KNG1 Ile581Thr and susceptibility to venous thrombosis

Activated partial thromboplastin time (aPTT) is a global coagulation test that has been used during the last 50 years as a standard screening test in clinical laboratories throughout the world.¹  aPTT levels are considered to reflect global coagulation activity. This test cumulatively explores factors belonging to the classic intrinsic (FXI, FIX, and FVIII) or common (FII and fibrinogen) coagulation cascade. Besides its sensitivity toward variation levels of these coagulation factors, aPTT also is associated with age, female sex, estrogen use, and obesity.²  Shortened aPTT levels have been proved reliable as a predictor of venous thrombosis (VT).^2,3  Very recently, the authors of a genome-wide association study (GWAS) reported 3 single nucleotide polymorphisms (SNPs) associated with aPTT levels.⁴  These 3 SNPs were located in 3 coagulation cascade genes, F12 (rs27431672), HRG (rs9898), and KNG1 (rs710446). We reasoned that alleles associated with shortened aPTT levels might be associated with an increased risk of VT. To test this hypothesis, the genotype distributions of rs27431672, rs9898, and rs710446 observed in 2 independent samples of VT patients from the MARseille THrombosis Association study (MARTHA) 08 (n = 972) and MARTHA10 (n = 570) were compared with those observed in 1110 healthy subjects from the Three-City Study.⁵ 

MARTHA patients are unrelated Europeans consecutively recruited at the Thrombophilia center of La Timone Hospital (Marseille, France) between January 1994 and October 2005 (supplemental Table 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). All participants provided written informed consent, and the protocol was approved by the ethics committee of each participating institution.

As part of an ongoing GWAS on VT risk, MARTHA08 patients were typed in 2008 with the Illumina Human610-Quad Beadchip, whereas the Illumina Human660W-Quad Beadchip was used in early 2010 for typing MARTHA10 patients. Control patients from the Three-City Study also were typed with the Illumina Human610-Quad Beadchip. The 3 tested SNPs were available on the GWAS arrays, and none of them showed significant deviation from Hardy-Weinberg equilibrium at P < 10⁻². The genotype success rates were greater than 99.8% for all SNPs, except for rs9898 in MARTHA10 patients, where it only reached 98.1%. Replication of the results observed in MARTHA was investigated in the Facteurs de Risque et de Récidives de la Maladie Thromboembolique Veineuse (FARIVE) study,⁶  where genotyping was performed with the use of TaqMan technology.

Association between SNPs and disease status was assessed by use of the Cochran-Armitage trend test and logistic regression, whereas linear regression analysis was used to test the association with aPTT levels. Haplotype association analyses were performed by the use of THESIAS⁶  and Gridhaplo software.⁷ 

As indicated in Table 1, no association with VT risk was observed for the F12 rs2731672 or the HRG rs9898. Conversely, the KNG1 rs710446-C allele was found more frequently in MARTHA08 and MARTHA10 patients (0.455 and 0.451, respectively) compared with healthy patients of the Three-City Study (0.409). When the 2 sets of VT patients were combined and compared with the Three-City Study, the rs710446-C allele was associated with an increased odds ratio (OR) for VT of 1.196 (95% confidence interval [CI] 1.071-1.336, P = .0012) Adjustment for sex, ABO blood group (tagged by the ABO rs8176746, rs8176704, and rs505922),⁸  and FV Leiden (tagged by rs2420371) with the use of logistic regression analysis did not modify this association because the corresponding allelic OR became 1.204 (95% CI 1.071-1.354, P = .0019).

To further replicate the association of KNG1 rs710446 with VT risk, we investigated its effect in a sample of 590 healthy patients and 596 VT patients who were part of FARIVE.⁸  Because of technical problem of genotyping this SNP with a Taqman assay, we instead genotyped the rs698078 (genotype success rate of 99%), which serves as a perfect proxy (r² = 1.00 according to the SNAP database⁹ ). As indicated in Table 2, the genotype distribution of the rs698078 paralleled that observed for rs710446 in Table 1. The rs698078-G allele that corresponds to the rs710446-C allele was found more frequently in VT patients compared with control patients (0.456 vs 0.422, P = .113). After adjusting for age, sex, ABO blood group, and FV Leiden mutation, we found that the OR for VT associated with the rs698078-G allele was 1.171 (95% CI 0.889-1.541, P = .059). The association was borderline, and this finding is likely because of the modest power (estimated to 37% by use of the CatS software¹⁰ ) the FARIVE study has to detect such an association.

Among the 3 newly discovered SNPs found genome-widely associated with aPTT in the work by Houlihan et al,⁴  in our study we found KNG1 rs710446 to be associated with VT but not F12 rs2731672 or HRG rs9898. We cannot exclude that the latter 2 could be associated with VT risk, but the MARTHA project is likely underpowered to detect their effects, if any. According to the strength of associations observed in Table 1, the power to detect at P = .05 the effects of rs2731672 and rs9898 was less than 35% for both. Other SNPs mapping the HRG and F12 genes were available as part of our ongoing GWAS on VT, but none of them showed evidence for association with the disease (supplemental Table 2). Conversely, other KNG1 SNPs demonstrated evidence of association with VT, but their association was because of their linkage disequilibrium with rs710446 (supplemental Table 3).

Altogether, these results provide strong evidence that the KNG1 rs710446, a nonsynonymous variant (Ile581Thr) predicted to be damaging,⁴  is associated with the risk of VT. Several arguments are in favor of the implication of KNG1 locus, which encodes high molecular weight fibrinogen (HK) in VT physiopathology. HK plays an important role in blood coagulation by positioning prekallikrein and FXI near factor XII.¹¹  In addition, KNG1 knockout mice demonstrated prolonged aPTT and delayed arterial thrombosis.¹²  Moreover, antibody against mouse FXI, by directly interfering with the FXI-HK interaction, prevented arterial occlusion induced by FeCl3 to a similar degree to total FXI deficiency.¹³  HK is also an important member of the plasma kallikrein-kinin system,¹⁴  the activation of which may contribute to the manifestations of disorders such as hereditary angioedema,¹⁵  sepsis,¹⁶  ulcerative colitis,¹⁷  and Alzheimer disease.¹⁸ 

It would be tempting to assert that the effect of rs710446 on VT risk is mediated by its effect on aPTT levels recently discovered.⁴  We also observed a strong association between rs710446 and plasma aPTT levels in MARTHA08 (R² = 5.96%, P = 8.74 × 10⁻¹²) and MARTHA10 (R² = 9.22%, P = 9.01 × 10⁻¹¹), with the rs710446-C allele associated with decreased aPTT levels (supplemental Table 4). However, because aPTT measurements were not available in the Three-City Study, it was not possible to really test this hypothesis.

Other KNG1 SNPs also were found to influence aPTT levels in MARTHA (supplemental Table 2) but, again, the strongest association was observed for rs710446. We were also able to confirm the results of Houlihan et al⁴  by noting that, at the F12 and HRG loci, the strongest associations were observed for rs2731672 and rs9898, respectively (supplemental Table 2). To a lesser extent, F11 also was suggested to modulate aPTT levels,⁴  and we also observed this phenomenon in our samples (supplemental Table 2). Surprisingly, some, but not all, aPTT-associated F11 SNPs showed some promising evidence of association with VT, and the converse also was observed (supplemental Table 2). The association of F11 SNPs with VT risk is not new because 2 F11 SNPs, rs2036914 and rs2289252, were previously found to be associated with VT in the Leiden Thrombophilia study.¹⁹  None of these 2 SNPs was available in our GWAS dataset, nor were they in strong linkage disequilibrium with any of the F11 SNPs associated with VT in the present work (supplemental Table 5). In-depth haplotype analysis would be further required to disentangle the exact contribution of F11 SNPs to VT risk and aPTT variability, but this research is out of the scope of the present report. In conclusion, in this report we identify the KNG1 Ile581Thr variant as a new candidate risk factor for VT and encourage further study of the genetic determinants of pertinent phenotypic intermediate traits for identifying new risk factor of VT.

We thank Marie Billerey and Gwenaelle Burgos for their excellent technical assistance.

The FARIVE study was supported by grants from the Fondation pour la Recherche Médicale, the Program Hospitalier de recherche Clinique (PHRC 20 002; PHRC2009 RENOVA-TV), the Fondation de France, and the Leducq Foundation. The MARTHA project was supported by a grant from the Program Hospitalier de la Recherche Clinique. M.G. is supported by a grant funded by the Agence Nationale pour la Recherche (Project ANR-07-MRAR-021), and T.O.-M is supported by a grant from the Fondation pour la Recherche Médicale. The Three-City Study is conducted under a partnership agreement between Inserm, the Victor Segalen–Bordeaux II University, and Sanofi-Synthélabo. The Fondation pour la Recherche Médicale funded the preparation and first phase of the study. The Three-City Study is also supported by the Caisse Nationale Maladie des Travailleurs Salariés; Direction Générale de la Santé; MGEN; the Institut de la Longévité; Agence Française de Sécurité Sanitaire des Produits de Santé; the regional governments of Aquitaine, Bourgogne, and Languedoc-Roussillon; and the Fondation de France, the Ministry of Research-Inserm Program “Cohorts and collection of biologic material.” The Lille Génopôle received an unconditional grant from Eisai.

Contribution: P.-E.M., M.L., J.E., P.A., and D.A.T. designed the research analyzed data and wrote the paper; and T.O.-M., W.C., M.G., N.S., G.A., M.-C.A., M.B., A.-M.D., L.L., L.M.L., and J.-C.L. performed research and analyzed data.

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

L.M.L. has previously published under her maiden name, Lorna M. Houlihan.

Correspondence: Professor Pierre-Emmanuel Morange, Laboratory of Haematology, CHU Timone, 246, rue Saint-Pierre, 13385 Marseille Cedex 05, France; e-mail: pierre.morange@ap-hm.fr.