Quercetin ameliorated insulin resistance via regulating METTL3-mediated N6-methyladenosine modication of PRKD2 mRNA in skeletal muscle and C2C12 myocyte cell line

Background: N6-Methyladenosine (m6A) is involved in many of pathological processes including insulin resistance (IR). Quercetin (Que), a bioactive compound with strong antioxidant activity, has potential therapeutic effects on IR-related metabolic diseases. The aim of this study is to investigate the roles of m6A and Que in hyperinsulinemia. Methods: Male C57Bl/6 mice received a high fat diet (HFD) for 8 weeks to establish an IR model. Biochemical analyzes, oral glucose tolerance test and homeostasis model assessment of IR (HOMA-IR) were conducted. C2C12 cells were transfected with methyltransferase-like 3 (METTL3) siRNA before treatment with palmitate (PA). 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) amino]-2-deoxyglucose (2-NBDG) was used to detect glucose uptake. Malondialdehyde (MDA) level, superoxide dismutase (SOD) activity, and glutathione (GSH) level were used to evaluate the levels of oxidative stress. SRAMP, a predicted site of m6A modication, was used to predict the relationship between METTL3 and erine-threonine kinase protein kinase D2 (PRKD2). Then, cells were treated with Que or/and transfected with pcDNA-METTL3 or PRKD2 siRNA. indirect calorimetry, Western blotting and qPCR were used to analyze protein and mRNA expression. Results: Que treatment reduced the body weight, blood glucose, plasma TG and serum insulin, ameliorated IR, and decreased oxidative stress in HFD-fed mice. Que at noncytotoxic doses promoted PA-induced glucose uptake and inhibited oxidative stress in C2C12 cells. Moreover, METTL3 and PRKD2 was downregulated in PA-induced C2C12 cells, and SRAMP revealed multiple m6A modication sites in the PRKD2 mRNA sequence. Downregulation of METTL3 enhanced PRKD2 expression by reducing m6A level and mRNA stability in PRKD2 mRNA transcript. C2C12 cells treatment with Que after inducing with PA inhibited m6A, METTL3 and phosphorylated insulin receptor substrate 1 (p-IRS1) levels, increased the protein expression of PRKD2, glucose transporter type 4 (GLUT4) and p-AKT and glucose uptake, and reduced oxidative stress. However, METTL3 overexpression or PRKD2 silence reversed the inhibitory effects of Que on the levels of MDA, GLUT4 and p-AKT, and the promotive effects on glucose uptake, SOD, GSH and p-IRS1 levels.

insulin [4]. Hence, ameliorating IR is an effective way to treat hyperinsulinemia and a series of metabolic syndrome.
Oxidative stress is associated with a variety of metabolic diseases, including diabetes, hypertension and excess weight [5]. Study reported that hyperacetylation of the mitochondrial antioxidant superoxide dismutase 2 (SOD2) could lead to SOD2 inactivation and mitochondrial oxidative stress [6]. Oxidative stress mediated by high glucose was one of the major complications of diabetes mellitus [7]. Oxidative stress is also closely related to IR, because it can disrupt insulin signal transduction, cause the disorder of fat factors and induce IR [5]. Moreover, oxidative stress also can reduce insulin sensitivity [8]. Studies reported that oxidative stress may damage insulin secretion and insulin function in diabetic patients [9].
As the most prevalent internal chemical modi cation for mRNAs in eukaryotic cells, N6-methyladenosine (m6A) is catalyzed by a methyltransferase complex composed of methyltransferase like 3 (METTL3), methyltransferase like 14 (METTL14), and wilms tumor 1-associated protein (WTAP) [10]. m6A is involved in regulating many important biological processes, including cancer, cardiovascular and metabolic diseases [11]. However, the role of m6A in IR is still unclear.
Flavonoids are the most widespread polyphenols in human diet and have many biological activities. As one number of avonoids, the mechanisms of quercetin (QUE) are pleiotropic [12]. QUE could increase the expression of endothelial nitric oxide synthase (eNOS), decrease the level of inducible nitric oxide synthase (iNOS), and thus alleviate insulin resistance in diabetic rats [13]. In addition, QUE also has been reported to have strong antioxidant and anti-in ammatory effects, and has multiple protective effects on oxidative stress-related diseases, including diabetes, cardiovascular diseases, atherosclerosis and cancer [14]. Satyakumar et al. reported that QUE effectively reversed the symptoms of fatty liver by decreasing triacyl glycerol accumulation, IR, in ammatory cytokine secretion and increasing cellular antioxidants in osteoarthritis induced hepatic steatosis of C2C12 cells [15], but the distinct molecular mechanisms of QUE in IR and oxidative stress is unclear. Hence, in this study, palmitate (PA)-induced IR C2C12 cell model and high-fat diet (HFD)-induced mouse model were established to investigate the effect and underlying mechanism of QUE on IR.

QUE ameliorated HFD-induced metabolic disorders and IR of mice
Compared with control group, the body weight, fasting plasma glucose, and plasma TG of mice fed with HFD for 10 weeks were signi cantly increased ( Fig. 1A-C, P < 0.05). Furthermore, the degree of fasting insulin and HOMA-IR in HFD fed mice was signally higher than that in normal diet fed mice ( Fig. 1D and E, P < 0.05). In accordance, HE staining showed that obviously lipid vacuolation in hepatic tissues from HFD mice when compared to the mice of control group (Fig. 1F). QUE treatment could lose weight to some extent, reduce blood glucose, plasma TG, serum insulin and HOMA-IR, and attenuate the HFD-induced hepatic steatosis ( Fig. 1A-E, P < 0.05). After glucose or insulin injection, the HFD group had signi cantly higher blood glucose and area under the curve than the control group, whereas QUE treatment signi cantly alleviated mouse glucose tolerance and insulin resistance ( Fig. 1G-J, P < 0.05). QUE ameliorated IR through inhibiting oxidative stress and METTL3-mediated N6-methyladenosine modi cation in skeletal muscle In relation to control group, mice fed HFD notably reduced SOD ( Fig. 2A, P < 0.05) and GSH (Fig. 2B, P < 0.05) levels, increased ROS content (Fig. 2C, P < 0.05), elevated m6A methylated RNA level (Fig. 2D, P < 0.05) and METTL3 mRNA and protein expression ( Fig. 2E and F, P < 0.05), inhibited GLUT4 and p-AKT protein expression, and increased p-IRS1 protein expression ( Fig. 2G and H, P < 0.05). However, QUE treatment could markedly reverse the promotive effect on ROS content, m6A methylated RNA level, and METTL3 mRNA and protein levels and p-IRS1 protein expression, and the inhibitory effects of SOD and GSH levels, and protein expression of GLUT4 and p-AKT (P < 0.05). QUE ameliorated glucose uptake and suppressed oxidative stress in PA-induced C2C12 cells As shown in Fig. 3A, when the concentration of QUE was 20 µM, the viability of C2C12 cells was signi cantly inhibited (P < 0.05), whereas the other concentrations had no effect. Therefore, we chose lower concentrations (5 µM and 10 µM) of QUE for subsequent experiments. Figure 3B showed that cells treatment with 250 µM PA could reduce 2-NBGD uptake, while 100 nM insulin prominently increased the 2-NBGD uptake in C2C12 cells in relation to cells without insulin stimulation (P < 0.05). However, cells treatment with 250 µM PA and 100 nM insulin could not increase 2-NBGD uptake, which revealed that PA could stimulate IR (P < 0.05). Compared with PA + insulin group, 2-NBGD uptake was increased after cells were treated with 5 µM or 10 µM QUE, and the increased effect of 10 µM QUE was more obvious (Fig. 2B, P < 0.05). Cells induced by PA could signi cantly increase ROS content and METTL3 protein expression, and inhibit SOD activity and GSH level ( Fig. 2C-G, P < 0.05), while QUE treatment could reverse the increased effect of PA on ROS level and METTL3 protein level, and the inhibitory effect of SOD and GSH levels (P < 0.05). With the increase of QUE concentration, its inhibitory or promoting effect was more signi cant.
Downregulation of METTL3 increased glucose uptake and suppressed oxidative stress in PA-induced C2C12 cells C2C12 cells were transfected with METTL3 siRNA to investigate its role in glucose uptake and oxidative stress. The protein level of METTL3 was noticeably decreased after transfection with METTL3 siRNA in comparison with PA + Scramble group ( Fig. 4A and B, P < 0.05). Silencing METTL3 reversed the increase effect of PA on ROS content (Fig. 4F, P < 0.05), and the inhibitory effect on glucose uptake (Fig. 4C, P < 0.05) and the levels of SOD and GSH ( Fig. 4D and E, P < 0.05).
Overexpression of METTL3 reversed the improvement effect of QUE on PA-induced glucose uptake and oxidative stress The protein expression of METTL3 was observably reduced after C2C12 cells were treated with 10 µM QUE compared with 250 µM PA treatment ( Fig. 5A and B, P < 0.05). However, overexpression of METTL3 could reverse the inhibitory effect of 10 µM QUE on the protein level of METTL3 (P < 0.05). Moreover, upregulation of METTL3 could also reverse the promotive effect of 10 µM QUE on glucose uptake Silence of PRKD2 reversed the improvement effect of QUE on PA-induced glucose uptake and oxidative stress through GLUT4/IRS-1/AKT pathway C2C12 cells were transfected with PRKD2 siRNA to analyze the role of PRKD2 in glucose uptake and oxidative stress induced by PA. The protein level of PRKD2 was dramatically decreased after transfection with PRKD2 siRNA by comparison with QUE treatment ( Fig. 7A and B, P < 0.05). Downregulation of PRKD2 counteracted the promotive effect of 10 µM QUE on glucose uptake ( Fig. 7C, P < 0.05) and the levels of SOD and GSH ( Fig. 7D and E, P < 0.05), and the inhibitory effect on SOD content ( Fig. 7F, P < 0.05). Compared with control group, Western blot analysis revealed that the protein levels of GLUT4 and p-AKT were signi cantly reduced after PA stimulation, while the protein level of p-IRS1 was markedly increased ( Fig. 7G and H, P < 0.05). The levels of GLUT4 and p-AKT in PA and 10 µM QUE group were increased, and p-IRS1 protein expression was reduced (P < 0.05). However, overexpression of METTL3 or silence of PRKD2 could reverse the promotive effect of QUE on the protein expression of GLUT4 and p-AKT, and the inhibitory effect of p-IRS1 protein level (P < 0.05).
Overexpression of METTL3 or silence of PRKD2 reversed the improving effects of QUE on HFD-induced metabolic disorders and IR By comparison with HFD + QUE group, overexpression of METTL3 or silence of PRKD2 could remarkedly increase mouse blood glucose (

Discussion
Recent studies showed that m6A methylation modi cation had a wide range of functional effects on homeostasis [16]. Our study revealed that HFD mice had signi cantly higher levels of blood glucose, TG and HOMA-IR, and the homeostasis of HFD mice were disrupted. Moreover, m6A levels were signi cantly increased in skeletal muscle of HFD mice. Abnormalities in any m6A level can lead to dysfunction and disease [17].
METTL3 is one of the core subunits of the methyltransferase complex catalyzing m6A formation, and aberrant expression of METTL3 is involved in many pathological processes, including IR[18]. Both m6A methylated RNA and METTL3 levels were consistently upregulated in liver tissues of Type 2 diabetes patients, and hepatocyte speci c knockout of METTL3 improved insulin sensitivity and decreased fatty acid synthesis in HFD-fed mice [19]. Li et al reported that both METTL3 expression and m6A levels were increased in the liver of HFD-induced metabolic disorder mice. Overexpression of METTL3 aggravated HFD induced hepatic metabolic disorders and IR [20]. In our study, METTL3 and m6A levels were also upregulated in HFD-fed mice and PA-induced C2C12 cells, and silence of METTL3 could noticeably ameliorate IR and oxidative stress.
Flavonoids have been reported to promote insulin secretion. Rutin, the glycoside compound of QUE, increases insulin secretion in rat β cells and maintain glucose sensing ability under high glucose conditions; it can also stimulate glucose transport into muscle by activating the synthesis and transport of transporter GLUT4 [21]. QUE could acutely stimulate insulin release in rat and suppress INS-1 β cell viability [22]. It also reports that QUE could inhibit intestinal glucose absorption, insulin secretion, and insulin sensitizing activity, as well as improve glucose utilization in peripheral tissues. Previous study suggested that QUE had poor bioavailability, while recent reports revealed that QUE can be detected in plasma after consumption of foods or supplements containing QUE components, indicating that QUE has a relatively long half-life in humans [12]. In addition, avonoids and avonols such as apigenin, luteolin, kaempferol, QUE, and setin can inhibit glucose transport, glycogen synthesis, protein synthesis, and cell survival by inhibiting IR phosphorylation through their activity at the C3 position [23].
QUE, a polyphenol with two aromatic rings, is the optimal con guration for free radical scavenging and transition metal ion binding, so QUE is considered to be a powerful antioxidant [24]. In nonalcoholic fatty liver, QUE ameliorated in ammation, oxidative stress and lipid metabolism in Db/Db mice [25]. QUE inhibited oxidative stress and in ammation progression by attenuating NF-κB and stabilizing HIF-1α, abrogating the possibility of hypobaric hypoxia induced pulmonary edema in rats [26]. Moreover, QUE ameliorated the body's antioxidant capacity by increasing the level of GSH and SOD, which utilized GSH as a hydrogen donor to convert free radicals generated in the body captured by SOD into H 2 O 2 [27]. Hence, when C2C12 cells were treated with QUE, C2C12 cell glucose uptake, SOD activity and the level of GSH increased, but ROS content decreased.
The serine-threonine kinase protein kinase D2 (PRKD2) de ciency causes to IR and metabolic disorders by promoting β-cell insulin secretion, which leads to hyperinsulinemia [28]. Our study revealed that PRKD2 level was downregulated in C2C12 cells treated with PA, while increased after transfected with pcDNA-METTL3. What is more, there are multiple m6A modi cation sites in the PRKD2 mRNA sequence through SRAMP. Hence, silencing PRKD2 could reverse the promotion of QUE on blood glucose absorption and the inhibition of oxidative stress. The reduction of GLUT4 expression or activity leads to reduced glucose uptake by skeletal muscle and adipocytes and is also an important molecular basis of IR [29]. Hyperbaric oxygen improves insulin sensitivity by promoting GLUT4 expression in muscle [30]. It is reported that any factor that reduces IRS-1 phosphorylation or induces serine IRS-1 phosphorylation at the 307 sites could damage insulin signal transduction and lead to insulin resistance [31]. Benzyl isothiocyanate ameliorated IR via inhibiting phosphorylation of IRS-1, reducing oxidative damage and upregulating the expression of GLUT4 [32]. In our study, QUE could increase GLUT4 and p-AKT protein levels, and reduce p-IRS-1 protein expression. Overexpression of METTL3 or silence of PRKD2 reversed the promotive effects of QUE on GLUT4 and p-AKT protein expression and the inhibitory effect on p-IRS1 protein level.

Conclusions
Our study revealed that noncytotoxic doses of QUE could ameliorate IR and repress oxidative stress in skeletal muscle of HFD-fed mice and PA-stimulated insulin-resistant C2C12 cells through METTL3mediated N6-methylation of PRKD2 mRNA, suggesting that QUE may be a potential therapeutic target against hyperinsulinemia.

Animals
Male C57BL/6 mice at 3 months of age were obtain from laboratory animal center of Medicine College of Xi'an Jiaotong University and kept in the sterile cages under controlled conditions (20 ± 2ºC, humidity 40-50%, cycle 12h/12h light/dark) with free access to food and water. Mice were randomly divided into six groups (n = 8 per group): control group, high fat diet (HFD) group, HFD + QUE group, HFD + QUE + Lenti-NC + Scramble group, HFD + QUE + Lenti-METTL3 + Scramble group and HFD + QUE + Lenti-NC + PRKD2 siRNA group. The mice of control group were feed with a normal chow diet, which contained 70% carbohydrate calories, 20% protein calories, and 10% fat calories. HFD contained 20% carbohydrate calories, 20% protein calories, and 60% fat calories. For QUE treatment, HFD-fed mice for 8 weeks were intraperitoneally injected with QUE (10 mg/kg) daily for 10 weeks. Lenti-METTL3 and Lenti-NC were purchased from GeneChem (Shanghai, China) and injected to mice via the tail vein once a week for four weeks (1×10 9 units/100µL). Small interfering RNA targeting serine-threonine kinase protein kinase D2 (PRKD2 siRNA) were injected intraperitoneally into mice at 2mg/Kg. All mice were fasted for 12 h before sampling. After experiments, the mice were anaesthetized with an intraperitoneal injection of pentobarbital sodium (80 mg/Kg), and were euthanized with an intraperitoneal injection of excessive sodium pentobarbital. The research was approved by the ethics committee of the Second A liated Hospital of Xi'an Jiaotong University.

Glucose tolerance test (GTT) and insulin resistance test (IRT)
Glucose and insulin tolerance tests were conducted in mice after fasting for one night. Brie y, mice were intraperitoneally injected with glucose at a dose of 2 g/Kg body weight, and blood samples were collected from the tail tip at the time points of 30, 60, 90 and 120 min. For IRT, the mice were received an intraperitoneal injection of 0.75 IU/kg insulin, and blood glucose concentrations were measured after 30, 60, 90 and 120 min respectively. Glucose levels were analyzed with glucometer (Yuyan Instruments, Shanghai, China). The insulin resistance index was calculated using the equation: homeostasis model assessment of IR (HOMA-IR) = blood glucose (mmol/L) × serum insulin (µU/mL)/22.5.
Measurement of triglyceride (TG), superoxide dismutase (SOD), antioxidant no-enzymatic (GSH) and oxygen species (ROS) TG levels from mouse plasma were measured using an ELISA kit (ab65336, Abcam, Cambridge, UK). The levels of SOD and GSH from C2C12 cell suspensions and mouse skeletal muscle lysates were measured using ELISA under standard conditions, according to the manufacturer's instructions on the reagent kits. ROS content was detected with ROS assay kit (ab113851, Abcam, Cambridge, UK). The absorbance at the corresponding wavelength was measured with a microplate reader (Molecular Devices, Sunnyvale, CA, USA). Glucose uptake assay Glucose uptake assay was performed using the uorescent D-glucose analog 2-NBDG to monitor intracellular glucose uptake. C2C12 cells were seeded in 96-well plates at a density of 4×10 5  m6A MeRIP-qPCR Total RNA from C2C12 cells was isolated using PolyATtract mRNA isolation systems (Promega, Madison, USA). Then, mRNA was denatured at 70℃ for 10 min, and 5 µL of anti-m6A antibody (ab208577, Abcam, Cambridge, UK) or anti-IgG (ab172730, Abcam, Cambridge, UK) was conjugated to protein A/G magnetic beads in IP buffer (150 mM NaCl, 10 mM Tris-HCL, and 0.1% NP-40) for 2 h at 4 ℃. A total of 100 ng RNA was incubated with antibodies in a buffer containing RNase inhibitor and protease inhibitor. m6A RNA was eluted with 6.7 mM of N6-methyladenosine 5'-monophosphate sodium salt (Sigma, St. Louis, MO, USA) for 1 h at 4 ºC. Total RNA was eluted with elution buffer and puri ed by phenol chloroform extraction. The enrichment of m6A was determined with RT-qPCR analysis. Fold enrichment was calculated by calculating the 2 −ΔCt of eluate relative to the input sample.

Cell viability assay
Total RNA methylation quanti cation Total RNA from C2C12 cells was extracted using TRIzol reagent (Invitrogen, Carlsbad, USA). m6A level in the total RNA was measured with a commercial m6A RNA methylation quanti cation kit (ab185912; Abcam, Cambridge, UK). In brie y, 200 ng of RNAs were added into each assay well, followed by a capture antibody solution and a detection antibody solution, respectively. Colorimetric quanti cation was analyzed through reading the absorbance at 450 nm wavelength, and the level of m6A of each sample was colourimetrically measured according to the standard curve.

Hematoxylin-Eosin (HE) Staining
Liver tissues were xed in 4% paraformaldehyde overnight and then embedded in para n. Slices of liver sections were dewaxed with xylene, then dehydrated in ethanol. The slices were stained with HE (Synthgene, Nanjing, China). The histological images were captured by an inverse microscope equipped with an LCD camera (Olympus, Tokyo, Japan).
Immuno uorescence analysis C2C12 cells were xed with 4% paraformaldehyde for 15 min and then cells were permeabilized using Triton X-100 (0.1%) and subsequently blocked with 5% goat serum for 2 h at room temperature. Next, primary antibody anti-PRKD2 (ab51250, Abcam, Cambridge, UK) was added and cultivated overnight at 4 ℃. Alexa uor 488 uorescein conjugated secondary antibody (ab150077, Abcam, Cambridge, UK) was replenished and incubated at room temperature for 1 h. Finally, DAPI was incubated at room temperature for 10 min. Fluorescence intensity was evaluated by uorescence microscopy (MF43-N, Mshot, China).

Statistical Analysis
All experiment were repeated 3 times independently and all data were reported as the mean ± SEM using GraphPad Prism 7.0 statistical software (GraphPad Software, La Jolla, CA, USA). One-way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test was used for comparisons between three groups. ** P<0.05 was regarded signi cant difference.

Authors' contributions
Jiao Y performed the experiments, statistical analysis and prepared the manuscript; Albert Williams analyzed and proofread the data and prepared the gures; Wei N participated in the study design and conceptualization, supervised and directed the work, drafted and revised the manuscript.

Funding
This study was supported by Key R&D projects of Shaanxi provincial (2020SF-165).

Availability of data and materials
The datasets used or analyzed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate All subjects had not received preoperative radiotherapy and/or chemotherapy and were obtained with informed consent before surgery. The protocols were approved by the Ethics Committee of the Second A liated Hospital of Xi'an Jiaotong University.

Consent for publication
The manuscript is approved by all authors for publication.          Schematic representation illustrating the effects of QUE on the progression of IR through METTL3mediated N6-methyladenosine modi cation of PRKD2 mRNA and oxidative stress.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplementarymaterials.docx