Genetic Variants Associated With Obesity and Insulin Resistance in Hispanic Boys With Nonalcoholic Fatty Liver Disease

Abstract

Background and objectives: Nonalcoholic fatty liver disease (NAFLD) disproportionately affects Hispanic boys. Further, obesity and insulin resistance are major risk factors for NAFLD. No gene localization studies had been performed on children with biopsy-proven NAFLD. This study aims to identify genomic variants associated with increased adiposity and insulin resistance in a population of children with varying histologic severity of NAFLD.

Methods: We conducted a genome-wide association scan (GWAS) including 624,297 single-nucleotide polymorphisms (SNPs) distributed among all 22 autosomal chromosomes in 234 Hispanic boys (up to 18 years of age) who were consecutively recruited in a prospective cohort study in the Nonalcoholic Steatohepatitis Clinical Research Network Studies. Traits were examined quantitatively using linear regression. SNPs with P value <10 and a minor allele frequency >5% were considered potentially significant.

Results: Evaluated subjects had a median age of 12.0 years, body mass index (BMI) of 31.4, and hemoglobin A1C (Hgb A1C) of 5.3. The prevalence of NAFL, borderline NASH, and definite NASH were 23%, 53%, and 22%, respectively. The GWAS identified 10 SNPs that were associated with BMI z score, 6 within chromosome 2, and 1 within CAMK1D, which has a potential role in liver gluconeogenesis. In addition, the GWAS identified 9 novel variants associated with insulin resistance: HOMA-IR (6) and HbA1c (3).

Conclusions: This study of Hispanic boys with biopsy-proven NAFLD with increased risk for the metabolic syndrome revealed novel genetic variants that are associated with obesity and insulin resistance.

Trial registration: ClinicalTrials.gov NCT01061684.

Conflict of interest statement

Conflicts of Interest:

John C Rausch, Joel E Lavine, Xiuqing Guo, Soonil Kwon, Jeffrey Schwimmer, Jean Molleston, Rohit Loomba, Elizabeth M Brunt, Yii-Der Ida Chen, Mark O Goodarzi, Kent D Taylor, Katherine P Yates, Aynur Unalp-Arida, and Jerome I Rotter: none

Naga P Chalasani: Dr. Chalasani has consulting agreements and research grants with pharmaceutical industry but none of them represent a conflict for this particular paper.

Figures

Figure 1. Manhattan Plots

Manhattan plots of BMI z-score, HOMA-IR, and hemoglobin A1C using linear regression genome-wide association studies. (A) represents BMI z-score using a dominant model, (B) represents BMI z-score using an additive model, (C) represents HOMA-IR using an additive model, and (D) represents hemoglobin A1C using an additive model. The y-axis is the –log10(p-value). The x-axis is the position on the 22 autosomal chromosomes. The different colored circles each represent individual single nucleotide polymorphisms (SNPs). The dotted line indicates significance at p-value of <10−6.

Figure 1. Manhattan Plots

Manhattan plots of BMI z-score, HOMA-IR, and hemoglobin A1C using linear regression genome-wide association studies. (A) represents BMI z-score using a dominant model, (B) represents BMI z-score using an additive model, (C) represents HOMA-IR using an additive model, and (D) represents hemoglobin A1C using an additive model. The y-axis is the –log10(p-value). The x-axis is the position on the 22 autosomal chromosomes. The different colored circles each represent individual single nucleotide polymorphisms (SNPs). The dotted line indicates significance at p-value of <10−6.

Figure 2. Quantile-Quantile Plots

Quantile–Quantile (QQ) plots for BMI z-score, HOMA-IR, and hemoglobin A1C in genome-wide association studies. (A) represents BMI z-score using a dominant model, (B) represents BMI z-score using an additive model, (C) represents HOMA-IR using an additive model, and (D) represents hemoglobin A1C using an additive model.

Figure 2. Quantile-Quantile Plots

Quantile–Quantile (QQ) plots for BMI z-score, HOMA-IR, and hemoglobin A1C in genome-wide association studies. (A) represents BMI z-score using a dominant model, (B) represents BMI z-score using an additive model, (C) represents HOMA-IR using an additive model, and (D) represents hemoglobin A1C using an additive model.