SEARCH

SEARCH BY CITATION

Keywords:

  • Germline;
  • Differentiation;
  • Tissue-specific stem cells;
  • Targeted gene disruption;
  • Transgenic mouse;
  • Stem cell–microenvironment interactions

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Stem cells are maintained by both stem cell-extrinsic niche signals and stem cell-intrinsic factors. During murine spermatogenesis, glial cell line-derived neurotrophic factor (GDNF) signal emanated from Sertoli cells and germ cell-intrinsic factor NANOS2 represent key regulators for the maintenance of spermatogonial stem cells. However, it remains unclear how these factors intersect in stem cells to control their cellular state. Here, we show that GDNF signaling is essential to maintain NANOS2 expression, and overexpression of Nanos2 can alleviate the stem cell loss phenotype caused by the depletion of Gfra1, a receptor for GDNF. By using an inducible Cre-loxP system, we show that NANOS2 expression is downregulated upon the conditional knockout (cKO) of Gfra1, while ectopic expression of Nanos2 in GFRA1-negative spermatogonia does not induce de novo GFRA1 expression. Furthermore, overexpression of Nanos2 in the Gfra1-cKO testes prevents precocious differentiation of the Gfra1-knockout stem cells and partially rescues the stem cell loss phenotypes of Gfra1-deficient mice, indicating that the stem cell differentiation can be suppressed by NANOS2 even in the absence of GDNF signaling. Taken together, we suggest that NANOS2 acts downstream of GDNF signaling to maintain undifferentiated state of spermatogonial stem cells. STEM CELLS 2012; 30:280–291.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Spermatogonial stem cells represent a stem cell population in adult testes, and their proper biological activities ensure a continuous production of mature spermatozoa. The stem cell function resides in the undifferentiated spermatogonia, consisting of Asingle (As; isolated single cells), Apaired (Apr; interconnected two cells), and Aaligned (Aal; interconnected 4, 8, 16, or 32 cells) [1–5]. The undifferentiated spermatogonia transform into A1 differentiating spermatogonia, which subsequently undergo six mitotic and two meiotic divisions to form haploid spermatozoa [5]. In addition to the morphological classification, undifferentiated spermatogonia are also characterized by the expressions of various marker genes [6–10] (Supporting Information Fig. 1A): promyelocytic leukemia zinc finger (PLZF) is expressed in the all types of undifferentiated spermatogonia; NANOS2 and glial cell line-derived neurotrophic factor family receptor α 1 (GFRA1) are expressed in large subsets of As and Apr; NANOS3 and Neurogenin3 (Ngn3) are known as markers for Aal spermatogonia. Although the entire undifferentiated spermatogonial pool might together constitute the stem cell population during homeostasis and regeneration [7, 11, 12], the heterogeneous gene expression within the population correlates strongly with the cell fates in the short term: Ngn3-expressing cells tend to differentiate, whereas NANOS2- or GFRA1-expressing cells are more likely to self-renew [7, 13]. Therefore, it would be valuable to ask how cellular states or fates of NANOS2+GFRA1+ and NANOS3+Ngn3+ cells are regulated during spermatogenesis. In this article, we define hereafter NANOS2+GFRA1+ undifferentiated spermatogonia as spermatogonial stem cells and upregulation of NANOS3 and/or Ngn3 as spermatogonial stem cell differentiation.

External niche stimuli and internal gene expression regulate self-renewal and differentiation of spermatogonial stem cells [13–22]. The growth factor glial cell line-derived neurotrophic factor (GDNF), a ligand for GFRA1, is secreted by somatic Sertoli cells and acts as one of the major niche signals for spermatogonial stem cells. GDNF signals via the GFRA1/RET coreceptor through activation of Src family kinases, Ras and phosphoinositide 3-kinase-Akt pathway and subsequently induces several target genes in spermatogonial stem cells [23–31]. Previous study has shown that Gdnf haploinsufficiency led to a progressive germ cell loss phenotype, and that testicular overexpression of Gdnf resulted in an accumulation of undifferentiated spermatogonia [19]. Although mice lacking Gdnf or its receptors exhibit neonatal lethality, transplanted testes from each of Gdnf-, Gfra1-, and Ret-null newborn mice showed severe spermatogonial cell depletion, which might result from deficits in spermatogonial proliferation and/or in their maintenance of an undifferentiated state [20]. It is as yet unclear how GDNF signaling influences adult spermatogenesis, as germ cells are lost by postnatal day (P) 7 in the testes of null mutants of Gdnf or its receptors [20].

An RNA-binding protein NANOS2 is an intrinsic factor required for the maintenance of spermatogonial stem cells [13]. Previously, we reported that the removal of Nanos2 gene in adult testes resulted in the depletion of spermatogonial stem cells, which was evidenced by the absence of PLZF and GFRA1 expression [13]. In contrast, continuous Nanos2 expression in the germline caused an accumulation of spermatogonia, which were positive for PLZF and GFRA1 but exhibited no or lower levels of NANOS3 or Ngn3 expression [13]. Furthermore, the Nanos2-overexpressing cells showed slower proliferation and wild-type levels of apoptosis [13]. These results indicate that NANOS2 suppresses proliferation and differentiation of spermatogonial stem cells.

As NANOS2 and GDNF (or its receptors) induced similar phenotypes upon loss of function and overexpression experiments with respect to spermatogonial stem cell maintenance, we examined the relationship between these two factors. We showed that GDNF signaling pathway induced/maintained NANOS2 expression but not vice versa. Furthermore, the stem cell loss phenotype of the Gfra1–conditional knockout (cKO) mice was partially rescued by overexpressed Nanos2. To verify this phenotype, we examined how stem cell properties were affected by Gfra1 depletion and by Nanos2-overexpression. Our results indicated that NANOS2 was capable of suppressing differentiation of spermatogonial stem cells in the absence of GDNF signaling.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Nanos2-Overexpression Induced by Ngn3-Cre

To characterize recombined cells by Ngn3-Cre, we crossed Ngn3-Cre with Rosa–yellow fluorescent protein (YFP) reporter (courtesy of Dr. R. Kopan) [32] (Supporting Information Fig. 1B) mice. We dissected Ngn3-Cre; Rosa-YFP double transgenic testes at P7 and 6 weeks (Supporting Information Fig. 1D).

For conditional activation of Nanos2, we intercrossed the CAG-floxed CAT-3xFlag-Nanos2 (Tg-Nanos2) transgenic (Supporting Information Fig. 1C) female mice with Ngn3-Cre (courtesy of Dr. S. Yoshida) [33] male mice to obtain Tg-Nanos2; Ngn3-Cre double transgenic male pups. The littermate Tg-Nanos2 transgenic males (without Cre) were used as controls. We used 6-week-old mice for phenotypic analysis (Supporting Information Fig. 1D).

Simultaneous Gfra1-cKO and Nanos2-Overexpression by a Tamoxifen-Inducible Cre

To induce conditional inactivation of Gfra1, Gfra1 floxed mice (courtesy of Dr. H. Enomoto) [34] (Supporting Information Fig. 4A) were crossed with Cre-Ert2 line that ubiquitously expressed a tamoxifen (TM)-inducible Cre recombinase (Artemis Pharmaceuticals GmbH). For conditional activation of Nanos2, we subsequently introduced transgene, CAG-floxed CAT-3xFlag-Nanos2 (Tg-Nanos2) [35] (Supporting Information Fig. 4B) in the Gfra1-cKO background. We generated in a total four kinds of mice harboring the following alleles: (a) Gfra1flox/+; Cre-Ert2, (b) Gfra1flox/flox; Cre-Ert2, (c) Gfra1flox/+; Cre-Ert2; Tg-Nanos2, and (d) Gfra1flox/flox; Cre-Ert2; Tg-Nanos2. Correspondingly, activation of Cre recombinase produces (a) Gfra1Δ/+; Cre-Ert2 (control), (b) Gfra1Δ/Δ; Cre-Ert2 (Gfra1-cKO), (c) Gfra1Δ/+; Cre-Ert2; Tg-Nanos2 (Nanos2-overexpression), and (d) Gfra1Δ/Δ; Cre-Ert2; Tg-Nanos2 (Nanos2-overexpression in the Gfra1-cKO background). Cre activity was induced by intraperitoneal injection of TM (75 mg/kg body weight; Sigma, St. Louis, MO, www.sigmaaldrich.com/united-states.html, dissolved in corn oil Σ) into 4-week-old mice for 5 consecutive days. The testes from each genotype of mice were dissected at indicated time points (Supporting Information Fig. 4C).

Whole-Mount Immunohistochemistry

After removing from the tunica albuginea, the seminiferous tubules were untangled physically. The tubules were fixed with 4% paraformaldehyde (PFA) for 2 hours at 4°C. After washing with PBS-T (0.3% TritonX-100/PBS), the specimens were incubated with 3% skim milk/PBS-T for 1 hour at room temperature (RT) and subsequently with primary antibodies at RT overnight. The following day, the tubules were washed with PBS-T and were incubated with secondary antibodies at 4°C overnight. After that, they were washed with PBS-T before mounting. The specimens were examined using a confocal laser scanning microscope (Zeiss, Germany, www.zeiss.de/en LSM510).

Primary antibodies were used at the following dilutions: mouse anti-FLAG (1:5,000, Sigma), rabbit anti-PLZF (1:1,000, Santa Cruz, Santa Cruz, CA, www.scbt.com), mouse anti-PLZF (1:500, Calbiochem, Darmstadt, Germany, www.merck-chemicals.com/life-science-research/calbiochem/c_m7eb.s1OGuMAAAEjT.h9.zf4), rabbit anti-NANOS2 [36] (1:500), goat anti-GFRA1 (1:200, R&D Systems, Minneapolis, MN, www.rndsystems.com), rat anti-green fluorescent protein (GFP; 1:500, Nacalai Tesque, Kyoto, Japan, http://www.nacalai.co.jp/global/index.html), and rabbit anti-NANOS3 [36] (1:500). All secondary antibodies were used at a 1:250 dilution: anti-rat, anti-rabbit, anti-goat, or anti-mouse IgG antibodies conjugated with either Alexa-488 or Alexa-594 (Molecular Probes, Eugene, OR, www.invitrogen.com/site/us/en/home/brands/Molecular-Probes.html).

Histological Methods

For immunofluorescence analysis, testes were fixed in 4% PFA at 4°C overnight before they were embedded in paraffin or OCT compound (Tissue Tek, Sakura, Tokyo, Japan, www.sakura-finetek.com/top_e.html). Paraffin sections (6 μm) were autoclaved at 105°C for 15 minutes in antigen unmasking solution (Vector Laboratories, Burlingame, CA, www.vectorlabs.com) after deparaffinization and hydration. Frozen sections (8 μm) were rinsed with PBS. After blocking with PBS containing 3% skim milk for 1 hour at RT, sections were rinsed and incubated overnight with primary antibodies at 4°C. The following day, the sections were washed in PBS and incubated with secondary antibodies for 1 hour at RT. After washing, these sections were counterstained with 4′,6-diamidino-2-phenylindole (Sigma) and enclosed.

Primary antibodies were used at the following dilutions: goat anti-GFRA1 (1:200, Neuromics, Edina, MN, http://www.neuromics.com), chick anti-GFP (1:500, Abcam, Cambridge, MA, www.abcam.com), chick anti-GFP (1:500, Aves Labs), rat TRA98 (1:5,000, courtesy of Dr. Y. Nishimune), rabbit anti-PLZF (1:500, Santa Cruz), mouse anti-PLZF (1:100, Calbiochem), mouse anti-FLAG (1:5,000, Sigma), mouse anti-cleaved PARP (1:200, Cell Signaling, Beverly, MA, www.cellsignal.com), rabbit anti–synaptonemal complex protein 3 (1:500, courtesy of Drs. S. Chuma and N. Nakatsuji), rabbit anti-phospho histone H3 (1:500, Millipore, Billerica, MA, www.millipore.com/index.do), and rabbit anti-NANOS2 (1:100). All secondary antibodies were used at a 1:250 dilution: anti-goat, anti-chick, anti-rat, anti-rabbit, or anti-mouse IgG antibodies conjugated with either Alexa-488 or Alexa-594 (Molecular Probes).

Statistics

In each graph, the mean values were shown with the standard error. All statistical analyses in this study were performed using the Student's t test. Statistical significance was defined as p < .05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Ectopic Expression of Nanos2 in GFRA1-Negative Spermatogonia Did Not Induce De Novo GFRA1 Expression

Our previous study showed that the majority of Nanos2-overexpressing cells expressed GFRA1 when we induced Nanos2 by Nanos3-Cre, in which Cre recombination occurs in the embryonic male gonocytes, precursors of spermatogonial stem cells [13]. The result raises two possibilities: (a) NANOS2 acts upstream of GDNF signaling and is directly involved in the induction/maintenance of GFRA1 expression, and (b) GFRA1 expression is maintained independently of NANOS2 where GFRA1 expression might be a coincidental but indirect consequence of the NANOS2-mediated differentiation block. To test whether NANOS2 directly affects GFRA1 expression, we investigated the effect of ectopic Nanos2-overexpression in GFRA1-negative spermatogonia.

Given that Ngn3 expression commences around 7 days after birth [33, 37] (Supporting Information Fig. 1D) and the expression is restricted in a GFRA1-negative undifferentiated spermatogonia in adult [7, 8] (Supporting Information Fig. 1A), we expect that Nanos2 should be induced in GFRA1-negative population from early postnatal stage by Ngn3-Cre. To confirm this, we crossed Ngn3-Cre mice with Rosa-YFP reporter mice (Supporting Information Fig. 1B) that allowed us to visualize the cells that had experienced Ngn3 expression. At P7, we detected YFP-positive germ cells on the basement membrane of seminiferous tubules (Supporting Information Fig. 2A to 2F), reflecting Ngn3-positive germ cells and their descendants at this stage. These Ngn3-lineage cells did not overlap with GFRA1-positive spermatogonia, as evidenced that more than 90% of GFRA1-positive cells were YFP negative (Supporting Information Fig. 2G). In agreement with the high efficiency of this Cre line [33, 37], the 6-week-old double transgenic testes revealed that the majority of the spermatogenic germ cells including spermatogonia, spermatocytes, and spermatids were positive for YFP, as a result of the Ngn3-Cre-mediated recombination that occurred in spermatogonial population by this age (Fig. 1A, 1D). However, when we carefully observed YFP expression, most of GFRA1-positive undifferentiated spermatogonia remained as YFP negative (Fig. 1A–1F, Supporting Information Fig. 2G). Hence, Cre recombination was indeed introduced mostly in GFRA1-negative spermatogonia by using Ngn3-Cre.

thumbnail image

Figure 1. Nanos2-overexpression in GFRA1-negative spermatogonia. (A–F): Immunostaining of 6-week-old Rosa-YFP; Ngn3-Cre testes with the indicated markers. YFP expression is observed in Ngn3-Cre expressing cells and their offsprings. GFRA1- (C) or PLZF- (F) single positive cells were indicated by arrowheads. (G–R): Whole-mount immunostaining of 6-week-old Tg-Nanos2; Ngn3-Cre testes with the indicated markers. The transgene-derived Nanos2 was detected by staining with either anti-FLAG or endogenous anti-NANOS2 antibodies. The dotted lines show outlines of seminiferous tubules. Scale bars = 100 μm. Abbreviations: Aal, Aaligned; Apr, Apaired; As, Asingle; DAPI, 4′,6-diamidino-2-phenylindole; GFRA1, glial cell line-derived neurotrophic factor family receptor α 1; Ngn3, Neurogenin3; PLZF, promyelocytic leukemia zinc finger; Tg, transgene; YFP, yellow fluorescent protein.

Download figure to PowerPoint

To address if NANOS2 could induce GFRA1 expression, we overexpressed Nanos2 in GFRA1-negative population by crossing Ngn3-Cre mice with CAG-floxed CAT-3xFlag-Nanos2 transgenic (hereafter referred to as Tg-Nanos2) mice (Supporting Information Fig. 1C). We examined marker expressions using whole-mount immunostaining in the control and Ngn3-Cre-driven Nanos2-overexpressing testes at age of 6 weeks. In controls, PLZF was expressed in undifferentiated spermatogonia including As, Apr, and Aal (Fig. 1G–1I). Whereas GFRA1 was expressed predominantly in As and Apr spermatogonia, in which endogenous NANOS2 was coexpressed (Fig. 1M–1O). In the Nanos2-overexpressing testes, NANOS2-positive cells were positive for PLZF (Fig. 1J–1L), indicating their property of undifferentiated spermatogonia. Intriguingly, most of these Nanos2-overexpressing cells remained as GFRA1 negative (Fig. 1P–1R). This result indicates that Nanos2-overexpression by Ngn3-Cre did not induce de novo GFRA1 expression. We quantified the results by using testis cross-sections and found that 72.5% of FLAG-NANOS2-positive cells were GFRA1 negative (Supporting Information Fig. 3A–3C, 3J). Furthermore, the proportion of GFRA1-positive cells in PLZF-positive population was not changed in the condition of Nanos2-overexpression (Supporting Information Fig. 3D–3I, 3K). From these results, we ruled out the possibility that NANOS2 acts directly in the induction of GFRA1 expression.

NANOS2 Suppresses Differentiation of GFRA1- Negative Undifferentiated Spermatogonia

NANOS3+Ngn3+ spermatogonia express neither NANOS2 nor GFRA1 (Supporting Information Fig. 1A) and might be in a more differentiated state compared with the NANOS2+GFRA1+ state. To test if NANOS2 could suppress spermatogonial differentiation and maintain stem cell-like state in the absence of GDNF signaling, we investigate the differentiation states of Nanos2-overexpressing spermatogonia driven by Ngn3-Cre, in which these Nanos2-overexpressing cells do not receive GDNF signals because of a lack of GFRA1-receptor expression (Fig. 1M–1R, Supporting Information Fig. 3).

Similar to the case in which Nanos2 was induced via Nanos3-Cre [13], the Ngn3-Cre-driven Nanos2-overexpression resulted in the accumulation of PLZF-positive spermatogonia in the periphery of the tubules, while TRA98-positive germ cells in luminal side of the tubules were dramatically reduced (Fig. 2C, 2D). This result suggests that ectopic Nanos2 disturbs the proper production of differentiating germ cells. Furthermore, the PLZF-positive cells in the Nanos2-overexpressing mice showed lower proliferation rates (Fig. 2B) and similar levels of apoptosis (Fig. 2A) compared with control spermatogonia. Hence, Nanos2-overexpression decreases proliferation and differentiation but does not affect apoptosis even in the GFRA1-negative spermatogonia.

thumbnail image

Figure 2. Characterization of Nanos2-overexpressing spermatogonia induced by Ngn3-Cre. (A, B): Quantification of apoptotic (A) and proliferative (B) spermatogonia in 6-week-old mice (N≧3). PARP is one of the main targets of CASPASE-3, therefore the PARP cleavage serves as a marker of cells undergoing apoptosis. **, p < .01. (C, D): Six-week-old testes were stained with the indicated markers. TRA98 is a germ cell marker that is expressed in spermatogonia, spermatocytes, and round spermatids. (E–J): Whole-mount immunostaining of 6-week-old testes with the indicated markers. Scale bars = 100 μm. Abbreviations: Aal, Aaligned; DAPI, 4′,6-diamidino-2-phenylindole; Ngn3, Neurogenin3; PARP, poly (ADP-ribose) polymerase; PH3, phospho histone H3; PLZF, promyelocytic leukemia zinc finger; Tg, transgene.

Download figure to PowerPoint

To further ask the effects of ectopic Nanos2 on spermatogonial differentiation, we examined the expression of NANOS3, an earliest differentiation marker. We found that Aal spermatogonia in control testes showed high levels of NANOS3 (Fig. 2E–2G); however, the majority of Nanos2-overexpressing cells were NANOS3-negative or -dim (Fig. 2H–2J). These results indicate that NANOS2 is able to suppress differentiation of spermatogonia without GDNF receptor expression.

NANOS2 Expression Was Lost Upon the cKO of Gfra1

We hypothesized that GDNF signaling pathway acts upstream of NANOS2 in the spermatogonial stem cells. To test this hypothesis, we examined NANOS2 expression in Gfra1-cKO mice. We crossed mice carrying a gene cassette composed of floxed Gfra1 cDNA followed by EGFP cDNA [34] (Supporting Information Fig. 4A) with Cre-ERT2 mice that ubiquitously express a TM-inducible Cre recombinase. We injected TM into 4-week-old of (a) Gfra1flox/+ (heterozygous for the Gfra1 floxed allele); Cre-Ert2 or (b) Gfra1flox/flox (homozygous for the Gfra1 floxed allele); Cre-Ert2 mice for 5 consecutive days, resulting in the generation of (a) Gfra1Δ/+ (Gfra1 heterozygous used as control) or (b) Gfra1Δ/Δ (Gfra1-cKO) mice, respectively (Supporting Information Fig. 4C). In the Gfra1 floxed mouse strain, the excision of the floxed Gfra1 gene results in the expression of a reporter EGFP gene, which is inserted into the Gfra1 locus (Supporting Information Fig. 4A). This allowed us to visualize Gfra1 enhancer/promoter activated cells that have undergone Cre recombination, through monitoring of GFP fluorescence. Therefore, we can compare Gfra1-positive, normal spermatogonial stem cells versus Gfra1-null spermatogonial stem cells just after the gene is deleted.

At 5 weeks of age (1 week after the first TM injection), we detected GFRA1 protein expression in GFP (Δfloxed Gfra1)-positive cells in control testes but not in Gfra1-cKO testes (Fig. 3A, 3B, 3I), indicating a successful deletion of Gfra1 upon TM injections. Then, we examined NANOS2 expression in the GFP-positive cells. Consistent with coexpressions of GFRA1 and NANOS2, more than 95% of GFP-positive cells were NANOS2-positive in the Gfra1Δ/+ mice (Fig. 3E, 3J). Upon the knockout of Gfra1, NANOS2 expression was dramatically decreased in these cells although they still showed undifferentiated cell morphology such as As or Apr (Fig. 3F, 3J). This result suggests that GFRA1-mediated GDNF signaling might induce/maintain NANOS2 expression.

thumbnail image

Figure 3. GFRA1 and NANOS2 expression in Gfra1-positive spermatogonia. (A–H): Whole-mount immunostaining of 5-week-old testes with the indicated markers. Scale bar = 100 μm. (I, J): The ratio of either GFRA1-positive (I) or NANOS2-positive (J) clusters (2n cells: 1, 2, 4, 8, 16) was counted in GFP-positive spermatogonial polulation (n = 3). **, p < .01. Abbreviations: As, Asingle; Apr, Apaired; Aal, Aaligned; GFP, green fluorescent protein; GFRA1, glial cell line-derived neurotrophic factor family receptor α 1; Tg, transgene.

Download figure to PowerPoint

Induction of Nanos2-Overexpression in the Gfra1-cKO Background

The essential role of GDNF signaling in the spermatogonial stem cell maintenance is shown by in vivo and in vitro studies [19, 20, 38, 39]. If NANOS2 were an essential downstream factor to achieve this function, we speculate that continuous expression of NANOS2 could negate the stem cell loss phenotype caused by the loss of GDNF signaling. To test this hypothesis, we overexpressed Nanos2 simultaneously with the cKO of Gfra1. We introduced CAG-floxed CAT-3xFlag-Nanos2 transgene (Tg-Nanos2) (Supporting Information Fig. 4B) in the Gfra1flox/+; Cre-Ert2 or Gfra1flox/flox; Cre-Ert2 mouse backgrounds. Injections of TM to (a) Gfra1flox/+; Cre-Ert2; Tg-Nanos2 or (b) Gfra1flox/flox; Cre-Ert2; Tg-Nanos2 mice resulted in the generation of (a) Gfra1Δ/+; Tg-Nanos2 (Nanos2-overexpression in control background) and (b) Gfra1Δ/Δ; Tg-Nanos2 (Nanos2-overexpression in Gfra1-cKO background), respectively.

As Cre-Ert2 is ubiquitously expressed from Rosa26 locus, we expected that Nanos2 expression would be induced in both somatic cells and all stages of germ cells. However, after TM injections, NANOS2 expression was predominantly observed in spermatogonia and a part of spermatocyte, while later stages of germ cells and somatic cells were negative for NANOS2 (Supporting Information Fig. 5). This might be due to a higher efficiency of Cre recombination in mitotic/meiotic cells or a post-transcriptional repression of NANOS2 in later stage germ cells and somatic cells. In spite of the differential efficiency of NANOS2 expression among cell types, Cre-Ert2 successfully induced Nanos2-overexpression in GFP-positive spermatogonia in both Gfra1Δ/+ and Gfra1Δ/Δ backgrounds (Fig. 3C, 3D, 3G–3J).

Partial Rescue of Gfra1-Knockout Phenotype by Overexpressed Nanos2

To determine the speed in which spermatogonial stem cells are lost upon Gfra1 deletion and the extent to which overexpressed Nanos2 could alleviate this phenotype, we examined the testes histology (Fig. 4) and changes in the numbers of PLZF (undifferentiated spermatogonia)- and GFP (Δfloxed Gfra1)-positive cells in each genotype at 5, 6, 8 and 12 weeks of age (Fig. 5). At 5 weeks, seminiferous tubules of control and all three mutant mice seemed similar, with multiple layers of germ cells (Fig. 4A, 4D, 4G, 4J). However, in the Gfra1Δ/Δ mice, in addition to the overall decrease of PLZF-positive spermatogonia by 8 weeks (Fig. 5A, blue bar, Supporting Information Fig. 6C), there was a severe depletion of GFP-positive spermatogonia (Fig. 5B, blue bar, Supporting Information Fig. 6D) compared with controls (Fig. 5A, 5B, white bars, Supporting Information Fig. 6A, 6B). We could detect GFP-positive cells in the Gfra1-cKO testes until 5 weeks of age but never observed thereafter (Fig. 5B, blue bar). By 12 weeks, all types of germ cells had disappeared from the Gfra1Δ/Δ testes, revealing Sertoli cell-only phenotypes (Fig. 4F). Hence, we confirmed that Gfra1 deficiency in adult testes resulted in the stem cell loss phenotype.

thumbnail image

Figure 4. Testicular histology of Gfra1-conditional knockout mice with or without Nanos2-overexpression. (A–L): Testes from each mutant were immunostained with the indicated markers. Scale bar = 100 μm. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; GFRA1, glial cell line-derived neurotrophic factor family receptor α 1; PLZF, promyelocytic leukemia zinc finger; Tg, transgene.

Download figure to PowerPoint

thumbnail image

Figure 5. Temporal changes in the number of spermatogonia in each mutant. The numbers of PLZF-positive (A) or GFP-positive (B) cells per seminiferous tubule were counted at the indicated time points (N≧3). The average numbers were normalized using the control of each stage (white bars). **, p < .01; *, p < .05. Abbreviations: GFP, green fluorescent protein; GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; PLZF, promyelocytic leukemia zinc finger; Tg, transgene.

Download figure to PowerPoint

On the contrary, if we compared the number of PLZF-positive spermatogonia in Gfra1-cKO testes with (Fig. 5A, green bar) or without (Fig. 5, blue bar) Tg-Nanos2, these cells were maintained for a prolonged period in the presence of Tg-Nanos2. At 8 weeks, we could observe PLZF-positive spermatogonia in Gfra1Δ/Δ; Tg-Nanos2 testes (Fig. 5A, green bar, Supporting Information Fig. 6G) but not in Gfra1Δ/Δ testes (Fig. 5A, blue bar, Supporting Information Fig. 6C). Consistently, GFP-positive cells were maintained in the 6-week- or 8-week-old Gfra1-cKO testes only when the Tg-Nanos2 was induced (Fig. 5B, green and blue bars, Supporting Information Fig. 6D, 6H). Eventually, Gfra1Δ/Δ; Tg-Nanos2 mice lost their germ cells including PLZF-positive spermatogonia by 12 weeks (Fig. 4L). From these results, we conclude that a partial or a short-term rescue of the Gfra1-cKO phenotype can be achieved by Nanos2-overexpression.

Apoptotic Cell Death Is Not Responsible for the Loss of Spermatogonial Stem Cells in the Gfra1-Knockout Testes

The stem cell state or “stemness” is regulated via the following cellular processes: (a) cell proliferation to expand the stem cell population; (b) suppression of differentiation to maintain stem cells in undifferentiated state; and (c) cell death/survival to control their quality and quantity. GDNF signaling might control these cellular events in a NANOS2-dependent or -independent manner. To address which cellular event(s) are affected by the knockout of Gfra1 and rescued by Nanos2-overexpression, we compared four different genotypes shown above.

First, we quantified the apoptosis of PLZF- (Supporting Information Fig. 9A) and GFP (Δfloxed Gfra1)- (Fig. 6A) positive cells immediately after the loss of Gfra1. At 5 weeks of age, apoptotic cells were detected at a very low frequency in the control testes and their numbers were not altered by Gfra1 deletion (Fig. 6A, Supporting Information Fig. 7A, 7B). In the Nanos2-overexpressing testes, regardless of the presence or absence of Gfra1, we observed an increased apoptosis in spermatocytes and a part of spermatogonia (Supporting Information Fig. 7C–7G), indicating that Nanos2-overexpression by the ubiquitous Cre-Ert2 may cause inappropriate effects on spermatogenic germ cells. Nevertheless, there were no significant differences in the proportion of cleaved PARP-positive cells within PLZF- (Supporting Information Fig. 9A) or GFP- (Fig. 6A) positive undifferentiated spermatogonia in any comparison we examined. We obtained similar results by using testes from 8-week-old mice (Supporting Information Fig. 10A). From these results, we ruled out the possibility that Nanos2-overexpression rescued the stem cell loss phenotype of Gfra1-cKO testes by interfering with the apoptotic pathway.

thumbnail image

Figure 6. Quantification of apoptosis and proliferation. The rates of apoptotic (A) and proliferative (B) spermatogonia were quantified within GFP-positive populations in 5-week-old mice (N≧3). **, p < .01 and *, p < .05. Abbreviations: GFP, green fluorescent protein; GFRA1, glial cell line-derived neurotrophic factor family receptor α 1; PARP, poly (ADP-ribose) polymerase; PH3, phospho histone H3; Tg, transgene.

Download figure to PowerPoint

Overexpressed Nanos2 Cannot Rescue the Reduced Proliferation of Gfra1-Knockout Spermatogonia

Given that GDNF signaling is implicated in the proliferation of spermatogonial stem cells [20, 23–25, 28, 38], we next addressed whether Nanos2-overexpression could rescue the proliferation defects of Gfra1-cKO mice. We quantified the proliferation of spermatogonial stem cells using phospho-histone H3 (PH3) as a marker of cell proliferation (Fig. 6B, Supporting Information Fig. 9B). Normally (i.e., in Gfra1Δ/+ testes), 30.9% of GFP-positive spermatogonia were positive for PH3 (Fig. 6B, Supporting Information Fig. 8A, 8B), whereas only 8.7% of Gfra1-cKO cells showed the signal (Fig. 6B, Supporting Information Fig. 8C, 8D). Hence, we confirmed that the proliferation of spermatogonial stem cells was highly dependent on GDNF-stimulated pathway.

Upon the overexpression of Nanos2, we found a significant reduction of proliferation in PLZF- (Supporting Information Fig. 9B) or GFP-positive cells (Fig. 6B) regardless of the presence or absence of a functional Gfra1 allele. Significantly, in the Gfra1Δ/Δ; Tg-Nanos2 mice, the proliferation of GFP-positive cells constituted only 1.5% of the total GFP-positive cells (Fig. 6B). Similar results were obtained by using testes from 8-week-old mice (Supporting Information Fig. 10B). Hence, spermatogonial proliferation might not be the reason why Nanos2-overexpression could rescue the stem cell loss phenotype of Gfra1-cKO mice. Rather, NANOS2 and GDNF signaling might regulate the proliferation of spermatogonial stem cells in different pathways: GDNF signaling stimulates the proliferation via known [23, 24, 27, 28, 30] or unknown pathways, while NANOS2 might delay cell cycle progression of the stem cells independent of GDNF signaling.

Overexpression of Nanos2 Blocks Stem Cell Differentiation Caused by cKO of Gfra1

The stem cell loss is also caused by the inability of stem cells to maintain an undifferentiated state, which may result in the depletion of the stem cell population through differentiation. Indeed, Gdnf-overexpression suppresses differentiation of spermatogonia, and the deletion of Gdnf or its receptors promotes differentiation of spermatogonia and causes the stem cell loss phenotype in vivo and in vitro [19, 20, 38]. To test if Nanos2-overexpression could suppress differentiation of Gfra1-deficient stem cells and contribute to the stem cell maintenance, we evaluated the differentiation state of Gfra1-deficient stem cells with or without overexpressed Nanos2. We performed whole-mount immunostaining for GFP (Δfloxed Gfra1) and NANOS3 at 5 weeks (Fig. 7). In control testes, NANOS3 expression was much higher in GFPlow or − spermatogonia (morphologically defined as Aal) than in GFP+ spermatogonia (morphologically defined as As or Apr) (Fig. 7A). In the Gfra1-deficient testes, however, most GFP-positive spermatogonia, including As or Apr expressed NANOS3 strongly (Fig. 7B). Quantitative analyses revealed that more than 80% of GFP-positive spermatogonial cluster in the Gfra1-cKO testes exhibited a strong NANOS3 signal, while 37.0% of GFP-positive cells in the control testes showed the signal (Fig. 7E). Consistent with this result, the proportion of NANOS3-positive cells was increased in the entire PLZF-positive spermatogonia upon the cKO of Gfra1 (Supporting Information Fig. 11A, 11B, 11E). These results suggest that GDNF signaling acts as a suppressive factor on differentiation of spermatogonial stem cells.

thumbnail image

Figure 7. Forced expression of Nanos2 suppresses differentiation of Gfra1-knockout spermatogonia. (A–D): Whole-mount immunostaining of 5-week-old testes with the indicated markers. Scale bar = 100 μm. (E): Frequency of NANOS3-positive spermatogonial clusters (2n cells: 1, 2, 4, 8, 16) in GFP-positive clusters was counted by whole-mount immunostainig of tubules (n = 3). **, p < .01. Abbreviations: As, Asingle; Apr, Apaired; Aal, Aaligned; GFP, green fluorescent protein; GFRA1, glial cell line-derived neurotrophic factor family receptor α 1; Tg, transgene.

Download figure to PowerPoint

On the contrary, Nanos2-overexpression decreased the level of NANOS3 expression in spermatogonia both in PLZF- (Supporting Information Fig. 11C, 11E) and GFP- (Fig. 7C, 7E) positive population. Notably, the elevated NANOS3 expression caused by Gfra1 deletion was suppressed in the cells overexpressing Nanos2 (Fig. 7D, 7E, Supporting Information Fig. 11D, 11E). In the Gfra1Δ/Δ; Tg-Nanos2 mice, only 9.8% of GFP-positive cells showed the NANOS3 signals (Fig. 7E). Hence, overexpressed Nanos2 can delay the differentiation progression of spermatogonial stem cells caused by Gfra1 deficiency. Taken together, we conclude that NANOS2 acts downstream of GDNF signaling to suppress the differentiation of spermatogonial stem cells.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

NANOS2 Is a Possible Target of GDNF Signaling and Suppresses Differentiation of Spermatogonial Stem Cells

In vivo studies showed that deletion of GDNF signaling components [20] (Figs. 5, 7) or Nanos2 [13] caused a precocious differentiation of spermatogonia and subsequent loss of stem cell population. In cultured spermatogonia, several genes related to differentiation, for example, Ngn3 or Kit, were upregulated by GDNF/GFRA1 removal or Gfra1-knockdown experiments [29, 38]. In contrast, overexpression of Gdnf resulted in the accumulation of undifferentiated spermatogonia due to a blocked differentiation [19]. Similarly, Nanos2-overexpression using Nanos3-Cre line showed that NANOS2 cell autonomously maintained spermatogonia as NANOS3-Ngn3 undifferentiated state in the presence of GFRA1 [13]. Therefore, GDNF signaling and NANOS2 might have suppressive roles for initial/early step of spermatogonial differentiation, probably at the transition point from NANOS2+GFRA1+ to NANOS3+Ngn3+.

In this study, we used several combinations of mutant mice and examined the interplay between NANOS2 and GDNF signaling (summarized in Supporting Information Fig. 12). We found that NANOS2 expression was dependent on GDNF signaling pathway, as cKO of Gfra1 resulted in an immediate loss of NANOS2 expression (Fig. 3E, 3F, 3J). When we manipulated NANOS2 to be continuously expressed after the knockout of Gfra1, NANOS2 suppressed differentiation of Gfra1-null spermatogonia (Fig. 7) and maintained the stem cell population for a longer period when compared with simple Gfra1-null spermatogonia (Fig. 5). NANOS2 also has an ability to repress the differentiation of GFRA1-negative, Ngn3-positive population (Fig. 2E–2J). From these results, we propose that NANOS2 is one of the strong candidates of downstream factor of GDNF signaling with respect to the suppression of spermatogonial stem cell differentiation.

Although the relationship between GDNF signaling and NANOS2 begins to be understood, we need further studies to answer the following questions. (a) Although NANOS3 expression was increased in Gfra1-cKO cells, it is still unclear whether these cells retain the ability to enter normal differentiation pathway just like endogenous NANOS3+Ngn3+ population. The long-term lineage tracing experiments using ubiquitous lineage tracing marker will be needed to tell the fate of the mutant cells. (b) We still do not know whether GDNF signaling directly regulates NANOS2 transcription/translation or indirectly affects NANOS2 expression as a result of maintaining proper cellular state of spermatogonial stem cells. We need to test whether NANOS2 expression is upregulated in the Gdnf-overexpressing mice; if so, the deletion of Nanos2 cause spermatogonial cell loss, due to de-repression of differentiation even in the presence of overexpressed Gdnf. The answers to these questions will surely lead to better understanding of the molecular mechanisms in spermatogonial stem cell regulation.

Opposite Roles of NANOS2 and GDNF Signaling in the Proliferation of Spermatogonial Stem Cells

Given that Nanos2-overexpression could not rescue the phenotype of Gfra1-cKO mice completely (Figs. 4, 5), GDNF signaling might also have a NANOS2-independent pathway to maintain spermatogonial stem cells. A possible explanation is that the proliferation of spermatogonial stem cells is highly dependent on the GDNF-stimulated pathway. It is supported by the analyses in our Gfra1-cKO mice (Fig. 6B) or Gdnf-, Gfra1- and Ret-null mice [20], in which the spermatogonial proliferation is significantly reduced. It is also shown that, the GDNF and its downstream factors are essential for the growth of spermatogonial stem cells in culture [23–25, 28, 38–40]. On the other hand, our previous [13] and current studies (Figs. 2B, 6B, Supporting Information Fig. 9B) revealed that Nanos2-overexpressing cells showed a lower proliferation rate than that of control spermatogonia. As Gfra1 is required for the promotion of proliferation of spermatogonial stem cells and the proliferation is further decreased by the presence of overexpressed Nanos2, the number of GFP (Δfloxed Gfra1)-positive cells in the Gfra1Δ/Δ; Tg-Nanos2 mice gradually reduced (Fig. 5B) due to the significantly lower growth rate (Fig. 6B) and normal level of apoptosis (Fig. 6A).

The suppressive role of NANOS2 on spermatogonial proliferation is an interesting feature of the stem cell regulation. Because, it has long been proposed that stem cells divide infrequently and are predominantly quiescent in several other tissues [41, 42]. However, it remains unknown (a) differences between NANOS2+GFRA1+ versus NANOS3+Ngn3+ population in terms of cell division frequency; (b) a link between slow-cycling nature and stem cell activity/potential in spermatogonial stem cells; and (c) crosstalk between pathways for regulating proliferation and differentiation. We will address these issues as future works, especially by focusing on the identification of NANOS2 targets, which are specifically involved in the regulation of either differentiation or cell cycle/division.

Nanos2-Overexpression Phenotypes by Using Different Cre Lines

In our previous study, we used Nanos3-Cre to induce Nanos2-overexpression, in which Cre expression commences in all embryonic male gonocytes [13]. After birth, the male gonocytes resume proliferation and migrate from their original central position toward the periphery in the seminiferous tubules and transform into spermatogonia by P7 [1, 3, 43]. We think that Nanos3-Cre-driven Nanos2-overexpression blocks differentiation of spermatogonia, and the Nanos2-overexpressing spermatogonia never proceed from NANOS2+GFRA1+ into NANOS3+Ngn3+ state (Supporting Information Fig. 12B). Consistent with this idea, the majority of Nanos3-driven Nanos2-overexpressing cells were identified as GFRA1-positive, As or Apr cells in the adult stage [13].

As shown in the results of lineage tracing experiment of Ngn3-Cre, most of GFRA1-positive cells are not marked by Ngn3-Cre in either P7 (Supporting Information Fig. 2) or adult (Fig. 1A–1F) stage, indicating that Cre recombination occurred in the cells that have already exited from NANOS2+GFRA1+ state and have become NANOS3+Ngn3+. Consistently, the Ngn3-driven Nanos2-overexpressing cells in the adult testes exhibited Aal-like properties reflecting the characteristic of Ngn3-expressing cells (Fig. 1G–1R). However, the effects of Nanos2-overexpression by Ngn3-Cre were similar to that observed in the Tg-Nanos2/Nanos3-Cre mice: NANOS2 slows down proliferation and suppresses cell differentiation (Fig. 2, Supporting Information Fig. 12C).

Nanos2-overexpression induced via Cre-Ert2 occurred in the entire undifferentiated spermatogonia (and much broader cells) (Supporting Information Figs. 5, 12E, 12F). In both Gfra1Δ/+; Tg-Nanos2 and Gfra1Δ/Δ; Tg-Nanos2 mice, GFP (Δfloxed Gfra1)-positive cells were As or Apr and showed no NANOS3 expression (Fig. 7) similar to the case of Tg-Nanos2/Nanos3-Cre [13]. Furthermore, similar to the case of Ngn3-Cre, Cre-Ert2 driven Nanos2-overexpression might suppress differentiation of GFRA1-negative (NANOS3+Ngn3+) population, as shown by NANOS3 downregulation in the entire PLZF-positive population (Supporting Information Fig. 11). The proliferation rate was also decreased in both GFP- and PLZF-positive cells by Cre-Ert2-driven Nanos2-overexpression irrespective of the presence or absence of Gfra1 (Fig. 6B, Supporting Information Fig. 9B). Together, these results strongly support our idea that NANOS2 can repress differentiation and proliferation of both GFRA1-positive cells and the cells that have lost GFRA1 expression, either by Gfra1-knockout or in a process of normal differentiation into NANOS3+Ngn3+ state.

Stem Cells in Murine Spermatogenesis

Although studies using functional transplantation assays along with lineage tracing experiments have demonstrated that the stem cell activity/potential is highly enriched in the undifferentiated spermatogonial population, an unequivocal identification of the stem cells within the subpopulation of undifferentiated spermatogonia is still under argument [6, 7, 9, 12, 13, 44]. In particular, several studies indicate that GFRA1 selection did not result in the enrichment of transplantable stem cells in adult testes [6, 12]. However, we denoted GFRA1+NANOS2+ spermatogonia as stem cells in this study, as our previous long-term lineage tracing analysis indicated that NANOS2-expressing spermatogonia (frequently coexpress GFRA1) acted as spermatogonial stem cells in vivo and the lack of Nanos2 led to a stem cell loss phenotype [13]. In addition, we showed for the first time that cKO of Gfra1 gene in the adult stage caused a depletion of spermatogonia, followed by a complete loss of spermatogenic germ cells (Figs. 4, 5), strongly indicating that GFRA1-positive spermatogonia constitute the stem cell population. As indicated by the recent study [44], one cannot exclude the possibility that ex vivo manipulation of germ cells and pretreatment of recipients involved in functional transplantation assays may have altered the isolated donor cells or the recipient tissues, which might affect cell behavior. Future experiments are required to determine the fates of GFRA1-positive cells in intact testes by using an inducible Cre-loxP system.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

This study shows a genetic interaction of NANOS2 and GDNF signaling for the maintenance of stem cells during murine spermatogenesis. We suggest that NANOS2 maintains the undifferentiated state of spermatogonial stem cells downstream of GDNF signaling. Our results offer a novel insight into the understanding of how spermatogonial stem cells are regulated by stem cell-intrinsic and -extrinsic mechanisms.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

We thank Dr. H. Enomoto for the Gfra1 floxed mice; Dr. S. Yoshida for the Ngn3-Cre transgenic mice; R. Kopan for the Rosa-YFP mice; Dr. Y. Nishimune for the TRA98 antibody; and Drs. S. Chuma and N. Nakatsuji for the SCP3 antibody. This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (Japan) to Y.S. and a Research Fellowship for Young Scientists from the Japan Society for the Promotion of Science to A.S.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICT OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
STEM_790_sm_supplFig1.pdf257KSupplemental Figure 1. Strategy of Nanos2-overexpression by Ngn3-Cre. (A): A schematic drawing of the relationship between the typical morphology and gene expression patterns in PLZF-positive undifferentiated spermatogonia. Asingle and Apaired spermatogonia contain a higher proportion of GFRA1+NANOS2+, while many of Aalinged are NANOS3+Ngn3+, which then become differentiating spermatogonia. Of note, the heterogeneity exists even in the cysts with the same number of cells with regard to their gene expression (e.g. a small subset of Asingle spermatogonia are GFRA1-NANOS2-NANOS3+Ngn3+). (B): Rosa-YFP reporter mice start to express YFP after removal of the inactivating sequence by Cre-mediated recombination. Because the recombination is irreversible and inheritable, progenies of the recombined cells can be marked by YFP. (C): A conditional transgenic mouse line, CAGfloxed CAT-3xFlag-Nanos2 (Tg-Nanos2) is capable of expressing Flag tagged-Nanos2 under the control of the ubiquitous CAG promoter after excision of the floxed CAT gene (an innocent gene) by Cre recombinase. (D): The experimental schedule. Ngn3-Cre starts to be expressed in spermatogonia around P7. We collected testes at the indicated time points. Abbreviations: PLZF, promyelocytic leukemia zinc-finger; GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Ngn3, Neurogenin3; Tg, transgene; CAG, cytomegalovirus (CMV) early enhancer/chicken beta-actin promoter; CAT, chloramphenicol acetyltransferase; P, postnatal day.
STEM_790_sm_supplFig2.pdf260KSupplemental Figure 2. Ngn3-Cre targets GFRA1-negative cells. (A-F): At P7, Rosa-YFP; Ngn3-Cre double transgenic testes were immunostained with the indicated markers. Most of YFP-positive spermatogonia (Ngn3-lineage cells) were not stained with GFRA1. Scale bar, 100 μm. (G): Quantification of YFP-positive cells per GFRA1-positive cells at P7 and 6 weeks (N=3). Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Ngn3, Neurogenin3; P, postnatal day.
STEM_790_sm_supplFig3.pdf311KSupplemental Figure 3. Nanos2 overexpressed by Ngn3-Cre did not induce de novo GFRA1 expression. (A-I): Immunostaining of 6-week-old Tg-Nanos2; Ngn3-Cre double transgenic testes with the indicated markers. Most of FLAG-positive cells did not express GFRA1. Scale bar, 100 μm. (J-K): Quantification of GFRA1-positive cells per FLAG- (J) or PLZF- (K) positive cells in 6-week-old mice (N=3). The proportion of GFRA1-positive cells was not changed after Ngn3-Cre driven Nanos2-overexpression. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; PLZF, promyelocytic leukemia zincfinger; Tg, transgene; Ngn3, Neurogenin3.
STEM_790_sm_supplFig4.pdf175KSupplemental Figure 4. Strategy of Nanos2-overexpression in the background of Gfra1- conditional knockout. (A): Floxed Gfra1 allele expresses Gfra1, thereby serving as a functional allele. Activation of Cre recombinase results in a removal of floxed Gfra1 sequences, simultaneously generating EGFP knock-in (Gfra1 knock-out) allele. (B): A conditional transgenic mouse line, CAG-floxed CAT-3xFlag-Nanos2 (Tg-Nanos2) is capable of expressing Flag tagged-Nanos2 under the control of the ubiquitous CAG promoter after excision of the floxed CAT gene (an innocent gene) by Cre recombinase. (C): The experimental schedule. Cre activity was induced by intraperitoneal injection of Tamoxifen into 4-week-old mice for 5 consecutive days and the testes were harvested at the indicated time points. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene; CAG, cytomegalovirus (CMV) early enhancer/chicken beta-actin promoter; CAT, chloramphenicol acetyltransferase.
STEM_790_sm_supplFig5.pdf394KSupplemental Figure 5. Successful induction of Nanos2-overexpression by Cre-Ert2. (A-I): Five-week-old testes were immunostained with the indicated markers. In control Gfra1δ/+ mice, NANOS2 was expressed in a small subset of spermatogonia (arrowheads). In Gfra1δ/+; Tg-Nanos2 or Gfra1δ/δ; Tg-Nanos2 mice, strong NANOS2 signal were detected in early germ cells located on the basal side of seminiferous tubules. Scale bar, 100 μm. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene.
STEM_790_sm_supplFig6.pdf341KSupplemental Figure 6. Partial rescue of Gfra1-conditional knockout phenotype by overexpressed Nanos2. (A-H): Eight-week-old testes were stained with indicated markers. PLZF- or GFP- positive spermatogonia are indicated by arrowheads. At 8-weeks, both PLZF- and GFP- positive spermatogonia were observed in Gfra1δ/δ; Tg-Nanos2 mice but not in Gfra1δ/δ mice. Scale bar, 100 μm. Abbreviations: PLZF, promyelocytic leukemia zinc-finger; GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene.
STEM_790_sm_supplFig7.pdf273KSupplemental Figure 7. Apoptotic cell death in each mutant. (A-D): The representative images of double immunostaining with anti-GFP (green, A-D) or anti-SCP3 (a marker of meiosis, green, E-G) and anti-cleaved PARP (magenta, A-G) antibodies at 5 weeks. Arrowheads indicate GFP-positive cells. Apoptotic cells were increased in Nanos2- overexpressing testes especially in meiotic germ cells (E-G), but the most of them were GFPnegative (A-D). Scale bar, 100 μm. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene; SCP3, synaptonemal complex protein 3; PARP, poly (ADP-ribose) polymerase.
STEM_790_sm_supplFig8.pdf290KSupplemental Figure 8. Detection of cell proliferation in Gfra1-conditional knockout and Nanos2-overexpressing cells. (A-H): The representative images of double immunostaining with anti-GFP (green) and anti-PH3 (magenta) antibodies at 5 weeks. GFP-positive spermatogonia are indicated by arrowheads. Scale bar, 100 μm. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene; PH3, phospho Histone H3.
STEM_790_sm_supplFig9.pdf157KSupplemental Figure 9. Quantification of proliferation and apoptotic cell death in PLZFpositive spermatogonia. Quantification of apoptotic (A) and proliferative (B) spermatogonia in 5-week-old mice (N≧3). **, P<0.01 and *, P<0.05. Apoptosis in the PLZF-positive cells didn't increase in any mutants (A), indicating that the mutant cells survived even after losing GFP expression. However, we did not find much difference in proliferation rate between control and Gfra1δ/δ testes in PLZF-positive population (B). Since the ratio of GFP-positive cells in PLZF-positive cells is low (20-25%), the decrease in the proliferation rate in the GFP-positive cells in Gfra1δ/δ was not reflected in the proliferation rate in the PLZF-positive population. Abbreviations: PLZF, promyelocytic leukemia zinc-finger; PARP, poly (ADPribose) polymerase; PH3, phospho Histone H3; GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene.
STEM_790_sm_supplFig10.pdf168KSupplemental Figure 10. Quantification of proliferation and apoptotic cell death at 8-weeks of age. Quantification of apoptotic (A) and proliferative (B) spermatogonia in 8-week-old mice (N4). **, P<0.01. No GFP-positive spermatogonia are observed in Gfra1δ/δ mice at 8- weeks. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; Tg, transgene; PARP, poly (ADP-ribose) polymerase; PH3, phospho Histone H3.
STEM_790_sm_supplFig11.pdf216KSupplemental Figure 11. Overexpression of Nanos2 by Cre-Ert2 suppresses differentiation of PLZF-positive spermatogonia. (A-D): Whole-mount immunostaining of 5-week-old testes with the indicated markers. Scale bar, 100 μm. (E): Frequency of NANOS3-positive spermatogonial clusters (2n cells: 1, 2, 4, 8, 16) were defined in PLZF-positive population (N=3). **, P<0.01. Differentiation rate in PLZF-positive cells in Gfra1δ/δ testes were increased as expected by the promoted differentiation in GFP-positive cells, while it was strongly suppressed by Nanos2-overexpression. Abbreviations: PLZF, promyelocytic leukemia zinc-finger; GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; As, Asingle ; Apr, Apaired; Aal, Aaligned; Tg, transgene.
STEM_790_sm_supplFig12.pdf603KSupplemental Figure 12. Summary of results. In the wild-type, undifferentiated spermatogonia can be classified at least two groups according to the cellular state and the marker gene expression: one is NANOS2+GFRA1+ (yellow) and the other is NANOS3+Ngn3+ (blue). (A): In normal NANOS2+GFRA1+ cells (yellow), NANOS2 suppresses differentiation in the downstream of GFRA1. The proliferation of these cells is balanced by a positive regulation of GDNF signaling and a negative regulation of NANOS2. After losing NANOS2/GFRA1 expressions (blue), spermatogonia might be in more mature state, and ready for the next steps of differentiation (red right arrow). (B): NANOS2 induced by Nanos3-Cre blocks differentiation of NANOS2+GFRA1+ cells (yellow) into NANOS3+Ngn3+ state (blue). (C): NANOS2 is induced only in NANOS3+Ngn3+ cells (blue) by Ngn3-Cre. Both differentiation and proliferation are suppressed even in the absence of GFRA1. NANOS2+GFRA1+ population (yellow) is normally maintained and enters into NANOS3+Ngn3+ state. (D): In Gfra1-cKO mice, differentiation-related genes, such as NANOS3 is abnormally up-regulated in NANOS2+GFRA1+ spermatogonia (yellow), which may cause precocious activation of differentiation pathway (red right arrow) and loss of stem cell population. (E): Nanos2-overexpression by Cre-Ert2 suppresses differentiation and proliferation of both NANOS2+GFRA1+ (yellow) and NANOS3+Ngn3+ (blue) population. (F): Nanos2-overexpression can block differentiation progression of Gfra1-null cells (yellow) and delay the stem cell loss phenotypes of these mice. However, spermatogonial population is eventually lost because of a strikingly lower level of proliferation and a constant level of apoptotic cell death. Proliferation of NANOS3+Ngn3+ cells is supported by unknown growth factor (X) in the absence of GFRA1. Abbreviations: GFRA1, glial cell line-derived neurotrophic factor family receptor alpha 1; OE, overexpression; Tg, transgene; Ngn3, Neurogenin3.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.