Abstract P196: A Large Genetic Effect of APOC3 A43T SNP on Serum Triglycerides is Mediated by Serum APO-CIII Levels in Amerindians

Abstract

Introduction: Using a population-based linkage analysis approach, we previously identified a major locus for serum triglyceride (TG) levels in Pima Indians, which largely reflects the effect of a functional SNP (p.(Ala43Thr), A43T, or rs147210663) in the APOC3 gene. This non-synonymous SNP is extremely rare in non-Amerindian populations and more common in the Pima Indians (frequency of the T43 allele≈3%). This SNP alters the structure of APO-CIII, which leads to loss of its function and to lower circulating levels of APO-CIII. Its effect size reduces TG level by approximately 1 SD unit (per allele), which is among the largest effects for any complex trait reported to date. Hypothesis: To further explore the functional mechanism underlying this association, we analyzed the relationships among A43T, TG and APO-CIII levels; we hypothesized that the APOC3 A43T SNP effect on TG is mediated through circulating APO-CIII in the blood. Methods: We measured serum APO-CIII level in 411 randomly selected Amerindian subjects from a larger cohort study with both the A43T genotype and TG levels measured; this included 22 carriers of a T43 allele (frequency=2.7%). All analyses were modeled using linear regressions based on cross-sectional data, with age, sex, diabetes status, and population admixture included as covariates. The natural logarithms of APO-CIII and TG levels were taken to reduce skewness. Results: The phenotypic correlation between APO-CIII and TG was 0.68. We found a strong association between the A43T SNP and TG levels (β=-0.78±0.21 SD unit per copy of the T43 allele, p=0.00026, explaining 3.2% variance) that was greatly attenuated with further adjustment for APO-CIII levels (β=-0.04±0.17 SD unit, p=0.81, explaining 0.01% variance). On the other hand, there was also a strong association between the A43T SNP and APO-CIII levels (β=-1.22±0.21 SD unit, p=1.2х10 -7 , explaining 6.5% variance), which remained significant even after adjustment for the TG effect (β=-0.61±0.16 SD unit, p=0.00016, explaining 3.4% variance). Formal mediation analysis, conducted by the Sobel method, suggested significant and virtually complete mediation of the A43T effect on TG through APO-CIII levels (p=5.2х10 -7 , percent mediation=95%). Conclusion: Our results suggest that the APOC3 A43T SNP effect on TG may be primarily mediated through APO-CIII, and lend further strength to the possibility of making APO-CIII a therapeutic target for CVD interventions. We are in the process of expanding our analysis to a much larger sample.