SPOC1 (PHF13) is required for spermatogonial stem cell differentiation and sustained spermatogenesis

Stem cells are fundamental for the regenerative capacity of organ systems (Dick, 2008). They are characterized by having the potential to self-renewal and to differentiate. The mechanisms regulating the balance between these processes remain poorly understood. An impressive example for the capacity of self-renewal and differentiation of stem cells comes from spermatogenesis, where millions of spermatozoa are produced each day in the postpubertal male. Thus, spermatogenesis is one of the most efficient cell-producing systems in the adult mammalian organism (Sharpe et al., 2003), making it an excellent model system to elucidate the general mechanisms of stem cell renewal and differentiation.

It is suggested that, in non-primate mammals, spermatogonial stem cells (SSCs) are single cells (A_single, A_s) attached to the basement membrane of the seminiferous tubules. Depending on whether the cells undergo self-renewal or differentiation, the A_s spermatogonia either divide into two separate (A_s) cells or into a pair of spermatogonia (A_paired, A_pr), respectively. The A_pr enters additional rounds of cell divisions after which the spermatogonia form chains and are called A_aligned (A_al) spermatogonia. A_pr and A_al spermatogonia are still considered undifferentiated (Nakagawa et al., 2007). The A_al cells then further differentiate into the A1–A4 spermatogonia, intermediate and B spermatogonia, which undergo a final round of replication and enter the first meiotic prophase, resulting in primary spermatocytes. After completion of first meiotic prophase, a reductional division (meiosis I) and a subsequent mitosis-like division (meiosis II) lead to the formation of haploid spermatids that develop into spermatozoa (Aponte et al., 2005; de Rooij, 2001; He et al., 2009; Olive and Cuzin, 2005; Russell et al., 1990). During differentiation, the germ cells migrate from the basal lamina into the adluminal compartment of the seminiferous epithelium by traversing the blood–testis barrier (BTB) formed by Sertoli cells. This process requires extensive interaction between Sertoli cells, as well as between Sertoli and germ cells. This tight coordination between germ cell movement and differentiation is required in order to avoid arrest of spermatogenesis and apoptosis during this passage (reviewed by Mruk and Cheng, 2004). The basal compartment provides a niche assuring a specialized microenvironment capable of generating the necessary balance between self-renewal and differentiation. This process is regulated by extrinsic niche stimuli secreted by the Sertoli cells, as well as by intrinsic gene expression in the SSCs. Although several extrinsic stimuli have been described (Chen et al., 2005; Hofmann, 2008; Oatley and Brinster, 2008) little is known about the intrinsic factors that regulate SSC self-renewal, proliferation and differentiation. In undifferentiated SSCs (A_s–A_al), some core transcription factors are known (e.g. OCT4 and SOX2) that are of general importance for the maintenance of the pluripotency of stem cells (Dann et al.,2008; Oatley and Brinster, 2008). Furthermore, three factors, promyelocytic leukaemia zinc-finger (PLZF, also known as ZBTB16 and ZFP145) (Buaas et al., 2004; Costoya et al., 2004; Falender et al., 2005; Hobbs et al., 2010), TATA-box-binding-protein-associated factor 4B (TAF4B) (Falender et al., 2005) and NANOS2 an RNA-binding protein (Sada et al., 2009; Saga, 2010) have also been shown to be important for SSC maintenance. Whereas TAF4B is involved in SSC formation and proliferation, PLZF and NANOS2 are directly involved in SSC self-renewal. PLZF functions both as a transcriptional repressor and as an activator of genes implicated in self-renewal of SSCs and in differentiation of undifferentiated A-type (A_s–A_al) spermatogonia (Buaas et al., 2004; Dadoune, 2007; Hobbs et al., 2010). Loss-of-function studies for PLZF and NANOS in the mouse result in animals with a progressive loss of germ cells that culminate in a Sertoli-cell-only phenotype in a variable percentage of tubules (Buaas et al., 2004; Costoya et al., 2004; Sada et al., 2009).

Recently, we identified the gene SPOC1 (for survival-time-associated PHD protein in ovarian cancer), also designated PHF13, and demonstrated that enhanced SPOC1 RNA levels in primary and recurrent epithelial ovarian cancers are correlated with a shorter survival time in a cohort of patients (Mohrmann et al., 2005). The SPOC1 protein localizes to the nucleus and contains a plant homeodomain (PHD), which is believed to regulate chromatin-specific interactions (Bienz, 2006). In agreement with this prediction, we have recently demonstrated that SPOC1 is dynamically associated with chromatin and that it plays a role in proper chromosome condensation and cell division (Kinkley et al., 2009). SPOC1 RNA is detectable in most tissues but the highest levels are present in the testis where it exclusively localizes to spermatogonia (Mohrmann et al., 2005). These findings suggest a testis-specific functional role of SPOC1 in spermatogonial stem cells. Beyond this, there is very little additional data available on SPOC1 expression and function. In particular, the role of SPOC1 during development and in carcinogenesis remains unclear. Therefore, in order to gain insight into the in vivo functions of the SPOC1 protein we created and analyzed mice whose Spoc1 gene has been disrupted by a gene-trap insertion.

Here, we report that mice deficient for SPOC1 display a progressive loss of germ cells, which is accompanied by apoptosis of pachytene spermatocytes. We provide evidence that this is not caused by defects in meiosis, but rather by a disturbed differentiation of SSCs, where SPOC1 is normally coexpressed together with PLZF. Thus, SPOC1 represents a new intrinsic factor indispensable for sustained spermatogenesis and the differentiation of the SSC pool.

Embryonic stem cells (ESCs) carrying a mutant Spoc1 (Phf13) locus were obtained from a library of ESC clones generated by random insertional mutagenesis. The Spoc1 locus in the selected ESC clone was disrupted by a gene-trap vector containing intronic sequences and a splice acceptor site of the engrailed 2 (En2) gene, as well as β-geo, a fusion of β-galactosidase and neomycin phosphotransferase II (BayGenomics, CA). The resulting insertional mutation leads to a fusion transcript containing sequences from gene-specific exons upstream of the insertion joined to the β-geo marker and terminated by a poly(A) site. As a result, expression of the trapped gene can be detected by lacZ staining and its sequence can easily be determined by 5′ rapid amplification of cDNA ends (5′ RACE). In the ESC clones Xb691 and Xa022, the 5′ RACE product contained sequences from exon 1 [nucleotides 303 to 354 of the Spoc1 (Phf13) cDNA (GenBank accession number NM_172705)] of the Spoc1 gene, indicating an insertion of the gene-trap vector into intron 1 (Fig. 1A). We further mapped the integration sites to between nucleotides +917 and +918 (Xb691) and +876 and +877 (Xa022) of intron 1 relative to the first base of intron 1. On the RNA level this insertion leads to the termination of Spoc1 gene transcripts after exon 1, which contains the translation start site.

Mutant mice were generated from both ESC clones by injection into blastocysts from C57BL/6 mice. Subsequent intercrosses of heterozygous animals generated a mixed genetic background. The phenotypes were identical for both mice strains and therefore the data are presented only for the mouse strain generated from ESC clone Xb691. For PCR genotyping (Fig. 1B) of mutant Spoc1, primers were generated to both the β-gal cassette and to intron 1, as shown in Fig. 1A. To determine the effect of the gene-trap insertion on splicing of Spoc1, RT-PCR was performed on RNA isolated from tail cuts of heterozygous and homozygous mice using primers corresponding to the β-galactosidase sequence of the gene-trap vector, as well as from exon 1 of Spoc1. Sequencing of the resulting RT-PCR products revealed splicing of exon 1 to the splice acceptor site of the gene-trap cassette in homozygous (−/−) and heterozygous (+/−) animals (data not shown). Heterozygous animals expressed β-galactosidase from the gene-trap construct at a low level in most tissues, demonstrating splicing and functionality of the gene trap (Fig. 1C).

In order to prove the efficient knockdown of endogenous SPOC1 in vivo, several approaches were undertaken. First, northern blot analysis of testis tissue (mice at 18 weeks of age) was performed with a probe corresponding to the Spoc1 3′ untranslated region (3′-UTR). There was no detectable Spoc1 mRNA in homozygous animals (Fig. 1D), whereas heterozygous animals showed an ~50% reduction of Spoc1 expression in comparison to wild-type animals.

Next, we performed quantitative real-time PCR (qRT-PCR) with a probe corresponding to the junction of exon 2 and 3 of wild-type Spoc1 cDNA. Accordingly, RNA was extracted from the testes of homozygous mice and their wild-type littermates at various ages (Fig. 1E). We observed a low expression of persisting wild-type Spoc1 transcript ranging from 2 to 14 % in the homozygous mutant mice compared with wild-type tissue in the different animals. The residual Spoc1 transcripts are probably due to skipping of the gene-trap construct in some transcripts. Nevertheless, western blot analysis of testis tissue from homozygous mice demonstrated that, at the protein level, no detectable SPOC1 was observed at different ages (5, 10 and 20 weeks). By contrast, a strong signal for SPOC1 was observed in wild-type animals, and an intermediate signal was detected from extracts of heterozygous animals (Fig. 1F). Immunoblotting against SPOC1 was performed using a monoclonal anti-SPOC1 antibody (Kinkley et al., 2009). These data indicate that mice with a homozygous gene-trap insertion represent a complete functional Spoc1-knockdown at the protein level and we therefore refer to these animals as Spoc1^−/− mice. This conclusion is further strengthened by the immunohistochemical studies described below (Fig. 2).

SPOC1 expression and localization was investigated by X-gal staining of Spoc1^+/− testis, which expresses β-galactosidase under the control of the Spoc1 promoter (Fig. 2A). X-gal staining showed a granular signal in a few peritubular basal cells in most tubules, consistent with the position of spermatogonia (Buaas et al., 2004; Costoya et al., 2004; Dadoune, 2007). A similar spermatogonia-specific pattern was observed when testes sections were stained with the monoclonal rat anti-SPOC1 antibody (Fig. 2B). Spoc^−/− gonocytes were PLZF positive (staining not shown) but were negative for SPOC1 (Fig. 2C), demonstrating the specificity of the antibody, as well as the knockdown of SPOC1, in the mouse model. The higher number of β-galactosidase-positive cells in comparison with cells stained with anti-SPOC1 antibody might be due to a greater sensitivity of the enzyme-linked β-galactosidase staining, to post-transcriptional degradation of Spoc1 in some Spoc1-RNA-containing cells or to diffusion of X-gal, and remains to be investigated.

To demonstrate further that SPOC1 is expressed in early spermatogonia, we evaluated immunohistochemically the coexpression of SPOC1 with PLZF. PLZF is a marker of undifferentiated A-type spermatogonia (A_s–A_al) and is a known intrinsic factor involved in SSC self-renewal (Buaas et al., 2004; Costoya et al., 2004; Payne and Braun, 2006). We observed the expression of both proteins in the nuclei of gonocytes, the progenitors of SSCs, in testis 3 days post partum (dpp) (Fig. 2C), as well as in spermatogonia of 10-week-old mice (Fig. 2D,E), demonstrating that SPOC1 is expressed in undifferentiated A-type spermatogonia. Quantitative analysis over a developmental window [postnatal day (P) P1 to 10 weeks] demonstrated that there was coexpression of PLZF in 77–100% of SPOC1-positive cells, whereas SPOC1 was found to be coexpressed in 97–100% of PLZF-positive spermatogonia (supplementary material Table S1 and Fig. S1). Furthermore, whole-mount immunofluorescence confirmed SPOC1 localization in seminiferous tubules of adult mice (12 weeks of age) and revealed that SPOC1 was expressed in single (A_s), paired (A_pr) and aligned (A_al) spermatogonia (supplementary material Fig. S2). Taken together these results demonstrate that SPOC1 is expressed, in a manner similar to PLZF, in undifferentiated spermatogonial stem cells.

Spoc1^+/− mice were indistinguishable from wild-type littermates on the basis of appearance and breeding capacity. Intercrosses of the heterozygous animals produced ~14% Spoc1^−/− animals, which deviates from the expected Mendelian ratio of 25% (supplementary material Fig. S3). Preliminary observations indicate intrauterine death [at embryonic day (E) E13.5–E15.5] as well as peri- and early postnatal lethality, explaining the divergent Mendelian ratio.

Gross phenotypic examination of Spoc1^−/− mice revealed that there was initially no visible abnormalities. However, homozygous males rapidly (at 5 weeks and older) developed a reproductive defect and required a substantially prolonged period of time to impregnate the females, whereas heterozygous animals bred normally (supplementary material Fig. S4). A more detailed examination of Spoc1^−/− mice at different ages revealed a substantially reduced testis size (Fig. 3A). Already at 5 weeks of age the relative testis weight (testis weight to body weight ratio, GSI) was dramatically reduced (56%) compared with wild-type animals (Fig. 3B). Correspondingly, we found a substantial reduction in the cross-sectional area of seminiferous tubules in Spoc1^−/− mice (Fig. 3C). Sperm number was decreased to 28.6% (5 weeks), 9.3% (10 weeks) and 10.2% (20 weeks) of the wild-type number, respectively. Histological investigation of testis from aged mice (59 weeks) revealed a complete germ cell loss in Spoc1^−/− animals, whereas heterozygous (Spoc1^+/−) and wild-type mice appeared normal (Fig. 3D). Together, these data indicate a vast breakdown of spermatogenesis, which explains the early infertility of the Spoc1^−/− mice.

The observed phenotype is solely caused by the absence of Spoc1 and not by additional non-specific gene-trap integrations. This can be concluded from the following observations: (i) mice generated from two independent ESC clones (Xb691 and Xa022) show similar phenotypes; (ii) the β-galactosidase expression pattern in Spoc1^+/− mice matches the expression of Spoc1, as determined by RNA in situ hybridzation and imunohistochemical detection; and (iii) in the mixed genetic background we observed a 100% penetrance between the Spoc1^−/− genotype and the observed phenotypic changes.

In contrast to Spoc1^+/− testis, which appeared phenotypically normal and indistinguishable from those in wild-type mice even after >12 months (Fig. 3D), histological examination of Spoc1^−/− testes revealed severe testis atrophy with profoundly altered spermatogenesis in a large number of tubules. Detailed histological investigation of Spoc1^−/− testes from mice at 5, 10 and 20 weeks of age revealed an increase in the number of seminiferous tubules that showed a reduced number of germ cells with progressing age. In 5-week-old animals spermatogenesis appeared normal despite a reduced cross-sectional area of seminiferous tubules (Fig. 3C, Fig. 4A), whereas tubules from 10-week-old animals were disorganized (Fig. 4A). In 20-week-old animals the number of tubules with abnormal spermatogenic epithelium had further increased and the testis became completely atrophic. Unfortunately, inter-individual variation prevented an exact quantification of this phenotype. Nevertheless, tubules with a normal germinal epithelium were rarely present at this age. Remarkably, several tubules (per testis) predominantly contained Sertoli cells and a few spermatozoa but no developmentally younger germ cells, whereas other tubules contained Sertoli cells together with chains of A_al spermatogonia (Fig. 7B) and still other tubules (25% of tubules at 20 weeks of age) showed a Sertoli-cell-only phenotype (Fig. 4A,B and Fig. 5B). A gross quantification of these changes is depicted in supplementary material Fig. S5. The Spoc1^−/− Sertoli cells were similar in number and size to those of wild type (Fig. 4) and expressed the Sertoli-specific protein GATA1 (Fig. 4C). When we determined the number of GATA1-positive cells per seminiferous tubule in sections from 10-week-old animals, there were no significant differences between wild-type and Spoc1^−/− testes (data not shown).

In summary, our histological data demonstrate a normal onset of spermatogenesis in juvenile animals followed by a successive loss of spermatogenetic cells from the germ cell epithelium after 5-weeks post partum. The loss of germ cells in early developmental stages, but the presence of later developmental stages of germ cells, as was observed in older animals, indicates an ongoing process of depletion of germ cells possible owing to a failure in stem cell renewal. Accumulation of A_al spermatogonia in the absence of more progressed germ cell stages might arise through a failure in their differentiation and/or their entry into meiosis.

To exclude that the phenotypes observed are simply caused by a reduced number of gonocytes before the formation of SSCs, histological evaluation of Spoc1^−/− and wild-type testes (3 dpp) was performed and these testes were stained for PLZF and SPOC1, respectively. These experiments demonstrated that Spoc1^−/− and wild-type animals displayed similar numbers of gonocytes (Fig. 2C), strongly arguing that the observed loss of germ cells in the postpubertal testis is a result of impaired SSC self-renewal and/or differentiation.

The observed reduction in testis size and seminiferous tubule cross-sectional area indicate either a reduced proliferation of germ cells or increased apoptosis. To address these possibilities we performed TUNEL assays and immunohistochemical stainings of apoptosis and proliferation markers. The TUNEL assays revealed a substantially increased overall rate of apoptosis in tubules from 5-week- and 20-week-old Spoc1^−/− animals (Fig. 5A–C), which was confirmed by staining of activated caspase 3 (supplementary material Fig. S6). Apoptosis was strongly increased in prophase I cells, especially in the pachytene stage at seminiferous epithelial stages IV to VII (Fig. 5D–G). Stage X tubuli served as internal positive controls for the assay because these tubuli always contain TUNEL-positive and TUNEL-negative cells (Voet et al., 2003). Next, testis sections were stained with antibodies against PCNA and Ki67, two widely used proliferation markers that are expressed in a subset of A-, In-, and B-spermatogonia, as well as in prophase spermatocytes (Costoya et al., 2004; Wrobel et al., 1996). PCNA staining for mice at 5, 10 and 20 weeks of age revealed a slightly reduced appearance in the mutant being significant only at 20 weeks (P=0.034) (supplementary material Fig. S7). By contrast, Ki67 staining revealed no significant difference (P>0.16) between Spoc1^−/− and Spoc1^+/+ mice at 5, 10 and 20 weeks of age (supplementary material Fig. S8). At 1 week of age no significant difference in the number of PCNA- or Ki67-positive cells was apparent (data not shown). These results strongly indicate that the demise of spermatogenesis is driven by enhanced apoptosis in the knockout.

Arrest of spermatogenesis can be elicited by the absence of numerous factors, some of which are required for passage through prophase I (Cooke and Saunders, 2002; Scherthan, 2003). To determine whether Spoc1 deficiency causes defects in spermatocyte development, we studied the course of prophase I by following meiotic differentiation through markers for double-strand break (DSB) formation and repair (phosphorylated H2AX, known as γH2AX) (Mahadevaiah et al., 2001; Meyer-Ficca et al., 2005) and the homologous recombination repair protein MRE11 that is expressed at high levels in meiotic cells (Eijpe et al., 2000). Determination of testes tubule stages was performed as described previously (Russell et al., 1990).

Normal patterns of γH2AX formation and localization were detected in prophase I cells of testis sections from 10-week-old Spoc1^−/− mice, with strong γH2AX signals being present in leptotene and zygotene nuclei and at the XY body of pachytene-diplotene spermatocytes (Fig. 6A,B). A physiological chromatin signal (Meyer-Ficca et al., 2005) was also seen in spermatids of Spoc1^−/− and wild-type animals (Fig. 6A,B). Strong expression of the DNA DSB repair protein Mre11 was present in early meiotic cells and at the XY body of pachytene spermatocytes of knockout (Fig. 6C) and wild-type (not shown) testes, which is consistent with previous studies on wild-type mice (Eijpe et al., 2000).

Next, we investigated meiotic and mitotic cell division by immunostaining for phosphorylated histone H3 serine 10, which labels condensing mitotic and meiotic chromosomes (Cobb et al., 1999; Hendzel et al., 1997). Meiotic metaphase I and II cells (found in the lumen of stage XII tubules; Fig. 6C) and mitotic metaphases (observed at the periphery of testis tubules in most stages (data not shown) were similar in wild-type and Spoc1^−/− testis, indicating that cell division is not affected. Furthermore, the normal patterns of γH2AX and Mre11 localization from leptotene to pachytene stages and the faithful XY-body formation suggest a normal course of DNA repair and chromosome pairing during prophase I in Spoc1-deficient meiocytes. In agreement, meiotic chromosome pairing and sister chromatid (SC) formation, was found to be normal by γH2AX and SYCP3 (Liebe et al., 2006) staining of spreads of wild-type and Spoc1^−/− spermatocytes (Fig. 6D). Together, these findings suggest that DNA repair and chromosome pairing progress normally in Spoc1^−/− mice.

Because the histological data indicate a defect in SSC self-renewal or differentiation, we specifically investigated the distribution of undifferentiated spermatogonia over time in the Spoc^−/− testes. PLZF is a marker for undifferentiated spermatogonia (A_s, A_pr and A_al) and is co-expressed with the stem cell markers OCT4 (also known as POU5F1) and GDNF family receptor α1 (GFRα1) (Buaas et al., 2004; Costoya et al., 2004; Payne and Braun, 2006). Using a mouse anti-PLZF antibody, we examined testis sections from wild-type and Spoc1^−/− mice at P0 (neonates) and 20 weeks of age. Tubules without PLZF-positive cells [measured according to Buaas et al. (Buaas et al., 2004)] and PLZF-positive cells per tubulus were counted, respectively. In the adult testis, only strongly PLZF-expressing cells at the periphery of the tubules (thus representing A_s–A_al spermatogonia) were scored (Fig. 7). The data obtained clearly demonstrated that the number of tubules with PLZF-positive cells was unchanged between neonatal (P0) Spoc1^+/+ and Spoc1^−/− mice, whereas there was a highly significant reduction (P<0.0001) in adult mice (Fig. 7A,C). Unexpectedly, the average number of PLZF-positive cells per tubule remained unchanged in both neonatal and adult testis (Fig. 7D). This result can be attributed to tubules exclusively present in Spoc1^−/− testis that showed a strong increase in the number of PLZF-positive cells (>10) clustered in the periphery of a few mutant tubules (Fig. 7B). Quantification revealed 3.9% of tubules with >10 PLZF-positive cells in Spoc1^−/− and 0.1% in wild-type testis in mice at 5 weeks of age At 10 weeks of age, the values were even more divergent, with 6.7% in Spoc1^−/− compared with 0.1% in wild-type testis. This accumulation of cells represents A_al spermatogonia and can be found in tubules that in addition contain only Sertoli cells.

Because Spoc1^−/− mice with a mixed 129 and C57BL/6 genetic background were used throughout most of the experiments, we sought to verify this unexpected result in a congenic gene-trap line in order to exclude genetic background effects. We therefore generated a congenic mouse strain by successive backcrossing of Spoc1^+/− animals with C57BL/6 wild-type animals. Testes from knockout mice of generation N₁₂ were investigated. The testis showed a comparable, but slightly more severe phenotype than in the mixed genetic background (data not shown). Again, PLZF-positive cells per tubule were counted in testis sections from mice at 5, 10 and 20 weeks of age (Fig. 8). The result clearly demonstrated an almost complete loss of PLZF-positive spermatogonia at 20 weeks of age in Spoc1^−/− mice (Fig. 8C). Remarkably, this was preceded by an increase in the number of tubules with extended chains of spermatogonia and thus an elevated number of A_al spermatogonia, which was already visible at 5 weeks of age (Fig. 8A,B). This result confirms our previous findings and demonstrates that although further differentiated germ cells are almost completely lost, undifferentiated spermatogonia still proliferate and accumulate, generating extended chains of A_al spermatogonia. With progressive aging these cells also eventually disappear, leading to a Sertoli-cell-only phenotype. The results in the congenic strain strongly indicate a profound defect in differentiation (and not in proliferation or self-renewal) of the SSCs, leading to the observed transient accumulation of undifferentiated spermatogonia in mutant tubules of older mice.

Here, we have shown that SPOC1 is essential for the maintenance of spermatogenesis in the postpubertal testis and is a prerequisite for sustained spermatogenesis and normal fertility in adult mice. Spoc1^−/− mice displayed a wild-type number of gonocytes and a normal first wave of spermatogenesis in juvenile animals, but a progressive loss of germ cells was initiated several weeks after puberty. This loss first affected non-SSC stages of germ cells and then, at a later time point, the undifferentiated spermatogonia, which finally resulted in a large number of Sertoli-cell-only seminiferous tubules and a nearly complete atrophy of the testes in animals older than 20 weeks. The testes tubules of younger Spoc1^−/− animals displayed an altered distribution of undifferentiated spermatogonia in tubules, with a transient increase in A_al spermatogonia and a successive loss of all germ cells with increasing age. Furthermore, there were substantially more apoptotic germ cells overall, but specifically in the pachytene stage of Spoc1^−/− testis, whereas there was no significant difference in the number of PCNA- or Ki67-positive (proliferating) cells or the number of primary gonocytes. Immunostainings of molecular markers for initiation and progression of meiosis were also normal. These results indicate that the seminiferous epithelium gradually loses its ability to start and/or complete new waves of differentiation, whereas the mutant epithelium has a compromised capacity to support the transit of spermatocytes I through the pachytene stage.

Overall, our findings strongly indicate that SPOC1 represents a new intrinsic regulating factor in undifferentiated and PLZF-positive spermatogonia (A_s–A_al) that is responsible for maintaining proper and sustained differentiation of cells from the self-renewing stem cell pool. This conclusion is based on several observations. First, the initial colonization of the seminiferous tubules by germ cells appears to be normal, given that the number of gonocytes after birth does not differ between the seminiferous tubules of Spoc1^−/− and wild-type animals. Thus, factors that regulate primordial germ cell migration and early stages of proliferation appear intact. Second, morphological and histological analysis, as well as immunohistochemical staining indicates that the observed testis atrophy is not caused by a failure in supporting cells or in hormone status because Sertoli cell number and seminal vesicles were normal. Third, the biologically active SSCs first appear in male mice of ~3–4 dpp (McLean et al., 2003). It has been suggested that two different populations of gonocytes are present in the postnatal period. In the first postnatal week gonocytes directly develop into Kit-expressing and, therefore, differentiating spermatogonia. These cells are the starting point of a first round of spermatogenesis. They do not undergo self-renewal and are therefore rapidly depleted (de Rooij, 1998; de Rooij and Russell, 2000; Yoshida et al., 2007). By contrast, a second subpopulation of gonocytes develops into a pool of spermatogonia, which undergo self-renewal and thus provides a stem cell population for all subsequent rounds of spermatogenesis (de Rooij, 1998; de Rooij and Russell, 2000; Yoshida et al., 2007). The fact that the testes of 5-week-old Spoc1^−/− mice show seminiferous tubules with almost unaffected spermatogenesis (although reduced in diameter) indicates that the first round of spermatogenesis is unaffected and the ensuing meiosis and spermiogenesis can proceed normally. Consequently, our data suggest a specific defect in the differentiation of spermatogonia descending from the second subpopulation of gonocytes, which is characterized by a self-renewing capability. Forth, we observed a consecutive loss of spermatogenic substages, with a loss of spermatocytes I preceding the loss of spermatids, and finally spermatozoa, eventually leading to Sertoli-cell-only tubules in animals >20 weeks of age. This sequential abrogation of spermatogenic stages, together with the transient increase in A_al spermatogonia and the subsequent loss of all undifferentiated spermatogonia, strongly suggests a failure of SSCs to transit to differentiating spermatogonia. Hyperproliferation of undifferentiated spermatogonia leading to a transient alignment of undifferentiated spermatogonia is not probable because PCNA- and Ki67-positive cells were not increased in Spoc1^−/− testis. It is tempting to speculate that the observed transient clustering of PLZF-positive spermatogonia along the periphery in some mutant tubules indicates an abrogated differentiation and migration of spermagonia, given that both processes are tightly coupled (reviewed by Lie et al., 2009; Mruk and Cheng, 2004). Finally, we noted an increased apoptotic rate in germ cells in the pachytene stage during stages IV to VII of the seminiferous epithelial cycle and in spermatogonia. However, most of the spermatocytes managed to complete prophase I and spermatogenesis. This agrees with the normal expression and localization of Mre11 and γH2AX (prophase I recombination markers), phosphorylated histone H3 (a meiotic and somatic metaphase marker) and the presence of a normal XY-body and normal chromosome pairing during prophase I in the Spoc1^−/− and wild-type mice. These data suggest that a pachytene (recombination) checkpoint at stage IV of the seminiferous cycle (Ashley et al., 2004; Barchi et al., 2005) is not responsible for the increased cell death of Spoc1^−/− spermatocytes. Because, in mitotic cells, SPOC1 is important for prophase condensation and chromosome alignment in metaphase (Kinkley et al., 2009), and given that prophase I of meiotic cells is equivalent to the G2 phase of mitotic cells and involves chromosome condensation into pachytene, it seems probable that SPOC1 could also play a role in meiotic prophase I chromatin function during pachytene chromosome condensation. Hence, the absence of SPOC1 in the knockout might underlie the elevated levels of pachytene apoptosis. The surviving spermatocytes do not seem to have defects in haploid (spermatid) development because spermatids and spermatozoa are the last form to occur before Sertoli-cell-only tubules form. The overall increase of apoptotic cells in Spoc1^−/− testis by 5 weeks of age probably explains the reduced testis size and tubular volume. Taken together with the differentiation defect of SSCs, this might explain the fast demise of spermatogenesis in the knockout.

An alternative explanation for the observed phenotype would be an altered differentiation and migration process in mutant spermatogonia and carry over of alterations that cause cell death in some pachytene spermatocytes. It is known that even the slightest deregulation affecting either the ability of the germ cell to traverse through the BTB or its migration through the adluminal compartment can cause arrest of spermatogenesis leading to germ cell degeneration and apoptosis (reviewed by Lie et al., 2009; Mruk and Cheng, 2004). Such a scenario in the Spoc1^−/− testes could feasibly result in a deregulation of the differentiation processes responsible for proper spermatogenesis and thus trigger a subsequent apoptosis during or right after traverse through the BTB or migration through the adluminal compartment. Of note, this process would also not affect the first wave of spermatogenesis because the BTB forms later during puberty (reviewed by Cheng et al., 2009). This explanation would be consistent with the expression of the SPOC1 protein in undifferentiated, PLZF-positive spermatogonia and with the altered (clustered) distribution of these cells even in tubules with an otherwise Sertoli-cell-only phenotype.

PLZF is a known intrinsic transcription factor important for the self-renewal of SSCs. Interestingly, PLZF and SPOC1 are coexpressed in undifferentiated SSCs and the phenotype of the Spoc1^−/− mice is very similar to the testis phenotype of Plzf^−/− (luxoid) mice (Buaas et al., 2004; Costoya et al., 2004). PLZF belongs to the BTB–POZ-ZF (for broad complex, tramtrack, bric-à-brac or poxvirus and zinc finger) family of transcription factors. It has been suggested that this group of proteins represents transcriptional repressors that act by epigenetically modifying chromatin (Kelly and Daniel, 2006) and that PLZF probably influences the epigenetic program of spermatogonial cells (Buaas et al., 2004). Similarly, SPOC1 is a PHD-finger-containing protein that is predominantly chromatin-associated and is involved in regulating chromatin structure and in mitotic chromosome condensation (Kinkley et al., 2009). Therefore, owing to the similarities between SPOC1 and PLZF, and because PHD-containing proteins are repeatedly being identified as new epigenetic readers and modifiers of chromatin, it is tempting to speculate that SPOC1 also affects stem cell differentiation in a manner similar to PLZF by altering the epigenetic program. Interestingly, another PHD finger protein (PYGO2) has recently been shown to modulate chromatin and to be involved in spermiogenesis (Nair et al., 2008). Of note, the reduced number of Spoc1^−/− offspring (compared with the expected Mendelian ratio) indicates a profound effect on viability with reduced penetrance, possibly due to the variation of epigenetic factors.

In summary, we have identified a new intrinsic factor in undifferentiated PLZF-positive spermatogonia that is required for the differentiation of self-renewing spermatogonial stem cells into spermatocytes, possibly through a chromatin-mediated effect on differentiation-inducing genes. A further study will thus be conducted to identify the effect of SPOC1 disruption at the expression level.

All animal experiments were performed according to approved guidelines. To develop a Spoc1-deficient mouse strain we used ESC lines, Xb691 and Xa022, which have a gene-trap insertion in the Spoc1 gene between exon 1 and 2 (Bay Genomics, San Francisco, CA; http://www.genetrap.org). According to JAX convention the gene-trap allele was designated as Phf13^Gt(XB691)Byg and Phf13^Gt(XA022)Byg but is here abbreviated to Spoc1⁻. The Spoc1 gene-trapped ESCs were microinjected into C57BL/6 host blastocysts to generate germline chimeras. The resulting Spoc1⁻ line was maintained as heterozygotes and intercrossed to obtain homozygotes (Spoc1^−/−). The genotypes were determined by PCR. The wild-type Spoc1 allele PCR was detected with primers in intron 1, WT-F (5′-CCTGGTGGAGCAAGTTGGAAC-3′) and WT-R (5′-CTCACCTTGTCGCTACCCCAAC-3′), which produced a 151-bp amplicon. To detect the gene-trap allele, the primers ko-F (5′-GGGTGCGTTGGTTGTGGATAAG-3′) and ko-R (5′-CATTCAGGCTGCGCAACTGTTGGG-3′) were used which produced a 332-bp amplicon.

Quantification of relative transcript levels of Spoc1 was performed with the QuantiTect Probe RT-PCR kit (Qiagen) in an ABI PRISM 7900 HT sequence detection system (Applied Biosystems). The assays were performed in triplicate in 384-well optical plates (Applied Biosystems) with a final volume of 20 μl each. For Spoc1 detection, an Assays-on-Demand system (Applied Biosystems) (1 μl 20 × probe assay, 10 μl 2 × qPCR mastermix and 100 ng cDNA) was used. To normalize for variations in RNA amount and reverse transcription efficacy, transcript levels encoding the housekeeping RNA-polymerase II (RPII) were determined. The following primers and probes were used: RpII Fw (5′-GAATCCGCATCATGAACAGTGAT-3′), Rev (5′-CATCCATTTTATCCACCACCTCTT-3′), probe (5′-CTCCTCTTGCATCTTGT-3′; 5′-FAM + 3′-TAMRA), Spoc1 probe (5′-AAAGGAGGAGCTCCCCCTGAGGAGC-3′). Analysis was performed with ABI PRISM 7900 HT SDS V2.1.1-Software.

For northern blot analysis total RNA from testis was isolated using TRIzol Reagent (Invitrogen), separated by formaldehyde agarose gel electrophoresis and blotted onto a nylon membrane (Amersham) as previously described (Tagariello et al., 2008). As a hybridization probe we used a mixture of two different PCR-products, amplified with the primers Spoc-mus-Fw1 (5′-GTGAGCAGATCAAACCGTGAGAG-3′) and Spoc-mus-Rev1 (5′-GTGGAAAGGCGGCTACACATTGTC-3′), as well as mSPOC-Fw-5 (5′-GCCACTGTTACACTGGGTGTC-3′) and mSPOC-Rv-5 (5′-GAAAGGTGACACACTGGCCAC-3′), respectively. The probes are located in the Spoc1 3′ UTR (exon 4).

Protein extracts were prepared from the ovary and testis of 18-week-old mice by homogenizing in RIPA buffer (~300 μl per 5 mg tissue; 50 mM Tris-HCl pH 8, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate and 0.1% SDS; freshly prepared with Complete Mini Protease Inhibitor from Roche). The proteins were immunoblotted and probed with a rat monoclonal anti-SPOC1 antibody (Kinkley et al., 2009) at a concentration of 1:1000 and a rabbit polyclonal anti-actin antibody (sc-1616, Santa Cruz Biotechnology) at a dilution of 1:3000. The membranes were further probed with secondary horseradish peroxidase (HRP)-conjugated rabbit anti-rat-IgG antibody (1:15,000) (P0450, Dako) or swine anti-rabbit-IgG antibody (1:15,000) (P0399, Dako) antibodies, respectively. The protein bands were visualized using ECL reagents (Pierce) as described by the manufacturer.

After perfusion, the testis were fixed in Ito-solution for at least 12 hours, washed in cacodylic buffer for 3 hours and incubated for 2 hours in 1% osmium tetroxide in cacodylic buffer. After washing in cacodylic buffer, the testis were dehydrated in ascending ethanol concentrations (70, 80, 90 and 100%) and embedded through acetone: 30 minutes in ethanol and acetone (1:1), 30 minutes in acetone, 2 hours in acetone and epon (1:1) and then epon for 2 hours. The epon was hardened for 24 hours at 60°C and then 48 hours at 90°C. The testis was sagittally cut using an Ultracut microtome (Reichert-Jung, Germany) assembled with a diamond knife (Drukker, Germany). The sections were stained with Toluidine Blue for 1 minute at 50°C, washed for 10 minutes in Xylene and mounted with Entellan (Merck, Germany). Sections were analyzed with an Olympus IX 70 microscope.

Testes were fixed in 10% formalin overnight, washed with tap water for 24 hours, embedded in paraffin and sectioned at 5-μm thickness. Immunohistochemical staining was performed on deparaffinized and rehydrated sections subjected to antigen unmasking using 10 mM citrate buffer 1 hour at 95°C. After inactivation of endogenous peroxidase (30 minutes in 0.3% H₂O₂, 10% methanol in PBS at room temperature) and three 5-minute washing steps in PBS, a permeabilization step (20 minutes in 0.5% Triton X-100 in PBS) was performed. After an additional three 10-minute washing steps in PBS, the further steps are executed with Vectastain Elite ABC Kit (Vector Laboratories).

Sections were incubated with rat monoclonal anti-SPOC1 antibody (1:10) (Kinkley et al., 2009) rabbit polyclonal anti-Mre11 antibody (1:200, Novus Biologicals, Littleton, CO), rabbit polyclonal anti-H2AX antibody (1:500, Biomol, Hamburg), rat monoclonal anti-GATA1 antibody (1:200, Santa Cruz Biotechnology, California, USA), anti-phosphorylated histone H3 (serine 10) antibody, mouse polyclonal anti-PLZF antibody (1:250, Santa Cruz Biotechnology), mouse monoclonal anti-PCNA antibody (PC10 Calbiochem) or rabbit polyclonal anti-Ki67 antibody (Bethyl Laboratoris) overnight at 4°C. Fluorescence detection of primary antibodies was performed as described (Barrionuevo et al., 2009). Nuclei were counterstained with hemalaun or DAPI. Sections were analyzed using Olympus IX 70 (peroxidase staining) or a Zeiss Axioplan 2 microscope, equipped with fluorescence filters for red, green and blue excitation (Chroma) (fluorescent staining).

TUNEL assays were performed using the In Situ Cell Death Detection Kit and Fluorescein (Roche) on paraffin-embedded tissue following the instructions of the manufacturer.

Preparation of spermatocyte spreads was performed as described previously (Scherthan et al., 2011).

Tissue was fixed in 4% paraformaldehyde for 2 hours at 4°C. After three washing steps in PBS (4°C), tissue was incubated in 30% sucrose in PBS at 4°C with slight agitation. For cryosections, the tissue was embedded in Tissue-Tek (Sakura, Staufen, Germany). Sections were washed in detergent solution (1 × PBS, 2 mM MgCl₂, 0.02 % NP-40 and 0.01% Na-deoxycholate in water) for 15 minutes before incubation with staining solution [1 × PBS, 20 mM Tris HCl, pH 7.5, 2 mM MgCl₂, 0.02% NP-40, 0.01% Na-deoxycholate, 5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆, 0.75 μg/ml X-Gal in DMSO] under optical control. To stop the staining reaction, sections were washed three times in post-washing solution (1 × PBS, 5 mM EDTA, pH 8 in water). Sections were mounted with Crystal Mount (Sigma, Taufkirchen, Germany) and Clarion (Biomeda). Sections were analyzed with the Olympus IX 70 microscope.

H. Scherthan acknowledges the technical assistance of D. Gassen, A.W. that of S. Richter, and H.W. that of K. Reumann, U. Matschl and N. Lohrengel. The work of H.W., H. Staege and S.K. was supported by a grant from the Deutsche Krebshilfe and Müggenburg-Stiftung. H. Scherthan received support from the DFG (SCHE350/10-1, SPP1384), The Heinrich-Pette Institut is supported by the Germany Federal Ministry of Health and the Freie und Hansestadt Hamburg.