SOX8 Regulates Permeability of the Blood-Testes Barrier That Affects Adult Male Fertility in the Mouse

ABSTRACT

Sertoli cells provide nutritional and physical support to germ cells during spermatogenesis. Sox8 encodes a member of the high mobility group of transcription factors closely related to Sox9 and Sox10. Sertoli cells express SOX8 protein, and its elimination results in an age-dependent dysregulation of spermatogenesis, causing adult male infertility. Among the claudin genes with altered expression in the Sox8^−/− testes, was claudin-3, which is required for the regulation and maintenance of the blood-testes barrier (BTB). Because the BTB is critical in restricting small molecules in the luminal compartment of the seminiferous tubules, the aim of this study was to analyze the level of tight junction proteins (claudin-3, claudin-11, and occludin) and BTB permeability in Sox8^−/− adult testes. The acetylation level of alpha-tubulin and microtubule organization was also evaluated because microtubules are critical in maintaining the microenvironment of the seminiferous epithelium. Western blot analysis shows that claudin-3 protein is decreased in Sox8^−/− testes. Chromatin immunoprecipitation confirmed that SOX8 binds at the promoter region of claudin-3. Claudin-3 was localized to the Sertoli cell tight junctions of wild-type testes and significantly decreased in the Sox8^−/− testes. The use of biotin tracers showed increased BTB permeability in the Sox8^−/− adult testes. Electron microscopy analysis showed that microtubule structures were destabilized in the Sox8^−/− testes. These results suggest that Sox8 is essential in Sertoli cells for germ cell differentiation, partly by controlling the microenvironment of the seminiferous epithelium.

INTRODUCTION

In adult mammalian testes, spermatogenesis takes place within the seminiferous epithelium, which consists of somatic Sertoli cells and migrating germ cells. Sertoli cells are in close contact with germ cells in the seminiferous epithelium, providing physical and nutritional support for spermatogenesis and male fertility. The seminiferous epithelium is divided into basal and adluminal compartments by tight junctions (TJs) between adjacent Sertoli cells, forming the blood-testes barrier (BTB). The BTB is crucial for spermatogenesis because of its selective permeability to nutrients, growth factors, a variety of ions, and other small molecules [1–6]. During spermatogenesis, Type B spermatogonia undergo the last mitotic division in spermatogenesis to give rise to preleptotene spermatocytes, the first stage in meiosis I, at stage VIII of the epithelial cycle [7], and these cells progress to leptotene spermatocytes in the basal compartment, traversing through the BTB to enter the luminal compartment.

Many of the functions attributed to Sertoli cells have microtubule-dependent components [8]. These include polarized movement of metabolites and binding proteins toward the lumen of the seminiferous tubules, selective targeting of paracrine factors and adhesion molecules to associated germ cell subsets [9–13], structural support [14–16] and positioning of spermatids within the seminiferous epithelium [13, 17–20], which are all dependent on microtubule activity.

Localization of SRY-related HMG-box8 (SOX8) protein in Sertoli cells dynamically changes during spermatogenesis. SOX8 protein is located in both the nucleus and cytoplasm of Sertoli cells within stage I–IX tubules, whereas SOX8 protein is limited into cytoplasm within stage X–XII tubules in the gonad [21]. Genetic loss of Sox8 results in progressive degeneration of the seminiferous epithelium because of disturbed physical interaction between Sertoli cells and the developing germ cells. Clear evidence of the male infertility phenotype due to loss of SOX8 in Sertoli cells appears at 2 mo of age. The young Sox8 homozygous knockout (Sox8^−/−) males show variation in their fertility but no fertile Sox8^−/− males were observed beyond 150 days of age. Sox8^−/− mice have a generalized defect in germ cell placement within seminiferous epithelium. Disturbance of the spermatogenic cycle is most clearly evident as inappropriate germ cell placement [21].

Altered function of Sertoli cells is implied in testicular dysgenesis syndrome [22]. Several integral and peripheral membrane proteins, including the occludins, claudins, and junction adhesion molecules, make up the Sertoli cell TJs [23–31]. Claudin-3, one of the reported 20 different claudin proteins produced in testes, regulates the permeability of the BTB [26, 32–34]. Our earlier transcriptome profiling indicated a number of molecules found in TJs and specialized anchoring junctions were altered in 2-mo-old Sox8^−/− mice testes, suggesting that SOX8 may regulate the assembly or function of Sertoli cell TJs [35]. This led us to investigate the TJ proteins as well as the structure and function of the BTB in adult Sox8^−/− testes. In addition, with the aim of investigating factors that may modulate the microenvironment of the seminiferous tubule, we analyzed the acetylation level of alpha-tubulin as well as microtubule organization, which are critical in maintaining the microenvironment of the seminiferous epithelium in the Sox8^−/− testes.

MATERIALS AND METHODS

Mice

The generation of Sox8 mutant mice and genotyping has been previously described [21, 35]. The National Institute of Environmental Health Sciences Animal Care and Use Committee approved all the animal studies. All the experimental data were collected from a minimum of three animals of each genotype and stage.

Western Immunoblot Analysis

Whole testes were dissected and homogenized in buffer-X [36]. After centrifugation, the supernatant were collected and protein separated by SDS PAGE and transferred to the nitrocellulose membrane. The blots were incubated with the below mentioned antibodies for claudin-3, SOX8 (sc-17338), and alpha-tubulin (monoclonal anti-acetylated alpha-tubulin clone 6-11B-1; Sigma), followed by horseradish peroxidase (HRP)-conjugated anti-mouse immunoglobulin G (IgG) or HRP-conjugated anti-rabbit IgG (GE Healthcare Bio-Sciences). Chemiluminescence was detected with ECL Western blot detection kits (GE Healthcare Bio-Sciences) according to the supplier's recommendations.

Chromatin Immunoprecipitation

The testes were removed from 2-mo-old wild-type (WT) mice. Chromatin immunoprecipitation (ChIP) was performed using SOX8 (H-95) X antibody (sc-20094; TransCruz). ChIP analysis was performed with Stratagene Mx300P and Brilliant SYBR Green quantitative PCR (QPCR) Master Mix. The average cycle threshold amplification values and percentage of sample input were calculated. PCR primers were designed from the promoter region and include claudin-3 PRO-FW: TCAGGCAGGAGCCAACACA, and claudin-3 PRO-RV: GTGAACAGTGGGAAAGAGATTTGA (−1127 to −1050); exon primers include mclaudin3-FWD: TCCAGATGGTGACAGACGACACAC, mclaudin3-REV: GGAAGGGCGAGGTTTCTTTG (+988 to +1132). The primer positions are denoted relative to the downstream of the transcriptional start site (+).

Immunohistochemistry, Immunofluorescent Microscopy, and Imaging

Testes from 2-mo-old WT and Sox8^−/− males were incubated overnight in 4% paraformaldehyde at 4°C for processing of histological samples, and 5-μm paraffin sections were used for immunohistochemistry (IHC) analysis. IHC analysis of mouse tissue was conducted using rabbit anti-claudin-3 (34–1700), rabbit anti-claudin-11 (36–4500), and rabbit anti-occludin (71–1500) (Invitrogen [Zymed], Carlsbad, CA). Staining for claudin-3 was performed using the Discovery XT Protocol and included the 32-min Protease Automated System (Ventana Medical Systems) antigen retrieval and heating with the OmniMap anti-Rabbit Detection Kit. The tissues were subsequently incubated with anti-claudin-3 (1:200), anti-claudin-11 (1:75), and anti-occludin (1:75) for 60 min. For negative control tissue, normal rabbit serum (Jackson ImmunoResearch Laboratories) was used in place of the primary antibody. Staining was visualized using 3,3′-diaminobenzidine chromagen (DakoCytomation). Slides were cleaned with an isopropanol solution prior to scanning. Slides were then scanned using the Aperio Scanscope XT Scanner (Aperio Technologies, Inc.), which uses line scanning technology to capture high-resolution, seamless digital images of glass slides. After scanning, the slides were viewed using the digital slide-viewing program Aperio Imagescope version 11.0.2.716 (Aperio Technologies, Inc.). Images were then captured for manuscript use using Aperio Imagescope. Measurements were taken from these images using a measurement scale that was internally calibrated for each image as it was scanned.

Biotin Tracer Studies

Using a biotin tracer [37], the permeability of the BTB was assessed. The biotin tracer studies were performed as described in Meng et al. [32]. Briefly, we anesthetized 2-mo-old WT and Sox8^−/− mice and injected 50 μl of 10 mg/ml EZ-Link Sulfo-NHS-LC-Biotin (Pierce Chemical Co.) into the interstitium of exposed testes. The testes were removed after 30 min and frozen in dry ice. Paraffin-embedded 5 μm thick sections were prepared, incubated, and treated with anti-streptavidin antibody conjugated to Alexa Fluor 568 (Molecular Probes) for 30 min at room temperature. The sections were rinsed with PBS and mounted in prolonged antifade mounting medium.

Biotin Analysis

An Olympus IX-71 inverted fluorescence microscope equipped with a LUCPlanFLN 20X 0.45NA objective lens was used for image acquisition. Fluorescent excitation was achieved by using an Exfo Excite 120 metal halide lamp with a 560/20-excitation filter. Emission light was collected through a 607/37-emission filter with a Photometrics Cool Snap ES camera. The software used to operate the imaging system was Metamorph for Olympus Basic version 7.5.6.0. A 5-sec exposure was used to capture a 14-bit grayscale image. These images were subsequently analyzed in the Metamorph software. For quantitation, a region of interest (ROI) was drawn and the region measurements drop-in was used to acquire the average intensity within each ROI.

Electron Microscopic Examination

On the day of necropsy, each mouse was euthanatized and then the whole body was perfused via the heart using 4% paraformaldehyde. Once perfused, the right testicle was taken for light microscopy while the left was collected and processed routinely for transmission electron microscopy [38]. Two resin blocks per animal were examined, and at least five digital photomicrographs were taken of each block, each of which illustrates a portion of a seminiferous tubule [39]. The basement membrane is composed of the basal lamina, connective tissue layers, a single myoid cell layer, and an endothelial layer.

RESULTS

Loss of SOX8 Decreased Claudin-3 Protein in the Testes

Because our previous results showed that SOX8 is specifically expressed in Sertoli cells in the testes [40], we analyzed by Western blot analysis the levels of SOX8 protein of neonatal and adult testes. SOX8 appeared relatively constant and likely unchanged throughout the timeframe examined, suggesting that SOX8 is possibly essential for the proper function of Sertoli cells in neonatal and adult animals (Fig. 1A). Claudins are an essential component of TJs and are expressed in a variety of tissues [41–44]. Our published transcriptome analysis previously determined that expression of the claudin gene family is altered in 2-mo-old Sox8^−/− testes, and among the altered claudins, Cldn3 was significantly downregulated in Sox8^−/− testes [35]. We therefore analyzed claudin-3 protein levels in postnatal and adult WT testes. We found that claudin-3 was produced constantly and appeared similar at all ages of WT testes (Fig. 1A). The protein density was evaluated by using ImageJ software, and P values from t-test comparing protein density values of each pair of groups (e.g., younger vs. adult) were not significant {data not shown). We further sought to examine the level of claudin-3 protein in Sox8^−/− versus WT testes. Western blot analysis (n ≥ 3) showed that the claudin-3 protein level appeared to be similar in all the WT testes and significantly decreased for all the examined time points in Sox8^−/− testes compared to WT (Fig. 1B). We decided to further explore how claudin 3 expression may be regulated by SOX8.

FIG. 1.

SOX8 regulates claudin-3 expression. A) Temporal quantitative protein level of SOX8, claudin-3, and an anti-tubulin antibody as a loading control in wild-type (WT) testes. B) Temporal quantitative protein level of claudin-3 in WT testes and Sox8^−/− testes probed with anti-claudin-3 and an anti-alpha-tubulin antibody as a loading control. C) ChIP analysis of SOX8 binding at the claudin-3 (Cldn3) promoter and coding region. ChIP-QPCR shows SOX8 occupancy on the claudin-3 promoter region and coding region compared with immunoglobulin-G (IgG) in adult testes. The error bars are expressed as mean ± SD.

SOX8 encodes a member of the high-mobility group of transcription factors. Sequence analyses of the promoter region of claudin 3 revealed a putative binding site for SOX9, which is a member of the SoxE subgroup along with SOX8 and SOX10. Thus, we speculated that SOX8 might bind to the SOX9-binding sites to regulate Cldn3 expression. We performed a ChIP assay on chromatin prepared from 2-mo-old WT testes using a SOX8 antibody and primers designed from the putative region for SOX-binding and -coding regions. As shown in Figure 1C, the ChIP result showed that SOX8 is associated with chromatin in the Cldn3 promoter region but not a more distant coding region that does not contain the putative SOX8 binding motif. This result is consistent with the view that SOX8 signals may directly influence Cldn3 gene expression.

Claudin-3 Is Localized to the BTB and Diminished in Sox8^−/− Testes

To determine the spatial localization of claudin-3, IHC was performed using an anti-claudin-3 antibody. Claudin-3 staining was present at Sertoli cell TJs in WT testes (Fig. 2, A and B). In contrast, claudin-3 staining was diminished at the Sertoli cell TJs in Sox8^−/− testes (Fig. 2, C and D). However, in the epididymis of the WT and Sox8^−/− animals, claudin-3 staining was strongly detected at the TJs (Fig. 2, E and F). This suggests that SOX8 is critical in Sertoli cells for maintenance of claudin-3 in Sertoli cell, but not epididymis, TJs. Occludin is present in mouse Sertoli cell TJs and appears to be essential for their function in the testes of mice [45]. Our earlier microarray analysis demonstrated that mRNA of claudin-11 (1.4-fold) and occludin (1.4-fold) were mildly upregulated in Sox8^−/− testes [35]. However, IHC shows that claudin-11 and occludin staining were prominent and equivalent at the TJs of Sertoli cells (WT: Fig. 2, G, H, M, and N; Sox8^−/−: Fig. 2, I, J, O, and P) and epididymis (WT: Fig. 2, K and Q; Sox8^−/−: Fig. 2, L and R) of the both WT and Sox8^−/− mice. The level of claudin-11 and occludin indicate that TJ structures were present in the seminiferous epithelium of Sox8^−/− mice testes. However, the diminished level of Claudin-3 suggests that the function of the TJs may be altered or compromised in Sox8^−/− mice.

FIG. 2.

Immunohistochemical analysis illustrates the expression of tight junction (TJ) proteins claudin-3, claudin-11, and occludin on serial sections from wild-type (WT) and Sox8^−/− testes. A, B, G, H, M, N) Claudin and occludin proteins are localized to TJs of the Sertoli cell in the basal compartment of the seminiferous epithelium. C, D) Expression of claudin-3 is diminished in the region of TJs of Sertoli cells in Sox8^−/− testes. I, O) Claudin-11 and occludin expression is retained in the TJs of the Sox8^−/− testes. E, F, K, L, Q, R) Immunohistochemical staining determines the localization of claudin-3, claudin-11, and occludin in the TJs of epididymis. The arrows in all the panels are pointing to TJs. Original magnification ×40.

BTB Permeability Is Increased in the Sox8^−/− Mice

Sox8^−/− testes exhibit reduction of claudin-3 protein compared to WT testes. This prompted us to examine the ultrastructure of Sertoli cell TJs and to test further whether BTB function was compromised in the Sox8^−/− testes by increased permeability to small molecules (<600 Da). Sertoli cells regulate the flow of nutrients, growth factors, and mitogens by maintaining TJs between adjacent cells, also called the BTB. Functional disturbance of the BTB has been shown to result in testicular dysgenesis or irregular spermatogenesis [1–4]. We first used electron microscopy to make ultrastructural observations of the junctional complexes in both WT and Sox8^−/− mice and found no overt changes in the BTB (Fig. 3, A–D, red arrow) at either 2 or 5 mo of age. Sertoli-Sertoli junctional complexes were uniform, evenly spaced electron-dense/electron-lucent unions in both genotypes. Likewise, Sertoli-spermatid and desmosome junctions were also normal. Further measurements of the basement membrane (Supplemental Fig. S1 and Supplemental Table S1; all the Supplemental Data are available online at www.biolreprod.org) showed no significant differences between WT and Sox8^−/− mice.

FIG. 3.

Ultrastructural examination exhibits junctional complexes in testes. The WT mice are images A (2 mo) and C (5 mo). Sox8^−/− animals are images B (2 mo) and D (5 mo). Note that at both 2 and 5 mo of age, neither WT nor Sox8^−/− mice had evidence of disruption in the junctional complexes, just minimal swelling was observed (blue asterisk) in these particular Sertoli-Sertoli junctional complexes (red arrow). Between the two arrowheads in images A and B, the basement membrane defines the boundaries of the basement membrane of the seminiferous tubule. M, mitochondria; LY, lysosome; N, nucleus. Loss of SOX8 in Sertoli cells increases the permeability of the BTB. E, F, G, H) Penetration of the seminiferous tubules of 2-mo-old WT and Sox8^−/− mice was determined by a biotin tracer 30 min after injection into the interstitial space. E, F) Biotin tracer restricted to the interstitial space and basal compartment of adult WT mice. G, H) Image displayed inclusion of the biotin tracer into the adluminal compartment of seminiferous tubules of Sox8^−/− mouse. Original magnification ×16 500 (A–D) and ×20 (E–H). I) Bar graph shows the average biotin intensity observed in WT and Sox8^−/− testes. The error bar is expressed as mean ± SD.

It has been demonstrated that a reduction of claudin-3 protein in Sertoli cell TJs leads to increased permeability of the BTB, causing adult male infertility [32]. We next injected a biotin tracer into the interstitial space of the testes of anesthetized live mice as described earlier [32]. The permeability of the seminiferous tubules was determined by using anti-streptavidin conjugated to Alexa Fluor 568 antibody. We observed structural differences in the seminiferous tubules with WT tubules appearing squiggly in comparison to the smooth appearance in the Sox8^−/− mice. In the WT testes, the streptavidin staining was detected more in the interstitial spaces and basal compartments than in adluminal compartments (Fig. 3, E, F, and I). In contrast, in the Sox8^−/− testes, the biotin tracer was detected in the adluminal compartment (Fig. 3, G, H, and I), even though occludin staining and ultrastructural analysis showed that TJ structures were present in interstitial spaces and basal compartments (as shown in Figs. 2, O and P, and 3, B and D). These data suggest that by controlling the claudin-3 protein level in testes, SOX8 is critical for the permeability of the BTB.

Microenvironment of the Seminiferous Epithelium Affected in Sox8^−/− Testes

Besides claudin-3, it is possible that other factors also contributed to adult infertility of Sox8^−/− males. Our previous gene expression analysis indicated upregulation of HDAC6 (1.3-fold) in Sox8^−/− testes [35]. HDAC6 is a member of the zinc-dependent HDACs [46, 47] that is localized in cytoplasm and selectively deacetylates alpha-tubulin and HSP90 [38, 48–54]. We then sought to determine whether loss of Sox8 affects acetylation of alpha-tubulin. We employed an anti-acetylated alpha-tubulin antibody that recognizes a short region encompassing acetylated Lys-40 in alpha-tubulin [55] to determine the levels of acetylated alpha-tubulin in different ages of WT and Sox8^−/− testes (Fig. 4). The acetylation level of alpha-tubulin was greatly reduced in Sox8^−/− compared to WT testes at Postnatal Day 45 and at 2 and 5 mo of age. These results suggest that SOX8 is also required for acetylated alpha-tubulin in testes.

FIG. 4.

Loss of SOX8 in adult testes results in attenuation of acetylated alpha-tubulin. A decline in the acetylated alpha-tubulin was detected in Sox8^−/− testes. The whole-cell lysates from the WT and Sox8^−/− testes were used for Western blot analysis. An anti-acetylated alpha-tubulin antibody (6-11B-1) was used to evaluate the extent of the protein.

Tubulin is essential for microtubule function in general. Microtubule-related components have been known to maintain the microenvironment of seminiferous epithelium and contribute to several aspects of Sertoli cell function, including positioning of spermatids within the seminiferous epithelium [13, 17–20]. We previously observed disorientation and inappropriate placement of germ cells within the epithelium and increased elongated spermatid retention, that is, spermiation failure [21]. Together with the results shown in Figure 4, we speculated that loss of SOX8 in Sertoli cells destabilized the microtubules in seminiferous epithelium and consequently affected differentiation of germ cells, which we evaluated by performing an ultrastructural analysis of WT and Sox8^−/− testes.

When compared to WT mice, Sox8^−/− mice showed abnormal microtubules, that is, manchettes, in spermatids at both 2 and 5 mo of age (WT: Fig. 5, A and C, red arrow; Sox8^−/−: Fig. 5, B and D). Microtubules are essential to Sertoli cells for nutritional and physical support of differentiating germ cells. Manchettes are organized microtubular structures that extend from the posterior region of the spermatid nucleus, and manchette abnormalities could contribute to sperm shape abnormalities [56]. The seminiferous epithelium of the Sox8^−/− testes revealed several histological abnormalities, including disorientation, inappropriate placement of germ cells within the epithelium, and increased elongated spermatid retention [21]. Numerous normal manchettes consisting of thin, straight, and consistently uniform microtubules, approximately 1 to 8 μm long, which originate from the neck/nuclear ring (Fig. 5, NR black arrows) of the spermatid head, were seen in all the WT mouse testes. In contrast, Sox8^−/− testes lacked normal manchettes in late-stage spermatids and elongated spermatids (Fig. 5, B and D).

FIG. 5.

Electron microscopic examination shows organelles of the testes. Photographs illustrate structure from the seminiferous tubule of WT and Sox8^−/− mice at 2 mo old. A, C) WT mice showing spermatids with prominent thin straight manchettes (MT; red arrow) extending from the nuclear ring (NR; black arrows). B, D) Sox8^−/− mice show abnormal manchettes. Bar = 5 μm; original magnification ×1900.

Other changes found in 5-mo-old Sox8^−/− mice were a minimal to mild increase in residual bodies, membrane damage, and cellular degeneration that indicated changes secondary to seminiferous tubular degeneration (data not shown). Cellular degeneration and an increase in residual bodies may indicate another cause for decreased fertility in Sox8^−/− mice. The flagella, which control sperm motility and affect male fertility, were normal in both WT and Sox8^−/− adult mice (Supplemental Fig. S2). Altogether, abnormal manchettes and cellular degeneration observed in the adult Sox8^−/− testes could affect the microenvironment of the seminiferous epithelium and potentially contribute to sperm abnormalities leading to decreased fertility.

DISCUSSION

Investigation of SOX8 target genes is important to elucidate the molecular mechanism underlying the role of SOX8 in adult male fertility. In male gonads, SOX8 is specifically produced by Sertoli cells and critical for the maintenance of adult male fertility [21]. Sox8^−/− mice are infertile due to the inability of the Sertoli cells to nurture the germ cells past the round spermatid stage [21]. Tight junctions between adjacent Sertoli cells form the BTB and divide the seminiferous epithelium into basal and adluminal compartments [3]. This physical barrier facilitates the distinct microenvironments required for the differentiated germ cells [3, 57]. The BTB also restricts the movement of intercellular molecules from the interstitial space into the adluminal compartment [3]. Claudin-3 [32], claudin-5 [28], claudin-11 [58], and occludin [45] maintain BTB integrity. Reduction of claudin-3 and deficient BTBs has been demonstrated in mice that are Sertoli cell-specific deletion of androgen receptor (AR) [32]. Our results further strengthen the role of claudin-3 protein in BTB function of Sertoli cells. Our studies illustrated that claudin-3 reduction in the TJs of Sox8^−/− testes enhanced the permeability of the BTB. Further, SOX8 binds to the promoter region of Cldn3, providing a regulatory mechanism for Cldn3 expression. Interestingly, bioinformatics analysis suggest that a putative SOX9-binding site is located downstream of the AR gene transcription start site and may hint at a molecular mechanism(s) for SOX8 in regulation of AR in adult testes. Our results clearly demonstrate that BTB function in Sox8^−/− testes is altered, allowing increased movement of small molecules between the testicular lymph and the adluminal compartment of the seminiferous epithelium. Our results are also consistent with and support the concept that claudin-3 has a BTB regulatory function, despite retention of TJ structures containing claudin-11 and occludin around the Sertoli cell TJs. This supports the view that decreased claudin-3 in Sox8^−/− mice may functionally alter the BTB and cause adult male infertility.

Multiple factors, including claudin-3, likely contribute to adult infertility of Sox8^−/− males. Microtubules, for example, are required to support many of the functions attributed to Sertoli cells in the seminiferous epithelium as mentioned earlier [8, 9–20]. Sox8^−/− males display various defects that collectively result in adult male infertility [21]. However, despite the critical importance of SOX8 in supporting spermatogenesis, little is known about precisely how SOX8 action in somatic Sertoli cells mediates progression of germ cell differentiation. SOX8 may act as transcriptional regulator in Sertoli cells, which direct instructive signals to differentiating germ cells. In addition, SOX8 could play an important role in the maintenance of the microenvironment of the seminiferous epithelium, facilitating permissive conditions rather than being instructive in germ cell differentiation.

Our analysis shows that cellular degeneration and spermatid microtubules integrity was compromised, presumably affected by the microenvironment of the seminiferous epithelium in Sox8^−/− testes and contributing to infertility. In addition, decreased acetylated alpha-tubulin and impaired microtubules in the testes of Sox8^−/− animals possibly altered microenvironments of the seminiferous epithelium, which can contribute to adult male infertility. Our studies suggest that SOX8 is critical in Sertoli cells and may contribute to the maintenance of the microenvironment of the seminiferous epithelium through TJ integrity. This is consistent with evidence that loss of specific Sertoli cell proteins impairs adult male fertility and supports our hypothesis that loss of SOX8 compromise Sertoli cells ability to provide a permissive environment to facilitate germ cell differentiation [32].