Genetic Contribution to Circulating Levels of Hemostatic Factors in Healthy Families With Effects of Known Genetic Polymorphisms on Heritability
Levels of fibrinogen, factor VII (FVII), factor XIII (FXIII), plasminogen activator inhibitor (PAI)-1, and tissue plasminogen activator have been associated with coronary artery disease as have genetic polymorphisms. Quantitative genetic analyses allow determination of the genetic contribution to phenotypic variation. We investigated familial influences on these hemostatic factors in 537 adults from 89 randomly ascertained healthy families of white North European origin. We used maximum likelihood analysis to estimate the heritabilities of these factors and effects of covariates on the factors in these families. After adjustment for age and sex, the factors showed considerable heritability, varying from 26% (PAI-1) to 47% (FXIII complex). The influence of known polymorphisms was negligible for fibrinogen and contributed 2% to the variance of the FXIII complex and PAI-1 and 11% to the variance of FVII coagulant activity. Age, sex, body mass index, lifestyle, and metabolic covariates explained between 10% (FXIII) and 44% (PAI-1) of phenotypic variance. Childhood household influences significantly affected FVII (11%) and FXIII (18%). A significant degree of phenotypic variance of several hemostatic factors can be explained by additive genes and known covariates. The impact of certain well-characterized polymorphisms to the heritability is small in this population of healthy families, indicating the need to localize new genes influencing hemostatic factor levels.
Developing statistical methodology to analyze data collected from complex pedigrees allows investigation of the traits that underlie the strong influence of family history on the risk of common conditions such as atherothrombotic disease and diabetes mellitus. The identification of heritability should be a prerequisite for the search for quantitative trait loci affecting such traits and thereby modulating the risk of disease. There is strong evidence for genetic familial influences on body mass index (BMI), and metabolic cardiovascular risk factors such as dyslipidemia and candidate genetic loci have been identified. To date, most information on the genetic heritability of circulating levels of the hemostatic vascular risk markers fibrinogen, plasminogen activator inhibitor (PAI)-1, tissue plasminogen activator (tPA), and factor VII (FVII) has been collected from single-sex twin studies or from families selected on the basis of being at increased risk of diabetes, heart disease through insulin resistance, or thrombosis.1–3⇓⇓ It is not clear whether the high estimates of heritability for hemostatic risk factors will apply to healthy mixed-sex families. Furthermore, polymorphisms identified in the genes encoding fibrinogen and PAI-1 appear to have a significant but small impact on the variance of circulating levels,4,5⇓ and although FVII gene polymorphisms influence protein levels more significantly, these results have been derived from population rather than family studies.6,7⇓
The present study aimed to estimate additive genetic heritability of the hemostatic risk factors fibrinogen, PAI-1, tPA, and FVII and also of coagulation factor XIII in representative healthy White European families in a variance components analysis, with adjustment for the influence of known covariates and for genotype at polymorphisms previously identified as contributing to factor levels and disease.
A list of potential subjects was generated at random from the community-based Family Health Authority register in Leeds, and the potential subjects were contacted by mail. Each subject (proband) who replied was included in the present study if healthy (defined as the absence of clinically overt vascular disease or diabetes mellitus) and if ≥4 first-, second-, or third-degree relatives were able to take part. All probands were aged ≥16 years, were of white European origin, and gave informed consent according to a protocol approved by the Leeds Teaching Hospitals Trust Research Ethics Committee. A total of 280 subjects were contacted, with 89 (32%) having appropriate family size and health. Participating family members were interviewed to obtain information regarding the individuals past medical history and smoking and alcohol consumption patterns. Systolic and diastolic blood pressures were measured to the nearest 2 mm Hg with each subject in a sitting position, and calculations were performed from the mean of 3 readings over 10 minutes. BMI was calculated from weight in kilograms divided by height in square meters.
To minimize the influence of the circadian rhythm on hemostatic factor levels, subjects were asked to arrive at the hospital between 7:00 and 11:00 am. After an overnight fast of at least 10 hours (including abstention from cigarette smoking), blood samples were drawn from an antecubital vein with a 19-gauge needle without venous stasis. Blood was taken into 0.9% citrate on ice for PAI-1 and tPA and into lithium heparin at room temperature for fibrinogen, FVII, and factor XIII (FXIII). These samples were centrifuged at 3000g for 30 minutes at 4°C for tPA and PAI-1. Aliquots of plasma from the spun samples were snap-frozen in liquid nitrogen and stored at −40°C until assay. Fibrinogen levels were measured by the method described by Clauss.8 The interassay coefficient of variation (CV) was 3.5%. Levels of FVII coagulant activity (FVII:C) were determined by a clotting assay with the use of FVII-deficient plasma. Clotting activity was measured on the ACL 300 (Instrumentation Laboratory) and expressed as a percentage of activity given by calibration plasma. This assay reflects FVII zymogen levels and does not appreciably detect activated FVII. The interassay CV was 3.2%. PAI-1 antigen and tPA antigen were measured by ELISA (Biopool International), with interassay CVs of 9.7% and 9.5%, respectively. FXIII A2B2 complex was measured on an anti-B/anti-A sandwich as described by Komanasin et al.9 The interassay CV was 6.8%.
Genomic DNA was extracted from 10 mL of whole blood by using a Nucleon DNA extraction kit (Nucleon Biosciences). The restriction fragment length polymorphisms were genotyped after amplification of relevant DNA regions by polymerase chain reaction and digestion with the appropriate restriction enzymes as previously described.10–13⇓⇓⇓
Levels of the hemostatic factors were logarithmically transformed to remove skewness. Reverse-transformed values are presented in descriptive tables for ease of interpretation. For the genetic analysis, pedigree and phenotypic data were prepared by using the computer package SPSS (SPSS Inc).
The heritability of each hemostatic factor was calculated by using standard quantitative genetic variance components analysis.14 After adjustment for covariates (including age, sex, current smoking habits, BMI, alcohol intake, fasting glucose, cholesterol, and triglycerides), the distribution of each trait within a pedigree was assumed to be multivariate normal with variance-covariance matrix given by the following: V=2ςA2φ+ςE2I, where φ is the known matrix of kinship coefficients, and I is the identity matrix. The variance components ςA2 (the additive genetic variance or heritability) and ςE2 (the residual variance) were estimated by maximum likelihood, implemented in the SOLAR software package (Southwest Foundation for Biomedical Research).15 In some models, an additional component of variance due to shared household effects (ςC2) was included. Thus, the variance-covariance matrix becomes the following: V=2ςA2φ+ςE2I+ςC2H, where H is a matrix consisting of the numbers 0 and 1, depending on whether or not the pedigree members were siblings and, if so, were thus assumed to have had a common childhood household environment. No ascertainment correction was necessary for the analyses because probands were ascertained independently of phenotype. Heritability is presented in the results as the proportion of total phenotypic variance (ie, variance before adjustment for covariates) that could be attributed to the additive effects of genes.
Five hundred thirty-seven subjects were recruited and examined from 89 pedigrees (mean subjects 6.0 per family) from 133 households. The structure of the pedigrees, which varied in size from 3 to 28 individuals, is shown in Table 1. Table 2 gives the clinical characteristics of the subject group. The mean age of the probands was 43.4 years, and the whole population (of which 40.4% were male) had a mean age of 43.2 years. Of the females, 63% were premenopausal (27.9% of whom were on estrogen-containing oral contraceptives), and 37% of postmenopausal women were on hormone replacement therapy. Seven subjects (1.3%) were known to have type II diabetes, and a further 2 subjects had fasting blood glucose >7.0 mmol/L. Of the subjects, 22 (4.1%) had overt ischemic heart disease, and 7 (1.3%) had cerebrovascular disease (previous stroke or transient ischemic attack). These subjects were excluded from the analysis. Sixteen subjects were taking β-blockers, 12 were taking calcium antagonists, 14 were taking ACE inhibitors, 25 were taking diuretics, and 23 were on lipid-lowering therapies. The mean BMI was 25.9 kg/m2, and the group showed a healthy atherosclerotic risk profile.
Levels of PAI-1 and tPA were significantly higher in men than in women, whereas fibrinogen was higher in women. After adjustment for age, the levels of fibrinogen (2.85 versus 3.07 g/L, P<0.01), FVII:C (102.8% versus 118.2%, P<0.01), and PAI-1 (9.6 versus 15.9 nmol/mL, P<0.01) rose after menopause. In premenopausal women, use of the combined oral contraceptive pill was associated with lower levels of PAI-1 and tPA (P<0.01), whereas in postmenopausal women, estrogen replacement therapy was associated with lower levels of fibrinogen and PAI-1 (P<0.01).
Table 3 shows the effect of adjustment of known (nongenetic) covariates. The largest incremental change in the model R2 was attributable to age and sex for fibrinogen, FXIII complex, and tPA. For FVII:C and PAI-1, BMI was the most significant covariate. Demographic, lifestyle, and metabolic variables and BMI explained 10% of the variance of the FXIII complex, 20% of the variance of fibrinogen, 23% of the variance of FVII:C, and 44% and 41% of the variance of PAI-1 and tPA antigen, respectively. Table 4 shows mean hemostatic factor levels according to genotype. The distributions of the fibrinogen and PAI-1 genotypes were not in Hardy-Weinberg equilibrium because of the familial nature of the population. Genotype had a small but significant effect on the levels of the FXIII complex (2% for Val34Leu), PAI-1 (2% for 4G/5G), and for FVII:C (11% for Arg353Gln).
We performed quantitative genetic analyses for each of the hemostatic factors before and after adjustment for age, sex, BMI, and the appropriate polymorphism. The heritability estimates are shown in Table 5 and are expressed as the proportion of the total phenotypic variability rather than residual variability after adjustment for covariates. Unadjusted, heritability estimates varied from 23% (tPA) to 52% (FXIII complex). Adjusting for age, sex, and known environmental covariates did not significantly influence the heritability estimates. The heritability of FVII:C was the only estimate to be significantly affected by adjustment for the relevant genetic polymorphism studied.
Table 6 shows the components of variance attributable to the measured covariates, additive genetic effects, shared childhood household effects, and residual environmental factors (including assay imprecision) for the hemostatic factors in this population. The proportion of total phenotypic variability accounted for by household effects tended to be considerably smaller than that accounted for by the genetic effects. In fact, shared childhood households were only significant for FVII:C (11%) and the FXIII complex (18%). Modeling the household on “current” address was significant only for fibrinogen (0.21, SE 0.07; P<0.01). The effect of childhood or current household did not contribute to the variance of either PAI-1 or tPA. The final model accounted for 48% of the variance in fibrinogen levels, 57% of the variance of the FXIII complex, and 58%, 65%, and 67% of the variance in PAI-1, tPA, and FVII:C, respectively.
We performed a variance components analysis on important hemostatic proteins from healthy white North European families, with the assessment of specific polymorphisms related to hemostatic factor levels and/or vascular disease. The heritabilities for all factors were significant: 37% for fibrinogen, 40% for FVII:C, 52% for the FXIII complex, and 28% and 23% for PAI-1 and tPA, respectively, indicating the importance of genetic factors. Furthermore, the previously identified and frequently studied polymorphisms accounted for only a small fraction of this heritability.
Fibrinogen is widely recognized as an independent risk factor for coronary artery disease, clusters with other cardiovascular risk factors,16 and may provide a link between vascular disease and smoking. As a result, there has been much interest in determining genetic and environmental influences on levels, although these have mainly been in population studies of subjects with vascular disease. Heritability estimates, which range from 27% to 51% in twin and pedigree studies,17–19⇓⇓ are broadly in keeping with our estimates, although twin studies may overestimate heritability because of greater behavioral sharing in monozygous compared with dizygous twins. Many studies have shown a relationship between common fibrinogen β chain polymorphisms (which are all in linkage disequilibrium) and levels.20,21⇓ However, the influence of these polymorphic variations on fibrinogen levels is weak and was not detected in the present study of healthy individuals, in common with other studies of healthy subjects.22,23⇓ The minimal impact of the polymorphisms on the total genetic contribution possibly underlies this. The strong contribution of current household, as modeled by current address, at 21% of variance, may result from shared household behavior, such as smoking habits, lack of exercise, or shared common infections.24
Our heritability estimate for FVII is similar to that found in Spanish families,3 although it is less than that found in twins.1 The gene coding for factor VII has 5 identified polymorphic sites that may contribute up to 30% of protein levels in population studies,6 a finding that is reinforced by the impact of the Arg353Gln polymorphism on the heritability estimate. The contribution of the Arg353Gln allele to levels in the present study of white families is similar to previous estimates in healthy southeastern Asians25 and whites,4 although because of strong linkage disequilibrium, we examined only 1 polymorphism. This is the first heritability estimate for the FXIII complex. FXIII is composed of 2A and 2B subunits circulating as a tetrameric structure, with the A subunit containing the active site while bound to the carrier B molecule.26 Previous heritability estimates of FXIII have given different results for each subunit, possibly as a result of buffering of a free pool of B subunit.1 Assaying the complex results in a more accurate measure of protein levels and, therefore, a better estimate of the genetic contribution to phenotypic variability. The Leu allele at the FXIII A subunit has been associated with an increased FXIII activation rate13 and protection against arterial and venous thrombotic events. We found that the Val34Leu genotype significantly affected complex levels despite this polymorphism having been previously shown to have no effect on subunit levels as measured by ELISA.27 We also found a significant contribution to variance of shared sibling environment. de Lange et al1 found that the common environment of twins contributed to the variance of the B subunit, and this may be reflected in our finding for FXIII complex levels. The mechanism for this is unclear at this time.
Our heritability estimate for PAI-1 antigen (26%) is similar to that in previous family studies in healthy families (32%)28 but less than that in twin studies.29 We also found that traits of the insulin resistance syndrome are the most significant contributors to phenotypic variance, in agreement with the findings of Henry et al.30 The 4G/5G polymorphism, situated 675 bp from the transcription start site, has a triglyceride responsive element.31 Some studies, particularly those in subjects with high triglyceride levels, have found higher PAI-1 levels in 4G/4G subjects.32 However, this polymorphism appears to contribute <5% to the variance of PAI-1. We found that possession of the 4G genotype was associated with higher levels of PAI-1, although the contribution of the genotype was small (2%), in keeping with the study by Henry et al.
tPA levels are related to the development of myocardial infarction.33 We showed a significant genetic contribution to variance similar to a previous family study in a different population.3 A relationship between the most closely studied polymorphism, the Alu insertion/deletion, and phenotype has yet to be established, and although it may affect release,34 it does not directly affect resting levels. The genetics underlying circulating levels of tPA are complicated by its interaction in the circulation with PAI-1, which has been shown to determine up to 38% of tPA variance.35 In the present study, PAI-1 contributed up to 26% to tPA levels, whereas the 4G/5G polymorphism also significantly affects tPA levels, possibly because of linkage or pleiotropy with the tPA gene.
We have demonstrated a significant genetic contribution to levels of a number of hemostatic factors, implicated in vascular disease, in healthy white families after allowing for the effect of known covariates. We have studied healthy white families not selected by the presence of pathology or increased risk, emphasizing the importance of the genetics of hemostasis in lower risk families. The overall contribution of specific polymorphisms, previously found to affect levels and coronary artery disease, is comparatively small and indicates why so many genotype/disease studies show variable results (reviewed by Lane and Grant36). It is likely that the majority of the remaining genetic influence on hemostatic factor levels is due to either several unidentified polymorphisms in known genes or as-yet-unidentified genes, possibly with a pleiotropic effect on multiple hemostatic and metabolic factors implicated in vascular risk.37,38⇓
These results indicate the need to localize (and determine the effect of) new genes accounting for the total genetic influence on hemostatic factor levels. Identifying genetic contribution to phenotypic variance, made possible by recent developments in genetic analysis and population genetic statistics, will improve the understanding of disease processes, allowing the development of novel therapeutic approaches in the prevention of vascular disease.
This research was funded by a grant from The Special Trustees of the Leeds Teaching Hospitals NHS Trust.
Received October 16, 2001; revision accepted December 7, 2001.
- ↵Souto JC, Almasy L, Borrell M, Gari M, Martinez E, Mateo J, Stone WH, Blangero J, Fontcuberta J. Genetic determinants of hemostasis phenotypes in Spanish families. Circulation. 2000; 101: 1546–1551.
- ↵Dawson SJ, Hamsten A, Wiman B, Henney A, Humphries SE. Genetic variation at the plasminogen activator inhibitor-1 locus is associated with altered levels of plasma plasminogen activator inhibitor-1 activity. Arterioscler Thromb. 1991; 11: 183–190.
- ↵van’t Hoof F, Silveira A, Tornvall P. Two common functional polymorphisms in the promotor region of coagulation factor VII determining plasma factor VII activity and mass concentration. Blood. 1993; 93: 3432–3441.
- ↵Clauss A. Gerinnungsphysiologische schnell methode zur bestimmung des fibrinogens. Acta Haematol. 1952; 17: 237–246.
- ↵Komanasin N, Futers TS, Ariens RAS, Grant PJ. A novel polymorphism in the factor XIII B subunit (His95Arg) relates to the dissociation of the A2B2 tetramer. Thromb Haemost. 1999; 82: 111a.Abstract.
- ↵Thomas A, Lamlum H, Humphris S, Green F. Linkage disequilibrium across the fibrinogen locus as shown by five genetic polymorphisms, G/A-455 (HaeIII), CT-148 (HindIII/AluI), T/G-1689 (AvaII) and Bcl1 (B-fibrinogen) and taq1 (a-fibrinogen), and their detection by PCR. Hum Mutat. 1994; 3: 79–81.
- ↵O’Hara PJ, Grant FJ, Haldeman BA, Gray CL, Insley MY, Hagen FS, Murray MJ. Nucleotide sequence of the gene coding for human factor VII, a vitamin K-dependent protein participating in blood coagulation. Proc Natl Acad Sci U S A. 1987; 84: 5158–5162.
- ↵Lynch M, Walsh B. Genetics and Analysis of Quantitative Traits. Sunderland, Mass: Sinauer Associates Inc; 1998.
- ↵Lee AJ, Lowe GDO, Woodward M, Tunstall-Pedoe H. Fibrinogen in relation to personal history of prevalent hypertension, diabetes, stroke, intermittent claudication, coronary heart disease, and family history: the Scottish Heart Health Study. Br Heart J. 1993; 69: 338–342.
- ↵van’t Hooft FM, von Bahr SJ, Silveira A, Iliadou A, Eriksson P, Hamsten A. Two common, functional polymorphisms in the promoter region of the beta-fibrinogen gene contribute to regulation of plasma fibrinogen concentration. Arterioscler Thromb Vasc Biol. 1999; 19: 3063–3070.
- ↵Tybjaerg-Hansen A, Agerholm-Larsen B, Humphries SE, Abildgaard S, Schnohr P. A common mutation (G-455→A) in the beta-fibrinogen promoter is an independent predictor of plasma fibrinogen, but not of ischemic heart disease: a study of 9,127 individuals based on the Copenhagen City Heart Study. J Clin Invest. 1997; 99: 3034–3039.
- ↵Saha N, Liu Y, Heng CK, Hong S, Low PS, Tay JS. Association of factor VII genotype with plasma factor VII activity and antigen levels in healthy Indian adults and interaction with triglycerides. Arterioscler Thromb. 1994; 14: 1923–1927.
- ↵Ichinose A; Bloom AL, Forbes CD, Thomas DP, Tuddenham EGD, eds. The Physiology and Biochemistry of Factor XIII. Edinburgh, UK: Churchill Livingstone; 1994.
- ↵Pankow JS, Folsom A, Province MA, Rao DC, Williams RR, Eckfeldt J, Sellers TA. Segregation analysis of plasminogen activator inhibitor-1 and fibrinogen levels in the NHLBI Family Heart Study. Arterioscler Thromb Vasc Biol. 1998; 18: 1559–1567.
- ↵Cesari M, Sartori MT, Patrassi GM, Vettore S, Rossi GP. Determinants of plasma levels of plasminogen activator inhibitor-1: a study of normotensive twins. Arterioscler Thromb Vasc Biol. 1999; 19: 316–320.
- ↵Henry M, Tregouet DA, Alessi MC Aillaud MF, Visvikis S, Siest G, Tiret L, Juhan-Vague I. Metabolic determinants are much more important than genetic polymorphisms in determining the PAI-1 activity and antigen plasma concentrations: a family study with part of the Stanislas Cohort. Arterioscler Thromb Vasc Biol. 1998; 18: 84–91.
- ↵Eriksson P, Nilsson L, Karpe F, Hamsten A. Very-low-density lipoprotein response element in the promoter region of the human plasminogen activator inhibitor-1 gene implicated in the impaired fibrinolysis of hypertriglyceridemia. Arterioscler Thromb Vasc Biol. 1998; 18: 20–26.
- ↵Ossei-Gerning N, Mansfield MW, Stickland MH, Wilson IJ, Grant PJ. Plasminogen activator inhibitor-1 promoter 4G/5G genotype and plasma levels in relation to a history of myocardial infarction in patients characterized by coronary angiography. Arterioscler Thromb Vasc Biol. 1997; 17: 33–37.
- ↵Jern C, Ladenvall P, Wall U, Jern S. Gene polymorphism of t-PA is associated with forearm vascular release rate of t-PA. Arterioscler Thromb Vasc Biol. 1999; 19: 454–459.
- ↵Chandler WL, Trimble SL, Loo SC, Mornin D. Effect of PAI-1 levels on the molar concentrations of active tissue plasminogen activator (t-PA) and t-PA/PAI-1 complex in plasma. Blood. 1990; 76: 930–937.
- ↵Lane DA, Grant PJ. Role of hemostatic gene polymorphisms in venous and arterial thrombotic disease. Blood. 2000; 95: 1517–1532.
- ↵Hong Y, Pedersen NL, Egberg N, de Faire U. Moderate genetic influences on plasma levels of plasminogen activator inhibitor-1 and evidence of genetic and environmental influences shared by plasminogen activator inhibitor-1, triglycerides, and body mass index. Arterioscler Thromb Vasc Biol. 1995; 17: 2776–2782.