Lack of association between VEGF polymorphisms and ALS in a Dutch population

Low levels of the vascular endothelial growth factor (VEGF) in gene-targeted mice cause progressive motor neuron degeneration, reminiscent of ALS.¹ In addition, in transgenic animal models of ALS, IM transfer of the VEGF gene as well as intracerebroventricular delivery of VEGF delayed onset of the disorder and prolonged survival of the animals.^2,3 In humans, a large association study in a geographically heterogeneous group of patients with ALS and controls was performed for three common polymorphisms in the VEGF promoter/leader sequence, known to be correlated with reduced VEGF expression.⁴ This study showed that two haplotypes (homozygosity for -2,578A/-1,154A/-634G [AAG] or -2,578A/-1,154G/-634G [AGG]) modestly increased the risk of developing ALS in a Belgian, Swedish, and British/Birmingham population, but not in another British population from the London area. In an attempt to further establish the association between VEGF polymorphisms and ALS, we investigated whether the at-risk haplotypes in the VEGF gene are associated with an increased risk of a population in The Netherlands developing ALS.

Methods.

Results.

The characteristics of the 373 patients with ALS and 615 controls are shown in table 1. VEGF genotyping of the -2.578 A/G, -1.154 A/G, and -634 C/G polymorphisms showed no difference in allele frequencies between patients with ALS and healthy controls, nor in any haplotypes, in particular the at-risk haplotypes AAG/AAG (0.11 vs 0.11, p = 0.91) and AGG/AGG (0.02 vs 0.02, p = 0.56) (table 2). All genotype variations in patients and controls were in accordance with Hardy-Weinberg equilibrium. No association was found in the subgroup analysis based on age or site of onset, sex, or disease duration (data not shown).

After pooling our results with those of Sweden, Belgium, and Britain,⁴ the strength of the association of the AAG/AAG haplotype was reduced compared with the previous meta-analysis (OR = 1.3 [1.1 to 1.7), p = 0.02 vs OR = 1.6 [1.2 to 2.3], p = 0.002) and no longer significant for the AGG/AGG haplotype (OR = 1.4 [0.9 to 2.3], p = 0.13 vs OR = 1.8 [1.0 to 3.3], p = 0.04) (figure). The meta-analysis of both AAG/AAG and AGG/AGG haplotypes showed an increased risk of ALS (OR = 1.38 [1.1 to 1.7), p = 0.005), although lower than previously reported (OR = 1.8 [1.3 to 2.2], p = 0.00004).⁴

• Download figure
• Open in new tab
• Download powerpoint

Figure. Meta-analysis of the ALS risk associated with the AAG/AAG (A), AGG/AGG (B), and combined genotypes (C) in individual and pooled populations. Data from Swedish (292 patients, 381 controls), Belgian (153 patients, 426 controls), and British Birmingham (90 patients, 96 controls) and London (158 patients, 143 controls) populations are adapted from Lambrechts et al.⁴ Controls in our study are age and gender matched and in the other studies age matched.

Discussion.

In a large sample of 373 patients with ALS and 615 controls, we did not find an increased risk in this Dutch population of developing ALS according to the at-risk haplotypes or for the individual polymorphisms. Our results are similar to the British/London population, which also failed to find an association between the VEGF genotype and ALS.⁴ Power calculation showed that the numbers of patients and controls in our study were sufficient to detect a relative risk of 1.6 at a 0.05 significance level, making it unlikely that sample size was the reason that no association was found. Furthermore, smaller sample sizes in Sweden (292 patients vs 554 controls), Belgium (153 patients vs 426 controls), and Britain/Birmingham (90 patients vs 96 controls) did show significant results.⁴

A difference in genetic background between countries may be an important factor leading to different results in genetic association studies in ALS. Also in familial ALS, the frequency and type of SOD1 mutations differ between populations, for example, autosomal recessive D90A-SOD1 mutations, are predominantly present in Sweden and the frequency of SOD1 mutations in familial ALS is more than 20% in England, Sweden, and Belgium but less than 5% in The Netherlands.^5–8 Because sporadic ALS is considered to be a complex disease, multiple genetic polymorphisms and environmental factors may eventually lead to motor neuron degeneration. Regional differences in the presence of susceptibility genes in sporadic ALS would be comparable to founder effects in familial ALS described in several geographic regions⁸ and suggests the possibility of a population-specific disease susceptibility in sporadic ALS.⁹

The lack of association between VEGF polymorphisms and ALS does not rule out a role for VEGF in the pathogenesis of ALS. Other sequence alterations in linkage dysequilibrium with the at-risk haplotypes may be the true modifiers. Decreased levels of VEGF in plasma and CSF were also detected in patients with ALS without at-risk VEGF genotypes, suggesting the involvement of other mechanisms lowering VEGF production.^4,10 Furthermore, VEGF has been shown to be neuroprotective in several animal and in vitro models for ALS, which makes it a potential therapeutic agent,^2–4 regardless of its possible role in susceptibility to ALS.

More studies and extended haplotype analysis in different geographical regions are needed to determine the exact role of VEGF polymorphisms in ALS susceptibility.