The gene regulatory landscape driving mouse gonadal supporting cell differentiation

Abstract

Gonadal sex determination relies on tipping a delicate balance involving the activation and repression of several transcription factors and signaling pathways. This is likely mediated by numerous noncoding regulatory elements that shape sex-specific transcriptomic programs. To explore the dynamics of these in detail, we performed paired time series of transcriptomic and chromatin accessibility assays on pre-granulosa and Sertoli cells throughout their development in the embryo, making use of new and existing mouse reporter lines. Regulatory elements were associated with their putative target genes by linkage analysis, and this was complemented and verified experimentally using promoter capture Hi-C. We identified the transcription factor motifs enriched in these regulatory elements along with their occupancy, pinpointing LHX9/EMX2 as potentially critical regulators of ovarian development. Variations in the DNA sequence of these regulatory elements are likely to be responsible for many of the unexplained cases of individuals with differences of sex development.

Fig. 1.. Generation of a new mouse line allowing pre-granulosa cell purification for multi-omics analysis.

(A) Enh8 was cloned upstream of the mCherry reporter gene controlled by the hsp68 minimal promoter. The obtained vector was injected into zygote mouse embryos by microinjection to obtain the Enh8-mCherry [Tg(Enh8-mCherry)] mouse line. (B) Binocular pictures of dissected XX Enh8-mCherry gonads at E11.5, E12.5, E13.5, and E15.5 in bright-field and fluorescence. go, gonad; ms, mesonephros. Scale bar, 500 μm. (C) Immunofluorescent staining of E15.5 Enh8-mCherry ovary sections. mCherry (transgene) is labeled in red; FOXL2, a marker of pre-granulosa cells is labeled in green; and TRA98, a marker of germ cells is labeled in gray. Some mCherry expressing pre-granulosa cells are highlighted with yellow arrowheads. Scale bar, 100 μm. (D) Gene expression enrichment of ovarian marker genes in Enh8-mCherry sorted cells at E11.5, E12.5, and E13.5 RNA-seq compared to whole-gonad RNA-seq [from Zhao et al. (52)]. Enrichment (log₂ of the ratio between Enh8-mCherry–sorted cells and whole-gonad gene expression) is represented with a blue-to-red gradient, and the expression levels of the Enh8-mCherry–sorted cells are represented as TPM (transcript per million) with the size of the dots. (E) Schematic experimental design. XX gonads from Enh8-mCherry and XY gonads from SOX9^IRES-GFP/+ mouse lines were collected at E11.5, E12.5, E13.5, and E15.5. Pre-granulosa and Sertoli cells were purified by fluorescence-activated cell sorting (FACS) at each embryonic stage, and the sorted cells were subjected to RNA-seq and ATAC-seq to constitute a time-series paired gene expression and chromatin accessibility data collection. (F and G) PCA of the obtained RNA-seq and ATAC-seq data. Samples are colored by sex and embryonic stage. Pre-granulosa samples were circled in yellow, and Sertoli cell samples in green. (H and I) Expression profiles of Runx1, Fst (pre-granulosa specific genes), and Sox9 and Amh (Sertoli specific genes) in purified cells (continuous line) and in whole gonads [from Zhao et al. (52), dashed line] along embryonic stages in both sexes (XX in yellow and XY in green).

Fig. 2.. Characterization of the supporting cell transcriptome during their differentiation.

(A) Number of overexpressed genes in pre-granulosa and Sertoli cells at each embryonic stage compared to the opposite sex. (B and C) Heatmaps representing the expression changes (z score) of the genes dynamically expressed during pre-granulosa and Sertoli cells, respectively. Genes were clustered by expression profiles. The clusters were labeled with letters on the left side of the heatmaps. Twenty-five TFs known to cause a gonadal or infertility phenotype were labeled on the right side of the heatmaps. (D) Venn diagram showing the overlap of the dynamically expressed genes in both sexes with examples of gene names present in each intersection. Expression (TPM) profiles of genes from the different sets are shown as examples.

Fig. 3.. Establishment of sexually dimorphic open chromatin regions in supporting cells during their differentiation.

(A) Number of differentially accessible regions in pre-granulosa and Sertoli cells at each embryonic stage compared to the opposite sex. (B and C) Upset plot representing the distribution of the sex-biased open chromatin regions across embryonic stages in pre-granulosa and Sertoli cells, respectively. Each column represents a specific intersection between sets. Dots connected by lines (bottom) indicate which sets are involved in the intersection. The height of the bars on the y axis (top) quantifies the number of elements in each intersection. (D) Genomic features overlapped by the sex-biased open chromatin regions across embryonic stages in pre-granulosa and Sertoli cells. (E to G) Genomic tracks showing the normalized ATAC-seq signal of sexually dimorphic open chromatin regions (OCRs) around the pre-granulosa factor Rspo1, the Sertoli factor Sox8, and the critical gonadal factor Zfmp2 (also known as Fog2). Non-sex–specific OCRs are highlighted in gray, sex-biased OCRs are marked by an arrowhead on top, pre-granulosa–biased OCRs are highlighted in yellow, and Sertoli-biased in blue. The bar plot on the right-hand side shows the expression level in TPM of the gene of interest. The error bars represent the SD between the replicates. (H and I) TF-binding motif enrichment analysis in pre-granulosa and Sertoli-biased open chromatin regions, respectively. Only TFs found expressed in the supporting cells are shown. Motifs were merged by sequence similarity, and the consensus logo is shown. Known gonadal TFs are highlighted in yellow and blue. Nonsignificant enrichments are colored in gray. n.s., nonsignificant.

Fig. 4.. Dynamics of open chromatin regions in supporting cells during their differentiation.

(A and B) Heatmaps representing the accessibility changes (z score) of the open chromatin regions during pre-granulosa and Sertoli cells, respectively. Regions were clustered by accessibility profiles. The clusters were labeled with letters on the left side of the heatmaps. (C) Venn diagram showing the overlap between pre-granulosa and Sertoli cell dynamic open chromatin regions. (D) Genomic features where the dynamic open chromatin regions are found in goth pre-granulosa and Sertoli cells at each stage. (E to G) Genomic tracks showing the normalized ATAC-seq signal of genomic regions loci containing dynamic open chromatin regions (OCRs) in pre-granulosa and/or in Sertoli cells in the vicinity of known gonadal genes. Non-dynamic OCRs are highlighted in gray, significantly dynamic OCRs are marked by an arrowhead on top, pre-granulosa–dynamic OCRs are highlighted in yellow, and Sertoli-dynamic in blue. The bar plot on the right-hand side shows the expression level in TPM of the gene of interest. The error bars represent the SD between the replicates. (H and I) Heatmap of the TF motifs differentially enrichment between the different open chromatin region clusters from pre-granulosa and Sertoli cells, respectively. Only TFs found expressed in the supporting cells are shown. Motifs were merged by sequence similarity and the consensus logo is shown. Known gonadal TFs are highlighted.

Fig. 5.. Prediction of gene-cis-regulatory region associations.

(A) Schematic representation describing the strategy to link open chromatin regions with their target genes. Open chromatin regions located within a ±500-kb window of a gene’s TSS, whose accessibility correlates with gene expression, are identified as putative enhancers. Conversely, open chromatin regions that anticorrelate with gene expression are linked as putative silencers. The number of positive and negative link, as well as the average number of linked open chromatin regions (OCRs) per gene are indicated. (B to D) Genomic tracks showing the predicted links between open chromatin regions and gene expression. The positive links are represented as red line arcs, and negative links as blue line arcs. The genomic tracks represent normalized ATAC-seq signal of genomic regions loci in pre-granulosa and/or in Sertoli cells. Noticeable genomic loci are indicated with an arrowhead. The bar plot on the right-hand side shows the expression level in TPM of the gene of interest. The error bars represent the SD between the replicates. (E) Experimental design of the PCHi-C experiment. E13.5 pre-granulosa and Sertoli cells have been purified by FACS, fixed, and have been subjected to PCHi-C. Interactions between gene promoters and genomic regions were called using a 5-kb bin resolution. (F and G) Genomic tracks showing the PCHi-C interactions found in either pre-granulosa or Sertoli cells around two sexually dimorphic genes at E13.5. The interactions contain open chromatin regions and overlap with the RNA-ATAC linkage analysis. The dash lines on top of the ATAC-seq signal signify that the scale has been cropped.

Fig. 6.. ATAC TF footprints on sexually dimorphic open chromatin regions.

(A) Differential TF-binding motif enrichment and occupancy (ATAC-seq footprints) in the combined (all developmental stages) sex-biased open chromatin regions when comparing one sex to the other. Only TFs found expressed in the supporting cells are shown. Motifs were merged by sequence similarity and the consensus logo is shown. Known gonadal TFs are highlighted. (B and C) Comparison of the aggregated ATAC-seq footprint signals at E13.5 in both sexes for the top pre-granulosa and Sertoli cell differentially bound TF-binding motifs recognized by EMX2 and LHX9, and DMRT1 and SOX and SRY factors, respectively. The number of bound motifs is indicated. The dashed lines represent the motif location. (D to J) Genomic tracks showing TF footprints of the gonadal factors highlighted in figure (A) and WT1 on different sex-biased accessible loci. Footprints found in any of the studied embryonic stages were aggregated. For concision, we only show the name of the known gonadal TFs and not the full list of TFs able to bind the occupied motifs.