Microcystin-Leucine Arginine Causes Cytotoxic Effects in Sertoli Cells Resulting in Reproductive Dysfunction in Male Mice

Microcystin-leucine arginine (MC-LR) is a potent toxin for Sertoli cells. However, the specific molecular mechanisms of MC-induced cytotoxicity still remain unclear. In this study, we performed a comprehensive analyses of changes of miRNAs and mRNAs in Sertoli cells treated with MC-LR. Through computational approaches, we showed the pivotal roles of differentially expressed miRNAs that were associated with cell metabolism, cellular growth and proliferation, cell-to-cell signaling and interaction and cellular movement. Ingenuity Pathway Analyses (IPA) revealed some differentially expressed miRNAs and mRNAs that may cause reproductive system diseases. Target gene analyses suggested that destruction in tight junctions (TJ) and adherens junctions (AJ) in testes may be mediated by miRNAs. Consistent with a significant enrichment of chemokine signaling pathways, we observed numerous macrophages in the testes of mice following treatment with MC-LR, which may cause testicular inflammation. Moreover, miR-98-5p and miR-758 were predicted to bind the 3′-UTR region of the mitogen-activated protein kinase 11 (MAPK11, p38 β isoform) gene which stimulates tumor necrosis factor-α (TNF-α) expression in Sertoli cells. TNF-α could interact with the tumor necrosis factor receptor 1 (TNFR1) on germ cells leading to induction of germ cell apoptosis. Collectively, our integrated miRNA/mRNA analyses provided a molecular paradigm, which was experimentally validated, for understanding MC-LR-induced cytotoxicity.

following exposure to chronic low-dose MC-LR 9 . Furthermore, we also found that MC-LR may exert its toxicity on cultured germ cells and Sertoli cells in vitro resulting in reduced cell viability [11][12][13][14] .
Testicular Sertoli cells play important roles in spermatogenesis as they nourish sperm cells and contribute to the formation of the blood-testis barrier (BTB) that depends on the existence of Sertoli-Sertoli cell tight junctions 15 . Our recent studies suggest that MC-LR can enter Sertoli cells and induce autophagy and apoptosis in Sertoli cells in vitro 11,16 . However, the underlying molecular mechanisms associated with Sertoli cell toxicity induced by MC-LR remain unclear.
miRNAs are small (~21 nucleotides) non-coding RNAs which can bind to the complementary regions in the mRNA molecules and then degrade target mRNAs or repress their translation 17 . They are expressed in a wide range of tissues in many species, and computational predications indicate that more than one third of all human genes may be miRNA targets 17 . As an important regulatory factor, miRNA regulates diverse cellular processes, including the proliferation, apoptosis, and responses to various stimuli 13,18,19 . Therefore, identification and characterization of the changes of miRNAs can reveal the dynamic regulation of cellular functions.
In order to decipher the upstream regulatory networks associated with Sertoli cell cytotoxicity induced by MC-LR, here we took the advantage of bioinformatics technology to present an integrative analyses of the whole gene expression pattern and its regulatory miRNA networks. The interacting networks revealed key miRNA/ mRNA interacting pairs that are correlated with cell apoptosis, tight junction (TJ) and adhere junction (AJ) destruction, and up-regulation of tumor necrosis factor-α (TNF-α ) expression in Sertoli cells. To further identify the miRNA/mRNA regulatory network, we conducted both in vivo and in vitro experiments. We observed that exposure to MC-LR caused BTB destruction, massive Sertoli cell and germ cell apoptosis, testicular inflammation, and autoantibody generation, resulting in oligospermia. Taken together, our integrative miRNA/mRNA analyses has provided a valuable tool for understanding effectively complex signaling networks associated with reproductive dysfunction induced by MC-LR.

MC-LR modulates miRNA profiles in Sertoli cells.
To confirm miRNA microarray data 20 , we assessed the expression of 10 miRNAs by quantitative PCR (q-PCR) (Supplementary Table S1). The data generated by the q-PCR assay were consistent with the microarray analyses, and the correlation-coefficient between the mean values of ten individuals generated by both techniques for each miRNA was statistically significant (Supplementary Figure S1A and Supplementary Table S1), indicating the reliability of the array data generated by miRNA microarray. In this study, many miRNAs associated with azoospermia, such as miR-199a-5p 21 , miR-181a 22 , miR-221 23 , miR-141 19 , and miR-429 19,24 , were found to be significantly modulated by exposure to MC-LR (Table 1). Moreover, some miRNAs involved in the mechanisms of other reproductive system diseases, including the urinary tract tumor, prostate cancer, and genital tumor, were also detected 25-28 . MC-LR induces differentially expressed mRNAs with possible molecular functions. We selected six genes for q-PCR validation; our results demonstrate that their expression changes were consistent with the microarray data (Supplementary Figure S1B). Categorizing the altered genes may facilitate the understanding of the biological processes regulated by differentially expressed genes in Sertoli cells. Gene Ontology (GO) enrichment analyses revealed that these genes with aberrant expression levels take part in many biological processes, including cell metabolism (e.g. carbohydrate metabolism, lipid metabolism, vitamin and mineral metabolism, and amino acid metabolism), cellular growth and proliferation, cell-to-cell signaling and interaction, cellular movement, gene expression, and the macromolecule biosynthetic process (Fig. 1A). Pathway enrichment analyses showed that communication between innate and adaptive immune cells, the p53 signaling pathway, tumor necrosis factor receptor 1 (TNFR1) signaling, and tight junction-related pathways were significantly enriched ( Table 2).

Effects of MC-LR on proto-oncogene expression in Sertoli cells and testes.
Immunofluorescence assay was used to identify cultured Sertoli cells. Cultured Sertoli cells were consistently positive for androgen receptor (AR), sex-determining region Y box 9 (SOX9), nuclear receptor subfamily 5, group A, member 1 (Nr5a1), and doublesex-related transcription factor 1 (DMRT1), suggesting a high purity of our cultured cells ( Fig. 2A). We examined c-Fos and c-Jun mRNA abundance in Sertoli cells following exposure to MC-LR for 1, 3, 6, 12, and 24 h by q-PCR. As shown in Fig. 2B, the mRNA levels of c-Fos and c-Jun increased immediately after 1 h of MC-LR treatment, and reached a peak at 6 h post-exposure (up to 8-fold and 5-fold for c-Fos and c-Jun, respectively. In contrast to early alteration of mRNA levels, protein expression of c-Fos and c-Jun was not significantly elevated until 6 h (Fig. 2C,D). Moreover, relatively moderate changes of their protein levels were observed compared to their mRNA abundance. Similarly, MC-LR treatment also induced c-Fos and c-Jun protein expression in testes (Fig. 2E,F).
Scientific RepoRts | 6:39238 | DOI: 10.1038/srep39238 Integrated miRNA/mRNA analyses reveal regulatory networks for differentially expressed miRNAs. We chose the top 20 up-regulated and top 20 down-regulated miRNAs as core miRNAs to construct miRNA regulatory networks. We predicted the targeted genes of miRNAs based on the transcriptome data.
When we compared the set of predicted target genes that were under the control of those top 20 up-regulated miRNAs with the set of down-regulated genes following MC-LR treatment, 258 overlapping genes were identified in these two sets (Supplementary data 1). Likewise, we identified 439 overlapping genes in the set predicted by the top 20 down-regulated miRNA and the set containing up-regulated genes following MC-LR treatment (Supplementary data 2). With respect to the molecular functions, GO analyses revealed that these target genes may participate in cellular growth and proliferation, cell-to-cell signaling and interaction, cellular assembly and organization, and macromolecule synthesis (Fig. 1B). Interestingly, GO analyses revealed overlapping of the target genes with those of differentially expressed genes, indicating that the target gene prediction is consistent with mRNA microarray analyses. Moreover, the biological functional analyses of target genes from the Ingenuity Pathway Analyses (IPA) database demonstrated that nonobstructive azoospermia was significantly correlated with these ectopic miRNAs and mRNAs (Fig. 3). Integrative network analyses of those significantly deregulated miRNAs and mRNAs generated some functional miRNA-mRNA networks modulated by MC-LR. We also constructed a sub-network of some miRNAs (Fig. 4). TNF receptor associated factor (TRAF)-interacting protein with a forkhead-associated (FHA) domain (TIFA), small glutamine-rich tetratricopeptide repeat (TPR)-containing β (SGTB) and matrix metalloproteinase-8 (MMP-8) were found to be core miRNA-regulated genes.
miRNA targets were involved in inducing Sertoli cell apoptosis. TIFA, with an up-regulated gene transcriptional level (Supplementary Figure S1B), was predicted to be a target gene of miR-409-5p and miR-128-3p; and these miRNAs all showed decreased expression upon MC-LR treatment (Supplementary Table S2). Up-regulated expression of TIFA may promote two independent apoptosis signaling pathways: the induction of   p53 causing cell cycle arrest, and the activation of caspase-8 and caspase-3 (Fig. 5A). SGTB plays a crucial role in mediating cell apoptosis via the caspase-3-dependent pathway 29 . In this study, we also observed up-regulated expression of SGTB, targeted by miR-133a-3p, miR-181a-5p, miR-409-5p, and miR-542-3p, at both the transcriptional and protein levels in Sertoli cells following exposure to MC-LR (Supplementary Figure S1B and Fig. 5B). Interestingly, MC-LR also induced increased expression of active caspase-3 in Sertoli cells following exposure to MC-LR, which may be associated with Sertoli cell apoptosis. As predicted, flow cytometric analyses showed that MC-LR treatment induced Sertoli cell apoptosis (Fig. 5C). Moreover, up-regulated expression of SGTB was also observed in the testes of mice treated with MC-LR (Fig. 5D). Immunohistochemical results confirmed that caspase-3 was located in Sertoli cells and germ cells in the testicular tissues of mice after intraperitoneal administration of MC-LR for 7 days (Fig. 5E). This is consistent with the analysis by terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling (TUNEL) assay demonstrating increased numbers of apoptotic cells with increasing doses of MC-LR administered (Fig. 5F,G).

Figure 2. Induction of proto-oncogene expression in Sertoli cells and testes by microcystin-leucine arginine (MC-LR). (A)
Mouse Sertoli cells were collected and cultured for 2 days. Representative immunofluorescence analyses of the marker proteins of Sertoli cells are shown (× 600). Nuclei were stained with DAPI (blue), and phalloidin (orange) was used to label the actin. Sertoli cells were treated with 500 nm MC-LR for various durations as indicated. (B) Changes of c-Fos and c-Jun transcript were examined by q-PCR, and relative mRNA abundance was calibrated based on the mRNA amount of the corresponding gene in the untreated group (mean ± SD. *P < 0.05, compared with the control). (C,D) Expression of c-Fos and c-Jun in Sertoli cells was detected by western blotting. Representative gel electrophoresis bands are shown (C), and the expression levels of the proteins were quantified by densitometry and normalized to the expression of GAPDH (D). Densitometry data are shown as mean ± SD. *P < 0.05, compared with the control. Male mice were injected intraperitoneally with saline or MC-LR (7.5 μ g/kg body weight (bw), 15 μ g/kg bw, and 30 μ g/kg bw) for 7 days. (E,F) The expression levels of c-Fos and c-Jun were analyzed by western blotting. Representative gel electrophoresis bands are shown (E), and the expression levels of the proteins were quantified by densitometry and normalized to the expression of GAPDH (F). Densitometry data are shown as mean ± SD. *P < 0.05, compared with the control.
Scientific RepoRts | 6:39238 | DOI: 10.1038/srep39238 MC-LR treatment causes testicular inflammation. MC-LR significantly increased TNF-α expression in a time-dependent manner in cultured Sertoli cells (Fig. 6A). Levels of TNF-α peaked 6 h after treatment with MC-LR. Moreover, TNF-α was also significantly enriched in the testes of mice 7 days after intraperitoneal administration of MC-LR (Fig. 6B). Despite the fact that TNF-α was predominantly localized in Sertoli cells in the testes of control mice (Fig. 6C), substantially enhanced signals of TNF-α were detected in the testes of treated mice. Based on these results, we can confirm that MC-LR is able to induce TNF-α expression in Sertoli cells. The fibrosis levels in testicular biopsies of mice after exposure to low-dose MC-LR for 90 days were evaluated with Masson's trichrome stain. Low-dose chronic exposure to MC-LR exhibited thickened basement membranes, suggesting enhanced fibrosis of the testicular tissue (Fig. 6C).
Monocyte chemoattractant protein-1 (MCP-1) expression levels in MC-treated Sertoli cells and testis tissue were significantly up-regulated compared with the control group (Fig. 7A,B). Immunostaining and flow cytometric analyses demonstrated that MC-LR exposure significantly modulated macrophage numbers in the testes of the treated mice compared with controls ( Fig. 7C,D).

Integrative miRNA/mRNA analyses reveal molecular alteration associated with BTB destruction induced by MC-LR.
In this study, we observed that expression levels of occludin, TUBB, β -catenin, and ZO-1 were significantly decreased in Sertoli cells following exposure to 500 nm MC-LR, suggesting damaged tight junctions between Sertoli cells (Fig. 8A). We next explored the effect of MC-LR on the BTB in testes. MC-LR significantly reduced the expression of occludin, TUBB, β -catenin, and ZO-1 assessed by western blotting (Fig. 8B) and immunofluorescence assay (Fig. 8C). BTB integrity can be evaluated by monitoring the diffusion of a biotin tracer from the interstitium to the seminiferous epithelium. In the present study, the biotin tracer was restricted to the interstitial space and the seminiferous tubule basal compartment in the saline-injected testes. In contrast, the biotin tracer was found to diffuse through the BTB into the seminiferous tubules after MC-LR treatment for 7 days, indicating disruption of BTB integrity by MC-LR (Fig. 8D). In this study, we also found that MMP-8 expression was up-regulated in Sertoli cells and testes following exposure to MC-LR (Fig. 8A,B).

BTB breakdown promotes generation of autoantibodies against male germ cell antigens.
Breakdown of the BTB exposes germ cells to immune surveillance, resulting in systemic autoimmunity. Sperm autoimmunity is an important contributing factor in male infertility. Sera of mice treated with MC-LR recognized Ingenuity pathway analysis (IPA) was performed to describe functional relationships among miRNAs and genes based on the known associations in the databases, and some differentially expressed miRNA and mRNA were found to be associated with nonobstructive azoospermia. The value below the nodes represents fold changes of expression levels of miRNA or mRNA between the treated and control groups. male germ cells in contrast to the negative staining by the sera from control mice (Fig. 9A). These results support the generation of autoantibodies against male germ cell antigens in mice treated with MC-LR. We next analyzed the number of B lymphocyte in the testicular interstitium of control and treated mice using flow cytometry. MC-LR significantly increased B-cell numbers in the testes, which may be associated with the generation of autoantibodies against male germ cells (Fig. 9B).

MC-LR induces TNF-α expression in Sertoli cells via the MAPK11/ATF2 signaling pathway.
MC-LR treatment promoted the expression of MAPK11, p-p38 MAPK, and p-ATF2 at the protein level in Sertoli cells following exposure to MC-LR (Fig. 10A,B). Furthermore, treatment with SB203580, a selective inhibitor of p38 MAPK, significantly suppressed the activation of ATF2 and the production of TNF-α (Fig. 10C). Taken together, our data support that MC-LR stimulated TNF-α expression via promoting the MAPK11/ATF2 signaling pathway. Through integrative miRNA/mRNA expression profiling, we found that miR-758 and miR-98-5p may regulate MAPK11 expression (Fig. 10D), and decreased expression of these two miRNAs in Sertoli cells (Fig. 10E) and testes of mice ( Fig. 10F) may promote MAPK11 protein expression. MAPK11 signaling increased not only the phosphorylation but also the protein levels of ATF2, thus activating the TNF-α gene transcription (Fig. 10G).

TNF-α induces germ cells apoptosis via interacting with TNFR1.
As MC-LR could stimulate the secretion of TNF-α by Sertoli cells, we were intrigued to confirm that TNF-α could bind to TNFR1 on the surrounding germ cells to induce apoptosis. To this end, mouse germ cells (GC-1 cells) were treated with TNF-α (20 ng/mL) for various durations. Furthermore, MC-LR caused a significant increase of TNFR1, cleaved caspase-8, and active caspase-3 in a time-dependent manner (Fig. 11A,B). We next performed coimmunoprecipitation (Co-IP) assay to confirm the interaction of TNF-α with TNFR1 which peaked at 6-24 h after MC-LR treatment (Fig. 11C,D), which coincided with apoptosis of GC-1 cells (Fig. 11E).

Discussion
In this study, we analyzed the miRNA expression profile in Sertoli cells after MC-LR treatment. A set of sensitive biomarkers for monitoring and assessing the toxic effects of MC-LR on organisms can be developed based on the miRNA response profiling, which may help to predict potential MC-LR-mediated diseases. Many miRNAs  associated with reproductive system diseases or tumors were found to be significantly modulated by exposure to MC-LR (Table 1). Zhao et al. reported that MC-LR can alter the expression of a number of miRNAs and proteins involved in several pathways related to tumorigenesis 30 . Combining these results, we infer a possible role of miRNAs in the development or establishment of MC-LR-mediated diseases. Moreover, GO enrichment analyses revealed that those differentially expressed genes in Sertoli cells following exposure to MC-LR were involved in many biological processes (Table 2). Zhao et al. reported several proteins regulated by MC-LR in mouse liver, which are mainly involved in metabolism, ion homeostasis, cell communication, cell binding and responses to chemical stimuli 30 . Overall, these results suggest that MC-LR exposure can cause cell death as a result of systemic dysfunction of various cellular aspects.
It is reported that the p53 signaling pathway dysfunction is evident in carcinogenesis through the regulation of a series of signal transduction pathways, resulting in the occurrence and development of a variety of malignant tumors 31 . More than 50% of tumors have been identified to involve aberrant p53 signaling pathways associated with the inactivation or mutation of the p53 protein [32][33][34] . Epidemiological studies suggest that MC-LR in drinking water is associated with a high risk for liver cancer in certain areas of China where drinking water sources are potentially contaminated with MCs 32-34 . Induction of the expression of proto-oncogenes is demonstrated to have tumor-promoting activity 35 . We showed here that exposure to MC-LR induced mRNA and protein synthesis of c-Fos and c-Jun, two important members of the proto-oncogene family, in Sertoli cells and testes. Taken together, this study has provided evidence that MC-LR exposure may induce carcinogenesis in the testes.
TNF-α belongs to the TNF superfamily of cytokines which play an important role in inflammation, innate and adaptive immunity, and cell death induction 36 . In the seminiferous tubules, TNF-α secreted by Sertoli cells and germ cells are capable of inducing reversible BTB disruption, through reducing the steady-state levels of occludin and ZO-1 at the BTB 37,38 . In this study, we not only demonstrated MC-LR-induced TNF-α expression by Sertoli cells in vitro, but also demonstrated TNF-α production by tissue Sertoli cells following treatment of the mice with MC-LR. Sertoli cell-derived TNF-α may possibly act on the surrounding germ cells through autocrine or paracrine modes of action. Once TNF-α binds to the TNFR1 of germ cells, the TNFR1 signaling pathway can be activated triggering germ cell death 39 , resulting in oligospermia or azoospermia in males 9,10 (Fig. 10G). Indeed we observed reduced sperm concentration after MC-LR treatment (Table 3). Using Co-IP we showed the interaction of TNF-α with TNFR1 on germ cells, which may enhance the expression of active initiator caspase-8 and active caspase-3, leading to GC-1 apoptosis. However, it should be noted that Sertoli cells are not the only source of TNF-α in the seminiferous tubules. Infiltrated macrophages are also capable of synthesizing TNF-α in the testes. Moreover, we have observed TNF-α positive germ cells in the testes after MC-LR administration, suggesting one more source of TNF-α . MAPK11 can phosphorylate the transcription factor ATF2, which stimulates TNF-α expression by binding to its promoter 40 . Through integrative miRNA/mRNA expression profiling, we found that miR-758 and miR-98-5p may regulate MAPK11 expression. MAPK11 signaling increased not only the phosphorylation but also the protein levels of ATF2, thus activating the TNF-α gene transcription (Fig. 9G).
Pathway enrichment analyses showed that some hepatic fibrosis-related genes were significantly modulated by MC-LR exposure. Fibrotic thickening in the walls of the seminiferous tubules in the testes of men with impaired spermatogenesis is observed, making fibrosis a hallmark of male infertility 41 . In testes, thickened basement membranes may disrupt diffusion of nutrients into the seminiferous tubules, resulting in germ cell apoptosis and loss 41 . In this study, thickened basement membranes were observed in mice following low-dose chronic MC-LR exposure. We believe that long-term exposure to low dosage MC-LR may initiate chronic inflammation in the testes, contributing to testicular fibrosis and finally resulting in oligospermia or azoospermia in males. For the immune system to recognize pathogens, tumor cells or dead cells and launch effector responses, leukocytes such as neutrophils, macrophages, and T-lymphocytes are recruited to the precise location of inflammation regulated by the interaction between chemokines and receptors 42 . In this study, the chemokine signaling pathway was found to be significantly enriched in the testes of mice following exposure to MC-LR. Increased numbers of macrophages were observed in the testes of the treated mice. Macrophages can secrete a variety of proinflammatory cytokines (e.g. TNF-α , IL-6, IL-1β ) that may induce testicular tissue damage, germ cell apoptosis and BTB disruption 36,38,[43][44][45] . MCP-1 is an important chemokine that can attract immune effector cells to the injured sites to regulate many physiological processes such as development, wound repair, and immunity 46 . Significant enrichment of MCP-1 in the testes of treated mice can promote immune cell infiltration into the testes. However, there was not significant change of transcriptional level of MCP-1, which is not consistent with the protein measurement. This discrepancy may be attributed to the post-transcriptional regulation of protein expression. In mammals, the BTB is formed by adjacent Sertoli cells in the seminiferous epithelium near the basement membrane via the coexisting specialized TJ, basal ectoplasmic specialization (a testis-specific atypical adherens junction), and the desmosome-like junction 15 . Several integral membrane protein complexes constitute the BTB in mammal testes, such as the occludin-ZO-1 complex at the TJ and the N-cadherin-β -catenin complex at the basal ectoplasmic specialization. In mammals, the BTB confers immune privilege upon maturing germ cells, whose disruption may lead to the production of anti-sperm antibodies, causing immunological infertility. Exposure to environmental toxicants such as cadmium, mercury, bisphenol A can lead to male infertility through the disruption of TJ as a result of redistribution of junction proteins (e.g. occludin, N-cadherin) from the cell-cell interface to the cytosol [47][48][49] . In this study, expression levels of occludin, TUBB, β -catenin, and ZO-1 were significantly decreased in Sertoli cells and the testes of mice following exposure to MC-LR; the biotin tracer assay showed that MC-LR can disrupt BTB integrity. These data support that MC-LR can perturb spermatogenesis by destroying the microenvironment of spermatogenesis.
Microtubules, located in the peripheral areas of Sertoli cells, can promote the successful transit of spermatogenic cells from the basal compartment to the apical compartment 50 . The absence of normal microtubule organization can lead to the failure of spermatogenic cell differentiation and development 51 . In Sertoli cells treated with MC-LR, the expression of TUBB targeted by miR-362-5p was significantly decreased, implicating compromised spermatogenesis by MC-LR through impeding germ cell transit from the basal compartment to the tubular lumen. AJ proteins (e.g. E-cadherin, β -catenin) are present in the seminiferous tubules to adhere germ cells onto the Sertoli cells, and their destruction causes loss of germ cells from the seminiferous epithelium 52 . Exposure to MC-LR may cause the internalization of AJ proteins; germ cells are eventually released from the seminiferous epithelium prematurely as a result of disruption of adhesion complexes (Fig. 8E). Exposure to MC-LR decreased the sperm concentration compared, which may be attributed to the release of germs from seminiferous epithelium and germ cell apoptosis ( Table 3).
As a core target gene of miRNAs, MMP-8 can degrade extracellular matrix proteins, and their aberrant expression may cause breakdown of the blood-brain barrier (BBB) inducing infiltration of peripheral immune cells, . Densitometry data are shown as mean ± SD. *P < 0.05, compared with the control. (C) Germ cell lysates were subjected to coimmunoprecipitation (Co-IP) with anti-TNFR1 antibody, and the blot was probed with anti-TNF-α antibody. Moreover, blots were re-probed with anti-TNFR1 antibody to confirm equal protein loading. Co-IP with rabbit IgG served as a negative control. The presence of TNF-α in the cell lysate was detected by western blotting, serving as a positive control. (D) Densitometric analysis of data shown in C is plotted, and the relative association at time 0 h was arbitrarily set as 1. Data are shown as mean ± SD. *P < 0.05, compared with the control. (E) The number of apoptotic cells was determined by flow cytometric analysis. Cells were stained with FITC-conjugated annexin V and PI and then were analyzed by flow cytometry. resulting in induction of neuroinflammation 53,54 . miR-193, miR-29b and miR-133a were predicted to regulate MMP-8 expression at the translational level, and a significantly down-regulated expression of these miRNAs was observed in MC-treated Sertoli cells, which may therefore promote MMP-8 expression.
Developing male germ cells produce abundant immunogenic autoantigens, which do not induce detrimental immune responses in the testes under physiological conditions because of the testicular immunoprivileged properties 44 . Once the testicular immune privilege is disrupted, immune responses against autoantigens are initiated, thereby perturbing male fertility. Generation of autoantibodies against male germ cell antigens in the sera might be due to the disruption of the BTB induced by MC-LR.
In summary, our data confirmed a miRNA-mRNA integrated network in Sertoli cells following exposure to MC-LR. Cytotoxicity induced by MC-LR may be subject to the regulation of TIFA, SGTB, MMP-8, Occludin, TUBB1, β -catenin, MAPK11, and TNFR1. The clinical relevance of these specific signature miRNAs and their targeting pairs still need further validation. Our work has laid a foundation for identifying potential molecular markers for the diagnosis and treatment of the related diseases.  Supplementary Table S3. TNF-α and MCP-1 ELISA kits were obtained from eBioscience (San Diego, CA). Annexin V-FITC apoptosis detection kit was purchased from BD bioscience (San Jose, CA).

Materials and Methods
Primary cell culture and MC-LR treatment. Sertoli cells were isolated and cultured as previously reported 11,55 . Testes from 5 mice (20-day-old) were aseptically removed, decapsulated, minced, and washed twice with sterile PBS. The minced tissues were then digested with 0.25% trypsin at 37 °C for 6 min, followed by digestion in 0.1% collagenase I at 37 °C for 10 min. Next, the homogenate obtained was filtered through a 150-μ m filter, and cells were collected by centrifugation at 300 × g for 5 min. After being washed with PBS for 3 times, the isolated Sertoli cells were re-suspended in culture medium containing 90% DMEM-F12 medium and 10% FBS and then plated on cell culture dishes. Cells were maintained in a humidified atmosphere of 95% air/5% CO2 (v/v) at 37 °C. Sertoli cells were adherent to the bottom of the dishes after culture for 2 days. Next, these cultures were subjected to a hypotonic treatment to lyse residual germ cells 15,55 . After 2 to 3 days, these cells formed a monolayer. The expression of marker proteins (AR, SOX9, Nr5a1, and DMRT1) was confirmed by immunofluorescence staining to identify the purity of cultured Sertoli cells. miRNA and mRNA expression by microarray analysis. The comparative analysis of microarray data between treated samples and control samples was carried out using independent t test. False discover rate (FDR) was further controlled by the adjustment of P value using Benjamini-Hochberg algorithm. All statistical tests and visualization of differentially expressed genes were done within R environment 56 . miRNAs with P < 0.05 and fold change ≥ 2 and ≤ 0.5 were determined to be statistically significant. Using miRNA microarray, we have previously identified 46 miRNAs that were significantly up-regulated (FDR < 0.05) and 69 miRNAs that were down-regulated (FDR < 0.05) in Sertoli cells following exposure to MC-LR for 24 h (Supplementary Table S1) 20 . The raw mRNA expression data were deposited in the Gene Expression Omnibus Database (GEO, https://www. ncbi.nlm.nih.gov/geo/) under accession number GSE87475. For analyzing mRNA microarray data, we would have obtained too few differentially expressed genes if we used a 2.0 fold-change as cutoff. In order not to lose any potential genes that were associated with the cytotoxicity induced by MC-LR, we chose to set 1.2 as the fold-change cutoff. A total of 2494 genes were differentially expressed in Sertoli cells after treatment with MC-LR, including 1037 up-regulated genes (fold change ≥ 1.20) and 1457 down-regulated genes (fold change ≤ 0.83) 20 . q-PCR validation analyses of miRNAs and target genes. Total RNA from the testes and Sertoli cells was isolated using Trizol reagent (Invitrogen). The purity of RNA was determined with a spectrophotometer (Hoefer, Holliston, MA). To examine the mRNA levels, one μ g of total RNA was then reversely transcribed using  Table 3. Quantitative analysis of sperm quality from mice treated with MC-LR. Data are shown as mean ± SD. Superscript letters (a-c) denote variances that are significantly different from each group (P < 0.05). *Response that is significantly different from the control (P < 0.05).
a first strand cDNA synthesis kit (Takara, Dalian, China). q-PCR was conducted using the SYBR Green qRT-PCR kit (Takara) on a ViiA 7 Q-PCR System (Applied Biosystems, Waltham, MA). The amplification efficiency of each pair of primers was tested by constructing the corresponding plasmid; only primers with a similar amplification efficiency were used in this experiment (Supplementary Table S4). The q-PCR was carried out in a 10 μ l reaction mixture, which consisted of 5 μ l of 2 × SYBR Green I Mix, 0.4 μ l of forward and reverse primers (10 μ M of each primer ), and 4.2 μ l diluted cDNA. Cycling conditions were as follows: Denaturation at 95 °C for 30 s, followed by 40 cycles of denaturation for 5 s at 95 °C, annealing for 30 s at 60 °C and extension for 30 s at 72 °C. Expression levels of 10 miRNAs randomly chosen was examined with the Taqman miRNA assay kit (Invitrogen) according to manufacturer's instruction (Supplementary Table S2). The relative quantification values for each miRNA and mRNA were calculated by the 2 −ΔΔct method using U6 and GAPDH as internal references, respectively. Based on the fold changes of miRNA assessed by miRNA microarray or q-PCR, we obtained normalized miRNA expression values by calculating the logarithmic value of fold changes. We performed correlation analysis for the two sets of data.
miRNA-mRNA regulatory network construction. TargetScan (http://www.targetscan.org/), miRBase (http://www.mirbase.org/), and miRanda (http://www.microrna.org/microrna/home.do) were used to predict the target genes of differentially expressed miRNAs. The predicted target genes were compared with the mRNA microarray results; only genes with an expression pattern that negatively correlated with its regulatory miRNAs were included. Cytoscape software (Version 3.0; http://www.cytoscape.org/) was used to construct the regulatory network 57 . Moreover, miRNA regulatory network analyses was performed using IPA to describe functional relationships among miRNAs and genes based on the known associations in the databases (IPA ® , QIAGEN Redwood City, www.qiagen.com/ingenuity), which mapped those differentially expressed genes to biological pathways 58 .

Gene ontology and pathway enrichment analyses of differentially expressed genes.
To identify potential biological processes regulated by the differentially expressed genes, we performed GO and pathway enrichment analyses based on the GO (http://geneontology.org/page/go-enrichment-analyses) and pathways (http://www.genome.jp/kegg/pathway.html) database. We calculated the P value of each GO term using right-sided hypergeometric tests, and Benjamini-Hochberg adjustment was used for multiple test correction 59,60 .
We consider those terms with a P value < 0.05 as significantly enriched terms.

Animals and treatment.
Male specific pathogen-free (SPF) BALB/c mice aged six weeks were purchased from Laboratory Animal Centre of the Academy of Military Medical Science, China. Mice were randomly divided into 4 groups: one control and three experimental groups with 20 mice in each group. Mice in the experimental group were injected intraperitoneally with MC-LR dissolved in saline at 7.5 μ g/kg, 15 μ g/kg and 30 μ g/kg body weight for 7 consecutive days; the control mice were treated with identical volumes of normal saline. One day after the last treatment, mice were sacrificed by CO 2 asphyxiation and testes were obtained for analyses of histopathology, q-PCR, and western blotting. In separate experiments, we investigated the toxic effects of chronic low-dose exposure to MC-LR on the male reproductive system. Mice were divided into 5 groups with 20 mice in each. For exposure, mice were given drinking water containing 1 μ g/L, 10 μ g/L, 20 μ g/L, or 30 μ g/L MC-LR for 90 consecutive days. Control mice were provided with only the blank water. Mice were sacrificed later and testes were obtained for histopathological analyses.
TNF-α-induced germ cell apoptosis. In order to confirm the effect of TNF-α on germ cell apoptosis, mouse GC-1 cells that share many of the characteristics of spermatogonia were treated with 20 ng/mL TNF-α in vitro for 3 h, 6 h, or 12 h. Cell apoptosis was measured by flow cytometry based on Annexin V -FITC and PI staining profiles. Protein was extracted from both control and treated cells for further analyses.
Western blotting, Co-IP and ELISA analyses. One hundred milligrams of tissue were homogenized for extracting proteins in 1 mL ice-cold RIPA buffer (50 mM Tris-HCL PH 7.4, 150 mM NaCl, 1% NP, 0.1% SDS, 1 × phosphatase/protease inhibitors). For some cells, 200 μ l protein extraction buffer was added into the dish to extract the total protein from cells. The protein concentration was determined using the BCA protein assay kit (Beyotime, Nantong, China). Western blotting was performed as described previously 13 . About 20 μ g of protein from each sample was separated on 10% SDS-PAGE and electrophoretically transferred to a polyvinylidene fluoride (PVDF) membrane. The membrane was then blocked in TBS buffer containing 5% bovine serum albumin (20 mM Tris-HCl (pH 7.6) and 150 mM NaCl) for 1.5 h at 37 °C. Next, these blots were incubated overnight at 4 °C with rabbit anti-c-Fos, rabbit anti-c-Jun, rabbit anti-SGTB, rabbit anti-β -catenin, rabbit anti-MAK11, rabbit anti-Occludin, rabbit anti-ZO-1, rabbit anti-p-p38 MAPK, rabbit anti-active caspase-3, rabbit anti-p-ATF2, goat anti-ATF2, mouse anti-caspase-8, and mouse anti-TNFR1. The blots that were incubated with rabbit IgG, mouse IgG, and goat IgG, served as negative controls. The blots were then incubated with species-matched horseradish peroxidase-conjugated secondary antibodies (Boster, Wuhan, China). The chromogenic signal intensity was detected using an Odyssey Scanning System (LI-COR, Lincoln, NE) and quantified using image J software (NIH, Bethesda, MD). Co-IP was performed by using 800 μ g of protein from lysates of GC-1. Anti-TNFR1 antibody was used as the precipitating antibody to isolate TNFR1 from lysates, followed by western blotting to identify whether TNF-α can bind to TNFR1 on the membrane using anti-TNF-α antibody. Blots were re-probed with anti-TNFR1 antibody to confirm equal protein loading. Co-IP with rabbit IgG served as a negative control. TNF-α and MCP-1 levels in cell supernatant and testis homogenate were measured by ELISA kits (eBioscience, San Diego, CA). All protocols were performed according to the manufacturer's instructions. Detection limits of both kits were < 10 pg/mL, intra-and inter-assay variation was < 5% and < 10%, respectively. Immunohistochemistry, Masson's trichrome stain and TUNEL assay. Testes were fixed in 4% paraformaldehyde solution for 24 h, embedded in paraffin and cut into 5-μ m-thick sections. The sections were incubated with PBS containing 3% H 2 O 2 for 10 min to inhibit endogenous peroxidase activity. After being blocked with 3% BSA in PBS for 1 h at 37 °C, the sections were subsequently incubated with rabbit anti-TNF-α and rabbit anti-active caspase-3 antibodies overnight at 4 °C and then with HRP-conjugated secondary antibodies (Boster) at 37 °C for 1 h. HRP activity was examined using the diaminobenzidine method according to the manufacturer's instructions (Zhongshan Biotechnology, Beijing, China). After counterstaining with haematoxylin, the sections were mounted with neutral balsam for observation under a microscope DXM12000F (Nikon, Tokyo, Japan). Masson's trichrome stain was used to determine the collagen deposits in the testes according to the instructions by the manufacturer (KeyGen, Nanjing, China). Apoptotic cells in testicular tissues were determined by TUNEL assay as described by the manufacturer (Roche Applied Science, Mannheim, Germany), and the nuclei were stained with DAPI (Sigma-Aldrich). The images were captured using a laser scanning confocal fluorescence microscope (FV10i; Olympus, Tokyo, Japan). At least 20 microscopic fields (× 600) randomly selected were evaluated for each testis. MC-LR-induced apoptotic index was calculated as the percentage of TUNEL positive cells in all cells in an image.
Immunofluorescence staining. Immunofluorescence analyses of testicular tissues were performed as previously described 12 . The following primary antibodies were employed: Rabbit anti-β -catenin, rabbit anti-occludin, and mouse anti-CD68. Alexa Fluor 488-conjugated goat anti-rabbit antibody or Alexa Fluor 594-conjugated goat anti-mouse antibody (Invitrogen) was used as a secondary antibody.
Biotin-tracing assay. Freshly dissected testes were injected with 10 mg/mL EZ-Link Sulfo-NHS-LC-LC-Biotin (Thermo, Waltham, MA) dissolved in PBS. After being incubated at 37 °C for 30 min, the testes were subsequently fixed in 4% paraformaldehyde for 4 h, followed by being immersed in a 30% sucrose solution.
The specimens were embedded in optimal cutting temperature (OCT, Sakura Finetek USA, Torrance, CA) and then were cut at a thickness of 8 μ m followed by fixation in acetone at 4 °C for 15 min. The sections were blocked with 3% BSA to avoid non-specific staining prior to incubation with Streptavidin-Alexa Fluor 594 (Invitrogen). After nuclear staining with DAPI, sections were observed under the laser scanning confocal fluorescence microscope FV10i.
Flow cytometric analyses. Flow cytometric analyses were used to determine whether immune cells were enriched in testicular tissues. The testes were de-capsulated in PBS and then gently pipetted to dissociate the seminiferous tubules, resulting in the release of interstitial cells. Cell suspensions were passed through 100-μ m copper meshes to remove the seminiferous tubules, after which the interstitial cells were collected by centrifugation by 300 × g for 5 min. Cells were incubated with PE-conjugated F4/80 antibodies or FITC-conjugated CD45R/B220 (BD Biosciences). The cells were analyzed using a FACScallbur flow cytometer (BD Biosciences). Cell apoptosis was analyzed by an Annexin V-FITC and PI staining kit (Vazyme, Nanjing, China) according to manufacturer's instructions.
Statistical analyses. SPSS 18.0 (SPSS, Chicago, IL) was used for statistical analysis. Experimental results were expressed as mean ± standard deviation. The Levene's test were employed to check the normality and homogeneity of variances in the data. The Student's t test was used for paired comparisons. For the comparison of three or more groups, one-way ANOVA was used for the comparison, which was followed by Duncan's post hoc test. Values of P < 0.05 were considered statistically significant.