Tandem single-cell epigenetic and transcriptomic analysis reveals spatiotemporally distinct genomic features among testis cell lineages Authors: Mustika Rahmawati, Tia M Hoisington, Blanca Lopez-Biladeau, Kanako Hayashi, Nathan C Law Published: 2026 ABSTRACT SummaryThe mammalian testis develops through coordinated differentiation of germline and somatic lineages. How chromatin maturation and epigenetic states guide testis development is not well understood. To address this, we generated paired single-nucleus chromatin and transcriptomic data from mouse testis cells from embryonic day 14.5 through postnatal day 4. Across lineages, epigenetic maturation differed temporally during development, spatially within the genome, and organizationally with bifurcating or linear trajectories. For instance, germ cell chromatin accessibility specifically at distal intergenic regions linked to genes implicated in spermatogonial stem cell function. Sertoli cells within seminiferous tubules or the rete displayed region-specific chromatin profiles. Among interstitial cells, epigenetic maturation marked canonical lineage drivers and modulated diverging cell fates. Finally, motif enrichment demonstrated cell-type specific activity among transcription factors, and cell-to-cell communication inferences illuminated interconnected networks among lineages. These findings highlight distinct maturations guiding testis development and identify candidate regulators with potential relevance to infertility.Subject areas: Epigenetics, Developmental biology, Transcriptomics Introduction Mammalian testis development is a largely asynchronous process, with diverse cell lineages reaching developmental milestones at distinct points during maturation.1,2 Much of this asynchrony is likely attributed to the breadth of cellular heterogeneity within the developing gonad and the unique differentiation pathways of each cell lineage. In general, these differentiation trajectories are incompletely understood. Furthermore, how testicular cells communicate to coordinate testis maturation is not clear. Yet, the formation of a functional germline and supportive somatic environment is essential to sustain spermatogenesis into adulthood.3,4 Ongoing studies generally point to three broad cell lineages within the developing testis: the germ, steroidogenic, and supporting cell lineages. First, the germ cell lineage arises prior to gonad formation around E6.25 in mice and 2–3 weeks of gestation in humans.5,6,7,8 Primordial germ cells (PGCs) are specified outside the gonad from the proximal epiblast before migrating through the embryo to colonize the forming gonad around E7-10 in mice and 5 weeks of gestation in humans.9,10 Around E10 in mice, the genital ridge (future gonad) begins to form from the thickening of the coelomic epithelium adjacent to the mesonephros which serves as the source of both steroidogenic and supporting cells within the gonad.11,12,13,14 Sex determination temporally overlaps early gonadal development, with expression of the key sex-determining gene Sry initiating around E10.5 in mice and 5–6 weeks gestation in humans.15,16,17 Sry triggers a cascading genetic program that directs the gonad toward a male fate, which includes the differentiation of PGCs into prospermatogonia and somatic cells into male-specific counterparts, including Sertoli, undifferentiated mesenchymal, and vascular cells. As the nascent gonad progresses through sex determination, morphological features become readily more apparent. Germ cells are encapsulated by Sertoli cells, and together these cells form elongating tubules that give rise to seminiferous cords.18 The space between seminiferous cords, the interstitium, is populated with a highly heterogeneous composition of cells. For example, peritubular myoid cells that ultimately line the outer surface of seminiferous cords arise from multipotent progenitors within the interstitium.14,19 Studies suggest that two types of steroid-producing Leydig cells also arise from this population of interstitial progenitors: Fetal Leydig cells form first during fetal development, while adult Leydig cells form postnatally.20,21 Endothelial cells originating from the vascular plexus at the gonad-mesonephros border migrate into the interstitium and ultimately form vascular components that will support postnatal gonadal function.18 Finally, immune cells, including yolk sac-derived macrophages and hematopoietic stem cell-derived monocytes, colonize the interstitial space to support testis morphogenesis and postnatal function.22 Perinatally, the developing testis will continue to grow and mature. Shortly after birth, many of the somatic cell types within the testis are largely specified and will continue to reach maturity while spermatogenesis initiates. During neonatal development within the germ cell lineage, prospermatogonia transition to form spermatogonia, which serve as the reservoir from which continuous sperm production will originate.23,24 Interestingly, studies in mice demonstrated that one subset of prospermatogonia directly forms the first round of postnatal sperm production, while the remaining prospermatogonia form spermatogonial stem cells (SSCs) and other undifferentiated spermatogonia.25,26,27,28,29,30 However, the underlying mechanisms of prospermatogonial fate decisions remain unknown.4,28,31 Sertoli cells provide the initial pulse of retinoic acid that promotes the start of spermatogenesis around postnatal day 3 (P3) in mice.32 Meanwhile, Sertoli cells will continue to proliferate until just prior to puberty.33,34,35 During neonatal development, the fetal Leydig cell population declines and is progressively replaced by adult Leydig cells that arise from a stem cell reservoir.2,36,37 Adult Leydig cells will support androgen production long into adulthood. Finally, peritubular myoid cells secrete factors that support spermatogenesis, including some in an androgen-dependent manner.38,39,40 Also, peritubular myoid cells will ultimately generate the contractile motions necessary to propel sperm toward the rete testis (RT), where all seminiferous tubules (ST) merge. Collectively, the broad developmental events that encapsulate early testis development, from morphogenesis to the onset of spermatogenesis, have been well-characterized. However, tissue development is driven by changes in chromatin architecture that systemically coordinate gene expression and cellular fate. In the developing testis, the chromatin remodeling that underlies tissue maturation is far less understood. Therefore, in the present study, we generated an atlas of paired transcriptomic and chromatin accessibility information at single-nucleus resolution for cell populations within the developing testis with the goal of linking transcriptomic changes indicative of maturation to the underlying chromatin architecture driving cellular identity. Our analyses revealed that the cell lineages within the developing testis experience stark patterns of epigenetic maturation, both temporally and spatially within the genome. Furthermore, these alterations in chromatin architecture are often, but not always, intimately linked with changes in gene expression, including canonical regulators of testis maturation. Finally, we leveraged this data atlas to identify transcription factors associated with developmental transitions within each cell lineage. In addition to conventional means of data sharing, we have also shared the paired transcriptomic and chromatin data on an accessible web-based platform located within the Washington State University Bioinformatics Datahub (https://datahub.vetmed.wsu.edu). Results Single-nucleus Multi-‘omic analysis captures a diversity of testis cell populations To broadly study testis development from the completion of sex determination to the onset of spermatogenesis, we performed tandem single-nucleus RNA-sequencing (snRNA-seq) and single-nucleus assay for transposase-accessible chromatin sequencing (snATAC-seq), referred to as snMultiOme herein. Testes from duplicate male mice were collected at E14, E18, and P4. Before performing snMultiOme, we devised a strategy targeted at enriching the germ cell population represented in the data (Figure 1A). The representation of germ cells is generally low during testis development and comprises approximately 3–10% of the total cell population.27,29,41 To enrich for germ cells, we adapted a magnetic activated cell sorting (MACS) approach using an antibody recognizing the cell surface protein epithelial cell adhesion molecule (EPCAM), which is a similar strategy used in other studies.42,43 However, to maintain representation of all cell types within the testis, we combined EPCAM-enriched cell suspensions with the total testis cell suspension to accomplish the overall goal of germ cell enrichment. Serendipitously, we also enriched for rare populations of support cells during testis development (discussed more later in discussion). The resulting cell suspensions from MACS enrichment were processed to capture both snRNA-seq and snATAC-seq information from within the same individual cell. Single-nucleus, multi-‘omic capture of germ-cell enriched cells during testis development (snMultiOme) (A) Schematic representation of the overall experimental design. (B–D) UMAP dimensional reductions of data from snRNA-seq (A), snATAC-seq (B), or joint snRNA- and snATAC-seq (C). (E) UMAP representation of cell clustering via transcriptomic data. (F) Heatmap representation of the top 200 differentially expressed genes among 22 clusters. (G) Violin plot of expression of Epcam within the 22 clusters. Single-nucleus libraries were sequenced at generous read depths to achieve an average of 171,439 sequencing reads (130,669 peaks) per cell for snATAC-seq and 144,092 sequencing reads (3,313 genes) per cell for snRNA-seq. After quality control and filtering (Figures S1A–S1D), a total of 16,936 cells were captured across duplicate developmental ages (Figures 1B–1D; Table 1). We then used transcriptomic information and unsupervised clustering to divide all cells into 22 clusters, which captured individual cell populations that generally separated by developmental age (Figure 1E). Interestingly, the global distribution of all cells differed slightly between uniform manifold approximation and projections (UMAPs) calculated from snRNA-seq data versus snATAC-seq data. However, clusters generally assembled in similar relative positions regardless of modality (Figures S2A–S1C), suggesting cell-type-specific transcriptomic and chromatin signatures. Table 1Sequencing metrics for each biological replicateReplicateCell NumbersnRNA-seqsnATAC-seqtotal read pairsreads per cellstotal read pairsreads per cellE14.5 - replicate 13860791168595204965749525302194177E14.5 - replicate 23736590699283158110604876237161904E18.5 - replicate 127442244811878180824333421088678E18.5 - replicate 24446511752312115103781247405175719P4 - replicate 1820130186656158764211250862257623P4 - replicate 24971724783273145802748255860150524 Sequencing metrics for each biological replicate To infer cell types within these 22 clusters, we utilized a combination of canonical marker gene expression and differentially expressed genes (DEGs) to annotate major cell lineages within the testis (Figures 1F and S2D; Table S1). We were able to resolve germ, Sertoli, Leydig, peritubular myoid, immune, and endothelial cells, as well as developmental lineages of interstitial and myoid progenitors, support-like cells, and cells of RT. While some clusters, such as clusters 02 and 11, were clear instances of co-capture among multiple cell types based on the dual expression of canonical marker genes (which we have experienced previously44), most clusters were highly enriched for individual cellular populations. Surprisingly, based on gene expression data, our EPCAM enrichment strategy likely enriched for germ cells as well as support-like cells and cells of the RT (Figures 1G and S2E). Germ cell maturation includes distinct modulation of putative enhancers To begin evaluating germ cells captured within our snMultiOme atlas, we isolated clusters 9 and 10 from the analysis above (Figures 2A–2C), which encapsulated populations enriched for canonical germline markers. Unfortunately, following quality control filtering, only germ cells from E14 and P4 were captured in our dataset. However, by comparing these flanking ages within our dataset and evaluating differentially accessible regions (DARs) between E14 and P4, we discovered a widespread shift in chromatin accessibility (Figure 2D). Interestingly, of the 70,858 DARs in total, 77.41% (54,850) were more accessible at P4 compared to E14, suggesting numerous regions of increased accessibility through development. DARs were primarily located at promoter and distal intergenic regions (Figure 2E). Remarkably, 95.25% of DARs (or 67,494 out of 70,858) were correlated to the expression of one or more genes, and over 80% of DEGs between E14 and P4 were linked to DARs of chromatin (Figure 2F), suggesting a strong link between differential gene expression and chromatin architecture. Gene ontology (GO) analysis highlighted that many DEGs between E14 and P4 linked to changes in chromatin architecture were enriched within male germline-related genes, including those implicated in stem cell development, reproductive development, and meiosis (Figure 2G). Unique genomic features of germ cell maturation (A–C) From the aggregate dataset, germ cells were isolated, data reassembled, and UMAP representations generated from snRNA-seq (A), snATAC-seq (B), and joint snRNA- and snATAC-seq (C) data. (D) Genome-wide map of differentially accessible regions (DARs) accessible at E14 and P4. (E) Distribution of genomic locations for DARs. (F) Total number of DARs, number of DARs linked to the expression of at least one gene, and number of DARs linked to the expression of a differentially expressed gene (DEG) between E14 and P4 in germ cells. (G) Select gene ontology (GO) terms for DEGs between E14 and P4 that are linked to DARs. (H–K) Coverage plot representation of gene accessibility (snATAC-seq) and expression (snRNA-seq) for genes necessary for postnatal stem cell function, including Gfra1 (H), Id4 (I), Ret (J), and Zbtb16 (K). Gray regions denote DARs, linkages between peaks and gene expression are colored based on the strength of correlation, and asterisks (∗) denote accessible regions strongly linked with gene expression. (L and M) Immunofluorescence staining of testis cross-sections at E14, E18, and P4 for GFRA1 (L) and ZBTB16/PLZF (M) alongside the germ cell marker DDX4. Scale bars, 10 μm. Images are representative of n = 3 images taken at each age. (N) Relative enrichment for transcription factor motifs (or motif “activity”) and expression of transcription factors with the greatest fold change in enrichment within DARs from E14 to P4. Data for motif activity are presented as relative means. From E14 to P4, one canonical developmental event within the male germline is the formation of the SSC population.4 A hallmark of this development is the onset of genes that are necessary for SSC functions, such as self-renewal or production of daughter cells competent to form sperm. Upon closer examination of genes necessary for SSC function that are differentially expressed from E14 to P4, including Foxc2, Gfra1, Id4, Ret, Pou3f1, and Zbtb1645,46,47,48,49,50,51,52(Figures 2H–2K, S3A, and S3B), we discovered that promoter regions were not differentially accessible. Strikingly, however, distal regions that may include enhancers were both differentially accessible and strongly linked with expression (denoted with asterisks). As a positive control of chromatin remodeling, we verified that Nanog is both downregulated and the promoter is epigenetically repressed (Figure S3C). To validate differential expression from E14 to P4, we performed immunofluorescence imaging of testis cross-sections stained with antibodies recognizing GFRA1 and PLZF (encoded by the gene Zbtb16) alongside the pan-germ cell marker DDX4 (Figures 2L and 2M; see Table 2 for antibody conditions). Notably, PLZF was not observed at E18, while some germ cells expressed GFRA1 at E18, consistent with previous studies.26 Table 2Information pertaining to antibodies used in this studyAntibodyVendor and Cat. #ApplicationConcentrationRat anti-EPCAM (CD326)Thermo/eBioscience, 14-5791-81MACS3 μgGoat anti-GFRa1R&D systems, AF560IF1:250Rabbit anti-DDX4Abcam, ab13840IF1:200Goat anti-PLZFR&D systems, AF2944IF1:500Rabbit anti-SOX9MilliporeSigma, AB5535IF1:1000Mouse anti-GATA4Santa Cruz - SC-25310IF1:200Rat anti-CDH1 - Alexa Fluor 647 ConjugatedBi M oLegend, 147307IF1:50Rabbit anti-PAX8Proteintech, 10336-1-APIF1:200Goat anti-GATA2R&D systems, AF2046IF1:100Donkey anti-Rabbit Alexa Fluor 488Invitrogen, A21206IF1:500Donkey anti-Goat Alexa Fluor 555Invitrogen, A21432IF1:500Donkey anti-Mouse Alexa Flour 555Invitrogen, A31570IF1:500Donkey SerumEMD Millipore, S30IF– Information pertaining to antibodies used in this study A powerful feature of snMultiOme is the ability to map transcription factor motifs to accessible regions of chromatin and quantify the enrichment of motifs within isolated cell populations. We performed this type of transcription factor enrichment analysis using DARs linked to genes differentially expressed from E14 to P4. We identified transcription factors with increased enrichment, or “activity,” and increased expression from E14 to P4, as well as transcription factors with diminished enrichment and expression (Figure 2N). Many of these transcription factors include those known to be downregulated in the germline during this window, such as TFAP2C, SOX2, and POU5F1, as well as transcription factors that play essential roles in postnatal germline function, such as DMRTB1, DMRT1, SALL4, and SMAD2/3.53,54,55,56,57,58 Importantly, previous studies demonstrated that in the germline DMRT1 and SALL4 interact and co-localize within the same genome motif identified in our analyses herein; thus, for accuracy, this motif is represented as dual DMRT1 and SALL4 activity in germ cells.57 Finally, given that increased chromatin accessibility at distal intergenic sites, including potential enhancers, was linked with the onset of genes underlying SSC function, we mapped transcription factor activity within these regions to identify potential factors involved in SSC development (Figure S3D). Unique motifs enriched within these distal intergenic regions included those corresponding to E2F2, ZFP161, and E2F3. Overall, our data provide a resource of transcription factors that may be critical for germline development. Neonatal germ cell fate may not be overtly connected to chromatin signatures During postnatal development and the onset of spermatogenesis, prospermatogonia in the male germline transition to adopt different fates.4,28,31 Prior studies demonstrated that some prospermatogonia form undifferentiated spermatogonia, including SSCs, while others directly enter terminal differentiation as differentiating spermatogonia.25,26,27,28,30 Given the distinct epigenetic features we observed as prospermatogonia transition to form spermatogonia, we asked if any of these features are connected to different fates among prospermatogonia. To accomplish this, we generated unsupervised clusters of P4 germ cells with the goal of enriching for the three fates of neonatal spermatogonia, including SSCs, progenitor spermatogonia, and differentiating spermatogonia (Figure 3A). We leveraged transcriptomic data and differential gene expression of established markers of these populations to demonstrate relative enrichment of these subtly different populations among P4 spermatogonia (Figure 3B). Heterogeneous neonatal germ cell fates do not appear to coincide with major differences in chromatin structure (A) UMAP representation of clustering (based on snRNA-seq gene expression signatures) for E14 and P4 germ cells. (B) Expression (snRNA-seq) of marker genes that are associated with postnatal germ cell fate, including genes enriched within SSCs (Gfra1, Foxc2, Id4, Ret), progenitors (Neurog, Sox3), and differentiating spermatogonia (Stra8, Kit). (C) Total number of DARs, number of DARs linked to gene expression, and the number of DARs linked to the expression of DEGs among the 3 postnatal clusters (clusters 3–5). (D–F) Coverage plots corresponding to accessibility (snATAC-seq) and expression (snRNA-seq) of Gfra1 (D), Id4 (E), and Stra8 (F) among clusters 3–5. Gray regions indicate DARs among the 3 clusters, and asterisks (∗) denote regions identified in Figure 2. (G) Heatmap of relative transcription factor motif enrichment (or motif “activity”) for transcription factors with differential enrichment among clusters 3–5. (H) Enrichment for motifs associated with retinoic acid signaling among clusters 3–5. Data are presented as mean motif enrichment within each cluster. Next, we asked to what extent heterogeneity in gene expression is connected to differences in chromatin state. Surprisingly, among 2,407 DEGs and 18,266 DARs between clusters, just 9.50% (1736) of DARs were linked to the expression of only 8.12% (195) of DEGs among the clusters (Figure 3C), suggesting that differences in gene expression are weakly connected to chromatin state. Consistent with overall weak correlations between gene expression and chromatin state, differentially expressed marker genes lacked significant DARs linked to gene expression (Figures 3D–3F and S4A–S4E). Notably, enhancer elements strongly linked to the onset of SSC-associated genes identified in our analyses above were not differentially accessible among fate clusters (indicated with gray asterisks in Figures 3D, 3E, and S4A–S4D). Given that these findings are largely dependent on clustering, we performed a second, semi-supervised clustering approach. Previous studies indicate that germ cells at each developmental age generally arrange along a continuum with SSCs at one end and the most differentiated cell type at the opposite end.27,59 Consistent with previous studies,27 we order P4 germ cells along a pseudotime trajectory (Figure S5A). Then, we divided the pseudotime continuum into 3 groups based on pseudotime values, to again enrich for the three neonatal spermatogonial fates (Figure S5B). Finally, we evaluated chromatin accessibility, DARs, and gene expression across the three groups. Consistent with results above, while canonical marker genes implicated in SSC function or differentiating spermatogonia were differentially expressed, chromatin accessibility was largely unchanged among the three groups (Figures S5C–S5I). Even though the differential expression of canonical marker genes was not linked to heterogeneity in chromatin architecture, we asked if transcription factor enrichment may differ between the first set of unsupervised clusters. Of note, a short list of motifs was differentially enriched between the three clusters, including DMRT1/SALL4, which have established roles in spermatogenesis (Figure 3G).60,61 Notably, motifs of transcription factors associated with retinoic acid signaling, which trigger terminal differentiation of spermatogonia, were highly enriched within cluster 5 germ cells that also express canonical markers of differentiating spermatogonia like Stra8 and Kit (Figures 3H and S4E). Sertoli cell epigenetic maturation is largely established in early testis development Sertoli cells play a central role in both testis development and postnatal spermatogenesis. To explore the epigenetic maturation of Sertoli cells from E14 to P4 in more detail, we first computationally isolated Sertoli cells from the aggregate dataset. Interestingly, Sertoli cells from E14, E18, and P4 appeared to form discrete sub-populations when dimensionally reduced based on transcriptome signatures compared to chromatin signatures (Figures 4A–4C). However, after clustering and differential gene expression analyses, we discovered that most of this transcriptional heterogeneity was likely derived from differences in cell cycle state within each age (Figures S6A–S6C). Sertoli cell maturation from E14 to P4 (A–C) UMAP representations of Sertoli cells isolated from the aggregate data based on snRNA-seq (A), snATAC-seq (B), and joint snRNA- and snATAC-seq (C) data. (D) Temporal dynamics of expression for select marker genes implicated in Sertoli cell function at E14, E18, and P4. Horizontal lines represent the mean. (E) Coverage plot of genomic accessibility and expression for Fshr. (F and G) Differences in the number of DARs and DEGs and the mean fold change in DAR accessibility and DEG expression between Sertoli cells and germ cells. (H) Motif enrichment (or “activity”) for the expression of transcription factors that change during development within Sertoli cells. Data in F, G, and H (motif activity) are presented as means. Surprisingly, many of the canonical Sertoli cell markers, including Amh, Fshr, Inha, Lgals1, Rhox8, and Cyp26b1, changed only subtly at both the transcriptome and chromatin levels through developmental timepoints captured in our dataset (Figures 4D, 4E, and S6D–S6G). Indeed, overall numbers of DARs and DEGs quantified in Sertoli cells were markedly lower than those of germ cells spanning the same developmental window (Figures 4F and 4G). Also, the mean fold change in DARs was significantly lower compared to germ cells across the same ages, suggesting that epigenetic maturation in Sertoli cells from E14 to P4 is far more muted. Despite modest epigenetic changes, we leveraged our snMultiOme data to map transcription factor motifs with differential activity and differential expression from E14 to P4 (Figure 4H). While some transcription factors known to regulate Sertoli cell function were identified, such as WT1, NR5A2, and DMRT1,61,62,63,64,65 previously unstudied transcription factors were also identified, including KLF7 and ZFP281. Distinct chromatin signatures delineate interstitial and myoid lineages Peritubular myoid, Leydig, and other interstitial cells are generally thought to arise from a common population of multipotent progenitors during testis development. To explore the epigenetic maturation of these populations during development, we computationally isolated the peritubular myoid, Leydig, and other interstitial populations based on clustering and marker gene enrichment from above (Figures 5A, 5B, and S7A). The resulting UMAP based on transcriptomic data generated a stark projection of branches that could be ordered along both developmental age and pseudotime trajectory inference (Figures 5C and S7B). Two of these branches terminated in populations of terminally differentiated cell types, either peritubular myoid or Leydig cells. By contrast, the identity of the third branch was not clear. While trajectories were not as clearly resolved with snATAC-seq data of these cells, individual clusters largely segregated based on chromatin profiles (Figures S7C and S7D). Myoid and Leydig cell lineage maturation revealed lineage branches to terminal populations (A and B) UMAP representation of snRNA-seq profiles for peritubular myoid, Leydig, and interstitial progenitor cells isolated from the aggregate data (A) and through developmental time (B). (C) Trajectory inference modeling resolved distinct branches of cell sub-lineages through pseudotime. (D) Heatmap representation of the top 200 DEGs among the trajectory branches resolved. (E) Gene expression of canonical markers for interstitial progenitor, peritubular myoid, and Leydig cells. (F) UMAP representation of Gata2 and Gata4 expression. (G) Immunostaining for GATA2 and GATA4 within testis cross-sections at E14, E18, and P4. Orange and white arrowheads identify representative GATA2+; GATA4-or GATA4+; GATA2-interstitial cells, respectively. Dotted lines outline seminiferous tubules. Scale bars, 10 μm. Images are representative of n = 3 biological replicates. (H) Schematic representation of marker expression related to different lineage branches. (I–M) Coverage plots of accessibility (snATAC-seq) and expression (snRNA-seq) for marker and previously unidentified genes in which changes in chromatin accessibility, particularly at promoter and putative enhancer regions, are highly correlated with changes in gene expression (denoted with asterisk [∗]). (N and O) Motif enrichment (or “activity”) for transcription factors known to play roles in interstitial lineage development (N) and transcription factors that demonstrated high enrichment within the Leydig and peritubular myoid branches (O). Data in N and O are represented as relative means. To determine if these branches are indicative of potential functional differences, we quantified DEGs within each of the three branches and performed GO analysis. DEG analysis indicated that, indeed, each branch possessed a distinct gene expression signature (Figure 5D). Upon closer examination of canonical markers of this lineage, the unknown branch appeared to share some commonality with both the myoid and Leydig branches (Figure 5E). For instance, cells of the unknown branch expressed Hes1, Nr2f2 (or COUP-TFII), Tcf21, Pdgfra, and Wnt5a, which are markers of interstitial progenitors that give rise to Leydig and peritubular myoid cells or secreted factors essential for interstitial cell differentiation.14,36,66,67,68,69 Interestingly, the unknown branch was positive for the primitive interstitial progenitor marker Wt1 as well as the more advanced differentiation marker Gli1,14 but displayed disparate expression of the canonical Leydig markers Cyp11a1, Cyp17a1, and Lhcgr or the canonical peritubular myoid markers Cyp26b1, Acta2, and Myh11. Notably, we were unable to confidently identify separate populations of fetal and adult Leydig cells based on marker gene expression (Figures S7E–S7J), and thus, we could not confidently implicate the unknown branch with adult Leydig cell development; this likely reflects the yet unresolved origins of the adult Leydig cell population. As an alternative, we leveraged GO analysis of DEGs among the three branches, which indicated that a large portion (∼66%) of associated GO terms were indeed unique among the three branches (Figure S7K). Interestingly, among DEGs that define the three branches, we discovered that the unknown branch distinctly lacked expression of the canonical testis development factor Gata4 but was highly enriched for its familiar counterpart Gata2 (Figure 5F). GATA4 is critical for sex differentiation in early gonad development and subsequently critical for testis cord development (reviewed in70). Furthermore, during fetal development and after birth, GATA4 is highly expressed in Sertoli cells as well as peritubular myoid and Leydig cells.71,72 By contrast, the role of GATA2 in testis development is much less explored. To validate the divergent expression of Gata2 and Gata4 within the interstitial lineage, we performed immunofluorescence staining of testis cross sections from developmental ages captured in our studies (E14, E18, and P4), which largely confirmed the non-overlapping expression patterns of GATA2 and GATA4 within the interstitial compartment (Figure 5G). Furthermore, the presence of GATA2+ interstitial cells declined at P4 compared to prenatal ages. Thus, our data indicate that Gata2 and Gata4 may be distinguishing features among the highly heterogeneous interstitial lineage (Figure 5H) and support the role of GATA4 as a critical fate-determining transcription factor in testis development. To further explore if this heterogeneity among branches is rooted in variations in chromatin architecture, we leverage the associated snATAC-seq data from each cell. First, we identified 20,872 DARs among the branches of the interstitial lineage. Strikingly, 94.9% of the 2,257 DEGs that distinguish the branches were linked to at least one DAR, suggesting that transcriptional heterogeneity is largely linked to the underlying chromatin architecture. To explore this in more detail, we focused on several markers within the lineage, including Gata2 and Gata4. Among marker genes, the unknown branch had chromatin signatures reminiscent of the root branch in some cases, like that of Wt1 (Figure 5I), but also lacked distinct signatures of the root branch, such as Upk3b (Figure 5J). Furthermore, the unknown branch lacked chromatin signatures of the myoid and Leydig branches, especially canonical markers such as Lhcgr and Gata4 (Figures 5K and 5L). Finally, upstream elements linked to Gata2 expression were differentially accessible in the unknown branch (Figure 5M). Thus, our data suggest that the unknown branch represents an interstitial sub-population either in a transition state or a unique population of mesenchymal cells with unknown fate. Further studies are needed to explore this. Some canonical markers within the interstitial and myoid lineage include transcription factors. Therefore, we mapped motif enrichment for all transcription factors within DARs that distinguish each branch to approximate a relative “activity” of each transcription factor within the branches. For some known transcription factors involved in this lineage, patterns in enrichment/activity matched gene expression trends, such as LEF1 and WT1 (Figure 5N). However, the activity of transcription factors among the branches created a greater depth of information beyond only annotating expression. For example, Tcf21 transcripts were largely uniform across this lineage, consistent with TCF21 marking most cells of the interstitial compartment,68 but transcription factor activity appeared more confined to the myoid and Leydig cell branches (Figures 5E and 5N). Likewise, Gli1 transcripts are enriched within the myoid, Leydig, and unknown branches but activity appears far more enriched within the myoid branch (Figures 5E and 5N). Finally, we leverage the motif enrichment quantification to identify previously uncharacterized transcription factors with activity enriched within the branches (Figure 5O). Surprisingly, few transcription factor motifs were positively enriched within the unknown branch, and those that were enriched had low enrichment over the other clusters (>1.0 log2 fold change). By contrast, several transcription factors defined the root/other, myoid, and Leydig branches. Clear epigenetic boundaries define rete and seminiferous tubule sertoli cells During our initial annotation of cell types captured within the snMultiOme data, we discovered that EPCAM enrichment improved germ cell representation but also captured support-like cells within the rete region of the developing testis, as evidenced by clusters with elevated expression of notable markers such as Pax8, Cdh1, Krt8, and Nr2f2 (encodes for COUP-TFII; Figures 6A, 6B, and S8A). First, we confirmed that markers, such as PAX8 and CDH1, indeed identify cells predominantly localizing to the RT (Figures 6C–6E). Then, to explore the underlying epigenetics of this population in greater detail, we isolated clusters 19–22 of the aggregate MultiOme data and reassembled the paired snRNA-seq and snATAC-seq data from these populations (Figures 6F and S8B–S8F). UMAP projections of snRNA-seq, snATAC-seq, or combined data showed that individual clusters based on transcriptomic data largely segregated based on chromatin profile as well. Epigenetic regulation of support-like cells, including Sertoli cells of the rete testis (A) UMAP representations of support-like and rete cells colored based on developmental age. (B) Expression of canonical marker genes for cells within the support-like cell lineage, including Pax8, Cdh1, Krt8, and Nr2f2. (C) Diagram of testis anatomy, particularly the spatial distinction between the rete testis (RT) and seminiferous tubules (ST). (D and E) Immunostaining of CDH1 and PAX8 within testis cross-sections. Scale bars, 100 μm. Dotted lines identify RT and ST regions. Images are representative of n = 3 biological replicates at E14 and E18. (F) UMAP representation of cell clusters 19–22 corresponding to the original clustering in Figure 1. (G) Expression of markers for Sertoli cells (Sox9), primitive interstitial cells (Wt1), and early gonadal markers of the support-like cell population (Hes1 and Wnt4). (H and I) Coverage plots of accessibility for marker genes in which chromatin accessibility is highly correlated with expression. (J) Immunostaining of testis cross-sections for SOX9 and GATA4. Arrowheads indicate representative SOX9+; GATA4- Sertoli cells of the RT. Dotted lines identify RT and ST regions. Scale bars, 10 μm. Images are representative of n = 3 biological replicates at E14 and E18. (K–P) Coverage plots of chromatin accessibility and expression for marker genes that distinguish Sertoli cells of the RT versus ST. (Q) Motif enrichment (or “activity”) for and expression of transcription factors within RT and ST Sertoli cells. Motif activity is presented as means. To annotate differences among the four clusters, we calculated DARs and DEGs between the clusters. We discovered that only 35.40% of DEGs were linked to DARs, suggesting a lesser link between epigenetic heterogeneity and gene expression. Next, we leveraged both GO analysis of DEGs and marker genes to infer cluster identities. GO analysis netted largely overlapping GO terms without clear biological differences (Figures S8C–S8G). However, using marker genes, we found that Cluster 19 was enriched with the expression of the Sertoli cell marker Sox9 (Figure 6G). Furthermore, Cluster 20 was largely enriched for the interstitial cell marker Wt1. Both markers, Sox9 and Wt1, contained DARs with increased accessibility within the cluster of expression (Figures 6H and 6I). Finally, Clusters 21 and 22 lacked clear markers to explain functional differences but appeared as two closely related clusters that separate primarily based on developmental age, and thus, may represent residual support-like cells at E14 and E18. Indeed, both clusters have residual expression of the early gonadal marker Hes1 and the support-like cell lineage marker Wnt4. Given that our EPCAM enrichment strategy improved representation of otherwise rare Sox9 + Sertoli cells of the RT, we leveraged this unique opportunity to explore the underlying epigenetic attributes of Sertoli cells from the RT compared to those within ST. RT Sertoli cells in our dataset (Cluster 19) were distinct from clusters representing ST Sertoli (Clusters 6–8) such that the expression of Gata4 was lacking, which we confirmed by co-immunostaining for SOX9 and GATA4 (Figure 6J). Indeed, SOX9+ GATA4-cells were located within the RT region of the developing testis. Next, leveraging our snATAC-seq, we explored the underlying epigenetic signatures that demarcate these two functionally distinct populations of Sertoli cells. Overall, of 3,171 DEGs between RT and ST Sertoli cells, 78.7% (n = 2,496) of genes were linked to differentially accessible chromatin. While Sox9 chromatin signatures and gene expression were comparable between the Sertoli cell types, Pax8 was largely epigenetically repressed in Sertoli cells of the ST (Figures 6K and 6L). By contrast, canonical regulators and markers of ST Sertoli cells, including Fshr, Gata4, and Dmrt1, were heavily epigenetically and transcriptionally repressed in RT Sertoli cells (Figures 6M–6O). Notably, we discovered that Hoxb4 was among the strongest differentially expressed and differentially accessible genes in RT Sertoli cells versus those with STs (Figure 6P). Finally, we quantified transcription factor enrichment between the two Sertoli types. Outcomes revealed that many notable transcription factors with both differential expression and clear differences in motif enrichment (Figure 6Q). Collectively, these studies illuminate underlying epigenetic differences that define each Sertoli cell sub-type despite originating from a common population of progenitors in earlier testis development. Cell-to-cell communication dynamically modulates testis development One strength of single-cell transcriptomic approaches is the ability to analyze ligand-receptor combinations to infer potential cell-to-cell communication using computational approaches. Given the breadth of our data, we utilized this computational approach and discovered a wide range of cell interactions at each developmental timepoint captured in our data (Figures 7A and 7B). Interestingly, the myoid and interstitial lineage was a central node of cell signaling at each developmental age (Figures 7C–7E). By contrast, germ cells functioned far less as “receivers” of communication at E14 and were much more receptive to signals at P4. Overall, we discovered a major shift in the signaling pathways across all cell populations from E14 to P4 (Figure 7F). Cell-to-cell communication dynamics within the testis during development (A and B) Quantification of the total number of interactions (A) and sum of interaction weights (B) among all cells captured in the aggregate dataset at E14, E18, and P4. Data are presented as sums. (C–E) Circle plots of summed ligand-receptor pairs between cell types at E14 (C), E18 (D), and P4 (E). Dot diameter for each cell lineage represents the aggregate reception of signals from other cells, while line width represents the aggregate signal from one cell type to another. (F) Relative “activity” of signaling pathways that change between E14 and P4. (G–I) Circle plots for receptor-ligand combinations within FGF (G), GDNF (H), and BMP (I) signaling pathways, and expression of the associated receptors and ligands detected that comprise those pathways. In the context of germ cell maturation, some pathways were expected, such as CXCL (C-X-C motif chemokine) pathways and integrin signaling at E14, which play roles in PGC migration.9 On the other hand, the activity of several pathways implicated in SSC function within the male germline was enhanced at P4, including fibroblast growth factor (FGF), GDNF (glial cell-derived neurotrophic factor), and WNT (wingless-type MMTV integration site) signaling (Figures 7G, 7H, and S9A).73,74,75,76 In particular, peritubular myoid cells appeared to provide significant FGF pathway activity to not only germ cells, as previously established,76 but also Leydig, Sertoli, other myoid and interstitial cells, and support-like cells at P4. In general, FGF signaling was far more active at P4 than at E14. Finally, at E14 and P4, germ and Leydig cells appeared to generate significant BMP (bone morphogenic protein) pathway activity toward many testis cell types (Figures 7I and S9B). While BMP signaling has been proposed to participate in germ cell differentiation postnatally,77,78 further studies are necessary to elucidate the functional significance of this pathway. In the context of somatic cell types, our analysis revealed a clear representation of hedgehog (HH) signaling emanating from Sertoli cells and acting primarily on the interstitial lineage, including Leydig and peritubular myoid cells (Figure S9C), consistent with prior studies.79,80 Interestingly, we discovered that vascular endothelial growth factor (VEGF) signaling was largely active at E14 (Figure S9D), which coincides with testis morphogenesis and vascularization. More specifically, VEGF ligands originated from multiple cell types that acted on endothelial cells, which supports the role of interstitial and support cells in coordinating testis vascularization. Collectively, we present a list of signaling pathways, both known and previously unexplored, that warrant further investigation to understand the interconnected developmental cues underlying testis development. Discussion In this study, we generated a data atlas spanning key stages of mouse testis development to investigate how chromatin accessibility and gene expression dynamics underlie cell lineage specification and maturation. Prior studies have leveraged snATAC-seq alone to elucidate chromatin shifts during a similar developmental window,81 but how chromatin underlies the expression of fate-determining factors and shapes the divergence of lineage trajectories within the testis remains incompletely studied. To address this gap, we generated tandem snRNA-seq and snATAC-seq data from individual cells of E14, E18, and P4 mouse testes. After annotating testis cell lineages captured in our data, we evaluated the connection between chromatin changes and gene expression through development within each cell type. Furthermore, we leveraged the paired chromatin and gene expression data to identify candidate transcription factors that may control developmental transitions within each lineage. Several important themes emerge from our findings. Chromatin dynamics vary markedly across cell lineages Our analyses revealed that chromatin remodeling occurs in a lineage-specific and temporally regulated manner. Germ cells in particular undergo extensive changes in chromatin accessibility between E14 and P4, coinciding with the transition from prospermatogonia to spermatogonia and the emergence of the SSC population.4,28 Interestingly, widespread DNA methylation occurs during this developmental window,82,83 which makes our observation of widespread increases in chromatin accessibility surprising, given that DNA methylation is generally associated with gene silencing and chromatin compaction.84,85 A striking majority of DARs in germ cells at this transition mapped to distal intergenic elements, implicating enhancer-driven regulation as a potential mechanism in SSC specification. Indeed, the expression of canonical SSC-associated genes such as Gfra1, Foxc2, and Id4 was linked to distal accessible regions, whereas the promoters of such genes remained epigenetically stable. This result highlights the importance of distal regulatory control, including regulation at enhancers, for genes critical for SSC function. In contrast, Sertoli cells exhibited a remarkably stable chromatin landscape over the same developmental window. While subtle changes in transcription and chromatin were observed between E14 and P4, overall levels of epigenetic remodeling were notably lower compared to germ cells. For example, genes central to Sertoli cell function, including Amh, Fshr, and Inha86, carried consistent chromatin accessibility from E14 to P4. These findings suggest that Sertoli cells acquire a mature chromatin state early in testis development, possibly reflecting their roles in initiating sex determination and coordinating seminiferous cord formation and early testis morphogenesis.87,88 This early epigenetic maturity may confer stability to the Sertoli cell transcriptome and enable consistent support of testis development across ages. Other somatic lineages, including the interstitial, Leydig, and peritubular myoid populations, demonstrated divergent patterns of chromatin and transcriptional remodeling. Furthermore, most (94.5%) of transcriptional changes that indicate the differentiation of the interstitial population were linked to changes in chromatin architecture. Originating from a common population of Hes1, Nr2f2 (COUP-TFII), Tcr21, Pdgfra, and Wt1 positive progenitors,14,36,66,67,68,69 we resolved clear lineage separations into terminal cell types using trajectory inference models. As interstitial populations differentiate from these progenitors, downregulation of early testis regulators is accompanied by reduced chromatin accessibility; examples of this include Wt1 and Upk3b. On the other hand, differentiation toward terminal populations led to increased promoter and enhancer accessibility and gene expression for canonical Leydig cell markers such as Lhcgr, Cyp11a1, and Cyp17a1, and canonical peritubular myoid markers such as Cyp26b1 and Myh11. Strikingly, we identified a third interstitial lineage trajectory that notably lacked expression of the key interstitial regulator Gata4 (GATA4).70,71,72 Furthermore, this branch lacked Leydig cell genes such as those encoding the LH receptor (Lhcgr) and steroidogenic enzymes (Cyp11a1 and Cyp17a1) or peritubular myoid cell genes involved in muscle contraction and peristalsis (Myh11, Acta2). Finally, we discovered that this branch uniquely expressed a different member of the GATA family, Gata2 (GATA2). GATA2 expression in the interstitial compartment did not overlap with GATA4 at E14, E18, and P4, suggesting that GATA2 marks a distinct interstitial population. Notably, the presence of GATA2+ interstitial cells was markedly lower at P4 compared to E14 and E18. While the role of GATA2 in the epididymis has been explored,89 to our knowledge, the fate of GATA2+ testis interstitial cells has not been explored. In our studies, these cells displayed distinct chromatin features and transcriptional signatures that suggest a unique function. However, GO terms did not provide clear evidence of a function related to spermatogenesis. Thus, whether this population contributes to adult Leydig or myoid lineages, or instead represents an alternative mesenchymal trajectory, remains to be determined. Transcription factor activity reflects lineage progression A major strength of our multi-‘omic approach is the ability to infer transcription factor activity from the enrichment of transcription factor motifs alongside expression profiling for each transcription factor. This allowed us to not only confirm the activity of established regulators but also to identify previously unstudied candidates that may play important roles in testis development. For instance, in the germline, we observed increased activity of DMRT1·SALL4 and DMRTB1 at P4, consistent with their known roles in spermatogonial development and differentiation.53,54,55,57,58 Interestingly, DMRT1·SALL4 and DMRTB1 share remarkably similar but indeed unique motifs within the genome (Figure 2N),53,57 and thus, may represent a core set of key developmental regulators of neonatal development. In contrast, motifs for SOX2 and TFAP2C, which are involved in fetal germline programming, were depleted at postnatal stages. In terms of transcription factors with previously uncharacterized roles in testis lineages, KLF7 and ZFP281 emerged as differentially expressed and enriched at later developmental stages in Sertoli cells, suggesting potential contributions to postnatal Sertoli cell maturation. Within the interstitial lineage, transcription factor motif enrichment revealed more nuanced patterns than gene expression alone. For example, Tcf21 was broadly expressed across progenitor and terminal populations, consistent with previous lineage tracing studies,68 but the activity of TCF21 was higher in differentiated myoid and Leydig cells. This suggests that epigenetic remodeling may precede the transcription factor-dependent responses in some instances. It is important to note that the quantification of transcription factor activity herein is a predictive model of motif representation within DARs of each target population. This approach can present some limitations. For example, some motif enrichment may be biased toward transcription factors with more widespread motif localization. For example, in germ cells, SALL4 was flagged as highly enriched at P4, and previous studies indicate that SALL4 has a broad binding repertoire across the genome in germ cells.57 Furthermore, our quantification of motif activity was presented as cellular level enrichment and not necessarily at the genomic level. To accurately evaluate the genomic localization of transcription factors, ChIP-seq or CUT&RUN studies are necessary. Together, however, these data identified a suite of transcription factors for future functional studies. Germ cell fate heterogeneity may not be linked to chromatin architecture An unexpected finding of our study was the relatively weak correlation between chromatin accessibility and transcriptional heterogeneity among spermatogonia of different fates. Despite clear transcriptomic distinctions between SSCs, progenitor, and differentiating spermatogonia at P4, only a small subset of DEGs were linked to DARs. Moreover, enhancer elements associated with the expression of SSC-associated genes were not differentially accessible across spermatogonia clusters, despite differences in expression. This suggests that while enhancer priming may be critical for establishing SSC identity during the prospermatogonia-to-spermatogonia transition, broad differences in chromatin accessibility may not be the primary driver of spermatogonial subtypes at P4. A notable limitation, however, is that our analysis focused on chromatin accessibility as opposed to individual epigenetic marks. A recent study of isolated spermatogonia separated into either SSC-enriched or progenitor-enriched populations revealed that SSCs were largely not distinguishable based on ATAC-seq signature; rather, specific histone marks often differentially deposited at certain genomic loci demarcate SSCs and progenitors.90 Thus, future studies using single-cell approaches, such as single-cell CUT&Tag, may resolve this heterogeneity through development as it pertains to spermatogonial fate. An important caveat to the conclusion that broad chromatin accessibility may not influence germ cell fate is that we observed some differences in transcription factor motif enrichment among spermatogonial subtypes. In particular, motifs related to retinoic acid signaling were highly enriched in differentiating spermatogonia, consistent with known roles for retinoic acid in triggering the onset of spermatogenesis. While the motif enrichment may likely be a response to retinoic acid, further studies are needed to determine if a subset of prospermatogonia are primed with retinoic acid receptor motifs. Regardless, the widespread increase in retinoic acid receptor activity may contribute to the ability of retinoic acid to trigger the irreversible differentiation of spermatogonia. Epigenetic signatures demarcate sertoli cell subtypes Our dataset also provided a unique opportunity to explore the chromatin differences between Sertoli cells within seminiferous cords/tubules and those of the RT. While both subtypes share the expression of core Sertoli markers such as Sox9, they differ markedly in chromatin accessibility at lineage-defining genes. Rete Sertoli cells lacked accessibility at canonical tubule Sertoli markers (Gata4, Fshr, Dmrt1) and instead displayed enhanced accessibility at Pax8 and Hoxb4, suggesting distinct regulatory programs. These chromatin signatures likely underpin functional specializations of Sertoli cells within the rete, a region critical for sperm transport and fluid dynamics, and reinforce the notion that early lineage bifurcation is epigenetically encoded. Cell signaling and inter-lineage communication Lastly, our computational analysis of ligand-receptor interactions uncovered dynamic changes in intercellular communication during testis development. Our data indicates that germ cells become increasingly responsive to cues from somatic cells between E14 and P4, which largely coincides with the formation of SSCs and the onset of spermatogenesis.4 Specifically, many signaling pathways previously implicated in SSC maintenance or differentiation during spermatogenesis, including FGF, GDNF, and BMP signaling,75,76,91,92 became notably more active in germ cells at P4 due in large part to ligand availability. For instance, FGF2 is known to support replication of SSCs,76,93,94,95 and the availability of FGF2 from sources such as myoid and interstitial cells only became evident at P4 in our data. Interestingly, previous studies indicated that endothelial cells are a source of FGFs in the adult testis.76 However, our analyses did not identify a clear contribution of endothelial cells as ligand sources in the prenatal or neonatal testis, which may indicate developmental differences in the source of FGFs. In addition to the general increase in FGF activity with the onset of spermatogenesis, we observed the activation of the SSC self-renewing cue GDNF from both Sertoli and peritubular myoid cells, consistent with previous findings.40,96 Interestingly, we also observed the onset of BMP signaling, which, intriguingly, indicated significant BMP6 contributions from the Leydig cell population. However, the role of BMP in the onset of spermatogenesis and SSC formation is much less clear (reviewed in91,92). Collectively, we observed 50 predicted interactions and pathways that were identified as significantly changed from fetal to neonatal development (Figure 7F), including many that have not been explored in testis development and thus offer hypotheses for how intercellular communication shapes lineage specification. Future work incorporating spatial transcriptomics or perturbation-based models will be necessary to functionally validate these inferred networks. In summary, our integrated single-nucleus transcriptomic and chromatin accessibility atlas revealed complex, lineage-specific epigenetic regulation of testis development. The data underscore the asynchronous maturation of distinct testicular lineages, the enhancer-mediated control of SSC-related genes, and the potential role of understudied regulators that warrant functional validation. This resource provides a valuable foundation for dissecting the regulatory framework of testis development and lays the groundwork for future studies into the developmental origins of male infertility and gonadal disorders. Limitations of the study It is important to highlight that many of the findings presented herein are observational in nature and warrant functional studies to evaluate further. Additionally, future studies are necessary to explore epigenetic overlap with humans. Resource availability Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Nathan Law (nathan_law@wsu.edu). Materials availability This study did not generate new or unique reagents. Data and code availability •Single-nucleus paired transcriptome and chromatin (snMultiOme) from testis cells are available at the GEO database (GEO no. GSE305215). Processed data relating to chromatin accessibility of individual genes across cell types during development is available through the open-source Washington State University Bioinformatics Datahub at https://datahub.vetmed.wsu.edu. Single-nucleus paired transcriptome and chromatin (snMultiOme) from testis cells are available at the GEO database (GEO no. GSE305215). Processed data relating to chromatin accessibility of individual genes across cell types during development is available through the open-source Washington State University Bioinformatics Datahub at https://datahub.vetmed.wsu.edu. •Any additional data generated in the enclosed study is available upon request. Any additional data generated in the enclosed study is available upon request. •This study did not generate original code. Data analyses only used publicly available tools. The details of all computational analyses are outlined in STAR Methods. This study did not generate original code. Data analyses only used publicly available tools. The details of all computational analyses are outlined in STAR Methods. •Additional information required to reanalyze data from the enclosed study is available from the lead upon request. Additional information required to reanalyze data from the enclosed study is available from the lead upon request. Acknowledgments The authors acknowledge the substantial contributions of the many researchers whose work has advanced the field of testis development, and we emphasize that any omission of references does not diminish the significance of those contributions. The authors thank all the members of the Law lab for their valuable input and discussion. This research utilized resources from the Center for Institutional Research Computing at Washington State University. This research was supported by grant R01HD113747 from the NICHD (to N.C.L.). Author contributions Conceptualization, N.C.L.; methodology, K.H. and N.C.L.; investigation, M.R., T.M.H, B.L.B., K.H., and N.C.L.; writing – original draft, N.C.L.; writing – review and editing, M.R., T.M.H, B.L.B., K.H., and N.C.L.; funding acquisition, N.C.L. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesRat anti-EPCAM (CD326)Thermo/eBioscienceCat# 14-5791-81, clone G8.8; RRID: AB_953624Goat anti-GFRa1R&D systemsCat# AF560; RRID: AB_2110307Rabbit anti-DDX4AbcamCat# ab13840; RRID: AB_443012Goat anti-PLZFR&D systemsCat# AF2944; RRID: AB_2218943Rabbit anti-SOX9MilliporeSigmaCat# AB5535; RRID: AB_2239761Mouse anti-GATA4Santa CruzCat# SC-25310; RRID: AB_627667Rat anti-CDH1 - Alexa Fluor 647 ConjugatedBioLegendCat# 147307; RRID: AB_2563954Rabbit anti-PAX8ProteintechCat# 10336-1-AP; RRID: AB_2236705Goat anti-GATA2R&D systemsCat# AF2046; RRID: AB_335123Donkey anti-Rabbit Alexa Fluor 488InvitrogenCat# A21206; RRID: AB_2535792Donkey anti-Goat Alexa Fluor 555InvitrogenCat# A21432; RRID: AB_2535853Donkey anti-Mouse Alexa Flour 555InvitrogenCat# A31570; RRID: AB_2536180Chemicals, peptides, and recombinant proteinsDonkey SerumEMD MilliporeCat# S30Fish Skin GelatinMilliporeSigmaCat# G70414% paraformaldehyde/PBSThermo Fisher ScientificCat# 28908Critical commercial assaysMACS Cell Isolation KitMiltenyi BiotechCat# 130-090-312Single Cell Multiome ATAC + Gene Expression Kit10× GenomicsCat# 1000285Deposited dataRaw and analyzed snMultiOme dataThis paperGEO: GSE305215Analyzed snMultiOme dataThis paperhttps://datahub.vetmed.wsu.eduExperimental models: Organisms/strainsC57BL/6 J mice (Mus musculus)Jackson LaboratoriesStock# 000664Software and algorithmsCellranger_arc (v 2.0.0)10× Genomicshttps://www.10xgenomics.com/support/software/cell-ranger-arc/latest/tutorials/installationR software (v 4.2)R Foundationhttps://www.r-project.orgSignac (v 1.9.0)Stuart et al.97https://stuartlab.org/signac/Seurat (v 4.1.3)Hao et al.98https://satijalab.org/seurat/Seurat Wrappers (v 0.3.1)Hao et al.98https://satijalab.org/seurat/Monocle3Cao et al.99https://cole-trapnell-lab.github.io/monocle3/clusterProfiler (v 4.4.4)Xu et al.100https://bioconductor.org/packages/release/bioc/html/clusterProfiler.htmlLasX (v 3.7.5)Leica Microsystemshttps://www.leica-microsystems.com/products/microscope-software/p/leica-las-x-ls/downloads/ Experimental model and subjects details Animals All protocols for the care and use of animals were approved by the Washington State University Animal Care and Use Committee (approval reference numbers 6791 and 6746). Mice for the studies described herein were maintained in a 12 h light/dark cycle at constant room temperature (22°C–24°C) with food and water provided ad libitum. Mice were generated from natural mating of sexually mature C57BL/6 J males and females (Jackson Laboratories, stock no. 000664). Identification of a copulatory plug signified embryonic day (E) 0.5 of gestation. Day of birth was considered postnatal day (P) 0. Testes from male mice were collected at E14.5, E18.5, and P4 for the studies herein. The studies herein focused on testis development; thus, females were not studied. Method details Experimental design of snMultiOme Parallel single-nucleus RNA- and ATAC-sequencing were performed from freshly prepared single-cell suspensions generated from both testes of E14.5, E18.5, and P4 male mice. Each developmental age was collected in duplicate from different litters, representing n = 2 biologically distinct replicates per age. To obtain sufficient cells, testes from E14.5 and E18.5 males were pooled (7 males per replicate for E14.5 and 2–3 males per replicate for E18.5). Paired testes from a single male were used for P4 replicates. Replicates of each age were processed side-by-side on the same experimental day, but each developmental age was collected on different experimental days. Single-cell suspensions and cell enrichment To generate single-cell suspensions from isolated testes, detunicated testes were incubated in a solution of 0.25% trypsin/EDTA (Thermo Fisher Scientific) and 2 mg/mL deoxyribonuclease I (MilliporeSigma) for 10 min at 37°C with occasional, gentle pipetting. The trypsin digest was quenched with the addition fetal bovine serum (FBS) to a final concentration of 10% and poured through a 40 μm filter. The resulting single-cell suspension was washed and resuspended in DPBS-S (1% FBS, 1 mg/mL glucose, 1 mM sodium pyruvate, 10 mM HEPES, 100 units/mL penicillin, and 100 μg/mL streptomycin in PBS [Thermo Fisher Scientific]). Approximately 1 × 109 cells were resuspended in DPBS-S and incubated with 3 μg of a mouse monoclonal antibody recognizing EPCAM (Thermo Fisher, cat. no. 14-5791-8; clone G8.8) for 30 min on ice. Antibody-bound cells were then isolated using a magnetic cell sorting (MACS) Cell Isolation Kit and MS columns (Miltenyi Biotech) according to manufacturer protocol. Equal parts of MACS-isolated cells and flow-through were used to generate an EPCAM-enriched population. Nuclear isolation The EPCAM-enriched cell suspension was then resuspended in ice-cold Nuclear Lysis Buffer (10 mM Tris-HCL, 10 mM NaCl, 3 mM MgCl2, and 0.1% NP-40), mixed by pipette thrice, and immediately spun down to remove the Nuclear Lysis Buffer. Pelleted nuclei were washed and resuspended in 1× Nuclei Buffer (10× Genomics, Inc.). snMultiOme library preparation and sequencing Libraries for parallel snRNA-seq and snATAC-seq were prepared from the isolated testis nuclei according to manufacturer protocols (10× Genomics, Inc.). Libraries for each assay were pooled at proportions targeting equal read depth across libraries (see Table 1 for library metrics) and sequenced with an Illumina NovaSeq machine (Novogene Co, Ltd.) with paired-end 150 base-pair reads. Data assembly and quality control Demultiplexing, alignment to the mouse genome/transcriptome, and read quantification was conducted using the 10× Genomics cellranger_arc pipeline (v. 2.0.0). Gene-cell (snRNA-seq) and peak-cell (snATAC-seq) matrices across all biological replicates (n = 2 replicates across 3 biological ages) were aggregated and analyzed further using R software (v. 4.2) and the Signac97 (v. 1.9.0) and Seurat98 (v. 4.1.3) packages. With these packages. low-quality cells and estimated duplicates were excluded from downstream analysis by selecting cells with at least 1000 genes and 1000 unique peaks detected, less than 50000 snRNA-seq reads and 100000 snATAC-seq reads, and less than 35% mitochondrial transcript representation in snRNA-seq data. Dimensional reduction and cell identification Dimensional reduction was accomplished within Signac and Seurat using principal component analysis (PCA) and uniform manifold approximation projections (UMAP) for snRNA-seq data, latent semantic indexing (LSI) and UMAP for snATAC-seq data, and weighted nearest neighbor and UMAP for combined snRNA-seq and snATAC-seq projections. Cell types within the aggregated data were inferred using a hierarchical clustering approach using the snRNA-seq data within Signac/Seurat software (n = 50 principal components, resolution = 0.2) to generate a total of 22 clusters. Differentially expressed genes (DEGs) were identified within each cluster using Seurat/Signac default settings (including use of Wilcoxon rank sum tests, p < 0.05) to identify marker genes and infer cell-type identity. Trajectory inference and pseudotime ordering Starting with Seurat/Signac objects, trajectory inference and ordering was accomplished using the SeuratWrappers package97,98 (v. 0.3.1) which leverages the monocle3 package99 (v. 1.3.1). Briefly, Seurat objects were converted to Monocle objects using the “as.cell_data_set” function. Then, data was compiled using the “preprocess_cds”, “cluster cells” (via UMAP reduction), and “learn_graph” functions before running “order_cells” using UMAP reduction. Immunofluorescent stains of testis cross-Sections Tissue preservation, sectioning, and staining was conducted as previous.101 Briefly, testes were collected and fixed in 4% paraformaldehyde/PBS (Thermo Fisher Scientific, 28908) at 4°C for specified time frames depending on the animal age,102 dehydrated in an ethanol/xylenes series, and embedded in paraffin wax. Tissue sections were cut at 5 μm, placed on glass slides, and incubated in a series of xylenes/ethanol washes followed by sodium citrate antigen retrieval. Slides were then washed with 1× PBS, briefly submerged in 0.1% Sudan black/EtOH to mitigate background autofluorescence, washed again, and incubated for 1 h in blocking buffer containing 1% serum matched to the secondary antibody host species, 0.1% fish skin gelatin (Sigma, G7041), and 1× PBST (PBS +0.01% Tween 20). Following blocking, slides were incubated with select primary antibodies (see Table 2 for antibodies) overnight at 4°C. Slides were then washed three times in 1× PBS and incubated with the appropriate fluorescent labeled secondary antibody diluted in blocking buffer for 2 h at RT. Slides were washed three times in 1× PBS prior to mounting with Vectashield HardSet medium (Vector Laboratories). Slides were imaged using a DMi8 inverted fluorescent microscope (Leica-Microsystems; LasX v3.7.5). Presented images are representative of n ≥ 3 biological replicates. Quantification and statistical analysis Differential genes and differential peaks For downstream analysis of different cell populations across developmental ages, DEGs within snRNA-seq data or differentially accessible regions (DARs) within snATAC-seq data were quantified using Seurat/Signac and the default settings (including use of Wilcoxon rank sum tests, p < 0.05). Gene ontology for DARs and/or DEGs was a performed using the clusterProfiler package100 (v. 4.4.4). Peak-gene linkage analysis Correlations between accessible peaks and expressed genes were quantified in the aggregate dataset which included all cell types as the most stringent approach to link peaks to genes. This Linkage analysis was performed using the LinkPeaks function in Signac with the default settings to ultimately generate z-scores for each peak and gene (p < 0.05). Motif enrichment DARs were subset to include only those linked with gene expression. Then, mouse transcription factor motifs from the JASPAR database were mapped to DARs linked to gene expression using the FindMotifs function in Signac. Next, transcription factor motif “activity” was calculated using the RunChromVar function in Signac and differential motif activity was quantified using the FindMarkers function in Signac. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.116452. Supplemental information Document S1. Figures S1–S9 Table S1. Differentially expressed genes among the 22 hierarchical clusters that capture testis cell types through development References Associated Data Supplementary Materials Document S1. Figures S1–S9 Table S1. Differentially expressed genes among the 22 hierarchical clusters that capture testis cell types through development Data Availability Statement •Single-nucleus paired transcriptome and chromatin (snMultiOme) from testis cells are available at the GEO database (GEO no. GSE305215). Processed data relating to chromatin accessibility of individual genes across cell types during development is available through the open-source Washington State University Bioinformatics Datahub at https://datahub.vetmed.wsu.edu. Single-nucleus paired transcriptome and chromatin (snMultiOme) from testis cells are available at the GEO database (GEO no. GSE305215). Processed data relating to chromatin accessibility of individual genes across cell types during development is available through the open-source Washington State University Bioinformatics Datahub at https://datahub.vetmed.wsu.edu. •Any additional data generated in the enclosed study is available upon request. Any additional data generated in the enclosed study is available upon request. •This study did not generate original code. Data analyses only used publicly available tools. The details of all computational analyses are outlined in STAR Methods. This study did not generate original code. Data analyses only used publicly available tools. The details of all computational analyses are outlined in STAR Methods. •Additional information required to reanalyze data from the enclosed study is available from the lead upon request. Additional information required to reanalyze data from the enclosed study is available from the lead upon request.