Circular RNA COL1A2 Mediates High Glucose-Induced Oxidative Stress and Pyroptosis by Regulating MiR-424-5p/SGK1 in Diabetic Nephropathy

Diabetic nephropathy (DN) represents a major diabetes-related complication, which could undermine renal function. CircCOL1A2 has been previously reported to show abnormal expression during DN. However, its functional role in the progression of DN, as well as the potential molecular mechanisms, remains unclear. The present work examined the expression of circCOL1A2 in the plasma of DN patients, and employed high glucose (HG)-challenged HK-2 cells as the in vitro cell model of hyperglycemia (HG)-induced DN. CircCOL1A2 was silenced using siRNA in HK-2 cells to clarify the functional engagement of circCOL1A2 in HG-induced DN. We examined the roles of circCOL1A2 in regulating oxidative stress by measuring reactive oxygen species (ROS), lipid peroxidation, and superoxide dismutase (SOD) levels. Besides, the effects of circCOL1A2 silencing on pyroptosis were investigated by RT-qPCR, western blot (WB), and ELISA assays. StarBase (version 2.0) was used to identify the downstream effector of circCOL1A2, and their interactions were further verified through dual-luciferase reporter analysis, RNA pull-down assays, and RNA immunoprecipitation (RIP) assay. CircCOL1A2 was highly expressed in DN patients and HG-induced HK-2 cells. Knocking down circCOL1A2 alleviated oxidative stress and pyroptosis upon HG treatment. In addition, we demonstrated that circCOL1A2 knockdown could promote miR-424-5p expression while inhibiting Serum/Glucocorticoid Regulated Kinase 1 (SGK1) level. Furthermore, miR-424-5p inhibitor or SGK1 overexpression impaired the effects of circCOL1A2 knockdown on HG-induced oxidative stress and pyroptosis. Hence, our results demonstrated that the circCOL1A2 mediates HG-exposed pyroptosis and oxidative stress through modulating miR-424-5p/SGK1 axis in diabetic nephropathy, indicating that silencing circCOL1A2 is a potential intervention strategy for DN management.


Introduction
Diabetes mellitus (DM) represents one cluster of metabolic diseases with the feature of persistently high blood glucose (BG) contents [1].Type 2 diabetes mellitus (T2DM) is associated with insulin-resistance development and it accounts for 90-95% of DM burdens globally [2].T2DM-related complications comprise renal failure, heart attack, vision loss, wound repair impairment, and nerve injury [3,4].Diabetic nephropathy (DN) represents a major factor for end-stage renal disease (ESRD) worldwide [5,6].Typical pathological features in the kidney tissues of DN patients include abnormalities in vascular, tubulointerstitial, and glomerular compartments, which are caused by cell apoptosis and hypertrophy, extracellular matrix (ECM) deposition, and basement membrane thickening.[7] The uncontrolled high glucose (HG) is the major factor leading to DM-related DN, retinopathy, and cardiomyopathy.Previous studies suggest that oxidative stress can be induced by the excessive production of reactive oxygen species (ROS) in HG-challenged kidney cells [8,9].The onset of inflammatory responses upon HG stress could also induce the over-expression of adhesion molecules and chemokines in the proximal tubules to trigger pyroptosis [10][11][12].Therefore, HG-induced oxidative stress and inflammatory activation contribute to the occurrence and progression of DN, culminating in the gradual loss of renal function.However, the specific mechanism by which HG triggers oxidative stress and pyroptosis in kidney tissues remains unclear.
Circular RNAs (circRNAs), the non-coding RNAs (ncRNAs) with close-loop structure, are related to pathogenic mechanisms of different disorders [13,14].Recently, accumulating evidence pinpoints the critical roles of circRNAs in DN progression [15,16].As reporter by Zhang et al.'s study, circ0081108 (circCOL1A2) expression was elevated by 6.85-time in diabetic retinas relative to healthy controls [17].In addition, circCOL1A2 was also found to become dysregulated in diabetic retinopathy [18].However, there is currently no studies reporting the roles of circCOL1A2 in HG-induced damages in kidney cells and in the progression of DN occurrence.
MicroRNAs (miRNAs) are important downstream factors mediating the role of circR-NAs in pathological conditions.In the studies of DN, circRNA-miRNA interactions have been reported to mediate HG-induced kidney cell damages [19].For example, circ0000309 is implicated in DN-related ferroptosis regulation by targeting miR-188-3p/GPX4 axis [20].In addition, circ010383 was reported to act as a molecular sponge for miR-135a to regulate renal fibrosis in DN [21].Therefore, understanding the role of circRNA-miRNA interaction in HG-induced renal damages is important for develop intervention strategy.
Serum/Glucocorticoid Regulated Kinase 1 (SGK1) is a serine/threonine protein kinase that plays an important role in cellular stress response [22].This kinase activates certain potassium, sodium, and chloride channels in the regulation of cell survival and renal sodium excretion [23].High levels of expression of SGK1 is suggested to contribute to pathological conditions such as hypertension and diabetic nephropathy [24,25].There is also evidence that SGK1 is engaged in glucose metabolism in the intestine and kidney during the progression of hyperglycemia-induced secondary organ damage [26].Nerveless, the molecular mechanisms underlying SGK1 dysregulation under hyperglycemia remain to be elucidated.
In the present work, we reported the elevated expression of circCOL1A2 in the plasma samples of DN patients.MiR-424-5p was identified as a miRNA target of circCOL1A2 in regulating HG-induced oxidative stress and pyroptosis in HG-treated renal tubular cells (RTCs).Interestingly, miR-424-5p has been reported to become dysregulated in diabetes-related disorders [27,28].Besides, miR-424-5p can regulate the proliferation and extracellular matrix deposition in high glucose-challenged human glomerular mesangial cells (HGMC), suggesting its implication in modulating kidney cell response to hyperglycemia stress [29].Furthermore, we found that the 3′UTR (untranslated region) of SGK1 mRNA harbors binding sites for miR-424-5p, and miR-424-5p levels could regulate SGK1 expression levels in HG-treated RTCs.Overall, our study highlighted the critical role of circCOL1A2/miR-424-5p/SGK1 axis in modulating HG-induced diabetic nephropathy, suggesting that targeting circCOL1A2 may be a potential intervention strategy for DN management.

Patients and Tissue Samples
Our study protocols gained approval from the Research Ethics Committee of the First Affiliated Hospital of Bengbu Medical College (2021-086).Every participant provided informed consent.A total of 22 DN patients with T2DM and 25 healthy controls from the First Affiliated Hospital of Bengbu Medical College were recruited for this study.Plasma was collected for transcription-quantitative PCR (RT-qPCR) analysis.The diagnosis of diabetes of all patients was confirmed by experienced pathologist.The diagnoses were based on the criteria of the American Diabetes Association (ADA, 2016) [30].The inclusion criteria of DN patients were as follows: patients age > 18 years; patients who fulfilled the diagnostic criteria of the ADA of 2016; patients with a history of diabetes for more than a year; urinary protein/ creatinine ratio > 30 mg/g; persistent urinary protein or renal function decline (GFR < 60 mL/ min/1.73m 2 ); and biopsy confirmation of diabetes-related renal damages.The inclusion criteria of healthy controls were as follows: no medical history of diabetes; normal fasting blood glucose, renal function, and renal ultrasonography.Exclusion criteria for both groups such as subjects with infectious diseases, chronic systemic inflammatory diseases, blood diseases, autoimmune diseases, malignant tumors, or cardiovascular diseases were excluded.

Cell Lines and Culture
HK-2 cells lines were obtained in the Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China) and cultivated within DMEM/F12 (HyClone, USA) that contained 10% (vol/vol) fetal bovine serum (FBS, Merck Millipore, Germany) after heat inactivation under the condition of 5% CO 2 and 37 °C.To mimic hyperglycemia-induced damages in DN, HK-2 cells were cultured for a 2-week period within the medium that contained 30 mM glucose, meanwhile 24.5 mM mannitol combined with 5.5 mM glucose was adopted to be a reference control [31,32].Mannitol 24.5 mM was used together with 5.5 mM glucose to maintain the similar osmotic concentration as 30 mM glucose treatment.

StarBase
StarBase (version 2.0) was employed for predicting the binding sites between miR-424-5p and circCOL1A2, and the 3′UTR of SGK1 mRNA.
After washing with TBS Tween 20, membranes were subject to further probing using HRP-linked secondary antibody (1:3000; Cell signaling technologies, MA, USA).An ECL chemiluminescence kit (Thermo Fisher Scientific, MA, USA) was employed to detect protein bands, and GAPDH was used for internal reference for protein quantification using ImageJ software (Bethesda, MD, USA).

Enzyme-Linked Immunosorbent Assay (ELISA)
ELISA kits (RD, Minneapolis, MN, USA) were utilized to measure IL-1β and IL-18 concentrations in the cell culture supernatant following specific instructions.Signal development and the optical density of samples and standards were measured at 450 nm using a microplate reader (BioRad, CA, USA).The concentration of each cytokine was measured based on the linear regression of the standards.

RNA Pull-Down Assay
This assay was conducted using Pierce™ Magnetic RNA-Protein Pull-Down Kit (#20164, Thermo Fisher Scientific, CA, USA).Biotin-labeled circCOL1A2 (Bio-circCOL1A2 probe) and the control (Bio-NC probe) were synthesized by Sangon Biotech (Shanghai, China).In brief, cell lysates were collected from 1 × 10 6 HK-2 cells using RIPA lysis buffer, with 10% lysate saving as the input.Thereafter, 200 nM Bio-NC or Bio-circCOL1A2 probe was added to the cell lysate for 1-h incubation, followed by addition of 100 μL streptavidinmagnetic beads.The mixture was incubated for another 2 h, and the associated RNAs on the magnetic beads were purified using TRizol reagent.The relative enrichment of each target was quantified by qRT-PCR.

Statistical Analysis
Results were displayed as mean ± SD of 3 separate assays.GraphPad Prism 5.0 software was adopted for statistical analyses.We applied unpaired Student's t-test and oneway ANOVA for comparing 2 or multiple groups, respectively.Pearson's correlation analysis was adopted for determining correlation between the expression levels of two molecules.P < 0.05 stood for statistical significance.

Increased CircCOL1A2 Level in the Plasma of DN Patients and HG-Induced HK-2 Cells
Increased circCOL1A2 expression has been reported in diabetic retinopathy [17,18].We wondered whether circCOL1A2 also became dysregulated in DN.We carried out RT-qPCR analysis for detecting circCOL1A2 levels in plasma of healthy people (NC = 25) and diabetic nephropathy patients (DN = 22).The circCOL1A2 expression in DN patients showed an average of 2.5 time increase relative to controls (Fig. 1A).Next, we established an in vitro cell model based on HG-induced HK-2 cells to mimic hyperglycemia condition.HK-2 cells were cultured for a 2-week period within medium containing 30 mM glucose, with the culturing in 5.5 mM glucose plus 24.5 mM mannitol as the control.In HG-challenged HK-2 cells, circCOL1A2 level increased by nearly 3-fold relative to control group (Fig. 1B).These data suggest that elevated circCOL1A2 expression may be implicated in the progression of DN.

Knocking Down CircCOL1A2 Mitigates Oxidative Stress in HG-Treated HK-2 cells
To explore the role of circCOL1A2 under hyperglycemia condition, three different circ-COL1A2 siRNAs were used to knock down circCOL1A2 in HK-2 cells.RT-qPCR analysis showed that si-circCOL1A2#1 showed the strongest silencing effect, with a 70% reduction of circCOL1A2 when compared to the control siRNA (Fig. 2A).We therefore used this siRNA as si-circCOL1A2 in subsequent studies.Next, we measured ROS level as the indicator of oxidative stress level in HK-2 cells.The results showed that HG treatment increased ROS production by more than 2-fold in HK-2 cells, while silencing circCOL1A2 significantly reduced ROS production (Fig. 2B).Likewise, the analysis of the lipid peroxidation by-product MDA revealed a 3-fold increase of MDA production upon HG challenge, and silencing circCOL1A2 suppressed HG-induced MDA level by more than 60% (Fig. 2C).In contrast, the ROS detoxification enzyme SOD activity was inhibited by 50% upon HG treatment, in si-circCOL1A2 partially restored SOD activity upon HG treatment (Fig. 2D).Taken together, silencing circCOL1A2 could alleviate HG-induced oxidative stress in HK-2 cells.
Next, we explored the expression level of SGK1 in the plasma samples of healthy people (NC = 25) and diabetic nephropathy patients (DN = 22).SGK1 mRNA level was increased by nearly 3-fold in DN patients (Fig. 5E).According to correlation analysis, miR-424-5p level was adversely associated with SGK1 mRNA level in plasma samples of DN patients (Fig. 5F).In contrast, there was a significant positive correlation between circCOL1A2 expression and SGK1 mRNA level (Fig. 5G).We further showed that SGK1 protein level was increased by nearly 4-fold within HG-treated HK-2 cells (Fig. 5H).Therefore, circ-COL1A2/miR-424-5p axis regulates SGK1 expression in HK-2 cells.

CircCOL1A2 Regulates HG-Induced Damages in HK-2 Cells Through MiR-424-5p/ SGK1 Axis
To validate whether miR-424-5p/SGK1 axis mediates the effect of circCOL1A2 on HGinduced cell injury in HK-2 cells, we established the following experimental groups: control, HG treatment, HG + circCOL1A2 silencing, HG + circCOL1A2 silencing plus miR-424-5p inhibitor, and HG + circCOL1A2 silencing plus SGK1 expression vector.WB analysis showed that HG treatment caused a dramatic over-expression of SGK1, and silencing circCOL1A2 significantly suppressed the over-expression of SGK1.Meanwhile, the co-administration of miR-424-5p inhibitor or SGK1 expression vector impaired the effect of circCOL1A2 silencing (Fig. 6A).
ROS measurement showed that the protective effect of circCOL1A2 knockdown on HG-induced oxidative stress was impaired by SGK1 expression or miR-424-5p inhibitor (Fig. 6B).Similar results were observed with lipid peroxidation by-product MDA measurement (Fig. 6C).Consistently, silencing circCOL1A2 could increase SOD activity upon HG challenge, while SGK1 expression vector or miR-424-5p inhibitor suppressed this effect (Fig. 6D).These results suggest that circCOL1A2 impinges on the redox homeostasis of HG-induced HK-2 cells via mir-424-5p/SGK1 pathway.
Furthermore, we performed WB and qRT-PCR assays to detect pyroptosis-associated factors in above experiment groups.As shown in Fig. 6 E and F, si-circCOL1A2 transfection suppressed the upregulation of pyroptosis markers in HG-treated HK-2 cells, and miR-424-5p inhibitor or SGK1 over-repression weakened this effect.Similarly, we found that the inhibitory effect of circCOL1A2 silencing on HG-triggered IL-18 and IL-1β production was dramatically impaired following SGK1 expression and miR-424-5p inhibition (Fig. 6G).Collectively, circCOL1A2 modulates HG-induced HK-2 cell injury via miR-424-5p/SGK1 pathway.

Discussion
DN has become a prevalent disorder in diabetic patients and has a complicated pathogenic mechanism during its progression into ESRD [33].Accompanied by high morbidity and mortality rates, DN has imposed tremendous physico-mental pressure and financial burdens on the patients and the society.The uncontrolled hyperglycemia has been recognized as a major risk factor for diabetic-induced nephropathy.As previously reported, chronic hyperglycemia could exacerbate ROS production and oxidative stress in kidney [8,9] and further induces pyroptosis [10,11].Nonetheless, the precise mechanisms by which HG triggers oxidative and inflammatory damages in renal tissues is still largely unclear.
The present work examined the mechanism underlying HG-triggered DN using HGexposed HK-2 cells and DN patient plasma samples.Recent works have highlighted the crucial roles of circRNAs in regulating the progression of DN [15,16].Through literature review, a previous study reported that circCOL1A2 expression level increased by 6.85-fold in patients with diabetic retinopathy relative to non-diabetic counterparts [17].In addition, circCOL1A2 could promote angiogenesis via regulating miR-29b/VEGF axis in diabetic retinopathy [18].Our data provided novel evidence regarding the role of circCOL1A2 in HG-induced damages in renal tubular cells.CircCOL1A2 expression is markedly elevated in HG-treated HK-2 cells and the plasma samples of DN patients.Subsequently, this work showed that silencing circCOL1A2 level via siRNA in HK-2 cells could mitigate HGinduced oxidative stress and pyroptosis.The knockdown of circCOL1A2 reduced ROS production within HG-treated HK-2 cells and pyroptosis-related protein markers such as Caspase-1, GSDMD-N, NLRP3, and IL-1β were heavily upregulated upon HG challenge.ELISA measurement of IL-1β and IL-18 in the supernatant from HK-2 cells showed a protective effect of circCOL1A2 silencing on inflammatory response.Therefore, targeting circCOL1A2 may be employed as an intervention strategy to mitigate HG-induced damages in DN.
StarBase (version 2.0) was used to identify miR-454-3P as a target of circCOL1A2.Their interaction was verified by RIP, dual-luciferase reporter, and RNA pull-down assays.Previous studies have revealed versatile roles of miR-424-5p on dictating cellular sensitivity towards cell ferroptosis, apoptosis, and oxidative stress [34][35][36].Interestingly, miR-424-5p level was found to be decreased during DN [29].In agreement with this, we also showed the reduced miR-424-5p in DN patients and HG-induced HK-2 cells.Furthermore, knocking down circCOL1A2 could promote miR-424-5p level, and there was a negative correlation between circCOL1A2 and miR-424-5p levels within plasma from patients with DN.Moreover, the application of miR-424-5p inhibitor was able to undermine the protective effect of circCOL1A2 silencing on HG-induced HK2 cell damages.Thus, miR-424-5p at least partially mediates the protective effect of circCOL1A2 on HG-induced HK2 cell damages.
However, the protective effect of silencing circCOL1A2 in mitigating hyperglycemiainduced renal damages needs further validation in the animal model of DN.Besides, the mechanism by which hyperglycemia induces circCOL1A2 upregulation needs to be elucidated.Equally important, how SGK1 regulates ROS production and pyroptosis under hyperglycemia challenge warrants further investigations.

Conclusion
In conclusion, we reported the over-expression of circCOL1A2 in DN patients and HG-challenged HK-2 cells.Knocking down circCOL1A2 suppressed HG-induced pyroptosis and oxidative stress in HK-2 cells.Mechanistically, circCOL1A2 knockdown restrains HG-induced oxidative stress and pyroptosis by promoting miR-424-5p expression, thereby inhibiting SGK1 expression.Our study provides novel insights into the mechanisms underlying SGK1 upregulation in DN progression, and highlights circCOL1A2 as a potential target for therapeutic intervention in DN management.