Exosome-Derived microRNAs in Sertoli Cells Inhibit Spermatogonial 1 Apoptosis 2

23 Background: Spermatogenesis is a highly complicated biological process that occurs in the 24 epithelium of the seminiferous tubules. It is regulated by a complex network of endocrine and 25 paracrine factors and juxtacrine testicular cross-talk. Sertoli cells (SCs) play a key role in 26 spermatogenesis due to their production of trophic, differentiation and immune-modulating factors. 27 However, many of the molecular pathways of SCs action remain controversial and unclear. Recently, 28 research has focused on exosomes as an important mechanism of intercellular communication. 29 Results: We found that the exosomes derived from SCs (SC-Exos) significantly inhibited the 30 apoptosis of the primary spermatogonial stem cells (SSCs). Total of 1016 miRNAs in SCs and 556 31 miRNAs in SC-Exos were detected using microRNA (miRNA) high-throughput sequencing. Further, 32 294 miRNAs were differentially expressed between SCs and SC-Exos. Based on the GO and KEGG 33 analyses, the target genes of 37 (high-expressed in exosomes and RPM>1000) selected miRNAs 34 were involved in multiple biological pathways, including the MAPK signaling pathway and 35 PI3K/AKT signaling pathway. And miR-10b is one of the top ten exosomes with relatively large 36 enrichment of microRNA. In addition, the overexpression of miR-10b down-regulated expression of 37 the target KLF4 to reduce spermatogonial apoptosis in SSCs or C18-4 cell line. 38 Conclusions: The study indicated a large number of small RNAs loaded in exosomes was secreted 39 form the donor SCs to target spermatogonial regulated the apoptosis. And miR-10b inhibits the 40 apoptosis of spermatogonia through the target gene KLF4. sequenced via high-throughput sequencing and the differences between exosomal microRNAs and Sertoli cellular microRNAs were comparatively analyzed by bioinformatics. Further, the function of the specific high-abundance exosomal microRNA was revealed by examining the expression of target gene which was identified involve in regulating proliferation and apoptosis of spermatogonial stem cells. The findings of this study would provide a new insight about the SCs regulating the niche homeostasis of spermatogonial stem cells via exosomes or exosomal microRNAs.

Introduction 46 spermatogonial stem cells (SSCs) develop sequentially into spermatogonia, spermatocytes, 8 were sputter coated with palladium gold and viewed under a SEM. For transmission electron 162 microscopy (TEM), samples of exosomes were observed using previously described methods [25]. 163

Nanoparticle tracking analysis 164
Size distribution and quantification of exosome were analyzed by NanoSight NS3000 165 instrument (Malvern, England) as described previously [26]. Briefly, exosomes from SCs were 166 diluted in 1mL of DPBS and disaggregated by using a syringe and needle. Then the sample was 167 injected into the chamber and three individual samples were tested. 168

RNA extraction, RT-PCR and real-time PCR 169
Total RNA was extracted from the testes, cells and exosomes using TRIzol reagent (Invitrogen) 170 and reverse transcription was performed for cDNA using Takara Reverse Transcription kit according 171 to the manufacturer's instructions. sRNA-specific stem-loop RT primers were used to synthesize 172 cDNA. The PCR mixture consisted of 1 μL of the cDNA sample, 10 μL primestar buffer, 4 μL dNTP 173 mixture, 1μL of each primer (0.1mM), 1 μL primestar hs DNA polymerase and 33 μL water. The 174 PCR reactions started at 98 °C for 3 min, and then denature at 98 °C for 30 s, anneal at 58-62 °C for 175 5 s and enlongate at 72 °C for 30 s for 30 cycle. 176 Transcript level of klf4, cd63, sox9, α-sma, plzf and 3β-hsd were performed by PCR in a 20 µL 177 volume, containing 10 μL 2×mix, 0.5 μL each of forward and reverse primers, 1 μL cDNA, and 8 μL 178 ddH 2 O. The PCR protocol was as follows: initial denaturation for 5 min at 98 °C; followed by 30 179 cycles of denaturation for 30 s at 94 °C, annealing for 30 s at 60 °C, extension for 30 s at 72 °C and a 180 final extension for 10 min at 72 °C, with subsequent cooling to 4 °C. The products were performed 181 by agarose gel electrophoresis. gapdh gene was used as internal control gene for mRNA expression 182 normalization. 183 9 TaqMan MicroRNA assays were used to quantify the level of mature sRNAs. The 20 µl volume 184 included 1 µl RT product, 10 μl of 2 × TaqMan Universal PCR Master Mix, 0.4 µM TaqMan probe, 3 185 µM forward primer and 1.5 µM reverse primer. The reactions were started at 95°C for 10 min, 186 followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. All reactions were run in triplicate. The 187 results were performed via the 2 -△△ ct method in terms of the protocol as described previously [27]. 188 The CEL-miR-39 spike-in control (GenePharma, China) was used as additional internal reference for 189 sRNA expression normalization in exosomes and SCs [28,29]. For this purpose, 50 fmol of CEL-190 miR-39 was added to the RT reaction and was measured by qPCR using specific primers. Differences 191 were considered statistically different at p < 0.05 and analysis were performed using t-test. All of the 192 primers are presented in Sup: Table S1. 193

Bioinformatics Analysis . 194
Total RNA was isolated from SCs and exosomes from the SCs, using the Trizol reagent (Takara,195 Japan), following the manufacturer's instructions. The quantity and quality of the total RNA were 196 determined using a NanoDrop® 174 ND-1000 spectrophotometer (Thermo Fisher Scientific, USA) at 197 260/280 nm (ratio>2.0), and its integrity was tested using 2100 Bioanalyzer and RNA 6000 Nano 198

Culture of the Sertoli cells and identification of its' exosomes 240
The purified SCs from testes of 5-7 dpn mouse exhibited typical morphology of epithelial cells 241 and strong growth activity by a serial of different plating ( Figure 1A). RT-PCR results displayed that 242 the purified cells expressed sox9 (a marker for SCs), cd63 (a marker for exosomes), while not plzf (a 243 marker for spermatogonia) and 3β-hsd (a marker for Leydig cells) ( Figure 1B). And the 244 immunofluorescence experimental showed that the SOX9 and GATA4 positive cells were more than 245 95% ( Figure 1C). The SCs were cultured for following experiments. 246 The EVs were isolated from the medium by a serial of gradient ultracentrifugation after SCs 247 were cultured in DMEM/F12 medium supplemented with 10% KnockOut SR-Multi-Species for 48 h 248 ( Figure 1A). The EVs were circular double-layer membrane vesicles by SEM and TEM ( Figure 2A  249 and 2B). the diameter size of EVs is about 122.5 ± 4.1 nm, the size of most EVs was at 100 nm by 250 the Nanosight assay and the concentration of EVs was about 10 12 particles/mL. ( Figure 2C). 251 12 Meanwhile, CD63 and HSP70 were detected in the EVs by Western blot ( Figure 2D). The results 252 showed that the EVs from in vitro primary SCs were a typical class of exosomes. 253

Exosomes from Sertoli cells maintain spermatogonial cells survival 254
To transwell membrane and be transferred to recipient cells through culture medium (Sup: Figure 1). 262 The final concentration 100 ng/μL DiI-labaled exosomes from SCs were added into serum-free 263 DME/F-12 medium for culturing primary spermatogonial cells or C18-4 cells. The fluorescence (DiI-264 labaled exosomes) was detected as a granular pattern surrounding cells ( Figure 3A). These 265 experiments indicated the exosomes secreted by mouse SCs can target the spermatogonial cells. 266 Comparing with control (the medium without exosomes), spermatogonial cells or C18-4 cells 267 added exosomes are in better state that the cells are plump and smooth after 4 d of culture, the cell 268 viability increased by CCK8 analysis ( Figure 3B). The rates of apoptosis of primary spermatogonial 269 cells and C18-4 cells treated with exosomes were significantly down-regulated by Annexin V/FITC 270 staining ( Figure 3C). However, effect of exosomes on proliferation of C18-4 cell intimated no 271 significant by the quantitative analysis of EdU staining (Sup: Figure 2). In conclusion, exosomes 272 secreted from SCs can target spermatogonia to inhibit apoptosis, but no effect on proliferation. 273 274   13 Exosomes from specific cells are involved in regulating the function of receptor cells through 275 sorting of specific miRNAs. Small RNA transcriptomes from SCs and the exosomes were sequenced 276 by means of Illumina HiSeq TM 2500 platform and the data were analyzed by bioinformatics. The 277 number of raw reads of the SCs were 14,271,751 and the exosomes were 15,549,820, respectively. 278

Small RNA profiles of Sertoli cells and exosomes
After removing low-quality sequences, smaller sequences, contaminants which were formed by 279 adapter-adapter ligation, the clean reads obtained from the SCs were 13,046,249 and from exosomes 280 libraries were 14,348,194 (Table 1). The mapped of reads were 12,586,353 (96.47% of clean reads) 281 and 11,286,869 (78.66% of clean reads) for samples of SCs and exosomes, respectively. 282 Pearson correlation analysis showed a moderate correlation between the SCs and exosomes 283 (correlation coefficient, 0.733) ( Figure 4A). The sequences common by SCs and exosomes were 284 3.59% (32,630), cell-specific were 65.26% (592,652), and exosomes-specific were 31.15% (282,834) 285 ( Figure 4B). Analyzed the size distribution between 17 and 40 nt of reads, showed that 3 peaks at 21-286 23 nt for both SCs and exosomes. In addition, exosomes had a peak at 33 nt ( Figure 4C). After 287 mapping to mouse genome, the proportion of annotated miRNAs was 15.8% in SCs, and 6.3% in 288 exosomes. The annotated tRNAs accounted for 2.5% in SCs, 1% in exosomes. The rRNA in 289 exosomes was 4.5% more than in SCs (8.4% in exosomes, 3.9% in SCs) ( Figure 4D). 290

Comparing analysis of miRNAs between the Sertoli cells and its' exosomes 291
The sequencing results displayed that there were total 1016 miRNAs in SCs and total 556 292 miRNAs in exosomes. 524 miRNAs of them were shared in SCs and exosomes ( Figure 5A). Besides, 293 we found that 32 miRNAs were only detected in exosomes and 492 miRNAs in SCs were not sorted 294 into exosome ( Figure 5A, Table 2). 295 Comparing the expression of miRNAs between the SCs and exosomes (p value < 0.05 and fold 296 change > 2), a total 294 miRNAs were differentially expression between the SCs and exosomes. 203 297 14 of which were more highly expressed in SCs and 91 of which were more highly expressed in the 298 exosomes ( Figure 5B and Sup: Table S2). 299 Furthermore, analyzing reads per million (RPM) of miRNAs, the expression abundance of the 300 top 10 miRNAs reached 82.33% and 89.23 % of total miRNAs RPM in SCs and in exosome, 301 respectively. Comparing the top 10 miRNAs between in SCs and in exosome, five of them in 302 exosomes did not appear in the top 10 miRNAs of SCs (such as miR-10b-5p), although other five 303 miRNAs in SCs occupied in the top 10 miRNAs of exosomes ( Figure 5C and 5D). 304

GO annotation and KEGG pathway analysis of the target genes of the high-expression 305 expressed miRNAs in exosomes 306
To understand the function of miRNAs in exosomes, the target genes of the high-expressed 307 miRNAs (RPM>1000) were predicted 3,809 (Sup: Table S3). The target genes of the high-expression 308 expressed miRNAs in exosomes were analyzed by GO annotation and KEGG pathway analysis (Sup: 309 Table S4 and S5). The results identified the more target genes of the candidate miRNAs were 310 associated with cellular component and biological process by GO annotation ( Figure 6A). In KEGG 311 pathway annotation, the number of miRNAs-target genes were mainly involved in the axon guidance, 312 pathways in cancer, FoxO signaling pathway, cell cycle, proteoglycans in cancer, Hepatitis B, 313 metabolic pathways, PI3K-Akt signaling pathway, endocytosis and MAPK signaling pathway 314 ( Figure 6B). 315

Experimental validation of miRNAs expression 316
In this study, the expression levels of the differentially expressed miRNAs showed similar 317 trends between the RNA-seq and RT-qPCR results (Figure 7). A total of 8 uniquely expressed and 318 differently expressed miRNAs that identified from miRNA-seq were selected for validation using 319 15 stem-loop RT-qPCR (Figure 7). Four DE miRNAs showed similar trend as detected by  Specially, the expressions of miR-26a-5p and let-7i-5p were highly expressed in the SCs ( Figure 7A  321 and 7B), while miR-148a-3p and miR-10b-5p were highly expressed in the exosomes (p<0.05) 322 ( Figure 7C and 7D). miR-10b-5p seemed to be more readily enriched by exosomes. Meanwhile, 323 miR-135b-5p, miR-221-5p and miR-30b-3p only detected in SCs (the results not shown). However, 324 the expression of miR-6538 could not be detected in both exosomes and SCs by RT-qPCR (CT value 325 < 35), which may be due to their low abundance and the relative low content of miRNAs in SCs and 326 exosomes ( Table 2). 327

Exosomal miR-10b derived from SCs inhibited SSCs apoptosis and targeted KLF4 328
The expression of miR-10b was significantly higher in exosomes than in SCs (Sup: Figure 2). 329 Interestingly, there are reports miRNA-10b inhibits apoptosis of SSC through Kruppel-like factor 330 4(KLF4) [41]. To corroboration the biological function of miR-10b in spermatogonial, we transfected 331 miRNA-10b mimics to C18-4 cells for 48h, the transfection efficiency was up to 50 times higher 332 ( Figure 8A). To determine whether exosomal miR-10b inhibits SSCs apoptosis, we performed CCK-333 8 and TUNEL assays, and showed that the survival of SSCs transfected with miR-10b mimics was 334 significantly increased (p <0.01) compared with those transfected with NC ( Figure 8B and 8C). The 335 pro-apoptotic protein BAX was downregulated while the anti-apoptotic protein BCL2 was 336 upregulated. Meanwhile, the cleavage of caspase-3 was decrease after miR-10b treatment 48h 337 ( Figure 8D). These data confirm that miR-10b inhibits SSCs apoptosis. According to the previous 338 . Our results demonstrated that miR-10b is relatively more 395 highly expressed in exosomes than in SCs. We also validated the functionality of the miR-10b in 396 SSCs, the results displayed miR-10b inhibit spermatogonial apoptosis. Conversely, it has been 397 reported that miR-10b can promote the proliferation of SSC [41], but in our experimental the 398 proliferation of SSC was not regulated by exosomes, which could be because the composition of 399 exosomes is more complex, and there are other miRNAs masking the effect of miR-10b. The results 400 also showed that miR-10b inhibits the expression of KLF4 gene mRNA and protein levels. It 401 indicated that miR-10b may exert its biological function through targeting KLF4 after being 402 delivered to spermatogonia by exosomes. Study has shown that KLF4 as a transcription factor can 403 induce the totipotency of spermatogonial stem cells, implying the important role of KLF4 in germ 404 cells [58]. In addition, miR-9-5p downregulates KLF4 to inhibit the apoptosis of liver cancer cells 405 through the AKT signaling pathway [59]. KLF4 could also act on the MAPK signaling pathway 406 through ectopic expression to promote the apoptosis of melanoma cells [60]. In this experiment, the 407 enriched significantly different signaling pathways include PI3K/AKT and MAPK signaling 408 pathways. This indicates that KLF4 may affect spermatogonia apoptosis through PI3K/AKT and 409 MAPK signaling pathways. 410

Conclusion 411
In conclusion, mouse SCs delivered exosomes protected the spermatogonia by decreasing cell 412 apoptosis, and this protective effect was partly generated through the exosome miRNAs by RNA 413 sequencing. GO and KEGG analyses showed that the significant difference signal pathways include 414 19 PI3K/AKT and MAPK signal pathways which involved in regulating the apoptosis of spermatogonia. 415 The results also suggest that miR-10b inhibits the apoptosis of spermatogonia through the target gene 416 KLF4 (Figure 9).     Table S1. Stem-loop RT-PCR and qPCR primers of miRNAs (XLS 26KB). 631