We previously performed scRNA-seq analysis on 2,486 (319 with the Tang protocol and 2,167 with a modified STRT-Seq protocol) human FGCs and their microenvironment somatic cell for 44 embryos between 4 and 26 weeks of gestation9. We comprehensively and extensively analyzed the transcriptomic features of FGCs and clearly identified four and two types of FGCs in females and males, respectively. In females, these types were mitotic-phase, RA signaling-response-phase, meiotic-prophase and oogenesis-phase FGCs, while in males, the mitotic-phase and mitotic arrest-phase FGCs were identified. However, insufficient numbers of gonadal somatic cells were analyzed; thus, the developmental characteristics of gonadal somatic cells were not fully explored in our previous studies.
In this study, we created a comprehensive map of cell populations in the human ovaries and testes, including both FGCs and somatic cells in their niche. To characterize the differentiation and maturation of FGCs and gonadal somatic cells, we generated developmental expression profiles of 56,399 gonadal cells for 16 human embryos and fetuses between 6 and 23W of gestation. Specifically, we identified 7 groups of FGCs, including newly identified FGCs that simultaneously expressed POU5F1 and SPARC. In addition to immune cells, endothelial cells and blood cells, we identified three and seven types of somatic cells in male gonads (testes) and female gonads (ovaries), respectively. In both testes and ovaries, we identified a new KRT19+ cell population that highly expressed the RA synthesis enzyme ALDH1A2 at early developmental stages (Fig. 4f). We have found and verified surface protein markers that can be used to isolate specific somatic cell types in future studies, such as CDH2 for Sertoli cells and PDGFRA for Leydig cells. In addition, we also combined the postnatal 55,404 gonadal single cells (20,676 ovarian single cells and 34,728 testicular single cells) from four published datasets, and comprehensively and systematically explored the gene expression pattern of gonadal cells10,11,13,14. Especially for the testis, we covered single cell transcriptomic data of male individuals from 4W to 25W before birth, and from 2 days to 42 years of ages after birth, and conducted in-depth research on the expression patterns of testicular somatic and germ cells9,11,13.
Second, we assessed the developmental origins of granulosa cells in females and of steroidogenic lineages in both sexes (Fig. 4f). Our data indicated that DLK1+ cells in the gonads of 7W of male and female embryos may further differentiate into steroidogenic cell lineages: Leydig cells in the testes and theca cells in the ovaries. However, DLK1 expression continued in Leydig cells of the testes in subsequent developmental stages, while it disappeared in theca-like cells of the ovaries. A human fetus does not undergo sex differentiation until approximately 7W. Our data indicated that in 7W embryos, the expression feature and cell types of both germ cells and gonadal somatic cells are very similar if not identical in male and female embryos. Next, by mapping our ovarian somatic cell data to published adult ovarian scRNA-seq data, we found that the expression profiles of TAC1+ cells and proGCs are very similar; thus, we speculate that TAC1+ cells in 7W female gonads may further differentiate into granulosa cells in later developmental stages. In addition, through RNA velocity analysis, we found that Mid cells that appearing after 13W have the potential to differentiate to KRT19+ cells and granular cells (FOXL2+ cells).
Although the molecular characteristics, cellular origins and developmental and functional links of fLCs and aLCs have been extensively studied in rodents, we are just beginning to unravel these details in humans due to the scarcity of samples and technical limitations. In mammals, there are two main types of Leydig cells, fLCs and aLCs. Leydig cells are the main sources of androgens, while Sertoli cells are generally accepted as nonsteroidogenic cells. However, in this study, we found that Sertoli cells, rather than Leydig cells express HSD17B3, an enzyme that mediates the final step of testosterone synthesis, in the fetal (18.5% in the first trimester and 34.5% in the second trimester) and neonatal periods (30%). Leydig cells did not express HSD17B3 until neonatal stages (6.8%) and the ratio of cells that expressed HSD17B3 gradually increased in the later stages. This finding indicates that the androstenedione produced by fLCs is transferred to fetal Sertoli cells and then converted to testosterone. Then, during the neonatal stages, testosterone production function is gradually taken over by Leydig cells. At the adult stage, Leydig cells can produce testosterone by themselves since they can express HSD17B3 and other steroidogenic enzymes simultaneously. Additionally, we elucidated the cell-cell interactions between Leydig cells and Sertoli cells. Firstly, we found that the Hh ligand genes DHH and its downstream receptor gene PTCH1 were expressed in Sertoli cells and Leydig cells, respectively. However, the DHH-PTCH1 ligand-receptor regulates mainly the interactions between Sertoli and peritubular myoid cells in mice21. Secondly, we found that Sertoli cells and Leydig cells interacted with each other through PDGFB-PDGFRA/PDGFRB ligand-receptor gene pairs. Notably, these interactions were also observed in neonatal and adult testes.
To the best of our knowledge, this study is the first to systematically and comprehensively compare the gene expression patterns and cell type compositions among monosomy X (45, XO), normal female (46, XX) and male (46, XY) embryos at the same developmental stage (7W). Globally, monosomy X caused abnormalities not only in the transcriptome of FGCs, but also in gonadal somatic cells. More importantly, monosomy X was associated with cell type-specific developmental defects in FGCs and gonadal somatic cells. Furthermore, it also caused the loss of a PCP4-positive cell population in XO gonad compared with normal XX gonads (Supplementary Fig. 3).
Altogether, our data identified and characterized new types of FGCs and gonadal somatic cells. We speculated that the DLK1+ cell population was a progenitor cell population of the steroidogenic cell lineage and that the TAC1+ cell population was a progenitor cell population of granulosa cells. In addition, we illustrated the transcriptomic changes in monosomy X embryos. Our data also have potential applications for the isolation of specific gonadal somatic cell populations in vivo. It deepens our view of both germ cell and gonadal somatic cell development in the testes and ovaries. Finally, our work provides insights and valuable data resources for further elucidation of the molecular mechanisms of infertility.