Loss of SLC9A3 decreases CFTR protein and causes obstructed azoospermia in mice

Mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene cause cystic fibrosis (CF) and are associated with congenital bilateral absence of the vas deferens (CBAVD), which is the major cause of infertility in male patients with CF. However, most Taiwanese patients with CBAVD do not carry major CFTR mutations. Some patients have a single copy deletion of the solute carrier family 9 isoform 3 (SLC9A3) gene. SLC9A3 is a Na+/H+ exchanger, and depleted Slc9a3 in male mice causes infertility due to the abnormal dilated lumen of the rete testis and efferent ductules. Furthermore, SLC9A3 interacts with CFTR in the pancreatic duct and functions as a genetic modifier of CF. However, SLC9A3 function and its relation to CFTR expression in the male reproductive tract in vivo remain elusive. In the present study, we found that CFTR expression was dramatically decreased in the epididymis and vas deferens of Slc9a3 knockout mice. Adult Slc9a3-/- mice showed not only significantly decreased epididymis and vas deferens weight but also increased testis weight. Furthermore, Slc9a3-/- mice developed obstructive azoospermia because of abnormal abundant secretions and calcification in the lumen of the reproductive tract. Ultrastructural analysis of the epithelium in Slc9a3–/–epididymis and vas deferens displayed disorganized and reduced number of stereocilia and numerous secretory apparatuses. Our data revealed that interdependence between SLC9A3 and CFTR is critical for maintaining a precise microenvironment in the epithelial cytoarchitecture of the male reproductive tract. The Slc9a3-deficient mice with impaired male excurrent ducts in this study provide proof for our clinical findings that some Taiwanese of CBAVD carry SLC9A3 deletion but without major CFTR mutations.


Pathology of cystic fibrosis and congenital bilateral absence of the vas deferens
Cystic fibrosis (CF), characterized by mutations in transmembrane conductance regulator (CFTR) gene, is the most common autosomal recessive disorder in Caucasians [1][2][3][4]. CFTR is an apical membrane Clchannel and is responsible for anion secretion in the lungs, pancreas, male reproductive tract, and other epithelial cells. Loss of CFTR activity causes dehydration of the apical membrane and impairs the clearance of mucus from the respiratory tract [5]. Most patients with congenital bilateral absence of the vas deferens (CBAVD) have mutations in and/ or susceptible variants of the 5T allele in intron 8 of CFTR [6][7][8]. Up to 78%-82% of genetic mutations or variants of CFTR have been detected in CBAVD patients from different countries [6,[8][9][10][11][12][13]. Abnormal atrophy of the tissue structure of the vas deferens and the corpus and cauda epididymis is the major cause of male infertility in patients with CBAVD [14,15]. However, most Taiwanese patients with CBAVD do not carry CFTR mutations, and this is consistent with the low incidence of CF in Asian populations including Taiwan [16]. We previously performed genome-wide mapping of copy-number variations through oligonucleotide arraybased comparative genomic hybridization (CGH) and identified loss of solute carrier family 9 isoform 3 (SLC9A3) allele in Taiwanese men with CBAVD (in two of seven Taiwanese men with CBAVD) [17].

Functional roles of SLC9A3
SLC9A3, a Na + /H + exchanger, is expressed in the apical membranes of epididymal, vas deferens, renal proximal tubule, and intestinal epithelium [18][19][20][21][22]. Zhou et al. indicated that SLC9A3 is also expressed in the nonciliated cells of the efferent duct, which connects the testis and epididymis, and that Slc9a3 -/male mice are infertile because of the abnormal dilated lumen of the rete testis and efferent ductules [23]. However, the role of SLC9A3 in the epididymis and vas deferens remain to be clarified. Another well-known function of SLC9A3 is regulation of ion homeostasis in the intestine and colon. SLC9A3 is mainly involved in the transepithelial absorption of Na + and water and often functionally couples with the Cl -/HCO3exchanger [24]. In one previous study, Slc9a3 -/mice showed elevated intestinal fluid and diarrhea because of decreased Na + and HCO 3 absorption [21].

CFTR interacts with SLC9A3
Ahn et al. was the first to demonstrate that SLC9A3 interacts with the C-terminal PDZ motif of CFTR in PS120 cells [25]. In that study, SLC9A3 and CFTR were colocalised in the pancreatic duct of wild-type (WT) mice and SLC9A3 expression decreased by 53% in the pancreatic duct of homozygous 4F508 mutation (4F/4F) Cftr mice. This reciprocal interaction between SLC9A3 and CFTR is regulated by sodium-hydrogen exchange regulatory cofactor 2 in a renal epithelial cell line [26]. Furthermore, loss of SLC9A3 activity increases survival and reduces the occurrence of intestinal obstructions in Cftr -/mice because it rescues the dehydration induced by impaired CFTR function in the intestinal epithelium [27].
Clinical significance of single nucleotide polymorphisms in SLC9A3 in CF Genetic studies have also supported the clinical association between SLC9A3 and CF. Single nucleotide polymorphisms in SLC9A3 in children with CF are significantly associated with two clinical manifestations, the early infection of Pseudomonas aeruginosa and worsened pulmonary function [28]. Genome-wide association studies have indicated that genetic variants of SLC6A14, SLC26A9, and SLC9A3 in patients with CF (n = 3,763) increased susceptibility to early meconium ileus [29,30]. Furthermore, five CF-modifier loci, including SLC9A3, were associated with lung disease severity in 6,365 patients with CF [31]. These comprehensive studies highlight the critical associations between genetic variants of SLC9A3 and clinical indices such as the penetrance of the phenotype and age of onset in patients with CF. Although previous studies have indicated that SLC9A3 is associated with CFTR and significantly affects the severity of CF-related diseases, the direct connection between SLC9A3 and CF-related diseases in vivo is unclear. In the present study, we found that Slc9a3 deficiency in mice induced CBAVD-like phenotypes.

Results
CFTR reduction may be responsible for the reproductive etiology of Slc9a3 -/mice Most Caucasian patients with CBAVD show genetic mutations or variants of CFTR [6]. However, genes associated with CBAVD in Asian and Taiwanese populations are unclear [16,32]. Our previous large-scale genetic screening suggested that SLC9A3 is a high-potential candidate gene for CBAVD [17]. SLC9A3 and CFTR are coexpressed in the pancreatic duct, and the amount of SLC9A3 was shown to be reduced in 4F/4F Cftr mice [25]. In our results, Slc9a3 -/mice were completely infertile compared with age-matched WT and heterozygous mice (Table 1). Therefore, The number of progeny from breeding Slc9a3 +/+ , Slc9a3 +/-, and Slc9a3 -/male mice is presented as the mean ± SEM (n = 5 for each group).
we evaluated whether CFTR expression in Slc9a3 -/mice was reduced and contributed to the sterility. In contrast to our expectations, CFTR expression was drastically reduced in the caput (95.2% ± 1.3) and cauda (85.7% ± 6.6) epididymis and vas deferens (90.4% ± 4.1) (Fig 1). These findings suggested that reduced CFTR expression was responsible for the reproductive tract pathology in Slc9a3 -/mice. Evaluating the causes of reduced CFTR expression and infertility in Slc9a3 -/males first requires the precise localization of SLC9A3. Until now, information on the localization of SLC9A3 in different regions of the epididymis and vas deferens in mice has been incomplete. Several studies have shown that SLC9A3 is expressed in the apical region of nonciliated cells of the efferent ducts in species including humans, mice, rats, hamsters, and roosters [19,22,23,[33][34][35][36][37][38]. Although the expression of SLC9A3 was detected in the principal cells of the epididymis except distal cauda epididymis and vas deferens in rats [22], additional details regarding SLC9A3 expression in mice remains to be revealed. We further verified the SLC9A3 localization in the epididymis and vas deferens of mice (Fig 2). The specificity of anti-SLC9A3 antibody used in our study was confirmed through immunofluorescence staining on the epididymal sections of two Slc9a3 -/mice (S1 Fig), and no signal was observed along the stereocilia in the epididymal epithelia of Slc9a3 -/mice (S1A and S1B Fig). SLC9A3 localization in the efferent ducts was identical to that in previous studies (Fig 2A) [23,35]. In the epididymis of 2-month-old WT males, SLC9A3 was localized in the apical stereocilia of the corpus and cauda epididymal Western blotting results were quantified through a morphometric analysis. For each organ, the CFTR intensity was normalized to that of β-actin expression, and the mean intensity in WT mice was used as a reference to calculate the fold change in expression in Slc9a3 -/mice. Each bar represents the mean ±SEM; n = 3 per genotype. *Significant difference compared with WT mice (***p < 0.0001, analyzed using the Student's t test).
https://doi.org/10.1371/journal.pgen.1006715.g001 SLC9A3 and CF-related CBAVD epithelia (Fig 2C and 2D). SLC9A3 signals were also detected along the stereocilia on the vas deferens epithelium (Fig 2E, enlarged box). By contrast, SLC9A3 was more widely distributed in the caput epididymal and vas deferens epithelia (Fig 2B and 2E). The SLC9A3 localization (Fig 2) corresponds to the significantly reduced expression of CFTR in Slc9a3-deficient epididymis and vas deferens (Fig 1). We speculate that the interaction between CFTR and SLC9A3 may be critical for protein stability in male excurrent ducts.
Global effects in the reproductive system of Slc9a3 -/mice Ten 2-month-old Slc9a3 -/and WT male mice were analyzed to evaluate the pathology of the reproductive system. Slc9a3-deficient mice displayed increased testis size [3.59 ± 0.08 (WT) vs.  (Fig 3). We suggested that an obstruction probably occurred in the reproductive ducts of Slc9a3 -/males and were similar to those observed in knockout (cf/cf) Cftr mice [39].

Histopathological patterns of the testes and efferent ducts in Slc9a3 knockout mice
To determine the effect of SLC9A3 deficiency on infertility, we first analyzed testicular sections of WT and Slc9a3 -/mice of various ages to determine the course of progressive changes. The structures of the interstitial tissue and seminiferous tubules in WT testis were integrally organized, and germ cells showed regular arrangement and complete development (Fig 4A and 4B). The composition of the germ cell population was comparable between 2-month-old WT and Slc9a3 -/mice, but the organization of germ cells in the Slc9a3 -/mice was slightly disordered (Fig 4C and 4D). However, the number of each germ-cell type was slightly lower in 4 months- Efferent ductules, one of the factors underlying the Slc9a3 -/testicular histopathology, are the connective bridge between the rete testis and epididymis and are involved in the reabsorption of approximately 90% of luminal fluid from the testes [40]. SLC9A3 was expressed on the epithelium of efferent ductules (Fig 2A). Compared with age-matched WT mice (Fig 5A and  5B), the luminal diameters of the efferent ductules in 4-month-old Slc9a3 -/mice (Fig 5C and  5D) were wider. This phenotype was consistent with previous study by Zhou et al. [23]. Moreover, unexpected calcification was observed in the efferent ductules of Slc9a3 -/mice at the age of 4 months (Fig 5C and 5D). According to these data, we speculated that the testes of Slc9a3 -/mice underwent atrophy, probably because of back pressure from fluid accumulation caused by the obstruction and dysfunction of the efferent ducts [41][42][43].

SLC9A3 depletion induces systemic defects in the excurrent ducts
Sterility in patients with CF is due to the loss of the vas deferens and corpus and cauda epididymis [6]. We found that CFTR expression and organ weight were significantly decreased in the epididymis and vas deferens of Slc9a3 -/mice (Figs 1 and 3). To characterize the physiological significance of SLC9A3 in the epididymis and vas deferens, we analyzed the histological changes in Slc9a3 -/males of various ages. The lumen was filled with spermatozoa throughout the entire epididymis in WT mice (Fig 6A and 6B represent the caput epididymis; Fig 7A and 7B represent the cauda epididymis). In 2-month-old Slc9a3 -/males, a reduced number of spermatozoa was observed in a few ducts of the caput epididymis ( Fig  6C and 6D and S3 Fig). Recognizable spermatozoa were nearly absent in the lumen of the caput epididymis of >2-month-old Slc9a3 -/males. Moreover, the level of abnormal secretions was augmented in the lumens of the caput epididymis with an increase in age of Slc9a3 -/mice (Fig 6E-6H).
Similar phenotypes were occurred in the cauda epididymis of Slc9a3 -/mice and were also found to progressively deteriorate with age (Fig 7). Compared with the caput epididymis of Slc9a3 -/mice, the elevated secretions were significantly more severe in the cauda epididymis. The vas deferens of Slc9a3 -/males also displayed elevated secretions (Fig 8). Spermatozoa were not transported from the epididymis into the vas deferens in 2-month-old Slc9a3 -/mice (S3E and S3F Fig). Aberrant secreted materials were observed in the entire vas deferens, beginning from the age of 2 month. (Fig 8 and S4 Fig). The epithelia of excurrent ducts are actively involved in the maintenance of proper luminal milieu through mechanisms including secretion and absorption [44]. Epididymal and vas deferens epithelium are lined by nonmotile stereocilia. These membrane-extended structures substantially increase the surface area and absorptive and secretive capacities of epithelial cells [44]. Because abundant secretions were existed in the excurrent ducts of Slc9a3 -/mice, we further examined the ultrastructure of the epithelium in these organs by transmission electron microscopy. The amount of stereocilia was dramatically fewer in the Slc9a3 -/caput epididymis (Fig 9A and 9B, arrows). Moreover, the stereocilia in the Slc9a3 -/cauda epididymis was less organized and fewer in number. Vesicles, a type of secretory apparatus, were more abundant in the epithelium of the Slc9a3 -/cauda epididymis (Fig 9C and 9D, light blue arrows). Distinct morphological changes were also observed in the vas deferens of Slc9a3 -/mice. Stereocilia were less prominent and disturbed (Fig 9E and 9F). Hence, the less well-ordered and reduced

Discussion
SLC9A3, one of CFTR interactor and genetic modifier of CF, is a candidate gene for CBAVD in Taiwanese patients. Our data indicated reduced CFTR levels and obstructed azoospermialike phenotypes in the reproductive ducts of Slc9a3 -/mice. Furthermore, the augmented secretions in the lumen are progressively worse through the epididymis to the vas deferens with an increase in the age. Slc9a3 -/males also showed testicular atrophy, calcification in the efferent ducts, and impaired cytoarchitectures of the epithelium of the excurrent ducts. These results indicated a direct association between SLC9A3 and CBAVD in vivo (Fig 10).

Progressive changes in the phenotypes of the efferent ductules of Slc9a3 -/mice of different ages
Zhou et al. indicated that the luminal diameter of the efferent ductules in 90-120-day-old Slc9a3 -/mice was dilated [23]. In our results, we further identified unexpected calcification in the lumen of the efferent ductules in Slc9a3 -/mice at the age of 4 months. The expression of SLC9A3 and CF-related CBAVD SLC9A3 is abundant in the apical membrane of the epithelium in the gastrointestinal tract and kidneys [45]. Pan et al. indicate that SLC9A3 plays a critical role not only in Na absorption but also in Ca 2+ homeostasis [46]. In that study, Slc9a3 -/mice exhibited reduced Ca 2+ reabsorption in proximal tubules. SLC9A3-deficient mice also displayed lower intestinal Ca 2+ absorption. Both of these contributed to the hypomineralized bones in Slc9a3 -/mice. We suggest that SLC9A3 might participate in the regulation of Ca 2+ homeostasis in the efferent ductules. SLC9A3 deficiency may have led to an imbalanced Ca 2+ concentration and further depositing in the lumen of the efferent ductules as the Slc9a3 -/mice aged. Because the efferent ductules are the bridge between the testis and epididymis in mice and they participate in the reabsorption of approximately 90% of luminal fluid from the testes, this obstruction in the efferent ductules caused the fluid accumulation and formation of back pressure to the testis of Slc9a3 -/mice [40][41][42][43]. This may be the one of the main reasons for testicular atrophy and obstructed azoospermia in Slc9a3 -/mice.

Functional roles of SLC9A3 in the male reproductive tract
Mammalian spermatozoa are developed in the testis and undergo concentration and maturation in the seminiferous tubules, efferent ducts, epididymis, and vas deferens [38]. Precise regulation of pH and ion homeostasis is critical for luminal milieu [47]. SLC9A1, SLC9A2, SLC9A3 and CF-related CBAVD SLC9A3, and SLC9A5, the Na + /H + exchangers, are critical motional proteins in these processes [44,48]. SLC9A1-SLC9A3 are expressed in the efferent ducts and epididymis, and SLC9A1 and SLC9A5 are expressed in mature spermatozoa [44]. Loss of only Slc9a1 or Slc9a3 in male mice induces infertility. Slc9a1-null mice show severely decreased sperm motility because of disturbed intracellular pH in sperms [49]. SLC9A3 cooperates with ion channels such as CFTR and NBC1 and regulates H + secretion and HCO 3 reabsorption in WT mice [50]. We found that Slc9a3 -/mice displayed obstructed azoospermia-like phenotypes, which may be partially contributed to by decreased CFTR expression, similar to that observed in knockout (cf/cf) Cftr mice [39]. In addition, the dilated efferent ducts of Slc9a3 -/mice were abnormally calcified. We suggest that aberrant secretion and calcification are likely due to the dysregulation of ion homeostasis and improper pH caused by SLC9A3 deficiency.

Interdependent effects of CFTR and SLC9A3 in vivo
Ahn et al. indicated that SLC9A3 colocalized and interacted with CFTR in PS120 cells and mice pancreatic ducts [25]. In that study, the reduced SLC9A3 levels (53%) and activity in the pancreatic ducts of 4F/4F Cftr mice were evaluated. The authors proposed that CFTR forms a complex with SLC9A3 and EBP53 to increase the stability of SLC9A3. In the present study, the markedly reduced CFTR expression in the epididymis and vas deferens caused by SLC9A3 deficiency was more deteriorated (85.7%-95.2%, Fig 1). This is consistent with and even more severe than that reported in the previous study [25]. We suggest that the obstructed azoospermia-like phenotypes in Slc9a3 -/mice were attributable to both SLC9A3 deficiency and reduced CFTR expression. The homozygous 4F508 mutation (4F/4F) and knockout (cf/cf) Cftr mice, which show reduced male fertility, are two major strains of genetically modified mice with Cftr mutations [39,51]. In these strains, the development and morphology of the epididymis and vas deferens are normal before 20 days of age. At 40-44 days, spermatozoa are present in the testes and epididymis of 4F/4F and cf/cf mice, but the lumen of the vas deferens contains abnormal secretions instead of spermatozoa and is narrower. These mouse strains are less fertile than the control mice, with only one in three cf/cf male mice being fertile. In the present study, we found that Slc9a3 -/mice displayed similar but worse phenotypes including impaired spermatogenesis and less spermatozoa and severe aberrant secretions in the epididymis. Moreover, these results are consistent with the deteriorated levels of CFTR in Slc9a3 -/mice reported in a previous study (Fig 1) [25]. Large-scale genetic studies have indicated that genetic variants of SLC9A3 are associated with early infection, lung infection severity, and susceptibility to meconium ileus in patients with CF [28][29][30][31]. However, the association between CBAVD, a mild phenotype of CF, and SLC9A3 is unclear. Recently, we detected the loss of a SLC9A3 copy in Taiwanese patients with CBAVD by performing CGH and real-time PCR [17]. In the present study, we determined the detailed reproductive phenotypes, which are similar to the defects of cf/cf mice, in Slc9a3 -/mice. These findings indicate a direct association between SLC9A3 and CBAVD in vivo. Because the phenotypes of deficiency or different mutations of CFTR in mouse models are quite distinct from those in swine models or humans [14,15,39,52], the pathogenesis in male excurrent ducts caused by SLC9A3 loss in different species should also be dissimilar. The association between SLC9A3 function and CBAVD must be clarified by collecting more CBAVD patients and performing clinical examinations of the epididymis and vas deferens in future.

Ethics statement
The animal use protocol were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of Fu Jen Catholic University (Approval number: A10064). The animal experiments were performed according to the international guidelines and regulations. SLC9A3 and CF-related CBAVD Animals and fertile ability test FVB.129(Cg)-Slc9a3 tm1Ges /J mice were purchased from Jackson Laboratory. The genotype of each male mouse was assayed by extracting genomic DNA from the tail and by performing PCR. The primers used in genotyping were as follows: F1 (5´-CATACAACATAGGACTA GCC-3'), R1 (5´-CACTACTAGTCAGGCACTCT-3'), and R2 (5´-CACTACTAGTCAGG CACTCT-3'). The primer ratio of F1, R1, and R2 was 2:1:1. More than 10 mice of each genotype were sacrificed at 2 months of age by anesthesia with isoflurane, and their organs, including the testes, epididymis, and vas deferens, were collected and weighed. The fertility and fecundity of WT and Slc9a3 -/mice were compared by placing 2-month-old males of each genotype with two WT female mice and by counting the number of pups from each pregnancy.

Histological analysis and immunofluorescence staining
WT and Slc9a3 -/mice were sacrificed at 2, 4, and 6 months of age, and their organs were collected. The testes were fixed in Bouin's solution (Sigma-Aldrich), and the epididymis and vas deferens were fixed in PBS containing 4% paraformaldehyde. Next, the tissues were processed for embedding in paraffin wax. Sections of these paraffin-embedded tissues were stained with hematoxylin and eosin (H&E) for histological analysis. For immunofluorescence staining, the dewaxed sections were boiled with 0.1 M sodium citrate buffer (pH 6.0) for antigen retrieval. Sections were incubated overnight at 4˚C with diluted primary antibodies including anti-SLC9A3 antibody (ab95299; Abcam, Cambridge, MA, USA) and anti-pan-keratin (4545; Cell Signaling, Beverly, MA, USA). Primary antibodies were detected with Alexa Fluor 488 and Alexa Fluor 568 fluorescent secondary antibodies (Invitrogen, Carlsbad, CA, USA) and followed by DAPI staining and mounted with Dako mounting medium. The acrosome was stained with lectin peanut agglutinin (L-32458; Invitrogen, Carlsbad, CA, USA). In all the experiments, at least three age-matched WT and Slc9a3 -/mice were analyzed.

Electron microscopy
Parts of the caput and cauda epididymis and vas deferens from 2-month-old mice were excised and immediately fixed with 4% paraformaldehyde and 0.1% glutaraldehyde overnight at 4˚C. Next, the tissues were rinsed with 0.1 M phosphate buffer (pH 7.2) and treated with 1% osmium tetroxide at room temperature for 2 hours. After being rinsed with phosphate buffer again, the tissues were gradually dehydrated by series of increasing concentrations of ethanol. The tissues were then embedded with Spurr's resin kit (cat-14300; EMS) overnight at room temperature. The embedded tissues were sectioned into 75-nm-thick sections by using an ultramicrotome (EM UC7, Leica Microsystems, Wetzlar, Germany) and mounted on copper grids. Ultramicrographs were acquired using a transmission electron microscope (JEM-1400; JEOL) at 100 Kva.

Statistical analysis
All results were obtained from experiments performed in triplicate (at least) and are presented as the mean ± SEM. Data were analyzed with a Student's t test to determine the significance between two groups. Differences with a p value of <0.05 were considered statistically significant.