Genetic variants in PLG, LPA, and SIGLEC 14 as well as smoking contribute to plasma plasminogen levels

Abstract

Plasminogen is the precursor of the serine protease plasmin, a central enzyme of the fibrinolytic system. Plasma levels of plasminogen vary by almost 2-fold among healthy individuals, yet little is known about its heritability or genetic determinants in the general population. In order to identify genetic factors affecting the natural variation of plasminogen levels, we performed a genome-wide association study and linkage analysis in a sample of 3456 young healthy individuals who participated in the Genes and Blood Clotting Study (GABC) or the Trinity Student Study (TSS). Heritability of plasminogen levels was 48.1% to 60.0%. Tobacco smoking and female sex were associated with higher levels of plasminogen. In the meta-analysis, 11 single-nucleotide polymorphisms (SNPs) in 2 regions reached genome-wide significance (P < 5.0E-8). Of these, 9 SNPs were near the PLG or LPA genes on Chr6q26, whereas 2 were on Chr19q13 and 5' upstream of SIGLEC14. These 11 SNPs represented 4 independent signals and collectively explained 6.8% of plasminogen level variation in the study populations. The strongest association was observed for a nonsynonymous SNP in the PLG gene (R523W). Individuals bearing an additional copy of this allele had an average decrease of 13.4% in plasma plasminogen level.

Figures

Figure 1

Meta-analysis result. (A) Genome-wide plot of −log10(P) for ∼742 000 SNPs. The red line marks the 5.0E-8 threshold of genome-wide significance. (B) Quantile-quantile plot of observed vs expected −log10(P) for PLG meta-analysis. The observed P < 5.0E-8 values are shown in red. (C) Regional plot for the associated region near PLG, LPA, SLC22A3, and SLC22A2 on Chr6. (D) Regional plot for the associated region near SIGLEC14 and SIGLEC12 on Chr19.

Figure 2

Independence of association signals. PLG levels of different SNP genotypes were analyzed in the combined TSS and GABC data, which were merged after adjusting for the significant environmental factors and making sure that they already have nearly the same distributions (Kolmogorov-Smirnov test, P = .07; Mann-Whitney U test, P = .53). (A) Boxplot of the distribution of PLG levels in various genotype combinations of the top 2 independent SNPs, rs4252129 and rs1084651, showing additive effects of rs1084651 in every rs4252129 genotype. The inserted table shows the counts of nonmissing genotypes for rs4252129 (row) and rs1084651 (column). One-way analysis of variance test by coding the available genotype combinations ordered by allelic effect directions as integers 1 to 6 revealed P = 3.5E-38. (B-D) Distribution of PLG levels in various genotype combinations of the top 3 independent SNPs on Chr6 and Chr19, showing additive effects of rs41272114 (B), rs783176 (C), and rs1041297 (D) in most genotype combinations of rs4252129 and rs1084651. One-way analysis of variance tests by coding the available genotype combinations as integers revealed P values of 8.0E-40 (B), 2.8E-42 (C), and 6.6E-43 (D), respectively.

Figure 3

Gene–environment (smoking) interaction in the TSS cohort. (A) Boxplot of log-transformed, outlier-removed, and age- and sex-adjusted PLG levels against smoking. (B) Boxplot of PLG levels against smoking status and genotypes of the top SNP, rs4252129. (C) Boxplot of PLG levels against smoking status and the genotype combinations of the top 2 independent SNPs, rs4252129 and rs1084651.

Figure 4

Linkage analysis in TSS and GABC sibs (n = 557 sibships) using 35 356 LD clusters. The mapping position of each LD cluster is marked by a vertical tick.