VEGF PROMOTER HAPLOTYPE AND AMYOTROPHIC LATERAL SCLEROSIS (ALS)

Abstract

Vascular endothelial growth factor (VEGF) is a cytokine essential for angiogenesis. A recent study found that haplotypes, determined by three SNPs ( − 2,578C/A, − 1,154 G/A, and − 634G/C) in the VEGF upstream promoter/leader sequence, were associated with risk of amyotrophic lateral sclerosis (ALS). We used samples and data from a case-control study to examine the relation of ALS to VEGF haplotype. Genotypes at each of the three polymorphic sites were determined using allele-specific primer extension reactions followed by MALDI-TOF. We found a 3-fold increased risk among individuals homozygous for the AAG or AGG haplotypes (95% CI = 0.7 − 13.4), consistent with the findings of the previous study. Given the wide confidence interval, our findings should be interpreted cautiously.

Keywords:

• vascular endothelial growth factor
• amyotrophic lateral sclerosis
• epidemiologic studies
• haplotypes

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease affecting motor neurons in the brain and spinal cord. Little is known regarding genetic factors involved in ALS. Vascular endothelial growth factor (VEGF) is a cytokine essential for angiogenesis. A recent study found that haplotypes involving three SNPs in the VEGF upstream promoter/leaser sequence were associated with ALS risk (Lambrechts et al., 2003). These intriguing findings may have important implications for understanding ALS pathophysiology, yet confirmation in other studies is needed. Therefore, given the importance of the hypothesis and paucity of molecular epidemiological studies of ALS, we examined the association between VEGF haplotypes ALS risk in a case-control study in New England.

Materials and Methods

This case-control study has been described in detail elsewhere (Kamel et al., 2002). In brief, sporadic cases were recruited from two major referral centers in New England. Diagnosis of ALS (cf. Brooks et al., 1994) was confirmed by a board-certified neurologist. Patients had been diagnosed within two years of enrollment (1993 to 1996) and were eligible for participation if they had lived in new England at least half the year, spoke English, and were mentally competent. Population controls were identified by random telephone screening, and similar eligibility criteria were applied. Potential controls were excluded if they had a physician diagnosis of any neurodegenerative disease, neuropathy, or post-polio syndrome. We frequency matched controls to cases so that the distributions of the variables age, sex, and region within New England were similar in the two groups. Over 95% of the study subjects were of white, non-hispanic ethnicity. Among those eligible, 71% of the cases and 76% of the controls enrolled in the study; 95% of enrolled cases and 41% of enrolled controls donated a blood sample. Controls subjects who were invited but declined to come in to the laboratory were similar in age, gender, education, physical activity, smoking, and alcohol use of those who participated (Kamel et al., 2002). The present analysis includes 104 cases and 39 control subjects for whom complete data were available.

Genomic DNA was isolated from frozen whole blood (cf. Kamel et al., 2003). We examined the SNPs identified by Lambrechts et al. (2003): − 2,578C/A, − 1,154 G/A, and − 634G/C; they found that the haplotypes AAG and AGG conferred increased risk. Genotype at each of the three polymorphic sites was determined using allele-specific primer extension reactions followed by Sequenom's matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF) (cf. Storm et al., 2003). Individuals homozygous for the minor allele, heterozygous, or homozygous for the common allele could always be clearly distinguished, with peaks correctly corresponding to predicted masses.

We detected no deviations from Hardy-Weinberg equilibrium (Weir, 1996) (p > 0.54 for each locus). In our sample, we observed 11 of the 27 distinct three-locus genotypes possible. Eight of these observed genotypes uniquely determined a pair of haplotypes (diplotype). For example, a subject with genotype AA/AG/GG at the three loci, respectively, must carry haplotype AAG on one chromosome and AGG on the other. Three observed genotypes were ambiguous as to diplotype. For example, a subject with genotype AC/AG/GG could carry either of two diplotypes: haplotypes AAG and CGG or haplotypes AGG and CAG. For ambiguous genotypes, we assigned the most likely diplotype based on haplotypes seen in the unambiguous genotypes. The resulting set of diplotypes was the same as that reported by Lambrechts and colleagues (2003) with one exception. The relation of ALS to VEGF diplotypes was analyzed by multiple logistic regression to derive odds ratios (OR) and confidence intervals (CI) (cf. Breslow and Day, 1980). All models included the matching variables age, sex, and region. The Institutional Review Boards of the National Institute of Environmental Health Sciences, New England Medical Center, Brigham and Women's Hospital, SRA-Batelle, and CODA approved the study. All participants gave informed consent.

RESULTS

We considered five haplotypes (AAG, AGG, CGG, CGC, and AGC) and 11 diplotypes (Table I).

TABLE I Associations of ALS with VEGF Diplotype

VEGF diplotype	Cases	Controls	Adjusted	OR (95% CI)
No copies of AAG or AGG^†	23	12	1.0	(referent)
One copy of AAG or AGG^‡	62	23	1.5	(0.5–4.0)
Two copies of AAG or AGG^§	19	4	3.0	(0.7–13.4)

*ORs were adjusted for age (as a continuous variable), sex, and region (Boston or other area of New England).

^†Includes the diplotypes (case/control) CGG/CGG (9/4), CGC/CGC (11/5), CGG/CGC (3/2), and CGC/AGC (0/1).

^‡Includes the diplotypes (case/control) AAG/AGG (13/5), AAG/CGG (9/7), AAG/CGC (22/6), AGG/CGG (7/3) and AGG/CGC (11/2).

^§Includes the diplotypes (case/control) AAG/AAG (18/2) and AGG/AGG (1/2).

We found and increased risk of ALS among individuals homozygous for the AAG or AGG haplotypes (Table I). For individuals with only one copy of either haplotype, the relative risk was increased only slightly. The results were not altered after excluding individuals with a family history of ALS, after excluding individuals for whom haplotypes were inferred, or after additional adjustment for factors previously associated with ALS risk in our data, including cigarette smoking (Kamel et al., 1999) and blood lead levels (Kamel et al., 2002). In addition, we found no meaningful interaction between VEGF haplotype and either smoking or markers of recent or cumulative exposure to lead (blood and bone lead levels, respectively).

DISCUSSION

The pathophysiology of ALS is poorly understood, although genetic mechanisms likely increase susceptibility, and oxidative damage has been hypothesized to play an important role (Cleveland, 1003; Strong & Rosenfeld, 2003). A relationship of VEGF to ALS has been suggested by human, animal, and mechanistic studies (Lambrechts et al., 2003; Cleveland, 2003). Tissue response to hypoxia includes production of transcription factors that activate hypoxia response elements (HREs) in the promoter region of VEGF, thereby increasing VEGF production and secretion (Lambrechts et al., 2003). Mice with a deletion of the HREs developed a motor neuron degeneration syndrome with many similarities to ALS (Oosthuyse et al., 2001). VEGF slowed neurodegeneration in several in vivo and in vitro models (Oosthuyse et al., 2001; Lambrechts et al., 2003), and protected motor neurons from hypoxia and oxidative stress (Oosthuyse et al., 2001).

Most recently, Azzouz and colleagues (2004) found that intramuscular injection of a VEGF-expressing lentiviral vector delayed onset and slowed progression of ALS in mice. In addition, a recent study found low VEGF levels in the cerebrospinal fluid of patients with early stages of ALS (Devos et al., 2004).

A recent study found that, although changes in HRE sequence were not related to ALS, individuals homozygous for one of two haplotypes involving three SNPs in the VEGF upstream promoter/leader sequence (the AAG or AGG haplotype) had decreased serum levels of VEGF and a 1.8-fold increase in ALS risk. (Lambrechts et al., 2003) We found a 3-fold increased risk in homozygotes, providing the first independent confirmation of their findings. However, confidence limits around this estimate included unity and, therefore, the finding should be interpreted cautiously. We found no interaction of VEGF haplotype with lead exposure or smoking, environmental exposures that increase oxidative stress, but the statistical power to examine interaction was low. Although limited by a small sample size, our findings support the view that VEGF haplotype should be examined more closely as a risk factor for ALS (Greenberg & Jin, 2004). The identification of genes involved in ALS susceptibility is an important step in understanding the pathophysiology of neuron degeneration and may help develop effective treatments for afflicted individuals.

We are indebted to Drs. Theodore L. Munsat and Jeremy M. Shefner for their assistance in recruiting ALS patients. The Institutional Review Boards of the National Institute of Environmental Health Sciences, New England Medical Center, Brigham and Women's Hospital, SRA-Batelle, and CODA approved the study. All participants gave informed content.