Update on Genetics of Stroke and Cerebrovascular Disease 2005 Hugh

Over the past 2 years this update reported the identification by the Icelandic Decode group of 2 novel genes associated with ischemic stroke: phosphodiesterase 4D gene (PDE4D) and 5-Lipoxygenase activating protein (ALOX5AP).1,2 These putative associations have been with specific haplotypes of each gene, but no disease-specific mutations in either gene have been identified. PDE4D, a cyclic nucleotide phosphodiesterase, selectively degrades second messenger cAMP (cAMP). Reduced cAMP levels are associated with increased smooth muscle cell proliferation and migration, key events in atherosclerosis, making an association with stroke pathophysiologically plausible. Consistent with this, the initial association was reported only with large artery and cardioembolic stroke subtypes.2 Studies over the last year attempting to replicate this association have produced diverse results. In a UK population no overall association was found with ischemic stroke, but possible associations were identified with cardioembolic and large artery stroke.3 An American study reported an association with ischemic stroke, particularly large artery stroke.4 In contrast, no association was found in a German stroke cohort,5 or a Swedish stroke cohort aged 75 years.6 A linkage study from a second Swedish population confirmed linkage to 5q12,6 but no linkage study could be found in an American population.4 No association was found with carotid intima-media thickness,3 suggesting PDE4D does not exert its effects via accelerating early atherosclerosis. ALOX5AP codes for 5-lipoxygenase activating protein which is essential for conversion of arachadonic acid to leukotriene A4, a process catabolized by 5-lipooxygenase. LTA4 is converted into LTB4, which plays a crucial role in leukocyte chemotaxis and inflammatory responses, key processes in atherosclerosis. This pathway was also implicated in a separate study reporting an association between 5-lipooxygenase itself and carotid intima-media thickness.7 A number of groups have attempted to replicate the association of ischemic stroke with the ALOX5AP gene. The DeCode group replicated the association in a Scottish stroke population.8 A case control study from Germany reported a weak association with an ALOX5AP polymorphism.5 In American populations, no association was found in a case control study, or linkage to this chromosomal region in an affected sibling pair study.3 There are a number of reasonable explanations for such disparate results. There might be significant genetic heterogeneity for ischemic stroke, thereby producing different results in different study populations. Population-specific genetic ancestral commonalities (and differences) might account for divergent results among patients in different countries. Lastly, random chance might produce spurious positive associations in some populations and studies, but not others. Further large studies in multiple populations are required to better define the biologic importance of these specific genes in ischemic stroke. There continues to be a large number of other candidate gene studies published over the last year reporting associations between polymorphisms in a wide variety of genes and stroke. Many of these have produced inconclusive results because of methodological problems and inadequate sample size. A recent review of these issues was published in Stroke and provides guidelines for future studies.9 An important consideration is the heterogeneity of stroke and potential for genetic variants to selectively predispose to particular stroke subtypes. The need for large well-phenotyped populations, ideally aged 65 to 70 years in whom the genetic component seems to be stronger, is becoming recognized, leading to a number of multicenter stroke DNA banks being established. Most stroke association studies have looked at single or a few single nucleotide polymorphisms (SNPs). In many cases, groups of nearby SNPs segregate as 1 genetic trait, with relatively little recombination or variation among such markers. A grouping of such SNPs is referred to as a haplotype. In 2005 a large multinational effort produced a haplotype map of the entire human genome (the HapMap Project), which involved analyzing over 1 000 000 SNPs in 269 individuals from Nigeria, China, Europe, Japan, and the United States.10 In some cases, the haplotypes demonstrated significant stability and lack of recombination. For example, 1 region that had 36 SNPs might have given rise to as many as 2 different haplotypes, yet only 7 different haplotypes were seen among 120 parental chromosomes. On the other hand, there are many well-defined ‘hot-spots’ for recombination throughout