MAMMALIAN spermatogenesis in vivo is a long and

0021-972X/98/$03.00/0 Vol. 83, No. 12 Journal of Clinical Endocrinology and Metabolism Printed in U.S.A. Copyright 1998 by The Endocrine Society Human Spermatogenesis in Vitro: Respective Effects of Follicle-Stimulating Hormone and Testosterone on Meiosis, Spermiogenesis, and Sertoli Cell Apoptosis JAN TESARIK, MAURIZIO GUIDO, CARMEN MENDOZA, AND ERMANNO GRECO Laboratoire d Eylau (J.T.), 75116 Paris, France; Department of Obstetrics and Gynecology (M.G.), Università Cattolica del Sacro Cuore, 00168 Rome, Italy; Department of Biochemistry and Molecular Biology (C.M.), University of Granada Faculty of Sciences, 18071 Granada, Spain; and Center of Reproductive Medicine, European Hospital (E.G.), 00149 Rome, Italy ABSTRACT In spite of the availability of abundant data about in vitro spermatogenesis in laboratory animals, studies on human in vitro spermatogenesis are scarce. This study employed a relatively simple culture system, involving all cell types of seminiferous tubules, to analyze the effects of FSH and testosterone (T) on different characteristics of human germ and Sertoli cells in culture. By using fluorescence in-situ hybridization, we show that in vitro reduction of germ cell ploidy can be stimulated by FSH but not by T. FSH, but not T, also induced unexpectedly rapid (24 48 h) morphological changes resembling spermiogenesis, although individual changes (spermatid nucleus condensation and protrusion, cell body elongation, and flagellar MAMMALIAN spermatogenesis in vivo is a long and complex process, which is controlled by multiple and mutually interacting mechanisms (1 3). There has been much effort to develop cell culture systems with which spermatogenesis could be achieved in vitro. Early studies made use of relatively simple systems in which whole segments of seminiferous tubules were maintained in culture for several days (4 7). Later on, the culture systems for in vitro spermatogenesis were progressively refined to better assess individual cellular and molecular interactions (1). Although these refinements facilitated the analysis of spermatogenesis mechanisms, the yield of germ cells undergoing developmental progression and the speed of this progression did not seem to be improved by these modifications. In fact, the contrary was true in most cases. In a preliminary series of experiments, we performed cultures of testicular tissues sampled from patients suffering from complete spermiogenesis failure (8), in conditions bearing some similarity to the above early animal studies. Surprisingly, round spermatids from some patients overcame the spermiogenesis block, and some of them showed signs of elongation (Tesarik et al., work in preparation). In this study, we address the question of whether the efficacy of this culture system can be improved by supplementing the culture medium with FSH and testosterone (T), the two hormones known to be directly implicated in the Received May 5, 1998. Revised August 14, 1998. Accepted August 31, 1998. Address all correspondence and requests for reprints to: Dr. Jan Tesarik, Laboratoire d Eylau, 55 Rue Saint-Didier, 75116 Paris, France. growth) proceeded in an uncoordinated way and mostly resulted in the development of abnormal forms of elongated spermatids. Though ineffective alone, T potentiated the effects of FSH on meiosis and spermiogenesis. These effects of T were probably caused by the prevention of Sertoli cell apoptosis, an effect that could not be mimicked by FSH. These data show that, in the presence of high concentrations of FSH and T, human spermatogenesis can proceed in vitro with an unusual speed, but the resulting gametes are morphologically abnormal. The potential practical relevance of these findings to assisted reproduction remains to be assessed. (J Clin Endocrinol Metab 83: 4467 4473, 1998) regulation of spermatogenesis by acting at Sertoli cells (9, 10). The effects of added hormones on the progression of meiosis, on spermiogenesis, and on the development of apoptotic DNA damage of Sertoli cells were analyzed. Materials and Methods Source and preparation of testicular cells Testicular biopsy was performed in 18 men suffering from obstructive azoospermia to obtain spermatozoa for intracytoplasmic sperm injection (ICSI). Pieces of testicular tissue were placed in GAMETE-100 medium (Scandinavian IVF, Goteborg, Sweden) and disintegrated mechanically by stretching between two microscope slides, followed by repeated aspiration into a 1-mL tuberculin syringe. Large tissue pieces were removed, and the remaining cell suspension was homogenized and distributed among individual treatment groups. Testicular cells in these suspensions either were isolated or formed small cell clusters (Fig. 1A). The appearance of these cell clusters was similar after 48 h of in vitro culture (see below), except for a reduction of Sertoli cells occurring in media that were not supplemented with T (Fig. 1B). For both fresh and cultured testicular cells, aliquots of this suspension were incubated at 37 C with 1000 U/mL collagenase IV (Sigma Chemical Co., C-5138, St. Louis, MO), as described (11), to achieve complete disintegration of the cell clusters into single cells. This preparation was used for quantitative evaluation of the proportion of individual types of germ cells but not for further culture. The same enzymatic treatment was applied to samples in all experimental groups (see below) at the end of culture, before final evaluation. In vitro culture protocols All cultures were carried out in GAMETE-100 medium (Scandinavian IVF), in a water bath set to 30 C. Recombinant human FSH (Puregon, Organon, Oss, The Netherlands) was added at final activity concentrations of 10 IU/L, 25 IU/L, 50 IU/L, or 100 IU/L. Water-soluble T (Sigma Chemical Co., T-5035) was added at a concentration of 1 mol/l (cal- 4467

4468 TESARIK ET AL. JCE&M 1998 Vol 83 No 12 Evaluation of spermatid cytology Collagenase-dissociated cells were smeared onto microscope slides, fixed with ethanol, and stained with the use of the Papanicolaou method (14). Normal spermiogenesis forms (Fig. 3) were classified as Sa, Sb1, Sb2, Sc, Sd1, and Sd2, according to the criteria described by de Kretser and Kerr (15). In addition, the occurrence of three abnormal forms of human spermatogenic cells (Fig. 3), termed Saf, Sbp, and Scp, was also evaluated. The Saf (Sa with flagellum) stage only differed from Sa by the presence of a flagellum, whereas the cell did not show any signs of elongation; and the nucleus was still round, uncondensed, or only slightly condensed and was surrounded by a continuous rim of cytoplasm. Sbp (Sb pathological) spermatids still retained the round cell shape, but the nucleus was already condensed, elongated, and protruding at one pole of the cell. Scp (Sc pathological) spermatids possessed a flagellum but still retained the round cell shape. Unlike Saf, Scp spermatids also had a more- or less-condensed and protruding nucleus (Fig. 3). Aliquots of the same samples as used for cytological analysis were subjected to supravital staining with eosine (14) to determine cell viability. Evaluation of Sertoli cell apoptosis The presence of apoptosis-related DNA strand breaks in Sertoli cell nuclei was evaluated by terminal deoxynucleotidyl transferase-mediated deoxyuridine 5-triphosphate nick end labeling (TUNEL) using the Cell Death Detection Kit (Boehringer, Mannheim, Germany) according to manufacturer s instructions. Quantitative evaluation and statistics Two hundred cells were evaluated in different types of smear preparation (FISH, immunocytochemistry, cytology, and TUNEL) for each culture period and each kind of hormone supplementation. Testicular cells, recovered from one patient, represented one replicate. Percentages were calculated for each cell category, as defined in individual experiments. Quantitative data (mean sem) were analyzed by 2 and Kruskal-Wallis tests. FIG. 1. Appearance of a typical cell cluster after mechanical disintegration of testicular tissue, as observed without fixation in culture medium. A, Cluster of round germ cells among which granulated Sertoli cell cytoplasm (arrow) is seen at the beginning of culture; B, the same cluster after 48 h of culture in medium lacking T. Note the disappearance of the Sertoli cell cytoplasm. Bar, 30 m. culated according to the weight proportion of T in the water-soluble complex). Evaluation of germ cell ploidy The ploidy of germ cells, at different times of in vitro culture, was evaluated using a method combining fluorescence in-situ hybridization (FISH), using a digoxygenin-labeled probe for human chromosome 15, with immunocytochemical detection of proacrosin; these experiments were performed with smears of collagenase-dissociated cells and with the same reagents and protocols as described (12). The proacrosinspecific 4D4 monoclonal antibody was a generous gift from Dr. Denise Escalier (University of Paris, France). This antibody recognizes human spermatogenic cells from the pachytene primary spermatocyte stage onward (13). The FISH signal was revealed with Texas red-conjugated antidigoxygenin antibody, whereas the proacrosin immunoreactivity was visualized with fluorescein isothiocyanate-labeled antimouse IgG secondary antibody (Sigma Chemical Co.), and cell nuclei were counterstained with 4,6-diamidino-2-phenylindol (Fig. 2). Results Effect of FSH on germ cell meiosis and spermiogenesis After 24 h of culture with different concentrations of FSH, no effect on the ploidy of 4D4-reactive germ cells was observed with an FSH concentration of 10 IU/L. In contrast, there was a significant decrease in the percentage of cells showing four hybridization spots (4N tetraploid) and an increase in the percentage of cells showing one spot (1N haploid) with higher FSH concentrations. Beginning with 50 IU/L FSH, an increase in cells with 2 spots (2N haploid) was also observed, whereas no differences were detected between FSH concentrations of 50 IU/L and 100 IU/L (Table 1). The lowest FSH concentration at which an effect on spermiogenesis was observed was also 25 IU/L, producing a shift in the proportion of elongated spermatids towards more advanced stages of spermiogenesis and, especially, an increase in the abnormal forms of elongated spermatids Sbp and Scp (Table 2). Further increasing FSH concentration did not augment this effect (Table 2). No differences in cell viability were detected between individual groups (Table 2). Preliminary experiments (n 6) did not show any increase in T concentration in culture media after 1 and 2 days of testicular cell culture, either in the absence of FSH or in the presence of FSH at concentrations of 25 IU/L and 50 IU/L (data not shown).

HUMAN SPERMATOGENESIS IN VITRO 4469 FIG. 2. Smeared testicular cells processed by a method allowing a simultaneous detection of ploidy and of a germline marker in the same cells. Cell ploidy is revealed by pink FISH spots, each corresponding to a chromatid of chromosome 15 visualized with a digoxygenin-labeled probe and Texas red-conjugated antidigoxygenin antibody, in the nucleus (counterstained blue with 4,6-diamidino-2-phenylindol). The germline marker (proacrosin) is visualized by immunocytochemistry using 4D4 antiproacrosin monoclonal antibody and fluorescein isothiocyanatelabeled antimouse IgG (yellow fluorescence). All figure parts are printed at the same magnification. Bar (D), 10 m. A, 4N tetraploid cell (spermatogonium or a somatic cell), showing four pink FISH spots and no proacrosin immunoreactivity. Because the antiproacrosin antibody used only detects proacrosin from the pachytene stage onwards, and in view of the size of the cell nucleus, this cell could be a spermatogonium or a somatic cell. B, 2N haploid secondary spermatocyte. The 2N haploid status of the cell nucleus is indicated by two pink FISH spots. The two proacrosin-immunoreactive compact structures and the fluffy area between them represent, respectively, two developing proacrosomal vesicles and a Golgi region in which the acrosomal contents are assembled. The seemingly nuclear localization of these structures is caused by projection of these cytoplasmic structures over the flattened nucleus in the smear preparation. C, 1N haploid round spermatid. The single pink FISH spot corresponds to the 1N haploid status of the cell. Proacrosin immunoreactivity is accumulated in a single proacrosomal vesicle. D, 1N haploid late elongated spermatid, showing a single pink FISH spot and proacrosin immunoreactivity localized in a cap-like acrosomal structure. Separated and combined effects of FSH and T on germ cell meiosis and spermiogenesis Supplementation of culture medium with 1 mol/l T did not produce any effect on either germ cell meiosis (Table 3) or spermiogenesis (Table 4). However, when added together with FSH, T potentiated the effect of FSH, both on meiosis (Table 3) and on spermiogenesis (Table 4), as early as after 24 h of culture, and these effects were even more marked after 48 h of culture. The potentiation by T of the FSH effect on meiosis was only detected as a decrease in the percentage of 4N tetraploid cells and as an increase in the percentage of 1N haploid cells, without any detectable change in the prevalence of 2N haploid cells (Table 3). The potentiation by T of spermiogenesis-related changes was reflected by an increased percentage of the Sd2 normal forms and Scp abnormal forms of spermatids, whereas the prevalence of the Sbp abnormal forms showed a slight decrease after 24 h of culture (Table 4). No changes were detected in other spermatid forms. Unlike the 24-h culture interval, the survival of cells after

4470 TESARIK ET AL. JCE&M 1998 Vol 83 No 12 FIG. 3. Stages of normal (A, C, and E) and abnormal (B, D, and F) spermiogenesis, as visualized in ethanol-fixed and Papanicolaou-stained cell smears. Erythrocytes (e) can also be seen in D and F. All figure parts are printed at the same magnification. Bar (F), 20 m. A, Sa spermatid with a still uncondensed nucleus (n) surrounded by a continuous rim of cytoplasm. B, Saf spermatid with a centrally located, partly condensed nucleus (n) and a long flagellum (arrows). C, Sb2 spermatid with a slightly protruding, partly condensed nucleus (n) and an elongated cell body with a short flagellum (arrow). D, Sbp spermatid with a highly condensed, protruding nucleus (n) and a still round cell body without flagellum (arrow). E, Sd1 spermatid with a highly condensed, protruding nucleus (n) and an elongated cell body with a long flagellum (arrows). F, Scp spermatid with a partly condensed, slightly protruding nucleus (n), a still round cell body, and a long flagellum (arrows). TABLE 1. Effects of different concentrations of FSH, added to medium used for 24-h in vitro culture, on the prevalence of postzygotene (4D4 mab-immunoreactive) germ cells with different degrees of ploidy as reflected by the number of autosome-related FISH spots FSH concentration (IU/L) % 4D4-reactive cells with different numbers of FISH spots a 4 spots 2 spots 1 spot 0 43.3 5.1 b 9.1 1.0 b 47.6 5.2 b 10 42.9 5.2 b 9.8 1.2 b 47.3 5.1 b 25 30.4 3.3 c 11.6 1.3 b 58.0 5.4 c 50 20.9 2.7 d 15.0 1.9 c 64.1 6.6 d 100 20.2 2.6 d 15.3 1.8 c 64.5 6.5 d a Data are mean SEM of 10 replicates performed with samples from 10 different patients. Two hundred 4D4-positive cells were evaluated in each replicate. Cells with no spot, with 3 spots, and with more than 4 spots represented together less than 10% of 4D4-positive cells in all replicates and were not included in this evaluation. bcd Values with different superscript within columns are different (P 0.05). 48 h of culture was also slightly (but significantly) improved in the presence of T, irrespective of the presence of FSH in culture medium (Table 4). Separated and combined effects of FSH and T on Sertoli cell apoptosis Unlike FSH, T had a strong inhibitory effect on the progression of apoptotic DNA fragmentation in cultured Sertoli cells, irrespective of whether it was added alone or together with FSH (Table 5). This effect was especially marked after 48 h of culture. The combined addition of FSH and T did not potentiate the antiapoptotic effect of T on Sertoli cells, as compared with T alone (Table 5). Discussion The results of this study suggest that FSH stimulates both germ cell meiosis and spermiogenesis during in vitro culture of human testicular cells. This conclusion can be made after taking into account all precautions related to the specific nature of our experimental system. One (and probably the most important) precaution is related to the open and dynamic character of the system under study, contrasting with the static nature of the descriptive criteria to which the changing elements of this system had to be accommodated. However, a synthetic view of the data described in this study makes the conclusion of accelerated germ cell meiosis and spermiogenesis caused by the action of FSH the most consistent interpretation. First of all, the marked decrease in 4N diploid germ cells and the increase in 1N haploid germ cells during culture with FSH were not accompanied by a corresponding increase in either 2N haploid forms or in the earliest stages of spermiogenesis. Because no measurable decrease in the overall cell viability occurred during 2 days of culture with FSH and T, these observations must reflect an acceleration by FSH of both the first and the second meiotic division, with a concomitant depletion of the earliest forms of 1N haploid round spermatids by a simultaneous acceleration of spermiogenesis. In fact, a major shift in the prevalence of individual stages of spermatogenesis after 2 days of culture was from 4N primary spermatocytes towards the Sd2 late form of spermiogenesis, as well as towards the Scp abnormal elongatedspermatid form representing a typical product of in vitro spermiogenesis in the presence of FSH. Further, T did not show any similar effect. Finally, recombinant FSH was used in this study, thus excluding the possibility of Leydig cell stimulation by LH contaminating urinary FSH preparations.

HUMAN SPERMATOGENESIS IN VITRO 4471 TABLE 2. Effects of different concentrations of FSH, added to medium used for 24-h in vitro culture, on the occurrence of different normal and abnormal stages of spermiogenesis and on cell viability FSH concentration (IU/L) Prevalence of selected stages of spermiogenesis a % Normal forms % Abnormal forms Sa Sb1 Sb2 Sc Sd1 Sd2 Saf Sbp Scp Cell Viability (%) 0 2 1 b 1 1 b 1 1 b 1 1 b 7 2 b 21 3 b 4 2 b 1 1 b 1 1 b 81 11 b 10 3 1 b 1 1 b 1 1 b 1 1 b 7 2 b 23 3 b 3 1 b 1 1 b 2 1 b 82 11 b 25 3 1 b 1 1 b 1 1 b 1 1 b 9 3 b 32 3 c 2 1 b 5 1 c 8 2 c 86 11 b 50 4 2 b 2 1 b 1 1 b 1 1 b 10 3 b 35 4 c 2 1 b 5 2 c 9 3 c 85 10 b 100 3 1 b 1 1 b 1 1 b 1 1 b 9 2 b 34 4 c 2 1 b 6 2 c 9 3 c 88 11 b a Data are mean SEM of 10 replicates performed with samples from 10 different patients. Two hundred 4D4-positive cells were evaluated in each replicate. Percentages of postzygotene primary spermatocytes and of secondary spermatocytes are not included. bc Values with different superscript within columns are different (P 0.05). TABLE 3. Separated and combined effects of FSH (50 IU/L) and T (1 M), added to medium used for 24-h and 48-h in vitro culture, on the prevalence of postzygotene (4D4 mab-immunoreactive) germ cells with different degrees of ploidy as reflected by the number of autosome-related FISH spots Component added Time of culture % 4D4-reactive cells with different numbers of FISH spots a 4 spots 2 spots 1 spot None 24 h 43.9 5.3 b 8.7 1.4 b 47.4 5.0 b FSH 24 h 26.1 2.8 c 14.5 1.6 c 59.4 6.1 c T 24 h 43.0 5.5 b 8.4 1.1 b 48.6 4.9 b FSH T 24 h 15.7 2.2 d 15.8 1.8 c 68.5 6.8 d None 48 h 41.2 4.4 b 10.3 1.2 b 48.5 5.5 b FSH 48 h 23.3 2.5 c 15.3 1.6 c 61.4 6.5 c T 48 h 42.4 2.3 b 10.6 1.3 b 47.0 4.4 b FSH T 48 h 10.9 1.2 e 13.0 1.5 c 76.1 7.6 d a Data are mean SEM of 10 replicates performed with samples from 10 different patients. Two hundred 4D4-positive cells were evaluated in each replicate. Cells with no spot, with 3 spots, and with more than 4 spots represented together less than 10% of 4D4-positive cells in all replicates and were not included in this evaluation. bcde Values with different superscript within columns are different (P 0.05). TABLE 4. Separated and combined effects of FSH (50 IU/L) and T (1 M), added to medium used for 24-h and 48-h in vitro culture, on the occurrence of different normal and abnormal stages of spermiogenesis and on cell viability Cell viability (%) Component added Time of culture (h) Normal forms Abnormal forms % Selected stages of spermiogenesis a Sa Sd2 Sbp Scp None 24 2 1 b 20 3 b 1 1 b 1 1 b 77 5 b FSH 24 3 1 b 32 4 c 5 1 c 8 2 c 79 5 b T 24 2 1 b 21 3 b 2 1 b 1 1 b 81 5 b FSH T 24 2 1 b 39 3 d 2 1 b 12 2 d 83 6 b None 48 1 1 b 21 3 b 1 1 b 2 1 b 65 5 c FSH 48 4 1 c 31 4 c 2 1 b 9 1 c 66 5 c T 48 2 1 b 19 3 b 1 1 b 2 1 b 80 6 b FSH T 48 4 1 c 44 5 d 1 1 b 19 2 d 80 6 b a Data are mean SEM of 10 replicates performed with samples from 10 different patients. Two hundred 4D4-positive cells were evaluated in each replicate. Percentages of postzygotene primary spermatocytes and of secondary spermatocytes are not included. Stages in which no differences were observed are not included. b,c,d Values with different superscript within columns are different (P 0.05). Our findings concerning the effects of FSH on in vitro germ cell meiosis and spermiogenesis and of the absence of T effects on these processes are unexpected, and they challenge the widespread idea, according to which T is the main hormone regulating mammalian spermiogenesis with only a marginal, if any, contribution of FSH (16 22). It has to be noted, however, that a relatively high concentration of FSH (25 IU/L) was needed to produce perceptible changes in both germ cell ploidy (Table 1) and spermatid morphology (Table 2). This concentration is higher than the normal FSH concentration in male blood plasma, although even higher concentrations were detected in some men who suffered from nonobstructive azoospermia but with whose testicular spermatozoa it was still possible to fertilize and achieve pregnancies after ICSI (23). As to T, the concentration chosen in this study (1 mol/l) was slightly higher than the physiological one in the seminiferous tubules. A previous study (24) demonstrated a significant suppression of human male germ cell apoptosis by T at concentrations of 1 mol/l and 0.1 mol/l. The higher of those two concentrations was given preference in this study to make more likely the achievement of release of an effective concentration of free T from the water-soluble complex (see Materials and Methods). The effects of FSH and T on in vitro spermatogonial proliferation

4472 TESARIK ET AL. JCE&M 1998 Vol 83 No 12 TABLE 5. Separated and combined effects of FSH (50 IU/L) and T(1 M), added to medium used for 24-h and 48-h in vitro culture, on apoptosis-induced Sertoli cell DNA fragmentation as revealed by TUNEL assay Component added Time of culture % TUNEL-positive Sertoli nuclei a None 24 h 28.1 3.4 b FSH 24 h 26.0 2.8 b T 24 h 19.3 2.2 c FSH T 24 h 18.5 2.0 c None 48 h 70.9 8.2 d FSH 48 h 68.8 7.7 d T 48 h 20.3 2.3 c FSH T 48 h 20.0 2.5 c a Data are mean SEM of 10 replicates performed with samples from 10 different patients. Two hundred Sertoli cell nuclei were evaluated in each replicate. bcd Values with different superscript within columns are different (P 0.05). and the transition of spermatogonia to spermatocytes were not addressed in this study, and they remain to be evaluated. These findings are even more surprising, in view of the unusual speed with which spermiogenesis was progressing in vitro, as compared with the in vivo timing (25). One possible explanation for this discrepancy is the abrogation, under in vitro conditions, of many developmental control checkpoints, which normally do not allow further progression of spermiogenesis until previous steps have been completed. The frequent development of abnormal spermatid forms may be the price to be paid for this developmental speed. The relevance, to normal human spermatogenesis, of the rapid changes in germ cell ploidy and spermatid populations during in vitro culture is not clear. Work is in progress to evaluate the fertilizing ability of in vitro-developed spermatids after their injection into oocytes (26). In agreement with our data, two other recent studies (27, 28) have also reported a rapid flagellar growth in human round spermatids in vitro. In this study, these changes were usually accompanied by nuclear condensation and protrusion, but only when effective concentrations of FSH were present in culture medium. This explains why these nuclear changes were not observed in the two previous studies (27, 28), in which the culture medium was not supplemented with FSH. The main effects of T observed in this study were related to the improvement of Sertoli cell survival in culture with the inhibition of the apoptotic pathway leading to DNA fragmentation. Although rat and human Sertoli cells seem to be relatively resistant to in vivo apoptosis (24, 29, 30), explanted and in vitro cultured Sertoli cells fall prey to apoptosis with a much greater ease (31). In agreement with the results of the present study, T has previously been shown to protect human Sertoli cells against experimentally induced in vitro apoptosis (24). Taken together, the results of this study suggest that human germ cells can undergo meiosis and spermiogenesis with an extraordinary speed when they are cultured in vitro in the presence of FSH, and that T can potentiate this FSH effect by preventing apoptosis in the cocultured Sertoli cells. The possibility of inducing in vitro spermatogenesis in men with spermatogenesis arrest and of using the resulting in vitro cultured germ cells for assisted reproduction is currently under investigation. Acknowledgment The authors wish to thank Dr. Denise Escalier for kindly providing samples of 4D4 monoclonal antibody. References 1. Kierszenbaum AL. 1994 Mammalian spermatogenesis in vivo and in vitro: a partnership of spermatogenic and somatic cell lineages. Endocr Rev. 15:116 134. 2. Sharpe RM. 1994 Regulation of spermatogenesis. In: Knobil E, Neil JD, eds. The physiology of reproduction. New York: Raven Press; 1363 1434. 3. Gnessi L, Fabbri A, Spera G. 1997 Gonadal peptides as mediators of development and functional control of the testis: an integrated system with hormones and local environment. Endocr Rev. 18:541 609. 4. Steinberger A, Steinberger E, Perloff WH. 1964 Mammalian testes in organ tubules. Exp Cell Res. 36:19 27. 5. Steinberger A. 1975 In vitro techniques for the study of spermatogenesis. In: Hardmann JG, O Malley BW, eds. Methods in enzymology. Vol 39. New York: Academic Press; 283 296. 6. Parvinen M, Wright WW, Phillips DM, Mather JP, Musto NA, Bardin CW. 1983 Spermatogenesis in vitro: completion of meiosis and early spermiogenesis. Endocrinology. 112:1150 1152. 7. Toppari J, Parvinen M. 1985 In vitro differentiation of rat seminiferous tubular segments from defined stages of the epithelial cycle: morphologic and immunolocalization analysis. J Androl. 6:334 343. 8. Amer M, Soliman E, El-Sadek M, Mendoza C, Tesarik J. 1997 Is complete spermiogenesis failure a good indication for spermatid conception? Lancet. 350:116. 9. Griswold MD. 1993 Action of FSH on mammalian Sertoli cells. In: Russel LD, Griswold MD, eds. The Sertoli cell. Clearwater, FL: Cache River Press; 493 508. 10. Sar M, Hall SH, Wilson EM, French FS. 1993 Androgen regulation of Sertoli cells. In: Russel LD, Griswold MD, eds. The Sertoli cell. Clearwater, FL: Cache River Press; 509 516. 11. Crabbé E, Verheyen G, Tournaye H, Van Steirteghem A. 1997 The use of enzymatic procedure to recover testicular germ cells. Hum Reprod. 12:1682 1687. 12. Mendoza C, Benkhalifa M, Cohen-Bacri P, Hazout A, Ménézo Y, Tesarik J. 1996 Combined use of proacrosin immunocytochemistry and autosomal DNA in situ hybridization for evaluation of human ejaculated germ cells. Zygote. 4:279 283. 13. Escalier D, Gallo J-M, Albert M, et al. 1991 Human acrosome biogenesis: immunodetection of proacrosin in primary spermatocytes and of its partitioning pattern during meiosis. Development. 113:779 788. 14. World Health Organization. 1992 WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction. 3rd ed. Cambridge: Cambridge University Press; 1 107. 15. de Kretser DM, Kerr JB. 1988 The cytology of the testis. In: Knobil E, Neill J, eds. The physiology of reproduction. New York: Raven Press; 837 932. 16. Huang HFS, Marshall GR, Rosenberg R, Nieschlag E. 1987 Restoration of spermatogenesis by high levels of testosterone in hypophysectomized rats after long-term regression. Acta Endocrinol (Copenh). 116:433 444. 17. Sun Y, Irby D, Robertson D, de Kretser D. 1989 The effects of exogenously administered testosterone on spermatogenesis in intact and hypophysectomized rats. Endocrinology. 125:1000 1010. 18. Sun Y, Wreford NG, Robertson DM, de Kretser DM. 1990 Quantitative cytological studies of spermatogenesis in intact and hypophysectomized rats: identification of androgen-dependent stages. Endocrinology. 127:1215 1223. 19. Awoniyi CA, Zirkin BR, Chandrashekar V, Schlaff WD. 1992 Exogenously administered testosterone maintains spermatogenesis quantitatively in adult rats actively immunized against gonadotropin-releasing hormone. Endocrinology. 130:3283 3288. 20. McLachlan RI, Wreford NG, Meachem SJ, de Kretser DM, Robertson DM. 1994 Effects of testosterone on spermatogenic cell populations in the adult rat. Biol Reprod. 51:945 955. 21. O Donnell L, McLachlan RI, Wreford NG, Robertson DM. 1994 Testosterone promotes the conversion of round spermatids between stages VII and VIII of the rat spermatogenic cycle. Endocrinology. 135:2608 2614. 22. O Donnell L, McLachlan RI, Wreford NG, de Kretser DM, Robertson DM. 1996 Testosterone withdrawal promotes stage-specific detachment of round spermatids from the rat seminiferous epithelium. Biol Reprod. 55:895 901. 23. Gil-Salom M, Remohi J, Minguez Y. 1995 Pregnancy in an azoospermic patient with markedly elevated serum follicle-stimulating hormone levels. Fertil Steril. 64:1218 1220. 24. Erkkilä K, Henriksén K, Hirvonen V, et al. 1997 Testosterone regulates

HUMAN SPERMATOGENESIS IN VITRO 4473 apoptosis in adult human seminiferous tubules in vitro. J Clin Endocrinol Metab. 82:2314 2321. 25. Heller CG, Clermont Y. 1964 Kinetics of the germinal epithelium in man. Recent Prog Horm Res. 20:545 575. 26. Tesarik J, Mendoza C. 1996 Spermatid injection into human oocytes. I. Laboratory techniques and special features of zygote development. Hum Reprod. 11:772 779. 27. Tanaka A, Tanaka I, Nagayoshi M, Awata S, Mawatari Y, Kusunoki H. 1997 Comparative study of embryonic development of human oocytes injected with fresh round spermatids and injected with in vitro cultured round spermatids with flagella. In: Gomel V, Leung PCK, eds. In vitro fertilization and assisted reproduction. Bologna, Italy: Monduzzi Editore; 705 710. 28. Aslam I, Fishel S. 1998 Short-term in vitro culture and cryopreservation of spermatogenic cells used for human in vitro conception. Hum Reprod. 13:634 638. 29. Lin WW, Lamb DJ, Wheeler TM, Lipshultz LI, Kim ED. 1997 In situ endlabeling of human testicular tissue demonstrates increased apoptosis in conditions of abnormal spermatogenesis. Fertil Steril. 68:1065 1069. 30. Sinha-Hikim AP, Wang C, Lue Y, Johnson L, Wang X-H, Swerdloff RS. 1998 Spontaneous germ cell apoptosis in humans: evidence for ethnic differences in the susceptibility of germ cells to programmed cell death. J Clin Endocrinol Metab. 83:152 156. 31. Dirami G, Ravindranath N, Kleinman HK, Dym M. 1995 Evidence that basement membrane prevents apoptosis of Sertoli cells in vitro in the absence of known regulators of Sertoli cell function. Endocrinology. 136: 4439 4447.