Genetic risk factors, gene-environment and gene-gene interaction in venous thrombosis.

a. overall aim and key objectives 1. To assess the joint effect of common risk factors for thrombosis a. interaction of genes and environment b. interaction of genes and genes 2. To study the effect of relatively rare genetic variants Venous thrombosis is a common disease (1-2 per 1000 per year) with a multicausal aetiology. Defects that were identified several decades ago (protein C- , protein S-, antithrombin deficiency) increase risk over 10-fold, but are rare (<0.3% in the general population). Recently, abnormalities were identified that are moderate risk factors (relative risks 3-8), and that are more common (2-4%), such as factor V Leiden and prothrombin 20210A, and finally there are common, but weak risk factors, ie, hyperhomocysteinemia, high levels of clotting factors VIII, IX, and XI, MTHFR 677T and FXIII val34leu (>10%, RR 1.3-2). Environmental risk factors, such as surgery, malignancies, pregnancy, hormone use, may be common and increase risk 2-10 fold.17 The key question is why a certain individual, with a given genetic make-up, develops thrombosis at a certain point in time. It is likely that this is the result of the joint effect of many factors. While common factors overall may have small effects, they may be strong modifiers of risk when combined with other factors. Joint presence of several risk factors will be relatively rare. The ultimate combination can be seen as a rare but very strong risk factor, that gives a high (to infinite) risk in comparison to those without the complete combination (cf, a single monogenetic risk factor that invariably leads to disease). Our basic hypothesis is that common diseases happen in one individual and not in the other because of the presence of rare combinations of common risk factors. An example of such interaction is the use of female hormones, where we reported a 35-fold increased risk for oral contraceptives and factor V Leiden, and a 15-fold increased risk for postmenopausal hormones and factor V Leiden. Or course, they can also happen because of the presence of rare but strong single genetic risk factors. Our study aims at both. We have set up a study to investigate the interaction of common genetic and environmental risk factors (table). Since we are looking at combinations of risk factors, the study is large. The MEGA study began in 1999 and will include 2500 patients with venous thrombosis and 2500 controls. Because of the major effort required to see so many people for a blood draw, funding sofar has only concerned the logistic framework. The current proposal aims at acquiring Ph.D. students for the scientific analyses of the primary question (1a), and to take the study further, by enlarging it to 10 000 people with risk factor data and DNA This answers questions 1b and 2, the search for rare and strong genetic variants. It will be done as efficiently as possible, by mailed questionnaires and DNA from buccal swabs, and by ‘smart designs’ to restrict the number of assays. prevalence relative risk Factor V Leiden 4% 3-8 Prothrombin 20210A 2% 3 MTHFR 677T (homozygous) 10% 1.3 factor XIII val34leu 6% 0.7 factor VIII (>150IU/dl) 10% 3 factor IX (>129 IU/dl) 10% 3 fibrinogen (>4 g/l) 8% 2 homocysteine (>18.5µM/l) 5% 2 surgery/immobilisation (<6 months prior) 5% 10 pregnancy/puerperium (15-49 yr) 2% 5 oral contraceptives (15-49 yr) 35% 4 malignancies 1-2% 5-10 b. approach In the MEGA study, we include consecutive patients of six anticoagulation clinics (Leiden, Den Haag, Rotterdam, Utrecht, Amsterdam, Amersfoort), aged 18-70, with a first deep-vein thrombosis or pulmonary embolism. Controls are spouses of patients, while a second control group consists of a random population sample included by random digit dialing. The choice for these two control groups is based on the consideration that for genetic factors all are equally appropriate, while for environmental factors it varies with the variable studied. Overall response is over 75%, and when the first funding ends, in April 2003, we will have included the intended number of 5000 participants. In most subjects we perform an in-person interview and venapuncture for plasma and DNA. In individuals reluctant to participate in full (~10%), we obtain DNA by self-administered mailed buccal swabs. For the second part of the study, we will limit data collection to questionnaires and buccal swabs. Because we will extend the study to 10 000 individuals, we will be able to analyse not only gene-environment interactions of common risk factors (factor V Leiden, high levels of procoagulant factors, environmental factors), but also to study gene-gene interactions (factor V Leiden, prothrombin 20210A, MTHFR 677T), and multiple (>2) risk factor interactions. For the statistical analysis main effects will be estimated as odds ratios. Joint effects will be assessed by the relative risk for those with two or more risk factors. To increase statistical power, we will also estimate the synergy index in cases only. Finally, we will study the effect of new candidate risk factors. As the humane genome project comes to an end millions of SNPs (Single Nucleotide Polymorphisms) are identified. These SNPs are in part localized in the ~500 candidate genes that may affect thrombotic risk, which includes genes involved in blood coagulation, fibrinolysis, inflammation and vessel wall integrity. Not all known SNPs in this set of genes will be examined, for this would translate into millions of assays. Therefore, only SNPs that have a relatively low allelic prevalence, up to ~2 %, will be considered. The reason is that it is far more likely that these relatively rare SNPs are of functional importance and carry a significant thrombotic risk. Priority will be given to SNPs known to be associated with an intermediate phenotype such as plasma or tissue levels at rest or after stimulation. When such SNPs are not known, preference will be given to SNPs that predict amino acid replacements. When these are also not available, we will resort to SNPs in other functional segments of the gene, like the promoter region. We will refrain from testing a particular candidate gene when useful SNPs are not available. Examination of such candidate genes has to await further sequencing