Common And Rare Genetic Variants Associated With Venous Thrombosis 81
remains to be done to resolve remaining genetic determinants. Adopting a non-hypothesis-driven approach based on genome-wide association scans is likely to yield additional variants in potentially unexpected genes whose magnitude of effect may be smaller, but whose importance for our ability to define risk and understand disease pathogenesis is likely to be significant.
In the normal physiological state a balance exists between pro-coagulant and anticoagulant mechanisms, allowing extravascular blood to clot while maintaining blood flow within the circulation. Central to this process is thrombin, which promotes haemostasis when generated by vascular injury with physiological control of bleeding through vasoconstriction, platelet aggregation and coagulation, leading to fibrin polymerization and clot formation. However on binding to the endothelial cell membrane protein thrombomodulin, thrombin promotes anticoagulant effects through activation of protein C, which cleaves and inactivates specific coagulation factors Va and Villa to limit clot formation. This protease activity of protein C is dependent on a cofactor, protein S.
Rare deficiencies of specific inhibitors of the procoagulant system (the natural anticoagulants) have been recognized for many years among families with a strong history of recurrent thrombotic events. In 1965, Egeberg and colleagues identified deficiency of antithrombin as a cause of thrombophilia (Egeberg 1965). Antithrombin is an inhibitor of thrombin and thought to be the most important physiological inhibitor of the coagulation pathway. Multiple rare variants have been described in the SERPINC1 (previously known as AT3) gene encoding antithrombin at chromosome 1q23-q25.1 including missense, nonsense, and deletion variants affecting the function or levels of antithrombin protein. Homozygosity for antithrombin deficiency is thought to be lethal; heterozygotes have a ten-fold increased risk of thrombosis and are thought to be present among one in 2000 individuals. Protein C deficiency (Griffin et al. 1981) and protein S deficiency (Schwarz et al. 1984) are further conditions due to many different specific, very rare variants associated with risk of thrombosis in the heterozygous state but whose effects are typically modulated by other genetic and environmental factors. Unlike antithrombin deficiency they are found in the homozygous form, presenting with severe thrombosis soon after birth as neonatal purpura fulminans. Cumulatively, variation in the genes encoding antithrombin, protein C, and protein S are thought to account for only a small proportion of the genetic factors underlying thrombophilia, for example representing less than 5% of patients with familial forms of disease.
2.6.1 Factor V Leiden
A major breakthrough in this field was the discovery of a common variant involving the F5 gene encoding coagulation factor V, which conferred resistance to cleavage by activated protein C, and was found to account for a substantial proportion of the genetic contribution to risk of venous thrombosis in the general population (Bertina et al. 1994). The discovery had its origins in the recognition that a patient with a recurrent history of venous thrombosis showed a poor anticoagulant response to activated protein C (Dahlback et al. 1993). This led to the development of a new assay to screen for this phenotype, 'activated protein C resistance', that was shown in the Leiden thrombophilia case-control study to occur in 5% of healthy individuals but in 21% of unselected consecutive patients with a first episode of deep vein thrombosis (Koster et al. 1993) and in over 40% of those with a recurrent or family history of thrombosis (Griffin et al. 1993; Svensson and Dahlback 1994).
Family studies suggested that resistance to activated protein C was inherited in an autosomal dominant fashion and the molecular basis for this was found to involve coagulation factor V (Bertina et al. 1994; Dahlback and Hildebrand 1994). Linkage analysis using microsatellite markers implicated chromosome 1q21-q25, specifically the F5 gene locus, with activated protein C resistance. Resequencing of gene regions encoding the proposed activated protein C binding site and cleavage site in the coagulation factor V protein resolved a specific SNP in exon 10 of F5. The SNP, a G to A transition altered the codon from CGA to CAA leading to an amino acid substitution from arginine to glutam-ine at position 506 (p.R506Q) (c.1517G>A) (rs6025) (Bertina et al. 1994). The protein variant was named factor V Leiden after the Dutch city of Leiden where the study was based. Complete segregation was found in the affected family pedigree between heterozygosity for the SNP and activated protein C resistance.
Analysis of the Leiden thrombophilia case-control cohort showed that of the 64 patients and six controls with a phenotype of activated protein C resistance, 56 possessed the allelic variant while none of those without activated protein C resistance had the rarer A allele (Bertina et al. 1994). The variant modulated an activated protein C cleavage site, rendering the factor V Leiden protein significantly less sensitive to degradation and inactivation, and promoting a hypercoagulable state (Aparicio and Dahlback 1996). What was particularly striking was that this was a common SNP, originally reported as present in 2% of the Dutch population, with subsequent studies showing it to be common in those of European ancestry, with a prevalence of up to 15%, but rare or absent in African, East Asian, or other population groups (Rees et al. 1995; Dahlback 2008). Possession of one copy of the A allele (being heterozygous for factor V Leiden) is associated with a five-fold increased risk of venous thrombosis; having two copies leads to a 50-fold increased risk (Dahlback 2008). All those with the variant allele have the same haplotype at F5 with evidence that this was a recent mutational event, estimated as occurring 21 000 years ago, after the proposed migration out of Africa (Zivelin et al. 2006). Whether there was any selective advantage to possession of this variant or associated alleles is unclear; reduced risk of severe bleeding after child birth has been proposed (Lindqvist et al. 1999; Dahlback 2008).
2.6.2 Genetic diversity and thrombophilia: insights and applications
Soon after this discovery, a candidate gene approach led to the identification of a further common variant associated with risk of venous thrombosis. Here the candidate was prothrombin, encoded by the F2 gene at chromosome 11p11-q12, which is a precursor of thrombin. Poort and colleagues sequenced the 5' and 3' regions of the F2 gene, together with the exons, and found a G to A transition in the 3' untranslated region of the gene at nucleotide 20210, which was present in five out of 28 individuals with a personal or family history of venous thrombosis but none of the five controls (Poort et al. 1996). Analysis of the Leiden thrombophilia study case-control cohort, which included 424 unselected patients with deep vein thrombosis and 474 controls, showed 6.2% of cases and 2.3% of controls were heterozygous for the SNP (Poort et al. 1996). Possession of the SNP was associated with an
OR of 2.8 (1.4-5.6) for venous thrombosis and increased circulating levels of prothrombin, suggesting this or a linked polymorphism was acting as a regulatory variant modulating gene expression (Poort et al. 1996). Like the factor V Leiden SNP, in certain populations such as in southern Europe, the prothombin SNP is present at a relatively high frequency (2-4% of healthy individuals) but is rare in populations of non-European ancestry, with evidence of being a relatively recent mutation arising some 24 000 years ago (Zivelin et al. 2006).
The manifestation of disease in those with particular genetic variants is, however, highly variable such that some patients may never have a venous thrombosis while others will do so at a young age and have recurrent events. This reflects the multifactorial nature of the disease, with gene-gene and gene-environment interactions being important. A large meta analysis of 2130 cases and 3204 controls from eight case-control studies showed that possession of the factor V Leiden SNP was associated with an OR of 4.9 (4.1-5.9) for venous thrombosis, while the prothrombin SNP had an OR of 3.8 (3.0-4.9); possession of a copy of the variant allele for both SNPs was associated with an OR of 20 (11.1-36.1) for venous thrombosis and a significantly earlier age of having a first thrombotic event, suggesting a multiplicative rather than additive effect (Emmerich et al. 2001). Use of the oral contraceptive pill was associated with an odds ratio for venous thrombosis of 2.3 (1.7-3.0) which increased to 10.2 (5.7-18.4) in the presence of the factor V Leiden variant, and to 7.1 (3.4-165.0) with the prothrombin variant.
Research in this area has given us important insights into pathophysiology. Indications for thrombophilia screening are however currently controversial. It has been suggested that screening should be undertaken in those presenting with venous thrombosis at a young age (less than 50 years of age) or those presenting without apparent cause, or with a history of recurrent throm-botic events (Whitlatch and Ortel 2008). Others have argued that for the most common mutations, management is not altered and so screening rarely alters clinical care. Patients heterozygous for the factor V Leiden SNP or the prothrombin SNP alone do not have a clinically significantly increased risk of recurrence (Ho et al. 2006; Marchiori et al. 2007). Screening might however help in the counselling of asymptomatic relatives in terms of preventative measures or use of the oral contraceptive pill.
Those individuals possessing a copy of the risk allele for both the common SNPs, or being homozygous for factor V Leiden, or having antithrombin deficiency, will be rare but may be candidates for lifelong anticoagulation, similarly those with a history of recurrent or life threatening thrombotic events, antiphospholipid antibodies, or underlying cancer (Whitlatch and Ortel 2008).
2.7 Summary
The search for genetic determinants of disease has required remarkable perseverance and scientific endeavour, a task which has been facilitated by recent advances in human genetics, notably the completion of sequencing of the human genome and our understanding of the nature and extent of human genetic variation. Prior to this, the availability of increasingly polymorphic genetic markers and associated genetic and physical maps set the stage for linkage analysis and positional cloning to unlock the secrets of many rare monogenic diseases showing a clear phenotype with a mendelian pattern of inheritance. From the 1980s onwards over 1200 genes involved in such diseases have been defined (Botstein and Risch 2003). In this chapter a series of different diseases has been reviewed to highlight the theoretical basis of linkage analysis and positional cloning, and different issues that have arisen in their application. Pioneering work in cystic fibrosis, for example, highlighted the very considerable hurdles that had to be overcome to define the candidate region, identify and assemble clone DNA fragments, and define, prioritize, and interrogate transcribed regions before resolving a specific novel transcript in which it was possible to identify and screen for specific mutations. In most diseases showing mendelian inheritance investigated by linkage and positional cloning, a series of rare mutations were found to underlie the observed disease phenotype, usually altering the structure and function of the encoded protein. The situation was unusual for cystic fibrosis where a 3 bp deletion (the delta-F508 deletion) was found to account for 66% of cystic fibrosis chromosomes worldwide, although there are over 1600 other rare mutations of the CFTR gene identified.
For more common multifactorial traits not showing a clear mendelian segregation of inheritance, linkage analysis was less successful. Here it was becoming clearer that several genes were likely to be involved and many variants, with an individually modest magnitude of effect. Genetic association studies were extensively employed using a candidate gene approach although only a relatively small number of examples of robustly replicated associations were described. There have been significant limitations to many such studies relating to definition of the disease phenotype, a low prior probability of success, underlying population stratification, the choice of a limited number of markers for analysis, and underpowered study design.
The cataloguing of the extent, nature, and coinherit-ance of genetic diversity is continuing to advance our ability to define the genetic basis for common disease, notably in terms of SNP diversity with high throughput genotyping technologies and the availability of large panels of clinical samples and controls enabling genome-wide association studies as reviewed in Chapter 9. The distinction and similarities of using genetic markers to define disease regions in linkage analysis and association studies have been reviewed, the former utilizing a limited number of informative recombination events in the family pedigrees analysed to have a relatively broad resolution of disease regions. This contrasts with association studies where historical recombination and mutational events, and varying linkage disequilibrium between variants, provide increased resolution on analysing individuals sampled from a population.
In this chapter several different diseases have been reviewed in detail and other examples are found elsewhere in this book. Alzheimer's disease provides an elegant example of a complex multifactorial trait in which a proportion of cases with early onset familial disease show a mendelian inheritance that has allowed rare variants in at least three different genes (APP, PSEN1, and PSEN2) to be defined by linkage. Gain of function mutations in these three genes appear to lead to common mechanisms involving accumulation of AP protein. Gene dosage effects involving the normal protein are also apparent for APP at chromosome 21q21 as seen with trisomy 21 in Down syndrome and gene duplication events involving APP. In late onset Alzheimer's disease a common variant, the e4 allele, is significantly associated with age of onset of disease. This involves increased aggregation and reduced clearance of AP, but inheritance of this allele is neither necessary nor sufficient for disease to occur.
Research into the genetic basis of monogenic diseases showing mendelian inheritance has enabled the development of clinical genetics as a medical speciality, with application across medical disciplines to facilitate diagnosis and allow clear delineation of genetic risk. By contrast, research to determine specific genetic factors contributing to common multifactorial traits is in many ways less advanced. New insights into underlying disease pathogenesis have been defined but it is a complex picture, such that application to personalized medicine and precise delineation of individual risk associated with inheritance of particular alleles remains a significant hurdle to be overcome. Over the course of Chapters 3-9 different classes of genetic diversity are reviewed, set in the context of population and molecular genetics with examples relating to human disease and evolution as well as diversity in other species. We return specifically to the theme of defining the genetics of common disease in detail in Chapter 9, when SNP diversity is considered in greater detail and the steps which have led to genome-wide association studies are reviewed.
2.8 Reviews
Reviews of subjects in this chapter can be found in the following publications:
|
Topic |
References |
|
Linkage and positional cloning |
Collins 1991, 1992, 1995; Ballabio 1993; Risch and Merikangas 1996; Botstein |
Post a comment