Distinct and Cooperative Roles of amh and dmrt1 in Self-Renewal and Differentiation of Male Germ Cells in Zebrafish

Abstract

Spermatogenesis is a fundamental process in male reproductive biology and depends on precise balance between self-renewal and differentiation of male germ cells. However, the regulative factors for controlling the balance are poorly understood. In this study, we examined the roles of amh and dmrt1 in male germ cell development by generating their mutants with Crispr/Cas9 technology in zebrafish. Amh mutant zebrafish displayed a female-biased sex ratio, and both male and female amh mutants developed hypertrophic gonads due to uncontrolled proliferation and impaired differentiation of germ cells. A large number of proliferating spermatogonium-like cells were observed within testicular lobules of the amh-mutated testes, and they were demonstrated to be both Vasa- and PH3-positive. Moreover, the average number of Sycp3- and Vasa-positive cells in the amh mutants was significantly lower than in wild-type testes, suggesting a severely impaired differentiation of male germ cells. Conversely, all the dmrt1-mutated testes displayed severe testicular developmental defects and gradual loss of all Vasa-positive germ cells by inhibiting their self-renewal and inducing apoptosis. In addition, several germ cell and Sertoli cell marker genes were significantly downregulated, whereas a prominent increase of Insl3-positive Leydig cells was revealed by immunohistochemical analysis in the disorganized dmrt1-mutated testes. Our data suggest that amh might act as a guardian to control the balance between proliferation and differentiation of male germ cells, whereas dmrt1 might be required for the maintenance, self-renewal, and differentiation of male germ cells. Significantly, this study unravels novel functions of amh gene in fish.

MALE sex determination is initiated by sex-determining genes that activate downstream factors essential for testis development and spermatogenesis, such as amh, dmrt1, and sox9 (Herpin and Schartl 2015; Mei and Gui 2015). Spermatogenesis is a dynamic process, including self-renewal of spermatogonia via mitotic division, meiotic division of spermatocytes, and metamorphosis of spermatids (Xu et al. 2010). In the seminiferous tubules where spermatozoa are produced, the chief cells are male germ cells and Sertoli cells (Rossitto et al. 2015). Male germ cells are able to differentiate into spermatozoa, and their development relies on the direct membrane contact with Sertoli cells or paracrine signals secreted by Sertoli cells, such as bone morphogenetic proteins that are members of the transforming growth factor β (Tgfβ) family (Saitou 2009; Dimitriadis et al. 2015; Wang et al. 2016).

The subtle balance between self-renewal and differentiation of male germ cells is essential for normal male reproductive development and function, loss of which usually leads to male infertility (Rossitto et al. 2015). In mammals, germ cell maintenance requires some factors produced by Sertoli cells including GDNF and cyp26b1 (Meng et al. 2000; Bowles et al. 2006), and other important factors expressed in spermatogonial stem cells, such as plzf, nanos2, and dmrt1 (Tsuda et al. 2003; Buaas et al. 2004; Zhang et al. 2016). Follicle-stimulating hormone is a principal hormonal regulator of spermatogenesis and its receptor is also expressed by Sertoli cells (Marshall et al. 1995). Importantly, mutation of follicle-stimulating hormone receptor (fshr) led to female-to-male sex reversal, whereas double mutation of fshr and LH receptor gene (lhcgr) resulted in infertile males in zebrafish (Zhang et al. 2015b).

In teleost fish, amh, dmrt1, and their orthologs had been observed to involve male sex determination and spermatogenesis (Xia et al. 2007; Li et al. 2014a,b). Amh, a member of the Tgfb superfamily, was shown to express in Sertoli cells and to be responsible for regression of Müllerian ducts in mammals (Kobayashi et al. 2011). Amh orthologs were widely identified in teleost fish. In medaka, amh expression was observed to start from somatic gonadal mesoderm when the migrating PGCs meet with somatic gonadal precursors (Nakamura et al. 2006). In zebrafish, amh was revealed to preferentially localize in Sertoli cells around type A spermatogonia, whereas its expression weakened and finally became absent when the primary spermatogonia differentiated, suggesting that amh in Sertoli cells might inhibit differentiation of male germ cells (Skaar et al. 2011). Dmrt1 orthologs were also investigated in the model fish. Dmy, the duplicate of dmrt1, was confirmed as male sex-determining gene in medaka (Matson and Zarkower 2012). And, dmrt1 mutation was revealed to lead a male-to-female sex reversal in the XY mutant gonads (Masuyama et al. 2012). In zebrafish, dmrt1 was found to express in both germ line and Sertoli cells of the testes, and its mutation generated by TALEN technology was recently demonstrated to lead to testis development defects and female-biased sex ratio (Webster et al. 2017). However, the functional and cooperative roles between amh and dmrt1 in germ cell development remained largely unclear in vivo. Here, we used CRISPR/Cas9 technology (Li and Wang 2017; Xiong et al. 2017) to construct amh and dmrt1 mutants and thereby investigate the functional roles and regulative mechanisms of amh and dmrt1 in male germ cell development.

Materials and Methods

Zebrafish lines and maintenance

AB line zebrafish used in this study were reared in the zebrafish facility of the Institute of Hydrobiology, Chinese Academy of Sciences. The maintenance, breeding, and staging of zebrafish were performed according to standard methods (Kimmel et al. 1995; Westerfield 2000). The use of zebrafish was approved by the Animal Research and Ethics Committees of the Institute of Hydrobiology.

Targeted gene disruption by CRISPR/Cas9

Gene knockout by CRISPR/Cas9 was performed as described (Xiong et al. 2017). The guide RNA (gRNA) target sites were designed with an online tool, ZIFIT Targeter (http://zifit.partners.org/ZiFiT/CSquare9Nuclease.aspx), and gRNA was transcribed with the TranscriptAid T7 High-Yield Transcription Kit (Thermo Fisher Scientific, Waltham, MA). For Cas9 messenger RNA (mRNA) synthesis, pCS2-Cas9 expression vector was linearized with XbaI, purified, and transcribed using the mMESSAGE mMACHINE T7 ULTRA kit (Ambion) according to the manufacturer’s instructions. Cas9 mRNA (300 ng/µl) and gRNA (20 ng/µl) were co-injected into one-cell-stage zebrafish embryos, and the mutations were analyzed by genomic PCR and sequencing (primers used were listed in Supplemental Material, Table S1 in File S1) as described (Yang et al. 2017).

Fertilization assessment

Fertilization rate assessment of male zebrafish was performed as described (Brion et al. 2004) with minor modifications. In the morning, dmrt1 and amh mutant male zebrafish were crossed with wild-type (WT) females, respectively. After breeding, the eggs were collected, and the total number of embryos were recorded and placed in the incubator. The unfertilized eggs and dead embryos were counted and removed out at 4 and 24 hr postfertilization. Fertilization rate is the number of fertilized embryos divided by the total embryos, in percentage. Each experiment was performed more than three times.

Tissue collection and histological analysis

For histological analysis, intact testes at different developmental stages were carefully dissected after anesthetization and fixed in 4% paraformaldehyde for up to 24 hr at 4°, dehydrated, and embedded in paraffin. Testes were sectioned at 4-μm thickness and stained with hematoxylin and eosin (H&E) following the manufacturer’s instructions. Histological analyses and the staging of spermatogenesis were performed as described previously (Leal et al. 2009; Sun et al. 2013). Each experiment was performed on at least three individual zebrafish in triplicate, and a series of sections was observed.

Antibody preparation and immunofluorescence analysis

The partial ORF of insl3 was fused with pET-21b (Novagen) and the recombinant fusion protein was expressed in Escherichia coli. After purification, the expressed protein (1 μg/μl) was emulsified in isopyknic Freund’s complete adjuvant (Sigma, St. Louis, MO) and injected into the lymph of rabbits. Two weeks later, the rabbits were subcutaneously booster immunized three times every 14 days, with equal purified protein emulsified in Freund’s incomplete adjuvant. Ten days later, antiserum was collected and stored at −80° as described (Peng et al. 2009; Li et al. 2017).

WT and mutant testes at different stages were collected and fixed as above. After deparaffinization, slides were rehydrated in PBS for 20 min, repaired in EDTA antigen retrieval buffers, and incubated for 1 hr with 5% defatted milk diluted in PBS at room temperature to avoid nonspecific binding of antibodies. Then, slides were incubated overnight at 4° with primary antibodies: anti-Vasa antibody (rabbit, 1:50) (Liu et al. 2015), anti-Insl3 antibody (rabbit, 1:100), phosphohistone 3 (PH3) antibody (rabbit, 1:1000) (3377S, Cell Signaling Technology, Danvers, MA), anti-Sycp3 antibody (rabbit, 1:200) (ab150292) diluted in 1% defatted milk in PBS. After washing, the slides were incubated for 1 hr with (FITC)- or tetramethyl rhodamin isothiocyanate (TRITC)-conjugated secondary antibody (1: 150) (Thermo Fisher Scientific) in the dark at room temperature. Finally, the samples were counterstained with DAPI (Sigma) to visualize cell nuclei and photographed using a Leica SP8 confocal microscope (Leica, Jena, Germany).

Probe synthesis and in situ hybridization

Zebrafish nanos2 complementary DNA (cDNA) fragment was amplified by specific primers with a T7 RNA polymerase promoter (Table S1 in File S1) and a DIG RNA labeling kit (Roche, Mannheim, Germany) was used for probe synthesis. The intact testes were fixed with 4% paraformaldehyde in PBS at 4° overnight and immersed in 30% saccharose-PBS for up to 24 hr at 4°. Then the samples were embedded in optimal cutting temperature (OCT) and sectioned at 7-μm thickness with freezing microtome (Leica). After drying at 37° for 1 hr, the sections were stored at −80°. The mRNA detection by in situ hybridization was performed as previously described (Wang et al. 2004, 2013; Dranow et al. 2016), and the sections were photographed on a Zeiss Axio Observer A1 inverted microscope (Leica).

RNA isolation and quantitative real-time PCR

Testis RNA isolation and quantitative real-time PCR (qPCR) analysis were performed to evaluate the expression changes of sex-associated genes, between WT and the corresponding mutant testes at different stages. Total RNA samples were isolated from zebrafish testes at different stages (WT and amh mutant testes at 180 days postfertilization (dpf); WT and dmrt1 mutant testes at 27, 33, 43, 54, 75, and 180 dpf) using the SV Total RNA Isolation System (Promega, Madison, WI). The isolated RNAs were respectively reverse transcribed with the GoScript Reverse Transcription System (Promega) according to the manufacturer’s instruction. Then, qPCR reactions were performed on the CFX96 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA) using iQ SYBR Green Supermix (Bio-Rad) as previously described (Neumann et al. 2011). Elongation factor 1α (ef1α) was used as the endogenous control.

The selected sex-associated genes included germ cell markers, vasa (Yoon et al. 1997), dnd (Weidinger et al. 2003), piwil1 expressed in all generations of type A spermatogonia (Chen et al. 2013), dazl mainly expressed in type B spermatogonia (Chen et al. 2013), plzf expressed in both type A and B spermatogonia (Ozaki et al. 2011), nanos2 expressed in premeiotic spermatogonia (Beer and Draper 2013), and sycp3 expressed only in spermatocytes (Ozaki et al. 2011). The Sertoli cell-related genes are amh, sox9a (Rodríguez-Marí et al. 2005; Skaar et al. 2011), dmrt1 (Webster et al. 2017), gsdf (Gautier et al. 2011), and ar (de Waal et al. 2008). The steroidogenic genes expressed in Leydig cells are star, hsd3b, cyp11b, and cyp17a1 (Wang and Orban 2007; Arukwe 2008; Hinfray et al. 2011; Baudiffier et al. 2012). Each experiment was analyzed in triplicate and the data were analyzed using the 2^−ΔΔCt program. Primer sequences are shown in Table S1 in File S1.

Analysis of apoptosis

Apoptosis in WT and mutant testes was detected by TUNEL using the In-Situ Cell Death Detection Kit, Fluorescein (Roche) according to the manufacturer’s instructions. TUNEL and specific antibody double labeling was performed as described (Sweeney et al. 2012). The slides were photographed on a Leica SP8 confocal microscope and the number of apoptotic cells on each section was counted as described (Mei et al. 2008).

Statistical analysis

All experiments were performed at least in triplicate. More than three animals were analyzed in each experimental group. The data were shown as mean ± SD, and P < 0.05 was considered statistically significant after using Student’s t-test to identify statistical differences between the groups. Statistical analysis was performed with SPSS software and graphs were plotted using GraphPad Prism 6 (La Jolla, CA).

Data availability

Strains used are available upon request. File S1 contain the Figures S1–S7 and Table S1. All primers used in this study are listed in Table S1 in File S1.

Results

Generation and characterization of amh-mutated zebrafish

Zebrafish amh encodes a protein with 549 amino acids, containing TGF-β domain at the C terminus and AMH domain in the middle. To investigate the function of amh, CRISPR/Cas9 was performed to generate amh-mutated zebrafish. The targeting site was chosen in the sixth exon to disrupt both the AMH domain and TGF-β domain. The sequencing results showed that several types of mutations have passed through the germline (data not shown). Finally, 5-bp deletions (named amhΔ5) and 17-bp insertions (named amh+17) were identified and chosen to establish mutant lines (Figure S1A in File S1). The genotypes of WT and two homozygotes were confirmed by genomic sequencing (Figure S1B in File S1). The mutations in amhΔ5 and amh+17 resulted in the ORF shift and caused premature termination codons. Both mutants lost the AMH domain and TGF-β domain (Figure S1C in File S1).

Loss of amh function caused the dysregulation of sexual development. We found all homozygous amh mutant fish displayed female-biased sex ratios. The percentage of females was ∼71% (46 females and 19 males) in the homozygotes, whereas the percentages of females were 46% (51 females and 59 males) in the heterozygotes and 40% (22 females and 33 males) in WT (Figure S2 in File S1). The observed genotype distribution among the progenies was as expected (χ² = 0.545, n = 230, P = 0.761), and the observed sex distribution of homozygotes was significantly different from that of heterozygotes (χ² = 8.887, n = 175, P = 0.003) or WT (χ² = 10.267, n = 120, P = 0.001). Thus, we confirm that amh is required for normal sex ratios in zebrafish.

Loss of amh function causes dysregulation of germ cell development

At 2.5 months of age, amh mutant males had the normal body shape of WT, but an enlarged abdomen was observed at 4 and 6 months of age (Figure 1A). WT zebrafish males exhibited normal secondary sex characteristics, including slender body, darker yellow pigmentation, and presence of breeding tubercles (Dranow et al. 2013). However, the yellow fin pigmentation was significantly reduced in amh-mutated males (Figure 1A). WT zebrafish testes almost achieved a maximum size at 2.5 months old. In contrast, amh mutant testes were found to continue growing after sexual maturity (Figure 1B). Amh mutant males had normal reproductive capacity at 2.5 months old, even though the GSI (gonadosomatic index) was slightly increased. However, the homozygous mutated males gradually lost reproductive capacity by 3 months of age (Figure 1C). After dissection, we observed that abdomens of amh mutants were occupied by hypertrophic testes, while the other internal organs were squeezed (Figure 1, D and E). To identify the testicular structure defects in amh mutants, we analyzed the histology of hypertrophic testes with H&E and immunofluorescent staining. At 6 months of age, WT testes exhibited normal testicular structure, consisting of spermatogenic cysts with spermatogonia, spermatocytes, spermatids, and mature spermatozoa. And, successive differentiation stages of germ cells were observed (Figure 1F). In contrast, abnormal structures were detected in the hypertrophic testes, in which numerous spermatogonium-like cells were filled, whereas only a few spermatids and spermatozoa were seen (Figure 1, G and H). Anti-Vasa immunofluorescence localization indicated that the hypertrophic testes in amh mutants were filled with much more Vasa-positive germ cells and fewer spermatids and spermatozoa (Figure 1, J and K) than the control testes (Figure 1I).

Figure 1.

Loss of amh leads to the dysregulation of male germ cell development. (A) Comparison of the body shape and fin pigmentation between WT males and amh mutant males (amh^−/−) at 2.5, 4, and 6 months of age. (B) Gonadosomatic index (GSI = gonad weight / total body weight, in percentage) changes of WT and amh mutant males at 2.5, 4, and 6 months of age (mean ± SD, n = 6). (C) Fertilization rate changes of WT and amh mutant male zebrafish by crossing with WT female at 2.5, 3, 4, and 6 months of age (mean ± SD, n = 5). (D) Anatomical observation of WT and amh mutant testes (arrows). (E) Morphological comparison of WT and amh mutant testes. (F–H) Histological comparisons of testes from WT, amhΔ5, and amh+17 by H&E staining at 6 months of age. (I–K) Immunofluorescence of testis sections stained with anti-Vasa antibody at 6 months of age. Sg, spermatogonia; Sc, spermatocytes; St, spermatids; Sz, spermatozoa. Bar, 100 μm.

Phenotypes and gonadal tissues of amh-mutated females were also investigated. Homozygous females had similar phenotypes to homozygous males, including enlarged abdomen and hypertrophic ovary (Figure 2A). At 6 months of age, WT ovaries showed all stages of oocyte maturation from primary growth oocyte to mature oocyte. In contrast, the hypertrophic ovaries of homozygous females were filled with lots of primary growth oocytes and the oocyte maturation was arrested at an early stage, indicating that loss of amh function resulted in an excessive proliferation of germ cells and a profound oocyte maturation defect (Figure 2A). And, the homozygous females had a significant increase in GSI compared with the control (Figure 2B). We further confirmed that these accumulated germ cells are premeiotic oogonia by RNA in situ hybridization with a premeiotic germ cell marker nanos2 (Beer and Draper 2013) at 4 months of age (Figure 2C, a and b). Taken together, amh deficiency causes overproliferation and impaired differentiation of germ cells. Thus, amh is essential for regulating germ cell differentiation in both male and female zebrafish.

Figure 2.

Amh mutant females develop hypertrophic ovaries because of overproliferation and differentiation damage of premeiotic oogonia. (A) Gross morphology (a and b) and ovary histological examination (c and d) of WT female and amh mutant at 6 months old. (B) GSI changes of WT and amh mutant females at 5 months old (mean ± SD, n = 4). (C) nanos2 RNA in situ hybridization of WT (a) and amh mutant (b) ovaries at 4 months old. Premeiotic oogonia (PO) are marked by arrows. Bar, 100 μm. *** P < 0.001.

We further checked the cell proliferation by immunofluorescence with PH3 antibody, a mitosis marker for cell proliferation, in both WT and amh mutant testes. Two adjacent sections (4 μm) were immunohistologically stained with PH3 and Vasa antibody, respectively. As expected, there was a large increase in the number of PH3-positive cells in amh-deficient testes (Figure 3A). Almost all of the PH3-positive cells were Vasa-positive in amh mutants (Figure 3A, d–f). Statistical analysis showed that the average number of PH3-positive cells per section in amh mutants (mean = 8496, n = 3) was ∼21-fold higher than in WT (mean = 400, n = 3), suggesting that the germ cells might have an excessive proliferation (Figure 3B). Then we examined the state of germ cell differentiation in the amh-mutated testes by immunohistological staining with antibody of Sycp3, a meiosis marker expressed in spermatocytes. Two adjacent sections (4 μm) were stained with Sycp3 and Vasa, respectively (Figure 3C). The results showed that Sycp3-positive cells were enriched in the WT testes, and most Vasa-positive cells were Sycp3-positive (Figure 3C, a–c). By contrast, in amh-deficient testes, the germ cells lacked or had very low levels of Sycp3 expression, and only a few Vasa-positive cells were Sycp3-positive (Figure 3C, d–f), suggesting that the accumulated germ cells should be premeiotic germ cells. Statistical analysis indicated that the average number of both Sycp3- and Vasa-positive cells per section in WT (mean = 8167, n = 3) was ∼7-fold higher than in amh mutants (mean = 1166, n = 3) (Figure 3D), revealing a severely impaired differentiation of male germ cells.

Figure 3.

Amh deficiency causes overproliferation and differentiation damage of male germ cells. (A) Immunofluorescence staining of WT (a–c) and amh mutant (d–f) testis sections at 6 months old with PH3 antibody (a) and Vasa antibody (b), respectively. c is the merge of a and b. (B) Quantification of average PH3-positive cell numbers per section in WT and amh mutants (mean ± SD, n = 3). (C) Immunofluorescence staining of WT (a–c) and amh mutant (d–f) testis sections at 6 months old with Sycp3 antibody (a) and Vasa antibody (b), respectively. c is the merge of a and b. (D) Quantification of average Sycp3-positive cell numbers per section in WT and amh mutants (mean ± SD, n = 3). Bar, 75 μm. *** P < 0.001.

Histologic analysis of testes was performed at 3 months of age when amh mutants began to lose the reproductive capacity. In WT males, successive spermatogenic cells, including spermatogonia, spermatocytes, spermatids, and spermatozoa (from the edge to the center of lobules) were evident, and small clusters of spermatogonia (arrows) were observed to locate on the testicular lobule edge (Figure 4A), whereas in both amhΔ5 and amh+17 mutated testes, ectopic clusters of spermatogonium-like cells (arrows) were detected among the spermatozoa (Figure 4, B and C). Moreover, these ectopic spermatogonium-like cells among the spermatozoa were confirmed as Vasa-positive (Figure 4, D–F) and PH3-positive (Figure 4, G–I). These data indicate that amh function loss causes dysregulation of germ cell development and lead to the overproliferation of spermatogonia in amh-mutated testes.

Figure 4.

Amh mutant testes produce a number of ectopic spermatogonium-like cells. (A–C) H&E staining of testis sections from WT, amh+17, and amhΔ5 zebrafish at 3 months old. (D–F) Immunofluorescence staining of WT, amh+17, and amhΔ5 testis sections with anti-Vasa antibody and DAPI at 3 months old. (G–I) Immunofluorescence staining of mutant testis sections with PH3 antibody and DAPI at 3 months old. I is the merge of G and H. Spermatogonia in WT and ectopic spermatogonium-like cells in amh mutants are marked by arrows. Bar, 25 μm.

Abnormal expression changes of male germ cell differentiation-related genes and gonadal somatic cell development-related genes in amh mutant testes

Although amh mutant males had normal reproductive capacity at ∼2.5 months of age, histologic analysis of testes observed the number increase and cluster formation of spermatogonia (arrows) at 2 months of age (Figure 5, A and B). And, the result was confirmed by significantly increasing expression signal of nanos2, a marker enriched in zebrafish spermatogonia. Compared with the WT control (Figure 5C), ectopic clusters of nanos2 mRNA expression (arrowhead) were detected within the lobules (Figure 5D). Insl3 is a somatic Leydig cell marker and stimulates spermatogonial differentiation in adult zebrafish testes (Good-Avila et al. 2009; Assis et al. 2016). We further checked the expression of Insl3 by immunohistochemical staining. In WT testes, many clusters of Leydig cells were present in the interstitial area (Figure 5, E–G), whereas only a few single Leydig cells among the disorganized germ cells were detected in amh mutants (Figure 5, H–J), suggesting that Leydig cell differentiation failed in amh-deficient testes.

Figure 5.

Amh deficiency causes dysregulation of male germ cell differentiation-related genes and gonadal somatic cell development-related genes. (A and B) H&E staining of WT testis (A) and amh mutant testis (B) at 2 months old. (C and D) In situ hybridization for spermatogonium-specific marker nanos2 in WT testis (C) and amh mutant testis (D) at 2 months old. (E–J) Immunofluorescence staining of WT (E–G) and amh mutant (H–J) testis sections with Insl3 antibody at 6 months old. (K) Expression changes of male germ cell differentiation-related genes and gonadal somatic cell development-related genes in amh mutant testes at 6 months old. Spermatogonia in WT are marked by arrows and ectopic clusters of spermatogonia in amh mutants are marked by arrowheads. Bar, 100 μm. * P < 0.05, ** P < 0.01, *** P < 0.001.

Moreover, we investigated the expression changes of several male germ cell differentiation-related genes and gonadal somatic cell development-related genes in amh mutant testes by qPCR at 6 months of age (Figure 5K). In seven germ cell marker genes, six spermatogonium marker genes, namely, vasa, dnd, piwil1, dazl, plzf, and nanos2, were significantly up-regulated, whereas meiosis marker gene sycp3, only expressing in spermatocytes, was downregulated in the amh-mutated testes (Figure 5K, left). These results were consistent with our histological observation that spermatogonia had excessive proliferation and spermatogenic cell differentiation was severely impaired. Four Sertoli cell marker genes, including dmrt1, sox9a, gsdf, and ar, were significantly upregulated in amh-mutated testes. However, four Leydig cell marker genes, namely, star, hsd3b, cyp11b, and cyp17a1, were significantly downregulated, suggesting that amh might be required for androgen synthesis because of their association with steroidogenesis. These data suggest that amh should play crucial roles in germ cell differentiation and spermatogenesis because loss of amh function leads to abnormal expression changes of male germ cell differentiation-related genes and gonadal somatic cell development-related genes.

Generation and characterization of dmrt1-mutated zebrafish

To investigate the functions of dmrt1 in zebrafish germ cell development, the CRISPR/Cas9 system was used to generate dmrt1-mutated zebrafish. The targeting site was chosen in the first exon and closed to the start codon. Finally, 14-bp deletions (named dmrt1Δ14) and 11-bp insertions (16-bp insertions plus 5-bp deletions, named dmrt1+11) were identified and chosen to establish mutant lines (Figure S3A in File S1). The genotypes of WT and two kinds of homozygotes were confirmed by genomic sequencing (Figure S3B in File S1). The mutations in dmrt1Δ14 and dmrt1+11 resulted in the ORF shift and caused premature termination codons (Figure S3C in File S1).

The dmrt1 mutants displayed female-biased sex ratios, because the percentages of females were ∼47 and 79% in the heterozygotes and homozygotes, respectively (Figure S4 in File S1). The observed genotype distribution among the progenies was as expected (χ² = 0.020; n = 102; P = 0.889) and the observed sex distribution of homozygotes was significantly different from that of heterozygotes (χ² = 10.141; n = 102; P = 0.001). We found that the homozygous dmrt1 mutant females were fertile and exhibited the same normal ovarian development as WT females (Figure S5 in File S1). Significantly, both the dmrt1Δ14 and dmrt1+11 mutant males were viable (Figure 6A), but thin thread-like testicular structures were observed in the mutated adults compared with typical testes in the control (Figure 6B). And, all the dmrt1-mutated males were infertile (Figure 6C). We also found a significant GSI decrease in homozygous mutants compared with WT (Figure 6D). Compared with normal testicular architecture (Figure 6E) in WT control adults, both the dmrt1Δ14 and dmrt1+11 mutant testes showed severe histological abnormalities, including some disorganized fibroblast-like somatic cells and diffused vacuolation in the severely hypoplastic testes (Figure 6, F and G). At 8 months old, the WT testes still had two normal lobes (Figure 6H); however, ∼50% heterozygous males displayed partially regressed testis lobes relative to WT (Figure 6I), and all homozygous testes had severely regressed lobes with a mass of adipose tissue (Figure 6, J and K). In addition, H&E staining showed a large number of disorganized fibroblast-like somatic cells in the regressed testes (Figure S6 in File S1).

Figure 6.

Phenotypic and testis tissue analysis of dmrt1 mutants. (A) Gross morphology comparisons between WT and dmrt1 mutant males. (B) Anatomical observation of WT (left) and dmrt1 mutant (right) testes at 75 dpf. (C) Fertilization rate difference of WT and dmrt1 mutant males by crossing with WT females at 90 dpf (mean ± SD, n = 8). (D) GSI difference of WT and dmrt1 mutant males at 75 dpf (mean ± SD, n = 7). (E–G) Histological observation of testes from WT (E), dmrt1Δ14 (F), and dmrt1+11 (G) by H&E staining at 90 dpf. (H–K) Anatomical observation of testes in WT, heterozygotes (dmrt1^+/−), and homozygotes (dmrt1^−/−) at 240 dpf. Testes are marked by black arrows and adipose tissue associated with testes is marked by white arrows. Sg, spermatogonia; Sc, spermatocytes; St, spermatids; Sz, spermatozoa. Bar, 20 μm. *** P < 0.001.

Dmrt1 deletion impairs male germ cell development and testicular lobule formation

Subsequently, we investigated the detailed defects of dmrt1 deletion during testis development and spermatogenesis. At 28 dpf, WT testes showed abundant Vasa-positive male germ cells, including different stages of spermatogonia (sg) (Figure 7, A and B), whereas dmrt1-mutated testes displayed fewer spermatogonia and residual primary oocytes (po) (Figure 7, C and D). At 34 dpf, testicular lobules with spermatogonia, spermatocytes (sc) and spermatids (st) appeared in WT testes (Figure 7, E and F). Male germ cells were significantly reduced in dmrt1-mutated testes, and the degenerated oocytes (do) still existed, indicating that morphological alteration from ovarian follicle to testis cord failed (Figure 7, G and H). At 43 and 54 dpf, WT testes showed typical testicular structure, including well-organized spermatogonia, spermatocytes, and spermatids (Figure 7, I, J, M, and N). However, mutant testes were completely devoid of normal testicular structure and lost all Vasa-positive germ cells, with only disorganized somatic cells (Figure 7, K, L, O, and P). Compared with WT testes, there was an apparent increase in the number of Leydig cells in the mutant testes, as revealed by immunofluorescent staining with antibody of Insl3, a marker of Leydig cell (Figure 7, Q and R). These results indicate that dmrt1 deletion impairs male germ cell development and testicular lobule formation.

Figure 7.

Dmrt1 is required for male germ cell development and testicular lobule formation. (A–P) Histological observation of WT and dmrt1 mutant testes at 28 dpf (A–D), 34 dpf (E–H), 43 dpf (I–L), and 54 dpf (M–P) by H&E staining (A, C, E, G, I, K, M, and O) and Vasa antibody immunofluorescence staining (B, D, F, H, J, L, N, and P) of WT and dmrt1 mutant testes. (Q and R) Immunofluorescence staining of WT (Q) and dmrt1 mutant (R) testis sections with anti-Insl3 antibody at 54 dpf. (S) H&E staining of dmrt1 mutant testis at 54 dpf. Numbers of animals in each experimental group: n = 4. Sg, spermatogonia; Sc, spermatocytes; St, spermatids; Sz, spermatozoa; Po, primary oocytes; Do, degenerated oocytes. Bar, 25 μm.

Dmrt1 is required for proliferation and maintenance of male germ cells

To better understand the effects of dmrt1 deficiency on testis development, we first performed immunofluorescent staining with PH3 antibody, a mitosis marker for cell proliferation. At 28 dpf, PH3-positive proliferating cells were enriched in the WT testes (Figure 8A, a and b). In contrast, few PH3-positive cells were detected in the dmrt1 mutant testes (Figure 8A, c and d). At 33 and 53 dpf, the depletion of PH3-positive proliferating cells was observed in the dmrt1 mutants (Figure 8A, g, h, k, and l), compared to a number of positive proliferating cells in the WT testes (Figure 8A, e, f, i, and j). Statistical analysis indicated that the average number of PH3-positive cells in WT (mean = 47.14, n = 7) was ∼64-fold higher than in dmrt1 mutants (mean = 0.73, n = 11) at 33 dpf, revealing a significant decrease of proliferating cells in the dmrt1 mutant testes (Figure 8B).

Figure 8.

Defective cell proliferation and abnormal male germ cell development in dmrt1 mutant testes. (A) Immunofluorescence staining of WT and dmrt1 mutant testis sections with PH3 antibody and DAPI at 28 dpf (a–d), 33 dpf (e–h), and 53 dpf (i–l). (B) Quantified results of average PH3-positive cell numbers per section in WT and dmrt1 mutant testes at 33 dpf (mean ± SD, numbers of animals used: WT = 7 and mutant = 11). (C) Immunofluorescence staining of WT and dmrt1 mutant testes with Sycp3 antibody and DAPI at 18 dpf (a–d), 26 dpf (e–h), 30 dpf (i–l), and 40 dpf (m–p). (D) Quantified results of average Sycp3-positive cell numbers per section in WT and dmrt1 mutants at 18, 26, 30, and 40 dpf (mean ± SD, n = 3). Bar, 25 μm. *** P < 0.001.

The development of male germ cells was further examined by immunohistological staining with antibody of Sycp3, a meiotic marker of germ cell. At 18 dpf, we did not observe any Sycp3-positive cells in the bipotential gonads of both WT and mutant zebrafish (Figure 8C, a–d). However, at 26 dpf, the number of observed Sycp3-positive cells was less in dmrt1-mutated testes compared with the control (Figure 8C, e–h). At 30 dpf, male germ cells were significantly reduced in dmrt1-mutated testes as we only detected a few of Sycp3-positive cells. In contrast, WT testes exhibited more signals of Sycp3 (Figure 8C, i–l). At 40 dpf, although Sycp3-positive cells were enriched in the WT testes, we did not find any Sycp3 signals (Figure 8C, m–p). Thus, the gradual loss of Sycp3 expression in mutant testes supported the idea that dmrt1 is essential for normal development of male germ cell in zebrafish testes (Webster et al. 2017).

The apoptosis of male germ cells and somatic cells were evaluated by fluorescent TUNEL analysis. At 27 and 31 dpf, compared to the sporadic signals in the testes of WT control (Figure 9A, a, b, e, and f), we found that the number of TUNEL-positive cells was significantly increased in dmrt1-mutated testes (Figure 9A, c, d, g, and h), in spite of the reduced mRNA expression of p53 (Figure S7 in File S1). Quantification of TUNEL-positive cells in testis sections showed that the average number of apoptotic cells in dmrt1 mutant testes (mean = 26.13, n = 8) was about four-fold higher than in WT (mean = 6.8, n = 5) at 31 dpf, suggesting an abnormal increase of cell apoptosis in dmrt1 mutant testes (Figure 9B). To determine the type of apoptotic cells, we co-stained Vasa-positive cells by antibody immunofluorescence and apoptotic cells by TUNEL in both WT and dmrt1 mutant testes at 28 dpf. Almost no apoptotic signals were detected in the germ cells of control testes (Figure 9C, a–c), but a large number of germ cells were found to be co-labeled with Vasa and TUNEL signals within dmrt1 mutant testes (Figure 9C, d–f), indicating that dmrt1 is required for the maintenance of early male germ cells. In addition, some TUNEL-positive cells were not observed to be Vasa-positive, implying apoptosis occurrence of some somatic cells (Figure 9C, d–f). Thereby, we further checked the status of somatic cells by double staining with anti-Insl3 (Leydig cell marker) antibody and TUNEL. Many more TUNEL-positive cells were found in dmrt1-deficient testes compared with the control at 27 dpf, but none of them was Insl3-positive, suggesting that these apoptotic cells were germ cells and other somatic cells such as Sertoli cells in mutant testes (Figure 9C, a–d). At 54 dpf, when mutant testes were completely devoid of germ cells, only a few TUNEL-positive cells were detected in both mutant and WT testes, and they were not Insl3-positive (Figure 9C, e–h), whereas a large number of Insl3-positive Leydig cells did not undergo apoptosis in dmrt1-deficient testes. Therefore, these results indicate that dmrt1 is specifically required for proliferation and maintenance of male germ cells.

Figure 9.

Apoptosis is induced in the dmrt1 mutant testes. (A) TUNEL and DAPI staining in testis sections of WT and dmrt1 mutants at 27 dpf (a–d) and 31 dpf (e–h). (B) Quantified results of average number of apoptotic cells assessed by TUNEL in WT and dmrt1 mutants at 31 dpf (mean ± SD, numbers of animals used: WT = 5 and mutant = 8). (C) Double immunofluorescence staining for TUNEL (in green) and germ cell marker Vasa antibody (in red) in WT and dmrt1 mutants at 28 dpf. (D) Triple fluorescence staining with anti-Insl3 antibody, TUNEL, and DAPI in WT and dmrt1 mutant testes at 27 dpf (a–d) and 54 dpf (e–h). Bar, 25 μm. *** P < 0.001.

Loss of dmrt1 leads to dysregulation of germ cell-related genes and gonadal somatic cell-related genes

To demonstrate that dmrt1 plays a key role in both germ line and somatic cell development of zebrafish testes, the expression changes of some germ cell-related genes and gonadal somatic cell-related genes were evaluated by qPCR between WT and the corresponding dmrt1-mutated testes during gonad development (Figure 10). At 27 dpf, the germ cell marker genes, including vasa, piwil1, dnd, and sycp3, were upregulated, whereas dazl was downregulated in dmrt1-mutated testes, compared with WT control. However, at 33 dpf, the expression of all germ cell marker genes was significantly reduced, except that dnd was slightly upregulated in dmrt1-mutated testes. At the later stages of testis development, including 43, 54, and 75 dpf, there was almost no transcript expression of these five germ cell marker genes in dmrt1-mutated testes (Figure 10, A–E). Next, we checked the expression dynamics of four Sertoli cell-related genes. Sox9a, gsdf, and ar were first upregulated at 27 dpf, then gradually downregulated at 33, 43, and 54 dpf. Finally, sox9a disappeared at 54 dpf and gsdf disappeared at 75 dpf. Strikingly, the expression of amh was absent at all stages (Figure 10, F–I). Further, we analyzed expression changes of two key steroidogenic genes hsd3b and cyp17a1 that were proposed to express in Leydig cells (Figure 10, J and K). Their dynamic expression did not show a relatively consistent trend, while cyp17a1 expression displayed an increasing trend in the dmrt1-mutated testes compared to the control, especially at later stages. Moreover, we constructed a graphical heat map to show the expression changes of all 11 investigated genes during testis development when the function of dmrt1 was lost (Figure 10L). Our data indicated that a compensation effect might be activated at early stages, such as 27 dpf in all cell lineages of the dmrt1-mutated testes, but the expression changes of three kinds of genes were consistent with our histological results that germ cells and Sertoli cells were absent or impaired (Figure 7), whereas numerous Insl3-positive Leydig cells existed in residual testes of the dmrt1 mutants (Figure 9). Therefore, their expression changes suggest that dmrt1 should be required for the maintenance of both germline genes and Sertoli cell genes during testis development.

Figure 10.

Disruption of dmrt1 expression causes dysregulation of germ cell-related genes and gonadal somatic cell-related genes. (A–K) qPCR analysis of these genes including vasa, piwil1, dazl, dnd, sycp3, sox9a, amh, gsdf, ar, cyp17a1, and hsd3b, between WT and the corresponding dmrt1 mutant testes at 27, 33, 43, 54, and 75 dpf (n = 3). (L) Heat map constructed to compare the expression changes of all 11 investigated genes during testis development.

Discussion

Germ cells have been extensively studied in recent years, because they provide the genetic linkage between generations in multicellular organisms (Xu et al. 2010; Gui and Zhu 2012). Here we investigated the roles of amh and dmrt1 in zebrafish germ cell development by CRISPR/Cas9 technology. All the male and female amh mutants developed hypertrophic gonads. Amh-mutated males showed hypertrophic testes with excess proliferation of Vasa-positive spermatogonium-like cells within the lobules. Moreover, the hypertrophic testes were filled with disorganized nanos2-positive spermatogonia and the number of Sycp3-positive cells was severely reduced, suggesting the uncontrolled proliferation and impaired differentiation of male germ cells. Dmrt1-mutated testes displayed a progressive loss of spermatogenesis, in which both germ cell and Sertoli cell marker genes were significantly downregulated or completely inhibited at later stages, whereas Leydig cell marker genes were upregulated (Figure 10), and the increase expression was consistent with immunohistochemical observation by Leydig cell marker Insl3 antibody (Figure 9D). Based on these findings, we propose a hypothesized model in which amh and dmrt1 cooperatively regulate zebrafish male germ cell self-renewal and differentiation. As shown in Figure 11, dmrt1 is required for the maintenance and self-renewal of early male germ cells, because dmrt1 mutation impairs male germ cell development and testicular lobule formation and leads to regressed testis, whereas amh is decisive for controlling the balance between proliferation and differentiation of male germ cells, because amh function loss causes overproliferation and accumulation of spermatogonia and results in hypertrophic testis (Figure 11A). Therefore, dmrt1 activates amh expression, whereas amh negatively regulates dmrt1 expression. Moreover, dmrt1 maintains self-renewal and proliferation of early male germ cells, while amh inhibits excessive proliferation of spermatogonia and maintains normal differentiation. In other words, amh and dmrt1 cooperatively keep the balance of spermatogenesis by regulating self-renewal, proliferation, and differentiation of male germ cells (Figure 11B).

Figure 11.

Schematic diagram for the cooperative regulation between amh and dmrt1 in zebrafish male germ cell self-renewal and differentiation. (A) Histological and cytological characteristics of WT, dmrt1^−/−, and amh^−/− testes at different stages. (B) Detailed diagram about the cooperative interactions and distinct functions of dmrt1 and amh in keeping the balance of spermatogenesis by regulating self-renewal, proliferation, and differentiation of male germ cells. Sg, spermatogonia; Sc, spermatocytes; St, spermatids; Sz, spermatozoa.

Initially, amh was detected in Sertoli cells and responsible for regression of Müllerian ducts in mammals (Kobayashi et al. 2011). What are the functions of amh orthologs in fish species? In vitro studies indicated that zebrafish Amh production in Sertoli cells was preferentially localized around type A spermatogonia and inhibited the differentiation of male germ cells (Skaar et al. 2011). In medaka hotei mutants that lack the function of Amh-type II receptor (AmhrII), one-half of the adult XY homozygous fish were sex reversed into females, and the other half displayed hypertrophic testes (Morinaga et al. 2007). In the hypertrophic testes of medaka hotei mutants, a hyperproliferation of undifferentiated type A spermatogonia was detected in both juvenile and adult stages, but the number of mitotically quiescent germline cells was normal (Morinaga et al. 2007; Nakamura et al. 2012). So far, the specific receptors for Amh, such as AmhrII, had been found only in a couple of fish species (Rocha et al. 2016). However, AmhrI and AmhrII had not been found in the zebrafish genome. Similar to the phenotypes in medaka hotei mutants (Morinaga et al. 2007), our zebrafish amh mutants showed hypertrophic testes (Figure 1) and highly uncontrolled proliferation of premeiotic spermatogonia (Figure 1 and Figure 3), indicating the conserved roles of Amh/Amhr signaling in male germ cell maintenance and differentiation. Particularly, our studies revealed that the germ cells in amh-deficient testes lacked or had very low levels of Sycp3 expression (Figure 3, C and D), confirming the deficiency of spermatogonium differentiation. Besides its roles in testis development, the Amh signal was also revealed to regulate female germ cell differentiation in zebrafish ovary, as an overproliferation of premeiotic oogonia was observed in the hypertrophic ovaries of 4-month-old amh-deficient females (Figure 2, C and D). Therefore, our current results suggest that amh might act as a guardian to regulate differentiation of both male and female germ cells in zebrafish, because its loss of function leads to overproliferation of both premeiotic spermatogonia and oogonia and thereby forms hypertrophic tumor-like testes or ovaries in the amh-deficient males or females (Figure 1 and Figure 2).

Different from mammals and most other fish species, the undifferentiated gonads in zebrafish first differentiate as “bipotential juvenile ovaries,” and then half of the juvenile ovaries undergo apoptosis and regression and eventually develop into testes during 20–30 dpf (Wang et al. 2007; Orban et al. 2009). In zebrafish, primary germ cells are required for female sex differentiation, because loss of primary germ cells led to an all-male sexual phenotype in adults (Slanchev et al. 2005). Recently, several genes have been revealed to be involved in zebrafish male sex determination and testicular differentiation. Disruption of zebrafish fancl and fancd1 (brca2) gene functions caused female-to-male sex reversal by p53-mediated germ cell apoptosis (Rodríguez-Marí et al. 2010, 2011). However, nr0b1 (DAX1) mutation in zebrafish resulted in female-to-male sex reversal in a p53-independent manner (Chen et al. 2016). Disruption of foxl2a and foxl2b function led to the sex reversal from females to males (Yang et al. 2017). Dmrt1 and its paralog gene dmy are involved in male sex determination and differentiation in medaka. Dmrt1 mutation resulted in the sex reversal of XY gonads into ovary type (Matsuda et al. 2002; Masuyama et al. 2012). In this study, the dmrt1-mutated zebrafish generated by CRISPR/Cas9 approach also displayed female-biased sex ratios (Figure S4 in File S1) as recently described by Webster et al. (2017). In comparison with the recent report that germ cells were still present in adult dmrt1 mutant males, our current data observed severely regressed testes with only a mass of adipose tissue (Figure 6, J and K) and complete loss of all germ cells at later stages (Figure 7). The phenotypic difference might be the reflection of different genetic backgrounds between the previous TU strain and the AB strain. Additionally, we performed anit-Vasa immunofluorescence localization, and our current results were confirmed by both general H&E staining and anit-Vasa immunofluorescence staining. Therefore, our current data suggest that dmrt1 should be required for proliferation and maintenance of male germ cells, because dmrt1 deletion resulted in loss of all male germ cells and destroyed testicular lobule formation (Figure 6 and Figure 7), and the germ cell damage and testicular tissue destruction were confirmed by severe expression suppression of some germ cell-related genes and gonadal somatic cell-related genes at later stages (Figure 10).

Expressed in Sertoli cells and germ cells, mouse dmrt1 plays a germ cell-autonomous role and a Sertoli cell-nonautonomous role to promote germ cell development (Kim et al. 2007; Agbor et al. 2013; Zhang et al. 2016). In dmrt1-mutated zebrafish testes, we detected the reduced expression of sox9a, amh, and gsdf. Therefore, it is concluded that the male germ cell defects in dmrt1 mutants are due to loss of dmrt1 function in both of these cell types. Regardless of male germ cell defects in dmrt1 mutants, our data showed that the expression of two key steroidogenic genes, hsd3b and cyp17a1, had no obvious change, which is why we could observe the normal secondary sex characteristics, including slender body, yellow fin pigmentation, and breeding tubercle. It has been shown that Sertoli cells can be reprogrammed to other cell types, such as pluripotent stem cells and fetal-like Leydig cells (Sun et al. 2014; Zhang et al. 2015a). In dmrt1-mutated testes, there are more Insl3-positive Leydig-like cells than the control testes. Whether these Leydig-like cells are reprogrammed from other cell types need further studies.

In conclusion, our current data indicate that dmrt1 positively regulates amh expression, whereas amh negatively regulates dmrt1 expression, and thereby forms a negative feedback loop of amh and dmrt1. Significantly, dmrt1 mutants were simultaneously found to lead to complete inhibition of amh expression in the examined testes (Figure 10, G and L), implying that dmrt1 mutants might have same phenotypes with double mutants of dmrt1 and amh. Therefore, amh and dmrt1 cooperatively regulate self-renewal and differentiation of male germ cells in zebrafish.

Supplementary Material

Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.117.300274/-/DC1.