ROS-generating oxidase Nox3 regulates the self-renewal of mouse spermatogonial stem cells 1

Transcription

1 BOR Papers in Press. Published on May 6, 2015 as DOI: /biolreprod ROS-generating oxidase Nox3 regulates the self-renewal of mouse spermatogonial stem cells 1 Hiroko Morimoto, 3 Mito Kanatsu-Shinohara, 3,4 and Takashi Shinohara 2,3 3 Department of Molecular Genetics, Graduate School of Medicine, Kyoto University, Kyoto, Japan 4 Japan Science and Technology Agency, PRESTO, Kyoto, Japan 1 This research was supported by the Japan Science and Technology Agency (PRESTO, CREST) and the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan. Financial support was also provided by Uehara Memorial Foundation and Funding Program for Next Generation World-Leading Researchers. 2 Correspondence: Takashi Shinohara, Department of Molecular Genetics, Graduate School of Medicine, Kyoto University, Yoshida Konoe, Sakyo-ku, Kyoto , Japan. tshinoha@virus.kyoto-u.ac.jp Running title: Nox3 in spermatogonia self-renewal Summary sentence: ROS-generating oxidase Nox3 regulates spermatogonial stem cell selfrenewal. Keywords: Spermatogenesis, Spermatogonia, Spermatogonial stem cells, Stem Cells, Testis ABSTRACT Spermatogonial stem cells (SSCs) represent a unique population of germ cells with selfrenewal potential. Although reactive oxygen species (ROS) are considered toxic to germ cells, we recently showed that moderate levels of ROS are required for SSC self-renewal, and that Nox1 is involved in ROS generation. In this study, we show that self-renewal factor treatment induces Nox3 to triggers SSC self-renewal. Nox3 was transiently expressed in cultured spermatogonia by FGF2 and GDNF stimulation, while Nox1 was predominantly expressed during the stable phase of proliferation. Nox3 inhibition by short hairpin RNA reduced cytokineinduced ROS generation and limited the proliferation of cultured spermatogonia. While Nox3 overexpression revealed no apparent effect, depletion of Nox3 decreased the number of SSCs in both cultured spermatogonia and freshly isolated testis cells. Our results suggest that selfrenewal of SSCs is regulated by the sequential activation of different Nox genes and underscore the complexity of ROS regulation in the self-renewal division of SSCs. INTRODUCTION Spermatogonial stem cells (SSCs) are the only stem cells present in the germline and provide the foundation of spermatogenesis. Although their number in the testis is very low, they undergo self-renewal division and produce numerous spermatozoa [1,2]. SSC self-renewal is driven by several cytokines, including glial cell line-derived neurotrophic factor (GDNF). Mice with GDNF overexpression (OE) contain clusters of undifferentiated spermatogonia, whereas decreased GDNF levels reduce SSC self-renewal and compromise spermatogenesis [3]. Due to Copyright 2015 by The Society for the Study of Reproduction.

2 their small population and lack of an SSC-specific marker, studying SSC self-renewal machinery in vivo is very difficult and the mechanism of SSC self-renewal has long remained an enigma. However, the advent of SSC culture techniques has created the possibility of investigating the mechanism of self-renewal in vitro [4,5]. The addition of several cytokines, including GDNF, promotes SSC-self-renewal in vitro and produces grape-like colony formation. These cultured cells, designated germline stem (GS) cells, proliferate in a logarithmic manner and initiate spermatogenesis upon introduction into the seminiferous tubules of infertile mouse testes. The molecular mechanism of self-renewal machinery was analyzed using GS cells. GDNF is now believed to activate HRAS via Src family kinase molecules, and cells transfected with activated HRAS undergo self-renewal without exogenous cytokines [6]. Chemical inhibition of AKT or MAP2K1, both of which are downstream molecules of HRAS, abrogate GS cell proliferation [7-9], suggesting that they are necessary for self-renewal division. These signals upregulate Etv5 and Bcl6b [10,11], which work in combination with other constitutively expressed transcription factors, such as Zbtb16 or Taf4b, to drive SSC self-renewal [5]. In our search to find additional molecules involved in SSC self-renewal, we recently found that supplementation of reactive oxygen species (ROS) inhibitors to GS cells severely compromised their growth [12]. In contrast, while high concentrations of hydrogen peroxide killed GS cells, its addition at modest levels increased GS cell proliferation, and GS cells transfected with activated Hras also showed increased ROS expression. These observations led us to study the function of Nox genes, since they are a major source of ROS [13]. Nox1 appeared to be responsible for ROS generation because Nox1 was not only strongly expressed in stably proliferating GS cells, but its depletion by short hairpin RNA (shrna) also reduced SSC activity [12]. Depletion of other Nox genes, such as Nox2-Nox4, showed no effect. Based on these results, we analyzed Nox1 knockout (KO) mice and found that spermatogonia do not actively proliferate. Moreover, although the number of SSCs was comparable to that of wild-type mice, serial transplantation experiments revealed that the self-renewal of SSCs in Nox1 KO mice does not occur as efficiently as those from the control. Positive effects of ROS on SSC self-renewal were unexpected because ROS are generally assumed to be harmful for spermatogenesis [14]. In this study, we analyzed the role of Nox3 in SSC self-renewal. In 2001, Nox3 was cloned as a homologue of Cybb [15] and is predicted to encode a protein of ~65 kda with six conserved predicted transmembrane alpha helices containing a putative heme-binding region as well as a flavoprotein homology domain. It is thought to function together with Cyba as an enzyme that produces superoxide [13]. Nox1 or Nox3 expression promotes CYBA transport to the plasma membrane and both oxidases can be inhibited by mutations in the CYBA binding sites (SH3 domains) of the so-called Nox organizers (NOXO1 or NCF1)[16]. Although we found no apparent effect of Nox3 depletion in GS cells in our previous study, we observed that GS cells exhibited dynamic changes in Nox gene expression when their proliferation was stimulated by self-renewal factors following cytokine deprivation. This finding led us to examine the role of Nox3 in SSC self-renewal by functional assay. Materials and methods Animals and cell culture GS cells used in the present study were derived from the transgenic mouse line C57BL/6 Tg14(act-EGFP)OsbY01 (a gift from Dr. M. Okabe, Osaka University, Osaka, Japan), which were backcrossed onto a DBA/2 background for at least seven generations. Mice in the C57BL/6 (B6) background were also used for transfection experiments and histological analysis.

3 GS cells were cultured as described previously using Stempro-34 (Invitrogen, Carlsbad, CA)[4]. GS cells were maintained on mouse embryonic fibroblasts (MEFs) that had been treated with mitomycin C. Proliferation assays were performed by plating cells in 12-well plates. Cells were recovered at the indicated time points. GS cells were recovered using 0.25% trypsin/1 mm EDTA. PD (3 μm) and LY (33 μm; both from Selleck Chemicals, Houston, TX) were added to cells at the indicated time points. To determine the cell recovery, cells in a well (including MEFs) were incubated in trypsin/edta for 5 min, and the number of single cells/well was determined using a hemocytometer after stopping the trypsin reaction and vigorous pipetting. Transplantation For transplantation, 4-week-old B6 DBA/2 F1 (BDF1) mice were injected peritoneally with busulfan (44 mg/kg). At least 1 month following the busulfan treatment, animals were anesthetized and cells were transplanted into seminiferous tubules via the efferent duct [17]. Each injection filled 75-85% of the seminiferous tubules. The Institutional Animal Care and Use Committee of Kyoto University approved all animal experimentation protocols. Lentivirus infection For Nox3 OE, mouse Nox3 cdna (a gift from Dr. B. Bánf, University of Iowa, Iowa City, IA) was cloned into CSII-EF-IRES2-Puro. CSII-EF-Eyfp-IRES2-Puro served as a control. All knockdown (KD) vectors are listed in Supplemental Table S1 (Supplemental Data are available online at A mixture of five different clones was used to produce the culture supernatant for KD experiments. For the production of lentivirus particles, 293T cells were plated at a density of cells/55 cm 2 in Dulbecco modified Eagle medium /10% fetal bovine serum (FBS). The next day, the lentivirus vector and packaging plasmids in alpha modified Eagle minimum essential medium were mixed with polyethyleneimine MAX (Polysciences, Warrington, PA) and incubated for 30 min at room temperature before addition to 293T cells. The culture medium was changed the following day, and the virus supernatant was collected at 72 h and 120 h post-transfection. Virus particles were concentrated by ultracentrifugation at 50,000 g for 2 h. The virus titer was determined by infection of 293T cells using a Lenti-X p24 Rapid Titer Kit (Clontech, Mountain View, CA) according to the manufacturer s instructions. Virus particles were transfected into cells with 10 μg/ml polybrene by centrifugation at 3,000 g for 1 h, as described previously [18]. Virus-containing medium was removed the following day in all GS cell transfection experiments. The multiplicities of infection (MOIs) were adjusted to 4.0. Terminal deoxynucleotidyl transferase-mediated dutp nick end-labeling (TUNEL) Staining For TUNEL staining, GS cells were incubated in phosphate-buffered saline (PBS)/0.1% Triton-X/0.1% sodium citrate for 2 min, and labeled using an in situ cell death detection kit: TMR red (Roche Applied Science, Mannheim, Germany) according to the manufacturer s protocol. Cells were counterstained with Hoechst (Sigma, St. Louis, MO) and analyzed under a fluorescence microscope. Antibodies used are listed in Supplemental Table S2.

4 Flow cytometry ROS levels were measured using 10 μm 2,7 -dichlorodihydrofluorescein diacetate (H 2 DCFDA; Invitrogen), as described previously [12]. In brief, single cell suspensions were incubated for 30 min at 37 C with 10 μm H 2 DCFDA in PBS/1% FBS. After washing twice with PBS/1% FBS, cells were incubated for 60 min in PBS/1% FBS before FACSCalibur (BD Biosciences, Franklin Lakes, NJ) analysis. For the analysis of cell surface markers, GS cells were incubated with allophycocyanin (APC)-conjugated antibodies listed in Supplemental Table S2. Analysis of recipient testes Recipient mice were killed 6-8 weeks posttransplantation. Donor cell colonies were counted under UV light, and donor cell clusters were defined as colonies when the entire basal surface of the tubule was occupied and were at least 0.1 mm in length. Immunostaining Single-cell suspensions were concentrated on glass slides by centrifugation with Cytospin 4 (Thermo Electron Corporation, Cheshire, UK). Cells were then treated with 0.1% Triton-X 100 and 0.1% sodium citrate in PBS, and then fixed in 4% paraformaldehyde for 1 h. Testis were fixed in 4% paraformaldehyde for 2 h at 4 C, embedded in Tissue-Tek OCT compound (Sakura Finetek, Tokyo, Japan) and processed for cryosectioning. Sections of 10 μm thickness were then prepared. Rhodamine-labeled peanut agglutinin (PNA; Vector Laboratories, Burlingame, CA) was used to detect the acrosome. Antibodies used are listed in Supplemental Table S2. Hoechst (Sigma) was used for counterstaining. Gene expression analysis Total RNA was isolated using TRIzol (Invitrogen). In case of GS cell cultures, cells were incubated on gelatin-coated plates for 2 h to remove MEFs before sample collection. First-strand cdna was produced using a Verso cdna synthesis kit for reverse transcription-polymerase chain reaction (RT-PCR) (Thermo Fisher Scientific, Waltham, MA). PCR conditions were as follows: 94 C for 5 min, followed by 35 cycles at 94 C for 1 min, 60 C for 1 min, and 72 C for 1 min. For real-time PCR, StepOnePlus real-time PCR system (Applied Biosystems, Cheshire, UK) and FastStart Universal SYBR Green Master (Rox) (Roche Applied Science, Mannheim, Germany) were used according to the manufacturers protocols. Transcript levels were normalized to those of Hprt. PCR conditions were as follows: 95 C for 10 min, followed by 40 cycles at 95 C for 15 s and 60 C for 1 min. Each PCR was performed at least in triplicate. PCR primers used are listed in Supplemental Table S3. Statistical analyses Results are presented as the means ± SEM. Data were analyzed using Student s t-tests. Multiple comparison analyses were performed using ANOVA followed by Tukey s HSD test. RESULTS Nox Gene Expression During GS Proliferation GS cell proliferation is influenced by self-renewal factors, including GDNF and FGF2. These cells form typical grape-like colonies and expand by about 3- to 5-fold over a period of 5 days. However, the absence of self-renewal factors altered the speed of GS cell proliferation (Fig. 1A, B). Unlike stably proliferating conditions, the recovery of GS cells was relatively poor

5 following cytokine deprivation (Fig. 1C). When cells were starved for 4 days and restimulated by GDNF and FGF2, only ~30% of the cells were recovered after 5 days compared to the control cells that had been cultured without cytokine starvation. Although grape-like colonies were present, they were smaller than those found under stably proliferating conditions. These changes in cell proliferation and colony morphology suggest that gene expression profiles are significantly different between cytokine-induced restimulation of proliferation and stable proliferation. To examine the expression of Nox genes in GS cells, we first examined their expression when cells were actively proliferating by real-time PCR. Consistent with our previous study [12], Nox1 was strongly expressed in GS cells. Nox2 was barely detectable and Nox3 was weakly expressed; Nox4 was expressed more strongly than Nox1 (Fig. 1D). However, when cells were cultured without FGF2 and GDNF for 4 days, Nox gene expression patterns changed dramatically (Fig. 1E). In contrast to stably proliferating cells, Nox1, Nox2 and Nox4 were significantly upregulated by a lack of cytokines. When GDNF and FGF2 were added to stimulate proliferation, Nox1 and Nox2 expression was suppressed. However, only Nox3 was upregulated by cytokine treatment. The effect of MAP2K1 and PIK3-AKT pathways on Nox gene expression was examined, since both of these pathways have been implicated in Nox gene expression [19,20]. Real-time PCR analysis showed that inhibition of these pathways differed significantly depending on Nox genes (Fig. 1F). While MAP2K1 suppression by PD did not influence Nox1 and Nox2 expression, this treatment suppressed Nox3 upregulation. It weakly increased Nox4 expression. In contrast, LY , a PIK3 inhibitor, slightly increased Nox4 expression, but inhibited expression of other Nox genes. Among the four Nox genes, Nox3 was inhibited by both pathways. These results suggest that Nox3 is positively regulated by both MAP2K1 and PIK3-AKT pathways and may be responsible for the reinitiation of proliferation by self-renewal factors. Effect of Nox3 Gene Dosage During Stimulation of GS Proliferation Although Nox3 depletion did not affect stably growing GS cells in our previous study [12], the results in the preceding section indicate that Nox3 expression increases concurrent with cytokine-induced GS cell proliferation. GS cells were first transduced with Nox3 cdna by lentivirus vector infection. Real-time PCR revealed a 60.3-fold increase in Nox3 mrna level at 5 days postinfection (n = 4)(Fig. 2A). When cells were stably proliferating, transfection of Nox3 caused no apparent effect when compared with those transfected with a control empty vector (Fig. 2B). We then determined whether Nox3 OE affected cells following cytokine starvation. Cells were transfected and deprived of cytokines for 4 days followed by supplementation of GDNF and FGF2 for 4 days. Assuming that all cells are single cells and that all cells were recovered by trypsin digestion, no significant changes in the number of cell recovery were found even after cytokine stimulation (Fig. 2C), suggesting that Nox3 OE alone does not influence GS cell proliferation. We next examined the effect of Nox3 depletion. GS cells were infected with a lentivirus expressing shrna against Nox3, and cells were deprived of cytokines for 4 days (Fig. 2D). Nox3 mrna expression decreased by 57% 9 days post-transfection (n = 11)(Fig. 2A). Cells were then stimulated with FGF2 and GDNF, and recovered 4 days after restimulation. Although Nox3 depletion did not influence stably proliferating GS cells, the same treatment caused a significant decrease in cell recovery when the cell number was determined 4 days after cytokine

6 stimulation (Fig. 2E, F), and 60.2 ± 0.1% and ± 0.1% of the input cells were recovered from Nox3-depleted and control cells that had been transfected with a control shrna, respectively (n = 5). These results suggest that Nox3 is necessary but insufficient to initiate GS cell proliferation. Effect of Nox3 Depletion on the GS Cell Phenotype To clarify the effect of Nox3 depletion on GS cells, TUNEL staining was performed to examine apoptosis after Nox3 KD and compared with cells that had been transfected with a control scrambled shrna (Fig. 3A). When apoptotic cells were examined after Nox3 KD, Nox3- depleted cells showed a 2.9-fold increase in TUNEL-positive cells (Fig. 3B, C). In contrast, when the proliferation status was examined by MKI67 staining, Nox3-depleted cells showed a 53.4% decrease in MKI67-positive cells (Fig. 3D, E). Since the addition of PD or LY to GS cells decreased ROS levels following cytokine stimulation of starved GS cells compared with control DMSO treatment (Fig. 3F), Nox3, whose expression was inhibited by the MAP2K1 and PIK3-AKT pathways, is likely involved in ROS generation after cytokine stimulation. Consistent with this hypothesis, GS cells that had been depleted of Nox3 showed reduced ROS expression compared with those transduced with a control scrambled shrna (Fig. 3G). We also carried out real-time PCR and flow cytometry to analyze the phenotype. Although we did not find significant effect by flow cytometry (Fig. 3H), real-time PCR analyses showed that Nox3 KD significantly decreased expression of Id4, Etv5, Nanos3, Neurog3, Bcl6b, Zbtb16, Cdkn1a, Ccnd2 and Ccnd3 genes when compared with a control vector. Ccnd1, Sohlh1, and Cdkn1b were upregulated significantly after Nox3 KD (Fig. 3I). These results suggest that the reduction of ROS levels by Nox3 depletion causes apoptosis and compromises proliferation of GS cells. Nox3 Depletion Decreases SSC Self-Renewal in GS Cells Because the phenotypic analysis did not clearly show the effect of Nox3 KD on SSCs, we directly examined the effect of Nox3 KD on self-renewal activity using a functional transplantation assay. Aliquots of cells were transplanted into seminiferous tubules of busulfantreated mice to examine SSC activity (Fig. 4A) [21]. GS cells were transfected with lentivirus vectors that express shrna against Nox3 and the results were compared with those transfected with a control shrna vector. When cells were recovered 4 days after transfection, 46.6% and 59.0% of the input cells were recovered from Nox3-depleted and control cells, respectively (n = 4), Although cell recovery was less for Nox3-depleted cells, the difference was not significant (Fig. 4B). To assess the number of SSCs, aliquots of recovered cells were transplanted into seminiferous tubules of busulfan-treated mice. In addition, the number of SSCs after an additional 4 days of culture with FGF2 and GDNF were also examined. FGF2 and GDNF were added to both the control and Nox3 KD groups following cytokine deprivation. Cell recovery for Nox3-depleted and control cells was 47.1% and 111.7% that of the transfected cells, respectively (n = 4), which was significantly different (Fig. 4B). These cells were also transplanted to evaluate the increase in SSC number during cytokine stimulation. Analysis of recipients 2 months after transplantation showed that the numbers of colonies generated by Nox3-depleted GS cells and control cells before cytokine stimulation were not significantly different (Fig. 4C, D). Similarly, while Nox3-depleted cells produced a larger number of colonies when transplanted after cytokine stimulation, the difference was not significant, suggesting that the concentration of SSCs was unaffected by Nox3 depletion.

7 However, when the total number of SSCs was calculated (cell recovery concentration), the total number of SSCs found in Nox3-depleted GS cells was 27.1% that of the control, indicating a significant difference (Fig. 4E). Immunohistochemical analysis revealed the expression of SYCP3 in recipient testes in both Nox3-depleted and control cells, indicating the progression of meiosis (Fig. 4F). Acrosome formation was also confirmed by PNA staining. Expression and Function of Nox3 Gene in Fresh Testis Cells To clarify the role of Nox3 in vivo, we carried out immunohistological staining of testes samples from 1-, 10-, and 35-day-old mice (Fig. 5A). The samples were also stained with an antibody against CDH1, which is a marker for SSCs. Double immunohistochemistry showed that all CDH1-expressing cells are positive for NOX3 signal. In 1-day-old pup testes, CDH1 + gonocytes, which are precursors of spermatogonia, expressed NOX3. NOX3 expression was also found in all of the CDH1 + spermatogonia and along the basement membrane in 10-day-old pup testes, which contain spermatogonia and spermatocytes. In 35-day-old testes that contain all stages of germ cells, NOX3 expression was more widely expressed in the seminiferous tubule, but the signal was relatively strong along the basement membrane and all CDH1 + undifferentiated spermatogonia expressed NOX3. These results suggested that NOX3 is expressed in CDH1-expressing gonocyte and spermatogonia, which contain SSCs. To confirm the involvement of Nox3 in SSC self-renewal, fresh testis cells from ~7- to 10-day-old mice were analyzed because SSCs are relatively enriched in these mice due to a lack of differentiating germ cells [22]. In the first set of experiments, the impact of Nox3 OE was examined. Pup testis cells were transfected with a lentivirus expressing Nox3 cdna. The results were compared with those infected with an empty control vector. When cells were recovered 2 days after transfection, Nox3 transfection clearly did not alter the number of cells (Fig. 5B). The recovered cells were then transplanted into busulfan-treated mice. Analysis of the recipients at 2 months posttransplantation showed no significant effect of Nox3 OE on germ cell colony number. The number of colonies generated by cells transfected with the Nox3 OE and control vectors was 6.3 and 6.4 colonies per 10 5 transplanted cells, respectively (Fig. 5C, D). These results confirmed our results using GS cells and suggest that Nox3 OE does not influence SSC activity. In the second set of experiments, the effect of Nox3 KD was examined. Dissociated pup testis cells were transfected with Nox3 shrna or control shrna. Although cell recovery had decreased, we were unable to detect any significant difference in cell recovery 2 days posttransfection (Fig. 5E). However, Nox3-depleted pup testis cells and control cells produced 1.4 and 4.8 colonies per 10 5 transplanted cells, respectively, after transplantation into busulfantreated mice (Fig. 5F, G), which was significantly different. Immunohistological analysis of recipient testes showed normal spermatogenesis (Fig. 5H) and indicates that Nox3 is also involved in the self-renewal division of SSCs in both GS cells and freshly isolated testis cells. DISCUSSION We previously showed the importance of ROS in SSC self-renewal [12]. Although an excessive amount of ROS was detrimental and killed GS cells, moderate levels of ROS are required for GS cell proliferation and SSC self-renewal in vivo. Due to the apparent effect of Nox1 deficiency, we initially hypothesized that Nox1 is the only Nox family gene that is responsible for self-renewal. However, our current study suggests that Nox gene expression changes dynamically according to exogenous cytokines, and that Nox3 also contributes to SSC

8 self-renewal when cells are stimulated with cytokines. The difference in cell recovery between starved cells and actively proliferating cells suggests that different genes are used in these processes. Changes in Nox gene expression patterns suggest that Nox1 is not involved in the initial stimulation of GS cell proliferation after starvation. The expression of Nox1 as well as Nox2 and Nox4 was downregulated by cytokine stimulation, which was unexpected, since we previously showed that GDNF or FGF2 stimulation increases ROS levels after cytokine deprivation [12]. Nox3 was the only upregulated gene, which raised the possibility that Nox3 is responsible for ROS generation and promotion of self-renewal division by cytokine stimulation. Little is known about the promoter region of Nox3, which is highly restricted with respect to tissue distribution and is predominantly expressed in the ear, but the role of Nox3 in other tissues has not been well investigated [13]. Due to the importance of the PIK3-AKT and MAP2K1 pathways in SSC self-renewal, we hypothesized that these pathways play critical roles in the regulation of Nox3 expression. As expected, inhibition of the PIK3-AKT or MAP2K1 pathways decreased Nox3 expression. However, these pathways do not necessarily stimulate Nox gene expression, as Nox1 expression was upregulated by MAP2K1 pathway inhibition, suggesting that Nox1 and Nox3 are regulated by different mechanisms. The regulation of Nox gene expression differs significantly among different cell types. Although whether the Nox3 system is constitutively active or activationdependent has remained controversial [13], our results provide evidence that Nox3 is similarly activated by exogenous cytokine stimulation. Nox3 depletion can successfully inhibit ROS generation and decrease GS cell proliferation. This suggested that Nox3 is involved in SSC self-renewal via ROS generation. This result contrasts with our previous observation that depletion of Nox1, but not Nox3, can suppress actively proliferating GS cells [12]. Therefore, Nox1 and Nox3 appear to collaborate to promote self-renewal division in different phases of GS cell proliferation. Several reports have revealed the involvement of different Nox genes in the same cell type, including stem cells [23,24]. However, the roles of different Nox genes in self-renewal are still unclear. Our results are the first to provide evidence that different Nox genes play distinct contributions to stem cell self-renewal. They also support our previous suggestion that stimulation of self-renewal and maintenance may not be the same. Indeed, testes of Nox1 KO and wild-type mice contain comparable numbers of SSCs, but the poor self-renewal activity of Nox1-deficient SSCs became evident only after SSCs were stimulated to undergo self-renewal division by serial transplantation. We speculate that ROS generation is triggered by Nox3, and that once ROS have accumulated, their levels are maintained by Nox1. The latter process may involve MAPK14, which was phosphorylated only when GS cells were stably proliferating [12]. This possibility must be tested in future experiments. In contrast, Nox3 OE does not enhance GS cell proliferation. While Nox3 OE may not have been sufficient to drive proliferation, Nox3 may be necessary but insufficient for triggering the proliferation of GS cells. This idea is consistent with a previously proposed model of Nox3 function, demonstrating that Nox3 depends on several enzymes, including Cyba, Noxa1, Ncf1, Ncf2 and Noxo1 [13.25]. In particular, Noxo1 is considered the key subunit for Nox3 activation because inactivation of Noxo1 mimics the phenotype of Nox3 mutant mice [26,27]. Therefore, the amount of these enzymes may not have been sufficient to enhance GS cell proliferation even when Nox3 was transfected. Due to a lack of suitable antibodies for Nox3, little is known about Nox3 subcellular localization, but these previous observations suggest that the co-expression of other enzymes, such as Noxo1, dramatically enhances ROS production, and is probably

9 necessary for ROS generation in GS cells by Nox3 OE. Nox3 mutant mice have vestibular defects that are responsible for the head-tilt mutation [26]. These defects result in the complete congenital absence of otocinia and profound vestibular dysfunction, thereby causing a failure of the perception of balance and gravity. Another study also showed its involvement in insulin resistance [28]. In db/db mice, increased expression of Nox3 and generation of ROS were demonstrated in the liver, which was accompanied by an increased accumulation of lipids and reduced glycogen content. However, thus far, no apparent abnormalities have been reported with regard to spermatogenesis or fertility. This may not be surprising given the relatively normal phenotype of Nox1 KO mice [12], which produce sperm and retain fertility. Moreover, only spontaneous and chemically induced mutants have been reported for Nox3, which may retain residual activity. While relatively subtle phenotypes of spermatogenesis in these mice may be caused by the redundancy of Nox genes, we suspect that the initial formation of the SSC pool or regeneration of spermatogenesis may be somewhat delayed in Nox3 mutant mice, given its role in triggering proliferation. While our studies have confirmed the importance of ROS in SSC self-renewal, determining how Nox-mediated ROS generation coordinates with total cellular ROS regulation will be important. Mice with mutations in Atm and Foxo1 genes exhibit severe defects in spermatogenesis [29,30], which may be caused by increased levels of ROS. Therefore, the signaling mechanism by which self-renewal factors regulate the web of oxidant and antioxidant molecules needs to be evaluated. In addition, the testis environment is hypoxic and essentially all cells within the seminiferous tubules show positive staining with pimonidazole, a chemical that detects cells experiencing an oxygen concentration less than 1.5% [31]. These results raise questions on how the oxygen level is maintained and regulated in the germline niche, in which SSCs undergo self-renewal. Since ROS can diffuse into other cells, understanding the balance is even more complicated in vivo. Thus, our understanding of ROS regulation requires analysis of not only intracellular machinery, but also extracellular microenvironments, and such analysis will open up another dimension in SSC research, which could have practical implications in the prevention of infertility. ACKNOWLEDGMENTS We thank Ms. Y. Ogata for technical assistance and Dr. B. Bánfi (University of Iowa) for providing us with Nox3 cdna. REFERENCES 1. de Rooij DG, Russell LD. All you wanted to know about spermatogonia but were afraid to ask. J Androl 2000; 21: Meistrich ML, van Beek MEAB. Spermatogonial stem cells. In: Desjardins CC, Ewing LL (eds), Cell and molecular biology of the testis. New York: Oxford University Press; 1993: Meng X, Lindahl M, Hyvönen ME, Parvinen M, de Rooij DG, Hess MW, Raatikainen-Ahokas A, Sainio K, Rauvala H, Lakso M, Pichel JG, Westphal H, et al. Regulation of cell fate decision of undifferentiated spermatogonia by GDNF. Science 2000; 287: Kanatsu-Shinohara M, Ogonuki N, Inoue K, Miki H, Ogura A, Toyokuni S, Shinohara T. Long-term proliferation in culture and germline transmission of mouse male germline stem cells. Biol Reprod 2003; 69: Kanatsu-Shinohara M, Shinohara T. Spermatogonial stem cell self-renewal and development.

10 Annu Rev Cell Dev Biol. 2013; 29: Lee J, Kanatsu-Shinohara M, Morimoto H, Kazuki Y, Takashima S, Oshimura M, Toyokuni S, Shinohara T. Genetic reconstruction of mouse spermatogonial stem cell self-renewal in vitro by Ras-cyclin D2 activation. Cell Stem Cell 2009; 5: Lee J, Kanatsu-Shinohara M, Inoue K, Ogonuki N, Miki H, Toyokuni S, Kimura T, Nakano T, Ogura A, Shinohara T. Akt mediates self-renewal division of mouse spermatogonial stem cells. Development 2007; 134: Oatley JM, Avarbock MR, Brinster RL. Gial cell line-derived neurotrophic factor regulation of genes essential for self-renewal of mouse spermatogonial stem cells is dependent on Src family kinase signaling. J Biol Chem 2007; 282: Braydich-Stolle L, Kostereva N, Dym M, Hofmann MC. Role of Src family kinases and N- Myc in spermatogonial stem cell proliferation. Dev Biol 2007; 304: Ishii K, Kanatsu-Shinohara M, Toyokuni S, Shinohara T. FGF2 mediates mouse spermatogonial stem cell self-renewal via upregulation of Etv5 and Bcl6b through MAP2K1 activation. Development 2012; 139: Oatley JM, Avarbock MR, Telaranta AI, Fearon DT, Brinster RL. Identifying genes important for spermatogonial stem cell self-renewal and survival. Proc Natl Acad Sci USA 2006; 103: Morimoto H, Iwata K, Ogonuki N, Inoue K, Ogura A, Kanatsu-Shinohara M, Morimoto T, Yabe-Nishimura C, Shinohara T. ROS are required for mouse spermatogonial stem cell selfrenewal. Cell Stem Cell 2013; 12: Bedard K, Krause KH. The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology. Physiol Rev 2007; 87: Saalu LC. The incriminating role of reactive oxygen species in idiopathic male infertility: an evidence based evaluation. Pak J Biol Sci 2010; 13: Chen G, Cao Z, Xu X, van Meir EG, Lambeth JD. Homologs of gp91 phox: cloning and tissue expression of Nox3, Nox4, and Nox5. Gene 2001; 269: Ueyama T, Geiszt M, Leto TL. Involvement of Rac1 in activation of multicomponent Nox1- and Nox3-based NADPH oxidases. Mol Cell Biol 2006; 26: Ogawa T, Aréchaga JM, Avarbock MR, Brinster RL. Transplantation of testis germinal cells into mouse seminiferous tubules. Int J Dev Biol 1997; 41: Kanatsu-Shinohara M, Muneto T, Lee J, Takenaka M, Chuma S, Nakatsuji N, Horiuchi T, Shinohara T. Long-term culture of male germline stem cells from hamster testes. Biol Reprod 2008; 78: Lee SH, Park DW, Park SC, Park YK, Hong SY, Kim JR, Lee CH, Baek SH. Calciumindependent phospholipase A2β-Akt signaling is involved in lipopolysaccharide-induced NADPH oxidase 1 expression and foam cell formation. J Immunol 2009; 183: Adachi Y, Shibai Y, Mitsushita J, Shang WH, Hirose K, Kamata T. Oncogenic Ras upregulates NADPH oxidase 1 gene expression through MEK-ERK-dependent phosphorylation of GATA-6. Oncogene 2008; 27: Brinster RL, Zimmermann JW. Spermatogenesis following male germ-cell transplantation. Proc Natl Acad Sci USA 1994; 91: Shinohara T, Orwig KE, Avarbock MR, Brinster RL. Remodeling of the postnatal mouse testis is accompanied by dramatic changes in stem cell number and niche accessibility. Proc Natl Acad Sci USA 2001; 98: Peterson JR, Burmeister MA, Tian X, Zhou Y, Guruju MR, Stupinski JA, Sharma RV,

11 Davisson EL. Genetic silencing of Nox2 and Nox4 reveals differential roles of these NADPH oxidase homologues in the vasopressor and dipsogenic effects of brain angiotensin II. Hypertention 2009; 54: Piccoli C, D Aprile A, Ripoli M, Scrima R, Lecce L, Boffoli D, Tabilio A, Capitanio N. Bone-marrow derived hematopoietic stem/progenitor cells express multiple isoforms of NADPH oxidase and produce constitutively reactive oxygen species. Biochem Biophys Res Commun 2007; 353: Banfi B, Malgrange B, Knisz J, Steger K, Dubois-Dauphin M, Krause KH. NOX3, a superoxide-generating NADPH oxidase of the inner ear. J Biol Chem 2004; 279: Paffenholz R, Bergstrom RA, Pasutto F, Wabnitz P, Munroe RJ, Jagia W, Heinzmann U, Marquardt A, Bareiss A, Laufs J, Russ A, Stumm G, Schimenti JC, Bergstrom DE. Vestibular defects in head-tilt mice result from mutations in Nox3, encoding an NADPH oxidase. Genes Dev 2004; 18: Kiss PJ, Knisz J, Zhang Y, Baltrusaitis J, Sigmund CD, Thalmann R, Smith RJH, Verpy E, Bánfi B. Inactivation of NADPH oxidase organizer 1 results in severe imbalance. Curr Biol 2006; 16: Gao D, Nong S, Huang X, Lu Y, Zhao H, Lin Y, Man Y, Wang S, Yang J, Li J. The effects of palmitate on hepatic insulin resistance are mediated by NADPH Oxidase 3-derived reactive oxygen species through JNK and p38mapk pathways. J Biol Chem 2010; 285: Takubo K, Ohmura M, Azuma M, Nagamatsu G, Yamada W, Arai F, Hirao A, Suda T. Stem cell defects in ATM-deficient undifferentiated spermatogonia through DNA damage-induced cell-cycle arrest. Cell Stem Cell 2008; 2: Goertz MJ, Wu Z, Gallardo TD, Hamra FK, Castrillon DH. Foxo1 is required in mouse spermatogonial stem cells for their maintenance and the initiation of spermatogenesis. J Clin Invest 2011; 121: Gruber M, Mathew LK, Runge AC, Garcia JA, Simon MC. EPAS1 is required for spermatogenesis in the postnatal mouse testis. Biol Reprod 2010; 82: FIGURE LEGENDS Figure 1. Nox gene expression during stimulation of GS cell proliferation (A) Experimental strategy. Cell recovery during a 5-day culture period was compared between stably growing GS cells and those deprived of FGF2 and GDNF 4 days prior to replating and cytokine restimulation. B) Comparison of GS cell appearance. While stably proliferating cells make large colonies by 5 days, cells deprived of cytokines for 4 days produced smaller colonies during the same period. C) The fold increase in cell number after cytokine stimulation (n = 12). Cells ( /3.8 cm 2 ) were plated and recovered 5 days after replating. Results of 3 experiments. D) Real-time PCR analysis of Nox gene expression during the stable growth phase (n = 9). Results of 3 experiments. E) Real-time PCR analysis of Nox gene expression after 4 days of cytokine deprivation followed by cytokine restimulation for 4 h (NF + FG)(n = 9). Expression levels were compared with stably proliferating GS cells (FG) and those after 4 days of cytokine deprivation (NF). Results of 3 experiments. F) Effect of PD and LY on Nox gene expression patterns during cytokine restimulation (n = 9). Cells were starved for 4 days and restimulated by cytokines. Inhibitors were added 1 h prior to cytokine restimulation. Samples were recovered 4 h after cytokine stimulation. Results of 3 experiments. PD, PD ; LY, LY294002; FG, FGF2 + GDNF. Asterisks indicate statistically significant differences (p < 0.05). Bar = 100 μm (B).

12 Figure 2. Effect of Nox3 gene dosage on GS cell proliferation. A) Real-time PCR analysis of Nox3 expression following transfection of Nox3 OE (n = 12) or KD (n = 11). For Nox3 OE, culture medium was removed the following day after transfection and cells were cultured for 4 days with FGF2 and GDNF before sample collection. For Nox3 KD, culture medium was removed the following day after transfection and cells were cultured without cytokines for 4 days. Cells were then cultured with FGF2 and GDNF for 4 days before sample collection. Results of 4 and 3 experiments for Nox3 OE and KD, respectively. B) Cell recovery after Nox3 OE in stably growing GS cells (n = 6). Cells ( /9.6 cm 2 ) were transfected and recovered 7 days post-transfection. Results of 2 experiments. C) Cell recovery after Nox3 OE following cytokine deprivation and restimulation (n = 6). Cells ( /9.6 cm 2 ) were transfected 1 day prior to cytokine removal, and starved for 4 days before cytokine restimulation. Cells were recovered 4 days after restimulation. Results of 2 experiments. D) Experimental strategy for KD experiments. E) Appearance of Nox3-depleted cells. Stably growing GS cells without starvation are also indicated for comparison of colony size. F) The fold increase in cell number after transfection. Cells ( /9.6 cm 2 ) were starved for 4 days followed by cytokine restimulation for 4 days before recovery (n = 5). Cell recovery from stably growing GS cells that did not undergo transfection is also shown for comparison. Results of 3 experiments. Asterisks indicate statistically significant differences (p < 0.05). Bar = 100 μm (E). Figure 3. Phenotype of Nox3-depleted cells. A) Experimental strategy. Cells ( /9.6 cm 2 ) were starved for 4 days followed by cytokine restimulation for 4 days. B) TUNEL staining counterstained with Hoechst C) Quantification of apoptosis. At least 3,325 cells in 20 fields were counted. D) MKI67 staining. E) Quantification of proliferating cells. At least 2,224 cells in 20 fields were counted. F) Flow cytometric analysis of ROS levels in GS cells cultured with PD or LY FGF2 and GDNF were added 1 h after the supplementation with PD or LY ROS levels were determined by flow cytometry 4 h after cytokine stimulation. G) Flow cytometric analysis of ROS levels in GS cells transfected with the Nox3 KD lentivirus. H) Flow cytometric analysis of cell surface markers. I) Real-time PCR analysis of spermatogonia markers (n = 9). PD, PD ; LY, LY Asterisks indicate statistically significant differences (p < 0.05). Bar = 50 μm (B, D). Figure 4. Functional analysis of the SSC activity of GS cells transfected with Nox3 KD vectors. A) Experimental strategy. B) The fold increase in cell number at the indicated time points (n = 4). Cells ( /9.6 cm 2 ) were starved for 4 days followed by cytokine restimulation for 4 days. No cytokines were added for the first transplantation experiment (day 0), but the second transplantation experiment was carried out after cytokines were added to both control and Nox3 KD cultures (day 4). Cell number was determined before and after cytokine stimulation. Results of 4 experiments. C) Macroscopic appearance of recipient testes. Seminiferous tubules with green fluorescence indicate spermatogenesis from cultured cells. D) Colony count [n = 20 for Nox3 (0 day); n = 22 for control (0 day); n = 24 for Nox3 (4 days); n = 23 for control (4 days)]. Results of 4 experiments. E) Changes in SSC number during culture. F) Immunohistochemical section of recipient testes stained with rhodamine-pna (red) for acrosomes and with anti-sycp3 antibody (cyan) for meiotic cells. Counterstained with Hoechst (blue). Asterisks indicate statistically significant differences (p < 0.05). Bar = 1 mm (C), 20 μm (F). Figure 5. Functional analysis of the SSC activity of pup testis cells following Nox3 OE or KD.

13 A) Double immunohistochemistry of NOX3 and CDH1 in postnatal mouse testes. Arrows indicate CDH1 + germ cells (green) expressing NOX3 (red). Counterstained with Hoechst (blue). B) Cell recovery after Nox3 OE (n = 4). All recovered cells ( /9.6 cm 2 ) were used for transfection. Cell number was determined 2 days after transfection for transplantation. Results of 4 experiments. C) Macroscopic appearance of recipient testis transplanted with testis cells after Nox3 OE. D) Colony count (n = 17 for Nox3; n = 18 for control). Results of 4 experiments. E) Cell recovery after Nox3 KD (n = 4). All recovered cells ( /9.6 cm 2 ) were used for transfection. Cell number was determined 2 days after transfection for transplantation. Results of 4 experiments. F) Macroscopic appearance of recipient testis transplanted with testis cells after Nox3 KD. G) Colony count (n = 15 for Nox3; n = 12 for control). Results of 4 experiments. H) Immunohistochemical section of a recipient testis stained with rhodamine-pna (red) for acrosomes and with anti-sycp3 antibody (cyan) for meiotic cells. Counterstained with Hoechst (blue). Asterisks indicate statistically significant differences (p < 0.05). Bar = 10 μm (A), 1 mm (C, F), 20 μm (H).

14

15

16

17

18