Observations on the Histogenesis of Ovarian Tumors Produced in Mice by X-Rays

OBSERVATIONS ON THE HISTOGENESIS OF OVARIAN TUMORS PRODUCED IN MICE BY X-RAYS

J. S. BUTTERWORTH, M.D. (Frorr~the Urfiarinzerzt of Pathology, Cornrll Ultiver,\ity Mrrlicul College, NFW York)

The occurrence of ovarian tumors in x-rayed mice has been described in a previous paper (1). These tumors were an incidental finding in a colony of 2068 nice, of which 775 were irradiated to increase the incidence of leu- kemia. The ovarian tumors were 15 times more frequent in the irradiated stock than in the control stock (2). In general the tumors were of two types: granulosa-cell tumors, including luteomata, and tubular adenomata.' All of them occurred in senile mice and, with the exception of the tubular adenomata, the same types were found in smaller numbers in the non-irradiated control mice. The present paper deals with the histogenesis of these tumors, which, like the histogenesis of the components of the normal ovary, has long been a perplexing problem.

Granulosa-cell Tumors: Considerable interest has recently been shown in granulosa-cell tumors of the ovary because of their endocrine activity. Von Kahlden (4) and Mengershausen (5) during the last decade of the nineteenth century were among the first to describe this type of tumor. Although it is a rather uncommon tumor in women, Bland and Goldstein (6) have recently reviewed the literature of the subject up to 1935, and found reports of 248 cases. Many different names have been applied to the growth, but the term granulosa-cell tumor, supplied by Werdt (7), is now in general use. The endocrine significance of the granulosa-cell tumor was first recognized by Schroder (8), Neumann (9), and Robert Meyer (lo), who reported cases showing cystic hyperplasia of the uterine mucosa, hypertrophy of the uterus, metrorrhagia, and other signs of excessive hormone production. Hormones have been demonstrated experimentally by injecting extracts of the tumors into immature or ovariectomized mice ( 11, 12, 13). The human granulosa-cell tumors are comn~onlysupposed to be derived from embryonic rests of granulosa cells, as first suggested by Robert Meyer in 1918 (14). His concept is based upon the observation that mature follicle cells proliferate only in association with ova, and upon the finding of nests of granulosa cells (Robert Meyer, 12) and of a small granulosa-cell tumor (Te Linde, 15) in the medulla of adult ovaries. Granulosa cells located in this

I This investigation has been supported by a grant from the International Cancer Research Foundation. 2 These tumors differ from the aclenoma tubulare ovarii testiculare first described by Pick (3) in that the cells are smaller and low-cuboidal in type, with scanty cytoplasm, and bear 110 re- semblance to the tubules of the testes. 86 J. S. BUTTERWORTH region are believed by most authors to be embryonic rests and not derivatives of adult follicle cells. The occurrence of a large proportion of these tumors after the menopause, when normal follicles are absent, is often cited in sup- port of this theory. Robinson (16), however, has taken exception to this theory and believes that granulosa-cell tumors develop from the epithelium of the ovarian follicles. In an ovary with a granulosa-cell tumor he describes, in the part not occupied by the tumor, follicles of which granulosa cells were undergoing proliferation and invading the stroma. Ewing and Schroder concur in this view. For further reference to literature on granulosa-cell tumors the reader is referred to the paper by Bland and Goldstein (6). The luteoma is a rare type of ovarian tumor. In 1921 Glynn (17) col- lected 14 cases of lutein-cell tumors which had been reported as " ovarian hypernephromata." He believes that there has been no verified example of hypernephroma in the ovary. Novak and Te Linde (18) accept most of Glynn's cases as true luteomata. Ewing (19) states that the peculiar structure of the tumor makes its origin from lutein cells probable, although no positive proof of the relation has been furnished. Novak and Brawner (20) and others have recently suggested that luteomata arise from lutein transformation of granulosa-cell tumors. The similarity of the human and mouse granulosa-cell tumors has recently been pointed out by Dr. H. F. Traut (2 1). Eflect of X-Rays on the Ovary: The effect of x-rays on the ovaries of both immature and adult mice has been carefully studied by Brambell, Field- ing and Parkes (22-26). Their work, however, covered a period of only four or five months after irradiation and no mention is made of the changes in the ovaries of senile mice irradiated early in life. When the mice were irradiated at three weeks of age (puberty occurs at six to seven weeks of age in the mouse) with a sterilizing dose of x-rays, Bram- bell, Fielding and Parkes (22) found that the graafian follicles underwent complete atrophy and were usually entirely reabsorbed. In a few cases some of the larger follicles were invaded by growth of the theca interna and membrana granulosa and eventually formed corpora lutea atretica which persisted indefinitely but seemed to have no physiological function. As these degenerative processes occurred, the germinal epithelium prolif. erated, forming a parenchymatous tissue which was a usual constituent of the ovary sterilized before puberty. In some cases this tissue became luteinized so that the entire ovary resembled a single large corpus luteum and inhibition of the estrous cycle occurred. Occasionally a second proliferation occurred from the germinal epithelium which formed cords of cells resembling anovular follicles . These authors also report three cases in which another kind of proliferation of the germinal epithelium occurred and resulted in the formation of epithelial canals, each of which, growing into the ovary, carried with it a layer of the tunica albuginea cells as a sheath. These canals were found in mice 98, 98 and 195 days after irradiation. The authors offer no explanation for these " anomalous " structures. Irradiation of the adult mouse (24) led to approximately the same histo- logical changes, but in this case the tissues forming the cortex of the ovary at the time of irradiation constituted the bulk of the ovary after irradiation. All ova disappeared and the ovary ultimately consisted of tissue derived from cells of anovular follicles formed by disintegration of the ovum and proliferation of the membrana granulosa of primordial and mature follicles. These tissues seldom underwent lutein transformation in the adult irradiated ovaries. Parkes (27) has found that the estrous cycle persists in animals with irradiated ovaries, even after complete disappearance of all normal follicles. This is attributed to the elaboration of estrus-producing hormones by the cells that proliferate in the sterilized immature animals shortly after irradiation, and by the cells derived from follicles in the sterilized adult animals.

The material used in this investigation was from three series of experiments : (1) The ovarian tumors previously reported (1) were studied from the point of view of histogenesis. (2) Thirty-nine female mice were irradiated at about sixty days of age and some of them were killed; the ovaries of others were removed by biopsy, at intervals from 3 days to 412 days after irradiation. As controls, 10 female mice of the same strains and the same age were kept under identical conditions. Part were killed; the ovaries of the others were removed by biopsy, at intervals of 35 to 320 days after irradiation. These irradiated and control female mice were never in contact with male mice. (3) Fifty-two female mice irradiated at ages varying from 30 to 130 days, and 15 non-irradiated female mice of various strains were killed or died at intervals from 17 to 745 days after irradiation and were used to study the later effects of irradiation. The tissues were ~reservedin Kaiserling's fixative and were stained with hematoxylin and eosin. In a few cases sections were stained for fat with Sudan 111. In several instances interrupted serial sections of the ovaries were made, and the hypophyses, adrenals, uterine horns, and vaginas of many animals were sectioned. The mice were irradiated over the entire body in groups of 15 to 25. The irradiation factors were as follows: 190 kv., 30 ma., 50 cm. distance, 0.5 mm Cu + 1 mm. Al. Three of the mice in this series were exposed to gamma rays emanating from 0.1 mg. of radium. These mice were exposed from the ages of 47, 60, and 64 days to the ages of 377, 676, and 495 days respectively, according to a technic previously described (28).

In Table I all irradiated mice in the second and third series are grouped in order of the time which elapsed between the first irradiation and the examination of the ovaries. Many of the mice were exposed to x-rays more than once and each figure in the fourth column indicates a separate dose. Multiple doses were given a month or more apart. The last column in the table indi- J. S. BUTTERWORTH

TABLEI Mice Ex@osed to X-Rays

Time in days between Age in days Doses of Degree of Mouse irradiation and at first x-rays in proliferation No. examination of irradiation r units of germinal the ovaries epithelium

Af 644 200 0 Ak 375 200 0 Af 665 200 0 Af 669 200 0 Af 656 200 0 Slb 181 50-50-50 0 Arb 157 50 X 4 0 Af 666 200-300 0 Af 677 200-250 0 Af 640 200-400 0 Ak 370 200-300 0 Af 679 200-300 0 Af 633 200-400 0 Ak 366 200-400 f Af 674 200-400 0 Ak 383 200-300 0 Af 680 200-300 0 Rg 326 400 0 Rg 337 400 0 Rg 329 400 0 Rg 338 400 f Af 665 200-400 0 Rg 336 400 0 Rg 345 400 0 Af 655 200400 0 Ak 388 200-300 + A 9238 400 +f Ak 372 200-300 0 Rg 334 400 + Af 649 200-400 + Ak 379 200-400 + Rg 331 400 + Ak 375 200-300 + Af 653 200-400 + Rg 332 400 + Af 674 200-400 ++++ Af 681 200-400 f Rg 333 400 0 Af 670 200400 + Aqe 17 200-250 ++ Aqe 20 200-250 +++ Af 671 200-400 ++ Af 644 200-400 + Rfb 124 400 ++ Rgc 38 200-300 + Rfb 89 450 + cates the extent of tubular down-growth of the germinal epithelium: + indicates a thickened epithelium with no definite down-growths; + indicates short down-growths; $ + and ++ longer down-growths ; ++ + $. indicates that the entire ovary is replaced by tubules. HISTOGENESIS OF OVARIAN TUMORS PRODUCED IN MICE BY X-RAYS 89

Time in days between Age in days Doses of Degree of Mouse . irradiation and at first x-rays in proliferation No. examination of irradiation r units of germinal the ovaries epithelium

Af 676 Af 636 Af 632 Af 634 Ar 252 A 9350 Rxa 17 Rg 343 Rxc 35 Arb 32 A 9351 Rg 330 Ak 374 Ak 365 Af 660 Af 652 Af 646 Af 647 Af 18 Afb 146 Slb 370 Afb 419 Rfb 404 Aka 70 Ak 385 Rg 344 Rg 341 Rg 115 Afb 153 Sgc 96 Rmb 22 Af 598 Ard 93 Sgc 98 Akd 19 Ara 27 Aka 86 Rg 344 Rg 103 Axa 20 Rxc 2 Rg 106

Radium treated mice

------* ET indicates an early ovarian tumor.

Histologic Changes Following Irradiation of the Mozise Ovary: The early changes in the ovary that follow irradiation with x-rays are well known (22-26, 31) and need not be described in detail here. The later changes that precede 90 J. s. BUTTERWORTH the developnlent of the ovarian growths may be illustrated by the following examples :

Mouse Af 666 (26 days after irradiation) : The centers of the larger follicles were filled with loose granulosa cells embedded in a pink-staining colloid material. No normal ova were present in the sections but some small irregular areas of pink-staining material sur- rounded by granulosa cells represented necrotic ova. The greater part of the ovary was composed of solid masses of cells derived from follicles. In these cells, which appeared partially luteinized, the granular material in the nuclei did not stain as deeply as that in the nuclei of normal granulosa cells; the cytoplasm was more abundant and took a light pink stain with hematoxylin and eosin. Under the tunica albuginea were several nests of follicle cells apparently derived from small or primordial follicles after degeneration of the ova. These cells resembled normal granulosa cells. The germinal epithelium was flat and showed no tendency to proliferate. The uterine horns seemed to be in a resting stage. Mouse Ak 370 (47 days after irradiation): This ovary is illustrated in Fig. 1. There is very little change from the 28-day stage. Under the germinal epithelium several anovular follicles can be seen, while the rest of the ovary is made up of lighter staining cells derived from the degenerated follicles. Several pink-stained hyalin masses can be seen in the central part, which are remains of degenerated ova. There is no proliferation of the germinal epithelium. Mouse Ak 366 (107 days after irradiationj: In this ovary the cells derived from the degenerated follicles have become larger, the cytoplasm is abundant, granular, and stains lightly if at all. These cells appear to be identical with the " clear" cells described in the previous communication (1). There are numerous nests of granulosa cells at the periphery of the organ and the germinal epithelium is thickened in places but no definite down- growths can be found. Mouse Ak 388 (195 days after Zrradiatiou): 'I'his ovary is the first in the series to show unmistakable down-growth of the germinal epithelium forming tubules which penetrate a short distance into the ovary. The other structures show no conspicuous change and the ovary is composed mainly of clear lutein cells and a few anovular follicles.

All ovaries in this series underwent similar changes after irradiation and it seems unnecessary to describe each case in detail. The different doses of x-rays used made no difference in the picture as far as we could observe, but the ovaries of mice irradiated several weeks after puberty seemed to contain more luteinized stroma and fewer anovular follicles than those of mice irradiated at, or shortly before, puberty. The ovaries of the three animals exposed to radium were indistinguishable from the ovaries of x-rayed mice of the same ages. The tubular down-growths were a very consiant feature of the ovaries which had been irradiated for more than 180 days (Table I). The extent of these down-growths varied from a few tubules at the periphery of the ovary to almost complete replacement of the ovary by tubules. In the entire series the ovaries of only 5 mice that lived more than 180 days after irradiation (Ak 372, Rg 333, Afb 146, Sgc 96, Ard 93) failed to show any down- growths in the available sections, From the foregoing brief survey it can be seen that the irradiated adult ovary is composed of structures that persist or are derived from the normal ovary (anovular follicles, luteinized cells, stroma) and newly formed structures from the germinal epithelium (tubular down-growths). Early Ovarian Tumors: It is often difficult to decide when the late changes in these irradiated ovaries are sufficient to warrant a diagnosis of tumor. In FIG. 1. Mouse Ak 370: Ovary 47 days after irradiation, showing small anovular follicles near the periphery and larger, partially luteinized masses of follicle cells. In the central part are several small irregular areas of hyaline material resulting from degeneration of ova. X 60. (All magnifications are approximate.) FIG. 2. Mouse Af 674: 105 days after irradiation. The upper part of the illustration shows several small anovular follicles. A mitotic figure is present in the small follicle in the upper right hand corner. X 300. FIG. 3. Mouse Afb 153: 448 days after irradiation. Several nests of granulosa cells (anovular follicles) can be seen scattered through the cortex of the ovary. The deeply stained area is an early mixed type of tumor composed of granulosa cells and tubules from the germinal epithelium. X 60. FIG. 4. Mouse A 9351: Mitotic figure in a clear cell. X 550. 9 2 J. S. BUTTERWORTH this series. 10 of the animals had ovaries that, because of their size and micro- scopic appearance, can be classified as early tumors. The majority of these tumors were a mixture of the granulosa-cell type and tubular adenoma. Mouse Rg 103: Lzitcomcz of the Oziczry, Lcion~yonztaof the Uterzu, Hyprrplasiu of Il~e Eftdonbetrizim (745 days after irradiution): One ovary of this inouse measured 1.0 cm. in longest diameter and was yellow in color. Microscopically it was composed mainly of lutein cells and in one area there was a round, circumscribed mass 4 mm. in diameter, composed of cells which seemed to be intermediary between granulosa cells and lutein cells. A fat stain showed that these cells contained considerable fat but not as much as the larger lutein cells with clear cytoplasm present in this ovary. At one edge of the growth was a small nodular mass composed of granulosa cells with interlacing tubules that did not show the presence of fat in the section stained with Sudan 111. The other ovary measured 0.3 cm, in longest diameter and n~icroscopicallywas a typical senile x-rayed ovary. The endometrium of the uterine horns was thickened; the glands were dilated to form cysts and in some areas the cells were secreting a pink-shining colloid material. There was a leiomyoma of the uterine wall measuring about 0.2 cm. in diameter. The hypophysis was enlarged and very vascular, with an increase In the relative number of chromophobe cells. The greater part of this tumor is similar to that found in Mouse Rr 294 and illustrated on page 71 (Fig. 3) of the previous communication (I). It is interesting to note that this animal received seven daily doses of 0.1 C.C.of Antuitrin-S one month before death and this prepara- tion may have caused the transformation of a granulosa-cell tumor into the partially luteinized cells described. A somewhat similar case in a woman has recently been reported by Cutler (29). Mouse Rg 330: Bilateral Grunulosa-cell Tzunors of the Ovary: Thls mouse had bilateral ovarian tumors, each 3 mm. in diameter, and composed entirely of solid masses of granulosa cells very similar to the ovarian tumor of Mouse Af 196, illustrated on page 81 of the previous communication (1 ). The uterine horns were in the resting stage; the adrenals showed slight subcortical fibrosis, and the hypophysis was not unusual. Mouse Rg 344: Bilateral Granzllosa-cell Tumors of the Ovary, Cystic Hyperplasia of tlte Endometriunt with Adenomyosis: This mouse had bilateral ovarian tumors, of which one was removed 418 days after irradiation and the other was examined when the mouse died 578 days after irradiation. Both tumors were essentially the same and were composed of solid masses of granulosa cells between which were scattered numerous cleft- or duct-like spaces lined by low-cuboidal epithelium. The endometrium of the uterine horns was thickened and the glands were widely dilated. There was slight subcortical fibrosis of the adrenal glands and the hypophysis showed no marked change. This tumor is similar to the ovarian tumor of Mouse Afa 12, illustrated on page 81 of the previous communication (1). This mouse also received seven daily injections of Antuitrin-S beginning 395 days after irradiation. A mild luteinizing effect was noted in the ovary removed at biopsy 15 days following the end of the Antuitrin-S treatment, but the ovary examined 175 days after the Antuitrin-S treatment did not show any luteinizing effect. Mouse Rg 115: Tubular Adenofnu of the Ovary (443 days after irradiation): The ovarian tumors found in this mouse were bilateral and cystic, measuring about 1.0 cm. in diameter. Microscopically the solid parts of the ovaries measured about 4 mm. across and were composed entirely of interlacing tubules as illustrated in Fig. 8. Mouse AfD 153: Mixed Tumor (448 days after irradzation): This tumor is illustrated in Fig. 3. The ovary is composed of luteinized stroma with a few nests of follicle cells as seen in most old irradiated ovarics. The darker parts of the illustration depict masses of granulosa cells through which tubules from the germinal epithelium grow in miscellaneous array. Mice Afb 419, Aka 86, Rfb 404, Rg 341 : Mixed Tzunors: These mice all had ovarian growths of the mixed type similar to that of Mouse Afb 153, described above, and also to that of Mouse Af 732, illustrated on page 87 of the previous communication (1). Histogenesis of Tubular Adcnomata: The histogenesis of this type of tumor can clearly be traced to down-growths of the germinal epithelium which com- FIG.5. Mouse Rg 334: 210 days after irradiation. The greater part of the ovary is composed of clear cells but early down-growths from the germinal epithelium can be seen at the periphery. X 70. FIG.6. Mouse Aqc 20: 287 days after irradiation. A later stage in the tubular down-growths from the germinal epithelium. X 60. FIG.7. Mouse Af 634: 320 days after irradiation. The entire ovary is composed of tubules derived from the germinal epithelium. X 60. FIG.8. Mouse Rg 115: 443 days after irradiation. This is a low-power illustration of a pure tubular adenoma. X 30. 94 J. S. BUTTERWORTH mence about 6 months after irradiation. Fig. 5 shows an early stage of the growth in the ovary of Mouse Rg 334. A few invaginations of the surface epithelium may be seen, and an occasional tubule is present. The greater part of the ovary is composed of luteinized stroma. A later stage is pictured in Fig. 6, where the tubules have penetrated the periphery of the ovary and form a ring of adenomatous growth about a core of luteinized stroma (Mouse Aqc 20). An advanced stage of this tubular invasion of the surface epithelium is shown in Fig. 7, where almost the entire ovary is replaced by the growth (Mouse Af 634). All sections are from the middle parts of the ovaries. Other ovaries in the series show all degrees of invasion between these ex- tremes. Often these ovaries are no larger than normal ovaries (2 to 3 mm. in largest diameter) but occasionally the growth continues and a tumor results. Such an adenomatous tumor was found in Mouse Rg 115, described above and illustrated in Fig. 8. In general this type of growth does not reach the size of granulosa-cell tumors or luteomata and seems to be devoid of excessive hormone activity. No cystic hyperplasia of the uterine mucosa or keratinization of the vaginal epithelium was noted in 5 mice in which the tubular down-growths from the germinal epithelium were conspicuous, and no granulosa or lutein cells were present in the sections. Rather frequently this tubular growth is mixed in with other types and a great many of our granulosa-cell tumors showed some formation of tubules unrelated to the remainder of the growth. Histogenesis of Granulosa-Cell Tumors: Although it is impossible to picture the actual formation of granulosa-cell tumors, we are able to demonstrate stages in their development that we believe serve to clarify their mode of origin. After irradiation of the ovaries in our experiments the nuclei of the ova became swollen and indistinct; the cytoplasm was shrunken and stained a deep pink. Eventually the ova entirely disappeared or were represented by irregular pink-staining hyaline masses. The large follicles became luteinized or in some cases many of the follicle cells became necrotic and disappeared. The smaller follicles, however, were not affected to the same extent, and at the periphery of the ovaries shortly after irradiation one could see many small nests of granulosa cells (anovular follicles, Figs. 1 and 2) which were formed either by destruction of the ovum in small or in primordial follicles, or possibly by growth of undifferentiated cells at the periphery of the ovary. We have never been able to demonstrate any connection between germinal epithelium and these anovular follicles. The small, round, circumscribed nests of cells persist near the periphery of the ovary and many of them increase slowly in size, as evidenced by the presence of mitotic figures (Fig. 2). These follicle-like nests can be found in mice of all ages, Figs. 1, 2 and 3 illustrate them in mice, 47, 105 and 448 days respectively after irradiation of the ovary. From these nests of slowly proliferating follicle cells it is but a short step to the stage in which there is an evident overgrowth of the follicular ele- ments. Fig. 11 illustrates the ovary of Mouse Sgc 98, 513 days after irradiation, showing a generalized hyperplasia of all the follicles. The last stage is that in which the cells of these follicles begin to proliferate FIG.9. Mouse Rra 20: 787 days after irradiation. The greater part of the ovary is composed of luteinized cells. In the upper left corner is a darker staining nest of granulosa cells which are actively proliferating. X 30. FIG.10. A high-power magnification of the area of granulosa cells (Fig. 9), showing a mitotic figure and a gradual transition toward lutein cells in the lower part of the illustration. X 400. FIG.11. Mouse Sgc 98: 513 days after irradiation. An ovary showing marked hyperplasia of anovular follicles. X 50. FIG. 12. Mouse Akd 18: 490 days after irr'adiation. The upper part of this ovarian tumor is composed of clear cells and tubules while the lower, darker stained part is made up of solid masses of granulosa cells which show many mitotic figures. X 22. 96 J. S. BUTTERWORTH more rapidly and invade the surrounding stroma, forming granulosa-cell tumors. The early granulosa-cell tumor of Mouse Rra 20 is illustrated in Fig. 9. Here one sees a small area of granulosa cells which are no longer sharply circumscribed and under high power can be seen to be proliferating rather rapidly, as evidenced by numerous mitotic figures (Fig. 10). At the periphery of this nest of granulosa cells there is a gradual transition to the lutein type of cell, which can be seen at the bottom of this figure. Fig. 12 illustrates the ovarian tumor of Mouse Akd 18, which is a more advanced granulosa-cell tumor. The upper third of this growth is composed of clear luteinized cells, among which are interspersed tubular down-growths from the germinal epithelium, as seen in any senile irradiated ovary. The lower two-thirds of the growth is composed of solid, nodular masses of granulosa cells which are expanding and compressing the tubules and luteinized cells. It is easy to visualize this granulosa-cell tamor as the result of unre- strained growth of the granulosa cells of anovular follicles. Histogenesis 01 Luteomata: The granulosa cells of tumors, like those of the normal ovary, seem to be the parent cells of the lutein cells, and transitional forms from one type to the other can be found in many of our tumors, such as that from Mouse Rra 20 (Fig. lo), already mentioned. In a few instances, however, we have found mitotic figures both in the lutein cells of luteomata (page 71, Fig. 4, of the previous communication, 1) and in the clear lutein- like cells of the irradiated ovary (Fig. 41, showing that each of these cells is capable of multiplying independently. Thus, it is possible that in a few cases luteomata may be produced by proliferation of lutein cells. It is quite possible that corpora lutea may have been present in some of the ovaries at the time of irradiation and subsequently gave rise to luteomata. In this respect Mouse Ax 100 is of interest. The left ovary was replaced by a solid granulosa-cell tumor (page 90 of the previous communication, I), while the right ovary was composed almost entirely of lutein cells. This is an added point in favor of the development of luteomata from lutein cells because any stimulation of the ovaries by the luteinizing hormone of the anterior hypophysis would presumably affect both ovaries. Nevertheless, Mice Rg 344 and Rg 103, which were given injections of Antuitrin-S (anterior pituitary- like hormones), showed evidences of lutein transformation of granulosa cells. Ovaries of Senile Unirradiated Mice: The structure of the normal mouse ovary is well known and need not be described at length here. Up to the age of about 15 months the control ovaries in our series preserved their usual char- acters. After this time most of the ovaries showed a decrease in the number of maturing follicles and an increase in the amount of ovarian stroma, so that the senile ovaries were composed mainly of stroma, corpora lutea and albi- cantia, pigment cells, and a few follicles. No down-growths from the germinal epithelium were seen in the ovaries of these mice.

Brambell, Fielding and Parkes, in their studies on the effect of x-rays on the ovary, found that no proliferation from the. germinal epithelium occurred in adult mice. Unfortunately, they do not state the age at which these mice HISTOGENESIS OF OVARIAN TUMORS PRODUCED IN MICE BY X-RAYS 97

were irradiated, but they say that the mice were adult and were observed for only 74 days after irradiation. In 3 mice, however, irradiated at 3 weeks of age, the ovaries were found to contain a few epithelial canals connected with the germinal epithelium and probably corresponding to our tubular down- growth. Our material clearly shows that proliferation from the germinal epithelium of the ovary does occur in most irradiated mice, 6 or more months after irradiation, and it is probable that most of the animals of Brambell, Fielding and Parkes would also have shown these down-growths if they had been Mesenchyme Ge~minalepithei ium

/' \\\ / / ,,,* \ /----,' &'/ -- '*A- ,, Ova Follicle cells -Persistin Tubular follicle ceyls dwngmvths I DisintegratedI Lutein1 cells Granuiosa-ce11 Tubular1 ova tumors adenomata

Luteomata

No~malprocesses in the ovary ----- Represents conflicting views -Processes induced or acceierated by X-pays

observed for a sufficient time. It must be borne in mind, however, that our dose of x-rays (average about 450 r units) was larger than that used by Brambell, Fielding and Parkes, which they describe as corresponding roughly to B tint on Levi's pastilles (equivalent to about 125 r), and that different strains of mice were used. Hormone Production: The production of follicular and corpus luteum hormones by granulosa-cell tumors and luteomata has been shown in our mice by proliferation and secretion of the uterine mucosa and keratinization and mucous transformation of the vaginal epithelium (1 ) . In 5 mice with pure tubular adenomata the uterine horns were in a resting stage, indicating that there was no conspicuous production of follicular hormone. If, on the one hand, these tumors have a derivation in common with granulosa cells (germinal epithelium), it would seem that their cells may produce small amounts of hormone or hormones. On the other hand, failure of these cells, which are derived from the germinal epithelium, to produce hor- 98 J. S. BUTTERWORTH mone may be interpreted in favor of the older view, recently revived by Fischel (30) and others, that the follicular epithelium arises from the ovarian mesenchyme. Further study will be necessary to settle this question. Relation of Human and Mouse Granulosa-cell Tumors: While most of the tumors that we have described have appeared in irradiated mice, we have found the same types of tumors, with the exception of tubular adenomata, in smaller numbers in unirradiated mice, and there is evidence that granulosa-cell tumors of the human female and the mouse are essentially the same type of growth (21). Thus, while it is possible for granulosa-cell tumors to develop from em- bryologic rests of granulosa cells (Granulosaballen of Robert Meyer), our observations do not indicate that it is necessary to postulate the presence of these rests of cells in adult and senile ovaries to explain the origin of granulosa- cell tumors. Chart I represents the derivation of tubular adenomata, granulosa-cell tumors, and luteomata in irradiated mice.

The ovaries of 91 mice irradiated with x-rays and 25 control mice have been studied at various ages from 5 weeks to 25 months in an effort to de- termine the histogenesis of tubular adenomata, granulosa-cell tumors, and luteomata. In irradiated mice tubular adenomata are formed by down-growths of the germinal epithelium of the ovary. The earliest down-growths appear about 6 months after irradiation. Granulosa-cell tumors are derived from nests of follicle cells which form at the periphery of the ovary after irradiation and in senile mice proliferate to form tumors. These nests of cells are probably derived from normal follicles following degeneration of the ova. Luteomata are usually formed indirectly by luteinization of granulosa-cell tumors, but it is possible that they are occasionally formed directly by division and proliferation of lutein cells. Granulosa-cell tumors and luteomata of mice are similar to those of women and it is unnecessary to postulate embryonic rests of granulosa cells to ac- count for the histogenesis of granulosa-cell tumors.

NOTE: The author wishes to express his gratitude to Dr. Jacob Furth for continuous encouragement and advice.

1. FURTH,J., AND BUTTERWORTH,J. S.: Am. J. Cancer 28: 66, 1936. 2. FURTH,J., AND FURTH,0. B.: Am. J. Cancer 28: 54, 1936. 3. PICK, LUDWIG:Arch. f. Gynak. 76: 191, 1905. 4. VON KAHLDEN,C.: Centralbl, f. allg. Path. u. path. Anat. 6: 257, 1895. 5. MENGERSHAUSEN: Inaugural Dissertation, Freiburg, 1894. Quoted by Bland and Goldstein (6). 6. BLAND,P. B., AND GOLDSTEIN,L.: Surg., Gynec. & Obst. 61 : 250, 1935. 7. WERDT,F.: Beitr. z. path. Anat. u. z. allg. Path. 59: 453, 1914. 8. SCHR~~DER,H.: Zentralbl. f. Gynak. 46: 195, 1922.