CircSMAD4 alleviates high glucose-induced inflammation, extracellular matrix deposition and apoptosis in mouse glomerulus mesangial cells by relieving miR-377-3p-mediated BMP7 inhibition

Background Diabetic nephropathy (DN) is a common complication of diabetes mellitus. Accumulating studies suggest that the deregulation of circular RNA (circRNA) is involved in DN pathogenesis. This study aimed to investigate the role of circSMAD4 in DN models. Methods Mice were treated with streptozotocin to establish DN models in vivo. Mouse glomerulus mesangial cells (SV40-MES13) were treated with high glucose to establish DN models in vitro. The expression of circSMAD4, miR-377-3p and bone morphogenetic protein 7 (BMP7) mRNA was measured by quantitative real-time PCR (qPCR). The releases of inflammatory factors were examined by ELISA. The protein levels of fibrosis-related markers, apoptosis-related markers and BMP7 were checked by western blot. Cell apoptosis was monitored by flow cytometry assay. The predicted relationship between miR-377-3p and circSMAD4 or BMP7 was validated by dual-luciferase reporter assay or pull-down assay. Results CircSMAD4 was poorly expressed in DN mice and HG-treated SV40-MES13 cells. HG induced SV40-MES13 cell inflammation, extracellular matrix (ECM) deposition and apoptosis. CircSMAD4 overexpression alleviated, while circSMAD4 knockdown aggravated HG-induced SV40-MES13 cell injuries. MiR-377-3p was targeted by circSMAD4, and miR-377-3p enrichment partly reversed the effects of circSMAD4 overexpression. BMP7 was a target of miR-377-3p, and circSMAD4 regulated BMP7 expression by targeting miR-377-3p. MiR-377-3p overexpression aggravated HG-induced injuries by suppressing BMP7. Conclusion CircSMAD4 alleviates HG-induced SV40-MES13 cell inflammation, ECM deposition and apoptosis by relieving miR-377-3p-mediated inhibition on BMP7 in DN progression.


Introduction
Diabetic nephropathy (DN), a serious complication of diabetes, is the most common cause of chronic kidney disease (CKD) and end-stage renal disease (ESRD) in the world [1]. In the past 20 years, the morbidity and mortality of DN have risen rapidly in the global population [2].
Clinically, DN manifests as increased urinary albumin excretion (> 300 mg/day), decreased glomerular filtration rate (GFR), diabetic glomerulopathy, and increased arterial blood pressure [3,4]. Pathologically, there are many changes in the kidney, including extracellular matrix (ECM) deposition (mainly in the mesangium), thickening of the glomerular basement membrane, interstitial fibrosis, and glomerular sclerosis [5]. DN has significantly negative impacts on society in the fields of public health and socioeconomics. Therefore, more research needs to be carried out to clarify the pathogenesis of DN to prevent the disease and improve treatment.
Circular RNA (circRNA) has been a hot research topic in the field of medicine in recent years. As a member of non-coding RNAs, circRNA is famous for its circular closed structure, which makes them more stable compared to linear molecules [6]. CircRNA is widely expressed in tissues, cells and multiple body fluids, such as blood, urine and saliva [7]. CircRNAs have been recognized as promising biomarkers for disease diagnosis and management due to wide expression and easy-todetect [6,8]. Mounting studies identify that numerous circRNAs are implicated in the progression of DN. For example, Hu et al. established DN cell model by treating mouse mesangial cells (SV40-MES13) with high glucose (HG) and then found that circRNA_15698 promoted ECM accumulation in HG-induced SV40-MES13 cells [9]. Peng et al. discovered that circRNA_010383 suppressed HG-induced ECM accumulation and blocked proteinuria and renal fibrosis in DN mice [10]. A previous circRNA expression profile identified several differentially expressed circRNAs in kidney tissues from DN mice, and we noticed that mmu_circ_0000894 was poorly expressed in DN mice [11]. The data from circbase database display that mmu_circ_0000894 is derived from mouse SMAD4 gene by "backsplicing", also named circS-MAD4. However, the role of circSMAD4 in DN progression is unclear and needs to be further investigated.
As for molecular mechanism, numerous circRNAs were shown to serve as a competing endogenous RNA (ceRNA) to regulate gene expression by competing for microRNA (miRNA) binding site [12]. For instance, cir-cRNA_15698 acted as miR-185 sponge to elevate TGF-β1, thus enhancing ECM accumulation in HG-treated SV40-MES13 cells [9]. The exploitation of bioinformatics databases makes it easy to predict the potential targets of circRNA or miRNA. Therefore, we hypothesize that circSMAD4 governs the miRNA/mRNA axis in DN development and further verify this.
In this study, we established DN mouse model and DN cell model and examined the expression of circSMAD4 in these models. Besides, we investigated the function of circSMAD4 on inflammatory responses, ECM accumulation and apoptosis in SV40-MES13 cells. Moreover, the interactions among circSMAD4, miR-377-3p and bone morphogenetic protein 7 (BMP7) were determined to provide a functional mechanism of circSMAD4.

Materials and methods
Animal models C57BL/6 mice (female, 6-week-old) were purchased from Vital River Laboratory Animal (Beijing, China) and housed in pathogen-free conditions. These mice were divided into two groups (n = 6/group), control and diabetic mice (DM). Mice in the DM group were administered with high-fat and high-glucose diets. Then, mice were intraperitoneally injected with streptozotocin (45 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) to increase the burden on the kidney for consecutive 7 days. Streptozotocin was dissolved in 0.1 M citrate acid solution. Mice were considered to have DN until fasting blood-glucose reached 16.5 mmol/L, urinary albumin level over 30 mg/24 h. Mice in the control group were administered with standard laboratory chow and injected with citrate acid solution (45 mg/kg) for consecutive 7 days. Mice were sacrificed after 0, 1, 2 and 4 months, and kidney tissues were excised for further analysis. Animal studies were performed in agreement with the guidelines of the Animal Care and Use Committee of Affiliated Hospital of Inner Mongolia University for Nationalities.

Flow cytometry assay
SV40-MES13 cells were collected at 48 h post-transfection to determine the number of apoptotic cells using the FITC Annexin V Apoptosis Detection Kit (BD Biosciences, San Jose, CA, USA). According to the protocol, cells were resuspended in Annexin V-FITC binding buffer and next stained with Annexin V-FITC and propidium iodide (PI). The apoptotic cells were sorted and distinguished using a flow cytometer (BD Biosciences).

Quantitative real-time PCR (qPCR)
Trizol reagent (Cwbio) was applied for total RNA isolation. Subsequently, cDNA was assembled using the Pri-meScript 1st strand cDNA Synthesis Kit (Takara, Dalian, China) or microScript microRNA cDNA Synthesis Kit (Norgen Biotek, Thorold, Canada), followed by qPCR amplification using the SYBR reagent (Sigma-Aldrich). Using β-actin or U6 as the internal reference, the foldchange of relative expression was calculated using the 2 −ΔΔCt method. The primer sequences were shown as below: circSMAD4 (

RNA pull-down assay
SV40-MES13 cells were transfected with biotinylated miR-377-3p (Bio-miR-377-3p; 50 nM; Ribobio) or biotinylated miR-NC (Bio-miR-NC) and maintained for 24 h. The cells were then harvested and lysed in lysis buffer (Invitrogen). Cell lysates were incubated with streptavidin magnetic beads (Invitrogen) for 4 h. The beads were washed, and RNA compounds on beads were eluted and isolated for qPCR analysis.

Statistical analysis
Data were collected from three independent experiments for each assay and operated using GraphPad Prism 7 (GraphPad, La Jolla, CA, USA). The differences in different groups were determined using Student's t-test or using analysis of variance followed by the Tukey posttest. The data were displayed as the mean ± standard deviation (SD). P < 0.05 was considered to be statistically significant.

CircSMAD4 was downregulated in DN mouse models
The data showed that the blood glucose of mice from DN group was significantly increased compared with that in control group (Fig. 1A). Besides, 24 h urinary albumin was strikingly enhanced in DN mice compared to control  1B). Moreover, urinary 8-OH-dG was also markedly reduced in DN mice compared to control (Fig. 1C). Meanwhile, the expression of circSMAD4 was significantly declined in DN mice compared to control, and its expression was negatively correlated with the severity of DN (Fig. 1D). The data suggested that circSMAD4 was poorly expressed in DN mouse models.

CircSMAD4 was downregulated in HG-treated SV40-MES13 cells, and HG promoted inflammation, cell apoptosis and ECM deposition
The expression of circSMAD4 was shown to be notably decreased in HG-treated SV40-MES13 cells compared with that in mannitol-or NG-treated cells ( Fig. 2A). In function, HG largely promoted the secretion of proinflammatory factors, including IFN-γ, MCP-1, IL-6 and TNF-α ( Fig. 2B-E). Besides, fibrosis-related proteins, including fibronectin and collagen IV, were shown to be upregulated in SV40-MES13 cells with HG treatment compared to NG or mannitol (Fig. 2F). In addition, flow cytometry assay presented that the number of apoptotic cells was strikingly increased in the HG-treated group compared to other groups (Fig. 2G). The expression of Bcl-2 was declined, while the expression of Bax was reinforced in HG-treated SV40-MES13 cells (Fig. 2H). These results manifested that HG triggered SV40-MES13 cell inflammation, ECM deposition and cell apoptosis.

HG-induced SV40-MES13 cell inflammation, ECM deposition and cell apoptosis were alleviated by circSMAD4 overexpression
Gain-and loss-function assays were performed to determine the role of circSMAD4 in HG-treated SV40-MES13 cells. The efficiency of circSMAD4 overexpression and knockdown was first checked, and the data showed that circSMAD4 expression was markedly enhanced in SV40-MES13 cells transfected with circSMAD4 but notably reduced in cells transfected with sh-circSMAD4 (Fig. 3A). The expression of circS-MAD4 decreased in HG-treated SV40-MES13 cells was partly recovered by circSMAD4 transfection, while its expression was further suppressed by sh-circSMAD4  (Fig. 3B). In function, the releases of IFN-γ, MCP-1, IL-6 and TNF-α induced by HG were effectively alleviated by circSMAD4 overexpression but further stimulated by circSMAD4 knockdown (Fig. 3C-F). The protein levels of fibronectin and collagen IV promoted by HG were largely suppressed by circSMAD4 overexpression but further strengthened by circSMAD4 knockdown (Fig. 3G). HG-induced SV40-MES13 cell apoptosis was largely alleviated by circSMAD4 overexpression but further promoted by circSMAD4 knockdown (Fig. 3H). Additionally, the protein level of Bcl-2 decreased in HG-treated SV40-MES13 cells was restored in HG-treated SV40-MES13 cells transfected with circSMAD4 but further declined in HG-treated SV40-MES13 cells transfected with sh-circSMAD4, while the protein level of Bax was opposite to Bcl-2 expression (Fig. 3I). These results suggested that circSMAD4 overexpression alleviated HG-induced SV40-MES13 cell inflammation, ECM deposition and cell apoptosis.

Discussion
Our study mainly discovered that circSMAD4 expression was strikingly decreased in DN mice and HGtreated SV40-MES13 cells. HG-induced SV40-MES13 cell inflammation, ECM accumulation and apoptosis were largely alleviated by circSMAD4 overexpression but further aggravated by circSMAD4 knockdown. Mechanism analysis revealed that circSMAD4 positively regulated BMP7 expression by acting as miR-377-3p sponge. Accordingly, circSMAD4 governed the miR-377-3p/BMP7 axis to prevent the progression of DN. Mounting circRNAs are displayed to be aberrantly expressed in DN [13], and some of them are confirmed to be involved in DN pathogenesis, suggesting that cir-cRNAs are novel class of regulators in DN. HG-treated mesangial cells are widely used as DN models in vitro [14][15][16]. Previous studies manifested that HG evoked inflammation, ECM accumulation and oxidative stress in mesangial cells [14]. Besides, HG was shown to induce mesangial cell injury by stimulating cell apoptosis, thus impairing cell functions [17]. We thus treated SV40-MES13 cells with HG and found that HG largely induced SV40-MES13 cell inflammation, ECM accumulation and apoptosis. CircSMAD4 was previously shown to be downregulated in DN mice by circRNA sequencing analysis [11]. Consistently, our study verified that circSMAD4 expression was remarkably reduced in streptozotocinadministered mice and HG-treated SV40-MES13 cells. Functional analysis uncovered that circSMAD4 overexpression largely alleviated HG-induced SV40-MES13 cell inflammation, ECM accumulation and apoptosis, while circSMAD4 knockdown aggravated these negative effects. Our study was the first to partly determine the role of circSMAD4 in DN and defined that circSMAD4 played a protective role in DN development.
To explore the regulatory mechanism of circSMAD4, we identified the target miRNAs of circSMAD4 and confirmed that miR-377-3p was a target of circSMAD4. MiR-377-3p was previously shown to be upregulated in DN mice and HG-treated mesangial cells, and it largely promoted the production of fibronectin and thus promoted mesangial cell fibrosis [18,19]. MiR-377-3p was also proposed as a biomarker of DN [20]. Besides, miR-377-3p also gained considerable attention in inflammatory responses in human diseases [21,22]. Consistent with these findings, miR-377-3p enrichment partly reversed the effects of circSMAD4 and thereby recovered HGinduced inflammation, ECM accumulation and apoptosis in SV40-MES13 cells, suggesting that miR-377-3p drove the development of DN. Our study clarified that miR-377-3p was involved in circSMAD4 network in DN.
Additionally, we showed that BMP7 was a target gene of miR-377-3p. BMP7 played an indispensable in DN. Wang et al. demonstrated that BMP7 overexpression reversed diabetes-induced renal hypertrophy and recovered the function of glomerular filtration rate, urinary albumin excretion and glomerular histology [23]. Besides, forced BMP7 expression inhibited glomerular fibrosis and interstitial collagen accumulation in diabetic mice [24]. In addition, BMP7 overexpression attenuated oxidative stress and inflammatory responses in diabetic kidney disease [25]. In agreement with these findings, our results presented that the reintroduction of BMP7 partly abolished the effects of miR-377-3p enrichment and effectively suppressed HG-induced inflammation, ECM accumulation and apoptosis in SV40-MES13 cells. Our study further determined that miR-377-3p inhibited the expression of BMP7, however, circSMAD4 acted as miR-377-3p sponge to relieve the inhibition of miR-377-3p on BMP7, indicating that circSMAD4 promoted the expression of BMP7 by targeting miR-377-3p, thus preventing DN development.
Given that various cell types of glomerulus are all involved in DN pathogenesis, such as mesangial cells, podocytes and endothelial cells, the data regarding the functions of circSMAD4 in podocytes and endothelial cells are lacking in our present study. We only used mesangial cells as DN model in vitro, which was a main limitation. Future work should be prepared to address the role of circSMAD4 in podocytes and endothelial cells.

Conclusion
Collectively, circSMAD4 was downregulated in DN models in vivo and in vitro. CircSMAD4 alleviated HGinduced inflammation, ECM deposition and apoptosis in mouse glomerulus mesangial cells by increasing BMP7 expression via targeting miR-377-3p. We first proposed that circSMAD4 prevented the progression of DN partially by governing the miR-377-3p/BMP7 pathway, aiming to provide a new strategy for DN treatment.