Evaluation of potential gene expression as earlymarkers of insulin resistance and non-alcoholic fatty liver disease in the Indonesian population

Early detection of insulin resistance (IR) or non-alcoholic fatty liver disease (NAFLD) is crucial to preventing future risks of developing chronic diseases. The Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), Liver Fat Score (LFS), and Fatty Liver Index (FLI) are generally employed tomeasure severity stages of IR andNAFLD. The study of gene expressions could explain themolecularmechanisms that occur early on in IR andNAFLD; thusprovidingpotential earlymarkers for both diseases. This studywas conducted to evaluate the gene expressions that could potentially be earlymarkers of IR and NAFLD. All participants (n = 21) had normal blood glucose and were categorized as without hepatosteatosis (n = 10), at higher risk of hepatosteatosis (n = 6), and hepatosteatosis (n = 5). Gene expression analysis was performed using the 2-∆∆CT relative quantification method. There were significant differences in galnt2 (p < 0.002) and sirt1 (p < 0.010) expression between thefirst and the third tertiles ofHOMA-IR; and in ptpn1 (p <0.012) expression between thefirst and the second tertiles of LFS. In conclusion, the expressions of galnt2 and sirt1 could be used as early markers of IR, while the expression of ptpn1 could be employed as an early marker of NAFLD.


Introduction
Metabolic syndrome is a disorder of energy use and storage.People with metabolic syndrome have an increased risk of developing chronic diseases-such as type-2 diabetes mellitus (T2DM) and non-alcoholic fatty liver disease (NAFLD) (Mendrick et al. 2017), that lead to morbidity and mortality (IDF 2005).According to the eighth edition of Diabetes Atlas, a higher prevalence of diabetes was observed in developing countries than in developed countries.In 2017, there were 7.6 million people with undiagnosed diabetes, 10.3 million people (6.2% of the population) with diabetes, and 27.7 million people with impaired glucose tolerance (IGT) in Indonesia (IDF 2017).In 2002, it was reported that about 30% of the population in Indonesia was developing NAFLD (Hasan et al. 2002).
Type-2 diabetes mellitus and NAFLD do not develop suddenly but are more likely caused by a prolonged unhealthy lifestyle.These diseases can be prevented by adopting a healthier lifestyle.Dietary patterns that are high in glucose could eventually lead to hyperinsulinemia and insulin resistance (Olokoba et al. 2012).Insulin resistance is considered to play a role in the formation of NAFLD due to abnormal fat metabolism in which hyper-insulinemia triggers triglyceride synthesis and accumulation in the liver.Non-alcoholic fatty liver disease is defined as the accumulation of triglycerides >5% (steatosis) in the liver of individuals who rarely consume alcoholic beverages (Gaggini et al. 2013).In people with a normal BMI, NAFLD is more commonly found in Asian populations and is referred to as a "metabolically obese" condition (Wong and Ahmed 2014).Therefore, the early detection of insulin resistance and NAFLD prior to diagnosis is important to prevent the future risk of developing chronic diseases (Preethi et al. 2011).
Some markers have been commonly used as clinical measurement standards, such as the homeostatic model assessment for insulin resistance (HOMA-IR) (Singh and Saxena 2010) and the Liver Fat Score (LFS) (Kahl et al. 2014) and the Fatty Liver Index (FLI) for fatty liver disease (Du et al. 2014).These markers are widely employed to diagnose and categorize developed stages of insulin resistance or NAFLD.For example, tertile analysis of HOMA-IR shows an association with glycemic control in the lean, non-diabetic Asian population (Hirata et al. 2015).However, these markers could not explain the molecular mechanism underlying the early conditions of insulin resistance and NAFLD.All blood parameters used should be collected after a 10-12 h overnight fast.ALT: alanine aminotransferase; AST: aspartate aminotransferase; BMI: body mass index; FPG: fasting plasma glucose; FPI: fasting plasma insulin; GGT: gamma-glutamyl transferase; MetS: metabolic syndrome; TG: triglyceride; T2D: Type-2-Diabetes; WC: waist circumference; WHR: waist-hip ratio.
Several genes are known to be involved in the development of insulin resistance; thus, they could be employed as potential molecular markers for insulin resistance.The relevant genes are galnt2, sirt1, and ptpn1.The expression of galnt2, a gene that encodes N-acetylgalactosaminyltransferase, could decrease the expression of enpp1, which encodes the inhibitor protein involved in insulin receptor signaling (inhibits insulin and insulin receptor (IR) interaction) (Marucci et al. 2013a).The genes sirt1 and ptpn1 have been studied for their role in the development of insulin resistance.The ptpn1 gene encodes protein tyrosine phosphatase 1B (PTP1B), a protein that could catalyze dephosphorilation of IR and insulin receptor substrates (IRS) tyrosine residues, and cause a disturbance of the insulin signaling pathway (Stull et al. 2012).Sirtuin 1 was involved in the deacetylation of PTP1B, which further deactivates PTP1B as a negative regulator of insulin, and was able to improve insulin sensitivity under conditions of insulin resistance (Sun et al. 2007).Thus, this study evaluates the gene expression of galnt2, sirt1, and ptpn1, which could potentially be early markers of insulin resistance and NAFLD in the Indonesian population.

Materials and methods
Study participants were selected from employees of PT Nutrifood Indonesia.A total of 21 participants consisting of 10 men and 11 women were involved in this study.Inclusion criteria included healthy adults aged 23-40 years, with a fasting blood glucose <100 mg/dL, no history of hepatitis, and not smoking, pregnant or breastfeeding.Subjects were asked to sign a medical action agreement, and they underwent anthropometric and blood biochemical parameters testing according to ethical clearance approved by Research Ethics Commission of UNIKA Atma Jaya.The anthropometric parameters measured were body mass index (BMI), visceral fat area, waist circumference, and hip circumference.Body composition was measured using InBody 230 (Biospace) according to the tool protocols.For fasting blood sampling, subjects were asked to fast for 10-12 h (overnight).The blood biochemical parameters (other than fasting insulin) were measured using a commercial laboratory service (Prodia).The values of HOMA-IR, LFS, and FLI were calculated using the formulas presented in Table 1.Fasting insulin was measured using Ultrasensitive Insulin ELISA (Mercodia).
The collected data were statistically analyzed using SPSS Statistics 22 (IBM, Armonk, New York, United States).Normality, outlier, and descriptive analyses were performed for each fasting blood biochemical parameter and index and for gene expression.For the gene expression analysis, HOMA-IR, LFS, and FLI were sorted from the lowest value to the highest value and were divided into tertiles.Statistical analysis was performed among tertiles of referred indices, in which this method of tertile analysis could be used to evaluate associations of markers (Hirata et al. 2015).An independent T-Test mean difference was performed on the expression of galnt2, sirt1, and ptpn1 between tertiles of HOMA-IR, LFS, and FLI.

Results and discussion
Characteristics of all study participants are presented in Table 2.All study participants were normoglycemic (FPG <100 mg/dL).Normality and outlier analysis resulted in valid HOMA-IR values for further analysis (n = 18) ranged from 0.45-2.20,LFS (n = 20) ranged from −3.99-0.71,and FLI (n = 21) ranged from 3.40-83.21.Two participants were categorized as insulin resistant based on the cutoff value of HOMA-IR ≥ 2.04 for the diagnosis of insulin resistance in Indonesia (Purnamasari et al. 2010).Thirteen participants were categorized as non-NAFLD, three participants were estimated to have elevated liver fat, four participants were predicted to have NAFLD, and no participants were diagnosed with NAFLD, based on the cut-off values of LFS < −1.413 to exclude NAFLD, LFS > −0.640 to predict NAFLD, and LFS > 1,257 to diagnose NAFLD (Kotronen et al. 2009).Ten participants were categorized without hepatosteatosis, six participants were identified as having a higher risk of hepatosteatosis, and five participants were characterized as having hepatosteatosis, based on the cut-off values of FLI < 30 to exclude NAFLD and FLI ≥ 60 to diagnose hepatosteatosis (Bedogni et al. 2006).
Based on the U.S. National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III), the triglyceride concentrations of our study population (52-242 mg/dL) were classified as normal to high (NCEP 2001).The study population levels of GGT and HDL were categorized as normal.Increased ALT levels in some participants might occur once in a while, intermittently, or be caused by steatohepatitis (Aragon and Younossi 2010).
The HOMA-IR, LFS, and FLI values were divided into tertiles.The lowest values are the first tertile (T1), and the highest values are the third tertile (T3).Analyses of study participants' characteristics relative to HOMA-IR, LFS, and FLI are presented in Table 3, 4, and 5, respectively.An analysis of gene expression among each tertile group is presented in Table 6.
There were significant mean differences of galnt2 (p < 0.002) and sirt1(p < 0.010) expression between the first tertile (T1) and the third tertile (T3) of HOMA-IR.According to study participants' characteristics among tertiles of HOMA-IR, there was a significant difference in fasting insulin among tertiles, but there was no significant difference in fasting glucose (Table 3).This result indicates that although T3 of HOMA-IR was categorized as insulin sensitive due to its normal glucose level, it was somewhat closer to insulin resistance condition than T1.Therefore, T3 of HOMA-IR can be referred to as the early condition of insulin resistance (Preethi et al. 2011).These results suggest that galnt2 and sirt1 expression can be used as markers of the early condition of insulin resistance.
The significant difference in galnt2 expression between the first and the third tertiles of HOMA-IR indicates that an increasing value of HOMA-IR in normoglycemic participants was followed by the increased expression of galnt2 (Table 6).This result seems to be contradictory to the previous study by Marucci et al. (2013b), which demonstrated that galnt2 mRNA expression levels in pe-ripheral whole blood cells were significantly reduced from control to obese to diabetics.In-vitro study in human monocytes had identified hyperglycemia as a major cause of galnt2 down-regulation in patients with T2D (Marucci et al. 2013b).Our study population had a wide range of fasting insulin levels, but they were still categorized as normoglycemic (fasting glucose level <100 mg/dL); thus, the expression of galnt2 was not expected to be heterogeneous among the subjects.Interestingly, in the T3 group, where insulin resistance levels were higher, the expression of galnt2 was higher than in the T1 group.This condition was might be due to in the early condition of insulin resistance, galnt2 would be expressed higher to inhibit enpp1, an inhibitor for insulin signaling (Marucci et al. 2013b), thus, enabling the insulin to work optimally.This result suggests that galnt2 could be a molecular marker for the early development of insulin resistance.
The significant difference in sirt1 expression between the first and the third tertiles of HOMA-IR also indicate that the increasing value of HOMA-IR is associated with a significantly higher expression of sirt1 (Table 6).This result seems to be contradictory with the previous study by Song et al. (2011) that demonstrated that sirt1 mRNA expression levels in the T2DM group were lower compared with the healthy group.SIRT1 deacetylase activity consumes NAD+, and a high blood sugar level in T2DM patients will decrease the NAD+/NADH ratio.The decrease of NAD+/NADH caused by altered homeostasis of glucose metabolism will be followed by a subsequent decrease of sirt1 expression (Kitada and Koya 2013).Therefore, as for galnt2, a hyperglycemia might also be the cause of sirt1 down-regulation in patients with T2D.Our study population had a wide range of fasting insulin levels, but they were still categorized as normoglycemic (fasting glucose level <100 mg/dL); thus, the expression of sirt1 was not supposed to be heterogeneous among the subjects.Instead, our study illustrates that a higher fasting plasma insulin (T3) within the normoglycemic population is associated with increased sirt1 expression.This result is in accordance with SIRT1's positive role in the insulinsignaling pathway by inducing insulin secretion, in which SIRT1 activation increases glucose uptake and insulin signaling (Song et al. 2011).In the early condition of insulin resistance, sirt1 would be expressed higher to induce insulin secretion until the body was unable to maintain normal blood sugar homeostasis, after which the expression of sirt1 would drop.Thus, sirt1 could be a molecular marker of the early development of insulin resistance.
There was no significant mean difference in ptpn1 expression across the tertiles of HOMA-IR.This result probably showed that ptpn1 expression would be up-regulated in the later stage of insulin resistance (Stull et al. 2012), and not in the early condition of insulin resistance.In this study, a difference of ptpn1 expression (p < 0.012) was only observed between the first and the second tertile of LFS, where the second tertile was significantly higher than the first tertile.Higher expression in T2 than in T1 in this study was in accordance with another report that found ptpn1 to be up-regulated in liver biopsies of Non-Alcoholic SteatoHepatitis (NASH) patients (Sanderson and Smyrk 2005).Higher expression of ptpn1 could also be caused by hepatocyte inflammation of TNF-α, as reported in the previous study of hepatocyte inflammation in vitro (Zabolotny et al. 2008).Therefore, these results suggest that changes in ptpn1 expression (which is a protein tyrosine kinase) in blood might also be studied as molecular markers for the early stages of NAFLD.No  significant mean difference in galnt2 and sirt1 expression among tertiles of LFS and FLI was found.This study was designed as a preliminary study, in which researchers were aware of the limitations of using a relatively small number of participants.However, this study provides insight for further exploration of potential markers for the early stages of insulin resistance and NAFLD in healthy individuals.

Conclusions
Further study is needed for galnt2 and sirt1 expression as potential molecular markers for the early stages of insulin resistance in a larger population.And although ptpn1 expression was not correlated with early insulin resistance, ptpn1 expression could be studied further as a molecular marker for early non-alcoholic fatty liver disease.
* p < 0.05; † p < 0.01; ‡ p < 0.001.The expression level analysis was performed using the expression ofactb as a reference to normalize gene expression level of target genes.