Introduction

Valves are critical anatomical structures that ensure unidirectional fluid flow in the circulatory and lymphatic systems. They function passively to prevent backflow, maintaining efficient fluid dynamics within vessels (Moore and Bertram, 2018; O’Donnell and Yutzey, 2020). When valve integrity is compromised, it distorts the luminal architecture, reduces intravascular pressure, and may result in regurgitation.

In the mammalian testis —an organ characterized by active, oscillatory fluid flow (Fleck et al., 2021)— a valve-like structure called the Sertoli valve (SV) emerges at the termini of seminiferous tubules (STs) during puberty (Aiyama et al., 2015; Nagasawa et al., 2018; Imura-Kishi et al., 2021). The SV comprises a specialized subset of Sertoli cells that exhibit elevated levels of phosphorylated AKT and STAT3. These cells extend cytoplasmic projections enriched with acetylated tubulin-positive (ace-TUB⁺) microtubules into the rete testis (RT) lumen, forming a dynamic and structurally distinct valve-like barrier (Russell, 1993). Recently, we identified SOX17, a transcription factor expressed in RT epithelial cells, as essential for SV formation and for spermiogenesis in the ST. In mice with RT-specific conditional knockout of Sox17 (Sox17 cKO), the SV fails to form properly before puberty, allowing retrograde fluid flow from the RT into the convoluted ST (Uchida et al., 2022). These findings imply the critical role of the RT-SV structure in maintaining proper fluid dynamics essential for spermatogenesis.

Interestingly, the SV epithelium provides a specialized niche for undifferentiated spermatogonia (A_undiff) and immature/proliferative Sertoli cells (Aiyama et al., 2015; see reviews by Figueiredo et al., 2021; Seco-Rovira et al., 2025). The basal compartment of the SV epithelium supports self-renewal of A_undiff spermatogonia by maintaining high expression of growth factors such as FGFs (Takashima et al., 2015; Kitadate et al., 2019; Yang et al., 2021; also see the review by Hofmann and McBeath, 2022) and GDNF (Meng et al., 2000; Uchida et al., 2016), together with by locally suppressing retinoic acid (RA)-induced differentiation (Imura-Kishi et al., 2021). Furthermore, SV Sertoli cells display Cyclin D1-positive states in mice and hamsters, and in rats and hamsters, some of these cells proliferate and contribute to the convoluted ST even at adult stage (Aiyama et al., 2015; Figueiredo et al., 2019), suggesting that the SV acts as a niche for proliferative Sertoli cells in hamsters and rats. Specification of these SV Sertoli cells is thought to be regulated by paracrine signals such as RSPO1, FGFs etc. from adjacent SOX17-positive RT cells (Imura-Kishi et al., 2021; Uchida et al., 2022; see review by Han et al., 2026). However, how SOX17-positive RT cells regulate SV formation in the proximal region of rodent testes remains unclear.

To address this question, we generated a novel transgenic (Tg) mouse line that exhibits excess SV formation by the Sertoli cells ectopically expressing Sox17 under the control of the human AMH promoter (SOX17⁺ Sertoli cells). Such SOX17⁺ Sertoli cells enhance the expression of WNT4 and RSPO1 in SV and RT regions, respectively. Moreover, we showed a partial rescue by the SOX17⁺ Sertoli cells in the Sertoli valve formation and aberrant spermiogenesis in the cKO testes with Sox17-null RT. These findings highlight the importance of SOX17-downstream paracrine actions in the SV and RT for the proper valve formation, rather than the SOX17 roles in the RT formation in itself.

Results

Overexpression of SOX17 in Sertoli cells induced hyperplasia of the SV region

To investigate the downstream paracrine effects of Sox17, we generated AMH-Sox17 transgenic (Tg) mice that ectopically express Sox17 in Sertoli cells under the control of the human anti-Müllerian hormone (AMH) promoter (Fig. 1A, S1). Among three Tg founder lines, one male (#27) exhibited infertility and was smaller in size of testes compared to his wild-type littermates. In this Tg male, the Sertoli valve (SV) structure appeared to be expanded in the terminal segments of the seminiferous tubules (Fig. 1B, C). Additionally, abnormal aggregates of Sertoli cells were frequently observed on the luminal side of some seminiferous tubules (indicated by asterisks in Fig. 1B). Immunostaining for SOX17 and E-cadherin in the #27 Tg testis revealed ectopically high levels of SOX17 expression in Sertoli cells throughout both the convoluted seminiferous tubules and the SV-like regions, similar to those observed in the RT epithelium (Fig. 1D). This ectopic, high-level expression of SOX17 in Sertoli cells appears to disrupt their normal function across the entire testis, ultimately leading to infertility in the #27 Tg male.

Fig. 1.

Establishment of the AMH-Sox17 Tg Line (#26) with Excess SV Formation

One Tg male (#26) was fertile, and the AMH-Sox17 transgene was successfully transmitted to its offspring (Fig. S1A, B). The morphology and weight of the testes in Tg mice were comparable to those of their wild-type littermates. Although the AMH-Sox17 transgene was moderately activated in a subset of Sertoli cells during both fetal and postnatal stages (Fig. S1C, D), an expansion of the acetylated tubulin-positive Sertoli valve (SV) region was observed in the proximal testis of this line (Fig. 2A), similar to that seen in the infertile #27 Tg male with a more severe phenotype (Fig. 1D). Because this Tg line could be stably maintained by backcrossing beyond the 8th generation, line #26 was used for all subsequent experiments (hereafter referred to as “Tg”).

Fig. 2.

In this Tg line, no appreciable alterations in the RT-SV region are observed during the immature stages up to 1 week after birth. However, following sexual maturation, immunostaining for acetylated tubulin and Cyclin D1 clearly marks an expanded SV region at the terminal segments of the seminiferous tubules (Fig. 2A). This expansion was further confirmed by electron microscopy, which revealed multiple cytoplasmic layers containing bundled microtubules extending distally from Sertoli cells into the rete testis (RT) lumen (blue asterisks, Fig. 2B). At the boundary between the SV and RT regions, SV Sertoli cells were supported by a multilayered basal lamina, consistent with previous reports (Fig. S2A; Tainosho et al., 2011; Uchida et al., 2022). Notably, in the Tg testes, the SV Sertoli cells displayed vacuole-like intercellular spaces in the basal portion of the terminal segment, and some appeared detached from the basal lamina (red asterisks, Fig. 2B). Furthermore, these cells were often found protruding into the RT lumen (designated as “pSV” in Fig. 2A, 2B, S2B), where necrotic or damaged cells were observed at the distal tips of the protrusions. These findings suggest that the protruding SV Sertoli cells may undergo necrosis and be sloughed off into the RT lumen.

Partial rescue of defective spermatogenesis and infertility in RT-specific Sox17 cKO mice by AMH-Sox17 Tg

In RT-specific Sox17 conditional knockout (cKO) testes, deformation of the Sertoli valve (SV) leads to backflow of rete testis (RT) fluid into the seminiferous tubules, resulting in defective spermiogenesis and male infertility (Uchida et al., 2022). In contrast, AMH-Sox17 Tg mice exhibit excessive SV formation due to ectopic expression of SOX17 in SV Sertoli cells; however the AMH-Sox17 transgene is not activated in the RT epithelium. To evaluate whether this ectopic expression could rescue the SV defects in the RT-specific Sox17 cKO background, we examined the testis phenotype of double mutant mice (Sox17 cKO; AMH-Sox17 Tg, referred to as cKO;Tg) (Fig. 3). By crossing Tg; Sox17^f/f males with SF1-Cre;Sox17^f/f females, we generated four genotypes for comparison: control (Sox17^f/f), cKO (SF1-Cre;Sox17^f/f), Tg (Tg; Sox17^f/f), and cKO;Tg (Sf1-Cre;Sox17^f/f; Tg). As expected, control and Tg males showed normal spermatogenesis and the presence of spermatozoa in the epididymis. In contrast, cKO males exhibited abnormal spermiogenesis with a complete absence of epididymal spermatozoa (Fig. 3A), consistent with previous findings (Uchida et al., 2022). Interestingly, cKO;Tg double mutants showed improved spermatogenesis, with the presence of epididymal spermatozoa and a partial restoration of fertility compared to cKO males. Quantitative assessment of spermatogenesis using a 10-point scoring system (7 points for seminiferous tubule integrity + 3 points for presence of advanced germ cells) confirmed a significant improvement in cKO;Tg males (Fig. 3B) related to cKO males. This was further supported by an increased ratio of elongated spermatids in these cKO;Tg testes. Notably, epididymal spermatozoa were detected in 2 of 7 cKO;Tg individuals at 8 weeks of age. Furthermore, mating trials using four independent cKO;Tg males resulted in successful offspring production in 2 out of the 4 cKO;Tg males.

Fig. 3.

These results demonstrate that ectopic SOX17 expression in Sertoli cells partially rescues the structural and functional defects of the SV caused by the absence of SOX17 in the RT, thereby improving spermatogenesis and restoring fertility in otherwise infertile males.

Ectopic SOX17 expression in Sertoli cells rescues SV deformation in RT-specific Sox17 cKO testes

In RT-specific Sox17 conditional knockout (cKO) testes, disruption of the SV formation leads to defective spermiogenesis and male infertility (Uchida et al., 2022), which could be partially rescued by the ectopic SOX17 expression in Sertoli cells, as described above (Fig. 3). To assess whether such ectopic SOX17 expression could rescue the SV defects caused by Sox17 deficiency in the RT, we examined the RT-SV phenotypes of Sox17 cKO;AMH-Sox17 Tg double mutant (cKO;Tg) mice (Fig. 4). Double immunostaining for SOX17 and SOX9 confirmed that in cKO;Tg testes, SOX17 expression was restricted to Sertoli cells in the terminal segments of the seminiferous tubules, and was absent from the RT epithelium—opposite to the expression pattern seen in control testes, where SOX17 was confined to RT epithelial cells (Fig. 4A). In Tg testes, both RT epithelial cells and SV Sertoli cells were SOX17-positive. An expanded region of SOX17/SOX9 double-positive SV Sertoli cells was evident in Tg testes, while this region appeared to be reduced in cKO;Tg testes, suggesting partial normalization of the SV region (yellow area, Fig. 4A).

Fig. 4.

Furthermore, acetylated tubulin-positive valve structures were properly formed in cKO;Tg testes, and normal spermatogenesis including SCP3-positive spermatocyte and Hsp70-positive elongated spermatids was observed in the convoluted seminiferous tubules (Fig. 4B). Quantitative analysis of SV Sertoli cell density revealed that the number of SOX9-positive nuclei in the RT-SV region was significantly increased in Sox17 Tg testes and significantly decreased in Sox17 cKO testes compared to control testes (Fig. 4C). Interestingly, in the cKO;Tg double mutant testes, SV Sertoli cell density was significantly higher than that in cKO testes, showing an intermediate level between Tg and cKO testes, comparable to that of controls (Fig. 4C). These findings indicate that ectopic SOX17 activity in SV Sertoli cells can rescue the structural defects of the SV in RT-specific Sox17 cKO testes, in a Sox17 dose-dependent manner.

Previous studies have shown that Rspo1 is highly expressed in RT epithelial cells in wild-type mice, but its expression is markedly reduced in cKO testes (Uchida et al., 2022). Along with the known expression of Wnt4 in SV Sertoli cells, we finally examined the expression patterns of Rspo1 and Wnt4 in the RT-SV region across different genotypes using RNAscope in situ hybridization (Fig. 5). Fixation, sectioning, and in situ hybridization were performed under identical conditions on serial sections. Expression profiles of RT-specific Rspo1 and SV-specific Wnt4 were generally comparable among control, Tg, cKO, and cKO;Tg mice (Fig. 5A), which confirm no ectopic Rspo1 expression in SOX17⁺ SV Sertoli cells. In Tg testes, Wnt4 expression appeared to be enhanced in the SV region, consistent with the expanded SV phenotype in these mice (Fig. 2). Notably, RT-specific Rspo1 expression was increased in Tg testes and slightly elevated in cKO;Tg mice relative to cKO mice, suggesting a potential paracrine mechanism, in which ectopic SOX17 activity in Sertoli cells might influence Rspo1 expression in adjacent RT epithelial cells.

Fig. 5.

Discussion

A previous study using RT-specific Sox17 cKO mice demonstrated that Sox17 expression is critical for Sertoli valve (SV) formation and proper spermiogenesis. Deletion of Sox17 in the rete testis (RT) disrupted the specification of modified Sertoli cells that form the SV, leading to backflow of RT fluid and abnormal detachment of immature spermatids within the seminiferous tubules (Uchida et al., 2022; Fig. 6). These findings established the essential role of SOX17⁺ RT epithelia in SV formation and clarified the physiological function of the SV in modulating fluid dynamics within the seminiferous tubules in vivo.

Fig. 6.

In the present study, we demonstrated that ectopic expression of SOX17 in Sertoli cells induces excessive SV formation in the seminiferous tubules, along with increased Rspo1 expression in the RT. Moreover, this ectopic SOX17 expression partially rescued epididymal sperm production and restored fertility in cKO;Tg males (Fig. 6). These results suggest that SOX17 activity in the RT can be, at least in part, functionally substituted by its ectopic expression in Sertoli cells. The observed fertility recovery in cKO;Tg males, even in the absence of SOX17 in RT cells, implies that the primary role of SOX17 in the RT is restricted to SV formation in the terminal segments and is not essential for sperm transport and maturation in the RT-to-epididymis pathway in mice.

Ectopic SOX17 expression in Sertoli cells may mimic paracrine signaling normally mediated by SOX17⁺ RT epithelia in the SV region. One possible explanation is that RT epithelial cells and Sertoli cells share lineage similarities (Malolina and Kulibin, 2019; Uchida et al., 2022; Sato et al., 2026). Several key transcription factors of Sertoli cells —including WT1, NR5A1, SOX9, GATA4, and, in part, DMRT1— are also expressed in the RT epithelia (Aiyama et al., 2015; Malolina and Kulibin, 2019; Kulibin and Malolina, 2022; Uchida et al., 2022; Malolina et al., 2023). During sexual maturation, ectopic SOX17 expression in Sertoli cells may induce paracrine and/or autocrine signaling programs resembling those of RT cells, thereby promoting hyperplasia in Tg testes and facilitating SV recovery in cKO;Tg testes.

The Sox17 transgene is ectopically activated in Sertoli cells not only near the SV but also throughout the testis, whereas the SV phenotype is restricted to the RT-adjacent region of Tg testes (Figs. 1C, 2A, 4A). One possible explanation for this spatial discrepancy is that SOX17 may act on RT-lineage Sertoli-like cells located in the terminal segments adjacent to the RT. A recent study by Mayère et al. (2022) reported that PAX8⁺ RT progenitor cells, referred to as early supporting-like cells, contribute to the proximal region of the seminiferous tubules, suggesting heterogeneity in the origin of Sertoli cells near the RT during embryogenesis, even though PAX8 expression in fetal Sertoli cells is downregulated shortly after testis differentiation. It is therefore possible that ectopic SOX17 expression in these rare supporting-like cells may induce paracrine signaling similar to that of RT epithelia, thereby leading to SV phenotypes in Sertoli cells near the RT in Tg mice.

Interestingly, Rspo1 and Wnt4 signaling is activated in the RT-SV region (Uchida et al., 2022; this study). In Wnt4-null embryos, the RT region is underdeveloped at the seminiferous tubule–RT junction, leading to a failure in connecting the RT and seminiferous tubules, similar to the rete ovary (RO) phenotype (Mayère et al., 2022). These findings suggest that Wnt4–Rspo1 signaling is essential for proper connection of the terminal segment of seminiferous tubules to the RT and efferent ducts, which is critical for the integrity of the male reproductive tract. Our in situ hybridization analyses showed enhanced Rspo1 and Wnt4 expression driven by SOX17⁺ Sertoli cells in Tg mice (Fig. 4), which likely contributed to the rescue of SV formation observed in cKO;Tg mice (Fig. 5). The presence of protruding SV cells expressing Cyclin D1 supports the idea that SOX17-mediated paracrine activity may involve localized Rspo1-Wnt4 signaling during SV formation. Considering the potential involvement of other signaling pathways such as FGFs and TGFs, further studies are warranted to delineate the specific roles of individual paracrine factors in the RT–SV region prior to SV formation.

In conclusion, by generating a novel AMH-Sox17 Tg mouse line, we demonstrate that the size of the SV structure in the terminal segment of the seminiferous tubules can be experimentally manipulated. Furthermore, SV formation was successfully rescued in Sox17 cKO testes by introducing the AMH-Sox17 transgene into RT-specific Sox17 cKO mice. These genetically engineered models — exhibiting SV overdevelopment or restoration — provide valuable tools for investigating paracrine signaling and elucidating the biological significance of the SV and RT in the male reproductive system of mammals.

Materials and Methods

Animals

AMH-Sox17 Tg mice were generated by pronuclear injection of the plasmid-based transgene into the C57BL6 background, as described previously (Shinomura et al., 2014). The transgene was prepared by inserting the entire coding region of the Sox17 gene (Kanai-Azuma et al., 2002) downstream of the hAMH promoter-driven cassette (Fig. 1A). We obtained three independent F0 Tg males (lines #13, 26, and 27), among which a fertile mouse line #26 with a modest phenotype was mainly used in this study. F0 males of line #27 were infertile, possibly due to excessive detachment of the SV-like Sertoli cells with high SOX17 expression. The RT-specific Sox17 cKO mice (i.e., SF1-Cre;Sox17^flox/flox mice; Dhillon et al., 2006; Kim et al., 2007) were produced from the Jackson Laboratory (Uchida et al., 2022). The SF1-Cre:Sox17^flox/flox;AMH-Sox17 Tg double mutant (cKO;Tg), AMH-Sox17 Tg;Sox17 ^flox/flox (Tg), SF1-Cre;Sox17^flox/flox (cKO) and Sox17 ^flox/flox (control) mice were mainly generated by crossing the AMH-Sox17 Tg; Sox17 ^flox/flox males with SF1-Cre:Sox17 ^flox/flox females. These mutant males at 4 and 8 weeks of age were used for the subsequent phenotypic analyses.

All animal experiments were performed following the Guidelines for Animal Use and Experimentation at the University of Tokyo, and all experimental procedures performed herein were approved by the Institutional Animal Care and Use Committee, Graduate School of Agricultural and Life Sciences, University of Tokyo (approval IDs: P13-762, P13-764, P22-147). The PCR list for the genotyping was shown in Table S1.

Histology and immunohistochemistry

Testes were fixed in 4% PFA at 4°C overnight, dehydrated through an ethanol series, and embedded in paraffin. The paraffin sections (4 μm thickness) were subjected to hematoxylin and eosin (H&E) staining, PNA lectin (acrosome) staining and immunohistochemistry. For immunohistochemistry, the sections were incubated overnight at 4°C with the primary antibodies listed in Table S2. Immunoreactions were visualized using Alexa fluor-488/555/594/680-conjugated secondary antibodies (Abcam). Samples were analyzed by fluorescence microscopy (BX51N-34-FL-2) or Leica TCS SP8 confocal laser microscopy. Testes from at least four animals in each group were examined, and tissue sections from cKO, Tg and cKO;Tg and their littermate controls were processed in parallel to ensure the reproducibility of the results.

Electron microscopy

Testes were trimmed and fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) overnight, post-fixed with 0.5% OsO4 suspended in 0.1 M phosphate buffer (pH 7.3) for 30 min, and dehydrated through a graded ethanol series. After passing through propylene oxide, the tissues were embedded in Epon 812. Ultrathin sections were cut, stained with uranyl acetate and lead citrate, and observed by TEM (JEM-1010; JEOL).

RNA in situ hybridization

Testes were fixed in 10% formalin, dehydrated through an ethanol series, and embedded in paraffin. The paraffin sections (4 μm thickness) were processed for RNA in situ detection using the RNAscope 2.5HDAssay Red Manual (ACDBio) as described previously (Imura-Kishi et al., 2021; Uchida et al., 2022). The probes used in this study are listed in Table S3.

Evaluation of Spermatogenesis score

To evaluate spermatogenesis in mouse testes, we performed the fluorescence staining with PNA lectin (a marker for acrosome; Nakata et al., 2015) on tissue sections (Control: n = 4; cKO: n = 4; Tg: n = 5; cKO;Tg: n = 5 males). Fluorescent images of entire testis sections were acquired using a confocal microscope (TCS SP8 Leica Microsystems) with the Tiling function in LAS X software. A two-axis scoring system was applied: Group A categorized the morphological condition of the seminiferous tubules, and Group B assessed the most advanced germ cell stage present within each tubule. Scores were assigned as maximun 10 points as follows: Group A evaluated the morphological condition of the seminiferous tubules, while Group B assessed the most advanced germ cell stage present in each tubule. Scores were assigned with a maximum of 10 points (Group A [tubule condition]: normal [7 points]; disrupted spermatocyte stage [5 points]; partial cell dropout or loss [3 points]; Sertoli cells only [0 points]) plus Group B [most advanced germ cell type]: elongated spermatids [3 points]; round spermatids [2 points]; spermatocytes [1 point]; spermatogonia only [0 points]) (see Fig. 3B).

Quantitative morphometric analyses

The number of Sertoli cells in the SV region was analyzed by using SOX9/E-cad/DAPI immunostained sections. In brief, the nuclear density of the SOX9-positive/E-cad-negative cells within RT-SV region was counted in each terminal segment of the seminiferous tubule (~100 μm from the edge of the RT; Fig. 4C). Only the SV sectioned at the mid-sagittal plane was subjected to testicular section analysis.

Statistical analysis

For comparisons among four genotype groups, one-way ANOVA followed by Tukey’s multiple comparisons test was performed. Quantitative data are presented as means ± standard deviation (SD) and shown as bar graphs with dot plots. Statistical analysis was performed using GraphPad Prism (v.10.4.2; GraphPad Software, San Diego, CA, USA). P < 0.05 was considered indicative of significance, and levels of significance are shown as *P < 0.05 and **P < 0.01. In addition, single comparisons between two groups, as shown in Fig. S1A, were were analyzed using a two-tailed Student’s t-test (data are presented as means ± standard error of the mean).

Data availability

All data generated or analyzed during this study are included in the manuscript and supporting files. Source data for all quantitative analyses are provided with the figures. No large-scale datasets were generated. Additional information is available from the corresponding author upon reasonable request.

Acknowledgements

The authors wish to thank Y. Hirate, and Y. Kuroda for their technical advice and support; Ikuno Oike for her secretarial assistance. The authors acknowledge Y. Akimoto, S. Matsubara, and J. Hayakawa for their excellent TEM technical support.

Additional information

Author contributions

X.H., A.U., K.N., T.E., A.U., M. K-A., Y.K. designed the study. X.H., A.U., S.L., K.N., K.T., A.Y., M. K-A., A.K.,and R.H. performed the biological experiments. The manuscript was written by X.H., A.U., S.L., and Y.K.

Funding

This work was supported by JSPS KAKENHI Grant Number 19H05241, 20H00445; 21H00227, 24H02036, 24H00537 (for Y. K); 22K07886, 25H01350 (for A.Y.); 22K06022 (for R.H.); 18H02361; 21H02387 (for M. K-A.). JSPS Research Fellowship to X.H. (DC2).

Funding

JSPS KAKENHI (19H05241)

JSPS KAKENHI (20H00445)

JSPS KAKENHI (21H00227)

JSPS KAKENHI (24H02036)

Additional files

Sup Figs1,2, Sup Table S1-3