty. It cannot be denied that gonadal maturation largely coincides with the onset of strong axillary gland activity. Yet the glands of infants are clearly recognizable and larger than the eccrine glands. The glands of female children or juveniles which are not yet menstruating are often large and functional (Klaar, 1926). The glands, then, do not respond specifically to ovarian hormonal stimula- tion; this is also reflected in the ageing processes. Although castration and menopause bring on some ageing changes quickly, many entire glandular units remain intact and apparently normal for years after these events. In some of the glandular units, regressive or ageing changes take place in some coils of the tubule while others remain normal. Thus, ovarian hormones play an obvious role in initiating their maturation and maintaining in part the functional state of the axillary apocrine glands, but other hormones, perhaps from the adrenal cortex, are probably much more closely impli- cated in this. What makes this assumption more probable is the fact that in ageing men the glands show far less striking regressive changes than those of women. REFERENCES Bommer, S. (1929). Acta derm. -verier eol., Stockholm, 10, 253. Bunting, H. (1948). Anat. Rec, 101, 5. Bunting, H., Wislocki, G. B., and Dempsey, E. W. (1948). Anat. Rec, 100, 61. Cavazzana, P. (1947). Riv. ital. Ginec, 30, 114. Cornbleet, T. (1952). Arch. Dermal. Syph., Chicago, 65, 12. Female Axillary Apocrine Sweat Glands 199 Homma, H. (1925). Arch. Derm. Syph., Berlin, 148, 463. Homma, H. (1926). John Hopk. Hosp. Bull., 38, 365. Horn, G. (1935). Z. mikr.-anat. Forsch., 38, 318. Ito, T., Tsuchiya, K., and Iwashige, K. (1951). Arch. anal, jap., 2, 279. Iwashige, K. (1951). Arch. hist, jap., 2, 367. Klaar, J. (1926). Wien. klin. Wschr., 39, 127. Loeschcke, H. (1925). Virchows Arch., 255, 283. Manca, P. V. (1934). G. ital. Derm. Sif., 75, 187. Montagna, W., Chase, H. B., and Lobitz, W. C, Jr. (1953). Amer. J. Anat., 92, 451. Richter, W. (1932). Virchows Arch., 287, 277. Schaffer, J. (1926). Wien. klin. Wschr., 39, 1. Schiefferdecker, P. (1922). Zoologica, Stuttgart, 72, 1. Sperling, G. (1935). Z. mikr.-anat. Forsch., 38, 241. Talke, L. (1903). Arch. mikr. Anat., 61, 537. Waelsch, L. (1912). Arch. Derm. Syph., Wien, 114, 139. Woollard, H. H. (1930). J. Anat., Lond., 64, 415. Zorzoli, G. (1950). Boll. Soc. ital. Biol, sper., 26, 1. Zorzoli, G. (1952). Boll. Soc. ital. Biol, sper., 28, 1078. DISCUSSION Medawar: This is a most interesting paper. The problem is obviously a very complex one, because there seem to be two levels of the ageing process: there is the ageing in the individual gland, which is at least partly a function of the cell population in the gland, and there is also the ageing of the population of glands considered as a whole when the number of functional units is decreasing. I would like to ask Prof. Montagna if it is possible that these two processes are independent; and I also want to ask him if he has any evidence that these glands atrophy or deteriorate and are then replaced. I do not mean replaced de novo but replaced inside the old collagen framework. I do not know whether his serial biopsy technique would reveal such a process if it occurred; but if there were a progressive deterioration in the replacement power of these glands, that might at least account for the fall off in the number of functional units in the population. Montagna: It is evident from our material that some glands become aged more rapidly than others. I do not know whether or not the same gland can recover. Mitotic activity can occur in the secretory cells of aged tubules ; it is not abundant, but it is not abundant even in young glands. It is possible that there might be some replacement of lost segments. One point that should be made clear is that some glands do fold up completely; not very many of them and not as often as one would suspect, but some do perish. Harrison: You say you find no change or no evident change during the menstrual cycle. How about during pregnancy, have you any evidence of that? 200 Discussion Montagna: This is a controversial point. We collected specimens once a month from eight women from about the third month of pregnancy through parturition and lactation, until their menstrual cycle had been restored. We found nothing in the axillary glands that we could say correlates with these changes. The only other person who has collected repeated biopsy specimens from the same subjects over the menstrual cycle, is Klaar. We have confirmed his work and we are the only ones who agree with him. We are in disagreement with all others, who have relied entirely upon autopsy material. Strauss: Prof. Montagna's magnificent paper is a very important morphological contribution to the understanding of the life curve of the human sex appeal! He mentioned very briefly, but very importantly for the anatomist, that he does not like to call the glands "apocrine glands". I would like to ask him why not? Montagna: This is a matter of definition. In apocrine glands the head- portion of the secretory cells should pinch off and disintegrate to form the secretion. Whereas this takes place in some cells, most cells secrete by giving off little globules of secretion at the brush-border. I think that that is the way most of the secretion takes place. Strauss: You would call them eccrine glands? Montagna: No! Since one cannot combat terminology by creating new terms I stick with the old terms! Amoroso: An essentially similar view has been propounded by Richardson in respect of the secretion of milk. Huggett: I am interested in certain points. I gather that you have not been able to test these more mysterious portions on animals — you might get animal material. I was wondering what is the distribution of these glands in animals and whether you might not be able to satisfy — shall I say that sex appeal — by getting animals who are prepared to supply their skin? Montagna: Although we have done a great deal of work on other mammalian glands, they are not nearly so interesting. Most tubular glands in other mammals are fairly dull ; they secrete very deliberately, with the exception of those of the horse and the pig which secrete fairly profusely. The inguinal glands of the rabbit are large but they secrete only traces of a musky smelling substance. Huggett: This is a species story? Montagna: I am afraid it is. Huggett: And you have not got an animal species you can use as an experimental subject? Montagna: No. Even among the Primates, the lemurs and the Platyrrhines have only apocrine, whereas the Catarrhines have both eccrine and apocrine. The chimpanzee has more eccrine glands than apocrine, and the orang has very few eccrine glands. Huggett: To what extent is this a sex-distributed characteristic? You spoke of the male as having deliberate changes — now I do not know what the word "deliberate" means in that context. Montagna: The change is very much slower. I mentioned that in Discussion 201 women at menopause there seems to be a precipitation of these changes whereas there is no event in the male which is comparable to that. The changes are all very slow, gradual ones. Huggett: Well, you mentioned that in the girl the apocrine gland begins to show its differentiation about 8 or 9 years of age; does anything similar happen at that date in the boy? Montagna: Not as pronounced. These observations were all made with our noses! Most parents, and particularly the mothers, probably know this. This is an empirical observation and I have nothing to substantiate it. I was talking about this with Dr. Rothman and he made the delightful comment that dermatologists were not aware of this. Zuckerman: Among Primates you looked at, did you examine the pigtailed macaque? Montagna: I have seen very few primates other than the rhesus. Zuckerman: Then I suggest that that might be different — merely because of your remarks about smell. THE METABOLISM OF SENESCENT LEAVES E. W. Yemm Department of Botany, University of Bristol Introduction An important feature of the metabolism of leaves and other plant organs during their senescence is the predomin- ance of catabolic processes in which many of the complex constituents, such as proteins and polysaccharides, are broken down to simpler soluble products. Yellowing of the leaves, associated with the loss of chlorophyll, is an obvious external sympton of ageing which usually precedes abscission and leaf fall. The products of catabolism, consisting of amino acids, amides and other soluble metabolites, are to a large extent translocated to other parts of the plant, where they may be stored as reserve substances in seeds or other perennating structures. In the account which follows, chief attention will be given to the changes of tissue proteins and other nitrogenous constituents and their relation to the respiratory activities of senescent leaves. Within the cell, proteins are important components of the protoplasmic structure and their break- down during senescence is therefore of particular interest. Moreover, there is now much evidence which suggests that in plant tissues the proteins are maintained by continuous synthetic activity, which is closely coupled with respiratory oxidations. The Catabolism of Proteins in Senescent Leaves It is well established that yellowing in old leaves is accom- panied by a marked breakdown of cytoplasmic and chloro- plastic proteins of the cell. For example, Michael (1935) 202 The Metabolism of Senescent Leaves 203 studied the changes of chlorophyll and protein content in leaves of nasturtium (Tropaeolum majus) and demonstrated a close correlation between the amounts of chlorophyll and protein over a wide range of conditions. Most of the detailed information concerning the biochemical mechanism underlying protein catabolism in senescent leaves has been obtained with leaves detached from the plant. Under these conditions many of the changes associated with normal ageing are greatly accelerated, so that after relatively short periods of time yellowing and the breakdown of protein can readily be detected. Experiments with detached leaves of tobacco (Vickery, Pucher, Wakman and Leavenworth, 1937), various species of grass (Wood and Cruickshank, 1944) and barley (Yemm, 1937, 1950) have shown consistently that appreciable protein catabolism occurs within a few hours of detaching the leaf. Moreover, in old leaves the breakdown is relatively independent of the conditions under which the leaves are kept ; it occurs in the dark and in the light and is not prevented when sugars are artificially supplied to the tissues. Both cyto- plasmic and chloroplastic proteins are involved, although in some species there is evidence that the plastid proteins may be at first more rapidly attacked. Investigations of the pro- ducts of protein breakdown, which are formed in the leaves, strongly suggest that the catabolic changes extend beyond simple hydrolysis of the proteins. In barley leaves, for ex- ample, it has been found that the amount of the amide, glutamine, which accumulates in freshly detached leaves commonly exceeds the amount which could arise directly by proteolysis, thus indicating that amino acids are rapidly drawn into secondary changes. It has been further shown that the extent of these secondary changes increases with ageing of the leaf; much greater quantities of glutamine accumulate when old as distinct from young leaves are detached from the growing plant and treated under comparable conditions (Yemm, 1949). Many features of the metabolism of amides, and in particular glutamine, indicate that they play an important part in both the synthesis and breakdown of protein in plants. 204 E. W. Yemm The rapid formation of amides in ageing tissues may repre- sent a gradual alteration in the balance between synthesis and breakdown. There is now much evidence that the proteins and other nitrogenous constituents of plant tissues are maintained in a continuous dynamic equilibrium. The operation of inde- pendently controlled processes of synthesis and breakdown was earlier suggested by Mothes (1933) and by Gregory and Sen (1937), and the use of isotopic nitrogen (15N) has provided direct evidence of the dynamic state of leaf proteins. The experiments of Hevesy and others (1940) and of Vickery, Pucher, Schoenheimer and Rittenberg (1940) demonstrated the incorporation of the isotope into tissue proteins in con- siderably greater quantities than could be accounted for by primary synthesis. More decisive evidence has been obtained from the experiments of Chibnall (Chibnall and Wiltshire, 1954) with leaves of runner bean and from our own experi- ment with barley leaves. It has been shown that, even under conditions in which a net loss of protein occurs in detached and senescent leaves, there is appreciable synthetic activity indicated by the active incorporation of 15N into the tissue protein. Some of the results of an experiment, in which a dilute solution of ammonium phosphate, containing about 30 atom per cent excess of 15N, was supplied to the cut surface of detached barley leaves, are summarized in Table I. The protein separated from the leaves was subjected to group analysis by methods similar to those described by Yemm (1950) and the abundance of 15N determined in each group. Under the conditions of this experiment a net loss of about 20 per cent of the tissue protein occurred during the period of 24 hours over which the experiment extended. Neverthe- less, it is evident from the data of Table I that considerable synthetic activity continued in the leaves; a significant incorporation of isotopic nitrogen was detected in each of the main groups of amino acids into which the protein was analysed. The highest abundance of 15N was observed in the The Metabolism of Senescent Leaves 205 dicarboxylic amino acids and in the amide N, again emphasiz- ing the high activity of glutamic and aspartic acids and their amides in the protein metabolism of the leaves. It may be concluded from the evidence outlined above that the protein content of leaves is the result of continuous Table I Incorporation of 15N into Proteins of Barley Leaves Protein separated from leaves 24 hours after supplying them with 0-04 M-NH4H2P04 containing 29-3 Atom % excess 15N. Atom % Excess 15N Total N Dicarboxylic amino acid Basic amino acids Other amino acids Amide N 0-183 0-210 0-080 0 160 0-441 synthetic and catabolic activities. In ageing tissues the loss of protein is attributable to a decline in synthetic capacity, leading to an adverse balance and the predominance of catabolic changes as yellowing of the leaf proceeds. Respiratory Activities of Senescent Leaves The changes of respiratory activity in yellowing leaves have been recently reviewed by James (1953). Much of the available data with leaves of different species has been ob- tained with detached leaves often kept for long periods in the dark. As already noted, senescence occurs rapidly under these experimental conditions, but in the present context it is of interest to examine briefly the general relation between the rate of respiration and yellowing of the tissues. The pioneer research of F. F. Blackman (see James, 1953), carried out with leaves of Cherry Laurel and Tropaeolum, showed that during yellowing of mature leaves a marked increase in the rate of respiration took place. In leaves of different species 206 E. W. Yemm the time scale of these changes varies widely ; the rise in C02 production and rapid yellowing occurred after about 25 days in Cherry Laurel and after 8-10 days in Tropaeolum leaves. With mature leaves of barley and of grasses the changes are much more rapid and yellowing often occurs 2-3 days after detachment from the plant. Despite these wide differences in the rate of ageing, many of the same features can usually be recognised; during yellowing relatively high or rising rates of respiration are generally observed with detached leaves of different species. More limited data are available concerning the respiration of leaves when they undergo normal senes- cence while still attached to the plant. However, Arney (1947) in experiments with attached strawberry leaves has shown that an appreciable rise in rate of C02 output occurs during yellowing. A detailed interpretation of the respiratory drifts in ageing leaves will not be attempted here. There is strong evidence that the respiratory mechanisms are considerably modified; with detached leaves of barley and other species, chemical analyses indicates that during yellowing a substantial part of the C02 lost from the tissues is formed by the oxidation of the carbon skeletons of amino acids. Godwin and Bishop (1927) found that a loss of relatively stable glycosides and probably polysaccharides took place during the senescence of Cherry Laurel leaves. High rates of respiration seem, therefore, to be closely related to the breakdown of proteins and other complex constituents in senescent leaves. Soluble metabolites, such as amino acids formed in this way, may be important respiratory substrates which sustain the cellular oxidations associated with the rapid release of C02 from ageing tissues. In addition to leaves, some other senescent plant organs show an acceleration of respiration during ageing. The so- called climacteric rise, observed in apples when they ripen and become yellow (Kidd and West, 1930), has been shown to occur in many other ripening fruits. The metabolic changes which take place in apples have been extensively studied (see, for example, the review by Pearson and Robertson 1954); The Metabolism of Senescent Leaves 207 the high rates of respiration in the ripening fruits are associ- ated with the breakdown of starch to sugars, although a slight synthesis of protein can be detected during yellowing. The experimental data briefly considered above indicate that the catabolic processes in senescent plants are commonly associated with high rates of cellular respiration. It is appro- priate now to try and distinguish some of the basic changes in cellular metabolism which accompany ageing of plant tissues, and to bring the data into relation with some current views on the regulation of respiration and its coupling with synthetic activities. The Relation between Cell Respiration and Protein Metabolism in Ageing Tissues Gregory (1937), Richards (1936) and their collaborators have demonstrated a close connection between respiration and protein metabolism in leaves of barley. In a tentative hypo- thesis Gregory and Sen (1937) suggested that a continuous protein cycle operates in leaves and that under normal cir- cumstances much of the respiratory C02 arises from the oxida- tion of the carbon skeletons of amino acids. The hypothesis has been further developed by Steward and his collaborators (see, for example, Steward and Thompson, 1954), and, along these lines, the rapid respiration of senescent leaves may be related to the high levels of available substrate, resulting from the breakdown of tissue proteins to amino acids. There are, however, other cellular mechanisms which may link respiration and protein metabolism and which seem to offer a more comprehensive interpretation of the catabolic changes in ageing leaves. It is now widely recognised that phosphorylation and phosphate carrier systems constitute important mechanisms regulating glycolysis and respiration and coupling them with endergonic synthesis. Experimental work with both animal and plant tissues indicates that the rate of respiration may be strongly influenced by the 208 E. W. Yemm availability of free phosphate and/or phosphate acceptors, both of them essential components of the enzymic systems engaged in glycolysis and oxidation. The general importance of phos- phorylation in regulating cell respiration in plants has been discussed by Lardy (1952), by Millerd and Bonner (1953) and, with particular reference to protein metabolism, by Folkes and Yemm (1954) and Yemm (1954). Evidence that the level of phosphorylation may regulate the rate of respiration in plant cells rests mainly upon the action of so-called uncoupling agents, such as dinitrophenols. As reviewed by Simon (1953), these reagents at low concen- trations strongly inhibit phosphorylation, whereas enzymatic oxidations are but little affected: under suitable conditions synthetic activities, particularly protein synthesis, are pre- vented, while a marked increase in cellular respiration gener- ally occurs. Thus treatment of the tissues with nitrophenols and other uncoupling agents produces experimentally changes in metabolism similar to those associated with senescence. Observations such as these have led to the view that ageing in plant tissues is associated with a partial failure of the linkage between phosphorylation and cell oxidations. As a result of this failure, the regulating action on cell respiration is impaired and there is a decline in the synthetic activities whereby the proteins and other complex constituents of the cell are maintained. Pearson and Robertson (1954) have reviewed some evidence that changes, such as those outlined above, occur during the ripening of apples. They have shown that the relatively low rate of cell respiration during the pre-ripening stages could be substantially increased by 2 : 4-dinitrophenol, but that after the climacteric rise and yellowing similar treatment had little effect. Some further evidence that changes of the phos- phorylating mechanisms were associated with ripening was obtained by Biale (1954) from a study of cytoplasmic particles, probably mitochondria, prepared from the avocado fruit. But, as yet, no comparable data have been obtained concern- ing the metabolism of senescent leaves. The Metabolism of Senescent Leaves 209 Conclusions With regard to respiratory activities and the metabolism of proteins in ageing leaves the following conclusions may be drawn from the data : 1. The predominance of catabolic changes in the tissues results, in part, from a decline in synthetic activity, and not solely from an increase in hydrolytic breakdown of complex constituents. 2. High rates of respiration are generally found in yellow- ing leaves; much of the carbon dioxide may arise from the oxidation of the carbon skeletons of amino acids. 3. It is possible that both the high rate of respiration and the loss of synthetic activity result from changes in the phos- phorylating mechanisms of the cells. REFERENCES Arney, S. F. (1947). New Phytol., 46, 68. Biale, J. B. (1954). VHIe Congres International de Botanique (Section 11), p. 390. Paris. Chibnall, A. C, and Wiltshire, G. H. (1954). New Phytol., 53, 38. Folkes, B. F., and Yemm, E. W. (1954). Biochem. J., 57, 495. Godwin, H., and Bishop L. R. (1927). New Phytol., 26, 295. Gregory, F. G., and Sen, P. K. (1937). Ann. Bot., Lond., 51, 521. Hevesy, G., Linderstr0m-Lang, K., Keston, A. S., and Olsen, C. (1940). C. R. Lab. Carlsberg (Serie chim.), 23, 213. James, W. O. (1953). Plant Respiration. Oxford University Press. Kidd, F., and West, C. (1930). Proc. roy. Soc. B, 106, 93. Lardy, H. A. (1952). The Biology of Phosphorus. Michigan State College Press. Michael, G. (1935). Z. Bot., 29, 385. Millerd, A., and Bonner, J. (1953). J. Histochem. Cytochem., 1, 254. Mothes, K. (1933). Ber. dtsch. bot. Ges., 51, 31. Pearson, J. A., and Robertson, R. N. (1954). VHIe Congres Inter- national de Botanique (Section 11), p. 380. Paris. Richards, F. J., and Templeman, W. G. (1936). Ann. Bot., Lond., 50, 367. Simon, E. W. (1953). Biol. Rev., 28, 453. Steward, F. C, and Thompson, J. F. (1954). In The Proteins, vol. II, pt. A., p. 513, Ed. by H. Neurath and K. Bailey. New York: Academic Press Inc. VlCKERY, H. B., PUCHER, G. W., SCHOENHEIMER, R., and RlTTENBERG, D. (1940). J. biol. Chem., 135, 531. 210 E. W. Yemm Vickery, H. B., Pucher, G. W., Wakman, A. J., and Leavenworth, C. S. (1937). Bull. Conn, agric. Exp. Sta., No. 399. Wood, J. G., and Cruickshank, D. H. (1944). Aust. J. exp. Biol. med. Sci., 22, 111. Yemm, E. W. (1937). Proc. roy. Soc. B, 123, 243. Yemm, E. W. (1949). New Phytol, 48, 315. Yemm, E. W. (1950). Proc. roy. Soc. B, 136, 632. Yemm, E. W. (1954). Proceedings of the Seventh Symposium of the Colston Research Society, p. 51 . London: Butterworth Scientific Publications. DISCUSSION Bourliere: I would just like to ask Dr. Yemm a few questions. First ; what happens in permanent leaves of evergreen plants ; do you observe the same modification as in deciduous leaves but occurring at a slower rate ? And second ; are some ecological factors, for instance temperature, able to modify the growth and ageing rate of the leaves in any way ? Yemm: Now, the first point is the question of different kinds of leaf. Of the two examples which I gave you, one was an annual plant (Tropae- olum) and the other one was what is effectively a much longer-lived evergreen leaf, the Cherry Laurel. There is a general relationship between the length of time which the leaf normally lives on the plant and the rate at which the senescence changes occur. I think the Cherry Laurel leaf takes about twice as long to pass through the yellowing phase as the Tropaeolum. The second one: ecological factors and leaf senescence really are extremely important. One of the things which increases the rate of senescence of leaves very markedly in cereal plants, which I have been interested in particularly, is water supply. One gets much earlier yellowing and, of course, earlier ripening of the leaves under conditions of restricted water supply. And there are many other factors — the supply of nitrogen to the plant is an important one ; under conditions of limited nitrogen supply again the leaves tend to become senescent much earlier. Bourliere: And what about temperature? Yemm: Well, temperature is very difficult to distinguish from humidity and water supply because it simply increases the rate of water loss from the plant so the two factors interact. High temperatures and relatively deficient water supply are both factors which tend to increase the rate of senescence. Huggett: Does it happen at high temperatures with humidity? Yemm: It certainly does in apples. Temperature is a very important factor in the ripening and senescence of apples. I do not think this has been examined very carefully with leaves but the yellowing and ripening of apples is very temperature-dependent. Huggett: What about leaves on tropical plants in humid climates like Nigeria ? Discussion 211 Yemm: Most of these leaves are not deciduous. They have quite a long life and are commonly relatively resistant leaves of the type of Cherry Laurel. Montagna: The factor of light is also rather important. I made some years ago what I thought to be a very perspicacious observation when I noticed that when in the autumn leaves of maple trees become yellow, all the leaves which face towards street lights remain green! When I confided this to my botanical friends they told me that everybody knows that! At any rate, in these deciduous plants, the leaves near the light remain much longer than those away from the light. Yemm: The evidence from more exact studies under controlled conditions is a little conflicting! Certainly with old leaves you get very little retardation of protein breakdown by keeping them in the light, compared with in the dark. BourlUre: Has any study been made on the ageing of seeds? What happens, for instance, in seeds which retain their germinative power for a very long time ? Yemm: There are certainly gradual changes in seeds. The longest- lived seeds, which may survive for something like 200 years, do, never- theless, gradually lose their ability to germinate. And seeds which are not quite so resistant have been shown to have quite appreciable res- piratory activities; they are not entirely dormant structures, and there are biochemical changes in certain cell constituents. One of the quite interesting suggestions which has recently been made is that there is a gradual oxidation of glutathione and possibly other sulphydryl com- pounds; when this has reached a certain point then the seed may no longer possess the capacity to germinate. Krohn: I wonder if Dr. Yemm would see any similarity in the life of the antlers that we have heard about this morning and the life of his leaves. Do they fall off in the same sort of way? What I particularly wondered was whether the petiole, which joins the leaf on to the rest of the plant, atrophies there and shrinks up first and then the changes in the leaf are secondary to that ? Yemm: No. The process of leaf abscission is quite an active one. Krohn: As in the antler? Yemm: Yes. It involves the production of an abscission tissue. Cell division begins in the petiole of the leaf and leads to the formation of a layer of abscission cells which ultimately cut off the leaf from the rest of the plant. But, of course, this is not an essential feature of senescence because in many annual and other plants the leaf undergoes senescence while still attached to the plant. Krohn: The other thing I wanted to ask was, do you think that in- fections with virus diseases and so forth, which are clearly persistent in the leaves, cause an accelerated ageing of the leaf? They obviously make them go yellow. Yemm: Well, I think that is a part of the same general process. Virus infection, generally speaking, probably leads to a diversion of protein synthesis from the synthesis of normal cytoplasmic elements in the cell into the special viral elements. In other words, protein synthesis is 212 Discussion diverted from its normal course and consequently plastid structures and pigments in the cells tend to break down, so that yellowing is here a pathological effect resulting from the diversion of protein metabolism rather than extensive catabolism of protein. Villee: I would like to congratulate Dr. Yemm on his work; it is very elegant and extremely interesting. I was interested in the com- parisons which can be drawn between protein metabolism in animal tissues and plant tissues. I noted from your 15N data that the picture is quite similar in the two, that the highest rate of incorporation was into aspartic and glutamic acids, other than the amide N, and that the basic amino acid lysine had a very low rate, suggesting that, as in animals, lysine probably does not exchange its nitrogen. You stated that gluta- mine accumulates rapidly, more than can be accounted for by that present in the protein ; is that as glutamine or as glutamine plus glutamic acid? Yemm: As glutamine alone. Practically the whole of the soluble glutamic acid in these tissues is combined as the amide. Villee: I see. It would suggest a de novo synthesis of glutamic acid and the formation of the amine from that. Yemm: Yes. The deviation of these data depends upon a knowledge of the composition of the leaf proteins. We have separated tissue pro- teins from the leaf and estimated their average composition in terms of amino acids. We then measured how much protein disappeared over a given period of time and also the amount of glutamine that accumulated. When the two are compared, it is found that glutamine accumulates in much greater quantities than can be accounted for simply by hydrolysis of the proteins. A similar result was obtained for the other common amide of plants, asparagine. Villee: This, of course, is quite understandable biochemically because of the amination of the corresponding keto acids. Now, I was also interested in the increase in respiration which occurs. Since we know that substituted phenols are accelerators of respiration, I wonder whether the breakdown of pigments such as anthocyanins might in some way release either nitrophenols or chlorophenols which could be a stimulating factor at this time. The coincidence of this with the yellow- ing of the leaf suggests that perhaps in the yellowing, some of the pig- ments themselves break down and release physiologically active substances. Yemm: I have no comment on it. I think that it is difficult to account for its occurrence in different types of leaf, very different in composition and in their anthocyanin pigments ; yet the acceleration of respiration seems to be a general phenomenon in senescent leaves and is closely correlated with yellowing and with protein metabolism. Villee: Have you made any estimates of phosphorus and oxygen ratios to see whether there is, in fact, a change in this ratio which would be the crucial point as to whether phosphorylation has altered later on ? Yemm: I can only say that we have tried it without success. The situation in plants, I think, is here perhaps a little more complicated than in animals. It seems that there are, in all probability, a large Discussion 213 number of polyphosphates besides the relatively simpler ATP/ADP system found in animal tissues, and it is a problem of some complexity to separate these from one another in analytical determinations. We have tried the methods developed for animal tissues, such as barium salt precipitation, without success. Villee: The other thing I would like to mention is that in point of fact you measure COa production, and it might be a little dangerous to call this "respiration ". Of course, in plants you do have this special problem of the simultaneous production and utilization, so to speak, of oxygen and carbon dioxide, and it is very difficult to separate respiration from photosynthesis. Yemm: Yes, but of course all the respiratory data are obtained in the dark. The two processes are separated in that way. In point of fact, with barley and some of the other leaves, respiratory quotients have been measured. The respiratory quotient of the yellowing leaf tends to be less than one, so that if one considered oxygen uptake it would show an even more striking increase during senescence than C02. The respira- tory quotient of less than one, I think, just reflects the fact that other substrates such as the carbon skeletons of amino acids are oxidized in ageing tissues besides sugars. Villee: Yes. That is what I was leading around to, because the COa may come from many different things. And, of course, the oxygen which leaves the plant as COa does not necessarily come in as gaseous oxygen. Williams: I would like to go into the question of ageing in general. Dr. Yemm seems to be dealing with the ageing of transient tissues but as far as I can gather it seems to be rather for the benefit of the per- petuation of the plant, does it not ? For instance, in a thing like barley, the products are transferred to the seed. I gather, too, that in a de- ciduous tree you are presumably using up metabolites or storing them for the next season. Now, in that sense I should have thought that this was more like the metabolism of the muscle in a salmon, for instance, in the breeding season which is presumably for the benefit of the salmon rather than an ageing process which is detrimental as it is in most animal systems. What happens, for instance, in perennial garden plants which are not everlasting? Is there, in fact, a reproductive life and then after- wards is there a senile stage in those sorts of plants where reproduction does not persist but the plant lives for some years before dying, and in those cases when the plant does die are there the same sort of changes ? Yemm: Well, your first point: I think that senescence of the leaves is a part of the ontogeny of the plant as a whole, and I think it is perhaps desirable that senescence in animals should be regarded in the same way. There is no such thing as an isolated senescent tissue — it seems necessary to consider it in relation to what is happening in the rest of the organism. The second point : I do not think there is true senescence of whole plants which propagate themselves vegetatively. It was for a long time thought that this was the case in potatoes, strawberries and other plants, but it has turned out in many cases that it is the tendency to develop 214 Discussion viral infections and pathological conditions which leads to the gradual decline in the vigour of the plants. I think that many plants can reproduce themselves vegetatively for a very long time without showing "senescence" in the clone as a whole. Rowlands: What happens in, say, the coniferous trees — does the cycle of yellowing go on in the pine needle ? Yemm: I do not think it has ever been examined experimentally. But I think there is no doubt there is an extensive breakdown of protein materials before the pine needles drop because of the extremely low protein and nitrogen content of pine needles after leaf fall. Huggett: I was interested in your finding that you have some synthesis accompanying senescence breakdown, as shown by your 15N data, be- cause that fits in with some observations which I have not reported in which 14C-labelled glucose incorporates into the rabbit placental glycogen, not only before the peak period but in the period after the peak period when it is normally disappearing. Therefore, senescence really appears to be a balance of synthesis and catabolism and, as you pointed out earlier, that balance is changing. The only thing I would like to know arising out of that is : are there any new synthetic mechanisms occurring during senescence that are not present in full adulthood or young life — in other words, replacement mechanisms that are distinctive of old age ? Yemm: Nothing comes to mind. Certain syntheses become more obvious, but whether it is a result of a change in the whole balance of metabolism one does not know. The one I was mentioning — glutamine — now this is a synthesis which becomes much more conspicuous in older leaves, but in our view the increase of glutamine may result from an alteration in the balance between synthesis and breakdown of protein as the leaf ages. Huggett: Are you entirely happy that oxidative metabolism rises with yellowing? It looked to me from the trace of curves you showed that there are several other confusing factors that did not seem to be correlated ? Yemm: All I would say, in general, is that there is very strong evidence that the rate of cellular respiration is high and probably rising in senescence. The point is this, that I think we cannot explain the decline in synthetic activity as being due to a loss in respiratory or oxidative capacity of the tissue ; the changes are, if anything, the reverse. It seems necessary to postulate some change in the efficiency of the coupling between respiration and cellular metabolism. THE PHYSICAL INSTABILITY OF HUMAN RED BLOOD CELLS AND ITS POSSIBLE IMPORTANCE IN THEIR SENESCENCE* J. E. Lovelock National Institute for Medical Research, London According to Comfort (1954) in a review of the biological aspects of senescence: "All theories of senescence at the pre- sent time must be based on unwarrantable assumptions in the absence of concrete answers to the essential questions of fact". One assumption which seems generally acceptable however is that senescence is in some way connected with "wear and tear", or in physical terms with the natural tend- ency for disorganization or entropy to increase with the passage of time. The notion that senescence is due to wear and tear in a mechanistic sense, namely, that living organisms wear out as a result of their continued activity, is probably erroneous but lies behind the widespread belief that the slowing or arrest of metabolism at low temperatures would lead to suspended viability. In recent years a considerable body of information has accumulated on the effects of low temperatures on living cells, and this resulted primarily from the important discovery of Polge, Smith and Parkes (1949) that spermatozoa could be protected against the otherwise fatal effects of freezing by the addition of glycerol to their suspending medium. Since that time many types of living cell have been preserved in the frozen state for periods greater than their normal life- span. It is found, however, that living cells do not necessarily survive for long periods when their temperature is lowered to a level sufficient to reduce metabolism to a negligible value. The progressive loss of red blood cells stored at —78° has been reported by Mollison, Sloviter and Chaplin (1952), and at this * This paper was presented by Dr. A. S. Parkes on behalf of Dr. Lovelock. —Ed. 215 216 J. E. Lovelock temperature ovarian tissue may also fail to survive even one week, although under otherwise similar conditions it survived a year or more at —190° (Parkes and Smith, 1953). The survival of red blood cells at — 20° is poor compared with that at —78°, although at both temperatures glycolysis has slowed to a point where there is a negligible consumption of glucose during the survival of the cells (Chaplin et ah, 1954). The notion that the cessation of activity would lead to the suspension of vital processes assumes not only that the harmful effects of wear and tear result directly from activity, but also that the cell is stable in a physical sense in the absence of activity. Nowadays it is more usual to regard living cells as dynamic entities. It is thought that their contents, and possibly also their structures, are maintained at a constant level by a balance between the accretion of synthesis and the losses due to chemical change and to passive diffusion. If a cell, which is in dynamic equilibrium at its normal temperature, is cooled to a temperature where metabolism ceases, or is greatly slowed, it will continue to suffer the loss and disorientation of its components by random molecular movement and by dif- fusion, for these physical processes are only slightly affected by a fall in temperature. It follows that the life-span of a cell, which is normally in dynamic equilibrium, will be short if it is stored at a temperature where the balance between synthesis and diffusion is no longer on the credit side. It may well be significant that the successful cold storage of living cells has so far only been achieved under conditions where diffusion itself is greatly slowed, namely at very low temperatures, in the presence of viscous substances such as glycerol, or in the dry state. The concept of the living cell as a steady state system is helpful not only in emphasizing the inadequacies of the purely mechanistic approach to the manifestations of entropy, but also in suggesting the more probable nature of the wear and tear process suffered by the cell, namely, the loss and disorienta- tion of its substance by diffusion. Diffusion is usually regarded Human Red Cell Instability and Senescence 217 as a relatively slow process, but at cellular dimensions it can proceed with considerable rapidity. This is illustrated by a calculation (Ponder, 1948) showing that 90 per cent of the haemoglobin of a red cell can diffuse in four seconds through a hole occupying only one thirty-thousandth of the cell area. In comparison with the information available concerning the metabolism of various types of living cell, little is known of the spontaneous diffusion processes to which they are con- tinuously subjected. One exception, however, is the red blood cell; it is well established that its internal contents are not in thermostatic equilibrium with the environment but are maintained at a steady level by active processes. While some of the structural components and the haemoglobin are appa- rently inert (Muir, Neuberger and Perrone, 1952), it has been shown that the cell lipids are in a condition of rapid turnover both by metabolism (Muir, Perrone and Popjak, 1951 ; Altman, 1953), and by passive diffusion (Gould, 1951). The integrity of the red cell depends upon the presence of the lipids, and since these substances are free to diffuse away, then to some extent the cell itself can be regarded as a steady state system maintained intact by its metabolic activity. This information, together with the knowledge that the normal life-span of the red cell is known with considerable precision and the fact that it is particularly convenient for experimental purposes, decided the choice of this cell for the investigation of its physical stability, even though it was real- ized that the red blood cell is highly specialized and cannot be considered representative of living tissue. This paper describes some experiments on the diffusion of components from the red cell and its ensuing dissolution. A detailed account of the experimental material and methods is described elsewhere (Lovelock, 1955a). The loss of sub- stance was accelerated by suspending the cells in a medium maintained unsaturated with respect to the components of the cell membrane. This was achieved either by repeated washing or by including in the medium a neutral adsorbent substance, namely alumina. A few observations are also included on the 218 J. E. Lovelock dissolution of red cells during their storage at low tempera- tures. The progress of the dissolution is described and the structure and state of the cell membrane discussed in the light of the experimental observations. The possible relevance of these observations to the problem of senescence is also discussed. The dissolution of red cells in the temperature range 0° to +40° The suspension of human red cells in 0-16M-NaCl is fol- lowed by a loss of lipid components from the cell membrane and to a lesser extent by haemolysis (Lovelock, 1954). The loss of lipids is rapid at first, but after 3 minutes an equilibrium level is reached, and the rate of loss falls to a low value. The equilibrium concentration of lipids does not appear to repre- sent a simple saturation of the medium with these substances, 30r 5 10 15 NUMBER OF WASHINGS 20 Fig. 1. The removal of components from the red cell by repeated washing. The cells were suspended in 0 16 M-NaCl for 10 min. at 37°, separated by centrifuging and resuspended in fresh 0-16 M-NaCl. This was repeated 20 times. + Lipoprotein ; □ Volume ; 3 Cholesterol ; X Phospholipid; O Haemoglobin. Human Red Cell Instability and Senescence 219 but depends upon the previous treatment of the cells. It increases with the length of time the cells have been stored in vitro and decreases with the number of times they have been washed. The effect of repeated suspensions in 0-16 M-NaCl is shown in Fig. 1. Even after 20 washes only 6 per cent of the cells have haemolysed, and the proportion haemolysing was con- stant for each washing. During the first 7 washes, however, 10 20 30 40 50 TIME (MINUTES) 60 Fig. 2. The rate of removal of components from the red cell by exposure to alumina. Cells suspended in 0 16 M-NaCl were exposed to powdered alumina (500 mg. A1203 per ml. of packed cells) at 37° 0 Dry weight of lipids ; + Lipoprotein ; (J Cholesterol ; X Phospholipid ; O Haemoglobin. a considerable amount of the cell membrane was dispersed in the medium and shrinkage of the cells took place. There- after the losses of cell components proceeded at a rate equi- valent to the loss of cells by haemolysis. Figs. 2 and 3 show that the effects of exposure to alumina are closely similar to those of repeated washing. The haemo- lysis of the cells is directly related to the time of exposure and concentration of alumina. The loss of lipids, however, bears a direct relationship to the time and intensity of exposure only 220 J. E. Lovelock 10 15 CONCENTRATION OF ALUMINA (QM./ ML. OF CELLS) 20 Fig. 3. The removal of components from the red cell by various concentrations of powdered alumina. Cells suspended in 0-16 M-NaCl were exposed for 15 min. at 37° to concentrations of alumina from 0-2 g. to 2-0 g. per ml. of packed cells. 0 Dry weight of lipids; + Lipoprotein; □ Volume; (J Cholesterol; X Phospholipid; O Haemoglobin. 25r -5 O 5 10 TEMPERATURE (°C) Fig. 4. The removal of components from the red cell by alumina at various temperatures. Cells suspended in 0 16 M-NaCl were exposed to 1 -0 g. of alumina per ml. of cells for 15 min. at temperatures between 15° and -5°. + Lipoprotein; (J Cholesterol; X Phospholipid; O Haemoglobin. Human Red Cell Instability and Senescence 221 after a considerable proportion of the membrane has been removed. This proportion appears to be more or less inde- pendent of the concentration of alumina. The results suggest that the dissolution takes place in two steps. First, there is a rapid loss of superficial or surplus material which is not immediately harmful. Thereafter the rate of loss of membrane components is closely proportional to the rate of haemolysis. This suggests that any further loss of material from individual cells leads to their rapid and complete destruction. Figs. 4 and 5 show the effect of temperature upon the rate of loss of lipids and the haemolysis of cells exposed to alumina. Between 40° and 5° the rate of haemolysis proceeds in a manner consistent with a process governed by simple diffusion, that is, with an activation energy of 5500 calories. Below +5° the rate of haemolysis and loss of lipoprotein begins to increase, but the loss of cholesterol and phospholipid continues to fall. Previous experiments (Lovelock, 1954) indicated that the loss of components from the red cell is not reversible in a simple manner. Cells which have been denuded of membrane components regain very little of their lost material after one hour in fresh plasma at 37°. The experiments just described were carried out using cells stored at 4° in "acid citrate dextrose" media for from 3 to 10 days. During this period of storage there is little or no change in the viability of the cells, as judged by their survival on transfusion (Loutit, Mollison and Young, 1943). It was noted, however, that the deleterious effects of both repeated washing and exposure to alumina increased during this period of storage. The experimental evidence shows that a considerable pro- portion of the red cell membrane is easily detached by washing or by exposure to alumina. The proportion detached does not appear to be a defined superficial layer or capsule, since the proportion lost varies with the temperature and duration of cold storage. The removal of this layer is not immediately 222 J. E. Lovelock harmful, and cells so treated do not haemolyse more rapidly than normal cells when suspended in 0 • 16 M-NaCl. The loss of components is, however, accompanied by a loss of volume. 1.0- 0.9 3 o z LU < I u. O U) fe a: o CJ O 08 CX7 0.6 05 3.1 3.2 3.3 3.4 3.5 3.6 3.7 33 40 30 20 10 0 -10 T'X I03 AND TEMPERATURE CO Fig. 5. The haemolysis of red cells in the presence of alumina at temperatures between 40° and —5°. Cells suspended in 0 16 M-NaCl were exposed to 2 0 g. of alumina per ml. of cells for 15 min. For temperatures below 0°, 2 0 m glycerol was included in the suspending medium to prevent freezing. Where glycerol was used the experiment is indicated ( X ), otherwise (£). The results are expressed as the log. rate of haemolysis and the reciprocal of the absolute temperature. This could result, for example, from changes in the internal composition of the cell due to alterations of its permeability. Recent preliminary measurement of the electrolyte and water Human Red Cell Instability and Senescence 223 content of the red cell, after treatment with alumina, suggests that no alteration in the internal composition takes place. If, however, the lost volume is simply that which was occupied by the detached components, it is possible to calculate the concentration of these components in the detached layer. The results suggest that the concentration of the removable layer is equivalent to that of a 2 per cent lipoprotein gel. This agrees well with the value suggested by Mitchison (1953), based on birefringence measurements, for the concentration of lipoprotein in whole membrane. Ponder (1954), however, suggests that the membrane is much less hydrated than this and gives an estimate of 33 per cent for the concentration of membrane material from observations of the volume of fragmented ghosts. It seems worth considering the possibility that the concen- tration of membrane components is not constant throughout its thickness but increases radially inwards from the surface. This could explain the ease of detachment of the tenuous surface layers, and to some extent resolve the discrepancies between the estimates of the membrane thickness and composition. The classical view of the structure of the red blood cell membrane implies a framework of stroma protein surrounded by a bimolecular layer of lipids, with a sprinkling of antigenic protein at the surface. It seems unlikely that a cell possessing such a structure could lose a substantial part of its membrane and still remain intact. Among the recent views upon the architecture of the red blood cell, that of Moskowitz and Calvin (1952) agrees best with the experimental results above. They envisage a membrane composed principally of fibrils of a lipoprotein, elenin, orientated parallel to the cell surface and cemented together by ether-soluble lipids. If the structure of the red cell membrane suggested by Moskowitz and Calvin is accepted, then the physical dissolu- tion of red cells might proceed as follows : when the cells are suspended in a fresh saline medium the surface lipids will dissolve or disperse in the medium ; the lipoprotein which was held in position by the lipids will then become detached and 224 J. E. Lovelock diffuse away. In a medium maintained continuously un- saturated with respect to the lipid components of the cell membrane this process will continue until so much of the lipoprotein has unravelled that the intact existence of the cell is no longer possible. The increase in the rate of haemolysis below 5° could result from a hardening of the cementing lipids, if the repair of small breaches in the membrane depended on their ability to flow. Recent investigations of the effects of thermal shock on red cells suggest that such a hardening of the lipids may in fact occur (Lovelock, 1955b). The free energy of the lipid components of a highly organized structure such as the red cell membrane is likely to be higher than that of the same components in their saturated solution. It follows that the red cell membrane is probably unstable in a physical sense, and its intact existence may well depend upon the continuous synthesis of lipid components, and upon the presence of a considerable reserve of membrane material. The effects of cold storage The experimental evidence so far shows that the red cell will disintegrate rapidly if kept in a medium which is not saturated with the components of the cell membrane. In their normal environment and during cold storage in their plasma the cells swim in the presence of a considerable excess of lipid and possibly other membrane components; in these circumstances it might be thought that damage by diffusion would not take place. The chemical analysis of red cells, after storage for a few months at 20°, indicated that changes very similar to those following exposure to alumina take place (Lovelock, 1954). Unless a physiologically abnormal medium is used for their storage, namely one which is acid and which contains a salt such as sodium citrate or lactate, a rapid dis- solution of the cells takes place. The beneficial effects of the acid citrate medium used for the routine cold storage of red cells are directly attributable to its poor solvent action towards stroma lipoprotein. Even in this medium, however, adverse changes do occur Human Red Cell Instability and Senescence 225 and are made apparent by a progressive increase in the ease of detachment of membrane material by washing or alumina and by the ultimate failure of the cells to survive when trans- fused. These changes can occur at all temperatures between 0° and — 78° and are therefore presumably physical in nature, for below — 20° metabolism has effectively ceased. In a personal communication Dr. A. Richardson Jones reports that it is possible to produce a specific anti-serum for the human red cell envelope by immunizing rabbits with alumina powder after its exposure to red cells. The anti- serum has the remark- able property of haemolysing freshly drawn cells but not cells which have been exposed to alumina, repeatedly washed, or subjected to prolonged cold storage. The experimental result Fig. 6. The relationship between the rate of spon- taneous haemolysis of red cells at 0° and the recipro- cal of the viscosity of their suspending medium. The cells were suspended in lightly buffered 0-16 M-NaCl, pH 7-0, containing 0-1 per cent glucose and various concentrations of glycerol between 0 and 4 • 0 m. AGEING VOL. 2 226 J. E. Lovelock strongly supports the notion that the changes in the cells during cold storage are similar to those induced by enforced diffusion. If the dissolution of the cell during cold storage is preceded by the diffusion and dispersion of its membrane, it would be expected that the rate of dissolution would be inversely related to the viscosity of the medium and directly related to the solubility of the membrane material. Some confirmation of this is provided by the results shown in Figs. 6 and 7 where 100 l(K\^5sSi5 *£ w 80 - Ur^£i^S,NOv O" z > 4>Vk H3i^52 ^Xrc > 60 tc XJto\),0-6l ^OA 216, 217 Subject Index 263 Red blood cells dissolution of, in temperature range 0° to +40°, 218-224 effects of cold storage on, 224— 229 enzyme content of, 235 human, physical instability of, and its possible importance in their senescence, 215- 232 life-span of, 233, 234 physical instability of, 217 problem of ageing of, 229-231 self-repair of, 234 storage of, 236-238 Reinke, crystalloids of, 92, 94, 95 Reproductive glands, accessory, role of foetal sex hormones in de- velopment of, 3-17 Reproductive tract, relation of foetal sex hormones to differenti- ation of, 3-17 Respiration and protein metabolism in ageing tissues, 207, 208 Seminal vesicles, development of, in cultured reproductive tracts 8,9 normal development of, in rat, 5 r61e of endogenous sex hormones in development of, 8, 9 Sertoli cells, 97, 98, 99 Sex hormones, foetal, relation to differentiation of reproductive tract, 3-17 Somatotrophic hormone, foetal development of rat after adminis- tration of, to the mother, 161-175 Spermatid, differentiation of, 87-91 electron microscope in study of, 87-96 Spermatocyte divisions, 87 Spermatogenesis, condensation of the karyoplasm, 90, 91 electron microscope in study of, 87-96 formation of aerosome and head cap, 88, 89 observations on, 87-96 spermatocyte divisions, 87 Stilboestrol, effect on ovaries of hypophysectomized rats, 60—64 Sweat gland, apocrine (see Apocrine sweat glands) Testis, human, cytomorphosis of germinal and interstitial cells of, 86-99 electron microscopy of, 86-99 interstitial cells, observations on, 91-96 in human foetus, 26 in rabbit foetus, 19, 20 of deer, seasonal changes in, 179, 180 Testis hormone, role in differenti- ation of reproductive tract, 3-17 Testicular control of prostate in foetus, 19, 20 function and sexual structures, 19, 20 Testosterone, masculinization of female urogenital sinus by, 18 role in differentiation of repro- ductive tract, 3-17 Thyroid, in the foetus, 21, 22 relationship between hypophysis and, 21, 22 Urogenital sinus, and prostate for- mation, 20 effects of endogenous sex hor- mones on, 3-17 Uterus, uptake of radioactive potas- sium by, in rat and goat, 148-160 Vaginal cornification, 49, 58 Villi, placental, in goat, 155-158 Wolffian ducts, development of, in cultured reproductive tracts, 5-8 effects of surgical injury or absence of testis, 4-8 normal development of, in rat, 5 role of endogenous sex hormones in development of, 3—17 role of endogenous sex hormones in retention or loss of, 3 testicular control of, 19 X- irradiation, effect on oocytes, 34 effect on secretory capacity of ovary, 49-52 of mouse ovary, 49, 57 of rat ovary, 49-52 Yolk sac, mitochondria of, 101, 102 Printed in Great Britain SPOTTISWOODB, BALLANTYNE & CO. LTD. London & Colchester H