Somatic cell-derived BMPs induce male germ cell meiosis initiation during embryonic stage via regulating Dazl expression

Germ cell fate is believed to be determined by the signaling from sexually differentiated somatic cell. However, the molecular mechanism remains elusive. In this study, ectopic initiation of meiosis in male germ cells was observed during embryonic stage by over-activating CTNNB1 in Sertoli cells. Somatic cell transcriptome and single germ cell RNA-seq analysis indicated that TGF-β signaling was activated after CTNNB1 over-activation. In vitro and in vivo experiments confirmed somatic cell-derived BMPs played crucial roles in germ cell meiosis initiation. Further studies revealed that Dazl was significantly increased in germ cells of CTNNB1 over-activated testes and induced by BMP signaling. DNMT3a and DNA methylation was also reduced in germ cells of CTNNB1 over-activated testes and increased by BMP signaling inhibitor treatment. Taken together, this study demonstrates that germ cell fate could be reprogrammed after sex determination. BMP signaling pathway is involved in germ cell meiosis initiation via up-regulating Dazl expression.


4
It's believed that germ cells enter meiosis spontaneously during embryonic stage 67 unless are specifically prevented by meiosis-inhibiting factors. CYP26B1 (Cyp26 68 family of cytochrome P450 oxidase 1) is first identified as a meiosis inhibitor which is 69 abundantly expressed in Sertoli cells during embryonic stage and catalyzes the 70 oxidization of RA to inactive forms (MacLean et al., 2001). Inactivation of CYP26B1 71 leads to the ectopic initiation of meiosis in male germ cells during embryonic stage 72 (Bowles et al., 2006;MacLean et al., 2007). Another meiosis-inhibiting substance is 73 FGF9 (fibroblast growth factor 9) also produced by Sertoli cells, which acts 74 antagonistically with RA to determine male germ cell fate (Bowles et al., 2010).  to-female sex reversal is observed in Fgf9-null mice, and could be rescued by Wnt4 76 (wingless-type MMTV integration site family, member 4) deletion (Jameson et al., 77 2012a). FGF9 also acts directly on germ cells by upregulating RNA binding protein 78 NANOS2 (nanos C2HC-type zinc finger 2) (Bowles et al., 2010). Nanos2 is male 79 meiotic gatekeeper specifically expressed in germ cells and plays pivotal roles in germ 80 cell sexual differentiation (Saga, 2010;Suzuki and Saga, 2008). Inactivation of Nanos2 81 results in meiotic initiation in male germ cells and upregulation of oocyte differentiation 82 associated genes (Suzuki et al., 2010;Suzuki et al., 2012). Inversely, female germ 83 cells fail to enter meiosis and start expressing male-specific genes after NANOS2 84 over-expression (Suzuki and Saga, 2008). 85 Several studies have demonstrated that TGF-β signaling is also involved in 86 regulating meiosis and germ cell development (Spiller et al., 2017;Wu et al., 2013). 87 Nodal/activin pathway is activated in both male germ cells and somatic cells which 88 5 induces Nanos2 expression. Disruption of Nodal/Activin signaling leads to male germ 89 cells meiosis and increased expression of female-specific genes in somatic cells 90 (Souquet et al., 2012;Spiller et al., 2013;Tassinari et al., 2015). Moreover, deletion of 91 Smad4 in germ cells results in female germ cells meiosis defect (Wu et al., 2016). 92 Moreover, recent study demonstrates that BMPs (bone morphogenetic proteins) direct  Our previous study demonstrates that over-activation of CTNNB1 in Sertoli cells 96 results in Sertoli to granulosa-like cells transformation (Li et al., 2017). Interestingly, in 97 this study, we find the ectopic initiation of meiosis in male germ cells of CTNNB1   Importantly, synaptonemal complex protein SYCP3, a typical marker of meiosis 126 was also abundantly expressed in the germ cells of COA testes. As shown in Figure   127 1 A, thread-like SYCP3 signal was detected in germ cells of COA testes at E16.5 and 128 E17.5 (Fig. 1Ai, j, white arrowheads), which was restrictedly expressed in the germ 129 cells of control ovaries during embryonic stage ( Fig. 1Ak-o, white arrowheads), but not 130 in the germ cells of control testes ( Fig. 1Aa-e, white arrows). To further determine the 131 meiotic stages of germ cells in COA testes, chromosome spreads of meiocytes were 132 7 prepared and stained with SYCP3 antibody. As shown in Figure 1B, germ cells at 133 leptotene, zygotene, pachytene, and diplotene stage of meiosis I were observed in 134 COA testes which was normally detected in germ cells of control ovaries during 135 embryonic stage (Fig. 1Ba). The numbers of germ cells at pachytene and diplotene 136 stages in COA testes was less than that in control ovaries (Fig. 1Bb, c). The mRNA 137 level of meiotic genes including Stra8, Dmc1, Rec8, Sycp3 and Sycp1 at E16.5 was 138 also analyzed by quantitative RT-PCR and was significantly increased in COA testes 139 compared to that of control testes, including the intrinsic factor Dazl (Fig. 1Cb). Male 140 germ cell specific gene (Dnmt3l) and pluripotency genes (Oct4 and Sox2) were 141 decreased in COA testes, whereas female germ cell specific gene Foxo3 and fetal 142 oogenesis-specific gene Sohlh2 was increased in COA testes (Fig. 1Ca). Collectively, 143 all these results indicated that germ cells in COA testes ectopically initiated meiosis 144 during embryonic stage.  (Chang et al., 2008;Zhang et al., 2015a). However, no STRA8 156 and γH2AX were detected in the germ cells of these two mouse models at E16.5 ( Fig.   157 S3D, E, G, H), only very weak SYCP3 was observed (Fig. S3F, I). These results   To explore the underlying mechanism which leads to the ectopic initiation of 9 meiosis in COA germ cells, RNA-sequencing assay was performed with isolated 177 Sertoli cells from control and COA mice at E13.5 and E14.5. We first profiled the 4Cb, e), and the signal of SYCP3 was decreased with XAV treatment (Fig. 4Cc, f).

230
Consistently, the mRNA level of other meiotic genes was also significantly decreased 231 after LDN-193189 treatment ( Fig 4D). These results suggested that both BMP and 232 WNT signaling are involved in the meiosis initiation of germ cells in COA testes, and 233 BMP signaling plays more important roles in this process.

234
The function of BMPs in germ cell meiosis was further examined using in vitro 235 cultured E13.5 control and DTA+/flox AMH-Cre testes and treated with RA, BMPs, and 236 smad4/smad5 pathway activator Kartogenin (KGN). As shown in Figure 4E, germ cells 237 with thread-like SYCP3 were detected in testes with RA+KGN treatment (Fig. 4Eb, d), 238 whereas only scattered SYCP3 foci were observed in germ cells of RA only treated 239 testes (Fig. 4Ea, c). Results of quantitative RT-PCR also showed that the mRNA level RA only treatment (Fig. 4Ga, e), whereas some germ cells with thread-like SYCP3 243 signal were observed after RA+BMP2 or RA+BMP4 treatment (Fig. 4Gb, f and c, g), 244 and RA+Kartogenin (Fig. 4Gd, h) treatment. Moreover, the mRNA level of meiotic 245 genes was also significantly increased after BMP2 and BMP4 treatment (Fig. 4H).

246
To further confirm the function of BMP signaling pathway in germ cell meiosis, Dazl is the intrinsic factor which is indispensable for RA response and meiosis 260 initiation (Lin et al., 2008). In this study, we found that the expression of Dazl was 261 significantly increased with RA+KAR treatment but not with RA treatment in E13.5 262 cultured testes (Fig. S5A). The expression of Dazl at E13.5 in DTA+/flox AMH-Cre testes 263 was also induced by BMP2 and BMP4 treatment (Fig. S5B). By contrast, the 264 13 expression of Dazl in control ovaries was significantly decreased with LDN treatment, 265 but not with XAV treatment (Fig. S5C). Moreover, the mRNA level of Dazl in RA treated 266 COA testes was significantly higher than that in RA treated control testes, which was 267 significantly decreased after LDN and XAV treatment (Fig. S5D). All together, these compared to control at E15.5 (Fig. S6Ba, c). Immunostaining results showed that 5mC 279 was significantly decreased in germ cells of COA testes at E15.5 compared to control 280 male germ cells, which was comparable to that of female germ cells (Fig. S6Ca-c).

281
The expression of DNMT3a and DNMT3L was significantly decreased in germ cells of 282 COA testes (Fig. S6Ce, n), whereas only DNMT3a was consistent with RNA-sequence 283 result and comparable to control female germ cells.

284
To test whether DNA methylation is regulated by RA and BMP signaling, 5mC level 285 and the expression of DNMT3a were examined by immunostaining in control and COA 286 14 testes. As shown, both 5mC and DNMT3a were detected in germ cells of control testes 287 (Fig. 6A and B a, g, m) and not repressed by RA treatment (Fig. 6A and Bb, h, n).

288
However, 5mC and DNMT3a were not detected in germ cells of RA treated COA testes 289 ( Fig. 6A and B c, i, o), whereas it was significantly increased after LDN and XAV+LDN 290 treatment ( Fig. A and B d, j, p, f, l, r). Additionally, the level of 5mC and DNMT3a was 291 not increased in COA germ cells with RA+XAV treatment (Fig. 6A and Be, k, q). These  , 2006;Koubova et al., 2006). It is widely believed that meiotic arrest of male germ 303 cells during embryonic stage is mainly due to the expression of RA oxidizing enzyme 304 CYP26B1 in Sertoli cells (Li et al., 2009). Interestingly, in this study, we found that 305 disruption of testicular cords or depletion of Sertoli cells in testes did not lead to 306 meiosis initiation of male germ cells during embryonic stages, suggesting that only RA 307 15 is not sufficient to initiate meiosis. There must be some other factors cooperate with 308 RA to regulate the meiosis initiation of germ cells.

309
The abnormal meiosis of male germ cells during embryonic stage is observed in 310 several mouse models with male-to-female sex reversal. In Sry knockout males, germ

322
In the present study, we found that the expression of meiosis associated genes  with AMH-Cre transgenic males (Lecureuil et al., 2002). DTA+/flox AMH-Cre mice were 374 obtained by crossing DTA+/flox (Brockschnieder et al., 2004) females with AMH-Cre 375 transgenic males. DNA isolated from adult tail tips and fetal tissues, and genotyping 376 was performed by PCR as previously described (Gao et al., 2006;Harada et al., 1999).

377
For administering LDN-193189 into pregnant mice, LDN-193189 (S2618, 378 Selleckchem) was dissolved in DMSO and then diluted in water and 2.5 mg/kg body 379 weight was injected intra-peritoneally every 12 hr from E13.5 to E16.5.  Green assay was performed with isolated RNA. Gapdh was used as endogenous 420 control. The relative level of candidate gene expression was calculated using the 421 formula 2-ΔΔCT as described in the SYBR Green user manual. The primers used for 422 RT-PCR are listed in Table S1.

442
The detailed procedure was performed as previously described (Yu et al., 2016).

443
Single-cell cDNA amplification and library construction 444 Single-cell cDNA amplification was carried out using the STRT protocol as 445 described previously (Dong et al., 2018;Fan et al., 2018). Libraries were prepared  Gene ontology and KEGG analysis was performed using DAVID Bioinformatics 454 Resources 6.8.

455
Gonads and re-aggregated tissues in vitro culture 456 The genital ridges were cultured on agarose blocks as described previously 457 (Coscioni et al., 2001;Zhang et al., 2015b). The agarose block was made with 2% 458 agarose and pre-balanced with culture medium up to 12 hrs. The fetal ovaries and 459 testes from E11.5 and E13.5 embryos were dissected and placed on the pre-balanced 460 agarose blocks. The gonads were cultured at 37°C in a saturated humidity incubator 461 22 infused with 5% CO2 in air. For in vitro treatment, LDN-193189 (S2618, Selleckchem), 462 XAV-939 (S1180, Selleckchem) and Kartogenin (S7658,Selleckchem) was dissolved 463 in DMSO and used with a respective concentration of 10 μM.

464
The genital ridges from female and male embryos were separated from 465 mesonephro and dissociated with 0.05% trypsin-EDTA at 37°C for 10 mins. After 466 neutralization with 10% FBS, wash twice with DMEM supplemented with 10% FBS.