Resveratrol improves skeletal muscle insulin resistance through downregulating lncRNA NONMMUT044897.2

Background: Long non-coding RNA (lncRNA) has proved to be crucial factors in the progression of insulin resistance (IR). Resveratrol (RSV) exhibits promising therapeutic potential for the IR. Nonetheless, whether RSV could inuence the expression of lncRNAs and the interaction mechanisms in IR remain unclear. Methods: We conducted high-throughput sequencing to detect the lncRNAs and mRNAs expression signatures and the co-expression network of lncRNAs and mRNAs in skeletal muscle after a high-fat diet (HFD)-induced IR mice model with or without RSV treatment, including hierarchical clustering, gene enrichment and gene co-expression networks analysis. Highly differentially expressed lncRNAs were selected and validated by RT-qPCR. Finally, the biological functions of the selected lncRNAs were investigated by silencing expressing the target genes through lentivirus transfection in C2C12 mouse myotubes cells. Results: We revealed that 338 mRNAs and 629 lncRNAs whose expression in skeletal muscle after a high-fat diet (HFD)-induced IR mice model was reversed by RSV treatment. Gene Ontology and Kyoto encyclopedia of genes and genomes databases indicated that the differential expression mRNAs modulate the insulin signaling pathway. After validating randomly selected lncRNAs via RT-qPCR, we found that lncRNA (NONMMUT044897.2) and Suppressor of Cytokine Signaling 1 (SOCS1) were up-regulated in the HFD group, and reversed by RSV treatment. Additionally, NONMMUT044897.2 was validated to function as a ceRNA of microRNA (miR)-7051-5p and SOCS1 was conrmed as a target for miR ‐ 7051-5p. We further performed lentivirus transfection to knockdown NONMMUT044897.2 in vitro and found that NONMMUT044897.2 silence inactivated SOCS1 and promoted the insulin signaling pathway. Importantly, RSV could mimic the effects of silencing NONMMUT044897.2. Conclusion: Our study revealed that resveratrol improves skeletal muscle IR might be via regulation of NONMUT044897.2. after reaching 80% conuence. Differentiated C2C12 cells were incubated for 24 hours with 0.25mM palmitate (PA)[21] (Aladdin Industries, China).At 0h, 8h, 16h, 24h after the intervention of PA, the glucose concentration in the culture medium was determined by the glucose oxidase assay to determine whether the insulin resistance model was established. Subcultured C2C12 cells were digested to prepare a cell suspension and then subcultured to a 96-well culture plate. When the cells grew to about 80%, add resveratrol medium with different concentrations of 100µM, 50µM, and 30µM. After 24 hours, add 10ul of CCK-8 to each well (need to be protected from light), culture for 20 minutes, measured the absorbance at 450 nm and calculated the cell survival rate. C2C12 cells in the logarithmic growth phase were subcultured in a 6-well plate and then transfected with lentivirus. Synthesized constructs including LV3-NC (5’ to 3’ TTCTCCGAACGTGTCACGT), LV3-NONMMUT044897.2 (5’ to 3’ GCTCTTTCAGATAAGCCTTGT) obtained from Genepharma Co., Ltd. China. Stable cell lines were obtained after puromycin selection for the PA-induced IR model and drug intervention experiments. The plated cells were grouped: control group (CON), PA group (PA), PA + shRNA-NONMMUT044897.2 negative control group (PA + shRNA-NC), PA + shRNA-NONMMUT044897.2 knockdown group (PA + shRNA-NONMMUT044897.2), and PA + RSV 30µM group (PA + RSV). After successful modeling, the glucose concentration was detected for 24 hours, and the NONMMUT044897.2 and miR7051-5p mRNA expression levels were measured by RT-qPCR. The cells were stimulated with insulin, and the protein was extracted 20 minutes after insulin stimulation for Western blot analysis and also cell cultures were extracted by TRIzol reagent (Invitrogen). RSV could improve IR by regulating the expression of lncRNAs in skeletal muscle remains undened. In this study, we further studied skeletal muscle, a different tissue than previous studies, and found a novel lncRNA NONMMUT044897.2 that may be involved in resveratrol's improvement of skeletal muscle IR in vivo. High-throughput sequencing showed that there were 3276 differential lncRNAs and 2118 mRNAs in HFD mice compared to the CON group, and 1640 differential lncRNAs and 604 mRNAs compared to the HFD + RSV group. We further provided 338 mRNAs and 629 lncRNAs whose expression was reversed between HFD and HFD + RSV groups, suggesting that RSV is under a strong role in the overall alteration of skeletal muscle gene expression. Moreover, via RT-qPCR uncovered that RSV improved IR by regulating the expression of lncRNAs in skeletal muscle. As mentioned above, the veried lncRNAs were consistent with the sequencing results and NONMMUT044897.2 was highly expressed. GO and KEGG analysis uncovered that the differential genes were part of the insulin signaling pathway. We found that NONMMUT044897.2 was associated with SOCS1, which is critically involved in the insulin signaling pathway, so we selected this lncRNA for further study, which hadn't been reported before. fat diet can result in the decrease of skeletal muscle IRS-1, P13K, AKT, GLUT4 gene expression, reduce p-AKT (ser473), p- GSK3β protein expression. Studies have reported that overexpression of SOCS1 can inhibit the phosphorylation and activation of IRS-1[32, 33], which in turn inhibits the activation of AKT, indicating that there is an important link between AKT and SOCS1. We found that in IR model mice, mRNA and protein expression levels of SOCS1 was signicantly increased, RSV treatment reversed this trend, thereby improving insulin resistance and decreasing blood glucose. the differential expression of lncRNAs between the IR mice model and RSV treatment and revealed the potential


Introduction
Insulin resistance (IR) has to be considered as a primary determinant for metabolic diseases, which reduce glucose uptake and utilization [1]. Skeletal muscle plays a signi cant part in the etiology of IR [2]. Resveratrol (3, 5, 4-trihydroxystilbene; RSV) is a kind of natural polyphenol, enriched in more than 70 kinds of plants [3]. Accumulating evidence has indicated that RSV displayed a diversity of biological activities [4,5], including anti-oxidative, anti-aging, anti-in ammatory, hypoglycemic property, and attenuate insulin resistance [6,7]. Several studies have indicated RSV possesses a satisfying anti-insulin resistance activity in skeletal muscle [8,9,10]. It is related to the activation of 5-adenosine monophosphate-activated protein kinase and restoration of glucose transporter 4 translocations [8,9]. Our earlier research also proved that skeletal muscle insulin resistance caused by a high-fat diet may be alleviated after RSV treatment [10].
Long non-coding RNAs (lncRNAs) are a class of RNA molecules that more than 200 nucleotides in length and have little or no protein-coding capacity [11]. Growing research has shown that lncRNAs widely participate in plenty of developmental and physiological processes [12,13], and are strongly correlated with the genesis and development of diseases, including coronary artery diseases [14], cancers [15], and metabolic diseases [16,17]. Recently, the functions of lncRNAs in IR have gained a lot of attention [18,19].
To date, however, there are limited studies on RSV regulating the function of lncRNAs. Given that, we intended to observe the effects and potential mechanisms of RSV on skeletal muscle IR of mice through measuring the whole genome expression pro le of lncRNAs and mRNAs and to probe the correlation between RSV and lncRNA (NONMMUT044897.2) in improving IR in vivo and in vitro, providing new therapeutic targets for IR.

Animal experiments
We developed to control group, the high-fat (HFD) group, and the HFD + RSV group model by using C57BL/6J background mice (n = 14 in each group). Total 42 healthy 6 weeks old clean-grade C57BL/6J male mice weighing around 22g were purchased from Beijing Viton Lihua Experimental Center and sustained on a standard 12h light-dark cycles, with temperature (20-25℃) and humidity (40-60%). HFD mice were developed by feeding with a diet containing D12492J feed (20% protein, 20% carbohydrate, 60% fat) for 8 weeks. HFD + RSV mice were intragastrically applied with 100 mg/kg/day RSV solution for 6 weeks. Dissolved RSV (Sigma Aldrich, USA) with dimethyl sulfolane (Sigma Aldrich; 30 mg.mL − 1 ) and then diluted with 0.9%NaCl in a ratio of 1:2. For control mice, were received normal chow (D12450J contains 20% protein, 70% carbohydrate, and 10% fat). Weight and food intake were measured weekly during feeding. After the feeding experiment, all mice fasting for 12 hours and then were received intraperitoneal injection of 50% glucose (1.5g.kg-1 bodyweight) for glucose tolerance tests (IPGTT). Blood glucose was detected from the tail vein with a glucose meter at 0, 15, 30, 60, and 120 minutes post-injection, and the model of insulin resistance was validated by the area under the curve (AUC). Animal studies were approved by the ethics committee of the Hebei General Hospital, and all animal experimental procedures complied with the National Institutes of Health guide for the care and use of laboratory animals.

Serum and tissue samples
Three mice in each group were randomly selected and given an intraperitoneal injection of 1.5U/40g of insulin (Sigma Aldrich). Cervical dislocation and euthanasia were performed 20 min later. The blood samples were gathered by puncture heart and then centrifuged at 3000×g at 4°C for 10min, then the serum was stored at -80℃. Collected skeletal muscles were withdrawn quickly and stored in liquid nitrogen for the follow-up study.

Serological indicators were tested
Detection kits for total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and free fatty acid (FFA) were acquired from Nanjing JianCheng Institute of Biological Engineering (Jiangsu, China). Serum insulin was obtained with an ELISA kit (ALPCO Diagnostics, USA). All steps followed the manufacturer's protocol.

Western blot
The same amount of protein with different groups was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, then transferred to PVDF membrane and sealed with 5% skimmed milk for 2 h. Diluted the primary antibodies in the blocking solution at the following concentration:β-actin: rabbit antibody, 1:5000; GAPDH: rabbit antibody, 1:10000; AKT: Rabbit antibody, 1:1000;p-AKT(Ser 473): Rabbit antibody, 1:750; GSK3β: Rabbit antibody, 1:750; p-GSK3β: Rabbit antibody, 1:750; GLUT4: Rabbit antibody, 1:5000, SOCS1: Rabbit antibody, 1:5000. The antibodies were acquired from Cell Signaling Technology (Danvers, USA) and Abcam (Cambridge, UK). PVDF membranes and primary antibodies were incubated at 4℃ for 24h. Washed and incubated membranes with second antibodies at RT for about 50 minutes and then washed three times for 10 minutes each time. Protein bands were calculated by densitometry taking advantage of the Image J software and normalized to β-actin or GAPDH levels.

RT-qPCR
Total RNAs were extracted with Trizol reagent and then tested for RNA purity and concentration using NanoDrop 2000 (Fisher Scienti c, USA).
The PrimeScript™ RT Reagent Kit with gDNA Eraser was reverted and ampli ed with SYBR® Premix Ex Taq™ II Kit (RR820A). Applied Biosystems 7500 real-time PCR systems to perform RT-qPCR, with a total of 41 cycles, including 3 minutes of pre-denaturation at 95°C, 5 seconds at 95°C, and 32 seconds at 60°C. The melting point curve was established at 60-95℃. β-actin and U6 were considered as an internal reference control for genes, respectively. The relative gene expression was quanti ed by the 2 −(ΔΔCt) method [20]. The speci c primers involved in this research are listed in Table 1. LncRNAs and mRNAs were quantitative analysis by Sinotech Genomics Co., Ltd (Shanghai, China). Illumina Novaseq 6000 was used to construct a high-throughput RNA sequencing based on the removal of ribosomal RNA. Fastp software was performed to lter the sequence to get the clean reads. Clean reads were mapped to the GRCM38 reference genome using Hisat2. Each gene fragment was counted using Stringtie software contrast and then normalized by using the TMM (trimmed mean of M values) algorithm and then calculated FPKM value of each gene.

Differential expression of lncRNAs and mRNAs analysis.
The differential expression of skeletal muscle in three groups was analyzed based on the edgeR software package. The threshold of the p-value was con rmed by controlling the False Discovery Rate. The screening criteria for differential expression of mRNAs and lncRNAs were P < 0.05 and fold change (FC) > 2.0.

Functional group analysis.
Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway were determined the potential role of the lncRNAs coexpressed with the differentially expressed mRNAs. GO analysis was implemented to establish signi cative annotations of genes and gene products in diversi ed organisms using the DAVID database (http://david.abcc.ncifcrf.gov). In addition, KEGG pathway analysis was used to make differential expression mRNAs in enriched pathways. P < 0.05 was identi ed as the signi cance threshold.
The co-expression network of NONMMUT044897.2 and microRNA and mRNA was analyzed by using Cytoscape software 3.6.0 (Cytoscape, USA).

C2C12 cell culture and treatments
C2C12 mouse myotubes cells were maintained in Dulbecco's modi ed Eagles medium (DMEM; Gibco, USA) containing 10% fetal bovine serum (San Diego, USA) and 1% penicillin/streptomycin (Wisent, China,) at 37 ℃ with 5% CO2. Cells differentiation was induced by incubation for 4 days in DMEM containing 2% FBS after reaching 80% con uence. Differentiated C2C12 cells were incubated for 24 hours with 0.25mM palmitate (PA) [21] (Aladdin Industries, China).At 0h, 8h, 16h, 24h after the intervention of PA, the glucose concentration in the culture medium was determined by the glucose oxidase assay to determine whether the insulin resistance model was established. Subcultured C2C12 cells were digested to prepare a cell suspension and then subcultured to a 96-well culture plate. When the cells grew to about 80%, add resveratrol medium with different concentrations of 100µM, 50µM, and 30µM. After 24 hours, add 10ul of CCK-8 to each well (need to be protected from light), culture for 20 minutes, measured the absorbance at 450 nm and calculated the cell survival rate. C2C12 cells in the logarithmic growth phase were subcultured in a 6-well plate and then transfected with lentivirus. Synthesized constructs including LV3-NC (5' to 3' TTCTCCGAACGTGTCACGT), LV3-NONMMUT044897.2 (5' to 3' GCTCTTTCAGATAAGCCTTGT) obtained from Genepharma Co., Ltd. China. Stable cell lines were obtained after puromycin selection for the PA-induced IR model and drug intervention experiments. The plated cells were grouped: control group (CON), PA group (PA), PA + shRNA-NONMMUT044897.2 negative control group (PA + shRNA-NC), PA + shRNA-NONMMUT044897.2 knockdown group (PA + shRNA-NONMMUT044897.2), and PA + RSV 30µM group (PA + RSV). After successful modeling, the glucose concentration was detected for 24 hours, and the NONMMUT044897.2 and miR7051-5p mRNA expression levels were measured by RT-qPCR. The cells were stimulated with insulin, and the protein was extracted 20 minutes after insulin stimulation for Western blot analysis and also cell cultures were extracted by TRIzol reagent (Invitrogen).

Statistical analysis
SPSS23.0 was used for data analysis and the results were indicated as mean ± SD. One-way ANOVA was performed for comparison between groups. P < 0.05 was de ned as statistically signi cant.

Establishment of IR animal model developed by a high-fat diet
Before the intervention, the body weight of the control group (CON) was similar to that of the high-fat diet group (HFD). After one week of dietary intervention, HFD mice were dramatically heavier than CON mice (Fig. 1A), while their average daily calorie intake was comparable (Fig. 1B).
Besides, the intraperitoneal glucose tolerance test (IPGTT) 8 weeks after dietary intervention showed that the blood glucose level of the HFD group was drastically elevated compared with the CON group( Fig. 1C) at 0, 30, 60, and 120 minutes after glucose saline injection. Meanwhile, the HFD group resulted in a remarkable augment of AUC in contrast to the CON group, (Fig. 1D), indicating that the insulin resistance model was successfully established.
3.2. RSV ameliorates body weight,insulin resistance and lipid levels in HFD-fed mice After 2 weeks of resveratrol (RSV) administration, the weight of the HFD + RSV group was greatly declined than that of the HFD group ( Fig. 2A), although the daily calorie intake between the groups was similar (Fig. 2B). Also, RSV treatment restricted the blood glucose level of IPGTT and AUC level (Fig. 2C, 2D). As compared with the CON mice, fasting blood glucose was robustly elevated in the HFD group as well as insulin levels ( Fig. 2E, 2F).RSV treatment showed a marked reduction in both of them (Fig. 2E, 2F). The quantitative insulin sensitivity index (QUICKI) of HFD mice was declined than that of CON and HFD + RSV groups (Fig. 2G). As compared with the CON group, the levels of TG, TC, LDL-C, and FFA were strikingly increased in the HFD group. Also, RSV treatment could abolish the up-regulation of TG, LDL-C, and FFA in the HFD group, while TC decreased but had no signi cance ( Fig. 2H-L). There was no difference in high-density lipoprotein cholesterol (HDL-C) of the mice.

RSV treatment reduced SOCS1and increased phosphorylation of AKT and GSK3β in HFD mice
In the CON, HFD, and HFD + RSV mice, no differences were found in the mRNA and protein levels of AKT and GSK3β (Fig. 3A, 3B, 3D, 3E, 3G). The HFD group dramatically repressed p-AKT and p-GSK3β protein levels compared with the CON group, and the RSV treatment showed a marked increase in p-AKT and p-GSK3β protein expression (Fig. 3D, 3F, 3H). Moreover, Suppressor of Cytokine Signaling 1 (SOCS1) in the HFD group was abnormally elevated, and the increase in the expression of SOCS1 was abrogated upon RSV treatment (Fig. 3C, 3D, 3I). These results suggest that RSV improves gene expression on the insulin signaling pathway.

RSV systematically modulates skeletal muscle gene expression
The expressions of lncRNAs and mRNAs in CON, HFD, and HFD + RSV were determined by high-throughput sequencing. Results demonstrated that after standardization, 58245 lncRNAs and 83089 mRNAs were identi ed in the skeletal muscle of mice. By comparison of HFD with the CON group, there were 3276 differentially expressed lncRNAs (1192 up-regulated and 2084 down-regulated) and 2118 differentially expressed mRNAs (314 up-regulated and 1804 down-regulated). Simultaneously, there were 1640 differential expression lncRNAs (921 up-regulated and 719 down-regulated) and 604 differential expression mRNAs (444 up-regulated and 160 down-regulated) in the HFD + RSV group compared with the HFD group. Among the up-regulated lncRNAs and mRNAs in the HFD group, 270 lncRNAs and 58 mRNAs were down-regulated in the HFD + RSV group. Among the lncRNAs and mRNAs down-regulated in the HFD group, 359 lncRNAs and 280 mRNAs were up-regulated in the HFD + RSV group. The top 30 differential expression lncRNAs and mRNAs are listed in Table 2 and 3. With FPKM 0 were eliminated.

Cluster analysis results and heat maps of lncRNAs and mRNAs
Through further cluster analysis of differential expression genes, data uncovered that there were signi cant differences in the expression patterns of lncRNAs and mRNAs among the three groups. Meanwhile, the cluster analysis results of the three groups of lncRNAs were similar to the mRNAs, suggesting that there was a close relationship between the expression of lncRNAs and mRNAs. (Fig. 4A,4B). Venn diagrams indicated that the HFD group had 3276 differentially expressed lncRNAs and 2118 differentially expressed mRNAs compared with the CON group. At the same time, the HFD + RSV group had 1640 differentially expressed lncRNAs and 604 differentially expressed mRNAs in comparison with the HFD group. (Fig. 4C, 4D) 3.6. Functional enrichment analysis of differentially expressed genes Page 10/20 The functions of these different mRNAs were studied by enrichment analysis. GO analysis classi ed differentially expressed mRNAs into three types: biological process (BP), molecular function (MF), and cellular component (CC). mRNA differential expression mainly took part in the following biological processes: transmembrane receptor protein tyrosine kinase signal transduction, secretion, hormone receptor, cytokine receptor, negative regulation of cell metabolism process, growth, receptor protease-related signaling pathway, cell response to hormone stimulation, and cell response to cytokine stimulation (Fig. 5A). KEGG analysis showed that the differential expression mRNAs involved in ubiquitin-mediated proteolysis, splice, endoplasmic reticulum stress protein, prolactin signaling pathway, MAPK signaling pathway, PI3K-Akt signaling pathway, osteoclast differentiation, microRNA in cancer, JAK-STAT signaling pathway, insulin signaling pathway, and estrogen signaling pathway (Fig. 5B). SOCS1 plays a vital role in the insulin signaling pathway, which was reported to involve in the development of IR.

RT-qPCR validation in vivo
To con rm the validity of the sequencing results, we next randomly picked 4 differential lncRNAs, two lncRNAs up-regulated in HFD, downregulation of HFD + RSV (NONMMUT044897.2, NONMMUT005295.2), and two lncRNAs with a reverse trend (NONMMUT128951.1, NONMMUT145909.1). Selected lncRNAs expression levels were consistent with the sequencing results ( Fig. 6A-D), but there was no statistical difference in the increase of NONMMUT128951.1 in the HFD + RSV group. Among the veri ed lncRNAs, NONMMUT044897.2 has a higher expression level. In addition, GO and KEGG analysis indicated that differential expression mRNAs enriched in the insulin signaling pathway.
Hence, we chose this pathway and predicted the closely related mRNA SOCS1 and NONMMUT044897.2 through pathway analysis, and the trends of the two were consistent. To elucidate the interaction between them, we constructed a related lncRNA-miRNA-mRNA network diagram.
The results revealed that this lncRNA regulates mRNA SOCS1 through miR-7051-5p and miR-762 (Fig. 6F). To further explore its potential molecular mechanism, we applied NonCode and miRBase database analysis and found that there was base pairing in the sequence of NONMMUT044897.2 and the sequence of miR-7051-5p. At the same time, according to the prediction results of Targetscan, it can be found that SOCS1 is a miR-7051-5p target (Fig. 6G). According to the above results, NONMMUT044897.2 may regulate SOCS1 through miR-7051-5p.

Establishment of a cell model of PA-induced IR
The cells were transferred to medium with and without 0.25mM PA, and glucose concentrations were determined at 0 h, 8 h, 16 h, and 24 h.
There was no signi cant difference in CON and PA group at 0 h, 8 h, and 16h, however, the glucose concentration of the PA group was drastically elevated at 24 h in comparison with the CON group, indicating that the IR model was successfully established (Fig. 7A). Besides, the glucose concentration was dramatically relieved at 24 h when RSV treatment (Fig. 7B).
The cell survival rate of C2C12 cells 24 hours after RSV of 30 ~ 100 µm was observed. The results showed that RSV of 30µm had no signi cant in uence on the survival rate of C2C12 cells (Fig. 7C). The cell survival rate of the PA group (84%) was lower than that of the Con group (89.6%), and the PA + RSV 30 µM group (85%) was higher than that of the PA group. However, there was no statistical difference among the three groups (Fig. 7D).

RT-qPCR validation in vitro
Two lncRNAs up-regulated in PA, downregulation of PA + RSV (NONMMUT044897.2, NONMMUT139818.1) and three lncRNAs with reverse trend (NONMMUT071570.2 NONMMUT065156.2 NONMMUT00000181045). The expression levels of selected lncRNAs were consistent with the sequencing results (Fig. 8A-E), but there was no statistical difference in the increase of NONMMUT00000181045 in the PA + RSV group. To verify the relationship between NONMMUT044897.2 and RSV in vitro, lentivirus transfected C2C12 cells. Results shown that as compared with the CON group, NONMMUT044897.2 expression was robustly increased in the PA group and PA + shRNA-NC group. The PA + shRNA-NONMMUT044897.2 group suppressed the expression of NONMMUT044897.2 when compared to the PA group. RSV administration also decreased the expression of NONMMUT044897.2 in comparison with the PA group (Fig. 8G).
Relative to the CON group, the miR-7051-5p mRNA in the PA group and the PA + shRNA-NC group was signi cantly declined, while knockdown of NONMMUT044897.2 and RSV treatment prominently enhanced miR-7051-5p mRNA level (Fig. 8F).The concentration of glucose in the medium of the PA group and the PA + shRNA-NC group was substantially increased when compared with the CON group. NONMMUT044897.2 silence and RSV treatment strikingly overturned the glucose concentrations in the medium (Fig. 8H). Knockdown of NONMMUT044897.2 distinctively upregulated the p-AKT, p-GSK3β, and GLUT4 protein levels, and greatly reduced the SOCS1 protein level, when compared with the PA group. RSV treatment had a similar effect on p-AKT, p-GSK3β, GLUT4, and SOCS1 protein levels which were comparable to the NONMMUT044897.2 silence ( Fig. 8I-N).

Discussion
Resveratrol is used in the eld of insulin resistance improvement [22,23]. RSV increases the expression of microencapsulated protein 3 (CAV-3), thereby allowing skeletal muscle cells to carry glucose the protein GLUT4 activates the transfer from the cytoplasm to the cell membrane, which in turn increases the ability of myocytes to transport glucose and improve insulin resistance [24]. In addition, RSV promotes the beta-oxidation process of fatty acids in the cell mitochondria [25], so that intracellular lipids are better metabolized. RSV also improves IR in skeletal muscle by reducing SNARE proteins in diabetic rats [26],which is involved in GLUT4 transport. Meanwhile, RSV attenuates insulin-stimulated AKT phosphorylation by eliminating insulin-induced ROS production in skeletal muscle [27].
A large amount of experiments has veri ed the therapeutic effect of RSV in IR. In this study, we demonstrated that the in uence of RSV in the improvement of insulin resistance in high-fat diet mice as reported previously [24,25,26,27]. Not only did blood glucose, insulin index, and area under the curve (AUC) decreased in high-fat mice after applying RSV, but also improved blood lipid levels. RSV treatment improved high-fat dietinduced mice insulin resistance by restoring insulin signaling pathway gene expression. After the intervention of RSV, the mRNA and protein expression levels of p-AKT, p-GSK3β and GLUT4 increased signi cantly after RSV application.
The expression of lncRNA has spatiotemporal speci city, and it participates in the process of gene regulation and biological function regulation in epigenetics, transcription, and post-transcription levels [28]. Abnormal expression of lncRNAs could affect the occurrence and development of human diseases, including non-alcoholic fatty liver, various cancers, and T2DM [29]. At present, the function of most lncRNAs is completely unknown. Previous studies [30,31]  High-throughput sequencing showed that there were 3276 differential lncRNAs and 2118 mRNAs in HFD mice compared to the CON group, and 1640 differential lncRNAs and 604 mRNAs compared to the HFD + RSV group. We further provided 338 mRNAs and 629 lncRNAs whose expression was reversed between HFD and HFD + RSV groups, suggesting that RSV is under a strong role in the overall alteration of skeletal muscle gene expression. Moreover, via RT-qPCR uncovered that RSV improved IR by regulating the expression of lncRNAs in skeletal muscle. As mentioned above, the veri ed lncRNAs were consistent with the sequencing results and NONMMUT044897.2 was highly expressed. GO and KEGG analysis uncovered that the differential genes were part of the insulin signaling pathway. We found that NONMMUT044897.2 was associated with SOCS1, which is critically involved in the insulin signaling pathway, so we selected this lncRNA for further study, which hadn't been reported before.
SOCS1 is a speci c negative regulator that regulates the JAK/STAT pathway [32]. The expression of SOCS1 increases under insulin resistance and overexpression of SOCS1 decreases the phosphorylation of IRS-1; overexpression of SOCS-1 can inhibit insulin-induced glycogen synthesis in L6 myotubes [33]. AKT has essential roles in many signaling pathways, such as cell survival and cell metabolism. AKT is the center of the insulin signaling pathway, which regulates glucose and lipid metabolism. Activated AKT can stimulate the translocation of insulin-sensitive GLUT4 to the cell membrane through its downstream substrate ASl60 to increase glucose uptake; it can also phosphorylate GSK3β to inhibit its activity, promote glycogen synthesis, lower blood sugar, and improve IR [34]. A high fat diet can result in the decrease of skeletal muscle IRS-1, P13K, AKT, GLUT4 gene expression, reduce p-AKT (ser473), p-GSK3β protein expression. Studies have reported that overexpression of SOCS1 can inhibit the phosphorylation and activation of IRS-1 [32,33], which in turn inhibits the activation of AKT, indicating that there is an important link between AKT and SOCS1. We found that in IR model mice, mRNA and protein expression levels of SOCS1 was signi cantly increased, RSV treatment reversed this trend, thereby improving insulin resistance and decreasing blood glucose.
Numerous studies have demonstrated that lncRNAs may involve in human diseases by regulating miRNA expression [35,36]

Conclusion
This research pro led the differential expression of lncRNAs between the IR mice model and RSV treatment and further revealed the potential regulated lncRNA NONMMUT044897.2. Much more work remains to be done to prove the relationship between the NONMMUT044897.2/miR-7051-5p/SOCS1and RSV in the IR model. There are also limitations. First, in vivo animal models are also needed to further perform by silencing NONMMUT044897.2 to verify the in uence on insulin resistance. Second, overexpression of NONMMUT044897.2 should be done in vitro. Third, knockdown of the NONMMUT044897.2 in the RSV group should be done to observe the in uence on the insulin resistance model. Overall, our data indicated that RSV could promote skeletal muscle insulin resistance, at least partially, via a lncRNA NONMMUT044897.2 /miR-7051-5p/SOCS1 pathway, thereby provides a new perspective for RSV treatment of IR in skeletal muscle.

Declarations
Ethics approval and consent to participate All animal experiments were performed with approval from the Institutional Animal Care and Use Committee of Hebei General Hospital, Hospital.
Consent for publication N/A.

Availability of data and materials
All data, analytic methods, and study materials presented within this article are available for other investigators from the corresponding authors on reasonable request.

Competing interests
The authors declare that they have no con icts of interest to this work.

Funding
This study was supported by the grant from the Natural Science Foundation of Hebei Province (No. H201830-7071).   Gene ontology and pathway analysis of differential expression genes in three groups. A. Top 30 GO terms of differential expression mRNAs; B.
Top 30 KEGG pathways of differential expression mRNAs.