Genetic variation in hyaluronan metabolism loci is associated with plasma plasminogen activator inhibitor-1 concentration

Matthew B. Lanktree, Christopher T. Johansen, Sonia S. Anand, A. Darlene Davis, Ruby Miller, Salim Yusuf, Robert A. Hegele


Elevated plasma plasminogen activator inhibitor-1 (PAI-1) concentration is associated with cardiovascular disease risk. PAI-1 is the primary inhibitor of fibrinolysis within both the circulation and the arterial wall, playing roles in both atherosclerosis and thrombosis. To define the heritable component, subjects within the population-based SHARE (Study of Health Assessment and Risk in Ethnic groups) and SHARE-AP (Study of Health Assessment and Risk Evaluation in Aboriginal Peoples) studies, composed of Canadians of South Asian (n = 298), Chinese (n = 284), European (n = 227), and Aboriginal (n = 284) descent, were genotyped using the gene-centric Illumina HumanCVD BeadChip. After imputation, more than 150 000 single nucleotide polymorphisms (SNPs) in more than 2000 loci were tested for association with plasma PAI-1 concentration. Marginal association was observed with the PAI-1 locus itself (SERPINE1; P < .05). However, 5 loci (HABP2, HSPA1A, HYAL1, MBTPS1, TARP) were associated with PAI-1 concentration at a P < 1 × 10−5 threshold. The protein products of 2 of these loci, hyaluronan binding protein 2 (HABP2) and hyaluronoglucosaminidase 1 (HYAL1), play key roles in hyaluronan metabolism, providing genetic evidence to link these pathways.


Activators and inhibitors of fibrinolysis are important regulators of atherosclerosis and arterial thrombosis, the key mechanistic pathways leading to myocardial infarction.1 Plasminogen activator inhibitor-1 (PAI-1) works as the primary inhibitor of fibrinolysis.1 Elevated PAI-1 concentration has been associated with acute myocardial infarction,2 though its contribution to atherosclerosis is complicated by associations with other cardiovascular risk factors.1 PAI-1 has pleiotropic effects, having been implicated in inflammation, extracellular matrix turnover, tissue repair, coagulation, angiogenesis, and cell adhesion and migration.1

Variation in plasma PAI-1 concentration contains a genetic component as indicated by heritability estimates in families.3 Complete PAI-1 deficiency has been described in a kindred carrying a rare frame-shift mutation in PAI-1 (alternative name SERPINE1).4 In unrelated members of the population, candidate gene studies of the PAI-1 locus have identified an associated 4G/5G promoter polymorphism, although reported estimates of its contribution to PAI-1 levels vary.5,6 A genome-wide association study of PAI-1 has yet to be reported. To uncover novel genetic associations with plasma PAI-1 concentration, we examined more than 2000 genes with a priori hypotheses for involvement in cardiovascular disease7 in the multiethnic Study of Health Assessment and Risk in Ethnic groups (SHARE) and Study of Health Assessment and Risk Evaluation in Aboriginal Peoples (SHARE-AP) populations.


Study participants

The SHARE and SHARE-AP studies are prospective population-based investigations of atherosclerotic risk factors across ethnic groups residing in Canada.8,9 Individuals were classified as South Asian (n = 298) if their ancestors originated from India, Pakistan, Sri Lanka, or Bangladesh; Chinese (n = 284) if their ancestors originated from China, Taiwan, or Hong Kong; European (n = 227) if their ancestors originated from Europe8; and Aboriginal (n = 284) if they were a Six Nations Band Member.8,9 SHARE was approved by the research ethics boards of McMaster University and the University of Western Ontario, and SHARE-AP was approved by a Six Nations Band Council Resolution and the McMaster University research ethics board. All participants provided informed written consent for DNA analysis in accordance with the Declaration of Helsinki.

PAI-1 measurement and SNP genotyping

Fasting blood samples were obtained from participants for the measurement of PAI-1 and DNA extraction. A 2-stage indirect enzymatic assay was performed using a Biopool Spectrolyse kit for measurement of PAI-1.10 As previously reported, PAI-1 measurements were log-transformed to obtain a normal distribution.6 Baseline characteristics of the SHARE and SHARE-AP cohorts are found in Table 1. All subjects were genotyped on the Illumina HumanCVD BeadChip, using the standard protocols of The Center for Applied Genomics (TCAG; Hospital for Sick Children, Toronto, ON; for SHARE and the Genome Quebec Innovation Center (McGill University, Montreal, QC; for SHARE-AP.7,11 Single nucleotide polymorphisms (SNPs) were selected to cover regions of interest at a density greater than afforded by genome-wide genotyping products.7,12 Each SNP was excluded if call rates were less than 95%, were not in Hardy-Weinberg equilibrium (P < .0001), or were found at low minor allele frequency (< .01), leaving 35 303, 31 751, 35 018, and 35 357 SNPs in South Asian, Chinese, European, and Aboriginal samples, respectively. In addition, the PAI-1 4G/5G promoter polymorphism (rs34857375) was genotyped in SHARE using a previously described protocol.13

View this table:
Table 1

Baseline characteristics of SHARE and SHARE-AP cohorts

Statistical methods

Identity-by-state and multidimensional scaling, as implemented in PLINK,14 indicate SHARE and SHARE-AP cluster appropriately with HapMap3 populations (see supplemental Figure 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). Imputation analysis was performed using MACH and SNP calls were retained with imputation quality of r2 more than .4, resulting in a total of 171 017 SNPs.15 The top 3 principal components of ancestry were calculated in each ethnicity independently using EIGENSTRAT.16 A codominant genetic model was selected, testing for an additive effect of each minor allele, with significance reported as the asymptotic probability of the t statistic in PLINK.14 All regression calculations were performed in each ethnicity independently, and after inclusion of clinical covariates including age, sex, body mass index, and current smoking status, as well as the top 3 principle components of ancestry. Percent of explained PAI-1 variation was calculated using a multivariate step-wise model selection procedure in SAS v9.1 (SAS Institute). Summary results of each ethnicity were combined using an inverse variance and sample size weighted meta-analysis as implemented in METAL ( For analysis of the HumanCVD BeadChip, a standard Bonferroni correction would yield a significance threshold of 1 × 10−6. We used a slightly relaxed significance threshold of 1 × 10−5 due to the higher prior odds of association due to the dense gene-centric nature of the array, as in a previous report.12

Results and discussion

A total of 7 SNPs in 5 loci were significantly associated with plasma PAI-1 concentration (Table 2, supplemental Figures 2-3). Rare variant rs11575741 in hyaluronan binding protein 2 (HABP2), present only in Europeans, displayed the strongest evidence for association (P = 3.3 × 10−7). Of note, rs11574741 is not found on, nor would be well imputed by, genome-wide genotyping arrays. Variants rs1283 and rs4855882 surrounding hyaluronoglucosaminidase 1 (HYAL1), an independent hyaluronan metabolism gene, were also associated (P = 4.5 × 10−6, P = 6.0 × 10−6, respectively). The second strongest association was with rs9469058, located in heat shock 70-kilodalton protein 1A (HSPA1A; P = 4.2 × 10−6). Two additional loci, membrane-bound transcription factor peptidase (MBTPS1) and T cell receptor gamma alternate reading frame protein (TARP) contained significantly associated SNPs (P = 5.2 × 10−6, P = 7.5 × 10−6, respectively). The 5G allele of the 4G/5G PAI-1 promoter polymorphism was marginally associated with lower PAI-1 concentrations (P = .051). Marginal association was also observed for a group of SNPs in perfect linkage disequilibrium within the 3′ UTR of PAI-1 (lead SNP rs11178: P = .042).

View this table:
Table 2

Genetic loci associated with plasma PAI-1 concentration

In line with previous knowledge of a very low density lipoprotein (VLDL) responsive element 5′ to PAI-1,17 plasma triglycerides accounted for the largest proportion of explained variation (supplemental Table 1). The HABP2 variant had the largest genetic effect explaining 10% of variation in plasma PAI-1 concentration in Europeans. The variation explained by MBTPS and TARP was approximately 1% in all ethnicities, while HSPA1A and HYAL1 ranged from 0.7% to 5%. Similar to previous reports the PAI-1 4G/5G explained 2.2% of PAI-1 variation in Europeans.5,6

Hyaluronan (synonyms: hyaluronic acid, hyaluronate) is a glucosaminoglycan found predominantly in the extracellular matrix of tissues including the vascular system.18 HABP2 was first identified through its affinity for hyaluronan.19 HYAL1 is an enzyme responsible for degrading the majority of high molecular weight hyaluronan into tetrasaccharide.20 Plasma hyaluronan concentrations are low, due to rapid receptor-facilitated uptake and catabolism by the liver.18 Hyaluronan is important for smooth muscle cell proliferation, locomotion and migration,21 thrombosis through inhibition of platelet aggregation, and evolving evidence suggests a pro-inflammatory effect.22 Providing evidence for direct functional link between hyaluronan and PAI-1, Marutsuka et al identified a hyaluronan dose-dependent increase in PAI-1 secretion from cultured vascular smooth muscle cells.23 A second paper described a positive correlation between concentrations of hyaluronan and PAI-1.24 In addition, HABP2 can form covalent complexes with PAI-1 directly.25

In conclusion, we report association between plasma PAI-1 concentration and genetic variation in 5 novel loci. It is unlikely that association between PAI-1 and genetic variation in 2 independent hyaluronan metabolism loci, found on different chromosomes, is the result of population stratification, linkage disequilibrium or confounding. Independent association of 2 hyaluronan metabolism genes provides genetic evidence to link hyaluronan to PAI-1 metabolism.


Contribution: S.S.A., S.Y., and R.A.H. designed the research, acquired study samples, and handled funding and supervision; A.D.D. and R.M. acquired study samples; M.B.L. and C.T.J. performed statistical analysis and analyzed and interpreted data; and M.B.L. drafted the manuscript. All authors provided critical revision of the manuscript for intellectual content.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Matthew B. Lanktree, Blackburn Cardiovascular Genetics Laboratory, Robarts Research Institute, University of Western Ontario, 4299-100 Perth Dr, London, ON, Canada N6A 5K8; e-mail: mlanktree2012{at}; or Robert A. Hegele, MD, FRCPC, FACP, Blackburn Cardiovascular Genetics Laboratory, Robarts Research Institute, University of Western Ontario, 4288A-100 Perth Dr, London, ON, Canada N6A 5K8; e-mail: hegele{at}


M.B.L. is supported by the Canadian Institutes of Health Research (CIHR) MD/PhD Studentship Award, the University of Western Ontario MD/PhD Program, and is a CIHR Fellow in Vascular Research. S.S.A. holds the Michael G. DeGroote/Heart and Stroke Foundation of Ontario endowed Chair in Population Health Research and the Eli Lilly–May Cohen Chair in Women's Health Research. S.Y. holds an endowed chair in Cardiovascular Research from the Heart and Stroke Foundation of Ontario. R.A.H. is a Career Investigator of the Heart and Stroke Foundation of Ontario, holds the Edith Schulich Vinet Canada Research Chair (Tier I) in Human Genetics, the Martha G. Blackburn Chair in Cardiovascular Research, and the Jacob J. Wolfe Distinguished Medical Research Chair at the University of Western Ontario. This work was supported by the Shared Hierarchical Academic Research Computing Network (SHARCNET), CIHR (MOP-13430, MOP-79523, CTP-79853), the Heart and Stroke Foundation of Ontario (NA-6059, T-6018, PRG-4854), Genome Canada through Ontario Genomics Institute, and the Pfizer Jean Davignon Distinguished Cardiovascular and Metabolic Research Award.


  • The online version of this article contains a data supplement.

  • The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

  • Submitted February 12, 2010.
  • Accepted June 3, 2010.


View Abstract