Is resumption of meiosis in the human preovulatory oocyte triggered by a meiosis-inducing substance (MIS) in the follicular fluid?

Transcription

1 FERTILITY AND STEmLITY Copyright c 1984 The American Fertility Society Vol. 41, No.3, March 1984 Prinred in U.8A. Is resumption of meiosis in the human preovulatory oocyte triggered by a meiosis-inducing substance (MIS) in the follicular fluid? * Lars Westergaard, M.D.t Anne Grete Byskov, Dr.Sci.:!: Claus Y ding Andersen, M. Tech. Sci.:f; JlIlrgen Grinsted, M.D.:!: Kenneth P. McNatty, Ph.D. Hvidovre Hospital, Hvidovre, Denmark, The Finsen Institute, Copenhagen, Denmark, and Wallaceville Animal Research Centre, Private Bag, Upper Hutt, New Zealand Aspirates from 31 ovarian follicles and 2 corpora lutea of 26 women at different stages of the menstrual cycle were investigated for activity of meiosis-inducing substance (MIS) and meiosis-preventing substance (MPS). The aspirated follicles were classified as dominant (i.e., preovulatory), healthy, or atretic according to their size, steroid hormone content, and stage of the menstrual cycle. To test for MIS and MPS activity, gonads of sexually undifferentiated fetal mice were cultured in media containing either 15% follicular fluid aspirate (test gonads) or 15% human blood serum (control gonads). MIS activity is present in follicular fluid if the test testes contain more meiotic germ cells than the control testes. MPS activity is present if the test ovaries have less meiotic germ cells than their controls. MIS activity was present only in healthy follicles of the late follicular phase (12 of 15 follicles). No MIS activity was seen in healthy or atretic follicles from other phases of the menstrual cycle. The MIS activity is apparently unrelated to the composition of steroids in the follicular fluid. MPS activity was not found in any of the follicles. It is proposed that the preovulatory resumption of meiosis may be triggered by a MIS in thefollicular fluid. Fertil Steril 41 :377, 1984 In man, as in most other mammals, oocyte meiosis is initiated during fetal life and is arrested at the diplotene stage of the prophase before birth. At this time, each oocyte is surrounded by a layer of granulosa cells. Normally, the oocyte meiosis is Received September 6, 1983; revised and accepted November 3,1983. *Supported by Nordisk Insulinfond (A. G. B.). treprint requests: Lars Westergaard, M.D., Department of Obstetrics and Gynecology, Afsnit 537, Hvidovre Hospital, Kettegard AIle 30, 2650 Hvidovre, Denmark. :f:finsen Laboratory, The Finsen Institute. Wallaceville Animal Research Centre. Vol. 41, No.3, March 1984 kept arrested at the stage of diplotene for many years and is first resumed in adult life in relation to the preovulatory gonadotropin surge. 1, 2 In the male gonad the COUrse of events is quite different; the germ cells became enclosed in the testicular cords. early during fetal life, and meiosis is not induced until puberty. From that time on, meiotic divisions take place continuously. It has been proposed that meiosis in mammalian gonads is regulated by a meiosis-inducing substance (MIS) and a meiosis-preventing substance (MPS).3 This theory is based on experiments in which the activity of MIS in different test media is shown by their ability to induce Westergaard et ai. MIS in human ovarian follicles 377

2 meiosis in fetal mouse testis in vitro, whereas MPS retards or prevents the beginning of meiosis in fetal mouse ovaries in vitro. Using this in vitro test system, MIS has been demonstrated in fetal and adult ovaries 4, 5 and MPS as well as MIS in fetal and adult testicular tissues of different mammals, including man.3, 6 10 Recently it was shown in the adult rat testis that MIS is present in higher concentrations in the parts of the seminiferous tubules that contain germ cells in stage VII to stage VIII of the spermatogenic cycle (i.e., the stages in which meiosis is initiated) than in the parts of the tubules containing germ cells in other stages. MPS was detected in all parts of the tubules but in relatively low and constant concentrations. 11 These results suggest that it is the ratio between MIS and MPS in the testicular tubules that regulates meiosis in the adult testis. Whether MIS and MPS or a ratio between the two also regulates meiosis in the adult human ovary is not known, but one may speculate that the meiotic arrest is caused by the presence of MPS in the follicle and that the resumption of meiosis at ovulation is caused by increasing amounts of MIS in the preovulatory follicle. An alternative proposal has been reported by Tsafriri and Channing. 12 Their experiments indicate that the oocyte is held in meiotic arrest by an oocyte maturation inhibitor (OMI), which is detectable in follicular fluid, and that a decrease in OMI as the follicle matures is one of the factors which allows meiosis to be resumed at ovulation (for a review, see Tsafriri et al. 13 ). Notwithstanding the above two proposals, a MIS-like activity has been demonstrated in human follicular fluid of large, preovulatory follicles. In these experiments, we showed that aliquots of follicular fluid stimulated meiosis in undifferentiated fetal mouse testes in vitro.5 The purpose of the studies reported herein was to test the MIS- and MPS-like activities offollicular fluid aspirates from adult women experiencing spontaneous, regular menstruation. In order to demonstrate potential physiologic roles of MIS and MPS, the aspirates were obtained at different stages of the menstrual cycle, and the follicles from which the fluids were recovered were classified as healthy atretic or healthy dominant according to their size and steroid content. Finally, we wanted to test the quality of the semiquantitative method traditionally used to evaluate MIS and MPS activity (i.e., the differ- 378 Westergaard et ai. MIS in human ovarian follicles ences in the distribution of nonmeiotic and meiotic germ cells in the fetal mouse test and control gonads) by comparing the results of this method with differential counts of the germ cell stages in the fetal mouse gonads. SUBJECTS MATERIALS AND METHODS Follicular fluid aspirates were obtained from 26 women (18 to 40 years of age; median age, 32 years) with spontaneous, regular menstruation who underwent laparotomy at varying stages of the menstrual cycle. The indications for surgery were either legal sterilization (17 women) or gynecologic disorders not involving the ovaries (9 women). Before the operation, all of the women were informed about the follicle aspiration procedure and the purpose of the study, and all gave their consent to participate. DATING THE MENSTRUAL CYCLE A sample of peripheral blood (collected before the aspiration of follicles) and an endometrial biopsy were obtained during the operation. The stage of the menstrual cycle was assessed from the date of the last menstrual period, the concentration in serum of luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E 2 ), and progesterone (P), and the endometrial biopsy. The menstrual cycle was divided into six phases: EF, early follicular (days -14 to -10); MF, midfollicular (days - 9 to - 5); LF, late follicular (days -4 to 0); EL, early luteal (days + 1 to +4); ML, midluteal (days + 5 to + 9); and LL, late luteal (days + 10 to + 14). Day -14 is the first day of menstrual bleeding, and day 0 is the day of expected ovulation. Of the 26 women participating in this study, 5 were in EF, 4 in MF, 13 in LF, 1 in EL, 1 in ML, and 2 in LL at the day of follicle aspiration. COLLECTION OF FOLLICULAR FLUIDS During the operation and with the ovaries in situ, fluids from 31 antral follicles and 2 cystic corpora lutea were obtained by aspiration using a 25-gauge needle (Butterfly-25 short, Abbott Laboratories AIS, Tr~r~d, Denmark). Before aspiration of the follicles, their diameters were measured. Fertility and Sterility

3 The aspirates were immediately transported on ice to the laboratory and centrifuged at 200 x g for 5 minutes. The supernatants were divided in two portions: one was assayed for steroid hormones, and the other (at least 0.5 ml) was used for the MISIMPS test. Both portions were stored at - 20 C until analyzed. THE MISIMPS TEST To test for MIS and MPS activity in follicular fluid, we used a modification of a previously described in vitro test system. 14 In this revised test, sexually undifferentiated gonads from 11 %-dayold mouse fetuses (Balb-C strain)-the day of the copulatory plug being day I-are dissected out together with the adjacent mesonephric tissue. The gonads are cultured for 6 days in 0.5 ml of culture medium, to which is added either follicular fluid (test medium) or serum (control medium) of peripheral blood from the same woman. Thus, the culture medium contained either follicular fluid or serum which was diluted to 15% with Eagle's modified minimum essential medium (Flow Laboratories, Edinburgh, Scotland), supplemented with streptomycin (50 J.Lg/ml), penicillin (50 IU/ml) and glutamine (2 mm). One gonad of each fetus was cultured in the test medium (test gonad), and the contralateral gonad (control gonad) in the control medium. After 6 days in culture at 35 C in 95% air and 5% CO 2 in a closed water-saturated chamber, the tissues were fixed in Bouin's solution, embedded in paraffin, serially sectioned at 5 J.Lm, stained with hematoxylin and periodic acid-schiff, and examined by light microscopy at x 1000 magnification. During the culture period gonadal sex differentiation occurred and was recognizable by light microscopy.6 As the sex of the undifferentiated gonads cannot be determined at the beginning of the culture period, between four and six pairs of fetal gonads were used for each test to ensure that at least one pair of testes and one pair of ovaries were set up. Aspirates of 20 antral follicles and 2 cystic corpora lutea contained a sufficient amount of fluid (i.e., 0.5 ml) to be tested individually. The aspirates of 11 atretic follicles were pooled into two pools (Le., one pool made of aspirates from 9 follicles, the other from aspirates of 2 follicles) and set up in two tests. Thus, a total of 33 aspirates were tested in 24 separate tests. For these tests a total of 240 fetal mouse gonads (59 pairs of ovaries and 61 pairs of testes) were cultured. Vol. 41, No.3, March 1984 A presence of MIS activity in the test media will be expressed in the testes, because these will contain more meiotic germ cells than seen in the control testes. A presence of MPS in the test media is observed in the ovaries, which in that case will contain fewer meiotic germ cells, and germ cells in less advanced stages of meiosis than the control ovaries. In the present study, MIS activity is defined to be present in ovarian aspirates when the fetal testes cultured in it show significantly (P < 0.05) more meiotic germ cells than the control testes cultured in serum. MPS activity is defined to be present in the ovarian aspirates when the fetal ovaries cultured in it contain significantly fewer meiotic germ cells than the control ovaries cultured in serum. Differential Counts All the counts were scored in a blind fashion. In every fifth section of the gonads the number of nonmeiotic and meiotic germ cells (i.e., leptotene, zygotene, pachytene, and diplotene stages of me i osis) were counted. The number of non meiotic and meiotic germ cells in test and control gonads were compared with the chi-square test. The total numbers of germ cells in test and control gonads were compared with Wilcoxon's rank sum test. Semiquantitative Evaluation Evaluation of MIS and MPS activity in the fetal gonads was done independently of two other investigators, who did not know the results of the differential counts. The procedure of this method is as follows: the relative frequency of nonmeiotic and meiotic germ cells is evaluated by looking on every section of the fetal test and control gonads. The most frequently occurring germ cell stage is given a score of 3, and the less frequent stages, scores of 2 or 1, and a score of 0 if that stage is absent. MIS activity is present when more meiotic germ cells are seen in test than in control gonads. The intensity of' the activity is rated +, + +, or MIS. A + MIS is given if the leptotene stages are more frequent in test testes, compared with the control testes, and are given higher scores than other meiotic stages. A + + MIS is given when the meiotic process has advanced further and the zygotene stages now represent a major part of the meiotic germ cells. A MIS is given when all or almost all germ cells are in meiosis and the major part has ad- Westergaard et al. MIS in human ovarian follicles 379

4 -:+ vanced to zygotene, pachytene, and diplotene stages. MIS activity is absent if the distribution of germ cell stages is similar for test and control testes and is rated with a -. A similar scoring system is applied on the test and control ovaries for evaluation of MPS activity, i.e., when the meiotic process is less advanced in the test ovaries than in the control ovaries. STEROID HORMONE ASSAYS OF THE FOLLICULAR FLUIDS P, androstenedione (il4a), testosterone (T), and E2 in the fluid of the individual follicles were measured by radioimmunoassay as reported elsewhere. 15 All assays were performed in triplicate, and the intraassay and interassay coefficients of variation for each steroid were < 15%. CLASSIFICATION OF FOLLICLES The follicles were classified as healthy atretic and healthy dominant according to the stage of the menstrual cycle in which they were aspirated, their size, and their content of steroids. 15, 16 A healthy follicle is characterized by having an intrafollicular concentration of E2 of> 200 ng/ml and a il4 A/E2 ratio of < 4. An atretic follicle is characterized by having an intrafollicular concentration of E2 of < 200 ng/ml and a il4 AlE 2 ratio of > 4. A dominant follicle is assumed to be the one destined to ovulate. It is characterized by being aspirated in LF and by being the largest healthy follicle in the pair of ovaries from which it was aspirated. RESULTS MIS activity was found only in follicles aspirated in the LF phase. No MPS activity was detectable in any of the aspirates. The diameters, classification, and steroid concentrations of the follicles and the result of the MIS/MPS tests are summarized in Table 1 for the 15 follicles aspirated in LF and in Table 2 for the remaining 16 follicles and 2 corpora lutea, which were aspirated in phases other than LF. In the two cases (follicles 21, 22, and 23 to 31) in which pools of follicular fluids were tested for MIS and MPS activity, diameters and steroid concentrations are given as the means of the individual follicular values. All of the 15 follicles aspirated in LF were classified as healthy. Twelve of the follicles were dominant, whereas the remaining 3 (9, 11, and 12) were second largest and situated in either the same or the contralateral ovary as dominant folliad, dominant healthy follicle; H, nondominant healthy follicle. btest gonad, undifferentiated fetal mouse testis cultured in 15% follicular fluid. CMeiotic cells, germ cells in the leptotene, zygotene, pachytene, or diplotene stages of the meiotic division. dcontrol gonad, undifferentiated fetal mouse gonad cultured in 15% human serum of peripheral blood. es, significant difference (P < 0.05) between the number of meiotic germ cells in test and control gonads; NS, no significant difference between the number of meiotic germ cells in test and control gonads. 380 Westergaard et ai. MIS in human ovarian follicles Fertility and Sterility

5 Table 2. Diameter, Cycle Phase, Classification, and Intrafollicular Concentration ofp, ~4A, T, E 2, and MIS Activity of 16 Follicles and 2 Corpora Lutea Aspirated in Phases Other than LF of the Menstrual Cycle Differential counts Steroid bormones Semiquanti Follicle Diam Cycle Control gonads': tative eval no. eter pbase" ~~'::t. P 114A T E2 Te:e1~ti:g/~~~o. no. meioticl uation of nonmeiotic no. nonmeiotic MIS activo germ cells germ cells ity (% meiotic) (%meiotic) mm nglml nglml nglml nglml 16 9 MF H /308 (28) 30/139 (22) (NS) EF H < /32 (25) 19/100 (19) (NS) EF A /67 (15) (12) (NS) MF A < <1 29/141 (21) 12/61 (20) (NS) LL A < < 7.5 <1 6/30 (20) 7/46 (15) (NS) 21,22 11 MF,EL A /167 (13) (14) (NS) EF,MF, A /56 (14) 7/56 (13) (NS) LL 32 ML CL < /47 (13) 7/42 (17) (NS) 33 ML CL < /62 (6) 4/48 (8) (NS) aef, MF, LF, early, mid, and late follicular phase; EL, ML, LL; early, mid, and late luteal phase. bh, healthy follicle, A, atretic follicle, CL, corpus luteum. CTest gonad, undifferentiated fetal mouse testis cultured in 15% follicular fluid. dmeiotic cells, germ cells in the leptotene, zygotene, pachytene, or diplotene stages of the meiotic division. 'Control gonad, undifferentiated fetal mouse gonad cultured in 15% human serum of peripheral blood. cles 8, 2, and 5, respectively. MIS activity was found to be present in 12 (75%) of the 15 LF follicles (Table 1). Of the 16 follicles aspirated in phases other than LF, 2 were classified as healthy and the remainder as atretic. Neither MIS nor MPS activity was found in any of these samples (Table 2). Tables 1 and 2 also show the results of the evaluation of MIS activity by differential counts and by the semiquantitative method. In follicles 1 to 12 significantly more meiotic germ cells were counted in the test testes than in the control testes (P < 0.05). For the remaining follicles and the two corpora lutea, the differences between test and control testes are insignificant. The semiquantitative evaluation confirmed the presence of MIS activity in follicles 1 to 12 and the absence of MIS activity in the remainder. The total number of germ cells (nonmeiotic + meiotic) counted in test testes (median, 123; range, 30 to 308) did not differ significantly from that counted in the control testes (median, 108; range, 42 to 482; P > 0.05). In Tables 3 and 4 details concerning the differential counts are shown. Table 3 shows the mean percentages ± standard error of the mean (SEM) of the nonmeiotic and meiotic stages of germ cells in the control testes and in the test testes that were cultured in fluid with MIS activity (follicles 1 to 12) and without MIS activity (follicles 13 to 33). The percentages of nonmeiotic, leptotene, and zygotene stages are significantly different (P < 0.01) between the first group and the latter two, which do not differ significantly from one another. The differential counts showed that the number of meiotic germ cells as a percentag~ of the total number of germ cells varied betwee. '- 99% and Table 3. Percentages (Mean ± SEM) of Nonmeiotic Germ Cells and Germ Cells in the Different Stages of Meiosis in Fetal Mouse Testes Cultured in Human Follicular Fluid or in Human Serum Culture media No. of tests Nonrneiotic Leptotene Germ cell stages Zygotene Pachytene Diplotene % Fluid of folliclesa ± 4SC 1-12 Fluid of folliclesb ± 2"' Human sera ± I"' afollicles 1 to 12 contain follicular fluid with MIS activity. bfollicles 13 to 33 contain follicular fluid without MIS activity. cs, s', indicate that the differences are statistically significant (P < 0.01). % 34 ± 2" 16 ± 2"' 14 ± I"' % % % 15 ± 2" 2 ± ± I"' ± I"' 0 0 Vol. 41, No.3, March 1984 Westergaard et a1. MIS in human ovarian follicles 381

6 Table 4. Percentages (Mean ± SEM) of Nonmeiotic Germ Cells and Germ Cells in the Different Stages of Meiosis in Fetal Mouse Ovaries Cultured in Human Follicular Fluid or in Human Serum Culture media No. of tests Nonmeiotic Germ cell stages Leptotene Zygotene Pachytene Diplotene Fluid of follicles ± Human sera ± 0.07 % % % % % 2.0 ± ± 2 78 ± ± ± ± 2 76 ± ± % in all test and control ovaries, with no significant differences in any of the 24 tests (Table 4). Thus, no MPS activity could be demonstrated in any of the aspirates, a finding which was confirmed by the results of the semiquantitative evaluation. The total number of germ cells in test ovaries (median, 398; range, 31 to 1367) did not differ significantly from the number counted in control ovaries (median, 321; range, 39 to 1892; P > 0.05). Table 4 shows the mean ± SEM percentages of the nonmeiotic and meiotic stages of the germ cells in test and control ovaries in all of the 24 tests. There are no significant differences between any of the stages between tests and controls. DISCUSSION The results of this study show that the follicular fluid of healthy human ovarian follicles aspirated in the LF phase promotes the onset of meiosis in fetal mouse testes in vitro. Twelve of the 15 LF follicles were classified as dominant and can be considered preovulatory, whereas the remaining 3 were next to largest. Because twin ovulations during the natural cycle are not uncommon,17 and because the size and hormonal contents of the 3 nondominant LF follicles are comparable to those of the 12 dominant LF follicles, it is possible that these 3 follicles also were destined to ovulate, i.e., preovulatory. Thus, MIS activity was demonstrated in the aspirates of 75% (12 of 15) of preovulatory follicles. No follicle, healthy or atretic, in phases other than LF showed any MIS activity, and the same was the case for the two corpus luteum aspirates. It is commonly agreed that the resumption of meiosis in preovulatory oocytes is dependent upon a preovulatory surge of gonadotropins. 1, 2 However, it is also believed that LH and/or FSH does not act directly on the oocyte but that its effects are mediated by one or more local substances, the nature of which is still uncertain (for a review, see Edwards 17 ). 382 Westergaard et al. MIS in human ovarian follicles Unfortunately, in the present study an exact dating of the LH surge was not possible, because only one peripheral blood sample was available from each patient for gonadotropin assay. In large healthy follicles that have been exposed to the preovulatory LH surge one might expect high levels of intrafollicular P. 1S, 19 The preovulatory follicles that showed MIS activity had highly variable levels of P, suggesting that MIS activity may not be related to a certain degree offollicular luteinization. In the preovulatory follicles devoid of MIS activity there is a tendency toward lower concentrations of Ll4 A and/or T than in those with MIS activity (Table 1). However, the number of follicles is too small to draw any conclusion from this finding. The intrafollicular concentrations of E2 showed less variation among the preovulatory follicles than the other steroids and with little difference between follicles with and without MIS activity. Thus, there seems to be no certain concentration of a single steroid or a composition of steroids in the follicular fluid that could serve as a marker for the presence of MIS activity. This, however, does not mean that the steroids are unimportant in relation to MIS activity. Whether MIS activity reflects the presence of one or more substances in the follicular fluid is uncertain. Preliminary purification studies have shown that the MIS has a molecular weight of < 5000 daltons and has the characteristics of a lipid. MIS may have some resemblance to P, because P has a weak MIS activity when added to the culture medium in concentrations of 2 ng/ml. However, with higher levels (100 ng/ml), the meiosis-inducing effect disappears.2o In the present study the follicular fluids showing MIS activity, as well as those of the preovulatory follicles not showing MIS activity, all had P concentrations well above 100 ng/ml, and the 15% dilutions, well above 2 ng/ml. Moreover, the aspirates of the two corpora lutea in which P is the dominating steroid did not show MIS activity. Thus, there seems to be no indication that P in vivo acts as the MIS. Further Fertility and Sterility

7 attempts to chemically characterize MIS are needed and are in progress. The results of the differential counts confirm the reliability of the much less time-consuming semiquantitative method of evaluating MIS and MPS activity. The data also show that a certain spontaneous meiotic activity is found in the control testes. The reason for this may be that the testes are cultured together with their mesonephroi. The occurrence of spontaneous meiosis in fetal testes cultured together with their mesonephric tissue has previously been described in another mouse strain (Bagg strain) in which the mesonephric tissue had a pronounced feminizing influence on the testis in vitro.21 The present study failed to demonstrate MPS activity in any of the aspirates from ovarian follicles or corpora lutea, whether healthy or atretic, preovulatory or nonovulatory. However, this does not mean that MPS is absent in all healthy follicles, because only large healthy follicles were tested in the present study. In a previous study,5 we reported that MPS-like activity could be identified in retentates of ultrafiltrated fluid of some large follicles but not in the unconcentrated fluids. Perhaps MPS is present in large follicles in decreasing amounts and needs to be concentrated before being detectable in our in vitro test system. Similarly, OMI activity of porcine follicular fluid appears to decline during the course of follicular development.22 Moreover, OMI-inactive follicular fluid oflarge follicles presented OMI activity in a low-molecular-weight fraction.13 Whether OMI and MPS are related to a parent compound is unknown, but not probable, since OMI has a molecular weight < 2000 daltons, and MPS has a molecular weight> 5000 daltons.2o In both cases their activities are low or absent in large follicles. However, since none of the follicles in the present study showed MPS activity, it can hardly be claimed that MPS plays an active role in preventing the resumption of oocyte meiosis in the human ovary. In conclusion, the data from the present study are consistent with the notion that the resumption of meiosis in the preovulatory oocyte requires active induction by an MIS in the follicular fluid rather than merely the removal of an MPS. Acknowledgments. We with to thank Mrs. Jette Lise Hansen, Mrs. Inga Husum, Mrs. Ulla Larsen, and Mrs. Ros Sapawi for skillful technical assistance, and Mrs. Else Wulff for typing the manuscript. Vol. 41, No.3, March 1984 REFERENCES 1. Edwards RG: Maturation in vitro of human ovarian 00- cytes. Lancet 2:926, Ayalon D, Tsafriri A, Lindner HR, Cordova T, Harell A: Serum gonadotrophin levels in proestrus rats in relation to the resumption of meiosis by oocytes. J Reprod Fertil 31:51, Byskov AG: Regulation of meiosis in mammals. Ann BioI Anim Biochim Biophys 19:1251, Fayer AB, Schneider JA, McCall D, Ances IG, Polakis SE: The induction of meiosis by ovaries of newborn hamsters and its relation to the action of the extraovarian structures in mesovarium (rete ovarii). Ann BioI Anim Biochim Biophys 19(4B):1273, Yding Andersen C, Westergaard L, Byskov AG, Grinsted J, Lauritsen JG: Possible regulation of meiosis in the human ovary. In Proceedings of the Third World Congress in Human Reproduction, Edited by K Semm, L Mettler. Amsterdam, Excerpta Medica, 1981, p Byskov AG, Saxen L: Induction of meiosis in fetal mouse testis in vivo. Dev BioI 52:193, WS, Baker TG: Initiation and control of meiosis in hamster gonads in vitro. J Reprod Fertil 48:399, Byskov AG: The meiosis-inducing interaction between germ cells and rete cells in the fetal mouse gonads. Ann BioI Anim Biochim Biophys 18(2B):327, Grinsted J, Byskov AG, Andreasen MP: Induction of me i osis in fetal mouse testis in vitro by rete tissue from pubertal mice and bulls. J Reprod Fertil 56:653, Grinsted J, Byskov AG: Meiosis-inducing and meiosispreventing substances in human male reproductive organs. Fertil Steril 35:199, Parvinen M, Byskov AG, Yding Andersen C, Grinsted J: Is the spermatogenic cycle regulated by MIS and MPS?. Ann NY Acad Sci 383:483, Tsafriri A, Channing CP: An inhibitory influence of granulosa cells and follicular fluid upon oocyte meiosis in vitro. Endocrinology 96:922, Tsafriri A, Dekel N, Bar-Ami S: The role of oocyte maturation inhibitor in follicular regulation of oocyte maturation. J Reprod Fertil 64:541, , Byskov AG: Regulation of the initiation of meiosis in fetal gonads. Int J Androl (Suppl) 2:29, Westergaard L, McNatty KP, Christensen I, Larsen JK, Byskov AG: Flow cytometric desoxyribonucleic acid analysis of granulosa cells aspirated from human ovarian follicles: a new method to distinguish healthy and atretic follicles. J Clin Endocrinol Metab 55:693, McNatty KP, Moore-Smith D, Makris A, Osathanondh R, Ryan KJ: The microenvironment of the human antral follicle: interrelationships among the steroid levels in antral fluid, the population of granulosa cells and the status of the oocyte in vivo and in vitro. J Clin Endocrinol Metab 49:851, Edwards RG: The adult ovary. In Conception in the Human Female. London, Academic Press, 1980, p McNatty KP: Follicular determinants of corpus luteum function in the human ovary. In Ovarian Follicular and Corpus Luteum Function, Edited by CP Channing, J Marsh, WA Sadler, New York, Plenum Press, 1979, p Brailly S, Gougeon A, Milgrom E, Bomsel-Helmreich 0, Papiernik E: Androgens and progestins in the human ovarian follicles: differences in the evolution of pre- Westergaard et ai. MIS in human ovarian follicles 383

8 ovulatory, healthy non-ovulatory and atretic follicles. J Clin Endocrinol Metab 53:128, Yding Andersen C, Byskov AG, Grinsted J: Partial purification of the meiosis-inducing substance (MIS). In Development and -Function of Reproductive Organs, Edited by AG Byskov, H Peters. Amsterdam, Excerpta Medica, 1981, p Byskov AG, Grinsted J: Feminizing effect of mesonephros on cultured differentiating mouse gonads and ducts. Science 212:817, Stone SL, Pomerantz SH, Schwartz-Kripner A, Channing CP: Inhibition of oocyte maturation from a porcine follicular fluid: further purification and evidence for reversible action. BioI Reprod 19:585, Westergaard et al. MIS in human ovarian follicles Fertility and Sterility