Di€Erential Expression and Regulation of the Retinoblastoma Family Of

Dierential expression and regulation of the retinoblastoma family of proteins during testicular development and spermatogenesis: roles in the control of germ cell proliferation, dierentiation and apoptosis

Wei Yan1,2, Jukka Kero1,2, Janne Suominen1 and Jorma Toppari*,1,2

1Department of Physiology, University of Turku, Kiinamyllynkatu 10, 20520, Turku, Finland; 2Department of Pediatrics, University of Turku, Kiinamyllynkatu 10, 20520, Turku, Finland

Normal spermatogenesis is highly dependent on well- hypophosphorylated form. During cell cycle it acts balanced germ cell proliferation, dierentiation, and through regulating a group of TFs that belong to E2F apoptosis. However, the molecular mechanisms that family. When the hypophosphorylated pRb is phos- govern these processes are largely unknown. Retinoblas- phorylated by cyclinD/CDK4/6 complexes during G1/S toma family proteins (pRb, p107 and p130) are transition and becomes hyperphosphorylated, it loses its potentially important regulators of cell growth, dier- ability to bind E2F family of TFs. Release of E2F TFs entiation and apoptosis. pRb has been shown to be results in activation of S phase genes and thus permits expressed in the rat testis and involved in the regulation the progression of cell cycle into S phase. pRb also of spermatogenesis. In the present study, the expression interacts with several dierentiation-speci®c TFs, such as and localization of the other two pRb family members, Myo D and NF-IL6, and activates transcription of genes p107 and p130, were analysed at both mRNA and that are required for cell cycle exit and onset of protein levels during testicular development and sperma- dierentiation program. The discovery of two relatives togenesis using Northern, Western blotting, immunohis- of pRb, p107 and p130, added complexity to our tochemistry, and in situ hybridization. Furthermore, understanding of the regulatory roles of pRb in the changes of levels and phosphorylation status of pRb control of cell cycle, apoptosis, and dierentiation family proteins in response to growth suppression and/or (Huppi et al., 1996; LeCouter et al., 1996; Mayol et al., apoptosis induction were investigated using a seminifer- 1993; Pertile et al., 1995). These three proteins share a ous tubule culture system and three animal models. Our highly homologous region that is composed of domains data suggest that: (1) pRb family proteins are dieren- A, B, and a spacer in between. It is also called pocket tially expressed in the rat testis and they function in a region and therefore they are often called pocket cell-type-speci®c manner during testicular development proteins. The pocket region is the functional domain of and spermatogenesis; (2) they participate in the control the proteins, since viral oncoprotein binding, tumori- of germ cell cycle and act in a cell cycle-phase-speci®c genic mutations and transcription factor interaction all fashion during germ cell proliferation, and (3) they are occur there. Intensive studies have been carried out to also involved in the regulation of apoptosis of germ cells analyse the physiological roles of this family of proteins and Leydig cells. Oncogene (2001) 20, 1343 ± 1356. over the past decade. Studies using knock-out models suggest that three pocket proteins are functionally Keywords: retinoblastoma protein family; spermato- redundant and inactivation of one pocket protein could genesis; cell cycle; apoptosis; dierentiation be compensated by the other two (LeCouter et al., 1998a,b; Mulligan and Jacks, 1998). On the other hand, a number of in vitro studies show that these three proteins Introduction have distinct biochemical and biological features in terms of binding partners of the E2F family TFs and The retinoblastoma susceptibility gene, Rb, was the ®rst timing of phosphorylation during cell cycle progression tumor suppressor gene described (Friend et al., 1987). and dierentiation (Dyson, 1998; Mayol et al., 1995). pRb has been well characterized as an important However, dierences in their physiological roles and regulator of cell proliferation, apoptosis and dierentia- potential coordination among the three members of pRb tion (for recent review see Lundberg and Weinberg, family still remains poorly understood. 1999; Kaelin, 1999). It functions mainly at transcrip- Spermatogenesis is a complex process in which germ tional level by binding transcription factors (TFs) in its cells undergo a number of mitotic divisions and two meiotic divisions, and ®nally dierentiate into a population of spermatozoa within the seminiferous epithelium. Normal spermatogenesis requires well *Correspondence: J Toppari, Department of Physiology, University balanced cell proliferation, dierentiation and apopto- of Turku, Kiinamyllynkatu 10, FIN-20520, Turku, Finland Received 13 November 2000; revised 5 January 2001; accepted 9 sis. Regulation of spermatogenesis involves endocrine, January 2001 paracrine, and autocrine factors, which form a complex Pocket proteins in the testis W Yan et al 1344 signaling network. Germ cells receive signals that are Localization of mRNAs and proteins of pRb family mediated by these factors and ®nally converge on the members in the immature and mature rat testes regulatory machinery that governs cell cycle entry, progression, and exit, as well as programmed cell In situ hybridization signals of p107 mRNA were death. However, the nature of the regulatory machin- detected near the basal compartment of seminiferous ery is still largely unknown. Previously, we have tubule (Figure 2A-A'). Higher magni®cations showed studied the expression and phosphorylation pattern of that the signals were mainly located in spermatocytes pRb during testicular development and spermatogen- and spermatogonia (Figure 2B-B'). p130 mRNA was esis (Yan et al., 1997). pRb appears to be involved in detected near the basal membrane of the seminiferous the regulation of spermatogonial proliferation and epithelium (Figure 3A-A'). Higher magni®cations Sertoli cell dierentiation in the adult rat testis. In revealed that they were localized in Sertoli cells and the present study, we made use of rat spermatogenesis spermatogonia (Figure 3B-B'). Sense controls gave as a model to analyse: (1) how the two retinoblastoma weak and homogenous background signals (Figure family proteins, p107 and p130, are expressed and 2C,C',D,D' and Figure 3C,C',D,D'). We also per- regulated during germ cell proliferation, dierentiation, formed in situ hybridization for pRb mRNA using a and apoptosis and (2) how their levels and phosphor- 32P-UTP-labeled riboprobe and the results were similar ylation status change during germ cell cycle progres- to that of our previous study using non-radioactive in sion and in response to growth-stimulating/inhibiting situ hybridization method (Yan et al. (1997); data not and apoptosis-inducing signals, both in vitro and in shown). vivo. In the immature rat testis at the age of 5 days of postnatal life (P5), pRb was detected in a proportion of Sertoli cells and spermatogonia; p107 was found in all Sertoli cells and some spermatogonia; p130 was Results expressed in all Sertoli cells, spermatogonia, and Leydig cells (1st panel, Figure 4). Upon P20, all three proteins Expression of pocket proteins p107 and p130 during were expressed in Sertoli cells at a higher level and in testicular development spermatogonia and spermatocytes at a lower level (2nd Both p107 mRNA and protein were detected through- panel, Figure 4). Upon p40, the localization of these out the testicular development, and the expression three proteins started to dierentiate: pRb was found in levels of p107 mRNA and protein were correlated with Sertoli cells and a proportion of spermatogonia; p130 each other (Figure 1A,B). Levels of both p107 mRNA was only expressed in somatic cell types including and protein were higher at days 10 and 20 of postnatal Sertoli cells, Leydig cells, and peritubular myoid cells; life than at other ages. The levels of hyperpho- p107 was localized to spermatogonia and meiotically sphorylated form of p107 changed synchronously with dividing germ cells including preleptotene, leptotene, those of total protein. In the adult rat testis, p107 pachytene spermatocytes (3rd panel, Figure 4). protein was detected in all stages of the seminiferous The dierential localization patterns remained as epithelial cycle (Figure 1C). However, the hyperpho- such until adulthood and all three proteins displayed sphorylated form of p107 appeared to be distributed in highly stage- and cell-speci®c expression pattern in the a stage-speci®c fashion, with higher levels at stages adult rat testis (4th and 5th panels, Figure 4). The XII ± XIV and lower levels at all other stages. localization and the mRNA and protein levels of these The expression pro®les of p130 mRNA and protein three pocket proteins were schematically summarized were similar to those of p107 during testicular in Figure 5. In Sertoli cells at stages VII ± VIII pRb development, with a peak at days 10 ± 20 of postnatal expression levels were the highest, while in spermato- life (Figure 1D,E). Levels of p130 mRNA correlated gonia of all stages the pRb levels were relatively well with those of protein. p130 protein could be constant (Figures 4 and 5A). The highest levels of p107 detected at all stages of the epithelial cycle with a were detected in spermatogonia of types A single (As) stage-speci®c phosphorylation pattern. More hyper- and A paired (Apr) at stages I ± III. p107 levels were phosphorylated forms were present at stages I ± VI than higher in spermatogonia of type A aligned (Aal), type at other stages (Figure 1F). A4 (A4), intermediate spermatogonia (In), and type B The expression levels of pRb mRNA during spermatogonia (B) at stages I ± VI. The expression testicular development and protein levels and phos- levels of p107 were relatively lower in spermatocytes phorylation status of pRb have been reported earlier than in spermatogonia. However, from preleptotene to (Yan et al., 1997). Here we analysed pRb expression late pachytene the expression levels of p107 increased and phosphorylation in the immature and mature rat and p107 protein expression ceased in step 1 testes using a dierent pRb antibody. pRb levels were spermatids right after meiotic cell divisions (Figures 4 continuously reduced during the testicular develop- and 5B). Higher expression levels of p130 were found ment, with higher levels prior to day 20 of postnatal in Sertoli cells at stages I ± VI than at other stages life and lower levels afterwards (Figure 1G). The stage- (Figures 4 and 5C). speci®c expression and phosphorylation pattern in the To see whether the undierentiated stem spermato- adult rat testis was similar to that previously reported gonia expressed p130, thin (3 mm) and consecutive (Yan et al., 1997); (data not shown). sections were stained with antibodies against p27kip1

Oncogene Pocket proteins in the testis W Yan et al 1345

Figure 1 Expression of mRNA and protein of pRb family members during testicular development and spermatogenesis. The ®gures show: Northern analysis of p107 (A) and p130 (D) mRNA expression during testicular development; Western blotting analysis of p107 (B), p130 (E), and pRb (G) expression and phosphorylation during testicular development; expression and phosphorylation of p107 (C) and p130 (F) at 10 stages of epithelial cycle of the seminiferous tubules. The hyperphosphorylated forms of pRb, p107, and p130 are shown as pRb-p, p107-p and p130-p. Roman Numerals represent stages of the rat seminiferous epithelium. Beta-actin was used as a loading control. NB, newborn; 1d ± 60d ; days 1 ± 60 after birth

(a Sertoli cell-speci®c protein, Beumer et al., 1999) and indicating that undierentiated spermatogonia did p130. Some ¯atten-shaped nuclei on the basal express p130 (Figure 6). Furthermore these cells were membrane are p27kip1 negative, but p130-positive, c-kit negative (Figure 6).

Oncogene Pocket proteins in the testis W Yan et al 1346

Figure 2 p107 mRNA localization in the rat testis. The bright Figure 3 p130 mRNA localization in the rat testis. The bright ®eld is shown in the left panel and the corresponding dark ®eld is ®eld is shown in the left panel and the corresponding dark ®eld is shown on the right. The speci®c signals are located in the basal shown on the right. The speci®c signals are con®ned to the basal compartment of the seminiferous epithelium (A, A'). Higher membrane of the seminiferous epithelium (A, A'). Higher magni®cation (B, B') shows the speci®c signals in spermatogonia magni®cation (B, B') displays the speci®c signals in Sertoli cells (Sg) and spermatocytes (Sp). The background signals were (Sc). The background signals were monitored according to the evaluated according to the sense probe (C, C', D, D'). Roman sense probe (C, C', D, D'). Roman Numerals represent stages of Numerals represent stages of the rat seminiferous epithelium. the rat seminiferous epithelium. Bars=50 mm Bars=50 mm

previously (Yan et al., 2000c). Total levels of p107 and Expression levels and phosphorylation status of pocket pRb increased and they both became more hyperpho- proteins in the cultured seminiferous tubules in the sphorylated after 48 h of culture in the presence of presence or absence of SCF SCF (Figure 7). In contrast, total protein levels of p107 When the seminiferous tubules from stage XII of the and pRb decreased dramatically and they became more epithelial cycle were cultured in the presence of SCF hypophosphorylated in the absence of SCF. No (100 ng/ml), the [3H]-thymidine uptake was signi®- signi®cant change was found in p130 levels and cantly elevated after 48 and 72 h culture in vitro in phosphorylation status (data not shown). comparison to the controls (Hakovirta et al., 1999). Moreover, in the absence of SCF, the number of Expression levels and phosphorylation status of pocket apoptotic germ cells was much higher than in the proteins in response to inhibition of spermatogonial presence of stem cell factor during 8 ± 72 h culture proliferation and induction of spermatogonial apoptosis (Yan et al., 2000c). To see how the expression levels by ACK-2 treatment and phosphorylation status of pocket proteins change in response to SCF/c-kit-mediated growth stimulating Administration of a monoclonal anti-c-kit antibody, and pro-survival signals in vitro, Western blotting ACK-2, causes reduction of BrdU-positive spermato- analyses were performed using proteins isolated from gonia in number and induction of apoptosis of the cultured seminiferous tubules at dierent time dierentiating spermatogonia (Tajima et al., 1994; points under the same culture conditions as described Yan et al., 2000a; Yoshinaga et al., 1991). As reported

Oncogene Pocket proteins in the testis W Yan et al 1347

Figure 4 Immunohistochemical detection of pocket proteins in the immature and mature rat testes. The ages of the testes are marked on the left. Sc, Sertoli cell; Sg, spermatogonia; Sp, spermatocyte; Lc, Leydig cell; Mc, peritubular myoid cell. Bars=50 mm previously, the number of BrdU-positive spermatogo- of spermatocyte apoptosis (Figure 9A). Spermatocytes nia started to decrease at 48 h after ACK-2 adminis- at stages VII ± XIV started to be TUNEL-positive at tration and was signi®cantly reduced after 72 h. 12 h after administration and the number of Concomitantly, TUNEL-positive spermatogonia sig- TUNEL-positive spermatocytes peaked at 24 h after ni®cantly increased in number 48 h after ACK-2 MAA treatment (Figure 9A). In the testis 12 h after treatment (Yan et al., 2000a). pRb level decreased MAA treatment, most of pachytene spermatocytes signi®cantly and it became more hypophosphorylated appeared to be weakly stained by anti-p107 antibody after 48 h of ACK-2 treatment (Figure 8A,B). Total and some were even negative, while in the control protein levels of p107 remained unchanged, but p107 testis pachytene spermatocytes were all strongly became more hypophosphorylated 72 h after the stained (Figure 9B). Western blotting analysis treatment (Figure 8C,D). p130 was not signi®cantly revealed that both total p107 protein levels and the changed (data not shown). hyperphosphorylated form of p107 decreased signi®- cantly at 12 and 24 h after MAA administration (Figure 9C). Levels of pRb and p130 were slightly Expression levels and phosphorylation status of pocket elevated but the phosphorylation status was not proteins during spermatocyte apoptosis induced by changed signi®cantly during the same time period methoxyacetic acid (MAA) (data not shown). Methoxyacetic acid (MAA) is able to induce speci®cally spermatocyte apoptosis within 24 h when Expression levels and phosphorylation status of pocket applied orally at a dose of 650 mg/kg in the rat proteins in EDS-treated rat testis (Krishnamurthy et al., 1998; McKinnell and Sharpe, 1997; Suter et al., 1998). TUNEL staining was Ethylene dimethane sulfonate (EDS) induces speci®- performed to monitor the time course of induction cally depletion of Leydig cells through apoptosis within

Oncogene Pocket proteins in the testis W Yan et al 1348

Figure 6 Immunohistochemical staining of three consecutive sections (3 mm thick) using antibodies against p27Kip1 (A), p130 (B), and c-kit (C). Sc, Sertoli cells; Sg, spermatogonia. Two bigger arrows point to two p130-positive, but p27- and c-kit- negative stem spermatogonia. All micrographs are in the same magni®cation. Bar=50 mm

after the treatment (Yan et al., 2000d). Following Figure 5 Schematic illustration of localization and expression Leydig cell depletion, germ cells undergo massive pattern of pRb (A), p107 (B), and p130 (C) in the rat seminiferous apoptosis during days 7 ± 15 due to androgen with- epithelium. Cells within the frame express a pocket protein and drawal (Figure 10A). BrdU incorporation study the width of the frame represents the relative abundance of the showed that spermatogonial proliferation was inhibited protein. The speci®c cell associations in the vertical columns represent speci®c stages (Roman numerals) of the epithelial cycle. during days 4 ± 20 (Figure 10B,C). After day 20, the Sc, Sertoli cells; A1-4, type A spermatogonia; In, intermediate number of BrdU-positive spermatogonia and prelepto- spermatogonia; B; type B spermatogonia; Pl, preleptotene tene and leptotene spermatocytes started to recover spermatocytes; L, leptotene spermatocytes; Z, zygotene sperma- and reached the control levels upon day 40 (Figure tocytes; P, pachytene spermatocytes; Di, diplotene spermatocytes; 10B,C). m, meiosis. Cells that are not marked are spermatids. Cells that undergo mitotic division are marked with asterisks Levels of p130 declined dramatically and it became more hypophosphorylated during the ®rst 4 days after EDS administration, when Leydig cells were undergoing massive apoptosis. Total protein levels 4 days and precursor Leydig cells start to proliferate and the hyperphosphorylated form of p130 started to from day 2 after EDS treatment, and ®nally, Leydig increase at day 20 after EDS treatment, when cell population can be fully regenerated around day 40 regenerated Leydig cell population became prominent

Oncogene Pocket proteins in the testis W Yan et al 1349

Figure 7 Expression and phosphorylation of p107 (A) and pRb (C) in the seminiferous tubule segments at stage XII of the epithelial cycle cultured in the presence (SCF+) or absence (SCF7) of stem cell factor (100 ng/ml). Quanti®cation of Western results of three independent experiments (B for p107 and D for pRb). Each bar represents mean+s.e.m. (n=3). ADU, arbitrary densitometric unit. *P50.05, as compared with controls (0 h)

Figure 8 Changes of expression levels and phosphorylation of p107 (A) and pRb (C) in response to inhibition of spermatogonial proliferation and induction of spermatogonial apoptosis by ACK-2 treatment. Histograms (B for p107 and D for pRb) show quantitative analysis of the Western results of three independent experiments. Each bar represents mean+s.e.m. (n=3). ADU, arbitrary densitometric unit. *P50.05, as compared with controls

and some of them started to produce testosterone 10F,G). Levels of total pRb did not change (Figure 10D,E). Total p107 levels decreased signi®- signi®cantly, but the hyperphosphorylated form of cantly and p107 became more hypophosphorylated pRb declined signi®cantly during days 7 ± 20 after during days 3 ± 10 after EDS treatment (Figure EDS treatment (Figure 10H,I).

Oncogene Pocket proteins in the testis W Yan et al 1350

Figure 9 Changes of expression levels and phosphorylation status of p107 during spermatocyte apoptosis induced by methoxyacetic acid (MAA). (A) TUNEL staining of the rat testes at 0 h (Control), 6, 12 and 24 h after MAA treatment. (B) p107 expression in the testis 12 h after MAA administration. Arrows point to p107-negative pachytene spermatocytes at stage IX in the MAA-treated testis (right), which are strongly p107-positive in the control testes (left). (C) Western analysis of p107 levels and phosphorylation status during induction of spermatocyte apoptosis by MAA. (D) Quantitative analysis of Western blotting results from three independent experiments. Each bar represents mean+s.e.mean (n=3). ADU, arbitrary densitometric unit. *P50.05, as compared with controls

Discussion tically proliferating spermatocytes and dierentiating spermatids. Germ cells at dierent stages are in In the present study, we found that pocket proteins dierent phases of cell cycle. Therefore, accurate pRb, p107, and p130 are tightly and dierentially staging enables us to determine the cell cycle phase regulated during testicular cell proliferation in the of a speci®c germ cell. For example, all pachytene immature testis and germ cell proliferation and spermatocytes are in G2 phase of meiosis and all type dierentiation during spermatogenesis in the adult A spermatogonia are at stage I, including A single testis. On the basis of our ®ndings, three major (As), A paired (Ap), A align (Aal), and A4 conclusions could be drawn. First, three pocket spermatogonia (A4), are in S phase of mitosis (de proteins act dierentially during germ cell cycle Rooij and Grootegoed, 1998). All spermatids have progression in terms of changes in their expression exited from cell cycle and all are in dierentiation levels and phosphorylation status. p107 appears to be status (G0). Preletotene spermatocytes at stages VII ± the only pocket protein that is expressed during VIII are in S phase of meiosis. meiosis. Secondly, pRb family proteins function in a cell-type-speci®c fashion during testicular development p130 Our in situ hybridization and immunohistochem- and spermatogenesis. For example, Leydig cells only ical staining clearly showed that p130 was only expressed express p130 and p130 is the only pocket protein that in terminally dierentiated somatic cells including Sertoli is involved in Leydig cell proliferation and dierentia- cells, Leydig cells, and peritubular myoid cells in the tion. Thirdly, dierent settings of changes in expression adult rat testis, suggesting that p130 may be responsible levels and phosphorylation status of pocket proteins for the exit of cell cycle and maintenance of the quiescent correspond to germ cell proliferation, dierentiation, G0-state of the cells. Previously it has been shown that and apoptosis. cell cycle exit caused by serum deprivation leads to an accumulation of active hypophosphorylated p130 protein and p130/E2F-4 complexes in numerous cell systems Differential expression and phosphorylation of pocket (Stiegler et al., 1998). Our ®nding further supports that proteins during germ cell cycle progression in the rat testis p130 is the main representative of the pRb family in G0- Rat spermatogenesis can be divided into 14 stages state cells. However, it is noteworthy that both total according to speci®c cellular associations at speci®c protein levels and phosphorylation status of p130 were locations of the seminiferous tubules. Each stage stage-speci®c, with higher expression levels and more contains mitotically proliferating spermatogonia, meio- hyperphosphorylated states at stages I ± VI than at the

Oncogene Pocket proteins in the testis W Yan et al 1351

Figure 10 Expression and phosphorylation status of pocket proteins in EDS-treated rat testis. (A) TUNEL staining of testes at dierent time points after EDS treatment. (B) Quantitative analysis of BrdU content in germ cells at stages VIII and I after EDS treatment. Insert A in (b) shows immunohistochemical detection of incorporated BrdU in seminiferous epithelium of stage VIII. Insert B in (b) shows a representative result of Southern-Western-based BrdU content assay for stages VIII and I. (C) Quantitative analysis of BrdU content in seminiferous tubules at dierent time points after EDS treatment. Insert shows a representative result of Southern-Western-based BrdU content assay. D±I Show Western analysis of protein levels, phosphorylation status, and quanti®cation of the Western results (D, E for p130, F, G for p107, and H, I for pRb). Each bar represents mean+s.e.m. (n=3). ADU, arbitrary densitometric unit. *P50.05, as compared with controls

Oncogene Pocket proteins in the testis W Yan et al 1352 rest of the stages of the epithelial cycle. It implies that have shown that pRb could interact with a number of p130 might be involved in the regulation of function of proteins that are responsible for terminating cell cycle these somatic cell types during spermatogenesis. p130 and inducing tissue-speci®c gene expression and there- has been identi®ed as a binding partner of HBP1, a TF by regulating function of the tissue, such as myoD, that plays an essential role during dierentiation of HBP-1 (Shih et al., 1998; Tevosian et al., 1997; Yee et myoblasts. More and more p130 binding partners are al., 1998). It has been recently established that in the being discovered, which are mostly TFs and function in prostate, pRb could directly bind androgen receptor regulating tissue-speci®c gene expression. It will be of (AR) and regulate AR-mediated gene transcription interest to see whether p130 could interact with these TFs (Gao et al., 1999; Lu and Danielsen, 1998; Yeh et al., and thereby control testis-speci®c gene expression in 1998). In the rat seminiferous epithelium, stages VII testicular somatic cells. and VIII are preferentially regulated by androgens (Parvinen et al., 1986). More intriguingly, the stage- p107 In the adult rat testis, strong expression of p107 speci®c expression pattern of pRb in Sertoli cells in mitotically and meiotically cycling germ cells, and (highest at stages VII ± VIII) disappeared during days absence in quiescent and dierentiated cells strongly 2 ± 20 after EDS treatment (data not shown), suggest- support that p107 is the most prominent pocket ing that pRb levels might be aected by androgens, protein during late G1 phase through G2/M. Interest- since testosterone levels were undetectable during this ingly, p107 levels were higher in the mitotically period of time (Henriksen et al., 1995; Tena-Sempere et dividing spermatogonia than in the meiotically dividing al., 1994). Therefore, it would be of interest to analyse spermatocytes. The levels of p107 increased with the whether pRb could also interact with AR in Sertoli progression of the meiotic cell cycle from G1 to S, G2 cells and thereby regulate Sertoli cell function. and M phases. These variable expression levels were more prominent in spermatogonia than in spermato- Cell type-specific expression of pocket proteins during cytes. This feature makes it dierent from pRb, which testicular development and spermatogenesis is also expressed in mitocally dividing spermatogonia and showed relatively constant levels during cell cycle Rat spermatogenesis is activated soon after birth. progression. Expression of p107 has been shown to be During the ®rst 5 days of postnatal life, mitotically strictly regulated by its E2F-dependent promoter arrested gonocytes within the testicular cords re-enter (Nevins, 1998). Our ®nding appears to support a the cell cycle, dierentiate into spermatogonia and current model in which p130 can actually repress p107 proliferate. At approximately P10, spermatogonia start expression, and accumulation of p130 upon cell cycle to dierentiate into spermatocytes and initiate meiosis. exit may account for the absence of p107 in quiescent By P20, round spermatids become evident and by or dierentiated cells (Grana et al., 1998; Stiegler et al., P35 ± 40, spermatozoa appear in the lumen of 1998). seminiferous tubules. Subsequent spermatogenic waves are maintained through continuous stem spermatogo- pRb The retinoblastoma protein was detected not nial self-renewal and re-entry of the proliferating pool only in all mitotically dividing spermatogonia, but also of spermatogonia. On the basis of dierential in terminally dierentiated Sertoli cells by using the localization in the immature and mature testes, we monoclonal antibody raised against amino acids 332 ± found that the pocket proteins function in a cell-type- 344 of human pRb. This result is partially consistent speci®c manner. In germ cells, undierentiated with our previous data showing that pRb was mainly spermatogonia that are in quiescent state express in the nuclei of Sertoli cells and spermatogonia and p130 but not pRb. Once they are activated to enter was also found in the cytoplasm of steps 14 ± 19 cell cycle G1 phase, those type A spermatogonia start spermatids using a polyclonal antibody raised against, to express pRb and p107 instead of p130. Therefore, c-terminus of pRb (Yan et al., 1997). The discrepancy p130 could be a marker for undierentiated germ appears to be caused by the dierent epitopes that cells. During mitotic cell cycle progression, both pRb these two antibodies were raised against, since in situ and p107 were expressed and the levels of pRb hybridization data of the present ([35S]-radioactive in remained relatively constant, but p107 levels increased situ hybridization) and previous (Dig+-based non- continuously with the progression of cell cycle. These radioactive in situ hybridization) studies consistently ®ndings are consistent with earlier report (Garriga et showed that Rb mRNA were present in Sertoli cell, al., 1998; Yan et al., 1997) and suggest that pocket spermatogonia, and spermatocytes, as well as early proteins may function not only through phosphoryla- spermatids. As reported in our earlier study (Yan et al., tion, but through changes in expression levels as well. 1997), a shorter transcript (2.7 kb) of pRb mRNA When type B spermatogonia dierentiate into pre- could be stored in spermatocytes and early spermatids leptotene spermatocytes and meiotic S phase starts, and translated into protein later in elongated sperma- p107 remained to be the only pocket protein. After tids. pRb expression levels did not change during cell meiotic cell divisions, spermatocytes become haploid cycle progression of spermatogonia. However, drastic spermatids. Neither p130 nor p107 was detected in elevation of pRb protein levels in Sertoli cells at stages spermatids, but pRb could be detected as reported VII ± VIII may suggest a potential role in the previously (Yan et al., 1997). p107 appears to be the regulation of Sertoli cell function. Numerous studies only pocket protein that is expressed during meiosis

Oncogene Pocket proteins in the testis W Yan et al 1353 and pocket proteins function in a cell type-speci®c phosphorylation status might be due to the similar manner in the rat testis. eect of SCF on meiotically dividing germ cells. During testicular development, all somatic cells, Administration of a monoclonal anti-c-kit antibody, including Sertoli cells and Leydig cells, are actively ACK-2, could inhibit BrdU incorporation of sperma- proliferating. Sertoli cells cease proliferation and start togonia and preleptotene spermatocytes, and induces to dierentiate by P20 in the rat. In proliferating apoptosis of dierentiating spermatogonia including Sertoli cells, three pocket proteins were all expressed, type A1±4 spermatogonia (Yoshinaga et al., 1991). In while p107 was absent when Sertoli cells ceased response to ACK-2 treatment, both pRb and p107 proliferation and started to dierentiate upon puberty. became more hypophosphorylated, in accordance with This ®nding supports the current thought that p107 is the reduced number of BrdU-positive spermatogonia at not expressed in terminally dierentiated cells, where stages I ± III, IV, VI and IX ± XIV and preleptotene p130 and pRb are frequently present (Stiegler et al., spermatocytes at stages VII and VIII. Severely reduced 1998). levels of pRb were correlated with massive spermato- Mesenchymal stem cells, present in the earliest stage gonial apoptosis and unchanged levels of p107 might of testicular development, serve as a source of Leydig be due to lack of spermatocyte apoptosis. This cells. Some of them dierentiate into fetal type of observation supports two recent studies showing that Leydig cells, while some proliferate and give rise to the stimulatory eect of SCF/c-kit on spermatogonial adult type of Leydig cells through a series of proliferation is actually mediated through PI3 kinase/ maturation steps including progenitor Leydig cells at Akt/p70 S6 kinase/cyclin D3/pRb pathway (Blume- P14 ± 28, immature Leydig cells at P28 ± 56, and ®nally Jensen et al., 2000; Feng et al., 2000; Kissel et al., adult type of Leydig cells from p56 onward in the rat 2000). The signaling pathway that is used for (Hardy et al., 1989). Interestingly, p130 was the only connecting SCF/c-kit interaction with germ cell pocket protein that was expressed in Leydig cells in apoptosis machinery remains unknown currently. both immature and mature testes. During days 3 ± 40 after EDS treatment, Leydig cell population could be Reduced levels and phosphorylation of p107 correspond to regenerated by precursor Leydig cells. During pre- spermatocyte apoptosis in MAA-treated rats cursor Leydig cell proliferation, we failed to detect either pRb or p107. P130 was steadily expressed in The apoptotic nature of the spermatocyte depletion by proliferating precursor Leydig cells. This further MAA has been shown by a number of groups supports that p130 may have cell-speci®c function in (Krishnamurthy et al., 1998; McKinnell and Sharpe, controlling Leydig cell proliferation and dierentiation. 1997; Suter et al., 1998). MAA-treated rats have Stage-speci®city of p130 expression in all somatic cells become a good model to study apoptosis of sperma- within the testis during spermatogenesis strongly tocytes although the mechanism by which MAA implicates that p130 may regulate functions of these induces apoptosis of spermatocytes in such a speci®c cells. fashion remains unclear. Only p107 changed in both protein levels and phosphorylation status, further con®rming that it is the only pocket protein that was Stem cell factor enhances pRb expression and expressed by spermatocytes. Signi®cantly reduced levels phosphorylation in vitro and blockade of SCF/c-kit of total protein and more hypophosphorylated status interaction in vivo suppresses levels and phosphorylation correspond to massive apoptosis of spermatocytes. of pRb: correlation with spermatogonial proliferation and Earlier we have shown that during spermatocyte apoptosis apoptosis induced by MAA, pro-apoptotic members, It is known that in the presence of stem cell factor including Bax and Bak, were upregulated and pro- (SCF) [3H]-thymidine incorporation of the seminiferous survival Bcl-2 family members, e.g. Bcl-w, was down tubules from stage XII was enhanced and apoptosis regulated (Yan et al., 2000a). Changes in p107 levels was suppressed signi®cantly after 48 ± 72 h culture in and phosphorylation status seem to support a role of vitro in comparison to controls (Hakovirta et al., 1999). p107 in protecting spermatocytes from apoptosis. A Recent studies also show that in the presence of SCF certain amount and more phosphorylated status of both mitotic and meiotic DNA synthesis of germ cells p107 correspond to meiotic cell cycle progression of could be maintained in vitro for at least 72 h and germ spermatocytes. Otherwise, they may undergo apoptosis. cells could pass through meiosis and further dier- pRb has a role in protecting cells from apoptosis entiate into spermatids (Vincent et al., 1998). In the (Harbour and Dean, 2000; Kasten and Giordano, present study, pRb and p107 levels increased and they 1998) although the mechanism has not been fully became more hyperphosphorylated during 24 ± 72 h of clari®ed. However, release of E2F-1 might at least culture in the presence of SCF. In contrary, their levels partially account for the induction of apoptosis by decreased and they became more hypophosphorylated inactivation of pRb (Kasten and Giordano, 1998). in the absence of SCF. Since pRb is mainly expressed p107 and p103 preferentially bind E2F-4 and E2F-5, in proliferating spermatogonia, it is likely that the while pRb binds E2F-1-3. E2F-1 could initiate changes of pRb expression levels and phosphorylation apoptosis, while other members of E2F family have status re¯ect the growth-stimulating and pro-survival not been shown to play a similar role. The present eect of SCF. Similarly, changes in p107 levels and ®nding that p107 might be involved in spermatocyte

Oncogene Pocket proteins in the testis W Yan et al 1354 apoptosis raised the possibility that either p107 could have not been demonstrated to be able to induce also interact with E2F-1, or E2F-4/5 has an apoptosis- apoptosis. It is therefore intriguing to ®nd that both inducing eect as well, or other p107 binding partner(s) p107 and p130 declined in expression levels and became may be capable of inducing apoptosis. more hypophosphorylated when cells undergo apoptosis, since it suggests that either E2F-4 or E2F-5 could possibly induce apoptosis, or p130 and p107 could also Reduced levels and less phosphorylated states of pocket bind E2F-1 or other factors, which could induce proteins correlate with proliferation inhibition and/or apoptosis. Clearly, studies on E2F family of proteins apoptosis of Leydig cells and germ cells in the during spermatogenesis are required to get more insight EDS-treated rats into the potential pro-survival function of p107 and Leydig cells are depleted within 4 days and testosterone p130. declined to undetectable levels at day 2 after EDS Taken together, we have shown in the present study treatment (Yan et al., 2000d). Precursor Leydig cells that the retinoblastoma family proteins are dieren- start to proliferate soon after mature Leydig cells tially expressed in the testis and they function in a cell undergo apoptosis, and the proliferation of Leydig cells type-speci®c way. Their levels and phosphorylation peaks at days 2 ± 7, day 10 and day 20 (Yan et al., 2000d). status are modulated during progression of cell cycle The Leydig cell population is ®nally regenerated around and apoptosis. Reduced levels and more hypophos- day 40 after EDS treatment. As a consequence of phorylated states correspond to suppression of pro- testosterone withdrawal, germ cells undergo massive liferation and/or apoptosis of germ cells and Leydig apoptosis between days 7 ± 15. A BrdU incorporation cells in the rat testis. This family of proteins may serve experiment revealed that DNA synthesis is inhibited as markers for evaluation of cell cycle status of germ during the same period of time. Since levels of both cells during spermatogenesis. gonadotropins, LH and FSH, are higher than normal levels between days 2 ± 20 after EDS treatment (Yan et Materials and methods al., 2000d), the germ cell abnormalities appear to result from testosterone withdrawal. It is noteworthy that Experimental animals and treatments during the ®rst 2 days after EDS treatment, when Sprague-Dawley male rats at the ages of 1 day, 5 days, 10 massive Leydig cell apoptosis took place, only p130 days, 20 days, 30 days, 40 days, and 2 ± 3 months, were used declined dramatically and became more hypophos- as experimental animals. They were housed two per cage in a phorylated. The changes suggest that p130 might be controlled environment at 218C with a 14 h lightness and capable of protecting Leydig cells from apoptosis in a 10 h darkness cycle with free access to water and food. normal situation. Since Sertoli cells and other somatic All animal experiments were approved by the Turku cells in the testis are morphologically and functionally University Committee on Ethics of Animal Experimentation. normal, it is unlikely that the changes of p130 were also To inhibit proliferation of spermatogonia and induce contributed by other somatic cells, even though it can apoptosis of dierentiating spermatogonia by blocking SCF/c-kit interaction (Tajima et al., 1994; Yoshinaga et al., not be excluded. More interestingly, signi®cantly re- 1991), a function-blocking anti-c-kit antibody, ACK-2 (kindly duced levels and more hypophosphorylated states of provided by Dr T Kunisada, Department of Immunology, p107 and pRb corresponded very well with inhibited Faculty of Medicine, Tottori University, Japan) was injected DNA synthesis of germ cells and massive germ cell i.v. at 3.5 mg/kg BW in physiological saline. The detailed apoptosis during days 7 ± 20 after EDS treatment. The procedure was described previously (Yan et al., 2000d). changes of p107 could be ascribed to spermatogonia and Brie¯y, injection was given twice, once every 48 h. Rats that spermatocytes, and changes of pRb may be due to received ACK-2 injection were sacri®ced at 24, 48, 72 and reduced number of BrdU-positive spermatogonia and 96 h after the ®rst injection, respectively. To monitor germ increased number of TUNEL-positive spermatogonia. It cell proliferation, 8-bromodeoxyuridine (BrdU, Boehringer is not surprising to observe that p130 levels and Mannheim, Germany) was i.p. injected at a dose of 50 mg/kg B.W. 1 h before sacri®ce. One testis was snap frozen in liquid phosphorylation status declined during days 7 ± 20 after nitrogen and then stored at 7708C for protein preparation, EDS treatment since precursor Leydig cells were actively the other was ®xed overnight at 48C in 4% paraformaldehyde proliferating and expression levels of p130 were lower (PFA) followed by dehydration and embedding into paran than in dierentiated status. However, it is indeed for immunohistochemistry and TUNEL staining. surprising that during precursor Leydig cells prolifera- For induction of spermatocyte apoptosis, methoxyacetic tion, neither p107 nor pRb was detected. Therefore, p130 acid (MAA, Aldrich Chemie, Steinheim, Germany) was appears to function alone during Leydig cell cycle diluted in physiological saline and administered orally at a progression, exit, and dierentiation of precursor Leydig dose of 650 mg/kg B.W. (Krishnamurthy et al., 1998; cells. The repopulation process of Leydig cells after EDS McKinnell and Sharpe, 1997; Suter et al., 1998). The control treatment mimics the adult-type Leydig cell development rats received physiological saline. The rats were sacri®ced at 6, 12 and 24 h, respectively. Sample collection was performed in the immature testis (Teerds, 1996). Therefore, we as described above. revisited the latter process and found that indeed neither To study the response of pocket proteins to Leydig cell process involved p107 or pRb. It is further con®rmed apoptosis after EDS treatment, and proliferation inhibition that p130 is the only pocket protein that is expressed in and apoptosis of germ cells due to testosterone withdrawal Leydig cells at least in the rat testis. Both p107 and p130 after EDS treatment, the rats were i.p. injected with a single preferentially bind E2F-4/E2F-5. However E2F-4 and -5 dose of EDS (75 mg/kg BW). EDS was synthesized as

Oncogene Pocket proteins in the testis W Yan et al 1355 previously described (Jackson and Jackson, 1984) and Immunohistochemistry dissolved in dimethylsulfoxide (DMSO)-water (1 : 3, Vol./ Vol.). Control animals (three rats/time point) for all time Immunohistochemical staining was carried out as described points received injection of vehicle. Rats (n=3/group) were previously (Yan et al., 2000a). The dilutions of the antibodies killed by cervical dislocation under CO2 anesthesia at 1, 2, 3, were 1 : 1000 for anti-pRb, 1 : 500 for anti-p130, and 1 : 250 4, 7, 10, 20 and 40 days after administration of EDS. To for anti-p107. Immunohistochemical detection of incorpor- monitor germ cell proliferation, 8-bromodeoxyuridine (BrdU, ated BrdU was performed using a monoclonal anti-BrdU Boehringer Mannheim, Mannheim, Germany) was i.p. antibody (at 1 : 200; Boehringer Mannheim, Mannheim, injected at a dose of 50 mg/kg B.W. 1 h before sacri®ce. Germany). A mouse anti-p27kip1 monoclonal antibody (F- Sample collection was performed as described above. 8) and a rabbit anti-c-Kit polyclonal antibody (M-14) were purchased from Santa Cruz Biotechnology Inc. (San Diego, Riboprobe preparation CA, USA). The primary antibodies preabsorbed with 1006 excess amount of corresponding immunizing peptides (from The pRb cDNA was kindly provided by Dr Rene Bernards the sources where the antibodies were purchased) were used (The MGH Cancer Center, USA). A 985-bp-long fragment for negative controls. (SacI/PstI) was sub-cloned into pGEM 4Z vector (Promega, Madison, WI, USA) for preparation of riboprobes for Northern and in situ hybridization experiments. Plasmids TUNEL staining containing p130 and p107 cDNA were gifts from Dr Michael TUNEL staining was performed as described previously (Yan A Rudnicki (McMaster University, Hamilton, Ontario, et al., 2000c). Canada) and Dr Loren J Field (Indiana University School of Medicine, Indiana, USA), respectively. A 2551-bp-long fragment from p130 cDNA and a 1845-bp-long fragment Southern ± Western-based detection of incorporated BrdU from p107 cDNA were subcloned into pGEM 4Z vector. DNA isolation, slot blotting, antibody incubation, and p130/pGEM4Z (XhoI/PuvII) and p107/pGEM 4Z (DraI/PstI) chemiluminescent detection were performed as described were linearized for preparation of antisen/sense probes for previously (Yan et al., 2000d). Northern and in situ hybridization. For Northern hybridization, [32P]-UTP (Amersham, Aylesbury, UK) was used for labeling the antisense riboprobe. For in situ hybridization, Quantitative analysis of Northern hybridization and [35S]-UTP (Amersham) was used for labeling both antisense Western blotting results and sense probes. Scanning and densitometric analysis were carried out as described previously (Yan et al., 2000b). After normalization Northern hybridization to b-actin protein, the densitometric value of an extra control RNA preparation, gel fractionation, and hybridization were sample was designated as 100% (excluded from statistical performed as described previously (Yan et al., 1999). analysis) and values of other controls and treated samples were expressed as the percentages of the designated one. In situ hybridization Statistical analysis Five-mm-thick sections were cut from paran-embedded testis samples and mounted onto SuperForst1 Plus glass The values from three independent experiments were pooled slides (Menzel-GlaÈ ser, Steinheim, Germany). The slides were for the calculation of the means and their standard errors incubated at 378C overnight and then stored at 48C before (s.e.m.) and for one-way analysis of variance and Duncan's use. In situ hybridization was performed as described new multiple range test to determine the signi®cant previously (Yan et al., 2000a). dierences between dierent experimental groups by using StatView 4.51 statistic program (Abacus Concepts Inc., Western blotting Berkeley, CA, USA). The P values less than 0.05 were considered statistically signi®cant. Western blotting was performed as described in details previously (Yan et al., 2000b). The primary antibodies used included a monoclonal mouse anti-human pRb antibody (at 1 : 500 dilution; Pharmingen, San Diego, CA, USA), a polyclonal rabbit anti-p130 antibody (at 1 : 250, C-20; Santa Cruz, CA, USA), or a polyclonal rabbit anti-p107 antibody (at 1 : 200 dilution, C-19; Santa Cruz, CA, USA), as well as a mouse anti-Actin monoclonal antibody (at 1 : 1000; ICN Acknowledgments Biomedicals Inc., Aurora, OH, USA). We would like to thank Drs Rene Bernards (The MGH Cancer Center, USA), Michael A Rudnicki (McMaster University, Hamilton, Ontario, Canada), and Loren J Field Tissue culture and stimulation with Stem cell factor (Indiana University School of Medicine, Indiana, USA), Transillumination-assisted microdisection and culture of for providing pRb, p107 and p130 cDNA plasmids seminiferous tubule segments from stage XII were performed respectively. Johanna Vesa is acknowledged for excellent as described previously (Yan et al., 2000c). SCF was technical assistance. This work was supported by the purchased from Genzyme Transgenics Corp. (Cambridge, Academy of Finland, Turku University Central Hospital, MA, USA) and the concentration used was 100 ng/ml. and EU contract.

Oncogene Pocket proteins in the testis W Yan et al 1356 References

Beumer TL, Kiyokawa H, Roepers-Gajadien HL, van den Mayol X, Garriga J and Grana X. (1995). Oncogene, 11, Bos LA, Lock TM, Gademan IS, Rutgers D, Ko A and 801 ± 808. de Rooij DG. (1999). Endocrinology, 140, 1834 ± 1840. Mayol X, Grana X, Baldi A, Sang N, Hu Q and Giordano A. Blume-Jensen P, Jiang G, Hyman R, Lee KF, O'Gorman S (1993). Oncogene, 8, 2561 ± 2566. and Hunter T. (2000). Nat. Genet., 24, 157 ± 162. McKinnell C and Sharpe RM. (1997). Int. J. Androl., 20, de Rooij DG and Grootegoed JA. (1998). Curr. Opin. Cell 171 ± 179. Biol., 10, 694 ± 701. Mulligan G and Jacks T. (1998). Trends Genet., 14, 223 ± 229. Dyson N. (1998). Genes Dev., 12, 2245 ± 2262. Nevins JR. (1998). Cell Growth Dier., 9, 585 ± 593. Feng LX, Ravindranath N and Dym M. (2000). J. Biol. Parvinen M, Vihko KK and Toppari J. (1986). Int. Rev. Chem., 275, 25572 ± 25576. Cytol., 104, 115 ± 151. Friend SH, Horowitz JM, Gerber MR, Wang XF, Bogen- Pertile P, Baldi A, De Luca A, Bagella L, Virgilio L, Pisano mann E, Li FP and Weinberg RA. (1987). Proc. Natl. MM and Giordano A. (1995). Cell Growth Dier., 6, Acad. Sci. USA, 84, 9059 ± 9063. 1659 ± 1664. Gao M, Ossowski L and Ferrari AC. (1999). J. Cell. Physiol., Shih HH, Tevosian SG and Yee AS. (1998). Mol. Cell. Biol., 179, 336 ± 346. 18, 4732 ± 4743. Garriga J, Limon A, Mayol X, Rane SG, Albrecht JH, Stiegler P, Kasten M and Giordano A. (1998). J. Cell. Reddy EP, Andres V and Grana X. (1998). Biochem. J., Biochem. Suppl., 31, 30 ± 36. 333, 645 ± 654. Suter L, Meier G, Bechter R and Bobadilla M. (1998). Arch. Grana X, Garriga J and Mayol X. (1998). Oncogene, 17, Toxicol., 72, 791 ± 797. 3365 ± 3383. Tajima Y, Sawada K, Morimoto T and Nishimune Y. (1994). Hakovirta H, Yan W, Kaleva M, Zhang F, Vanttinen K, J. Reprod. Fertil., 102, 117 ± 122. Morris PL, Soder M, Parvinen M and Toppari J. (1999). Teerds KJ. (1996). The Leydig cell. Payne AH, Hardy MP Endocrinology, 140, 1492 ± 1498. and Russell LD (ed). Cache River Press: Vienna, IL, USA, Harbour JW and Dean DC. (2000). Nat. Cell. Biol., 2, E65 ± pp. 203 ± 219. E67. Tena-Sempere M, Zhang FP and Huhtaniemi I. (1994). Hardy MP, Zirkin BR and Ewing LL. (1989). Endocrinology, Endocrinology, 135, 1018 ± 1024. 124, 762 ± 770. Tevosian SG, Shih HH, Mendelson KG, Sheppard KA, Henriksen K, Hakovirta H and Parvinen M. (1995). Paulson KE and Yee AS. (1997). Genes Dev., 11, 383 ± 396. Endocrinology, 136, 3285 ± 3291. Vincent S, Segretain D, Nishikawa S, Nishikawa SI, Sage J, HuppiK,SiwarskiD,MockBA,DosikJandHamelPA. Cuzin F and Rassoulzadegan M. (1998). Development, (1996). Mamm. Genome, 7, 353 ± 355. 125, 4585 ± 4593. Jackson CM and Jackson H. (1984). J. Reprod. Fertil., 71, Yan W, Linderborg J, Suominen J and Toppari J. (1999). 393 ± 401. Endocrinology, 140, 1499 ± 1504. Kaelin Jr WG. (1999). Bioessays, 21, 950 ± 958. Yan W, Samson M, Jegou B and Toppari J. (2000a). Mol. Kasten MM and Giordano A. (1998). Cell Death Dier., 5, Endocrinol., 14, 682 ± 699. 132 ± 140. YanW,SuominenJ,SamsonM,JegouBandToppariJ. Kissel H, Timokhina I, Hardy MP, Rothschild G, Tajima Y, (2000b). Mol. Cell. Endocrinol., 165, 115 ± 129. Soares V, Angeles M, Whitlow SR, Manova K and Besmer Yan W, Suominen J and Toppari J. (2000c). J. Cell Sci., 113, P. (2000). EMBO J., 19, 1312 ± 1326. 161 ± 168. Krishnamurthy H, Weinbauer GF, Aslam H, Yeung CH and Yan W, Kero J, Suominen J, Huhtaniemi I and Toppari J. Nieschlag E. (1998). J. Androl., 19, 710 ± 717. (2000d). Dev. Biol., 227, 169 ± 182. LeCouterJE,KablarB,HardyWR,YingC,MegeneyLA, Yan W, West A, Toppari J and Lahdetie J. (1997). Mol. Cell. May LL and Rudnicki MA. (1998a). Mol. Cell. Biol., 18, Endocrinol., 132, 137 ± 148. 7455 ± 7465. Yee AS, Shih HH and Tevosian SG. (1998). Front. Biosci., 3, LeCouterJE,KablarB,WhytePF,YingCandRudnicki D532 ± D547. MA. (1998b). Development, 125, 4669 ± 4679. Yeh S, Miyamoto H, Nishimura K, Kang H, Ludlow J, LeCouter JE, Whyte PF and Rudnicki MA. (1996). Hsiao P, Wang C, Su C and Chang C. (1998). Biochem. Oncogene, 12, 1433 ± 1440. Biophys. Res. Commun., 248, 361 ± 367. Lu J and Danielsen M. (1998). J. Biol. Chem., 273, 31528 ± Yoshinaga K, Nishikawa S, Ogawa M, Hayashi S, Kunisada 31533. T and Fujimoto T. (1991). Development, 113, 689 ± 699. Lundberg AS and Weinberg RA. (1999). Eur. J. Cancer, 35, 1886 ± 1894.

Oncogene