A common 936 C/T gene polymorphism of vascular endothelial growth factor is associated with decreased breast cancer risk

Abstract

A common 936 C/T polymorphism in the gene for the vascular endothelial growth factor (VEGF) has been associated with VEGF plasma levels. In our case-control study, we investigated the role of this polymorphism for breast cancer risk. VEGF genotype was determined in 500 women with breast cancer and 500 sex- and age-matched healthy control subjects. Carriers of a 936T-allele were more frequent among controls (29.4%) than among patients (17.6%; p = 0.000014). The odds ratio for carriers of a 936T-allele for breast cancer was 0.51 (95% confidence interval 0.38–0.70). Additionally, VEGF plasma levels were determined in 21 nonsmoking post-menopausal controls; carriers of a 936T allele had significantly lower levels (median 23 pg/ml; range 6–50 pg/ml) than noncarriers (37; 21–387; p = 0.034). We conclude that carriers of a VEGF 936T-allele are at decreased risk for breast cancer, this, however, requiring further confirmation in a larger study. © 2003 Wiley-Liss, Inc.

Breast cancer is the most frequently diagnosed cancer in Western Societies, with a lifetime incidence of about 10–13% among women.1, 2. The etiology of breast cancer is still not fully understood. Besides age at menarche and menopause, diet, reproductive history, estrogen administration and genetic factors have been suggested as risk factors.3, 4, 5, 6, 7 Only a small part of familial breast cancer cases can be explained by inherited mutations, the majority being most probably explained by a combination of common low-penetrance gene polymorphisms.8

Tumor growth requires the formation of new blood vessels, a process called angiogenesis. The most important regulator of angiogenesis is vascular endothelial growth factor (VEGF), a dimeric glycoprotein that is overexpressed in several tumor tissues. VEGF plasma levels are highly predictive for tumor growth and survival rate of patients.9, 10 Thus, therapeutic strategies blocking VEGF action can inhibit tumor growth.10

In a study including 23 healthy young men, a common 936 C/T polymorphism (Reference Single Nucleotide Polymorphism ID: NCBI|rs3025039) in the 3′ untranslated region of the VEGF gene has been associated with VEGF plasma levels, showing strongly reduced levels in carriers of a 936-T allele.11 We have therefore hypothesized that this genetic marker may be similarly associated with VEGF plasma levels in women and might influence the risk for breast cancer. The present large case-control study was designed to investigate these hypotheses.

METHODS

Subjects

Between January 2002 and July 2002, a total of 500 female patients with histologically confirmed prevalent breast cancers without synchronous and/or metachronous secondary malignancy were recruited at the Division of Oncology, Department of Internal Medicine, Karl-Franzens University, Graz, Austria. During the recruitment phase, 2 patients were not willing to participate in the study (participation rate 99.6%). For each patient, 1 healthy sex- and age-matched (± 1 years) control subject was included. Controls were selected from 2 Austrian population-based screening studies, the Salzburger Atherosklerose Präventionsprogramm bei Personen mit hohem Infarkt Risiko (SAPHIR, n = 295)12 and the Grazer Diabetes Screening Programm (n = 205). The study was performed according to the Austrian Gene Technology Act and to the guidelines of the Ethical Committee of the Universitätsklinik Graz. Written informed consent was obtained from all participating subjects. All subjects were Caucasian.

Determination of VEGF genotypes and plasma levels

For genetic analyses, genomic DNA was isolated from venous blood by standard methods and stored at 4°C. A 198 bp fragment containing the polymorphic site was amplified by polymerase chain reaction (PCR) using 5′-AAGGAAGAGGAGACTCTGCGC-3′ as forward and 3′-TATGTGGGTGGGTGTGTCTACAGG-3′ as reverse primer The PCR product was digested with restriction endonuclease NlaIII (New England Biolabs, Beverly, MA), fragments were analyzed on 2.0% agarose gels stained with ethidium bromide. The C allele remained uncut, while the T allele was cut into 2 fragments of 114 and 84 bp. For each set of reactions, a negative control containing H₂O instead of DNA to check for contaminations and a positive control (sample with homozygous TT genotype) to check for complete digestion was added. Fifty samples were re-analyzed; results were identical for all samples.

VEGF plasma levels were determined in 21 healthy, nonsmoking postmenopausal women. Analysis was performed using a commercially available enzyme immunoassay (human VEGF Quantikine, R&D Systems) as described previously.11 The assay was specific for VEGF165 and did not detect related molecules, e.g., platelet-derived growth factor or placental growth factor.

Statistics

p-values were calculated with Pearson's chi-square-test and Wilcoxon-Mann-Whitney-test using the exact test option in StatXact 4.0.1 (Cytel Software Corp., Cambridge, MA). Threshold for significance was p < 0.05. Odds ratio (OR) was calculated to estimate the risk for breast cancer.

RESULTS

At the time of diagnoses, patients were between 28 and 84 years of age, with a mean age of 57 ± 11 years. Controls were age-matched to patients (± 1 year), the mean age was 57 ± 11 years with a range of 28–84 years.

VEGF genotype was successfully determined in all patients and controls. Hardy Weinberg equilibrium was observed among genotypes of controls, but not among those of patients (χ² = 4.8, 1 degree of freedom). Carriers of a 936T allele (CT + TT genotypes) were more frequent among controls than among patients (Table I). The odds ratio for carriers of a 936T allele for breast cancer was 0.51 (95% confidence interval 0.38–0.70).

Table I. VEGF 936 C/T Genotype and Allele Frequencies

	Patients	Controls	p1
1 P-values were calculated with Pearson's chi-square-test.
CC, n (%)	412 (82.4)	353 (70.6)	0.000014
CT + TT, n (%)	79 + 9 (15.8 + 1.8)	137 + 10 (27.4 + 2.0)
C allele frequency, %	90.3	84.3	0.000070
T allele frequency, %	9.7	15.7

The potential associations of the VEGF genotype with tumor characteristics are presented in Table II. No correlation was found between the 936 C/T gene polymorphism of VEGF and tumor size, histological grading, lymph node status, estrogen receptor status, progesterone receptor status or age at diagnosis.

Table II. VEGF 936 C/T Genotype and Tumor Characteristics1

		CC	CT + TT	p
1 Numbers are n (%). p-value for age was analyzed by Student's t-test, all other p-values were determined by Pearson's chi-square-test.
Tumor size, n (%)	≤2 cm	210 (52)	43 (51)	0.81
	>2 cm	193 (48)	42 (50)
Histological grade, n (%)	1 and 2	189 (48)	48 (57)	0.12
	3 and 4	207 (52)	36 (43)
Lymph node metastases, n (%)	Negative	183 (46)	43 (51)	0.47
	Positive	217 (54)	42 (49)
Estrogen receptor, n (%)	Positive	303 (75)	63 (76)	1.00
	Negative	100 (25)	20 (24)
Progesteron receptor, n (%)	Positive	259 (65)	55 (66)	0.80
	Negative	141 (35)	28 (34)
Age at diagnosis, years		56.4 ± 11.0	57.1 ± 10.7	0.62

VEGF plasma levels of 21 healthy, nonsmoking postmenopausal women (12 CC, 6 CT and 3 TT) are presented in Figure 1. Levels were significantly lower in 936T allele carriers than in noncarriers.

Figure 1. VEGF plasma levels of controls stratified for VEGF genotype. Boxes indicate the range from lower to upper quartile values, the line inside the box represents the median. Error bars (“whiskers”) indicate range of the data. Carriers of a 936T allele had significantly lower levels (median 23 pg/ml; range 6–50 pg/ml) than noncarriers (37; 21–387; p = 0.034). Homozygous carriers tended to have lower values (9; 6–31) than heterozygotes (28; 20–50), although this difference was not statistically significant (p = 0.17).

DISCUSSION

Carriers of the VEGF 936T allele were less frequent in breast cancer patients than in healthy subjects, indicating that this genetic variant may be protective against breast cancer.

We did not detect a stronger protective effect in homozygous 936T carriers, which may be due to the low number of TT homozygotes (2.0% of patients and 1.8% of controls). The frequency of the 936T allele was 0.16 in controls of the present study, comparable to previous results from the same region (frequency 0.16)11 or Japan (frequency 0.14).13 Genotypes of patients of the present study were not in Hardy Weinberg equilibrium; the reason for this is unknown.

Owing to the retrospective design of the study, a survival bias can not be excluded. Nevertheless, assuming a higher survival rate in carriers of the protective 936T allele, this allele might have been over-represented among breast cancer cases of the present case-control study. This would have led to an underestimation of the true effect of this allele. Thus, in a prospective study an even stronger protective effect of the VEGF 936T allele might be detected, although this remains to be proven. Further studies are required to clarify this matter.

The VEGF 936 C/T polymorphism has previously been associated with lower VEGF plasma levels in healthy young men,11 a similar association was found in female controls of the present study. In contrast, in a recent study from Japan, no relation between this polymorphism and VEGF serum levels was found.13 This could be due to the different ethnic background of study subjects, or the fact that VEGF was measured in serum but not in plasma. Platelets release significant amounts of VEGF during aggregation; thus VEGF serum levels are typically 3 to 6-fold higher than plasma levels.14, 15 Effects of the VEGF genotype on VEGF levels might have been obscured by variations of platelet count or handling of samples before measurement.

The mechanism by which the 936 C/T polymorphism affects VEGF levels is currently not known. The 936-C allele carries a potential binding site for activator protein 4 (AP-4), which is abolished in the 936-T allele.11 AP-4 is a helix-loop-helix transcription factor-enhancing expression of several viral and cellular genes by binding to specific enhancer sites; the loss of this potential binding site could be responsible for decreased VEGF expression by the T-allele.16, 17 The 936 C/T polymorphism could be also in linkage disequilibrium with another functional polymorphism elsewhere. Nevertheless, no linkage disequilibrium between the 936 C/T site and any other known polymorphism has been detected until now.11, 13

The confidence to genetic association studies using single nucleotide polymorphisms (SNPs), such as the polymorphism presented here, has been undermined by a variety of publications showing positive results, which were later followed and refuted by contradictory studies. A list of desirable attributes for genetic case-control studies has therefore been presented recently.18 The present study conforms to a great extent to these standards (Table III). Nevertheless, the observed association of the VEGF genotype with breast cancer could be confounded by other risk factors, e.g., obesity/overweight, which might be linked to VEGF. In the present study analysis of such potential confounding was not possible; our results should therefore be interpreted cautiously until confirmed by other studies. Furthermore, the functionality of the polymorphic site needs to be analyzed by expression studies that use reporter systems.

Table III. Compliance of the Present Study with the “List of Desirable Attributes for Genetic Association Studies” as Suggested by Hegele [18]

Attribute		Fulfillment by present study
Optimized sampling?	+	Cases only eligible with clinically verified breast cancer.
	+	Controls selected from population-based screening studies (minimized selection bias).
	−	No prospective design.
Optimized phenotyping?	+	Breast cancer characteristics evaluated by experienced oncologist according to international standards.
	−	No data on other breast cancer risk factors (e.g., reproductive history, hormone replacement therapy or obesity). No analysis of potential confounding of the association between VEGF genotype and breast cancer risk by other risk factors.
Optimized genotyping?	+	Determination of genotypes blinded to case-control status.
	+	Repeated analyses for 50 samples with identical results.
		Uniform genotyping method for all subjects.
Minimized type 1 error (false positive)?	+	Low p-value for genotype difference (p < 0.0001), “chance finding” very unlikely.
Minimized type 2 error (false negative)?	+	High statistical power (0.98) to detect a relative risk of 0.5 even at low significance level (p = 0.01).
Functional consequences?	+	Repeatedly shown association of VEGF genotype with plasma level.
	−	No expression analysis in a reporter system. No association of VEGF 936 C/T polymorphism with serum levels in a Japanese study [13].
Appropriate interpretation?	+	Cautious interpretation until confirmation of results by other studies.
Replication?	−	No.
Publication?	(+)	Decision to publish results independent of positive or negative associations was made during the planning phase (avoidance of publication bias).

Further studies are needed to investigate the potential pharmacogenetic role of the VEGF 936 C/T polymorphism for antiangiogenesis approaches. Additionally, it might be interesting to analyze the implications of VEGF polymorphism for other malignant diseases.

Acknowledgements

We thank Markus Seitinger for his technical assistance.