Abstract
Sex determination of the gonads begins with fate specification of gonadal supporting cells into either ovarian pre-granulosa cells or testicular Sertoli cells. This fate specification hinges on a balance of transcriptional control. Here we report that expression of the transcription factor RUNX1 is enriched in the fetal ovary in rainbow trout, turtle, mouse, goat, and human. In the mouse, RUNX1 marks the supporting cell lineage and becomes pre-granulosa cell-specific as the gonads differentiate. RUNX1 plays complementary/redundant roles with FOXL2 to maintain fetal granulosa cell identity and combined loss of RUNX1 and FOXL2 results in masculinization of fetal ovaries. At the chromatin level, RUNX1 occupancy overlaps partially with FOXL2 occupancy in the fetal ovary, suggesting that RUNX1 and FOXL2 target common sets of genes. These findings identify RUNX1, with an ovary-biased expression pattern conserved across species, as a regulator in securing the identity of ovarian-supporting cells and the ovary.
Subject terms: Differentiation, Gene regulation, Reproductive biology
Proper ovarian differentiation requires RUNX1. Here, the authors show that double knockout of Runx1/Foxl2 results in masculinization of fetal ovaries, and that RUNX1 and FOXL2 jointly occupy common chromatin regions to maintain pre-granulosa cell identity in the fetal ovary.
Introduction
A critical step that shapes the reproductive identity of the embryo is the sexual differentiation of the bipotential gonads. Supporting cells in the fetal gonads are the first cell population to differentiate and dictate the fate of the gonads. As a consequence, defects in supporting cell differentiation have dire consequences on reproductive outcomes of the individual, from sex reversal to infertility. Supporting cells differentiate into either Sertoli cells, which drive testis development, or pre-granulosa cells, which control ovarian development. It has become clear that supporting cell differentiation and maintenance of their commitment requires a coordinated action of multiple factors that play either complementary, redundant, and even antagonistic roles1. For instance, fate decision and maintenance of ovarian identity relies mainly on two conserved elements: the WNT4/RSPO1/β-catenin pathway2–5 and the transcription factor FOXL26–8. These two elements synergistically promote expression of pro-ovarian genes and, at the same time, antagonize key pro-testis factors such as SOX9 and DMRT1. However, the combined loss of these two key pro-ovarian signaling only results in an incomplete inactivation of ovarian differentiation, suggesting that additional pro-ovarian factors are at play during gonadal differentiation9,10. Factors involved in gonad differentiation are generally conserved in vertebrates and even invertebrates, although their position in the hierarchy of the molecular cascade may change during evolution11. For instance, the pro-ovarian transcription factor FOXL2 is important for ovarian differentiation/function in human12, goat13, and fish14,15. The pro-testis transcription factor DMRT1 is highly conserved and critical for testis development in worms, fly16, fish17,18, and mammals19–21.
In this study, we set up to investigate the role of transcription factor RUNX1 in the mouse fetal ovary. In Drosophila melanogaster, the RUNX ortholog runt is essential for ovarian determination22,23. In the mouse, Runx1 mRNA is enriched in the fetal ovary based on transcriptomic analyses24. The RUNX family arose early in evolution: members have been identified in metazoans from sponge to human, where they play conserved key roles in developmental processes. In vertebrates, RUNX1 acts as a transcription factor critical for cell lineage specification in multiple organs and particularly in cell populations of epithelial origin25. We first characterize the expression profile of RUNX1 in the fetal gonads in multiple vertebrate species, from fish to human. We then use knockout (KO) mouse models and genomic approaches to determine the function and molecular action of RUNX1 and its interplay with another conserved ovarian regulator, FOXL2, during supporting cell differentiation in the fetal ovary.
Results
Runx1 expression pattern implies a role in ovary development
The runt gene, critical for ovarian determination in the fly22, has three orthologs in mammals: RUNX1, RUNX2, and RUNX3. Although all three RUNX transcription factors bind the same DNA motif, they are known to have distinct, tissue-specific functions26. In the mouse, Runx1 was the only one with a strong expression in the fetal ovary, whereas Runx2 and Runx3 were expressed weakly in the fetal gonads in a non-sexually dimorphic way (Fig. 1a). At the onset of sex determination (Embryonic day 11.5 or E11.5), Runx1 expression was similar in both fetal XY (testis) and XX (ovary) gonads before becoming ovary-specific after E12.5 (Fig. 1b), consistent with observations by others24,27. An ovary-enriched expression of Runx1 during the window of early gonad differentiation was also observed in other mammals such as human and goat, as well as in species belonging to other classes of vertebrates such as red-eared slider turtle and rainbow trout (Fig. 1c–f), implying an evolutionarily conserved role of RUNX1 in ovary differentiation.
To identify the cell types that express Runx1 in the gonads, we examined a reporter mouse model that produces enhanced green fluorescent protein (EGFP) under the control of Runx1 promoter28 (Fig. 2 and Supplementary Fig. 1). Consistent with Runx1 mRNA expression (Fig. 1b), Runx1-EGFP was present in both XX and XY gonads at E11.5, then increased in XX gonads and diminished in XY gonads at E12.5 onwards (Fig. 2). At E11.5 in both XX and XY gonads, Runx1-EGFP was present in a subset of SF1+/PECAM− somatic cell population, whereas it was absent in the SF1−/PECAM+ germ cells (Fig. 2a–d). In the XY gonads, these Runx1-EGFP+ somatic cells corresponded to Sertoli cells, as demonstrated by a complete overlap with SRY, the sex-determining factor that drives Sertoli cell differentiation29 (Fig. 2e and Supplementary Fig. 1). At this stage, there is no marker for ovarian-supporting cells that allow us to determine which subset of somatic cells were positive for Runx1-EGFP in XX gonads. However, at E12.5, when the sex of gonads becomes morphologically distinguishable, Runx1-EGFP was specifically detected in the supporting cell lineage of both sexes: strongly in FOXL2+ pre-granulosa cells of XX gonads (Fig. 2g) and weakly in SOX9+ Sertoli cells of XY gonads (Fig. 2f and Supplementary Fig. 1). Runx1-EGFP expression was eventually turned off in the fetal testis, while it was maintained in the ovary (Fig. 2h, i). Throughout fetal development of the ovary, Runx1-EGFP remained in FOXL2+ pre-granulosa cells (Fig. 3). Runx1-EGFP was also detected in the ovarian surface epithelium at E16.5 and birth (arrows in Fig. 3b, c), which gives rise to granulosa cells in the cortex of the ovary30,31. Runx1-EGFP was also expressed in somatic cells of the cortical region right underneath the surface epithelium at E16.5 and some of these Runx1-EGFP+ cells presented a weak expression of FOXL2 (Fig. 3g–i, arrowheads). In summary, Runx1 marks the supporting cell lineage in the gonads at the onset of sex determination and becomes pre-granulosa cell-specific as gonads differentiate.
XX Runx1 KO and Foxl2 KO share common transcriptomic changes
Its pre-granulosa cell-specific expression suggests that RUNX1, a factor involved in cell lineage determination25, could contribute to granulosa cell differentiation and ovarian development. To investigate its role in gonads and avoid early embryonic lethality as a result of global deletion of Runx132,33, we generated a conditional KO mouse model in which Runx1 was ablated in the SF1+ gonadal somatic cells34 (Fig. 4). We characterized the effects of Runx1 inactivation on XX gonad differentiation at E14.5, a stage where morphological differences between ovary and testis are already established. Although Runx1 expression was ablated successfully in XX gonads (Fig. 4d), ovarian morphogenesis appeared normal: E14.5 XX Runx1 KO gonads presented similar size and shape compared with XX control gonads (Supplementary Fig. 2), and differentiation and organization of different gonadal cell populations were similar to control XX gonads (Fig. 4a–c). For instance, XX Runx1 KO-supporting cells expressed the pre-granulosa cell marker FOXL2 but not Sertoli cell markers SOX9 and AMH (Fig. 4a, b). Supporting cells and germ-cell organization was similar to XX control gonads and did not form Laminin-outlined cord structures typical of testis differentiation (Fig. 4b). Similar to supporting cells, no difference was observed in the COUP-TFII+ interstitium organization between XX Runx1 KO and XX control gonads (Fig. 4b). Finally, E14.5 XX Runx1 KO germ cells had initiated meiosis, a typical feature of fetal ovary development, and normal expression of the germ-cell marker Mvh was observed (Fig. 4c, d). No significant change was detected for key genes involved in pre-granulosa cell differentiation (Foxl2 and Wnt4) or Sertoli cell differentiation (Sox9, Fgf9, Nr5a1, and Amh; Fig. 4d, e). However, pre-granulosa cell marker Fst and fetal Sertoli cell marker desert hedgehog (Dhh) were mis-expressed in E14.5 XX Runx1 KO gonads (Fig. 4d, e).
Similarly, at birth, XX Runx1 KO gonads maintained a typical ovarian shape, with FOXL2+ pre-granulosa cells scattered throughout the gonad and TRA98+ germ cells located mostly in the cortex (Fig. 4f). Despite their normal ovarian morphology, newborn XX Runx1 KO gonads exhibited an aberrant transcriptomic profile reminiscent of the transcriptome of newborn XX Foxl2 KO gonads, such as downregulation of pro-ovarian genes that are direct targets of FOXL2 (Fig. 4g and Supplementary Data 1). Foxl2 is involved in ovarian differentiation/maintenance in various vertebrate species. In the mouse, loss of Foxl2 results in normal ovarian morphogenesis at birth, despite aberrant ovarian transcriptome, and eventually leads to masculinization of the ovary postnatally6. We found that 41% of the genes differentially expressed in Runx1 KO were also misregulated in the absence of Foxl2 in newborn XX gonads (Fig. 4g and Supplementary Data 1). However, contrary to Foxl2 KO, loss of Runx1 in the ovary did not result in postnatal sex reversal and the females Runx1 KO were fertile. One possible explanation for these common transcriptomic changes is that Runx1 could be part of the same signaling cascade as Foxl2. However, analysis of Runx1 expression in Foxl2 KO newborn ovaries did not detect any changes in Runx1 expression in the absence of Foxl2 (Fig. 4h). Conversely, Foxl2 expression was not changed in the absence of Runx1. These results indicate that Runx1 and Foxl2 regulate common sets of genes but they are themselves regulated independently of each other in the fetal ovary.
Runx1 and Foxl2 double knockout results in partial masculinization of fetal ovaries
The common transcriptomic changes identified in Runx1 and Foxl2 KO newborn ovaries raised the question whether RUNX1 and FOXL2 could play redundant/complementary roles in supporting cell differentiation. We therefore generated Runx1/Foxl2 double KO mice (referred as DKO) and compared XX gonads differentiation in the absence of Runx1, Foxl2, or both (Fig. 5 and Supplementary Figs. 3 and 4). Abnormal development of XX DKO gonads became apparent around E15.5. At this stage, differentiation of supporting cells into Sertoli cells in the testis or pre-granulosa cells in the ovary has been established. For instance, the transcription factor DMRT1, involved in the maintenance of Sertoli cell identity19, is expressed in Sertoli cells but not pre-granulosa cells (Fig. 5a, e). At E15.5, DMRT1 is also present in a few germ cells in both the testis and ovary35. Similar to control ovaries, XX gonads lacking either Runx1 or Foxl2 had no DMRT1 proteins in the supporting cells (Fig. 5a–c). However, the combined loss of Runx1 and Foxl2 resulted in aberrant expression of DMRT1 in the supporting cells of XX gonads (Fig. 5d). At birth, a time when XX Foxl2 KO gonads still morphologically resemble ovaries6,7, XX Runx1/Foxl2 DKO gonads formed structures similar to fetal testis cords in the center, with DMRT1+ cells surrounding clusters of germ cells (Fig. 5i). Such structure was not observed in XX Runx1 KO or Foxl2 KO gonads with the exception that DMRT1 protein started to appear in a few supporting cells in the newborn XX Foxl2 KO gonads, in what appears to be one of the first signs of postnatal masculinization of Foxl2 KO ovaries at the protein level (Fig. 5h). Contrary to DMRT1, SOX9 protein, a key driver of Sertoli cell differentiation36,37, was not detected in XX Runx1/Foxl2 DKO newborn gonads (Fig. 6). Our results demonstrate that a combined loss of Runx1/Foxl2 induces partial masculinization of the supporting cells during fetal development of the ovary.
To further characterize the impacts of the combined loss of Runx1/Foxl2 on ovarian differentiation, we compared the transcriptome of newborn XX Runx1/Foxl2 DKO gonads with the transcriptomes of XX control, Runx1, or Foxl2 single KO gonads (Fig. 7 and Supplementary Data 2). The heat map for the 918 differentially expressed genes between XX Runx1/Foxl2 DKO and XX control gonads demonstrated allele-specific impacts: loss of Runx1 resulted in a mild and often nonsignificant effect on these genes, loss of Foxl2 had an intermediate/strong effect, and combined loss of Runx1/Foxl2 yielded the strongest effect (fold change > 1.5; p < 0.05 one-way analysis of variance (ANOVA); Fig. 7a and Supplementary Data 3). Gene ontology analysis revealed that the downregulated genes were associated with “ovarian follicle development” and “female gonad development,” whereas “male sex determination” was the most significantly enriched process for the upregulated genes (Supplementary Fig. 5). Conforming to the hierarchical clustering (Fig. 7a), comparison of the genes significantly changed in Runx1/Foxl2 single and DKOs suggested that Foxl2 was the main contributor to the transcriptional changes observed in Runx1/Foxl2 DKO (Fig. 7b–g and Supplementary Data 4): 61% of the genes downregulated in DKO were also downregulated in Foxl2 KO and 43% of the genes upregulated in DKO were also upregulated in Foxl2 KO. In addition, some genes appeared to be controlled by both Foxl2 and Runx1, and were significantly downregulated or upregulated in all three KOs (Fig. 7b, c). For instance, the genes Fst and Cyp19a, both involved in granulosa cell differentiation/function38,39, were downregulated in Runx1 KO, Foxl2 KO, and more repressed in Runx1/Foxl2 DKO (Fig. 7d). On the other hand, Dhh was upregulated in all three KOs, with the highest expression in the DKO (Fig. 7f). Finally, some genes were significantly changed in Runx1/Foxl2 DKO only, suggesting a cumulative effect of Runx1 and Foxl2 loss. For instance, Foxp1, a gene whose expression is enriched in pre-granulosa cells40, was significantly downregulated only in Runx1/Foxl2 DKO (Fig. 7e), whereas Fgf9, a Sertoli gene contributing to testis differentiation41, and Pdgfc were significantly upregulated only in Runx1/Foxl2 DKO (Fig. 7g).
In contrast to XX Foxl2 single KO gonads (Fig. 5), in which sex reversal only became apparent postnatally6, XX Runx1/Foxl2 DKO gonads exhibited masculinization with visible morphological changes before birth. To determine how the additional loss of Runx1 contributed to the earlier masculinization of Runx1/Foxl2 DKO ovaries, we identified the genes differentially expressed between XX Runx1/Foxl2 DKO and XX Foxl2 KO gonads (Fig. 7h and Supplementary Data 5). Expression of most of these genes was already altered in XX Foxl2 single KO gonads; however, the additional loss of Runx1 exacerbated their mis-expression. For instance, the pro-testis gene Dmrt1 and Nr5a1 were significantly upregulated, whereas the pre-granulosa-cell-enriched transcripts Fst and Ryr2 were further downregulated at birth (Fig. 7d, i). On the other hand, the additional loss of Runx1 did not cause further upregulation of the Sertoli genes Sox9 and Amh at birth, suggesting that Runx1 does not contribute to their repression in the ovary (Fig. 7j). Overall, the transcriptomic analyses of Runx1/Foxl2 single and DKOs revealed that Foxl2 is the main driver of the transcriptomic changes, and that the additional loss of Runx1 amplifies the mis-expression of genes already altered by the sole loss of Foxl2, leading to the failure to maintain pre-granulosa cell identity in the fetal ovary.
RUNX1 shares genome-wide chromatin occupancy with FOXL2
The masculinization of XX Runx1/Foxl2 DKO fetal gonads and the transcriptomic comparisons of newborn XX Runx1/Foxl2 single and DKO gonads suggest some interplay between RUNX1 and FOXL2 to control pre-granulosa cell identity. The fact that RUNX1 and FOXL2 are both transcription factors expressed in the pre-granulosa cells (Fig. 3 and Supplementary Fig. 1) raised the question whether this interplay could occur directly at the chromatin level. We have previously identified FOXL2 chromatin occupancy during ovarian differentiation by chromatin immunoprecipitation (ChIP) followed by whole-genome sequencing42 at E14.5, a time where FOXL2 and RUNX1 expression fully overlaps (Fig. 3). We performed additional de novo motif analyses on the genomic regions bound by FOXL2 in the fetal ovary, and discovered that several other DNA motifs were co-enriched with FOXL2 DNA motif (Fig. 8a). RUNX DNA motif was the second most significantly co-enriched motif. The other motifs were for CTCF, a factor involved in transcriptional regulation, enhancer insulation, and chromatin architecture43, and for the DNA motif recognized by members of the nuclear receptor family including liver receptor homolog-1 (LRH-1 encoded by Nr5a2) and SF1 (encoded by Nr5a1), a known co-factor of FOXL244,45. DNA motifs for TEAD transcription factors of the Hippo pathway, ETS, NFYA, and GATA4, another factor involved in gonad differentiation46, were also significantly enriched. The enrichment of RUNX motif with FOXL2-binding motif suggests that RUNX1, the only RUNX also expressed in pre-granulosa cells, could bind similar genomic regions to FOXL2 in the fetal ovary. To confirm this hypothesis, we performed ChIP-sequencing (ChIP-seq) for RUNX1 in E14.5 ovaries (Supplementary Data 6), the same stage as FOXL2 ChIP-seq42. The top de novo motif identified in RUNX1 ChIP-seq (p < 1e − 559) matched the RUNX motif47 (Fig. 8b) and corresponded to the motif that was co-enriched with FOXL2 in FOXL2 ChIP-seq (Fig. 8a). A total of 10,494 RUNX1-binding peaks were identified in the fetal ovary, with the majority of the peaks located either in the gene body (Fig. 8c; 25% exon and 22% intron) or close upstream of the transcription start site (TSS) (30% Promoter: < 1 kb of TSS; 12% Upstream: −10 to −1 kb of TSS). Comparison of genome-wide chromatin binding of RUNX1 and FOXL2 in the fetal ovary revealed significant overlap: 54% (5619/10,494) of RUNX1 peaks overlapped with FOXL2 peaks (Fig. 8d).
The transcriptomic data from Runx1/Foxl2 DKO ovaries provided us a list of genes significantly changed as a result of the absence of Runx1, Foxl2, or both (Fig. 7). To identify potential direct target genes of RUNX1 or/and FOXL2, we focused on the 918 genes differentially expressed in Runx1/Foxl2 DKO ovaries and determined which genes were nearest to RUNX1 or/and FOXL2-binding peaks (Fig. 9a and Supplementary Data 7). More than 50% of these genes (492/918; Fig. 9a) were the closest gene to RUNX1 or/and FOXL2 peaks. Some of these genes were nearest to only FOXL2 peaks (116 genes in Fig. 9a). For example, Pla2r1, a transcript enriched in pre-granulosa cells40 and similarly downregulated in both Foxl2 KO and Runx1/Foxl2 DKO fetal ovaries (Fig. 9b), contained two FOXL2-specific peaks, one in the promoter and one in the first intron. On the other hand, 102 genes (Fig. 9a) had RUNX1-specific peaks near their genomic locations. For instance, Ryr2, another transcript enriched in pre-granulosa cells40, was strongly downregulated in both Runx1 KO and Runx1/Foxl2 DKO fetal and newborn ovaries (Figs. 7i and 9c), and contained one RUNX1-specific peak in its intronic region. Finally, 274 genes were the closest genes to the peaks for both RUNX1 and FOXL2, with the majority of them (197 genes) nearest to overlapping peaks for RUNX1 and FOXL2 (Fig. 9a). Most of these genes were downregulated in Runx1/Foxl2 DKO ovaries (Supplementary Data 7). For instance, the pre-granulosa cell-enriched genes Fst and Itpr2, both downregulated in Runx1/Foxl2 single and DKO ovaries (Figs. 7d and 9d), contained common binding peaks for FOXL2 and RUNX1 (Fig. 9a, d). For Fst, this binding of RUNX1 and FOXL2 was located in its first intron, in the previously identified regulatory region that contributes to its expression42,48. On the other hand, Sertoli cell -enriched gene Dmrt1, which was strongly upregulated in Runx1/Foxl2 DKO (Figs. 5 and 7i), contained a common binding site for FOXL2 and RUNX1 near its promoter (Fig. 9a). Taken together, our results reveal that RUNX1, a transcription factor expressed in the fetal ovary of various vertebrate species, contributes to ovarian differentiation and maintenance of pre-granulosa cell identity through an interplay with FOXL2 that occurs at the chromatin level.
Discussion
The molecular events that specify pre-granulosa cell fate are complex and nonlinear, involving several signaling pathways that have redundant and complementary functions. This is in contrast to the fetal testis, where the molecular pathway driving its differentiation appears linear and sequential. Removal of one of the top pieces in the testis differentiation pathway has a domino effect that prevents induction of downstream events. This is exemplified by the complete gonadal sex reversal in gain- or loss-of-function mouse models for SRY or SOX9, the two transcription factors responsible for the initiation of the testis morphogenesis29,36,49. This is not the case in the mouse ovary, where no single-gene loss/mutation results in a complete ovary-to-testis sex reversal. For instance, defects in the WNT4/RSPO1/β-catenin pathway or loss of Foxl2 causes a late or postnatal partial ovary-to-testis sex reversal, whereas the combined loss of Foxl2 and elements of the WNT4/RSPO1/β-catenin pathway (Wnt4 or Rspo1) leads to ovary-to-testis sex reversal more pronounced than each single KO model in the mouse9,10. In this study, we demonstrated that Runx1 contributes to the molecular network controlling pre-granulosa cell differentiation. Loss of Runx1 in somatic cells of the ovaries altered ovarian transcriptome but did not affect ovarian morphogenesis during fetal development. In contrast, combined loss of Runx1 and Foxl2 compromised pre-granulosa cell identity. Loss of Runx1 or Foxl2 affected common sets of genes, suggesting some redundancy in their functions in the fetal ovary. These transcriptomic changes were enhanced in the absence of both genes, reaching a threshold that masculinized the fetal ovary. Based on the phenotypic and transcriptomic analyses, we propose that FOXL2 is the dominant player and RUNX1 acts as a supporting player. For instance, global KO of Foxl2 in XX gonads results in postnatal sex reversal6, whereas the XX Runx1 single KO has no such impacts. In addition, loss of Foxl2 leads to transcriptomic changes more prominent than loss of Runx1 in newborn ovaries, even though they affect common sets of genes. However, the additional loss of Runx1 amplifies the transcriptomic changes caused by the loss of Foxl2, resulting in fetal masculinization. One of the most striking changes in Runx1/Foxl2 DKO ovaries is the expression of DMRT1 in the fetal supporting cells. DMRT1 is a key driver of Sertoli cell differentiation and testis development in various species16,21. In the fly, doublesex (dsx), an ortholog of mammalian DMRT1, controls testis differentiation16. Intriguingly, runt, the fly ortholog of RUNX1, tips the balance toward ovarian determination by antagonizing the testis-specific transcriptional regulation of dsx22. In the mouse, testis differentiation is not controlled by DMRT1 but by SOX transcription factor SRY and its direct target SOX9. However, RUNX1 does not appear to contribute to the repression of the key pro-testis gene Sox9 in the fetal ovary and SOX9 protein was not detected in Runx1/Foxl2 DKO ovary at birth. This is in contrast with the phenotype of Wnt4/Foxl2 DKO newborn ovaries where SOX9 was upregulated, and as a consequence the ovaries were more masculinized10. Overall, our findings suggest that slightly different pro-ovarian networks control the repression of the evolutionary conserved pro-testis genes Sox9 and Dmrt1: Sox9, which plays a primary role in Sertoli cell differentiation in the mouse, is repressed by an interplay between the WNT4/RSPO1/β-catenin and FOXL29,10. On the other hand, Dmrt1, which has taken a secondary role in Sertoli cell differentiation in the mouse, is repressed by an interplay between RUNX1 and FOXL2. The fact that RUNX1 does not appear to control Sox9 may be the reason why RUNX1 only plays a secondary role in granulosa cell differentiation/maintenance, in contrast to the dominant roles of the WNT pathway and FOXL2. It would be interesting to determine the role of RUNX1 in species for which DMRT1 plays a more prominent role in initiation of differentiation of the testis.
Seeking the mechanisms underlying the interplay between RUNX1 and FOXL2 in the regulation of pre-granulosa cell identity, we identified that RUNX DNA-binding motif is significantly co-enriched with FOXL2 motif in genomic regions bound by FOXL2 in the fetal ovary. The fact that RUNX1 genome-wide chromatin occupancy partially overlaps with FOXL2 in the fetal ovary, that RUNX1 and FOXL2 are expressed in the same cells at the time of the ChIP-seqs, that their loss affects common set of genes, and that the DKO results in gonad masculinization, altogether support the model in which RUNX1 and FOXL2 jointly occupy common chromatin regions that control the maintenance of pre-granulosa cell identity. By themselves, RUNX proteins are weak transcription factors and they require other transcriptional regulators to function as either repressors or activators of transcription26. Interplay between RUNX1 and several members of the forkhead transcription factor family has been documented in different tissues. For instance, RUNX1 is a co-activator of FOXO3 in hepatic cells50. Similarly, an interplay between RUNX1 and FOXO1/FOXO3 was demonstrated in breast epithelial cells where a subset of FOXO target genes were jointly regulated with RUNX151. Another forkhead protein, FOXP3, acts with RUNX1 to control gene expression in T cells52 and breast epithelial cells53. Such cooperation in various tissues suggest that the interplay between RUNXs and forkhead transcription factors maybe an evolutionary conserved phenomenon. It would be interesting to determine whether the forkhead protein FOXL2 is also able to physically interact with RUNX1 in the fetal ovaries and how this interaction occurs. Our data and single-cell sequencing data27 demonstrate that RUNX1 starts to be expressed in the supporting cells earlier than FOXL2. In other tissues, it was demonstrated that RUNX1 can function as a pioneer factor that allows chromatin remodeling and recruitment of other factors to control gene expression and cell fate54. Therefore, it is possible that RUNX1 is first recruited to chromatin regions and the presence of RUNX1 facilitates the recruitment of FOXL2 to maintain pre-granulosa cell identity. Finally, in addition to the genes co-regulated by FOXL2 and RUNX1, we identified genes that were specifically mis-expressed in the absence of Runx1 but not Foxl2. Genome-wide analyses of RUNX1 binding in the fetal ovary also identified genomic regions bound by RUNX1 but not FOXL2. These results suggest that RUNX1 could also contribute to ovarian development or function independently of FOXL2.
RUNX1 contributes to cell-fate determination in various developmental processes such as hematopoiesis and hair follicle development. Depending on its interplay with other signal transduction pathways or co-factors, RUNX1 controls which path the precursor cells take when they are at the crossroad between cell proliferation/renewal and lineage-specific commitment25. We discovered that Runx1 has an ovary-biased expression during gonad differentiation in various vertebrate species, including turtle, rainbow trout, goat, mouse, and human. In mouse embryonic gonads, Runx1 is first detected in the supporting cells in a non-sexually dimorphic way at the onset of sex determination. Although its expression is maintained in the ovary, Runx1 appears to be actively repressed in the testis between E11.5 and E12.5 as the supporting cells commit to Sertoli cell fate. The suppression of Runx1 in the fetal testis is corroborated by previously published data from a time-course transcriptomic analysis during early gonad development55 and single-cell sequencing analysis of SF1+ progenitor cells27,56. The time course of Sertoli cell differentiation at the single-cell level revealed that Runx1 follows a similar spatiotemporal pattern of expression with Sry56. In the mouse, Sry expression in Sertoli cells is quickly turned off after the initiation of testis differentiation and it is suspected that the repression of Sry is due to a negative feedback loop by downstream pro-testis genes. The similar pattern of downregulation of Runx1 in the testis after E11.5 raises the possibility that Runx1 is downregulated by a similar signaling pathway. Regulation of Runx1 gene expression is complex and several enhancers that confer tissue-specific expression have been identified57. It remains to be determined how Runx1 expression is controlled in the gonads and how it is actively repressed in the fetal testis.
In contrary to the testes, fetal ovaries maintain expression of Runx1 in the supporting cells as they differentiate into pre-granulosa cells. During ovarian differentiation, granulosa cells arise from two different waves: the first cohort of granulosa cells arises from the bipotential supporting cell precursors that differentiate into either Sertoli cells or pre-granulosa cells during sex determination58. The second wave of granulosa cells that eventually populate the cortical region of the ovary appears later in gestation. This second wave arises from LGR5+ cells of the ovarian surface epithelium that ingress into the ovary from E15.5 to postnatal day 4 and eventually become LGR5-/FOXL2+ granulosa cells30,31. This timing of establishment of the second cohort of granulosa cells correlates with the expression of Runx1-EGFP in a subset of cells in the surface epithelium and in granulosa cells of the ovarian cortex at E16.5 and birth. These results suggest that Runx1 also marks granulosa cell precursors that will give rise to the second wave of FOXL2+ granulosa cells in the cortex. Therefore, both expression at onset of sex determination and at the surface epithelium/cortex at the time of the second wave of granulosa cells recruitment suggest that Runx1 is activated in cells that are primed to become supporting/granulosa cells.
Multiple transcription factors often form complex genetic regulatory networks that control cell-fate determination. Genomic sequence motifs or cis-regulatory elements for the supporting cell lineage in the testis were identified by combined analyses of SOX9 and DMRT1 ChIP-seq, and by motif prediction59. These “Sertoli cell signatures” are composed of binding motifs for transcription factors critical for Sertoli cell differentiation, including SOX9, GATA4, and DMRT1. These Sertoli cell signatures, present in mammals and other vertebrates, could represent a conserved regulatory code that governs the cascade of Sertoli cell differentiation, regardless of whether it primarily relies on SOX transcription factors such as SRY in mammals or on DMRT1 such as in several vertebrate species. Similarly, one would expect the presence of conserved “granulosa cell signature” genomic regions that confers granulosa cell differentiation. As FOXL2 is a highly conserved gene in granulosa cell differentiation in vertebrates, we used FOXL2 as an anchor factor to identify other factors that could take part in the regulatory network controlling granulosa cell differentiation/function. Unbiased analyses of the motifs co-enriched with FOXL2 motif in the fetal ovary identified the RUNX motif as one of the most co-enriched motifs. In addition to the RUNX motif, motifs for CTCF, nuclear receptors SF1/LRH-1/ESRRB, and transcription factors TEADs and GATAs were also significantly enriched with FOXL2 consensus motif in FOXL2-bound chromatin regions. For many of these transcription factors, their potential role in gonad differentiation is unknown or limited. For example, the transcription factors of the TEAD family belong to the Hippo pathway, which is involved in the regulation of Sertoli cell gene expression in the fetal gonads60. However, the potential involvement of the hippo pathway in granulosa cell differentiation has not been investigated.
In conclusion, we identified RUNX1 as a transcription factor involved in pre-granulosa cell differentiation/maintenance. RUNX1 first delineates the supporting cell lineage and then becomes pre-granulosa cell-specific during gonad development. RUNX1 plays redundant roles with FOXL2 through binding of common chromatin regions and control of common sets of genes to maintain pre-granulosa cell identity in the fetal ovary. Our findings provide insights into the genomic control of granulosa cell differentiation and pave the way for the identification of transcription factors and cis-signatures contributing to the fate determination of granulosa cells and the consequent formation of a functional ovary.
Methods
Mouse models
Tg(Runx1-EGFP) reporter mouse was purchased from MMRRC (MMRRC_010771-UCD) and CD-1 mice were purchased from Charles River (stock number 022). Runx1+/− (B6.129S-Runx1tm1Spe/J) and Runx1f/f (B6.129P2-Runx1tm1Tani/J) mice were purchased from the Jackson Laboratory (stock numbers 005669 and 008772, respectively). Sf1-CreTg/Tg mice34 (B6D2-Tg(Nr5a1-cre)2Klp) were provided by late Dr. Keith Parker and Foxl2+/− mice61 (B6;129-Foxl2 < tm1Gpil > ) were provided by Dr. David Schlessinger (National Institute on Aging), respectively. Runx1 KO mice (Sf1Cre+/Tg; Runx1f/−) were generated by crossing Runx1f/f females with Sf1-Cre+/Tg; Runx1+/− males. Controls were Sf1-Cre+/+; Runx1+/f littermates. Runx1/Foxl2 DKO mice (Sf1Cre+/Tg; Runx1f/−; Foxl2−/−) were generated by crossing Runx1f/f; Foxl2+/− females with Sf1Cre+/Tg; Runx1+/−; Foxl2+/− males. This cross also generated the single KOs for Runx1 (Sf1Cre+/Tg; Runx1f/−; Foxl2+/+) and Foxl2 (Sf1Cre+/+; Runx1f/+; Foxl2−/−), and control littermates (Sf1-Cre+/+; Runx1+/f; Foxl2+/+). Time-mating was set up by housing female mice with male mice overnight and the females were checked for the presence of vaginal plug the next morning. The day when the vaginal plug was detected was considered embryonic day or E0.5. All experiments were performed on at least four animals for each genotype. All animal procedures were approved by the National Institutes of Health Animals Care and Use Committee, and were performed in accordance with an approved National Institute of Environmental Health Sciences animal study proposal.
Immunofluorescences
For the Tg(Runx1-EGFP) mice, gonads were collected and fixed in 4% paraformaldehyde for 1–2 h at room temperature. Immunofluorescence experiments were performed on whole gonads at E11.5 and E12.5, and on sections for E14.5, E15.5, E16.5, and P0 (birth) gonads. The EGFP was detected in whole-mount gonads by direct fluorescent imaging and an anti-GFP antibody was used for immunofluorescences on sections. For the different KO models, gonads were fixed in 4% paraformaldehyde overnight at 4 °C and immunofluorescence experiments were performed on paraffin sections of E14.5, E15.5, and P0 gonads as previously described62. Briefly, the slides were rehydrated and citrate-based antigen retrieval was performed. The samples (slides or whole-mount samples) were blocked in blocking buffer (5% donkey serum/0.1% Triton X-100 in phosphate-buffered saline (PBS)) for 1 h at room temperature. The samples were then incubated overnight at 4 °C in the primary antibodies diluted in blocking buffer. The next day, the samples were washed three times in 0.1% Triton X-100 in PBS and were incubated for 1 h at room temperature in the secondary antibodies diluted in blocking buffer. The samples were then washed and counterstained with DAPI (4′,6-diamidino-2-phenylindole). The antibodies used in this study are listed in Supplementary Table 1. Whole gonads and sections were imaged under a Leica DMI4000 confocal microscope. For all immunofluorescence experiments, at least three independent biological replicates were analyzed and the image presented in the figures were representative of all replicates.
Real-time PCR analysis in the mouse
For the time-course kinetics of Runx1 expression, fetal gonads from CD-1 embryos at embryonic day E11.5, E12.5, E13.5, E14.5, E16.5, E18.5, and postnatal day P3 were separated from the mesonephros and snap-frozen. For each stage, three biological replicates were collected, with six gonads/replicate for the E11.5 stage and three gonads/replicate for the other stages. For Runx1 KO analysis, control and KO ovaries were collected at E14.5 (n = 4 biological replicates/genotype). For Runx1/Foxl2 DKO analysis, control, Runx1, and Foxl2 single and DKO ovaries were collected at E15.5 (n = 4/genotype) and P0 (n = 5/genotype). For all experiments, total RNA was isolated for each replicate using PicoPure RNA isolation kit (Arcturus, Mountain View, CA). RNA quality and concentration were determined using the Nanodrop 2000c and 300–400 ng of RNA was used for cDNA synthesis with the Superscript II cDNA synthesis system (Invitrogen Corp., Carlsbad, CA). Gene expression was analyzed by real-time PCR using Bio-Rad CFX96TM Real-Time PCR Detection system. Gene expression was normalized to Gapdh. The Taqman probes and primers used to detect transcript expression are listed in Supplementary Tables 2 and 3. Data were analyzed using Prism GraphPad Software by unpaired Student’s t-test or by ANOVA p < 0.05. Values are presented as mean ± SEM.
Runx1 expression in other species
For the rainbow trout, Runx1 expression during gonadal development was assessed by quantitative PCR63. Species-specific primers used are listed in Supplementary Table 3. For the red-eared slider turtle, Runx1 expression during gonadal development was assessed at Female-Promoting Temperature of 31 °C and at Male-Promoting Temperature of 26 °C by RNA-sequencing (RNA-seq)64. For the goat, Runx1 expression during gonadal development was assessed by quantitative PCR and two to three biological replicates were used for each stage of development65. Values are presented as mean ± SD. All goat-handling procedures were conducted in compliance with the guidelines on the Care and Use of Agricultural Animals in Agricultural Research and Teaching in France (Authorization number 91–649 for the Principal Investigator, and national authorizations for all investigators. Approval from the Ethics Committee: 12/045). For the human, Runx1 expression during gonadal development was assessed by RNA-seq. Human fetuses (6–12 GW) were obtained from legally induced normally progressing terminations of pregnancy performed in Rennes University Hospital in France. Tissues were collected with women’s written consent, in accordance with the legal procedure agreed by the National agency for biomedical research (authorization #PFS09-011; Agence de la Biomédecine) and the approval of the Local ethics committee of Rennes Hospital in France (advice # 11–48).
Microarray analysis
Gene expression analysis of control, Runx1 KO, Foxl2 KO, and Runx1/Foxl2 DKO ovaries was conducted using Affymetrix Mouse Genome 430 2.0 GeneChip® arrays (Affymetrix, Santa Clara, CA) on four biological replicates (one P0 gonad per replicate) for each genotype. Fifty nanograms of total RNA were amplified and labeled as directed in the WT-Ovation Pico RNA Amplification System and Encore Biotin Module protocols. Amplified biotin-aRNA (4.6 μg) was fragmented and hybridized to each array for 18 h at 45 °C in a rotating hybridization. Array slides were stained with streptavidin/phycoerythrin utilizing a double-antibody staining procedure and then washed for antibody amplification according to the GeneChip Hybridization, Wash and Stain Kit, and user manual following protocol FS450-0004. Arrays were scanned in an Affymetrix Scanner 3000 and data were obtained using the GeneChip® Command Console Software (AGCC; Version 3.2) and Expression Console (Version 1.2). Microarray data have been deposited in GEO under accession code GSE129038. Gene expression analyses were conducted with Partek software (St. Louis, Missouri) using a one-way ANOVA comparing the Robust Multichip Average (RMA)-normalized log2 intensities. A full dataset Excel file containing the normalized log2 intensity of all genes for each genotype and a graphic view of their expression is provided in Supplementary Data 2. In order to identify differentially expressed genes, ANOVA was used to determine whether there was a statistical difference between the means of groups and the gene lists were filtered with p < 0.05 and fold-change cutoff of 1.5. The heat map was created comparing the genes that were significantly different between control and Runx1/Foxl2 DKO ovaries. Venn diagrams were generated in Partek by comparing gene symbols between the lists of genes differentially expressed.
ChIP-seq assays and analysis
Ovaries from E14.5 CD-1 embryos were separated from the mesonephros, snap-frozen, and stored at −80 °C. RUNX1 ChIP-seq experiments and analyses in E14.5 ovaries were performed using the same protocol than FOXL2 ChIP-seq42. Two independent ChIP-seq experiments were performed by Active Motif, Inc., using 20–30 μg of sheared chromatin from pooled embryonic ovaries (n = 100–120 ovaries/ChIP) and 10 μl of RUNX1 antibody66 (provided by Drs Yoram Groner and Ditsa Levanon, the Weizmann Institute of Science, Israel). ChIP-seq libraries were sequenced as single-end 75-mers by Illumina NextSeq 500, then filtered to retain only reads with average base quality score > 20. Reads were mapped against the mouse mm10 reference genome using Bowtie v1.2 with parameter “-m 1” to collect only uniquely mapped hits. Duplicate mapped reads were removed using Picard tools MarkDuplicates.jar (v1.110). The number of uniquely mapped non-duplicate reads for each biological replicate was 8,932,674 and 15,036,698. After merging the replicate datasets, binding regions were identified by peak calling using HOMER v.4.967 with a false discovery rate < 1e − 5. Called peaks were subsequently re-defined as 300mers centered on the called peak midpoints and filtered for a fourfold enrichment over input and over local signal. Genomic distribution of RUNX1-bound regions was determined based on Refseq gene models as downloaded from the UCSC Genome Browser as of 9 August 2017. Enriched motifs were identified using HOMER findMotifsGenome.pl de novo motif analysis with parameter “-size given”. For RUNX1 and FOXL2 ChIP-seq comparisons, binding peaks that had at least 1 bp in common were considered overlapping. Peaks were assigned to the nearest gene based on RefSeq. Gene lists were analyzed for enrichment using the online tool EnrichR68. The ChIP-seq data are available in the ReproGenomics Viewer (https://rgv.genouest.org)69,70 and Gene Expression Omnibus (GSE128767; http://www.ncbi.nlm.nih.gov/geo/).
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Supplementary information
Acknowledgements
We thank late Keith Parker (UT Southwestern Medical Center, USA) for the Sf1-Cre mice, David Schlessinger (National Institute on Aging, USA) for the Foxl2+/− mice, Yoram Groner and Ditsa Levanon (The Weizmann Institute of Science, Israel) for the RUNX1 antibody for ChIP, Susan Brenner-Morton and late Thomas Jessell (Columbia University, USA) for the RUNX1 antibody for immunofluorescences, David Zarkower (University of Minnesota, USA) for the DMRT1 antibody, Ken Morohashi (Kyushu University, Japan) for the SF1 and SOX9 antibodies, and Dagmar Wilhelm (University of Melbourne, Australia) for the SRY antibody. We are grateful to the NIEHS Molecular Genomics Core, Integrative Bioinformatics Support Group, Comparative Medicine Branch, and Cellular and Molecular Pathology Branch for their services. This work was supported in part by the Intramural Research Program (ES102965) of the NIH, National Institute of Environmental Health Sciences in the U.S. The goat study was supported by Agence Nationale de la Recherche in France (ANR-16-CE14-0020). The human fetal gonads study was supported by l’Institut national de la santé et de la recherche médicale (Inserm), l’Université de Rennes 1 and l’Ecole des hautes études en santé publique (EHESP) in France. The study of red-eared slider turtle was supported by a grant (IOS-1256675) from National Science Foundation in the US.
Authors contributions
B.N. performed the experiments in the mouse. B.N. and H.H.-C.Y. designed the study, analyzed data, and wrote the paper. S.A.G performed bioinformatic analyses. F.C and E.L. analyzed RUNX1 expression in human fetal gonads. M.P. and E.P. analyzed RUNX1 expression in the goat. E.D.-D.-P. and Y.G. analyzed runx1 expression in rainbow trout. B.C. analyzed Runx1 expression in the red-eared slider turtle. S.A.G., E.L., F.C., M.P., E.P., E.D.-D.-P., Y.G., B.C. and H.H.-C.Y. edited the paper.
Data availability
The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information files or from the corresponding author upon reasonable request. Microarray and ChIP-seq data generated in this study have been deposited in the GEO database under accession codes GSE129038 and GSE128767, respectively. The ChIP-seq data are available in the ReproGenomics Viewer (https://rgv.genouest.org). Raw data underlying all reported mean values in graphs are provided in the Source Data File. All other relevant data supporting the key findings of this study are available in the Supplementary Information files. The source data underlying Figs. 1, 4d-e, 4h, 7d-g, 7i-j, and 9b-d are provided as a Source Data file.
Competing interests
The authors declare no competing interests.
Footnotes
Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Supplementary information is available for this paper at 10.1038/s41467-019-13060-1.
References
- 1.Capel B. Vertebrate sex determination: evolutionary plasticity of a fundamental switch. Genetics. 2017;18:675–689. doi: 10.1038/nrg.2017.60. [DOI] [PubMed] [Google Scholar]
- 2.Chassot AA, et al. Activation of beta-catenin signaling by Rspo1 controls differentiation of the mammalian ovary. Hum. Mol. Genet. 2008;17:1264–1277. doi: 10.1093/hmg/ddn016. [DOI] [PubMed] [Google Scholar]
- 3.Tomizuka K, et al. R-spondin1 plays an essential role in ovarian development through positively regulating Wnt-4 signaling. Hum. Mol. Genet. 2008;17:1278–1291. doi: 10.1093/hmg/ddn036. [DOI] [PubMed] [Google Scholar]
- 4.Vainio S, Heikkila M, Kispert A, Chin N, McMahon AP. Female development in mammals is regulated by Wnt-4 signalling. Nature. 1999;397:405–409. doi: 10.1038/17068. [DOI] [PubMed] [Google Scholar]
- 5.Harris A, et al. ZNRF3 functions in mammalian sex determination by inhibiting canonical WNT signaling. Proc. Natl Acad. Sci. USA. 2018;115:5474–5479. doi: 10.1073/pnas.1801223115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ottolenghi C, et al. Foxl2 is required for commitment to ovary differentiation. Hum. Mol. Genet. 2005;14:2053–2062. doi: 10.1093/hmg/ddi210. [DOI] [PubMed] [Google Scholar]
- 7.Schmidt D, et al. The murine winged-helix transcription factor Foxl2 is required for granulosa cell differentiation and ovary maintenance. Development. 2004;131:933–942. doi: 10.1242/dev.00969. [DOI] [PubMed] [Google Scholar]
- 8.Uhlenhaut NH, et al. Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation. Cell. 2009;139:1130–1142. doi: 10.1016/j.cell.2009.11.021. [DOI] [PubMed] [Google Scholar]
- 9.Auguste A, et al. Loss of R-spondin1 and Foxl2 amplifies female-to-male sex reversal in XX mice. Sex. Dev. 2011;5:304–317. doi: 10.1159/000334517. [DOI] [PubMed] [Google Scholar]
- 10.Ottolenghi C, et al. Loss of Wnt4 and Foxl2 leads to female-to-male sex reversal extending to germ cells. Hum. Mol. Genet. 2007;16:2795–2804. doi: 10.1093/hmg/ddm235. [DOI] [PubMed] [Google Scholar]
- 11.Herpin A, Schartl M. Plasticity of gene-regulatory networks controlling sex determination: of masters, slaves, usual suspects, newcomers, and usurpators. EMBO Rep. 2015;16:1260–1274. doi: 10.15252/embr.201540667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Crisponi L, et al. The putative forkhead transcription factor FOXL2 is mutated in blepharophimosis/ptosis/epicanthus inversus syndrome. Nat. Genet. 2001;27:159–166. doi: 10.1038/84781. [DOI] [PubMed] [Google Scholar]
- 13.Boulanger L, et al. FOXL2 is a female sex-determining gene in the goat. Curr. Biol. 2014;24:404–408. doi: 10.1016/j.cub.2013.12.039. [DOI] [PubMed] [Google Scholar]
- 14.Li M, et al. Efficient and heritable gene targeting in tilapia by CRISPR/Cas9. Genetics. 2014;197:591–599. doi: 10.1534/genetics.114.163667. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bertho S, et al. The unusual rainbow trout sex determination gene hijacked the canonical vertebrate gonadal differentiation pathway. Proc. Natl Acad. Sci. USA. 2018;115:12781–12786. doi: 10.1073/pnas.1803826115. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Raymond CS, et al. Evidence for evolutionary conservation of sex-determining genes. Nature. 1998;391:691–695. doi: 10.1038/35618. [DOI] [PubMed] [Google Scholar]
- 17.Matsuda M, et al. DMY is a Y-specific DM-domain gene required for male development in the medaka fish. Nature. 2002;417:559–563. doi: 10.1038/nature751. [DOI] [PubMed] [Google Scholar]
- 18.Nanda I, et al. A duplicated copy of DMRT1 in the sex-determining region of the Y chromosome of the medaka, Oryzias latipes. Proc. Natl Acad. Sci. USA. 2002;99:11778–11783. doi: 10.1073/pnas.182314699. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Matson CK, et al. DMRT1 prevents female reprogramming in the postnatal mammalian testis. Nature. 2011;476:101–104. doi: 10.1038/nature10239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Mello MP, et al. Novel DMRT1 3’UTR+11insT mutation associated to XY partial gonadal dysgenesis. Arq. Bras. Endocrinol. Metab. 2010;54:749–753. doi: 10.1590/S0004-27302010000800015. [DOI] [PubMed] [Google Scholar]
- 21.Raymond CS, Murphy MW, O’Sullivan MG, Bardwell VJ, Zarkower D. Dmrt1, a gene related to worm and fly sexual regulators, is required for mammalian testis differentiation. Genes Dev. 2000;14:2587–2595. doi: 10.1101/gad.834100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Duffy JB, Gergen JP. The Drosophila segmentation gene runt acts as a position-specific numerator element necessary for the uniform expression of the sex-determining gene sex-lethal. Genes Dev. 1991;5:2176–2187. doi: 10.1101/gad.5.12a.2176. [DOI] [PubMed] [Google Scholar]
- 23.Kramer SG, Jinks TM, Schedl P, Gergen JP. Direct activation of sex-lethal transcription by the Drosophila Runt protein. Development. 1999;126:191–200. doi: 10.1242/dev.126.1.191. [DOI] [PubMed] [Google Scholar]
- 24.Nef S, et al. Gene expression during sex determination reveals a robust female genetic program at the onset of ovarian development. Dev. Biol. 2005;287:361–377. doi: 10.1016/j.ydbio.2005.09.008. [DOI] [PubMed] [Google Scholar]
- 25.Groner Y., et al. RUNX Proteins in Development and Cancer (Springer, 2017).
- 26.Chuang LS, Ito K, Ito Y. RUNX family: regulation and diversification of roles through interacting proteins. Int J. Cancer. 2013;132:1260–1271. doi: 10.1002/ijc.27964. [DOI] [PubMed] [Google Scholar]
- 27.Stevant I, et al. Dissecting cell lineage specification and sex fate determination in gonadal somatic cells using single-cell transcriptomics. Cell Rep. 2019;26:3272–3283.e3273. doi: 10.1016/j.celrep.2019.02.069. [DOI] [PubMed] [Google Scholar]
- 28.Gong S, et al. A gene expression atlas of the central nervous system based on bacterial artificial chromosomes. Nature. 2003;425:917–925. doi: 10.1038/nature02033. [DOI] [PubMed] [Google Scholar]
- 29.Koopman P, Gubbay J, Vivian N, Goodfellow P, Lovell-Badge R. Male development of chromosomally female mice transgenic for Sry. Nature. 1991;351:117–121. doi: 10.1038/351117a0. [DOI] [PubMed] [Google Scholar]
- 30.Mork L., et al. Temporal differences in granulosa cell specification in the ovary reflect distinct follicle fates in mice. Biol. Reprod.86, 37 (2012). [DOI] [PMC free article] [PubMed]
- 31.Rastetter RH, et al. Marker genes identify three somatic cell types in the fetal mouse ovary. Dev. Biol. 2014;394:242–252. doi: 10.1016/j.ydbio.2014.08.013. [DOI] [PubMed] [Google Scholar]
- 32.Okuda T, vanDeursen J, Hiebert SW, Grosveld G, Downing JR. AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis. Cell. 1996;84:321–330. doi: 10.1016/S0092-8674(00)80986-1. [DOI] [PubMed] [Google Scholar]
- 33.Wang Q, et al. Disruption of the Cbfa2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis. Proc. Natl Acad. Sci. USA. 1996;93:3444–3449. doi: 10.1073/pnas.93.8.3444. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bingham NC, Verma-Kurvari S, Parada LF, Parker KL. Development of a steroidogenic factor 1/Cre transgenic mouse line. Genesis. 2006;44:419–424. doi: 10.1002/dvg.20231. [DOI] [PubMed] [Google Scholar]
- 35.Lei N, et al. Sex-specific differences in mouse DMRT1 expression are both cell type- and stage-dependent during gonad development. Biol. Reprod. 2007;77:466–475. doi: 10.1095/biolreprod.106.058784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Chaboissier MC, et al. Functional analysis of Sox8 and Sox9 during sex determination in the mouse. Development. 2004;131:1891–1901. doi: 10.1242/dev.01087. [DOI] [PubMed] [Google Scholar]
- 37.Vidal VP, Chaboissier MC, de Rooij DG, Schedl A. Sox9 induces testis development in XX transgenic mice. Nat. Genet. 2001;28:216–217. doi: 10.1038/90046. [DOI] [PubMed] [Google Scholar]
- 38.Jorgez CJ, Klysik M, Jamin SP, Behringer RR, Matzuk MM. Granulosa cell-specific inactivation of follistatin causes female fertility defects. Mol. Endocrinol. 2004;18:953–967. doi: 10.1210/me.2003-0301. [DOI] [PubMed] [Google Scholar]
- 39.Britt KL, et al. The ovarian phenotype of the aromatase knockout (ArKO) mouse. J. Steroid Biochem. Mol. Biol. 2001;79:181–185. doi: 10.1016/S0960-0760(01)00158-3. [DOI] [PubMed] [Google Scholar]
- 40.Jameson SA, et al. Temporal transcriptional profiling of somatic and germ cells reveals biased lineage priming of sexual fate in the fetal mouse gonad. PLoS Genet. 2012;8:e1002575. doi: 10.1371/journal.pgen.1002575. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Colvin JS, Green RP, Schmahl J, Capel B, Ornitz DM. Male-to-female sex reversal in mice lacking fibroblast growth factor 9. Cell. 2001;104:875–889. doi: 10.1016/S0092-8674(01)00284-7. [DOI] [PubMed] [Google Scholar]
- 42.Nicol B, et al. Genome-wide identification of FOXL2 binding and characterization of FOXL2 feminizing action in the fetal gonads. Hum. Mol. Genet. 2018;27:4273–4287. doi: 10.1093/hmg/ddy312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Ong CT, Corces VG. CTCF: an architectural protein bridging genome topology and function. Nat. Rev. Genet. 2014;15:234–246. doi: 10.1038/nrg3663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Park M, et al. FOXL2 interacts with steroidogenic factor-1 (SF-1) and represses SF-1-induced CYP17 transcription in granulosa cells. Mol. Endocrinol. 2010;24:1024–1036. doi: 10.1210/me.2009-0375. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Yang WH, Gutierrez NM, Wang LZ, Ellsworth BS, Wang CM. Synergistic activation of the Mc2r promoter by FOXL2 and NR5A1 in mice. Biol. Reprod. 2010;83:842–851. doi: 10.1095/biolreprod.110.085621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Manuylov NL, Smagulova FO, Leach L, Tevosian SG. Ovarian development in mice requires the GATA4-FOG2 transcription complex. Development. 2008;135:3731–3743. doi: 10.1242/dev.024653. [DOI] [PubMed] [Google Scholar]
- 47.Meyers S, Downing JR, Hiebert SW. Identification of AML-1 and the (8;21) translocation protein (AML-1/ETO) as sequence-specific DNA-binding proteins: the runt homology domain is required for DNA binding and protein-protein interactions. Mol. Cell Biol. 1993;13:6336–6345. doi: 10.1128/MCB.13.10.6336. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Blount AL, et al. FoxL2 and Smad3 coordinately regulate follistatin gene transcription. J. Biol. Chem. 2009;284:7631–7645. doi: 10.1074/jbc.M806676200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Barrionuevo F, et al. Homozygous inactivation of Sox9 causes complete XY sex reversal in mice. Biol. Reprod. 2006;74:195–201. doi: 10.1095/biolreprod.105.045930. [DOI] [PubMed] [Google Scholar]
- 50.Wildey GM, Howe PH. Runx1 is a co-activator with FOXO3 to mediate transforming growth factor beta (TGFbeta)-induced Bim transcription in hepatic cells. J. Biol. Chem. 2009;284:20227–20239. doi: 10.1074/jbc.M109.027201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Wang L, Brugge JS, Janes KA. Intersection of FOXO- and RUNX1-mediated gene expression programs in single breast epithelial cells during morphogenesis and tumor progression. Proc. Natl Acad. Sci. USA. 2011;108:E803–E812. doi: 10.1073/pnas.1103423108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Ono M, et al. Foxp3 controls regulatory T-cell function by interacting with AML1/Runx1. Nature. 2007;446:685–689. doi: 10.1038/nature05673. [DOI] [PubMed] [Google Scholar]
- 53.Recouvreux MS, et al. RUNX1 and FOXP3 interplay regulates expression of breast cancer related genes. Oncotarget. 2016;7:6552–6565. doi: 10.18632/oncotarget.6771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Lichtinger M, et al. RUNX1 reshapes the epigenetic landscape at the onset of haematopoiesis. EMBO J. 2012;31:4318–4333. doi: 10.1038/emboj.2012.275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Munger SC, Natarajan A, Looger LL, Ohler U, Capel B. Fine time course expression analysis identifies cascades of activation and repression and maps a putative regulator of mammalian sex determination. PLoS Genet. 2013;9:e1003630. doi: 10.1371/journal.pgen.1003630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Stevant I, et al. Deciphering cell lineage specification during male sex determination with single-cell RNA sequencing. Cell Rep. 2018;22:1589–1599. doi: 10.1016/j.celrep.2018.01.043. [DOI] [PubMed] [Google Scholar]
- 57.Nottingham WT, et al. Runx1-mediated hematopoietic stem-cell emergence is controlled by a Gata/Ets/SCL-regulated enhancer. Blood. 2007;110:4188–4197. doi: 10.1182/blood-2007-07-100883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Albrecht KH, Eicher EM. Evidence that Sry is expressed in pre-Sertoli cells and Sertoli and granulosa cells have a common precursor. Dev. Biol. 2001;240:92–107. doi: 10.1006/dbio.2001.0438. [DOI] [PubMed] [Google Scholar]
- 59.Rahmoun M, et al. In mammalian foetal testes, SOX9 regulates expression of its target genes by binding to genomic regions with conserved signatures. Nucleic Acids Res. 2017;45:7191–7211. doi: 10.1093/nar/gkx328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.Levasseur A, Paquet M, Boerboom D, Boyer A. Yes-associated protein and WW-containing transcription regulator 1 regulate the expression of sex-determining genes in Sertoli cells, but their inactivation does not cause sex reversal. Biol. Reprod. 2017;97:162–175. doi: 10.1093/biolre/iox057. [DOI] [PubMed] [Google Scholar]
- 61.Uda M, et al. Foxl2 disruption causes mouse ovarian failure by pervasive blockage of follicle development. Hum. Mol. Genet. 2004;13:1171–1181. doi: 10.1093/hmg/ddh124. [DOI] [PubMed] [Google Scholar]
- 62.Nicol B, Yao HH. Gonadal identity in the absence of pro-testis factor SOX9 and pro-ovary factor beta-catenin in mice. Biol. Reprod. 2015;93:35. doi: 10.1095/biolreprod.115.131276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.Marivin E, et al. Sex hormone-binding globulins characterization and gonadal gene expression during sex differentiation in the rainbow trout, Oncorhynchus mykiss. Mol. Reprod. Dev. 2014;81:757–765. doi: 10.1002/mrd.22344. [DOI] [PubMed] [Google Scholar]
- 64.Czerwinski M, Natarajan A, Barske L, Looger LL, Capel B. A timecourse analysis of systemic and gonadal effects of temperature on sexual development of the red-eared slider turtle Trachemys scripta elegans. Dev. Biol. 2016;420:166–177. doi: 10.1016/j.ydbio.2016.09.018. [DOI] [PubMed] [Google Scholar]
- 65.Elzaiat M, et al. High-throughput sequencing analyses of XX genital ridges lacking FOXL2 reveal DMRT1 up-regulation before SOX9 expression during the sex-reversal process in goats. Biol. Reprod. 2014;91:153. doi: 10.1095/biolreprod.114.122796. [DOI] [PubMed] [Google Scholar]
- 66.Umansky KB, et al. Runx1 transcription factor is required for myoblasts proliferation during muscle regeneration. PLoS Genet. 2015;11:e1005457. doi: 10.1371/journal.pgen.1005457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 67.Heinz S, et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell. 2010;38:576–589. doi: 10.1016/j.molcel.2010.05.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 68.Chen EY, et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics. 2013;14:128. doi: 10.1186/1471-2105-14-128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Darde T. A., et al. The ReproGenomics Viewer: a multi-omics and cross-species resource compatible with single-cell studies for the reproductive science community. Bioinformatics31, 3133–3139 (2019). [DOI] [PubMed]
- 70.Darde TA, et al. The ReproGenomics Viewer: an integrative cross-species toolbox for the reproductive science community. Nucleic Acids Res. 2015;43:W109–W116. doi: 10.1093/nar/gkv345. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The authors declare that all data supporting the findings of this study are available within the article and its Supplementary Information files or from the corresponding author upon reasonable request. Microarray and ChIP-seq data generated in this study have been deposited in the GEO database under accession codes GSE129038 and GSE128767, respectively. The ChIP-seq data are available in the ReproGenomics Viewer (https://rgv.genouest.org). Raw data underlying all reported mean values in graphs are provided in the Source Data File. All other relevant data supporting the key findings of this study are available in the Supplementary Information files. The source data underlying Figs. 1, 4d-e, 4h, 7d-g, 7i-j, and 9b-d are provided as a Source Data file.