Abstract
Non-alcoholic fatty liver disease (NAFLD) is prevalent in the aging society. Despite body weight reduction, the prevalence of NAFLD has been increasing with aging for unknown reasons. Here, we investigate the association of DNA methylation age acceleration, a hallmark of aging, with risk of NAFLD. Genome-wide DNA methylation profiles were measured in 95 participants who developed type 2 diabetes during 4-year follow-up, and 356 randomly sampled participants from Shanghai Changfeng Study. DNA methylation age was calculated using the Horvath's method, and liver fat content (LFC) was measured using a quantitative ultrasound method. Subjects with highest tertile of DNA methylation age acceleration (≥ 9.5 years) had significantly higher LFC (7.2% vs 3.1%, P = 0.008) but lower body fat percentage (29.7% vs 33.0%, P = 0.032) than those with lowest tertile of DNA methylation age acceleration (< 4.0 years). After adjustment for age, sex, alcohol drinking, cigarette smoking, BMI, waist circumference, and different type blood cell counts, the risk of NAFLD was still significantly increased in the highest tertile group (OR, 4.55; 95% CI, 1.06-19.61). Even in subjects with similar LFC at baseline, DNA methylation age acceleration was associated with higher increase in LFC (4.0 ± 10.7% vs 0.9 ± 9.5%, P = 0.004) after a median of 4-year follow-up. Further analysis found that 6 CpGs of Horvath age predictors were associated with longitudinal changes in LFC after multivariate adjustment and located on genes that might lead to fat redistribution from peripheral adipose to liver. Combination of the key CpG methylation related to liver fat content with conventional risk factors improves the performance for NAFLD prediction.