Rare Protein-Truncating Variants in APOB, Lower Low-Density Lipoprotein Cholesterol, and Protection Against Coronary Heart Disease

Supplemental Digital Content is available in the text.


Study participants
We recruited twenty-nine FHBL pedigrees to study associations of APOB PTV carrier status with lipid profiles from the Kanazawa University Mendelian Disease Registry in Japan.
Participants were diagnosed with FHBL if low serum LDL-C or apoB levels were observed (LDL-C < 70 mg/dL or apoB < 50 mg/dL). We used these thresholds of LDL-C and apoB for FHBL diagnosis in order not to miss detecting putative FHBL individuals harboring APOB PTVs 1 . Causative variants were identified by whole exome sequencing and APOB PTVs cosegregating with phenotype within each pedigree were identified. Identified causative variants were then confirmed through Sanger sequencing (primers shown in Supplemental Table 1).
Additionally, we sequenced the APOB gene in a total of 57,973 participants from the Myocardial Infarction Genetics Consortium (MIGen) of African, European, and South Asian ancestries (N=33,835), and from participants of European ancestry (N=24,138) in the Geisinger Health System and Regeneron Genetics Center DiscovEHR study who were recruited as part of the MyCode Community Health Initiative 2 ( Table 1). MIGen studies included the Italian Atherosclerosis Thrombosis and Vascular Biology (ATVB) study 3 , Bangladesh Risk of Acute Vascular Events study (BRAVE) 4 , the Exome Sequencing Project Early-Onset Myocardial Infarction (ESP-EOMI) study 5 , a nested case-control cohort of the Jackson Heart Study (JHS) 6 , the South German Myocardial Infarction study 7 , the Ottawa Heart Study (OHS) 8 , the Precocious Coronary Artery Disease Study (PROCARDIS) 9 , the Pakistan Risk of Myocardial Infarction Study (PROMIS) 10 , the Registre Gironi del COR (Gerona Heart Registry or REGICOR) study 11 , the Leicester Myocardial Infarction study 12 , and the North German Myocardial Infarction study 13 (Supplemental Table 5). Clinical data were assessed in each study.
All participants in the study provided written informed consent for genetic studies. The institutional review boards at the Broad Institute and each participating institution approved the study protocol.
In order to minimize the possibility of unintentionally sharing information that can be used to re-identify private information, a subset of the data generated for this study are available at dbGaP and can be accessed at through dbGaP Study Accessions: phs000814.v1.p1 (ATVB),

Phenotypes
In FHBL pedigrees, all blood samples were obtained after a 12-hour overnight fast. Apolipoprotein B was analyzed by a commercial immunoturbidimetric assay (Apo B auto N Daiichi, Sekisui Medical, Tokyo, Japan) 14 . Fatty liver was diagnosed by an independent liver ultrasound specialist when a participant was observed to have both liver brightness and hepatorenal echo contrast.
In MIGen, fasting LDL-C in mg/dL was used from the earliest available exam in each contributing study. LDL-C was calculated using the Friedewald equation 15,16 (LDL-C = total cholesterolhigh-density lipoprotein cholesterol [HDL-C] -(triglycerides/5)) for those with triglycerides <400 mg/dL. If triglycerides ≥400 mg/dL, calculated LDL-C was set to missing. In the DiscovEHR study, median lipid levels were calculated for each individual following removal of values that were > 3 standard deviations from the intra-individual median value for individuals with two or more measurements in the EHR 2 . In both MIGen and DiscovEHR, for those on lipid-lowering drug treatment, we replaced LDL-C by a value of the measured LDL-C divided by 0.7 and total cholesterol by a value of the total cholesterol divided by 0.8 17,18 . HDL-C and triglyceride levels were not adjusted for lipid-altering medication use, and triglyceride levels were natural logarithm transformed for statistical analysis.
In MIGen, early-onset CHD was defined as myocardial infarction, angiographic coronary artery disease, coronary artery bypass surgery, or percutaneous coronary revascularization in men ≤50 years or women ≤60 years. Details for clinical phenotypes for each cohort are available in Supplemental Table 5. In DiscovEHR, the present analysis was restricted to early-onset CHD cases and CHD-free controls (age <55 years for men or <65 years for women for both cases and controls). Participants were considered to have CHD if they had a history of coronary revascularization in the EHR, or history of acute coronary syndrome, ischemic heart disease, or exertional angina (ICD-9 codes 410*, 411*, 412*, 413*, 414*) with angiographic evidence of obstructive coronary atherosclerosis (>50% stenosis in at least one major epicardial vessel from catheterization report). CHD-free controls were defined as individuals without any case criteria or any single encounter or problem list diagnosis code indicating CHD.

Gene sequencing
Whole exome sequencing of MIGen was performed at the Broad Institute (Cambridge, MA, USA) as previously described 5 . Sequencing reads were aligned to a human reference genome (build 37) using the Burrows-Wheeler Aligner-Maximal Exact Match algorithm. Aligned nonduplicate reads were locally realigned, and base qualities were recalibrated using the Genome Analysis ToolKit (GATK) software 19 . Variants were jointly called using the GATK HaplotypeCaller program. The sensitivity of variant quality score recalibration (VQSR) threshold was 99.6% for single nucleotide variants and 95% for insertion/deletion variants, as we have previously reported 20 . We annotated all identified variants with the use of the Variant Effect Predictor software (version 88) 21,22 . PTV were defined as high confidence (https://github.com/konradjk/loftee) nonsense, splice-site, and frameshift mutations with minor allele frequency < 1% across contributing cohorts.
In the DiscovEHR study, APOB sequences were extracted from whole exome sequences generated as previously described 2 . Sequence reads were aligned to the human reference build GRCh37.p13. Single nucleotide variants (SNV) and insertion/deletion (indel) sequence variants were identified using the Genome Analysis Toolkit 23 and annotated using SnpEff 24 . PTVs were defined as any of the following: SNVs leading to a premature stop codon, loss of a start codon, or loss of a stop codon; SNVs or indels disrupting canonical splice acceptor or donor dinucleotides; open reading frame shifting indels leading to the formation of a premature stop codon.

Statistical Analysis
In FHBL pedigrees, the differences in cholesterol and hepatobiliary enzymes stratified by APOB PTV carrier status were analyzed using the Mann-Whitney U test while effect sizes were obtained from a linear regression associating carrying an APOB PTV on cholesterol adjusted for age and sex.
We performed linear regression with controls to associate APOB PTV carrier status with each blood lipid level in the MIGen studies and with LDL-C in the DiscovEHR study adjusting for age, sex, the first 5 principle components of ancestry, and indicators of cohort status.
In order to associate an aggregate of PTVs in the APOB gene with CHD risk, we performed an exact Cochran-Mantel-Haenszel analysis for stratified 2-by-2 tables 25 implemented in the meta R package. Heterogeneity was measured by the I 2 statistic, calculated in the meta R package. It describes the percentage of variation in association statistics across studies that is due to heterogeneity of the statistics rather than due to chance. We obtained pvalues for proportion of null allele counts in cases versus controls and odds ratios (OR) with 95% confidence intervals (CI). We considered a p-value less than 0.05 as statistically significant.
Additionally, since the Cochran-Mantel-Haenszel test does not allow adjustment for covariates, we performed a sensitivity analysis in the MIGen study using logistic regression adjusting for cohort, sex, and the first 4 principal components (PCs) of ancestry to control for potential population stratification 26,27 . The first 4 PCs of ancestry had p-values < 0.01 for association with CHD after adjusting for the cohort and sex. Abbreviations: CAD, coronary artery disease; CHD, coronary heart disease; CABG, Coronary artery bypass grafting; CVD, cardiovascular disease; EOCAD, early-onset coronary artery disease; EOMI, early-onset myocardial infarction; MI, myocardial infarction.