The Neonatal and Adult Human Testis Defined at the Single-Cell Level

Abstract

Spermatogenesis has been intensely studied in rodents but remains poorly understood in humans. Here, we used single-cell RNA sequencing to analyze human testes. Clustering analysis of neonatal testes reveals several cell subsets, including cell populations with characteristics of primordial germ cells (PGCs) and spermatogonial stem cells (SSCs). In adult testes, we identify four undifferentiated spermatogonia (SPG) clusters, each of which expresses specific marker genes. We identify protein markers for the most primitive SPG state, allowing us to purify this likely SSC-enriched cell subset. We map the timeline of male germ cell development from PGCs through fetal germ cells to differentiating adult SPG stages. We also define somatic cell subsets in both neonatal and adult testes and trace their developmental trajectories. Our data provide a blueprint of the developing human male germline and supporting somatic cells. The PGC-like and SSC markers are candidates to be used for SSC therapy to treat infertility.

Keywords: Leydig cells; Sertoli cells; germ cells; peritubular myoid cells; primordial germ cells; single-cell RNA sequencing; spermatogenesis; spermatogonia; spermatogonial stem cells; testes.

Figure 1.. Identification of Adult Human SPG Subsets

(A) tSNE plot of adult human testes cell clusters defined by scRNA-seq analysis. The fraction of the total cells in each subset is: spermatogonia (SPG) (39%), spermatocyte (SPC) (6%), spermatid (ST) (1%), Leydig cell (LC) (2%), peritubular myoid cell (PTM) (24%), endothelial cell (EC) and blood (27%), and macrophages (M) (1%). (B) SPG subsets defined by clustering analysis. The tSNE plot (top) was generated from the cells in the SPG subset defined in (A). The feature plots (bottom left) and violin plots (bottom right) show the expression pattern of markers (both known and new) that chart the progression of human SPG. (C) Cell cycle gene expression in SPG subsets. The ridge plots show the expression of S, G2-M, and S-G2-M phase genes (TOP2A, CDK1, and MKI67, respectively). The MKI67 tSNE features plot corroborates that Early diff-SPG and Diff-SPG cells tend to express higher levels of proliferation genes than spermatogonial stem cell 1 (SSC-1) and SSC-2 cells. (D) Heatmap of G2-M and S phase genes in the SPG subsets. (E) Developmental timeline of SPG cell subsets. The top shows Monocle pseudotime trajectory analysis of the SPG cell subsets defined in (B).The arrow indicates the developmental direction, based on the SSC-1 and Diff-SPG subsets being the least and most advanced, respectively. The red box labels a branch of cells that evidence suggests are TCs (see text). The expression pattern of marker genes is plotted along the pseudotime axis (bottom; same genes as those shown in B). The marker genes are predominantly expressed in the SPG subsets indicated in the vertical bar on the right. ED, early differentiating; D, differentiating.

Figure 2.. Identification and Characterization of Human Undifferentiated SPG Subsets

(A) tSNE plot from Figure 1B showing that TCs are predominantly at the juxtaposition point of SSC-1, SSC-2, and Early diff-SPG subsets (top). The feature plot shows the expression of some gene markers enriched in TCs (bottom). The arrows indicate the developmental directionality of the cell subsets, based on our model (see text). (B) Heatmap of top differentially expressed genes (DEGs) from the three cell subsets related to TCs compared to their expression in TCs themselves. DEGs that specifically mark the SSC-1, SSC-2, and Early diff-SPG subsets were defined as genes enriched in these 3 respective cell clusters when TCs were excluded. The blue and green boxes denote TCs expressing SSC-1A- and SSC-2-enriched genes. The red box marks TCs expressing genes enriched in SSC-1A, SSC-2, and Early diff-SPG. (C) tSNE plot from Figure 1B showing the three SSC-1 subsets identified using PAGODA2. (D) tSNE plot of only the 3 SSC-1 subsets. (E) Feature plot, violin plot, and Monocle pseudotime trajectory expression pattern of SSC-1A, -1B, and -1C marker genes. (F) Pseudotime trajectories of the adult SPG subsets shown in the key.

Figure 3.. Identification and Characterization of Primitive Undifferentiated SPG Markers

(A) IHC analysis of 5 candidate SSC marker proteins (representative images of 2 biological replicates). The genes encoding these proteins are predominantly expressed in the SSC-1B region, as shown by the tSNE plot. As a control, we included UTF1, a previously defined undifferentiated SPG marker (Di Persio et al., 2017) whose gene is broadly expressed in SPG, as shown in its tSNE plot. (B) IF analysis of adult testis sections stained with the antibodies shown. Anti-LPPR3 produced two different staining patterns: speckled (SPEK [pink arrow]) and whole nucleoplasm (NP [light blue arrow]). The cells were co-stained with an antibody against UTF1. Quantification of two biological replicates showed that both PIWIL4 and LPPR3 are expressed in a more selective set of cells than is UTF1. Scale bar: 50 μM. (C) IF analysis of intact human seminiferous tubules (whole mount) stained with the indicated antibodies (representative images of 2 biological replicates). Dashed lines outline the seminiferous tubules. The yellow arrows label PIWIL4+GFRA1+ and LPPR3+GFRA1+ cells, whereas the light blue arrows label GFRA1 single-positive cells. Scale bars: 100 (top) and 10 μM (bottom). (D) FACS plot of adult human testicular cells stained with the antibodies shown (left). The percentage of positive and negative cells are indicated in upper and lower regions, respectively; <0.1% positive cells stained with secondary antibody only (see Figure S3A). qPCR analysis was performed on sorted positive cells and unfractionated cells. The values shown are from FACS-purified cells relative to unfractionated cells (mean ± SD from two biological replicates), the latter of which was given a value of 1 (dotted line). The 10 genes listed first are known undifferentiated SPG markers or candidate SSC markers identified herein. The last gene VIM is a somatic cell-expressed gene and, thus, serves as a negative control.

Figure 4.. Identification and Characterization of Neonatal Germ Cells

(A) tSNE plot of neonatal human testicular cells analyzed by scRNA-seq analysis. Cell subsets were identified based on the expression pattern of known marker genes (Figure S4A). The fraction of the total cells in each subset is: germ cells (1%), Sertoli cell (SC) (59%), LC (32%), PTM (5%), and EC and blood (4%). (B) Neonatal germ cell subsets and markers defined by clustering analysis using Monocle. The tSNE plot (top) was generated from the cells in the germ subset defined in (A). Violin plots (bottom) show the expression pattern of selected PGCL and PreSPG gene markers. (C) Pearson correlation analysis of the cumulative expression of genes in PGCs at the gestational ages indicated with neonatal cell clusters. The values at top are the R² values. (D) Developmental timeline of PGCs and neonatal germ cell subsets. The top shows Monocle pseudotime trajectory analysis of the neonatal germ cell subsets defined in (B), as well as PGCs at the gestational ages indicated from Guo et al. (2015). The arrow indicates the developmental direction, based on the 4-week PGC and PreSPG subsets being the least and most differentiated, respectively. The expression pattern of marker genes is plotted along the pseudotime axis (bottom left) and with violin plots (bottom right). (E) Representative images of the expression pattern of 3 PGCL protein markers in a single neonatal (D7) human testis, as determined by IHC analysis. Red arrows mark positive cells. Gray dotted lines mark the approximate margins of the seminiferous tubules. As a positive control, cells were stained with an antibody against UTF1, which is encoded by a gene expressed in most neonatal germ cells (Figure 4D). The violin plot below shows quantification of positively stained cells. Scale bar: 50 μM. (F) Relationship of neonatal germ cells with PGCs and adult SSCs. tSNE analysis and violin plots of PGCs (from Guo et al., 2015), neonatal germ cell subsets (defined in B, above) and the SSC-1 subset. The violin plots show stage-specific marker genes. (G) The neonatal testis has adult SSC-like cells. The top shows Monocle pseudotime trajectory analysis of the indicated germ cell subsets. The arrow indicates the developmental direction. The expression pattern of marker genes is plotted along the pseudotime axis (bottom).

Figure 5.. Characterization of Somatic Cell Populations in Human Testes

(A) tSNE plot of adult and neonatal testicular somatic cells. Clusters were annotated based on known markers shown in (D)–(F). The percentage of the total in each subset is neonatal SC (20.0%), neonatal LC (31.3%), neonatal PTM (1.3%), adult SC (0.2%), adult LC (23.0%), and adult PTM (24.2%). (B) tSNE plot of the same clusters as in (A) with original sample source identities (neonatal and adult) indicated. (C) tSNE plot of the same clusters as in (A) marked by phase of cell cycle, as determined by average expression of G2-M and S phase gene expression (Kowalczyk et al., 2015) in each cell cluster relative to each other. (D) Heatmap showing the expression profile of neonatal and adult somatic cell marker genes. (E) LC markers used for cluster annotation. Monocle pseudotime trajectory of neonatal and adult LC subsets (top). Heatmap shows the expression profile of neonatal and adult LC markers (bottom). (F) PTM markers used for cluster annotation. Monocle pseudotime trajectory of neonatal and adult PTM subsets. Heatmap shows the expression profile of neonatal and adult PTM markers (bottom).

Figure 6.. Receptor-Ligand Signaling Interactions in Human Testes

(A and B) Notch ligand-receptor co-expression in the neonatal testis. tSNE plot of all neonatal cells (from Figure 4A) showing expression of xthe receptor NOTCH2 (in neonatal SC) and the ligands DLL3 (in neonatal germ cells) (A) and DLK1 (in neonatal PreSPG, LC, and PTM) (B).The inset shows tSNE plot of expression of only the ligand genes in neonatal germ cells (from Figure 4B). The dotplot shows expression of the indicated genes in the cell subsets shown. Note that percentage of positive cells indicated is a conservative estimate from scRNA-seq, an insensitive technique. (C) Kit ligand-receptor co-expression in the neonatal testis. KIT is expressed in neonatal germ cells and the ligand KITLG is expressed in neonatal SC and EC. (D) Hh ligand-receptor co-expression in the neonatal testis. The Hh receptor PTCH1 is expressed in neonatal PreSPG, LC, and PTM. The Hh ligand DHH is expressed in neonatal SC.

Figure 7.. Human Male Gametogenesis Model and Neonatal Signaling

(A) Model. Black arrows indicate the direction of differentiation. Grey arrows indicate transitions between different cell states. The expression pattern of key marker genes (both previously known and discovered in this study) are indicated by boxes (other marker genes are listed in Table S2). (B) Top: neonatal and adult somatic cells and selected marker genes. Bottom: signaling interactions between cells in the neonatal testes predicted from the scRNA-seq data shown in Figure 6. Gray arrows, somatic cell-somatic cell signaling. Black arrows, somatic cell-germ cell signaling.