RNA helicase DDX5 acts as a critical regulator for survival of neonatal mouse gonocytes

Abstract Objectives Mammalian spermatogenesis is a biological process of male gamete formation. Gonocytes are the only precursors of spermatogonial stem cells (SSCs) which develop into mature spermatozoa. DDX5 is one of DEAD‐box RNA helicases and expresses in male germ cells, suggesting that Ddx5 plays important functions during spermatogenesis. Here, we explore the functions of Ddx5 in regulating the specification of gonocytes. Materials and Methods Germ cell‐specific Ddx5 knockout (Ddx5 ‐/‐) mice were generated. The morphology of testes and epididymides and fertility in both wild‐type and Ddx5 ‐/‐ mice were analysed. Single‐cell RNA sequencing (scRNA‐seq) was used to profile the transcriptome in testes from wild‐type and Ddx5 ‐/‐ mice at postnatal day (P) 2. Dysregulated genes were validated by single‐cell qRT‐PCR and immunofluorescent staining. Results In male mice, Ddx5 was expressed in germ cells at different stages of development. Germ cell‐specific Ddx5 knockout adult male mice were sterile due to completely devoid of germ cells. Male germ cells gradually disappeared in Ddx5 ‐/‐ mice from E18.5 to P6. Single‐cell transcriptome analysis showed that genes involved in cell cycle and glial cell line‐derived neurotrophic factor (GDNF) pathway were significantly decreased in Ddx5‐deficient gonocytes. Notably, Ddx5 ablation impeded the proliferation of gonocytes. Conclusions Our study reveals the critical roles of Ddx5 in fate determination of gonocytes, offering a novel insight into the pathogenesis of male sterility.


| INTRODUC TI ON
In mice, as the foundation for the continuity of spermatogenesis, SSCs are derived from gonocytes that are originated from primordial germ cells (PGCs) residing in the proximal epiblast during gastrulation. [1][2][3] PGCs migrate to the developing gonad and differentiate into male or female germ cells following sex determination. [4][5][6][7][8] In the fetal male gonads, germ cells are referred to as gonocytes. 9 Testicular gonocytes stay in the G0/G1 phase of cell cycle from embryonic day (E) 13.5. 10,11 The quiescent gonocytes begin to relocate to the periphery from the central region of seminiferous tubules, and resume mitotic division shortly after birth. 10,12 Gonocytes give rise to undifferentiated spermatogonia or differentiating spermatogonia through gonocytes-to-spermatogonia transition (GST), which is the gateway to spermatogenesis in neonatal testes. Proliferation and migration of gonocytes are two crucial events that take place between P0 and P6. [13][14][15][16] Gonocytes and spermatogonia are closely related, but they have different morphology, transcriptome, DNA methylation and chromosome architecture. 12,15 Although tremendous efforts have been focused on studying the maintenance and differentiation of SSCs, the formation of SSCs (especially the proliferation, migration and transition of gonocytes), is still unclear. Sertoli cells have been reported to play important roles in regulating these processes. Platelet-derived growth factor (PDGF), synthesized and secreted by Sertoli cells, has the remarkable capability to drive the survival and migration of the gonocytes. 17,18 As a RNA-binding protein, DDX5 has been shown to be involved in multiple biological processes. Ddx5 has been reported to express in male germline [19][20][21] and plays a negative role in WNT signaling regulation of the GC-1 spg cell. 22 Disruption of Ddx5 in adult spermatogonia results in infertility. 23 Although it is clear that Ddx5 is required for the maintenance of spermatogonia in adult testis, we do not yet know whether it is required for the formation of the spermatogonia.
In this study, we found that Ddx5 was highly expressed in male germ cells from prenatal period. We generated germ cell-specific Ddx5 knockout mice by using transgenic mice expressing Cre recombinase driven by the Mvh promoter (Mvh-Cre) and found that Ddx5 -/mice showed significantly smaller and lighter testes than wild-type (WT) and heterozygous (Ddx5 +/-) mice. Furthermore, Ddx5 knockout resulted in a complete loss of germ cells, containing only Sertoli cells within 6 days after birth, and led to azoospermia and infertility in male adults. In addition, scRNA-seq experiments indicated that Ddx5 knockout resulted in aberrant expression of genes that were associated with cell cycle and GDNF pathway in gonocytes. Interestingly, we further found that Ddx5 knockout induced dysregulation of GDNF signaling pathway and then impeded proliferation of gonocytes and gonocytes-to-spermatogonia transition.

| Assessment of fertility and fecundity
To assess fertility and fecundity, one 3-month-old male mouse of either wild-type, Ddx5 +/or Ddx5 -/was placed into a cage with a wildtype female. Cages were monitored daily. The females were checked for the presence of vaginal plugs and pregnancy. The numbers of pups born in each mating set were recorded.

| Immunofluorescent staining of the frozen sections
with the diluted primary antibody and subsequently secondary antibody. Primary antibodies used in this study are listed in Table S2.
Immunoglobulin G (IgG) was used as a negative control for the primary antibody. A final concentration of 1 μg/mL DAPI (Beyotime, C1002) was included to stain nuclei. Fluorescent images were captured with a confocal microscope (LSM800, Carl Zeiss). Images were further processed with ZEN-2012SP2-blue software. Numbers of gonocytes and the area corresponding to testis cross-section were measured using ImageJ software. To minimize the difference caused by different testicular positions, the gonocytes were shown as unit area (mm 2 ).

| RNA extraction and quantitative RT-PCR (qRT-PCR)
Total RNA was extracted from mouse testes using TRIzol reagent (MRC, TR118) in accordance with the manufacturer's instructions.
Expression levels were normalized to the geometric mean of Gapdh and Actin. The relative expression level of candidate genes was calculated using the formula 2 -ΔΔCT as described. 26 The primers used are listed in Table S1. All experiments were repeated three times.

| Isolation of single testicular cells by fluorescence-activated cell sorting (FACS), single-cell reverse transcription and amplification
The testes at P2 were collected, de-capsulated and digested with TrypLE™ Express Enzyme (Thermo Fisher, 12605010). The cell suspensions were filtered through a 40 μm cell strainer (Corning, 22363547), and cells were collected by centrifugation at 300 × g

| Ddx5 expression during testicular development
To explore the roles of Ddx5 during male germ cell development, Ddx5 expression was evaluated in testes of postnatal mice, collecting every 5 days after birth according to the developmental process of spermatogenic cells. 28 The mRNA level of Ddx5 was increased accompanying mouse age, with more prominent expression at P20 and P25, similar to that of Mvh, a germ cell-specific marker ( Figure 1A,B).
Moreover, we performed western blotting for DDX5 and different markers of germ cells, respectively. As expected, DDX5 protein level was gradually increased from P0 to P90, and increasing levels as spermatocytes and round spermatids developed ( Figure 1C).
Western blotting of spermatogenic cells isolated using unit-gravity sedimentation indicated that DDX5 was predominantly expressed in spermatocytes and round spermatids, but not in elongating spermatids ( Figure 1D).
We further examined the expression and subcellular localization of DDX5 during earlier male germ lineage development by performing immunofluorescent staining ( Figure 1E). Transgenic mice which express the Oct4 promoter-driven GFP (hereafter referred to as Oct4-GFP) were used to mark germ cells. Our data indicated that DDX5 had lower level in PGCs at E9.5, but prominently expressed in gonocytes at E13.5 and E18.5 following sexual differentiation, as well as in postnatal germ cells ( Figure 1E). Moreover, we observed that DDX5 expression was relatively weak in somatic cells from embryonic period. DDX5 was expressed in the nucleus. Together, these data indicate that DDX5 is not only abundant in the nucleus of germ cells after birth, but also in gonocytes in prenatal period, suggesting that Ddx5 may regulate fate determination of germ cells at embryonic stages.

| Germ cell-specific Ddx5 knockout leads to infertility
DDX5 is a multi-functional protein and is involved in several cellular processes. 29,30 To further explore the role of Ddx5 in male ger- A and B, qRT-PCR to analyse the expression of Ddx5 and germ cell marker Mvh in the testes of mice at different stages (P0 to P90). Expression levels were normalized against geometric mean of Gapdh and Actin. Error bars correspond to means ± SD (*P < .05, **P < .01, ***P < .001). C, Western blotting to analyse DDX5 protein in the testes of mice at different stages (P0 to P90). AKAP3 was used as elongating spermatid marker. MIWI was used as spermatocyte and round spermatid marker. SYCP3 was used as spermatocytes marker. β-ACTIN was used as a loading control. D, Western blotting to analyse protein level of AKAP3, MIWI, SYCP3 and DDX5 in the fractionated testicular spermatocytes (SC), round spermatids (RS) and elongating spermatids (ES). β-ACTIN was used as a loading control. E, Immunofluorescent staining of DDX5 expression in germ cells from wild-type mice at E9.5 to P90. Germ cells were labelled with Oct4-GFP (green) in E9.5 and E13.5 gonad sections. Spermatogonia were labelled with anti-PLZF antibody in testis section at P10. All germ cells except elongating spermatids were labelled with anti-TRA98 antibody in testis section at P90. DNA was stained with DAPI (blue). Scale bars represent 20 μm , Ddx5 +/and Ddx5 -/adult mice. β-ACTIN was used as a loading control. C, Morphological analysis of testes in WT, Ddx5 +/and Ddx5 -/adult mice. Scale bar represents 5 mm. D, Assessment of testis to body weight ratio (mg/mg) from WT, Ddx5 +/and Ddx5 -/adult male mice. Error bars correspond to means ± SD. P = .4428 between WT and Ddx5 +/-, P = 3.644e-05 between Ddx5 +/and Ddx5 -/and P = 7.396e-07 between WT and Ddx5 -/-(***P < .001, n = 6). E, H&E staining of testis sections from WT, Ddx5 +/and Ddx5 -/adult mice.

| Ddx5 deletion results in the depletion of germ cells in adult male mice
To assess the reason of infertility in Ddx5-deficient male mice, we observed that the testes of Ddx5 -/mice were significantly smaller and lighter than those of wild-type and Ddx5 +/mice ( Figure 2C,D).
Haematoxylin and eosin (H&E) staining showed that although the architecture of seminiferous tubules largely preserved, all tubules were atrophy and empty in the testes of Ddx5 -/mice compared with multiple layers spermatogenic cells in the testes of wild-type and Ddx5 +/mice ( Figure 2E). In addition, Ddx5 -/mice resulted in loss of mature sperms in the cauda epididymides ( Figure 2F,G).
Consistently, germ cell-specific markers were not detectable in Ddx5 -/mice ( Figure 2B and Figure S1C). Meanwhile, flow cytometry analysis identified stage-specific subpopulations of spermatogenic cells in wild-type, but not in Ddx5 -/mice ( Figure S1B). The above data demonstrate that Ddx5 -/adult male mice have no capacity to produce germ cells.

| Ddx5 deletion results in a Sertoli cell-only phenotype in adult male mice
Although seminiferous tubules were empty in Ddx5 -/adult mice, there was still one layer of cells in peripheral regions of seminiferous tubules. It has been reported that Sertoli cells and SSCs are located in the basement membrane. 31 To examine the identity of the cells remained in seminiferous tubules, we performed immunofluorescent staining and observed that Ddx5 -/mice resulted in loss of TRA98positive germ cells compared with wild-type and Ddx5 +/mice, but only SOX9-positive Sertoli cells left ( Figure 3A,B). The phenotype that Ddx5 -/adult male mice were completely devoid of germ cells is consistent with Sertoli cell-only syndrome associated with human infertility. SSCs could not maintain in Ddx5 -/adult mice, we wondered whether Ddx5 affects the source of SSCs. We compared the testes in wild-type and Ddx5 -/mice at P6. Immunofluorescent staining with TRA98 indicated that SSCs were completely depleted in Ddx5 -/seminiferous tubules ( Figure 3C). Therefore, these results indicate that Ddx5 regulates male germline development at the early stage, even before spermatogonia pool is formed at P6.

| Germ cells are gradually lost in testes of Ddx5 -/neonates
Mvh-Cre transgenic mice express recombinase faintly at E15 and strongly at E18. To explore the effect of Ddx5 on germ cells during the embryonic stage, we collected testes at E18.5 and found that Ddx5 knockout has no effect on the number of gonocytes ( Figure S2A,B). Therefore, we focused on the effect of Ddx5 loss on postnatal days. We observed that Oct4-GFP-positive germ cells were reduced in the testes of Ddx5 -/mice at P0 and P2, and were totally absent at P6. In contrast, no change was observed in ovaries between wild-type and Ddx5 -/mice ( Figure 4A). Additionally, immunofluorescent staining revealed that Ddx5 -/mice had significantly reduced numbers of gonocytes in seminiferous tubules at P0 ( Figure 4B), and about 42% gonocytes were diminished in the testes of Ddx5 -/mice ( Figure 4C). We also analysed the percentage of Oct4-GFP-positive cells using flow cytometry at P2. Noticeably, Ddx5 -/mice had only 0.11% gonocytes in all testicular cells while wild-type mice had 1.31% ( Figure 4D). Taken together, these results demonstrate that gonocytes are gradually reduced and then completely depleted from E18.5 to P6 upon Ddx5 ablation, suggesting that Ddx5 is indispensable for the survival of gonocytes in postnatal testes.

| Single-cell RNA sequencing analysis reveals the effects of Ddx5 knockout on gene expression in gonocytes
To investigate the regulatory mechanism of Ddx5 in gonocytes, scRNA-seq analysis was performed on dissociated testicular cells, which were isolated from whole testes of wild-type and Ddx5 -/mice at P2 ( Figure 5A). We obtained 6328 and 4950 single-cell transcriptomes in the testes of wild-type and Ddx5 -/mice after quality control, respectively ( Figure S3A). Dimensionality reduction and visualization using uniform manifold approximation and projection (UMAP) 32 Table S3). The relative proportion of each cluster weighed by the percentage of total cells in both genotypes showed that the ratio of germ cells in the testes was dropped dramatically from 1.86% to 0.24% in Ddx5 -/mice compared with wild type, while the other cell clusters remained unchanged ( Figure S3D).

F I G U R E 3 Lack of germ cells in
Ddx5 -/testes. A, Immunofluorescent staining of DDX5 and TRA98 (germ cells) in testis sections from WT, Ddx5 +/and Ddx5 -/mice at P90. Null tubules were marked with asterisk. B, Immunofluorescent staining of DDX5 and SOX9 (Sertoli cells) in the testes from WT, Ddx5 +/and Ddx5 -/mice at P90. C, Immunofluorescent staining of DDX5 and TRA98 in testis sections from WT and Ddx5 -/mice at P6. Null tubules were marked with asterisk. DNA was stained with DAPI. Scale bars represent 20 μm To further elucidate the effect of Ddx5 loss on cellular responses in the testes, germ cells (gonocytes) were selected for further analysis. 54 germ cells (47 from wild type and 7 from Ddx5 -/-) were subjected to principal component analysis (PCA) based on the highly variable genes (n = 3817). We found a segregated distribution for wild-type and Ddx5 -/cells along the PC1 ( Figure 5D).

| Ddx5 maintains survival of gonocytes through regulating the expression level of genes related to GDNF pathway
After Ddx5 knockout, the most majority of differentially expressed genes (DEGs) were down-regulated, and many of them were critical  Figure 5G). GDNF is a growth factor secreted by testicular somatic cells, which is necessary for spermatogonia proliferation and survival. [33][34][35] GDNF forms a complex with GFRα1, a cell surface receptor of the gonocytes, 36 and subsequently binds to the transmembrane RET tyrosine kinase and activates intracellular signaling pathways ( Figure 6A). Our scRNA-seq data showed that the GDNF pathway-related genes were down-regulated in Ddx5 -/male mice ( Figure 5G). To confirm the dysregulation of GDNF pathway following Ddx5 knockout, we performed single-cell qRT-PCR analysis on Oct4-GFP-positive male germ cells, which were isolated from the testes of both wild-type and Ddx5 -/males at P2. Our data showed that the mRNA levels of Ddx5, Gfra1, Ret, Gdnf, Pou3f1, Fyn, Id4 and Etv5 were dramatically decreased in Ddx5-deficient gonocytes ( Figure 6B), which are consistent with the scRNA-seq results.
Moreover, immunofluorescent staining of cross-sections showed that GFRα1 protein level was significantly reduced on the surface of gonocytes in Ddx5 -/testes compared with that in wild type ( Figure 6C). Meanwhile, we barely observed proliferating gonocytes in the newborn Ddx5 -/mice by immunofluorescent staining with anti-Ki67 antibody ( Figure 6D). We also examined possible gonocyte apoptosis through TUNEL staining and found that apoptotic cells did not increase in Ddx5 -/neonates ( Figure S2C). In summary, these data suggest that Ddx5 knockout disrupts the GDNF signaling pathway and might further lead to phenotypic defects in proliferation and survival of gonocytes.

| D ISCUSS I ON
RNA-binding protein DDX5 is widely known to participate in various aspects of RNA metabolism. 30 In this study, we report that Ddx5 is required for male germ cell development and Ddx5-deficient male mice are infertile due to loss of gonocytes in neonates. The phenotype of Ddx5 -/mice is similar to the Sertoli cell-only syndrome (SCO) in humans, which shows the complete absence of germ cells in human testicular tissues. 37 While the potential mechanism of SCO is still unclear, DDX5 loss might contribute to SCO in humans.
During the process of our study, another independent laboratory reported UBC-Cre ERT2 mediated Ddx5 knockout affects the maintenance of spermatogonia. 23 Ddx5 knockout in spermatogonia at 8 to 12 weeks led to completely devoid of germ cells in seminiferous tubules. 23 Here, we generated Ddx5 conditional knockout mice by using Mvh-Cre to delete Ddx5 from embryonic period and found that the infertility of adult male mice after Ddx5 loss was due to low survival of gonocytes. Based on theirs and our study, both of us acquired the similar Sertoli cell-only phenotype and male infertility in adult mice. However, compared with the discoveries from Legrand et al, 23 we mainly focused on the roles of Ddx5 in gonocytes at the early stage from E18.5 to P6.
Indeed, our data clearly indicated that Ddx5 was important for the survival and transition of gonocytes. However, it remains unknown how Ddx5 regulates gene expression of gonocytes. DDX5 is widely known to be involved in several steps of RNA metabolism and ribosome biogenesis, 30,38,39 in which many genes were dysregulated in Ddx5 -/gonocytes. Further, the significantly enriched pathways of down-regulated genes were cell cycle-related pathways, which were consistent with the findings in spermatogonia. 23 Moreover, many genes involved in male germ cell development were significantly down-regulated in gonocytes of Ddx5 -/mice, including almost all of the genes related to GDNF-GFRα1-RET signaling pathway (eg, Gfra1, Ret, Etv5 and Id4) and spermatogenesisrelated genes (eg, Zbtb16, Sall4, Gtsf1 and Dmrt1). Among them, ZBTB16 and SALL4 are two important transcription factors for self-renewal and differentiation of spermatogonia. [40][41][42][43] In addition, both Gtsf1 and Dmrt1 are vital for spermatogenesis beyond the early meiotic phase. 4 4-47 Therefore, additional mechanisms for Ddx5 in regulating survival of neonatal mouse gonocytes could be pursued by investigating how Ddx5 modulates the expression of Zbtb16, Sall4, Gtsf1 and Dmrt1.
In conclusion, the findings presented here demonstrate that Ddx5 is required and essential for male fertility. Furthermore, Ddx5 is indispensable for survival and transition of gonocytes in male neonates at least through regulating the genes that associated with cell cycle and GDNF pathway.

CO N FLI C T O F I NTE R E S T
The authors declare no competing interests.

AUTH O R CO NTR I B UTI O N S
HY and QX initiated the study and designed the experiments. QX