Non-alcoholic fatty liver disease : the role of insulin resistance, inflammation and the PNPLA3 I148M variant

Non-alcoholic fatty liver disease (NAFLD) may result from obesity accompanied by insulin resistance and adipose tissue inflammation or from the common genetic variants in PNPLA3 (rs738409, C>G/I148M), TM6SF2 (rs58542926, C>T/E167K) and MBOAT7 (rs641738, C>T). These variants increase the liver fat content and the severity of NAFLD without features of insulin resistance. This thesis aimed to determine the following: i) whether NAFLD predicts type 2 diabetes independent of obesity and other known risk factors; ii) whether adipose tissue is inflamed in subjects homozygous for the PNPLA3 I148M variant; iii) whether obesity and insulin resistance rather than liver fat content increase coagulation factor activities and expression; and iv) which factors predict NAFLD and liver stiffness during an 11-year follow-up period. The present thesis includes one systematic review, two cross-sectional studies and one longitudinal study. Study subjects comprised Finnish adult men and women. Liver fat content was measured by proton magnetic resonance spectroscopy (studies II–IV). The gene expression of inflammatory markers in adipose tissue and that of coagulation factors in the liver were measured using qPCR (studies II and III). We conducted a systematic review of prospective longitudinal studies to determine if NAFLD predicts type 2 diabetes for aim i). Based on these studies ultrasound-diagnosed NAFLD and liver enzymes predict type 2 diabetes independent of confounders such as age and obesity. We found no studies indicating that NAFLD associated with genetic risk variants predicting future risk of type 2 diabetes. A group of 82 subjects were divided into two groups based on body mass index and PNPLA3 genotype for aim ii). The liver fat content was similarly increased in obese/insulin-resistant subjects and in carriers of the I148M variant compared to non-obese subjects and non-carriers of this variant. In obese subjects, the adipose tissue expression of pro-inflammatory chemokine MCP-1 was increased and anti-inflammatory ADIPOQ and TWIST1 were decreased compared with non-obese subjects, while these were comparable between carriers and non-carriers of the I148M variant. In study III, 92 non-diabetic subjects were divided into two groups based on insulin sensitivity (HOMA-IR) and PNPLA3 genotype. Coagulation factor activities (FVIII, FIX, FXIII, fibrinogen and VWF:RCo) were increased, and the prothrombin time and activated partial thromboplastin time were shortened in insulin-resistant subjects when compared to insulin-sensitive subjects; yet, these factors were similar in carriers and non-carriers of the I148M variant. The hepatic gene expression of FVIII, FIX and fibrinogen gamma-chain were higher in insulin-resistant subjects with NAFLD (n=13) compared with equally obese insulin-sensitive subjects without NAFLD (n=13). In study IV, 97 subjects were examined twice over an interval of 11 years. The baseline liver fat content independently predicted NAFLD and an increased liver stiffness (measured using transient elastography) at 11.3 years more accurately than routinely available clinical and biochemical parameters. Conclusions: Obesity, adipose tissue inflammation and insulin resistance rather than excess hepatic fat per se are related to an increased risk of type 2 diabetes and a pro-coagulant plasma profile observed often in NAFLD. However, the liver fat content emerges as more important predictor of advanced liver fibrosis than the associated metabolic abnormalities. These data support the view that NAFLD is heterogeneous disease.