Association of APOC3 Loss-of-Function Mutations with Plasma Lipids and Subclinical Atherosclerosis in the Multi-ethnic BioImage Study

To The Editor

In 2008, Pollin and colleagues identified one mechanism of lowering triglyceride-rich lipoproteins among the Lancaster Amish, loss of apolipoprotein C-III (APOC3) gene function, which was associated with reduced subclinical atherosclerosis (i.e., coronary arterial calcification on cardiac computed tomography).(1) We recently extended these observations to show that APOC3 loss-of-function mutations also reduce risk for clinical atherosclerotic cardiovascular disease (ASCVD).(2,3) Here, we address two questions related to mutations that reduce APOC3 function: (1) do these mutations also associate with lower plasma low-density lipoprotein cholesterol (LDL-C)? and (2) do these mutations reduce subclinical atherosclerosis in the general population, particularly in individuals with ancestry outside of Europe?

We studied 6,699 individuals of European, African, Asian, and Hispanic ancestries from the BioImage Study (NCT00738725), a prospective, observational study aimed at characterizing subclinical atherosclerosis in US adults (55–80 years) at risk for clinical ASCVD.(4) We genotyped four APOC3 (NM_000040.1) loss-of-function mutations (IVS2+1G→A, A43T, R19X, IVS3+1G→T) using the Illumina HumanExome BeadChip Array v1.1. Written informed consent was obtained from all study participants according to a protocol approved by the Western Institutional Review Board, Olympia, WA. Fasting blood lipids were measured at the baseline examination. Blood lipids were adjusted for the presence of statin medications to reflect the observation that statins, on average, reduce total cholesterol by 20% and LDL-C by 30%.(5) Non-invasive assessments for subclinical atherosclerosis (coronary arterial calcification (CAC), carotid plaque, and carotid intima media thickness (CIMT)) were performed at the baseline examination on a mobile imaging facility as previously described.(4)

6,395 subjects passed all quality-control measures. Variant calling was performed using GenCall (Illumina, San Diego, CA) and zCall.(5) 64 heterozygous carriers of APOC3 loss-of-function mutations were identified (25 IVS2+1G→A, 25 A43T, 13 R19X, 1 IVS3+1G→T; combined minor allele frequency of 0.5%) were identified. Principal components were derived from a set of high quality, independent variants on the genotyping array using Eigenstrat as has previously been done.(3,5) To minimize confounding from systematic differences in allele frequencies by trait, we reduced the observed genetic variation to the top eigenvectors derived from the sample covariance matrix. To test the association of APOC3 loss-of-function mutation with an outcome, linear regression was used for triglycerides, LDL-C, and high-density lipoprotein cholesterol (HDL-C), and CIMT; triglycerides and CIMT were natural log-transformed. And given the bimodal, skewed distributions of CAC (primary outcome) and carotid plaque, median quantile regression was used for these two variables. Age, sex, ethnicity, and principal components of ancestry were used as covariates in all analyses. Given a two-sided alpha threshold of 0.05, we have >80% power to detect an effect size of 0.16% of variance explained for analyzed traits.

Among carriers and non-carriers of the APOC3 loss-of-function mutations, there were no significant differences in age, sex, hypertension, diabetes mellitus, body-mass index, current smoking, aspirin use, or statin use. There were no significant differences in proportions of carriers amongst each ethnicity group (p-values > 0.20). We replicated the finding that APOC3 loss-of-function mutations were associated with reduced triglycerides (−43.7 %; p-value 1.83 × 10⁻²¹) and increased HDL-C (11.1 mg/dL; p-value 3.55 × 10⁻¹⁰), with a larger standardized effect on triglycerides compared to HDL-C (−1.17 standard deviations versus +0.73 standard deviations). When accounting for statin treatment, carriers did not have different LDL-C concentrations compared to non-carriers (p-value: 0.75).

APOC3 loss-of-function mutations were associated with decreased median CAC score (−27.9 units; 95% CI −51.08, −4.67; p-value 0.019) across all phenotyped participants (n = 5,631); this effect was consistent in those of European ancestry (−27.5 units; 95% CI −67.1, 12.1) and of non-European ancestry (−5.62 units; 95% CI −39.2, 27.9). Neither carotid plaque (p-value 0.79) nor CIMT (p-value 0.47) (n = 5,746) differed between carriers and non-carriers (Table 1).

Table 1.

Association of APOC3 Loss-of-Function Mutation Carrier Status with Blood Lipid Levels and Subclinical Atherosclerosis

	Non-Carriers (N=6,331)	Carriers (N=64)	Effect Estimate	95% CI	P-value
Blood Lipids
Triglycerides (mg/dL)*	166.1 ± 96.2	91.2 ± 44.1	−43.7%	−36.6, −49.9	1.83 × 10−21
HDL Cholesterol (mg/dL)	55.6 ± 15.2	66.6 ± 15.0	+11.1	7.6, 14.6	3.55 × 10−10
LDL Cholesterol (mg/dL)	130.3 ± 36.4	131.6 ± 35.8	+1.5	−7.4, 10.3	0.75
Subclinical Atherosclerosis
Coronary Arterial Calcification (Agatston Units)†	46.0 (0.0, 245.0)	29.0 (0.0, 227.5)	−27.9	−51.1, −4.7	0.019
Carotid Plaque (mm2)†	183.8 (0.0, 555.9)	112.8 (0.0, 367.2)	−8.7	−73.5, 56.0	0.79
Carotid Intima Media Thickness (mm)*	0.76 ± 0.16	0.74 ± 0.13	−1.7%	−6.3, 3.0	0.47

HDL = high-density lipoprotein; LDL = low-density lipoprotein; CI = confidence interval

Values are represented as effect estimates (95% confidence intervals) between APOC3 loss-of-function mutation carriers and non-carriers after adjustment for covariates. Covariates used were age, sex, ancestry, and principal components of ancestry.

^*Effect estimates for triglycerides and carotid intima media thickness were derived from natural log transformation and are expressed as percent difference in carriers compared with non-carriers.

^†Effect estimates for coronary arterial calcification and carotid plaque derived from median quantile regression and represent differences in medians. Therefore, summary statistics of the distributions in non-carriers and carriers are represented as median (interquartile range).

In a multi-ethnic study of US adults, APOC3 loss-of-function mutation carriers had reduced plasma triglycerides, higher HDL-C, and a decreased burden of coronary arterial calcification. These data support the notion that APOC3 deficiency reduces coronary atherosclerosis in the general population. Whether pharmacologic inhibition of APOC3 will reduce ASCVD risk remains to be tested.