Developmental Changes in Expression of Myeloid Cell Leukemia-1 in ...

0013-7227/02/$15.00/0 Printed in U.S.A.

The Journal of Clinical Endocrinology & Metabolism 87(7):3417–3427 Copyright © 2002 by The Endocrine Society

Developmental Changes in Expression of Myeloid Cell Leukemia-1 in Human Germ Cells during Oogenesis and Early Folliculogenesis P. S. HARTLEY, R. A. L. BAYNE, L. L. L. ROBINSON, N. FULTON,

AND

R. A. ANDERSON

Medical Research Council Human Reproductive Sciences Unit, Centre for Reproductive Biology, Edinburgh, Scotland, United Kingdom EH3 9ET The regulation of germ cell number in the developing ovary is central to female reproduction. Members of the Bcl-2 family of proapoptotic and antiapoptotic proteins have been implicated in this process in rodents. We investigated the expression of Mcl-1, Bcl-2, Bax, and BAD at 13–21 gestational wk in the human fetal ovary and of Mcl-1 in the adult ovary. mRNA expression of Mcl-1 and its short form Mcl-1s, Bcl-2, Bax, and BAD was demonstrated in fetal ovary by RT-PCR. Hybridization array analysis suggested a selective increase in Mcl-1 expression between 14 and 18 wk gestation, which was confirmed by quantitative PCR. There was a corresponding change in the expression of Mcl-1 protein, detected by immunohistochemistry, from germ cells at the periphery of the ovary at 14 –16 wk to the largest germ cells, including oocytes within newly formed primordial follicles, at 21 wk. Mcl-1 was

A

CENTRAL ASPECT of ovarian function is that the number of oocytes in the postnatal ovary is limited. This is determined by the balance between germ cell proliferation and loss during fetal development, and the rate of subsequent loss throughout adult life. Factors regulating this balance are therefore crucial to the determination of reproductive potential. Following a period of mitotic proliferation, the human fetal ovary at mid-gestation contains its maximal number of germ cells (1). Entry into meiosis and interaction with somatic cells to form primordial follicles is believed to be essential for germ cell survival during this period of development (2, 3). Several factors including c-kit and its ligand (kit ligand or stem cell factor), DAZLA, GDF-9, and neurotropins have been identified as potentially important regulators of germ cell survival in the ovary (4 –10). Nevertheless, the number of oocytes present in the ovary is reduced by some 85% by the time of birth (1). Follicular atresia during adult life has been established to be a regulated apoptotic process (11, 12). Prenatal loss of germ cells in the mouse has also been demonstrated to result from apoptosis (13, 14), and apoptosis has been reported in the developing human ovary (15, 16). The Bcl-2 family of evolutionarily conserved, proapoptotic, and antiapoptotic proteins are implicated in the survival or demise of numerAbbreviations: Cp, Cross-over point; DNase, deoxyribonuclease; GAPDH, glyceraldehyde phosphate dehydrogenase; Mcl-1s, short form of Mcl-1; Mcl-1L, long form of Mcl-1; PCNA, proliferating cell nuclear antigen; RT, Expand Reverse Transcriptase; TBS, Tris-buffered saline; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling.

also expressed by oocytes of primordial and preantral follicles in the adult. Bax and BAD immunostaining was detected in both somatic and germ cells in the fetal ovary, whereas Bcl-2 was restricted to somatic cells: no changes in expression were observed. Apoptotic cells, detected by terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling, were observed in all fetal ovaries but were infrequent. These results confirm that Bcl-2 family members are differentially expressed in several cell types within the developing human ovary. Increased mRNA expression and the changing distribution of Mcl-1 in germ cells as they develop into primordial follicles as well as persistence in the growing oocyte in the adult may indicate an important role for this survival/antiapoptotic factor throughout germ cell development and maturation. (J Clin Endocrinol Metab 87: 3417–3427, 2002)

ous cell types and have been identified as regulators of apoptosis within the mammalian ovary (17). Oncogenes and tumor suppressor genes implicated in the regulation of apoptosis have been identified in the first trimester human ovary (18), and Bax but not Bcl-2 was detected after 14 wk of gestation (16). Mcl-1 is a recently identified antiapoptotic member of the Bcl-2 family (19). Expression of Mcl-1 is upregulated by a variety of growth factors (20), including factors demonstrated to be of importance to germ cell survival (21). Very limited data are available, however, regarding its possible expression in the human ovary (22). We have examined the expression and localization of the apoptotic regulatory factors Mcl-1, Bcl-2, Bax, and BAD in mid-trimester human fetal ovary during this period of formation of the essential structures of the ovary. In the light of the high level of expression of Mcl-1 in the oocyte in the developing ovary, we extended our investigations to the developing follicle in the adult ovary. Materials and Methods Tissues Human fetal ovaries were obtained following medical termination of pregnancy. Women gave consent according to national guidelines (23), and the study was approved by the Lothian Paediatrics/Reproductive Medicine Research Ethics SubCommittee. Termination of pregnancy was induced by treatment with mifepristone (200 mg orally) followed by prostaglandin E1 analog (Gemeprost, Beacon Pharmaceuticals, Tunbridge Wells, UK) 1 mg every 3 h per vaginam. None of the terminations were for reasons of fetal abnormality, and all fetuses appeared morphologically normal. Gestational age was determined by ultrasound examination before termination and confirmed by subsequent direct measurement of foot length. A total of 30 specimens were used for this study.

3417

3418

J Clin Endocrinol Metab, July 2002, 87(7):3417–3427

Biopsies of 4 adult ovaries were also obtained from women undergoing gynecological surgery for benign disease. The women were aged 31–39 yr, of proven fertility, and all gave informed written consent and the study received ethical approval from the above-mentioned committee. Ovaries were dissected free and either fixed for immunohistochemical analysis or snap frozen and stored at ⫺70 C. Fixation was carried out in Bouins for 5 h, followed by transfer to 70% ethanol before processing into paraffin using standard methods.

Isolation of RNA and synthesis of cDNA Total RNA was extracted using either the RNeasy Mini Kit (QIAGEN Ltd., Crawley, West Sussex, UK) for RT-PCR or TRIReagent (Sigma, Poole, Dorset, UK) for array and quantitative PCR analysis according to the manufacturers’ instructions. To remove contaminating genomic DNA, 3 ␮g of total RNA were treated with deoxyribonuclease (DNase) using Amplification grade DNase-1 (Roche Molecular Biochemicals, West Sussex, UK) according to the manufacturer’s instructions. The RNA was primed for reverse transcription with oligo(deoxythymidine) (Genosys Biotechnologies, Pampisford, UK) at 65 C for 10 min. The entire reaction was added to a total volume of 57 ␮l containing deoxynucleotide triphosphates to 1 mm, dithiothreitol to 10 mm, 12 ␮l 5⫻ Expand Reverse Transcriptase (RT) buffer, and 120 U ribonuclease inhibitor (Promega Corp., Southampton, UK). One third (19 ␮l) of this reaction was added to 1 ␮l water (RT⫺), which acted as a negative control to establish the efficacy of the DNase treatment. One hundred units of Expand RT (Roche Molecular Biochemicals) were added to the remaining 38 ␮l (RT⫹), and both reactions were incubated for 2 h at 40 C. Reactions were stored at ⫺70 C until required.

Amplification of specific cDNAs by PCR Target-specific PCR was performed using 1 ␮l of the RT⫹, RT⫺, or H2O as template in a reaction volume of 25 ␮l containing 1 U of AGSGold DNA polymerase (Hybaid, Middlesex, UK) 2.5 ␮l 10⫻ reaction buffer (with MgCl2 at 15 mm), deoxynucleotide triphosphates to 200 ␮m, forward and reverse primers to 500 nm. PCR primer sequences were BAD forward 5⬘GTTTGAGCCGAGTGAGCAGG 3⬘; reverse 5⬘ATAGCGCTGTGCTGCCCAGA 3⬘; Bcl-2 forward 5⬘CCTTCTTTGAGTTCGGTGGG 3⬘; reverse 5⬘CCAGGAGAAATCAAACAGAGGC 3⬘; Mcl-1 forward 5⬘ATCTCTCGGTACCTTCGGGAGC 3⬘; reverse 5⬘CCTGATGCCACCTTCTAGGTCC 3⬘; Bax forward 5⬘TTCTGACGGCAACTTCAACTGG 3⬘; reverse 5⬘GAGGAAGTCCAATGTCCAGC 3⬘; GAPDH forward 5⬘GAACGGGAAGCTCACTGGCAT 3⬘; reverse 5⬘GTCCACCACCCTGTTGCTGTAG 3⬘. The identity of all PCR products was confirmed by direct sequencing using an PE Applied Biosystems (Foster City, CA) 373A automated sequencer.

cDNA array analysis First strand cDNA probes were generated by reverse transcribing 5 ␮g of total RNA from 14 and 18 wk gestation ovary in the presence of 50 ␮Ci [␣-32P]-deoxycytidine triphosphate (3000 Ci/mmol⫺1, Amersham Pharmacia Biotech) using the cDNA Synthesis reagents provided with the array which include apoptosis gene-specific primers. Identical array membranes (Human Apoptosis-4 GEArray, Super Array Inc., Bethesda, MD) were hybridized and washed according to the manufacturer’s recommendations and exposed for 3 d each to phosphorimage screens (for densitometry analysis) and autoradiography at ⫺70 C with intensifying screens. After scanning, phosphorimages were analyzed using ImageQuant Software (Molecular Dynamics Ltd., Buckinghamshire, UK). Arrays were normalized to account for differences in quantity of starting RNA by calculating densities as a proportion of GAPDH signal.

Hartley et al. • Developmental Changes in Mcl-1 Expression

quisition was performed for each of 45 amplification cycles followed by continuous melt curve analysis to ensure product accuracy. Primers (Sigma), were: GAPDH: forward GACATCAAGAAGGTGGTGAAGC, reverse GTCCACACCCTGTTGCTGTAG, Mcl-1L: forward ATCTCTCGGTACCTTCGGGAGC, reverse GCTGAAAACATGGATCATCACTCG. The homologous sequence for the reverse Mcl-1L primer is in exon 2, allowing specific detection of the full length form of Mcl-1 (denoted Mcl-1L), whereas the primers described above detect both full length and the alternatively spliced short form, Mcl-1s (24, 25). Standard curves for GAPDH and Mcl-1 were derived by making 10-fold (GAPDH) or 2-fold (Mcl-1L) dilutions of first-strand cDNA from 14 and 18 wk gestation ovary. When the number of cycles needed to first yield a fluorescent signal above background (the cross-over point, Cp) is plotted against the log of relative concentration using LightCycler Software (Molecular Dynamics Ltd., Buckinghamshire, UK), the dilutions yielded a straight line for each product, confirming that Cp is a good indicator of target concentration across several orders of magnitude. The slopes of these curves are a measure of the efficiency of the PCR. Subsequent quantification of ovary cDNA was performed on 1:50 dilutions of cDNA in duplicate reactions for each experiment. Both GAPDH and Mcl-1 amplification from individual samples were performed in the same experiment. To normalize differences in template cDNA concentration between ovaries to allow comparison, calculations for Mcl-1 amplification were made relative to GAPDH from the same sample. Allowance for differences in amplification rate for the two targets was achieved by determining the actual amount of amplification required to yield a signal for each target. Results were subjected to statistical analysis by ANOVA and Student’s t test.

Tissues, fixation, and immunohistochemistry Sections (5 ␮m) were mounted onto 3-aminopropyl triethoxy-silane (Sigma) coated slides that were subsequently baked overnight at 60 C. Slides were dewaxed with xylene, and rehydrated through graded ethanol solutions. Heat-induced epitope retrieval was performed for Bcl-2, BAD, and Bax, by pressure cooking for 2.5 min in 0.01 m citrate at pH 6. Slides were incubated in 3% H2O2 for 30 min to quench endogenous peroxidase activity and washed twice in Tris-buffered saline (TBS; 0.05 m Tris, 0.85% NaCl, pH 7.4). Sections were then blocked with 0.01 m avidin then 0.001 m biotin (both from Vector Laboratories, Inc., Peterborough, UK and both diluted in 20% normal serum/TBS) for 15 min each, with washes in TBS in between. Primary antibodies were applied to slides and incubated overnight at 4 C as follows; anti-Mcl-1 (S19; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at 1:100; anti-Bcl-2 (DAKO Corp., Glostrup, Denmark) at 1:30; anti-Bax (PharMingen, San Diego, CA) at 1:2500 and anti-BAD (Santa Cruz Biotechnology, Inc.) at 1:10. Unbound primary antibody was removed from slides by two TBS washes before the application of appropriate biotinylated secondary antibody (DAKO Corp.) at a concentration of 1:500. Following two washes in TBS, sections were incubated with avidin biotin horseradish peroxidase linked complex (DAKO Corp.) according to the manufacturers instructions. Bound antibody was visualized using 3,3⬘-diaminobenzidine tetra-hydrochloride (DAKO Corp.). Primary antibodies were omitted as negative controls for Bcl-2 and Bax, whereas for anti-Mcl-1 and anti-BAD primary antibodies were preadsorbed overnight at 4 C with 100-fold excess of the respective blocking peptide (Santa Cruz Biotechnology, Inc.). All sections were counterstained with hematoxylin, dehydrated, mounted, and visualized by light microscopy. Images were captured using an Olympus Corp. Provis microscope (Olympus Corp. Optical Co., London, UK) equipped with a Kodak DCS330 camera (Eastman Kodak Co., Rochester, NY).

Nuclear measurement and statistics Lightcycler quantitative PCR Quantitative PCR was performed using the Lightcycler system (Roche Molecular Biochemicals). Reverse transcribed RNA samples were diluted in water as indicated, and 1 ␮l of the dilution was added to a final volume of 10 ␮l containing 2 mm MgCl2 and 1 ␮m each of forward and reverse primer in 1⫻ LightCycler-Fast Start DNA Master SYBR Green 1 Master Mix (Roche Molecular Biochemicals). Signal ac-

The nuclear diameter of Mcl-1 immunopositive and immunonegative germ cells was measured in two dimensions for 14, 16, 18, and 21 wk gestation ovaries using Image Proplus Image Analysis software (Media Cybernetics, Silver Spring, MD). Mean diameter for each germ cell was calculated, and numbers of cells grouped in 1-␮m increments. Bias was avoided by systematically measuring all germ cell nuclei in sequential, nonoverlapping fields of view until either over one hundred cells or all


immunopositive cells in the case of later gestational ages were measured. Data were analyzed by ANOVA and Student’s t test.

Detection of apoptosis using in situ DNA terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) Tissue sections were prepared as for immunohistochemistry with heat-induced epitope retrieval as described for immunohistochemistry. Slides were incubated at 37 C in the presence of 1.5 ␮g/ml proteinase K in 0.05 m Tris, pH 7.4, for 15 min. Endogenous peroxidase, avidin and biotin were blocked as described for immunohistochemistry. Sections were incubated with TUNEL solution (30 mm Tris, pH 7.2; 140 mm Na-Cacodylate; 1 mm CoCl2) containing 5 ␮l/ml digoxigenin-deoxyuridine triphosphate (Roche Molecular Biochemicals) and 30 U/ml terminal deoxynucleotidyl transferase (Promega Corp., Southampton, UK) for 30 min at 37 C. TUNEL solution was washed off and replaced with sheep antidigoxigenin antibodies (Roche Molecular Biochemicals) at 1:100, incubated at room temperature for 90 min, and washed. Positive staining was visualized with avidin-biotin complex-horseradish peroxidase/diaminobenzidine tetra-hydrochloride as described above for immunohistochemistry. Positive and negative controls were included in each experiment. For positive controls, sections were treated with 40 U/ml DNase-1 (Promega Corp., Southampton, UK) for 20 min at 37 C. Negative controls were incubated without terminal deoxynucleotidyl transferase enzyme.

Western immunoblotting Protein was extracted from human fetal ovary by homogenization in an extraction buffer containing 62.5 mm Tris (pH 6.80), 1% sodium dodecyl sulfate, 10% glycerol, and a cocktail of protease inhibitors (Complete Mini Protease Inhibitor cocktail tablets, Roche Molecular Biochemicals, Mannheim, Germany). Ten micrograms of protein extract was boiled for 5 min in 4⫻ reduced sample buffer (250 mm Tris, pH 6.8; 8%

FIG. 1. Expression of mRNA for Mcl-1, Bcl-2, Bax, and BAD in human ovary. RT-PCR analysis of mRNA expression for Bcl-2, Bax, Mcl-1 (long and short forms), and BAD in 14 –19 wk gestation ovary and for Mcl-1 in adult ovary as indicated. Molecular weights of the amplicons are marked. The band at 198 bp for Mcl-1 represents the short form (24, 25), confirmed by direct sequencing. Lanes marked RT⫺ contained samples in which the reverse transcriptase enzyme was not included.

J Clin Endocrinol Metab, July 2002, 87(7):3417–3427 3419

sodium dodecyl sulfate; 20% glycerol; and 0.01% bromophenol blue) and separated by SDS-PAGE on a 10% Tris/glycine gel (Novex, Invitrogen, Paisley, UK) in parallel with prestained protein molecular weight markers (Rainbow markers, Amersham Pharmacia Biotech, Bucks, UK). Proteins were transferred onto polyvinylidenedifluoride transfer membrane (Hybond-P, Amersham Pharmacia Biotech). The membrane was blocked overnight at 4 C in 5% wt/vol powdered milk and 10% normal swine serum in TBS, pH 7.5, then incubated with the primary antibody (rabbit antihuman MCL-1: s-19, Santa Cruz Biotechnology, Inc.) diluted 1:200 for 2 h at room temperature. Bound antibody was detected using an alkaline phosphatase-linked secondary antibody (1:20,000; goatantirabbit alkaline phosphatase conjugate, Sigma) and visualized using the enhanced chemifluorescent system (Amersham Pharmacia Biotech). Primary antibody specificity was verified both by preabsorbing the primary antibody with the Mcl-1 blocking peptide (Santa Cruz Biotechnology, Inc.) and by omitting the primary antibody.

Results Expression of Mcl-1, Bcl-2, Bax, and BAD mRNA: RT-PCR and cDNA array analysis

The expression of Bcl-2 family members, Mcl-1, Bcl-2, Bax, and BAD in human fetal ovary was determined by RT-PCR analysis. Single products for Bcl-2, Bax and BAD were amplified from RNA extracted from fetal ovaries across the gestational range 14 –19 wk (Fig. 1). Two products were amplified for Mcl-1 from RNA extracted from both fetal (14 –17 wk gestation) and adult ovary specimens (Fig. 1). Although the major product corresponded to the full-length Mcl-1 cDNA, a fainter band was also identified that corresponded to the recently described short form splice variant

3420


Mcl-1s (24, 25). The identity of this PCR product was confirmed to be the short form spice variant by direct sequencing. To investigate differences in expression of these genes with gestational age, a mini cDNA array was hybridized with firststrand cDNA probes generated from a 14- and an 18-wk gestation human fetal ovary. Spots were analyzed by densitometry and normalized to the GAPDH signal. Bcl-2, Bad, and Bax appeared to be expressed at relatively low levels, yielding low hybridization signals (Fig. 2). These results suggested that there may be a small increase in signal for each of these genes in 18 wk compared with 14 wk ovary but of less than 1.5-fold. In contrast, Mcl-1 produced a stronger hybridization signal and showed a 2.02-fold greater intensity of signal in 18-wk ovary compared with 14-wk ovary when normalized to GAPDH. A 2-fold increase can be considered significant (26); thus, this result was indicative of up-regulation in Mcl-1 expression contemporaneous with primordial follicle formation in the human ovary. Quantitative PCR

cDNA array analysis, however, is only semiquantitative, and the increase in expression observed was modest. In addition, hybridization may not distinguish between the long and short forms of Mcl-1 mRNA that, given their antiapoptotic and proapoptotic functions, may make different contributions at different stages in development. It was therefore necessary to confirm this result by other means. We have used quantitative PCR in a LightCycler (Roche Molecular Biochemicals) using SYBR GREEN Dye chemistry to compare Mcl-1 expression relative to GAPDH in fetal ovary across the gestational range 14 –18 wk with primers specific for full-length Mcl-1 (Mcl-1L). Formation of product was ascertained in each reaction by melt curve analysis and confirmed by running sample reactions on a 2.0% agarose gel (Fig. 3, A–C). Standard curves for GAPDH and Mcl-1 (Fig. 3, D and E) were derived to determine the efficiency of each PCR and allow relative concentrations to be calculated (Fig. 3F). This confirmed an increase in Mcl-1 expression through mid gestation (P ⫽ 0.003), with a 2.6-fold increase in Mcl-1 mRNA levels between 14 –15 and 18 wk gestation, comparable with that found from the array analysis. Mcl-1 mRNA expression was significantly higher at 18 wk gesta-

FIG. 2. cDNA array analysis of Mcl-1, Bcl-2, Bax, and BAD expression. Phosphorimages of paired duplicate spots corresponding to BAD, Bax, Bcl-2, Mcl-1, and GAPDH from a human apoptosis-4 cDNA array probed with 32P-labeled first-strand cDNA synthesized from RNA from 14 and 18 wk gestation fetal ovaries as indicated. Mean spot signal densities for each Bcl-2 family member at the two gestations are expressed as a percentage of GAPDH signal, from which the ratio of level of expression between samples at 18 and 14 wk gestation was calculated for each mRNA.


tion than at all previous gestations examined, whereas Mcl-1 mRNA levels in 16 and 17 wk gestation ovaries were comparable to those at 14 –15 wk, suggesting that the increase in Mcl-1 mRNA occurs sharply at 17–18 wk gestation, at the time when primordial follicles are first observed. This analysis considered only the full-length form of Mcl-1 (Mcl-1L). A PCR assay specific for the short form (Mcl-1s) was also devised using a primer spanning the junction of exons 1 and 3 which will therefore only amplify a message in which these 2 exons are adjacent (i.e. Mcl-1s). This was validated by conventional PCR (not shown). However, this molecule appears to be present at much lower levels than full-length Mcl-1, as suggested by RT-PCR with nonspecific Mcl-1 primers (Fig. 1). The high Cp value in the LightCycler gave values close to background therefore quantitative comparisons were not possible. Immunohistochemical localization and immunoblotting of Mcl-1

Mcl-1 protein was specifically immunolocalized to the cytoplasm of germ cells in mid-trimester ovaries, with a change in the distribution across the gestational age range investigated (Fig. 4, A–D). At 14 and 16 wk gestation, a clear differential distribution within the ovary was observed with immunostained germ cells localized predominantly to the periphery of the ovary with few stained cells present centrally (Fig. 4, A). At later gestations (17–18 wk), there was a striking change in the pattern of immunostaining. Although the signal in the peripheral germ cells was maintained, larger germ cells within the more medullary region of the ovary showed intense immunostaining. This pattern of staining became even more marked at 18 and 21 wk gestation, where a proportion of the large germ cells showing intense Mcl-1 immunostaining were observed to be associated with somatic cells to form primordial follicles (Fig. 4, B–D). Furthermore, the intensity of immunostaining in the cytoplasm of germ cells located in the periphery of the ovary was much reduced at these later gestations. Weak Mcl-1 immunoreactivity was also detected in endothelial vascular cells (not shown), but no immunoreactivity was detected in other somatic cells of the fetal ovary, including the ovarian surface epithelium (Fig. 4A) at any gestation examined. No immuno-



FIG. 3. Real-time PCR quantification of Mcl-1 expression in fetal ovary. Expression of Mcl-1L mRNA was quantified in human fetal ovary specimens over the gestational range 14 to 18 wk. Data show (A) representative individual PCR melt curves of GAPDH and (B) Mcl-1L PCR products indicating single amplicons in each reaction; (C) agarose gel analysis of PCRs, indicating products with the expected sizes for GAPDH and Mcl-1L, the central lane showing molecular weight markers; (D) and (E) standard curves of input cDNA concentration vs. number of amplification cycles required to yield a signal above threshold for GAPDH and Mcl-1L respectively, demonstrating linearity; (F) calculated Mcl-1L mRNA expression as a percentage of GAPDH for ovaries of 14 –15, 16, 17, and 18 wk gestation (n ⫽ 4, 2, 3, and 4 respectively, mean ⫾ SEM). *, P ⫽ 0.003 by ANOVA.

staining was observed when the primary antibody was preabsorbed with the blocking peptide (Fig. 4E). Immunoblotting confirmed the presence of Mcl-1 protein in the fetal ovary (Fig. 5). A prominent single band of approximately 40 kDa corresponding to full-length Mcl-1 was identified in all specimens examined. The antibody used may also detect Mcl-1s (25); however, no band of the appropriate size was detected, indicating very little or no translation of this mRNA in these tissues. Quantitative analysis of Mcl-1 protein expression was not performed in view of limited tissue availability. Quantitative analysis was used, however, to investigate the size distribution of germ cells showing positive Mcl-1 immunoreactivity in the developing ovary to investigate the impression of increased Mcl-1 immunostaining in larger germ cells. Average germ cell nuclear diameter increased with increasing gestation across the range 14 –21 wk with, in addition, a wider range at later gestations (Fig. 6). Mean nuclear diameter increased for both Mcl-1 positive and negative cell populations at later gestations (both P ⬍ 0.001), although the increase was most marked for Mcl-1 positive cells. Thus, at 14 and 16 wk gestation, mean nuclear diameter of Mcl-1 positive and negative germ cells did not show statistically significant differences (respectively 8.6 ⫾ 0.1 ␮m, n ⫽ 130 vs. 8.5 ⫾ 0.1 ␮m, n ⫽ 118 at 14 wk; 8.4 ⫾ 0.1 ␮m, n ⫽ 102 vs. 8.8 ⫾ 0.1 ␮m, n ⫽ 101 at 16 wk). However, at 18 and 21 wk gestation a differential distribution was seen with Mcl-1 positive cells being significantly larger than Mcl-1 negative cells (15.8 ⫾ 0.3, n ⫽ 66 vs. 11.1 ⫾ 0.2 ␮m, n ⫽ 106 at

18 wk; 16.4 ⫾ 0.2, n ⫽ 54 vs. 11.3 ⫾ 0.2 ␮m, n ⫽ 108 at 21 wk; both P ⬍ 0.001, Fig. 6). Mcl-1 expression was also investigated in primordial and growing preantral follicles in the adult ovary. Intense immunostaining for Mcl-1 was detected in the cytoplasm of oocytes within all primordial and preantral follicles examined (Fig. 4, F–H). Weaker Mcl-1 immunostaining was also observed in the cytoplasm of granulosa and theca cells surrounding these oocytes in developing follicles (Fig. 4, G and H). In addition, Mcl-1 immunostaining was observed within the cytoplasm of ovarian surface epithelial cells (Fig. 4I), with weaker staining within the cytoplasm of endothelial vascular cells (not shown). The positive staining of adult ovarian surface epithelial cells was in contrast to the negative staining pattern of the surface epithelium observed in fetal tissues. No immunostaining was observed in any tissue sections incubated with primary antibody that had been preabsorbed with the blocking peptide (Fig. 4J). Immunohistochemical localization of Bcl-2, Bax, and BAD

The expression of Bcl-2 and Bax proteins in the fetal ovary showed a very different pattern to that of Mcl-1. Bcl-2 immunostaining was restricted to the cytoplasm of the somatic cell population in all samples examined (Fig. 7, A–C). Although some immunopositive cells were adjacent to germ cells and thus may be pregranulosa cells, other immunopositive cells were distributed throughout the stroma. Bax immunopositive staining was widespread and observed in

3422



FIG. 4. Immunohistochemical detection of Mcl-1 in human fetal and adult ovary. Immunohistochemical detection of Mcl-1 in human fetal ovary at (A) 14 wk, (B) 18 wk, and (C and D) 21 wk gestation, and in adult ovary in (F) primordial follicle, (G) secondary follicle, (H) preantral follicle, and (I) ovarian surface epithelium. E and J, Negative controls for fetal and adult ovary respectively, in which the primary antibody was preabsorbed with the blocking peptide. gc, Germ cell; sc, stromal cell; o, oocyte; pf, primordial follicle; th, theca; ose, ovarian surface epithelium. Scale bars, 50 ␮m (A, B, C, G, H); 20 ␮m (D, E, F, I).



Discussion

FIG. 5. Immunoblot of Mcl-1 in human fetal ovary. Total protein extracts (10 ␮g) of human fetal ovary at 13–17 wk gestation as indicated were separated by SDS-PAGE, transferred to polyvinylidenedifluoride membrane and incubated with anti-Mcl-1. The positions of molecular weight markers are shown. A single band of molecular mass approximately 40 kDa corresponding to full-length Mcl-1 was detected in each specimen. No signal was detected when the primary antibody was preabsorbed with blocking peptide or omitted (not shown).

all ovarian cell types (Fig. 7, D–F), although more intense staining was observed within somatic cells compared with germ cells (Fig. 7F). At earlier gestational ages (14 –17 wk), the intensity of staining within germ cells was not consistent; more intense staining was observed in a subset of germ cells nearer the periphery of the ovary (Fig. 7D). Positive Bax staining was also present in cells exhibiting apoptotic morphology (Fig. 7F), which was not observed for Bcl-2 or Mcl-1. Although the overall pattern of staining changed with the formation of primordial follicles at later gestations, there were no other major changes observed in the intensity of immunostaining for Bcl-2 or Bax across the range of gestational ages examined. Cells of the ovarian surface epithelium showed no staining for either Bcl-2 or Bax. Immunostaining for BAD was localized predominantly to the somatic cells of the fetal ovary in all samples examined (Fig. 7, G–I). The cytoplasm of a small number of germ cells was also immunopositive (Fig. 7I), but these did not appear to be restricted to any particular part of the developing ovary, and no pattern was detected. TUNEL of germ cells in human fetal ovary

TUNEL-positive nuclei were detected in all fetal ovary specimens examined. The cells showed the characteristic appearance of apoptosis (chromatin condensation and shrinkage of the cytoplasm) (27). The prevalence of these nuclei was very low, only 5–10 positive nuclei (⬍5% of germ cells) being detected per histological section (Fig. 7J), and there was no clear change in the abundance of these cells across the gestational range examined. It was noted, however, that none of the oocytes within primordial follicles were TUNEL-positive, although the absence of primordial follicles from the majority of specimens examined means that the great majority of germ cells studied were at earlier developmental stages. In positive controls, most cell nuclei were stained (Fig. 7K), whereas none were stained in negative controls.

These data demonstrate mRNA and protein expression of members of the Bcl-2 family Mcl-1, Bcl-2, Bax, and BAD in human fetal ovary during the second trimester. These antiapoptotic (Mcl-1 and Bcl-2) and proapoptotic factors (Bax and BAD) were expressed in different cell types. Mcl-1 was expressed exclusively by germ cells, whereas Bcl-2 was expressed by somatic cells and Bax and BAD by both cell types, although BAD expression was seen only in a few germ cells. The data indicate also that there is a pronounced gestational age-dependent pattern of Mcl-1 expression. We have demonstrated by two independent methods that Mcl-1 mRNA levels are higher at 18 wk gestation than at 14 –17 wk gestation, i.e. at the time when primordial follicles begin to form. At the protein level, at 14 –16 wk of gestation, only small, peripherally located germ cells showed immunostaining with no difference in size between Mcl-1 positive and negative cells. At later developmental stages, there was a wider range in cell size, with the larger germ cells intensely stained for Mcl-1 and the most intense staining seen in oocytes within primordial follicles. Oocyte-specific staining was also observed in primordial and developing preantral follicles in the adult ovary and, as such, is suggestive of Mcl-1 expression persisting in these cells through to adult life. The overall increase in Mcl-1 mRNA between 14 –17 and 18 wk gestation is consistent with the changes in Mcl-1 protein distribution observed by immunohistochemistry. However, the level of this increase (2.6-fold) is likely to be an underestimate for individual oocytes as the RT-PCR analysis was carried out on whole ovary samples, whereas only a proportion of the larger oocytes showed a marked increase in Mcl-1 staining by 18 wk. The period of ovarian development investigated presently spans germ cell proliferation by mitosis through to the entry of germ cells into meiosis and the association with somatic cells to form primordial follicles. In the human, the time scale of these processes is overlapping, with entry into meiosis being detected as early as 11 wk gestation (28, 29). Germ cell nuclear diameters were similar to those previously described, with increasing diameter associated with progression from mitosis into meiotic prophase (1). During these processes, there is believed to be increasing cell death, such that approximately only 15% of oocytes remain present at the time of birth. Apoptosis, a form of programmed cell death, is a universal system for achieving this goal. Although widely conserved between organisms, the intracellular mechanics and extracellular controls of apoptosis appear to be tissue and cell specific (17). The present results demonstrate that Bcl-2 family members are expressed throughout the second trimester and are likely to be involved in the regulation of germ cell development and survival at that time in development. In particular, Mcl-1 may be of central importance in suppressing apoptosis at the crucial time of germ cell-somatic cell interaction required for the formation of the primordial follicle. The demonstration of continued Mcl-1 expression by oocytes in the adult ovary both in primordial follicles and during early follicular growth may reflect a significant role in the regulation of follicular loss. At the stages of follicular development investigated in the present study, it is believed that follicular atresia is initiated by oo-

3424



FIG. 6. Size distribution of Mcl-1 immunopositive and immunonegative cells in the human fetal ovary. Histograms showing the frequency distribution of nuclear diameter of Mcl-1 immunopositive cells (filled columns) and immunonegative cells (open columns) in human fetal ovary at (A) 14, (B) 16, (C) 18, and (D) 21 wk gestation. n ⫽ 54 –130, as described in Results.

cyte apoptosis followed by granulosa cell death, whereas at later antral stages it is initiated by granulosa cell apoptosis (11, 30, 31). The significance of the differential expression of Bcl-2 family members by different cell types, i.e. Mcl-1 by germ cells, Bcl-2 by somatic cells and BAD and Bax by both, is consistent with the cellular specificity of apoptotic pathways (17) and may underlie the importance of interaction between these two cell types for their mutual survival (3). Distinct caspases have been suggested to mediate apoptosis in oocytes and granulosa cells within the ovary (32). In response to internal and external signals, Bcl-2 family proteins interact with each other and with other nonBcl-2 proteins (33, 34) to form heterodimers. These interactions, as well as posttranslational modifications (35), govern the subcellular localization of the Bcl-2 family members, their functional effects, and the ultimate balance between cell survival and apoptosis. Information regarding potential binding partners of Bcl-2 family members has been drawn from the yeast

two hybrid system (36). These studies identified Mcl-1 as an antiapoptotic protein capable of binding Bok, BAD, and Bax (37, 38). Mcl-1 was originally identified in human myeloblastic leukemia cells (19) and is up-regulated in hematopoietic cell models by various cell survival signals including stem cell factor (20, 39 – 41). Stem cell factor, the ligand for the c-kit proto-oncogene receptor expressed by primordial germ cells, is recognized to have an important antiapoptotic and proliferative role in the developing ovary (20, 21, 42). We have recently demonstrated the expression of c-kit by human ovarian germ cells before and during primordial follicle formation (7). It is therefore possible that Mcl-1 expression in germ cells leads to an increased antiapoptotic environment and that this process is stem cell factor/c-kit dependent. Up-regulation of Bax expression was also partially prevented by stem cell factor in cultured murine primordial germ cells (43), consistent with an increased antiapoptotic signal. It has been demonstrated recently in a hematopoietic cell



FIG. 7. Immunodetection of Bcl-2, Bax, BAD and of apoptotic cells by TUNEL in human fetal ovary. Bcl-2 immunostaining in fetal ovary at (A) 14 and (B and C) 18 wk gestation. Immunodetection of Bax in fetal ovary at (D) 15 and (E and F) 21 wk gestation. Immunodetection of BAD in fetal ovary at (G) 14 wk and (H and I) 18 wk gestation. J, TUNEL staining of an apoptotic germ cell in a 16 wk gestation fetal ovary. K, TUNEL positive control (DNase-treated). gc, Germ cell; sc, stromal cell; pf, primordial follicle; os, ovarian stroma; ap, apoptotic cell; t, TUNEL-positive; ose, ovarian surface epithelium. L–N, Negative controls for Bcl-2, Bax, and BAD, respectively. Scale bars, 50 ␮m (A); 20 ␮m (all others).

3426


line that Mcl-1 expression can be differentially controlled at both the transcriptional and the translational level (44). Cytokine-induced increase in Mcl-1 transcription is dependent upon ERK, whereas Mcl-1 protein up-regulation is dependent on phosphatidylinositol 3-kinase. Both of these kinases have been implicated as downstream effectors of a number of cytokine/ growth factor receptors including c-kit (45) and neurotropin receptors (46 – 48). As these receptor pathways have themselves been implicated in development of the human fetal ovary (7, 8), either or both of these mechanisms may therefore control Mcl-1 expression in the developing ovary. Recently, a splice variant of Mcl-1 has been identified in human placenta (24, 25). This form lacks the BH1, BH2, and transmembrane domains, while retaining the BH3 domain associated with a proapoptotic function. Overexpression resulted in induction of apoptosis and antagonism of the antiapoptotic effect of the full-length form of Mcl-1. The present results suggest that germ cells in the developing ovary and in the adult express both forms of Mcl-1 mRNA, although mRNA levels of the short form are much lower than those of the full-length message, and only the full-length protein was detected by immunoblotting. Thus, it is possible that the alternative splicing mechanism could be regulated in the oocyte at several stages of development to determine cell survival. Mcl-1 was also detected in the adult human ovary. As in the fetal ovary, Mcl-1 expression was intense in oocytes in primordial follicles and remained so during early follicular development. Granulosa and theca cells of growing preantral follicles were also demonstrated to express Mcl-1, as was the surface epithelium. Further investigation is required to clarify whether there is gene expression by granulosa cells, or whether the Mcl-1 protein detected is a product of the oocyte. Endothelial vascular cells showed weak Mcl-1 immunostaining, as previously described in cultured endothelial cells (49). Mcl-1 mRNA expression is up-regulated by gonadotropins in the rat ovary (50), indicating that it may have a continuing role during later stages of follicular development. Mcl-1 appears to be unique among the Bcl-2 family in that it appears to have a cell cycle control function in addition to but independent of its role in apoptosis. Mcl-1 was colocalized with proliferating cell nuclear antigen (PCNA) to the nucleus of a human osteosarcoma cell line (34). A mutant form of Mcl-1, incapable of binding to PCNA, was found to retain the antiapoptotic function but attenuated the inhibitory effect of wild-type Mcl-1 on DNA synthesis. Thus, Mcl-1 appears to inhibit DNA synthesis via a direct interaction with PCNA. Changing expression of Mcl-1 may therefore have several roles in ovarian germ cells as they progress from mitotic proliferation to entry into and arrest in meiosis. It is also possible the Mcl-1 may have several roles in the different cell types in the adult ovary in which expression was detected. Bcl-2 has been localized to stromal cells in the first and early second trimester human ovary (18), consistent with the present findings, although others were unable to detect Bcl-2 after 14 wk gestation (16). Bcl-2 and Bax have been identified in the adult human ovary, with both localized to granulosa cells (16, 51). The significance of Bcl-2 and Bax in mammalian ovarian development has been examined in murine loss-offunction models. Bcl-2⫺/⫺ mice display an abnormal ovarian phenotype (52), and BAX⫺/⫺ mice show an accumulation of


atretic follicles (53), although in BAX⫺/⫺ mice the number of primordial follicles did not appear to be affected. Overexpression of Bcl-2 leads to the suppression of apoptosis in murine ovarian cells (54), and similarly deficiency of caspase 2 results in an excess of surviving germ cells (55). Although Bcl-2 and Bax were undetectable during murine oocyte development, expression of both was increased in apoptotic germ cells in culture (43). Mcl-1 knockout results in periimplantation embryonic loss (56), although no information is available regarding the effect of heterozygosity. The presence of apoptosis in the developing human ovary was demonstrated morphologically and confirmed using the TUNEL technique. The prevalence of TUNEL-positive cells identified in the present study was low (⬍5% of germ cells), whereas others have reported 9 –17% of oocytes to be apoptotic using a similar technique (16). An analysis using morphological criteria gave much lower proportions of degenerating germ cells but indicated a marked increase from approximately 0.1–3.5% over the gestational range examined in the present study (28). In contrast, freshly obtained murine ovary at a comparable developmental stage (i.e. before primordial follicle formation) was found to contain no apoptotic germ cells, such cells only being detectable after in vitro culture (45). The wide variation and relatively high percentages of apoptotic germ cells reported for human fetal samples may therefore largely reflect differences in prefixation changes in the tissues as well as differences in the techniques used. The frequency of TUNEL-positive cells may, however, underestimate the incidence of apoptosis as DNA degradation is a late event in the sequence of cell death (57). It was noteworthy that we observed no TUNEL-positive germ cells among those that had already formed primordial follicles. Before primordial follicle formation, germ cells are arranged in clusters of cells in the human (58) and other mammalian and nonmammalian species (59, 60). In the mouse, it has recently been demonstrated that most germ cell loss occurs by apoptosis as clusters of cells break down to form primordial follicles (14). It is therefore likely that similar processes are involved in the human, with the majority of germ cell loss occurring at a comparable developmental stage. In conclusion, these data demonstrate the expression and immunodetection of members of the Bcl-2 family of apoptosisregulating proteins in the developing human ovary. Different family members were localized to different cell types and a striking developmental change was identified in the expression of Mcl-1 in oogonia and oocytes. Oocytes within adult preantral follicles also displayed Mcl-1 immunopositive staining. These data indicate that this antiapoptotic factor may have important roles both during germ cell maturation and germ cell/somatic cell interaction at the time of primordial follicle formation, and during subsequent follicle development in adult life. Acknowledgments We thank Mike Miller for his assistance with image analysis. Received December 28, 2001. Accepted March 25, 2002. Address all correspondence to: Dr. R. A. Anderson, Medical Research Council Human Reproductive Sciences Unit, Center for Reproductive Biology, 37 Chalmers Street, Edinburgh, Scotland, United Kingdom EH3 9ET. E-mail: [email protected].


References 1. Baker TG 1963 A quantitative and cytological study of germ cells in human ovaries. Proc R Soc Lond Ser B 158:417– 433 2. Hirshfield AN 1991 Development of follicles in the mammalian ovary. Int Rev Cytol 124:43–101 3. McLaren A 1991 Development of the mammalian gonad: the fate of the supporting cell lineage. Bioessays 13:151–156 4. Dong J, Albertini DF, Nishimori K, Kumar TR, Lu N, Matzuk MM 1996 Growth differentiation factor-9 is required during early ovarian folliculogenesis. Nature 383:531–535 5. Tisdall DJ, Fidler AE, Smith P, Quirke LD, Stent VC, Heath DA, McNatty KP 1999 Stem cell and c-kit gene expression and protein localization in the sheep ovary during fetal development. J Reprod Fertil 116:277–291 6. Dissen GA, Romero C, Hirschfield AN, Ojeda SR 2001 Nerve growth factor is required for early follicular development in the mammalian ovary. Endocrinology 142:2078 –2086 7. Robinson LLL, Gaskell TL, Saunders PTK, Anderson RA 2001 Germ cell specific expression of c-kit in the human fetal gonad. Mol Hum Reprod 7:845– 852 8. Anderson RA, Robinson LLL, Brooks J, Spears N 2002 Expression of neurotrophins and their receptors in the human fetal ovary. J Clin Endocrinol Metab 87:890 – 897 9. Ruggiu M, Speed R, Taggart M, McKay SJ, Kilanowski F, Saunders P, Dorin J, Cooke HJ 1997 The mouse Dazla gene encodes a cytoplasmic protein essential for gametogenesis. Nature 389:73–77 10. Elvin JA, Matzuk MM 1998 Mouse models of ovarian failure. Rev Reprod 3:183–195 11. Hsueh AJW, Billig H, Tsafriri A 1994 Ovarian follicle atresia: a hormonally controlled apoptotic process. Endocr Rev 15:707–724 12. Pru JK, Tilly JL 2001 Programmed cell death in the ovary: insights and future prospects using genetic technologies. Mol Endocrinol 15:845– 853 13. Coucouvanis EC, Sherwood SW, Carswell-Crumpton C, Spack EG, Jones PP 1993 Evidence that the mechanism of prenatal germ cell death in the mouse is apoptosis. Exp Cell Res 209:238 –247 14. Pepling ME, Spradling AC 2001 Mouse ovarian germ cell cysts undergo programmed breakdown to form primordial follicles. Dev Biol 234:339 –351 15. de Pol A, Vaccina F, Foraboscu A, Cavazzuti E, Marzoni L 1997 Apoptosis of germ cells during human prenatal oogenesis. Hum Reprod 12:2235–2241 16. Vaskivuo TE, Anttonen M, Herva R, Billig H, Dorland M, te Velde ER, Stenback F, Heikinheimo M, Tapanainen JS 2001 Survival of human ovarian follicles from fetal to adult life: apoptosis, apoptosis-related proteins, and transcription factor GATA-4. J Clin Endocrinol Metab 86:3421–3429 17. Hsu SY, Hsueh AJ 2000 Tissue-specific Bcl-2 protein partners in apoptosis: an ovarian paradigm. Physiol Rev 80:593– 614 18. Quenby SM, Gazvani MR, Brazeau C, Neilson J, Lewis-Jones DI, Vince G 1999 Oncogenes and tumour suppressor genes in first trimester human fetal development. Mol Hum Reprod 5:737–741 19. Kozopas KM, Yang T, Buchan HL, Zhou P, Craig RW 1993 MCL1, a gene expressed in programmed myeloid cell differentiation, has sequence similarity to BCL2. Proc Natl Acad Sci USA 90:3516 –3520 20. Huang H-M, Huang C-H, Yen JJ-Y 2000 Mcl-1 is a common target of stem cell factor and interleukin-5 for apoptosis prevention activity via MEK/MAPK and PI-3K/Akt pathways. Blood 96:1764 –1771 21. Manova K, Nocka K, Besmer P, Bachvarova RF 1990 Gonadal expression of c-kit encoded at the W locus of the mouse. Development 110:1057–1069 22. Sano M, Umezawa A, Suzuki A, Shimoda K, Fukuma M, Hata J 2000 Involvement of EAT/mcl-1, an anti-apoptotic bcl-2-related gene, in murine embryogenesis and human development. Exp Cell Res 259:127–139 23. Polkinghorne J 1989 Review of the guidance on the research use of fetuses and fetal material. London: Her Majesty’s Stationery Office 24. Bae J, Leo CP, Hsu SY, Hsueh AJW 2000 MCL-1S, a splicing variant of the antiapoptotic BCL-2 family member MCL-1, encodes a proapoptotic protein possessing only the BH3 domain. J Biol Chem 275:25255–25261 25. Bingle CD, Craig RW, Swales BM, Singleton V, Zhou P, Whyte MK 2000 Exon skipping in Mcl-1 results in a bcl-2 homology domain 3 only gene product that promotes cell death. J Biol Chem 275:22136 –22146 26. Rajeevan MS, Vernon SD, Taysavang N, Unger ER 2001 Validation of arraybased gene expression profiles by real-time (kinetic) RT-PCR. J Mol Diagn 3:26 –31 27. Kerr JFR, Wyllie AH, Currie AR 1972 Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 26:239 –257 28. Kurilo LF 1981 Oogenesis in antenatal development in man. Hum Genet 57:86 –92 29. Gondos B, Westergaard L, Byskov AG 1986 Initiation of oogenesis in the human fetal ovary: ultrastructural and squash preparation study. Am J Obstet Gynaecol 155:189 –195 30. Gougeon A 1996 Regulation of ovarian follicular development in primates: facts and hypotheses. Endocr Rev 17:121–155 31. Morita Y, Perez GI, Maravei DV, Tilly KI, Tilly JL 1999 Targeted expression of Bcl-2 in mouse oocytes inhibits ovarian follicle atresia and prevents spontaneous and chemotherapy-induced oocyte apoptosis in vitro. Mol Endocrinol 13:841– 850 32. Matikainen T, Perez GI, Zheng TS, Kluzak TR, Rueda BR, Flavell RA, Tilly JL 2001 Caspase-3 gene knockout defines cell lineage specificity for programmed cell death signaling in the ovary. Endocrinology 142:2468 –2480


33. Hsu SY, Hsueh AJ 1998 A splicing variant of the Bcl-2 member Bok with a truncated BH3 domain induces apoptosis but does not dimerize with antiapoptotic Bcl-2 proteins in vitro. J Biol Chem 273:30139 –30146 34. Fujise K, Zhang D, Liu J-L, Yeh ETH 2000 Regulation of apoptosis and cell cycle progression by Mcl-1. J Biol Chem 275:39458 –39465 35. Zha J, Harada H, Yang E, Jockel J, Korsmeyer SJ 1996 Serine phosphorylation of death agonist BAD in response to survival factor results in binding to 14 –3-3 not BCL-X(L). Cell 87:619 – 628 36. Hsu SY, Hsueh AJ 1998 Intracellular mechanisms of ovarian cell apoptosis. Mol Cell Endocrinol 145:21–25 37. Hsu SY, Kaipia A, McGee E, Lomeli M, Hsueh AJ 1997 Bok is a pro-apoptotic Bcl-2 protein with restricted expression in reproductive tissues and heterodimerizes with selective anti-apoptotic Bcl-2 family members. Proc Natl Acad Sci USA 94:12401–12406 38. Hsu SY, Lin P, Hsueh AJ 1998 BOD (Bcl-2-related ovarian death gene) is an ovarian BH3 domain-containing proapoptotic Bcl-2 protein capable of dimerization with diverse antiapoptotic Bcl-2 members. Mol Endocrinol 12:1432–1440 39. Chao JR, Wang JM, Lee SF, Peng HW, Lin YH, Chou CH, Li JC, Huang HM, Chou CK, Kuo ML, Yen JJ, Yang-Yen HF 1998 Mcl-1 is an immediate-early gene activated by the granulocyte-macrophage colony-stimulating factor (GMCSF) signaling pathway and is one component of the GM-CSF viability response. Mol Cell Biol 18:4883– 4898 40. Moulding DA, Quayle JA, Hart CA, Edwards SW 1998 Mcl-1 expression in human neutrophils: regulation by cytokines and correlation with cell survival. Blood 92:2495–2502 41. Fukuchi Y, Kizaki M, Yamoto K, Kawamura C, Umezawa A, Hata JJ, Nishihara T, Ikeda Y 2001 Mcl-1, an early-induction molecule, modulates activin A-induced apoptosis and differentiation of CML cells. Oncogene 20:704 –713 42. Pesce M, Farrace MG, Piacentini M, Dolci S, De Felici M 1993 Stem cell factor and leukemia inhibitory factor promote primordial germ cell survival by suppressing programmed cell death (apoptosis). Development 118:1089 –1094 43. Felici MD, Carlo AD, Pesce M, Iona S, Farrace MG, Piacentini M 1999 Bcl-2 and Bax regulation of apoptosis in germ cells during prenatal oogenesis in the mouse embryo. Cell Death Differ 6:908 –915 44. Schubert KM, Duronio V 2001 Distinct roles for extracellular-signal-regulated protein kinase (ERK) mitogen-activated protein kinases and phosphatidylinositol 3-kinase in the regulation of Mcl-1 synthesis. Biochem J 356:473– 480 45. Morita Y, Manganaro TF, Tao XJ, Martimbeau S, Donahoe PK, Tilly JL 1999 Requirement for phosphatidylinositol-3⬘-kinase in cytokine-mediated germ cell survival during fetal oogenesis in the mouse. Endocrinology 140:941–949 46. Greene LA, Kaplan DR 1995 Early events in neurotrophin signalling via Trk and p75 receptors. Curr Opin Neurobiol 5:579 –587 47. Yuen EC, Mobley WC 1999 Early BDNF, NT-3, and NT-4 signaling events. Exp Neurol 159:297–308 48. Roux PP, Bhakar AL, Kennedy TE, Barker PA 2001 The p75 neurotrophin receptor activates Akt (protein kinase B) through a phosphatidylinositol 3-kinase-dependent pathway. J Biol Chem 276:23097–23104 49. Karsan A, Yee E, Poirier GG, Zhou P, Craig R, Harlan JM 1997 Fibroblast growth factor-2 inhibits endothelial cell apoptosis by Bcl-2-dependent and independent mechanisms. Am J Pathol 151:1775–1784 50. Leo CP, Hsu SY, Chun SY, Bae HW, Hsueh AJ 1999 Characterization of the antiapoptotic Bcl-2 family member myeloid cell leukemia-1 (Mcl-1) and the stimulation of its message by gonadotropins in the rat ovary. Endocrinology 140:5469 –5477 51. Kugu K, Ratts VS, Piquette GN, Tilly KI, Tao XJ, Martimbeau S, Aberdeen GW, Krajewski S, Reed JC, Pepe GJ, Albrecht ED, Tilly JL 1998 Analysis of apoptosis and expression of bcl-2 gene family members in the human and baboon ovary. Cell Death Differ 5:67–76 52. Ratts VS, Flaws JA, Kolp R, Sorenson CM, Tilly JL 1995 Ablation of bcl-2 gene expression decreases the numbers of oocytes and primordial follicles established in the post-natal female mouse gonad. Endocrinology 136:3665–3668 53. Knudson CM, Tung KS, Tourtellotte WG, Brown GA, Korsmeyer SJ 1995 Bax-deficient mice with lymphoid hyperplasia and male germ cell death. Science 270:96 –99 54. Hsu SY, Lai RJ, Finegold M, Hsueh AJ 1996 Targeted overexpression of Bcl-2 in ovaries of transgenic mice leads to decreased follicle apoptosis, enhanced folliculogenesis, and increased germ cell tumorigenesis. Endocrinology 137:4837– 4843 55. Bergeron L, Perez GI, Macdonald G, Shi L, Sun Y, Jurisicova A, Varmuza S, Latham KE, Flaws JA, Salter JC, Hara H, Moskowitz MA, Li E, Greenberg A, Tilly JL, Yuan J 1998 Defects in regulation of apoptosis in caspase-2deficient mice. Genes Dev 12:1304 –1314 56. Rinkenberger JL, Horning S, Klocke B, Roth K, Korsmeyer SJ 2000 Mcl-1 deficiency results in peri-implantation embryonic lethality. Genes Dev 14:23–27 57. Collins JA, Schandl CA, Young KK, Vesely J, Willingham MC 1997 Major DNA fragmentation is a late event in apoptosis. J Histochem Cytochem 45:923–934 58. Gondos B, Bhiraleus P, Hobel CJ 1971 Ultrastructural observations on germ cells in human fetal ovaries. Am J Obstet Gynecol 110:644 – 652 59. Telfer WH 1975 Development and physiology of the oocyte-nurse cell syncitium. Adv Insect Physiol 11:223–319 60. Byskov AG 1986 Differentiation of mammalian embryonic gonad. Physiol Rev 66:71–117