Targeted sequencing of linkage region in Dominican families implicates PRIMA1 and the SPATA7-PTPN21-ZC3H14-EML5-TTC8 locus in carotid-intima media thickness and atherosclerotic events

Abstract

Carotid intima-media thickness (cIMT) is a subclinical marker for atherosclerosis. Previously, we reported a quantitative trait locus (QTL) for total cIMT on chromosome 14q and identified PRiMA1, FOXN3 and CCDC88C as candidate genes using a common variants (CVs)-based approach. Herein, we further evaluated the genetic contribution of the QTL to cIMT by resequencing. We sequenced all exons within the QTL and genomic regions of PRiMA1, FOXN3 and CCDC88C in Dominican families with evidence for linkage to the QTL. Unrelated Dominicans from the Northern Manhattan Study (NOMAS) were used for validation. Single-variant-based and gene-based analyses were performed for CVs and rare variants (RVs). The strongest evidence for association with CVs was found in PRiMA1 (p = 8.2 × 10^-5 in families, p = 0.01 in NOMAS at rs12587586), and in the five-gene cluster SPATA7-PTPN21-ZC3H14-EML5-TTC8 locus (p = 1.3 × 10^-4 in families, p = 0.01 in NOMAS at rs2274736). No evidence for association with RVs was found in PRiMA1. The top marker from previous study in PRiMA1 (rs7152362) was associated with fewer atherosclerotic events (OR = 0.67; p = 0.02 in NOMAS) and smaller cIMT (β = -0.58, p = 2.8 × 10^-4 in Family). Within the five-gene cluster, evidence for association was found for exonic RVs (p = 0.02 in families, p = 0.28 in NOMAS), which was enriched among RVs with higher functional potentials (p = 0.05 in NOMAS for RVs in the top functional tertile). In summary, targeted resequencing provided validation and novel insights into the genetic architecture of cIMT, suggesting stronger effects for RVs with higher functional potentials. Furthermore, our data support the clinical relevance of CVs associated with subclinical atherosclerosis.

Figure 1

Single Variant Analysis of Common Variants in Families and NOMAS. Each circle represents the p-values of a single variant analysis of common variants (CV, MAF ≥ 0.05) in Families (X-axis) and NOMAS (Y-axis) datasets. The two solid lines depict p = 0.05 in each sample set, respectively. CVs that are significant in both datasets are located in the upper right section of the plot. The strongest association was found with rs12587586 (p = 8.2 × 10⁻⁵ in Families, p = 0.01 in NOMAS) in PRiMA1. The second strongest association was found with a missense SNP rs2274736 (p = 1.3 × 10⁻⁴ in Families and p = 0.01 in NOMAS) in PTPN21.

Figure 2

Regional Plots of PRiMA1 and PTPN21. The X-axis displays the base pair location of each CV on chromosome 14; Y axis displays the –Log₁₀ (p-values) of QTDT for association with total cIMT in the Families. The index CV is displayed as a purple diamond. The color of the circles indicates LD estimated in the families between each CV and the index CV. The solid purple line displays the estimated recombination rate. In PRiMA1 (panel A.), the index CV is rs12587586, the top marker identified in the current study. The most significant marker in the previous study (rs7152362) is not in LD (r² = 0.02; D′ = 0.31) with rs12587586. In PTPN21, the index CV is rs2274736, the most significant CV in the region. It is located within a region with extended LD containing five genes, i.e. SPATA7, PTPN21, ZC3H14, EML5, and TTC8.

Figure 3

Gene-based Analysis of Exonic RVs in Families and NOMAS. Each vertical bar represents the p-values of gene-based test of exonic RVs in Families (X-axis) and NOMAS (Y-axis) datasets. All genes within the 5-gene cluster, except SPATA7, display evidence for association in at least one dataset. There is no evidence suggesting that exonic RVs in PRiMA1 are collectively associated with total cIMT (p = 0.7 in both datasets).