Male-to-female sex reversal has been previously reported in a Cbx2 mutant mouse model established on the C57BL/6Njcl (B6) genetic background [21]. These mice carry an insertion of a neomycin cassette in exon 5 of the Cbx2 coding sequence resulting in deletion of the C-terminal portion of the Cbx2 protein. Although expected Mendelian inheritance patterns were observed at birth, postnatal lethality affected up to 60% of the mutant offspring before 5–6 weeks of age. In this strain, all surviving chromosomally male Cbx2^(XY−/−) mice exhibited male to female sex reversal with female or intersex-type external genitalia. In addition, the internal reproductive organs consisted of bilateral ovaries containing follicles in 50% of males as well as the presence of both an ovary and a contralateral testis in 25% of mutant males [21].

Abnormalities of sex differentiation have also been reported in an alternative Cbx2 knockout mouse model generated on a BALB/C genetic background [18,21]. However, the extent of phenotypic sex reversal observed in this model is not known. Previous studies have documented striking differences in the level of sensitivity to phenotypic sex reversal amongst different mouse strains and genetic backgrounds [22]. Therefore, we set out to determine the type and severity of gonadal defects observed in BALB/C Cbx2^(−/−) knockout mice. This model [18] involved the targeted deletion of the 5′ untranslated region including the translation start site and the first four exons of the Cbx2 coding sequence leading to a null mutation of the Cbx2 protein. The majority of Cbx2^(−/−) mice in this model show neonatal lethality within hours after birth likely due to proliferative defects in several cell types and homeotic transformations [23].

Mice heterozygous for Cbx2 (Cbx2^(+/−)are viable and fertile and fetuses recovered on day 18.5 post coitum from heterozygous matings revealed a normal ratio of Cbx2^(+/+)/Cbx2^(+/−)/Cbx2^(−/−) genotypes of 31:69:33 (Figure 1A). The chromosomal sex of all 18.5 day post coitum (dpc) fetuses was determined by the presence or absence of Sry gene sequences as detected by PCR. Chromosomally female fetuses from all genotypes (n = 66) contained bilateral ovaries (Figure 1 A–C). However, Cbx2^(XX−/−) females exhibited small ovaries (Figure 1C) compared to control Cbx2^(XX+/+) females (Figure 1B). Importantly, in contrast to wild type controls (Figure 1D; n = 14) chromosomally male Cbx2^(XY−/−) mutant fetuses (n = 14) exhibited severe testicular growth retardation accompanied by impaired formation of testicular cords and in extreme cases unilateral sex reversal as reflected by the presence of a hypoplastic testis and a contralateral ovary in 28.6% of Cbx2^(XY−/−) fetuses (Figure 1E). As expected, the reproductive tracts obtained from wild type controls (Figure 1D) always exhibited bilateral testes at a caudal position towards the inguinal region. However, although the size and morphological appearance of the epididymis was similar to wild type controls, both the testis and adjacent epididymis remained in close proximity to the kidneys in Cbx2^(XY−/−) mutant fetuses (Figure 1E). Histological examination on day 18.5 pc confirmed a reduction in the size and number of testicular cords in Cbx2^(XY−/−) mutant males presenting bilateral testes as well as in the testis of mutant males showing unilateral sex reversal compared to wild type controls Cbx2^(XY+/+) (Figure 1F). Growth retardation and germ cell loss was apparent in both, testes (Figure 1F) as well as ovaries (Figure 1G) from Cbx2^(XY−/−) mutants compared to wild type gonads. Similarly, histological sections obtained from Cbx2^(XX−/−) females confirmed the presence of small ovaries compared to controls (Figure 1G).

These results indicate that, in addition to its role in sex determination, Cbx2 plays an important role in germ cell viability during fetal development. Importantly, our study reveals major differences in the sex reversal phenotype of Cbx2 knockout mice generated on the BALB/C genetic background [18] compared to the phenotype described in Cbx2 deficient mice on the C57BL/6Njcl (B6) background [21]. For example, the majority of Cbx2 null mice on the B6 background exhibit male to female sex reversal and approximately 50% of surviving adult mice exhibited bilateral ovaries. However, none of the knockout males analyzed contained bilateral testes [21]. In contrast, our study reveals that in the BALB/C model <30% of mice exhibit sex reversal. This is consistent with previous studies indicating an extreme sensitivity of the B6 background to genetic perturbations of the sex determination pathway due to higher levels of expression for several genes associated with female gonadal differentiation [22,24]. Interestingly, the majority of affected males in the BALB/C model exhibit bilateral testis and mutant females presented fully differentiated ovaries, making this strain uniquely suited for the analysis of the functional ablation of Cbx2 in mammalian germ cells during the critical window of epigenetic reprogramming and the mitotic to meiotic transition in the germline. These results provide the first evidence indicating a role for Cbx2 in germ cell development.

Next, we determined whether lack of Cbx2 function in the mammalian germline is associated with abnormal chromatin-related events during meiosis. Analysis of meiotic configurations on surface spread oocytes obtained on day 18.5 pc from the ovaries of Cbx2^(XY−/−) males exhibiting unilateral sex reversal revealed abnormal homologous chromosome synapsis involving one or two bivalents as well as several chromosome fragments as determined by synaptonemal complex staining with an anti-SYCP3 antibody (Figure 2A; red). Notably, phosphorylated histone H2AX (γ-H2AX) remained associated with the asynapsed X chromosome and, to a lesser extent, with the Y chromosome (Figure 2A; green). Immuno-FISH experiments using sex chromosome-specific paint probes revealed that in >80% of Cbx2^(XY−/−) oocytes the X (green, arrow) and Y (red, arrowhead) chromosomes fail to undergo partial synapsis (Figure 2B). Moreover, although γH2AX was found associated with the sex chromosomes in the remaining mutant oocytes that exhibited sex chromosome pairing, a typical XY body was not detectable (Figure 2C, D). These results indicate that XY oocytes from sex reversed Cbx2 null gonads exhibit meiotic abnormalities in the form of chromosome breaks in addition to the lack of synapsis of the sex chromosomes.

Surprisingly, our studies also revealed that germ cells obtained from the contralateral testis of Cbx2^(XY−/−) male fetuses on day 18.5 pc undergo precocious entry into meiosis demonstrated by bona fide synaptonemal complex formation and a chromosome configuration consistent with the zygotene stage of meiosis with asynapsed chromosomes and diffuse γH2AX staining (not shown). Strikingly, premature meiotic germ cells were also present in Cbx2^(XY−/−) mutant males with bilateral testes, albeit at less advanced stages of prophase I compared to XY ovarian oocytes (Figure 2E, F). These results indicate that Cbx2 plays a critical role in the pathway responsible for the pre-meiotic arrest in the mammalian male germline. Abnormal chromosome synapsis in Cbx2^(XY−/−) mutants suggests, moreover, a role for Cbx2 in meiotic chromosome synapsis in the male germline. However, whether chromosome synapsis and XY body formation are affected in Cbx2^(XY−/−) spermatocytes undergoing precocious meiosis remains to be investigated.

Our results are consistent with previous studies on the meiotic prophase configuration in XY oocytes obtained from the XY^tdym1 mouse strain, in which only 19% of oocytes exhibit partial synapsis of the sex chromosomes with no apparent sex body formation. Interestingly, in this model, histone posttranslational modifications such as ubiquitinated H2A remained associated with the X chromosome in 50% of oocytes [25]. However, the functional significance of these histone modifications for transcriptional silencing of the sex chromosomes in XY oocytes requires further investigation.

Notably, the identification of fetal male germ cells with precocious synaptonemal complex formation and a zygotene-like configuration on day 18.5 pc constitutes the first evidence indicating that Cbx2 may play a critical role in maintenance of the pre-meiotic arrest in fetal male germ cells. Previous studies elegantly demonstrated that treatment of embryonic testes with retinoic acid agonists or inhibitors of the RA metabolizing enzyme CYP26 induce premature Stra8 expression in cultured male gonads [3,4]. In addition, treatment of fetal testis with the histone deacetylase inhibitor TSA induced Stra8 expression and SYCP3-positive staining on histological sections [7]. However, the presence of meiotic figures in these experimental systems remained to be demonstrated. Our results show for the first time that in the absence of Cbx2 function, fetal testicular germ cells prematurely transit into meiosis and exhibit synaptonemal complex formation, albeit at a lower frequency compared to ovarian germ cells. Results obtained following exposure of fetal testes to histone deacetylase inhibitors suggest that, in addition to the stimulus provided by RA, epigenetic mechanisms might also contribute to prevent premature meiosis onset in the male germline [7]. Interestingly, valproic acid, a specific inhibitor of histone deacetylases, has been recently shown to reduce the expression of Cbx2 transcripts in mouse embryos on day 8 pc [26]. This suggests that the role of Cbx2 in modulating large-scale chromatin structure may be critical for antagonizing the potential effects of retinoic acid exposure on premature meiotic entry in the male germline. In support of this notion, previous studies indicate that Cbx2 is essential to control the developmental window of responsiveness to retinoic acid during fetal development [19].

Cbx2 associates with pericentric heterochromatin domains during mitosis in both murine somatic and embryonic stem cells. The protein also localizes to facultative heterochromatin at the inactive X chromosome suggesting a potential role in heterochromatin formation [27–29]. However, whether Polycomb group proteins, such as Cbx2, play a role in meiotic chromosome synapsis in the female germline is not known. Therefore, we undertook a detailed analysis of meiotic synapsis and recombination events in Cbx2 mutant oocytes. The dynamics of chromosome synapsis was determined following simultaneous staining of the lateral elements with SYCP3 (red) as well as the central elements of the synaptonemal complex with SYCP1 (green). At the pachytene stage, 80% of wild type oocytes (n = 90) exhibited fully synapsed chromosome bivalents (Figure 3A arrow, C). In contrast, mutant oocytes (n = 71) exhibited a significant increase in the incidence (57%) of abnormal synapsis (Figure 3B, arrowhead) and Figure 3C. Moreover, chromosomal associations between two partially synapsed bivalents, indicative of non-homologous chromosome interactions (arrows), were commonly observed in >30% of Cbx2^(XX−/−) oocytes (Figure 3B, C). These results indicate a previously unrecognized role of Cbx2 in chromosome pairing and homologous chromosome synapsis in female germ cells.

All animal experiments were approved by the institutional animal use and care committee of the University of Pennsylvania according to National Institutes of Health guidelines.

Mice heterozygous for a targeted deletion of Cbx2 (C.129P2-Cbx2^tm1Cim/J) [23] were obtained from the Jackson Laboratory (Bar Harbor, ME). To obtain fetuses carrying a homozygous null mutation (Cbx2^(−/−)), heterozygous mating pairs were set up and fetal gonads were collected on day 18.5 pc. Genotyping of Cbx2^(−/−) mice was conducted according to the Jackson Laboratory's genotyping protocol for Cbx2^tm1Cim, version 1.1 using the following primer pairs: “Cbx2-fwd AAC CGG AAG AGA GGC AAG AG”; “Cbx2-rev CCT GAA GGA GCA ACA AGA AAG”; “Neo-rev AGG TGA GAT GAC AGG AGA TC”. Genomic tail DNA was extracted by using the DNeasy Blood and Tissue Kit (QIAGEN Sciences, Germantown, MD) following manufacturer's instructions. The PCR conditions to determine the genotype were as follows: 94 °C for 3 min; 38 cycles of 94 °C for 30 s, 58 °C for 1 min and 72 °C for 1 min; followed by 72 °C for 10 min, and maintained at 4 °C until gel electrophoresis (Bio-Rad Laboratories, Hercules, CA). The chromosomal sex of each fetus was analyzed in parallel using the primer pair “Sry-fwd CTG TGT AGG ATC TTC AAT CTC T” and “Sry-rev GTG GTG AGA GGC ACA AGT TGG C” specifically amplifying a genomic region within the Y chromosomal SRY gene.

Fetal gonads were dissected from control and Cbx2^(−/−) fetuses on day 18 pc and immediately fixed in Bouin's solution (Sigma, St. Louis, MI). Following overnight fixation, the gonads were washed in PBS solution and processed for paraffin embedding and sectioning according to standard procedures. 5 μm serial sections were stained with H&E before microscopy and image acquisition. Fetal gonads recovered from fetuses on day 18.5 pc were also processed for the analysis of chromosome synapsis and meiotic recombination proteins on surface spread germ cells as described previously [70]. Briefly, gonads were dissected and transferred to a hypotonic solution containing 1% sodium citrate for 20 min, before spreading of meiotic configurations onto glass slides and fixation in a solution of 1% paraformaldehyde, 0.1% Triton X 100.

Unless otherwise specified, all primary antibody incubations were conducted overnight at 4 °C. The extend of meiotic chromosome synapsis in control and mutant germ cells was determined by immunochemical detection of the lateral elements of the synaptonemal complex protein SYCP3 using a 1:1000 dilution of a rabbit anti-SYCP3 antibody (abcam, Cambridge, MA). An anti-phosphohistone H2AX (Ser-139) antibody (Upstate, Charlottesville, VA) was used at a 1:500 dilution. The type of meiotic configuration present in control and mutant oocytes at the pachytene stage was analyzed by simultaneous staining of the central element of the synaptonemal complex with a rabbit anti-SYCP1 antibody (1:500) and a guinea pig anti-SYCP3 at a 1:250 dilution [71]. Antibodies directed against BRCA1 (goat polyclonal, Santa Cruz, CA) and ubH2A (mouse monoclonal, clone E6C5, Millipore, Billerica, MA) were used at a dilution of 1:400. The mouse monoclonal anti-Mlh1 (BD Pharmingen) and mouse polyclonal anti-Rad51 (Oncogene) antibodies were both used following an overnight incubation at 37 °C at a 1:50 dilution in combination with a rabbit anti-SYCP3 antibody (1:250). Immunodetection was performed using appropriate Alexa Fluor-conjugated secondary antibodies (Molecular Probes, Eugene, Oregon, USA) at a dilution of 1:1000 for 2 h at room temperature.

Following immunochemistry, slides were counterstained with 7 μL of antifading medium supplemented with DAPI (Vectashield; Vector Laboratories, Burlingame, CA). Epifluorescence analysis of meiotic chromosomes was conducted on a DMRX/E microscope (Leica Microsystems) using a 100× objective. Images were captured with a Leica DFC 350F camera using Openlab 3.1.7. software (PerkinElmer) and image processing was performed using Photoshop 2.0 (Adobe) for linear adjustments and cropping of fluorescent images. No gamma adjustments were made.

The position of the sex chromosomes in control and mutant germ cells was determined by Immuno-FISH analysis following immunochemistry on the same slide using a FITC-conjugated X and a Cy3-conjugated Y chromosome paint (STARFISH, Cambio) according to manufacturer's specifications. Briefly, surface spread interphase nuclei and metaphase chromosomes were denatured in 70% formamide (VWR International Ltd., Poole, UK) in 2 × SSC at 80 °C for 10 min and subsequently chilled in ice-cold 70% ethanol for 5 min. The X and Y chromosome-specific probes were denatured for 10 min at 85 °C, respectively, and incubated at 40 °C for 1 h. Overnight hybridization was carried out in a humidified chamber at 40 °C and stringency washes were conducted in a solution containing 50% formamide in 2 × SSC as previously described [64,72]. Co-localization of a fully synapsed bivalent with the signal provided by the X chromosome-specific probe was considered indicative of complete homologous pairing of the X bivalent in XX oocytes at the pachytene stage.

Data presented as percentage values were analyzed by one-way analysis of variance (ANOVA) following arcsine transformation. Comparison of all pairs was conducted by the Tukey-Kramer HSD or Student's t-test using JMP Start Statistics (SAS Institute Inc., Cary, NC). Variation among individual replicates is indicated as the standard deviation (s.d.). Differences were considered significant when P < 0.05 and are indicated by different superscripts.