Regulation of insulin resistance and type II diabetes by hepatitis C virus infection: A driver function of circulating miRNAs

Abstract Hepatitis C virus (HCV) infection is a serious worldwide healthcare issue. Its association with various liver diseases including hepatocellular carcinoma (HCC) is well studied. However, the study on the relationship between HCV infection and the development of insulin resistance and diabetes is very limited. Current research has already elucidated some underlying mechanisms, especially on the regulation of metabolism and insulin signalling by viral proteins. More studies have emerged recently on the correlation between HCV infection‐derived miRNAs and diabetes and insulin resistance. However, no studies have been carried out to directly address if these miRNAs, especially circulating miRNAs, have causal effects on the development of insulin resistance and diabetes. Here, we proposed a new perspective that circulating miRNAs can perform regulatory functions to modulate gene expression in peripheral tissues leading to insulin resistance and diabetes, rather than just a passive factor associated with these pathological processes. The detailed rationales were elaborated through comprehensive literature review and bioinformatic analyses. miR‐122 was identified to be one of the most potential circulating miRNAs to cause insulin resistance. This result along with the idea about the driver function of circulating miRNAs will promote further investigations that eventually lead to the development of novel strategies to treat HCV infection‐associated extrahepatic comorbidities.


| INTRODUCTION
It is estimated that 75%-80% of HCV patients will progress to chronic infection, the stage highly associated with the development of intrahepatic complications that subsequently lead to metabolic disorders. Mechanistic studies have revealed that HCV protein NS5A and the core protein directly inhibit microsomal triglyceride transfer protein (MTP) activity, thereby reducing very low-density lipoprotein (VLDL) assembly and inducing hepatic steatosis. In addition, HCV core protein is also associated with excessive reactive oxidative species (ROS) production via impairment of PPARc, thereby inducing oxidative stress, mitochondrial dysfunction and steatosis. Over time, accumulation of hepatic triglycerides leads to hepatic IR via decreased insulin-stimulated glycogen synthesis and enhanced hepatic gluconeogenesis; such conditions further cause peripheral IR in multiple organs through increased circulating insulin and free fatty acid levels. 4,5 Meta-analyses have further confirmed the positive relationship between HCV and IR. 2,6 Furthermore, nondiabetic and non-obese chronic HCV patients exhibit decreased peripheral glucose uptake and hepatic IR. 7,8 Decreased glucose abnormalities have also been reported among HCV patients who have SVR after interferon treatment, 9,10 further demonstrating the role of HCV in glucose metabolism. Proposed molecular mechanisms underlying this relationship include increased serum levels of pro-inflammatory cytokines TNF-a, TNFR1, TNFR2, IL-6, IL-10, decreased adiponectin and increased intramyocellular lipocalin-2. 8,[11][12][13][14][15] These factors activate an array of intracellular signalling pathways that result in increased ROS, cellular stress and metabolic dysregulation.
There are a number of investigations on the molecular mechanisms of regulation of insulin signalling by HCV infection. Insulintreated liver tissue from non-obese and non-diabetic HCV-infected patients results in increased levels of insulin receptor and insulin receptor substrate 1 (IRS-1), but showing decreased levels of IRS-1 tyrosine phosphorylation, PI3K activity and Akt phosphorylation. 16 Similarly, HCV core protein has been found to increase serine rather than tyrosine phosphorylation of IRS-1 in hepatocytes, resulting in its degradation and impaired downstream Akt signalling. 17 HCV core protein also stimulates IRS-1 serine phosphorylation via increasing mTOR levels, again resulting in decreased Akt signalling. 18 Reduced surface expression of glucose transporters GLUT1 and GLUT2 with consequential reduction in glucose uptake in HCV-infected hepatocytes has also been reported. 19 Impairment of PI3K and Akt phosphorylation by HCV has also been demonstrated by several studies to be mediated via the increased level of suppressor of cytokine signalling 3 (SOCS3), which inhibits PI3K and Akt phosphorylation and stimulates IRS-1 and IRS-2 proteasome degradation. [20][21][22] Additionally, increased ER stress response caused by HCV infection elevates protein phosphatase 2A (PP2A), also an inhibitor of Akt. 23 Despite these findings of IR development via direct effects on insulin signalling pathway, the complex relationship between intrahepatic HCV infection and extrahepatic IR remains elusive.
Interestingly, recent evidence has implicated changes in host microRNA (miRNA) expression profiles after HCV infection. Thus, we will focus on this relationship to analyse the development of extrahepatic IR associated with HCV infection in this study. miRNAs arẽ 22 nucleotide (nt)-long non-coding RNAs that function mainly by targeting the 3 0 untranslated regions (UTRs) of mRNAs, thus playing broad roles in regulating cellular functions such as development, differentiation, apoptosis, tumorigenesis, metabolism and host-pathogen interactions. 24 The association of specific miRNAs with certain disease states has therefore greatly accelerated the development of miRNAs into new biomarkers and/or as therapeutic targets. As miR-NAs also exist in serum and other body fluids, study of extracellular and circulating miRNA profiles has become an important approach to investigate the systematic roles of miRNA in disease development.
Although the functions of circulating miRNAs remain unclear, these miRNAs have been found to be highly stable under harsh conditions including boiling, pH change, long-term storage at room temperature and multiple freeze-thaw cycles, suggesting their important biological functions. 25 More evidence has emerged to support the idea that circulating miRNAs can function like hormones to regulate gene expression in distant tissues and organs. The changes in circulating miRNA profiles have also been revealed to be correlated with various diseases including cancers, cardiovascular disorders and diabetes. [25][26][27][28][29] Based on the analysis of current knowledge regarding the relationship between HCV infection and IR, we hypothesize that circulating miRNAs post-HCV infection can target peripheral tissues and modulate their gene expression patterns, thus serving as drivers in the development of IR and T2D. Comprehensive literature review, bioinformatics and structural analyses were carried out to examine our hypothesis. The general study approach was outlined in Figure 1.
We will first assess the overlap between circulating miRNAs from HCV infection and from T2D/IR, and then analyse the target genes of these miRNAs and their corresponding metabolic functions. In addition to the mechanisms described above on insulin signalling pathways, we will focus on the identification of metabolic effects associated with circulating miRNAs to improve our understanding of the complicated relationship between HCV infection and peripheral  The importance of miRNAs in liver pathology is also reflected on many aspects. While serum miR-122 levels increase at early stage of fibrosis with high inflammation, later during the progression of fibrosis to cirrhosis, both hepatic and serum levels of miR-122 are decreased by a large cohort study with 84 liver biopsies and 167 serum samples from chronic hepatitis C patients. 36 This is largely due to the loss of hepatocytes, which is also consistent with the inverse relationship between hepatic miR-122 and liver injuries under various conditions. 37-39 miR-122 is the most abundant miRNA in liver. 40 It is also a good biomarker for liver damage correlated with alanine aminotransferase (ALT) and aspartate transaminase (AST) activities in HCV mono-infection and HIV/HCV con-infection conditions. 41,42 In addition, the increase in circulating miR-122 is greater in HIV/HCV co-infection than HIV mono-infection 43 ; miR-122 is also correlated with liver injury and necroinflammation (eg IL-6) caused by HIV and HIV/HCV infections. 42,44 Based on these important findings of miR-122 in viral infection-induced liver diseases, we will focus on miR-122 in this study to discuss its role in linking HCV infection with extrahepatic comorbidities.

| Immune response
Specific changes of miRNAs with infection have been found to impact host immune response. miR-130a expression was increased in infected hepatocytes, and liver biopsy identified a target as the antiviral protein interferon-induced transmembrane protein 1 (IFITM1); subsequent exposure to anti-miR-130a increased IFITM1 expression. 45 HCV infection has also been reported to activate miR-21, thereby suppressing HCV-triggered type I IFN production and promoting HCV replication. Two identified targets for miR-21, mye- Overexpression of miR-196 was therefore proposed as a potential therapeutic strategy to protect against oxidative stress and HCV infection. 50 55 and miRNA profiling of streptozotocin-induced diabetic mice reveals miR-134 to be up-regulated specifically in mitochondria of liver samples. 56 Interestingly, a few of these miRNAs in T2D/IR have also been detected in circulating miRNA profiles of HCV-infected individuals ( Table 2) (Table 1).  differentially expressed in all tissues, to analyse which targets are differentially expressed in which peripheral tissues. 62 As shown in Figure 3, several targets were identified to be more expressed in the liver as well in several extrahepatic tissues including thyroid, pancreas, adipose, skeletal muscle and smooth muscle. Several genes differentially expressed in adipose, skeletal muscle and pancreatic tissues with metabolic functions are experimentally verified targets of miR-122 (Table 3)

| miR-122 target genes regulate IR and T2D in skeletal muscle and adipose tissue
To further evaluate miR-122 targets that play important roles in clinical IR tissues, we analysed the relationship between miR-122 targets and the down-regulated genes (≥2-fold decrease) in human IR tissues using the microarray data from 4 studies in Gene Expression Omnibus (GEO) 63 ; 2 assessing skeletal muscle and 2 assessing adipose tissue (Table 4) Table 4 in both skeletal muscle and adipose tissue have been previously experimentally verified as well.
F I G U R E 3 miR-122 targets IR-relevant genes differentially expressed in adipose, skeletal muscle and pancreas. Distribution of miR-122 predicted targets (511 input genes) differentially expressed in various tissues and compartments was analysed via GeneAnalytics LifeMap Tool, including adipose (22), skeletal muscle (8)  Interestingly, vesicle-associated miRNAs are found to be the minimal portion in circulation, whereas over 90% of miRNAs in serum and plasma are vesicle free and associated with AGO2. 76 Figure 4C). This confirmation may be ideal as it exposes only bases that are required for mRNA targeting.
Studies have also shown that miRNAs containing high percentage of adenine exhibit reduced mRNA targeting activity. It was also found that AGO2-associated miRNAs contain less adenine nucleotides at their 3 0 -end when compared to miRNAs extracted from the whole cell. 87 Consistent with this rule, we noticed that miR-122 contains more G (40.91%) and U (36.36%) bases than A bases (18.18%) in the regions outside of the 3 0 -end, further supporting miR-122 as a strong candidate for functioning through AGO2 ( Figure 4B). cles. 91 The discovery of specific miRNA 3 0 -end sequence motifs (EXOmotifs) favours the loading into exosomes through the interaction with nuclear ribonucleoprotein A2B1 (hnRNPA2B1). 92 93 In addition, another study also identified the association of AGO2 with plasma membrane in NK cells, 94 suggesting that AGO2 might have a potential membrane function to assist miR-122 uptake by target cells. AGO2 might also be able to interact with other endocytosis complex protein(s) to facilitate miR-122 uptake. This is an untested but significant topic that warrants further investigation. Predicted miR-122 targets were compared with down-regulated genes (≥2 fold, P < .05) in skeletal muscle and adipose tissue samples in 4 reported microarray studies using Gene Expression Omnibus (GEO). Targets that are more likely associated (direct or indirect) with IR-relevant phenotypes were determined using VarElect (phenotype query: insulin resistance, glucose intolerance, lipid metabolism, type 2 diabetes, glucose metabolism, lipid storage and inflammation).

| CONCLUSION S
a Experimentally verified targets (DIANA-microT-CDS). Lists of all overlapped genes from above 5 comparisons are available in Data S3.
F I G U R E 5 Working model for hepatic and circulating miR-122 to mediate IR and diabetes associated with Hepatitis C virus (HCV) infection. Key metabolic and cell signalling pathways impacted are illustrated, summarizing the findings from literature review and bioinformatics analyses in this study miR-122 in extrahepatic tissues. Our viewpoint regarding the "hormonal" function of miR-122 is supported by a recent study that obese mice adipose tissue macrophage-derived exosome can cause systemic insulin resistance and glucose intolerance in lean mice, wherein miR-155 is the causative factor. 95 This study also suggests that other miRNAs identified in Table 2 may exert functions in the association of HCV infection with IR and diabetes. Interestingly, miR-155 is exactly on the list as one candidate in Table 2.
miRNAs identified from this analytical review will be experimen-

ACKNOWLEDGEMENT
This study is partially supported by the start-up funds to JL and the summer research scholarship to AS from Geisinger Commonwealth School of Medicine.

CONF LICT OF I NTEREST
The authors declare that they have no conflict of interest.