Purpose: Angiogenesis is a necessary step in tumor growth and metastasis. Vascular endothelial growth factor (VEGF) is a major mediator of breast cancer angiogenesis. Therefore, we investigated the association of polymorphisms in the VEGF gene with breast cancer risk and prognostic characteristics of the tumors in a large case-control study.

Experimental Design: We examined three polymorphisms in the VEGF gene (−2578C/A, −1154G/A, and +936C/T) in 571 familial breast cancer cases from Poland and Germany and −2578C/A, −634G/C, and +936C/T polymorphisms in 974 unselected breast cancer cases from Sweden together with ethnically and geographically selected controls.

Results: None of the polymorphisms or any haplotype was significantly associated with either familial or unselected breast cancers. Our study suggests that the +936C/T polymorphism is unlikely to be associated with breast cancer. We also analyzed the unselected cases for genotypes or haplotypes that associated with tumor characteristics. The −634CC genotype and the −2578/−634 CC haplotype were significantly associated with high tumor aggressiveness (large tumor size and high histologic grade, P < 0.01) and the −2578AA genotype and the −2578/−634 AG haplotype with low histologic grade tumors (P = 0.04). The genotypes and haplotypes were not related with other tumor characteristics such as regional or distant metastasis, stage at diagnosis, or estrogen or progesterone receptor status.

Conclusions: Although none of the polymorphisms studied in the VEGF gene was found to influence susceptibility to breast cancer significantly, some of the VEGF genotypes and haplotypes may influence tumor growth through an altered expression of VEGF and tumor angiogenesis.

Angiogenesis is an important step in the development of cancer and is necessary for primary tumor growth, invasiveness, and metastasis (1). Vascular endothelial growth factor (VEGF) is believed to be important for the process of initiation of angiogenesis and is a major mediator of breast cancer angiogenesis (2). Overexpression of VEGF has been shown in various cancers (3). Several polymorphisms in the VEGF gene have been reported to affect the expression of the gene. The −2578CC, −2549 del/del, −1154GG, and −634CC have been shown associated with a higher VEGF production (47), whereas the +936 T allele has been shown to correlate with lower VEGF plasma levels (8, 9). In addition, the −634G/C polymorphism is located within a potential binding site of the MZF1 transcription factor (10) and the +936C/T polymorphism leads to a loss of a potential AP-4 binding site (8, 9). Recent studies have shown that some VEGF polymorphisms are associated with the development of cancer. The −1154AA genotype, for example, is associated with a decreased prostate cancer risk and less advanced melanomas (11, 12) and the +936 C/T polymorphism with a decreased breast cancer risk (9). In breast cancer, VEGF has been shown to be of prognostic importance (13). These data suggest that the polymorphisms involved in the angiogenic pathway may affect the progression or aggressiveness of the tumor, including breast tumors.

In the present study, we investigated the relationship between genetic polymorphisms in the VEGF gene and the development of breast cancer in patients from Poland, Germany, and Sweden. Five of the polymorphisms were located in the promoter region at positions −2578, −2549, −2489, −2447, and −1154, one in the 5′ untranslated region at position −634 relative to the translation starting site, and the seventh one in the 3′untranslated region at position +936, according to the numbering used by Renner et al. (ref. 8; Fig. 1). From these seven polymorphisms, four were selected for further analysis to evaluate their possible influence on the risk to breast cancer and the prognostic characteristics of the tumors.

Fig. 1.

Structure of the VEGF gene. Polymorphisms included in the present study. The translation starting site is marked by the ATG codon. The polymorphisms within the box are in complete linkage. The −2578, −1154, and +936 polymorphism were studied in the Polish and German familial breast cancer cases and controls and the −2578, −634, and +936 polymorphisms were analyzed in the Swedish unselected breast cancer series.

Fig. 1.

Structure of the VEGF gene. Polymorphisms included in the present study. The translation starting site is marked by the ATG codon. The polymorphisms within the box are in complete linkage. The −2578, −1154, and +936 polymorphism were studied in the Polish and German familial breast cancer cases and controls and the −2578, −634, and +936 polymorphisms were analyzed in the Swedish unselected breast cancer series.

Close modal

Subjects. Totally, 571 familial breast cancer cases (including 412 cases from Poland and 159 cases from Germany) as well as 974 unselected breast cancer cases together with ethnically and geographically selected controls were used in the study. To a large extent, the study populations consisted of Caucasians. The Polish and German cases were incident cases collected during the years 1997 to 2003 according to the criteria described earlier (14) through the Chemotherapy Clinics and the Genetic Counseling (Gliwice, Poland) and the Institute of Human Genetics, University Heidelberg (Germany). About 90% of patients approved participation in the study. The controls were recruited to earlier studies with comparable participation rate. Thirteen percent of the Polish breast cancer cases were BRCA1/2 mutation carriers. All the German cases used in this study were tested BRCA1/2 negative. All the familial cases were unrelated. These familial cases together with controls were analyzed at Karolinska Institute, Huddinge, Sweden. Unselected breast cancer cases together with controls were collected during the time period January 1990 to January 2001 by the Department of Public Health and Clinical Medicine/Nutritional Research in Umeå University, Sweden (15). The samples, including controls, within the Västerbotten intervention cohort, the mammary screening cohort, and the MONICA cohort were collected in a population-based manner. The cancer cases were retrieved by linkage to the regional cancer registry. Samples (n = 268) were taken before breast cancer diagnosis, and 449 were collected after diagnosis, being mainly prevalent cases. Samples (n = 257) were collected in a hospital-based manner from untreated patients referred to the Department of Oncology for newly diagnosed breast cancer; the controls were selected from the Västerbotten intervention cohort. Clinical data for the unselected breast cancer cases were retrieved from the registry managed by the Northern Sweden Breast Cancer Group (Table 1). Altogether, only 12 of 782 breast cancer cases had distant metastasis indicating that even the prevalent cases were recruited relatively shortly after diagnosis, excluding biases due to preferential survival. The unselected cases together with controls were analyzed at Umeå University. After DNA digestion, the coded samples were divided on the plates by randomly mixing cases and controls. The study was approved by the ethical committee of Karolinska Institute Syd.

Table 1.

Characteristics of the Swedish breast cancer samples at diagnosis

CharacteristicBreast cancer patients, n
Age at diagnosis (y), median (range) 54.3 (27.1-75.9) 
Distant metastasis  
    Negative 770 
    Positive 12 
Regional lymph node metastasis  
    Negative 574 
    Positive 237 
Stage at diagnosis  
    0 10 
    I 421 
    II 320 
    III 21 
    IV 13 
Histologic grade  
    Grade 1 142 
    Grade 2 322 
    Grade 3 250 
Estrogen receptors  
    Negative 146 
    Positive 358 
Progesterone receptor  
    Negative 165 
    Positive 275 
Tumor size (mm)  
    In situ 
    <20 510 
    21-50 223 
    >50 19 
CharacteristicBreast cancer patients, n
Age at diagnosis (y), median (range) 54.3 (27.1-75.9) 
Distant metastasis  
    Negative 770 
    Positive 12 
Regional lymph node metastasis  
    Negative 574 
    Positive 237 
Stage at diagnosis  
    0 10 
    I 421 
    II 320 
    III 21 
    IV 13 
Histologic grade  
    Grade 1 142 
    Grade 2 322 
    Grade 3 250 
Estrogen receptors  
    Negative 146 
    Positive 358 
Progesterone receptor  
    Negative 165 
    Positive 275 
Tumor size (mm)  
    In situ 
    <20 510 
    21-50 223 
    >50 19 

PCR amplification. The primer sequences for the polymorphisms in the VEGF gene were designed based on the published Genbank sequences M63971 and AF024710 and are available by the corresponding author. PCR was carried out as described earlier (16).

RFLP analysis. The −2578C/A, as a marker for the completely linked polymorphisms −2578C/A, −2549del/ins 18 bp, −2489C/T, and −2447del G, and +936C/T polymorphisms were analyzed in the familial sample sets using RFLP analysis. The assays were set up after PCR as described (16). PCR products were digested with the MvaI and Hsp92II restriction endonuleases, respectively, using the buffers and temperatures recommended by the manufacturers. The digested PCR products were resolved on a 10% polyacrylamide gel (Bio-Rad, Hercules, CA) and stained with ethidium bromide for visualization under UV light.

ALF express fragment analysis. ALF express fragment analysis was employed to analyze the −1154G/A polymorphism in the VEGF gene. The forward primer was labeled with a Cy5 fluorescence dye. After digestion with MnlI, PCR products were electrophoresed and detected using an automated ALF Express sequencer (Amersham Pharmacia). The results were analyzed using software package Fragment Manager as described earlier (17).

The genotyping results after the RFLP and ALF Express analyses were read by two independent individuals. About 10% of the PCR-RFLP and PCR-ALF assays were randomly repeated and the results were checked for concordance.

Taqman assay. The −2578C/A, −634G/C, +936C/T polymorphisms were analyzed in the Swedish populations by the Umeå Center for Genome Research, Medical and Clinical Genetics, Umeå University, Sweden, using Taqman PCR. The primers and probes were designed using the Assay-by-Design service (Applied Biosystems, Foster City, CA) and the sequences are available by the corresponding author upon request. The assays were carried out using the standard method recommended by Applied Biosystems. The results were automatically read using the ABI PRISM 7900 HT Sequence Detection System.

DNA sequencing. DNA sequencing on three random samples from each genotype for all the polymorphisms was used to confirm the genotypes following PCR-RFLP, PCR-ALF, and Taqman assays as described (16).

Haplotype and linkage disequilibrium analysis. Haplotypes and linkage disequilibrium between the polymorphisms of the VEGF gene were determined based on the genotypes of all individuals participating in the study using a Haploview program (version 2.05, http://www.broad.mit.edu/personal/jcbarret/haploview/).

Statistical analysis. The differences in the genotype and haplotype frequencies of the studied polymorphisms in the breast cancer cases and controls were compared for statistical significance by the Yates corrected χ2 test. The statistical significance for deviations from Hardy-Weinberg equilibrium was tested using the Pearson χ2 test. Odds ratios and 95% confidence intervals were calculated for associations between genotypes and breast cancer and tumor characteristics. The joint analyses were carried out using a Mantel-Haenszel adjustment with each series as a separate stratum. For a polymorphism with a variant allele frequency between 10% and 50%, the study had >90% power to detect a 1.8- to 1.5-fold increase of relative risk in all familial cases and a 1.5- to 1.3-fold increased risk in all cases. All the statistical tests were carried out using Epi Info 2000 software (http://www.cdc.gov/epiinfo).

We examined the effect of seven common polymorphisms in the VEGF gene on breast cancer development (Fig. 1). The polymorphisms −2578C/A, −2549del/ins 18 bp, and −2489C/T were amplified in the same fragment. The −2549del/ins 18 bp polymorphism resulted in an 18-bp larger PCR fragment, and the −2489C/T polymorphism abolished a restriction site for MvaI. The results of the RFLP analysis suggested that the −2549 and −2489 polymorphisms are in complete linkage disequilibrium. This was confirmed by sequencing 20 randomly chosen samples. Sequence analysis also confirmed the previously reported complete linkage between the −2578 and −2549 polymorphisms (18). A novel polymorphism, −2447 del G, which was completely linked to the −2578A allele, was also detected. Thus, the −2578C/A polymorphism was chosen as a marker for all the completely linked polymorphisms in this fragment. The polymorphisms −2578C/A, −1154G/A, and +936C/T were examined in familial breast cancer samples and controls using the RFLP assay. For the unselected breast cancer series, Taqman assay was used. Because the Taqman assay did not work for the −1154G/A polymorphism, we analyzed another functional polymorphism, −634G/C, instead (Fig. 1).

The genotype distribution of all studied polymorphisms followed the Hardy-Weinberg equilibrium in every sample set (P = 0.13-1.00). The genotype and allele distributions among the breast cancer cases and control subjects are shown in Table 2. The number of samples analyzed for each polymorphism was not exactly equal because of unsuccessful amplification of a few samples. No differences in the allele or genotype frequencies between the breast cancer cases and controls were detected in any population. Nor did the joint analysis show any differences between the breast cancer cases and controls (Table 2). The lack of association remained when the data were stratified by age (data not shown).

Table 2.

Allele and genotype distributions of the polymorphisms in the VEGF gene in different populations

Polish
German
Swedish
Joint
Polymorphism/genotypeCases (%)Controls (%)OR* (95% CI)Cases (%)Controls (%)OR* (95% CI)Cases (%)Controls (%)OR* (95% CI)OR* (95% CI)
−2578C/A           
CC 104 (25.3) 106 (25.1) 1.00 44 (28.8) 50 (30.9) 1.00 258 (27.5) 257 (27.3) 1.00 1.00 
CA 195 (47.4) 207 (48.9) 0.96 (0.68-1.36) 75 (49.0) 72 (44.4) 1.18 (0.68-2.06) 449 (47.8) 451 (48.0) 0.99 (0.79-1.24) 1.00 (0.84-1.19) 
AA 112 (27.3) 110 (26.0) 1.04 (0.70-1.54) 34 (22.2) 40 (24.7) 0.97 (0.50-1.86) 232 (24.7) 232 (24.7) 1.00 (0.77-1.29) 1.00 (0.82-1.23) 
A% 51.0 50.5  46.7 46.9  48.6 48.7   
−1154G/A           
GG 173 (42.1) 199 (47.0) 1.00 84 (52.8) 76 (46.6) 1.00    1.00 
GA 189 (46.0) 178 (42.1) 1.22 (0.91-1.65) 58 (36.5) 69 (42.3) 0.76 (0.46-1.25)    1.07 (0.83-1.38) 
AA 49 (11.9) 46 (10.9) 1.23 (0.76-1.97) 17 (10.7) 18 (11.0) 0.85 (0.39-1.89)    1.11 (0.74-1.66) 
A% 34.9 31.9  28.9 32.2      
−634G/C           
GG       488 (52.1) 492 (52.3) 1.00  
GC       363 (38.8) 367 (39.0) 1.00 (0.82-1.21)  
CC       85 (9.1) 82 (8.7) 1.05 (0.74-1.47)  
C%       28.5 28.2   
+936C/T           
CC 298 (72.3) 297 (70.4) 1.00 120 (78.4) 128 (78.5) 1.00 708 (76.6) 720 (77.1) 1.00 1.00 
CT 100 (24.3) 114 (27.0) 0.87 (0.63-1.21) 31 (20.3) 31 (19.0) 1.07 (0.59-1.93) 204 (22.1) 203 (21.7) 1.02 (0.81-1.28) 0.98 (0.82-1.17) 
TT 14 (3.4) 11 (2.6) 1.27 (0.53-3.05) 2 (1.3) 4 (2.5) 0.53 (0.07-3.46) 12 (1.3) 11 (1.2) 1.11 (0.45-2.71) 1.10 (0.62-1.94) 
T% 15.5 16.1  11.4 12.0  12.3 12.0   
Polish
German
Swedish
Joint
Polymorphism/genotypeCases (%)Controls (%)OR* (95% CI)Cases (%)Controls (%)OR* (95% CI)Cases (%)Controls (%)OR* (95% CI)OR* (95% CI)
−2578C/A           
CC 104 (25.3) 106 (25.1) 1.00 44 (28.8) 50 (30.9) 1.00 258 (27.5) 257 (27.3) 1.00 1.00 
CA 195 (47.4) 207 (48.9) 0.96 (0.68-1.36) 75 (49.0) 72 (44.4) 1.18 (0.68-2.06) 449 (47.8) 451 (48.0) 0.99 (0.79-1.24) 1.00 (0.84-1.19) 
AA 112 (27.3) 110 (26.0) 1.04 (0.70-1.54) 34 (22.2) 40 (24.7) 0.97 (0.50-1.86) 232 (24.7) 232 (24.7) 1.00 (0.77-1.29) 1.00 (0.82-1.23) 
A% 51.0 50.5  46.7 46.9  48.6 48.7   
−1154G/A           
GG 173 (42.1) 199 (47.0) 1.00 84 (52.8) 76 (46.6) 1.00    1.00 
GA 189 (46.0) 178 (42.1) 1.22 (0.91-1.65) 58 (36.5) 69 (42.3) 0.76 (0.46-1.25)    1.07 (0.83-1.38) 
AA 49 (11.9) 46 (10.9) 1.23 (0.76-1.97) 17 (10.7) 18 (11.0) 0.85 (0.39-1.89)    1.11 (0.74-1.66) 
A% 34.9 31.9  28.9 32.2      
−634G/C           
GG       488 (52.1) 492 (52.3) 1.00  
GC       363 (38.8) 367 (39.0) 1.00 (0.82-1.21)  
CC       85 (9.1) 82 (8.7) 1.05 (0.74-1.47)  
C%       28.5 28.2   
+936C/T           
CC 298 (72.3) 297 (70.4) 1.00 120 (78.4) 128 (78.5) 1.00 708 (76.6) 720 (77.1) 1.00 1.00 
CT 100 (24.3) 114 (27.0) 0.87 (0.63-1.21) 31 (20.3) 31 (19.0) 1.07 (0.59-1.93) 204 (22.1) 203 (21.7) 1.02 (0.81-1.28) 0.98 (0.82-1.17) 
TT 14 (3.4) 11 (2.6) 1.27 (0.53-3.05) 2 (1.3) 4 (2.5) 0.53 (0.07-3.46) 12 (1.3) 11 (1.2) 1.11 (0.45-2.71) 1.10 (0.62-1.94) 
T% 15.5 16.1  11.4 12.0  12.3 12.0   

Abbreviations: OR, odds ratio; 95% CI, 95% confidence interval.

*

Comparison with the genotype indicated by an odds ratio of 1.00.

Haplotypes for the VEGF gene were determined for the familial and unselected breast cancer cases separately because different polymorphisms were analyzed in these two series. The results revealed that the promoter polymorphisms −2578 and −1154 as well as the −2578 and −634 polymorphisms were in tight linkage disequilibrium (|D′| = 0.95 and 1.00, respectively), consistent with a previous report (11). The +936 polymorphism was not linked to the other polymorphisms (|D′| = 0.02-0.25). Haplotypes were created using genotyping data of the polymorphisms −2578C/A, −1154G/A, and +936C/T in the familial series and the polymorphisms −2578C/A, −634G/C, and +936C/T in the unselected series. Because of the tight linkage between the promoter polymorphisms, haplotypes CAC and CAT were only present among 0.8% of familial cases and controls. Similarly, haplotypes ACC and ACT were missing from the unselected breast cancer series, and haplotype CGT was present only in 0.3% of the samples. The frequencies of the haplotypes at positions −2578/−1154 and −2578/−634 in the cases and controls were similar as reported (11, 19). No significant differences in haplotype frequencies between the breast cancer cases and controls were detected (Fig. 2).

Fig. 2.

The observed frequencies of the haplotypes in the VEGF gene in (A) familial and (B) unselected breast cancer series. Nucleotides are listed in the order −2578C/A, −1154G/A, and +936C/T in (A) and −2578C/A, −634G/C, and +936C/T in (B).

Fig. 2.

The observed frequencies of the haplotypes in the VEGF gene in (A) familial and (B) unselected breast cancer series. Nucleotides are listed in the order −2578C/A, −1154G/A, and +936C/T in (A) and −2578C/A, −634G/C, and +936C/T in (B).

Close modal

When we stratified the unselected breast cancer cases by all tumor characteristics listed in Table 1, two interesting associations emerged (Table 3). Because of missing data, the numbers of cases in Table 3 were less than in Table 2. The −634CC genotype showed statistically significant association to higher tumor aggressiveness (large tumor size and high grade, P < 0.01). The −2578AA genotype was correlated with low-grade tumors (P = 0.04). Because the −2578 and −634 polymorphisms were in tight linkage disequilibrium, whereas the +936 polymorphism was not linked to these polymorphisms, the haplotype analysis was restricted to the −2578C/A and −634G/C polymorphisms (Table 4). The haplotype −2578/−634 CC showed significant association with higher tumor aggressiveness (large tumor size and higher grade, P ≤ 0.005). There was a significant trend to larger tumor size with the increasing copy number of the haplotype (Ptrend = 0.008), and individuals with two copies of the haplotype had both significantly larger and higher-grade tumors (P = 0.005). A less significant association was observed between the haplotype −2578/−634 AG and a lower tumor grade (P = 0.04) and only in carriers of two copies of this haplotype. Carriers of the −634 C allele tended to have more often estrogen receptor–positive tumors than carriers of the GG genotype. Because this effect was linked only to heterozygous carriers, the results are difficult to interpret. There were no indications that the polymorphisms would have any effects on the other tumor characteristics (Table 3). Addition of the information of the +936C/T polymorphism data did not change the results (data not shown).

Table 3.

Associations of the VEGF genotypes with tumor characteristics in the Swedish population

Polymorphism/genotype*
−634G/C−2578C/A
 GG (%) GC (%) CC (%) CC (%) CA (%) AA (%) 
Tumor size (mm)       
    ≤20 271 (53.8) 194 (38.5) 39 (7.7) 138 (27.3) 238 (47.0) 130 (25.7) 
    >20 101 (43.5) 99 (42.7) 32 (13.8) 73 (30.9) 106 (44.9) 57 (24.2) 
    OR (95%Cl) 1.00 1.37 (0.97-1.94) 2.20 (1.27-3.82) 1.00 0.84 (0.58-1.23) 0.83 (0.53-1.29) 
    P  0.08 0.004  0.41 0.44 
Histologic grade       
    Grade 1 79 (57.2) 51 (37.0) 8 (5.8) 28 (20.4) 67 (48.9) 42 (30.7) 
    Grade 2 164 (52.1) 126 (40.0) 25 (7.9) 90 (28.7) 147 (46.8) 77 (24.5) 
    Grade 3 111 (46.2) 96 (40.0) 33 (13.8) 77 (31.4) 114 (46.5) 54 (22.0) 
    Ptrend  0.44 0.009  0.19 0.04 
Stage       
    0 + I 206 (48.8) 174 (41.2) 42 (10.0) 118 (27.8) 208 (49.1) 98 (23.1) 
    II-IV 183 (53.5) 130 (38.0) 29 (8.5) 96 (27.7) 159 (46.0) 91 (26.3) 
    OR (95%Cl) 1.00 0.84 (0.61-1.15) 0.78 (0.45-1.34) 1.00 0.94 (0.66-1.34) 1.14 (0.76-1.72) 
    P  0.30 0.40  0.78 0.58 
Regional lymph node metastasis       
    Negative 281 (50.2) 224 (40.0) 55 (9.8) 158 (28.2) 266 (47.4) 137 (24.4) 
    Positive 114 (49.4) 92 (40.8) 25 (10.8) 65 (28.3) 116 (50.4) 49 (21.3) 
    OR (95%Cl) 1.00 1.01 (0.72-1.42) 1.12 (0.64-1.94) 1.00 1.06 (0.73-1.55) 0.87 (0.55-1.38) 
    P  0.99 0.77  0.82 0.60 
Distant metastasis       
    Negative 377 (50.5) 297 (39.8) 72 (9.7) 211 (28.1) 353 (46.9) 188 (25.0) 
    Positive 8 (72.7) 1 (9.1) 2 (18.2) 5 (45.5) 4 (36.4) 2 (18.2) 
    OR (95%Cl) 1.00 0.16 (0.00-1.20) 1.31 (0.13-6.74) 1.00 0.48 (0.09-2.25) 0.45 (0.04-2.79) 
    P  0.08 0.67  0.31 0.46 
Estrogen receptors       
    Positive 158 (45.0) 153 (43.6) 40 (11.4) 110 (31.3) 162 (46.0) 80 (22.7) 
    Negative 75 (56.0) 45 (33.6) 14 (10.4) 44 (31.9) 58 (42.0) 36 (26.1) 
    OR (95%Cl) 1.00 0.62 (0.39-0.97) 0.74 (0.36-1.50) 1.00 0.90 (0.55-1.46) 1.13 (0.64-1.97) 
    P  0.04 0.46  0.72 0.76 
Progesterone receptors       
    Positive 123 (45.2) 119 (43.8) 30 (11.0) 85 (29.5) 121 (44.5) 66 (24.3) 
    Negative 78 (51.7) 59 (39.1) 14 (9.3) 46 (29.5) 73 (46.8) 37 (23.7) 
    OR (95%Cl) 1.00 0.78 (0.50-1.22) 0.74 (0.35-1.55) 1.00 1.11 (0.68-1.82) 1.04 (0.58-1.84) 
    P  0.30 0.49  0.73 0.99 
Polymorphism/genotype*
−634G/C−2578C/A
 GG (%) GC (%) CC (%) CC (%) CA (%) AA (%) 
Tumor size (mm)       
    ≤20 271 (53.8) 194 (38.5) 39 (7.7) 138 (27.3) 238 (47.0) 130 (25.7) 
    >20 101 (43.5) 99 (42.7) 32 (13.8) 73 (30.9) 106 (44.9) 57 (24.2) 
    OR (95%Cl) 1.00 1.37 (0.97-1.94) 2.20 (1.27-3.82) 1.00 0.84 (0.58-1.23) 0.83 (0.53-1.29) 
    P  0.08 0.004  0.41 0.44 
Histologic grade       
    Grade 1 79 (57.2) 51 (37.0) 8 (5.8) 28 (20.4) 67 (48.9) 42 (30.7) 
    Grade 2 164 (52.1) 126 (40.0) 25 (7.9) 90 (28.7) 147 (46.8) 77 (24.5) 
    Grade 3 111 (46.2) 96 (40.0) 33 (13.8) 77 (31.4) 114 (46.5) 54 (22.0) 
    Ptrend  0.44 0.009  0.19 0.04 
Stage       
    0 + I 206 (48.8) 174 (41.2) 42 (10.0) 118 (27.8) 208 (49.1) 98 (23.1) 
    II-IV 183 (53.5) 130 (38.0) 29 (8.5) 96 (27.7) 159 (46.0) 91 (26.3) 
    OR (95%Cl) 1.00 0.84 (0.61-1.15) 0.78 (0.45-1.34) 1.00 0.94 (0.66-1.34) 1.14 (0.76-1.72) 
    P  0.30 0.40  0.78 0.58 
Regional lymph node metastasis       
    Negative 281 (50.2) 224 (40.0) 55 (9.8) 158 (28.2) 266 (47.4) 137 (24.4) 
    Positive 114 (49.4) 92 (40.8) 25 (10.8) 65 (28.3) 116 (50.4) 49 (21.3) 
    OR (95%Cl) 1.00 1.01 (0.72-1.42) 1.12 (0.64-1.94) 1.00 1.06 (0.73-1.55) 0.87 (0.55-1.38) 
    P  0.99 0.77  0.82 0.60 
Distant metastasis       
    Negative 377 (50.5) 297 (39.8) 72 (9.7) 211 (28.1) 353 (46.9) 188 (25.0) 
    Positive 8 (72.7) 1 (9.1) 2 (18.2) 5 (45.5) 4 (36.4) 2 (18.2) 
    OR (95%Cl) 1.00 0.16 (0.00-1.20) 1.31 (0.13-6.74) 1.00 0.48 (0.09-2.25) 0.45 (0.04-2.79) 
    P  0.08 0.67  0.31 0.46 
Estrogen receptors       
    Positive 158 (45.0) 153 (43.6) 40 (11.4) 110 (31.3) 162 (46.0) 80 (22.7) 
    Negative 75 (56.0) 45 (33.6) 14 (10.4) 44 (31.9) 58 (42.0) 36 (26.1) 
    OR (95%Cl) 1.00 0.62 (0.39-0.97) 0.74 (0.36-1.50) 1.00 0.90 (0.55-1.46) 1.13 (0.64-1.97) 
    P  0.04 0.46  0.72 0.76 
Progesterone receptors       
    Positive 123 (45.2) 119 (43.8) 30 (11.0) 85 (29.5) 121 (44.5) 66 (24.3) 
    Negative 78 (51.7) 59 (39.1) 14 (9.3) 46 (29.5) 73 (46.8) 37 (23.7) 
    OR (95%Cl) 1.00 0.78 (0.50-1.22) 0.74 (0.35-1.55) 1.00 1.11 (0.68-1.82) 1.04 (0.58-1.84) 
    P  0.30 0.49  0.73 0.99 

Abbreviations: OR, odds ratio; 95% CI, 95% confidence interval.

*

Comparison with the genotype indicated by an odds ratio of 1.00.

54% of samples were stage I and 41% were stage II.

Fisher exact test.

Table 4.

Associations of the VEGF haplotypes with tumor characteristics in the Swedish population

Tumor size (mm)
Histologic grade
≤20>20OR for tumor sizePGrade 1Grade 2Grade 3Ptrend
Haplotype         
−2578/−634*         
    CC 263 (26.9) 153 (35.0) 1.47 (1.15-1.89) 0.002 60 (23.4) 166 (27.6) 159 (34.1) 0.005 
    AG 479 (49.0) 196 (45.0) 0.85 (0.67-1.07) 0.18 140 (54.6) 287 (47.6) 208 (44.6) 0.04 
Haplotype in individuals         
No. CC copies in an individual         
    0 264 (54.0) 95 (43.6) 1.00  75 (58.6) 159 (52.8) 107 (45.9)  
    1 187 (38.2) 93 (42.7) 1.38 (0.97-1.97) 0.08 46 (35.9) 118 (39.2) 93 (39.9) 0.33 
    2 38 (7.8) 30 (13.8) 2.19 (1.24-3.86) 0.005 7(5.5) 24 (8.0) 33(14.2) 0.005 
No. AG copies in an individual         
    0 134 (27.4) 71 (32.6) 1.00  26 (20.3) 87 (28.9) 75 (32.2)  
    1 231 (47.2) 98 (45.0) 0.80 (0.54-1.18) 0.28 64 (50.0) 141 (46.8) 108 (46.4) 0.15 
    2 124 (25.4) 49 (22.5) 0.75 (0.47-1.18) 0.23 38 (29.7) 73 (24.3) 50 (21.5) 0.04 
Tumor size (mm)
Histologic grade
≤20>20OR for tumor sizePGrade 1Grade 2Grade 3Ptrend
Haplotype         
−2578/−634*         
    CC 263 (26.9) 153 (35.0) 1.47 (1.15-1.89) 0.002 60 (23.4) 166 (27.6) 159 (34.1) 0.005 
    AG 479 (49.0) 196 (45.0) 0.85 (0.67-1.07) 0.18 140 (54.6) 287 (47.6) 208 (44.6) 0.04 
Haplotype in individuals         
No. CC copies in an individual         
    0 264 (54.0) 95 (43.6) 1.00  75 (58.6) 159 (52.8) 107 (45.9)  
    1 187 (38.2) 93 (42.7) 1.38 (0.97-1.97) 0.08 46 (35.9) 118 (39.2) 93 (39.9) 0.33 
    2 38 (7.8) 30 (13.8) 2.19 (1.24-3.86) 0.005 7(5.5) 24 (8.0) 33(14.2) 0.005 
No. AG copies in an individual         
    0 134 (27.4) 71 (32.6) 1.00  26 (20.3) 87 (28.9) 75 (32.2)  
    1 231 (47.2) 98 (45.0) 0.80 (0.54-1.18) 0.28 64 (50.0) 141 (46.8) 108 (46.4) 0.15 
    2 124 (25.4) 49 (22.5) 0.75 (0.47-1.18) 0.23 38 (29.7) 73 (24.3) 50 (21.5) 0.04 
*

Comparison with all the other haplotypes.

Comparison with the noncarriers of the specific haplotype indicated by an odds ratio of 1.00.

Functional polymorphisms, which have an effect on the regulation of gene expression, can contribute to the differences between individuals in susceptibility to and severity of a disease. The effect may be seen by a polymorphism alone, or in combination with other polymorphisms. Several studies have shown that polymorphisms in the promoter as well as in the 5′ and 3′ untranslated regions of the VEGF gene are associated with the production of the VEGF protein (410). We analyzed the familial series for the −2578, −1154, and +936 polymorphisms using RFLP assay. For the unselected breast cancer series, a Taqman assay was used. Because the Taqman assay did not work for the −1154 polymorphism, we analyzed the −634 polymorphism instead.

The allele and genotype distributions of the four polymorphisms were in close agreement with those previously published for healthy Caucasian individuals (8, 11, 12, 1821). In a previous study among 500 Caucasian breast cancer cases and 500 controls, Krippl et al. have shown a decreased risk of breast cancer in individuals who were +936 T allele carriers (9). However, the genotypes in patients did not follow the Hardy-Weinberg equilibrium. In another study, no association between the +936 polymorphism and risk to breast cancer among 862 cases and 713 controls could be observed (21). Here, we did a large case-control study of 1,489 women with breast cancer, including 565 women with familial breast cancer from Poland and Germany and 924 unselected breast cancer cases from Sweden. The use of familial cases can substantially increase the power of association studies as shown earlier (22, 23). We observed no differences in the allele or genotype frequencies between either the familial or unselected breast cancer case and respective control groups, nor did the joint analyses show any differences between the cases and controls (odds ratio, 0.99; 95% confidence interval, 0.85-1.15; P = 0.93). To our knowledge, no other studies on the effect of the other polymorphisms on the risk of breast cancer have been published. In our study, no significant differences in the allele, genotype or haplotype distribution of the polymorphisms in the VEGF gene between the familial and unselected breast cancer cases and respective controls were detected. Being the largest study thus far, and with one third being familial cases, our study provided strong evidence that the +936 T allele or the other studied polymorphisms do not modify the risk of breast cancer. This result is not surprising, because VEGF, as a key mediator of angiogenesis, is more likely to alter the aggressiveness of the tumor than susceptibility to cancer.

Among the unselected breast cancer cases, we observed a significant correlation between the −634CC genotype and a larger tumor size and a higher histologic grade of the tumors. In addition, the haplotype −2578/−634 CC was associated with more aggressive tumors. In addition, our results showed that the −634GG genotype correlated with less aggressive tumors and the −2578AA genotype and the haplotype −2578/−634 AG with low-grade tumors. Our results are in contrast to the study of Howell et al. who have reported an association between the rare −2578/−1154/−634 CAC haplotype and a less advanced melanoma (11). However, our results are in agreement with the reports regarding the effect of the polymorphisms on VEGF production. The −634CC genotype has been reported to be associated with higher serum levels of VEGF than the GG genotype (7) and the two common haplotypes −2578/−1154/−634 AAG and AGG with a decreased transcription of the VEGF gene and lower circulating levels of VEGF (19). Moreover, the −2578CC genotype and the −2578C allele have been reported to correlate with a higher VEGF production than the A allele in vitro (5, 6).

In summary, the present study investigated polymorphisms in the VEGF gene in a large case-control study. None of the polymorphisms alone or in combination with each other was found to influence the risk of breast cancer, either in the familial or unselected cases. Our study provided evidence that the +936C/T polymorphism is not associated with breast cancer risk. However, some genotypes and haplotypes in the VEGF gene may have an effect on breast tumor growth. Functional studies of the haplotypes and an independent study are needed to confirm our results. The polymorphisms should also be studied in relation to metastasis and survival and whether they influence therapeutic effects.

Grant support: State Committee for Scientific Research grant 6P05A 14221 (E. Grzybowska) and EU grant LSHC-CT-2004-503465.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the Umeå Center for Genome Research and Pia Osterman (Umeå Center for Genome Research, Medical and Clinical Genetics, Umeå University, Sweden) for genotyping the Swedish samples; Åsa Ågren (Department of Public Health and Clinical Medicine/Nutritional Research, Umeå University, Sweden) for her efficiency and skill in keeping track of samples and data; the Northern Sweden Breast Cancer Group for providing the clinical data; Wallenberg Consortium North, Sweden; and Deutsche Krebshilfe headed by Prof. C.R. Bartram for the German samples.

1
Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors.
Nat Med
2003
;
9
:
669
–76.
2
Morabito A, Sarmiento R, Bonginelli P, Gasparini G. Antiangiogenic strategies, compounds, and early clinical results in breast cancer.
Crit Rev Oncol Hematol
2004
;
49
:
91
–107.
3
Nakamura M, Abe Y, Tokunaga T. Pathological significance of vascular endothelial growth factor A isoform expression in human cancer.
Pathol Int
2002
;
52
:
331
–9.
4
Yang B, Cross DF, Ollerenshaw M, Millward BA, Demaine AG. Polymorphisms of the vascular endothelial growth factor and susceptibility to diabetic microvascular complications in patients with type 1 diabetes mellitus.
J Diabetes Complications
2003
;
17
:
1
–6.
5
Mohammadi M, Ollier WE, Hutchinson IV. A functional association study of VEGF gene promoter polymorphisms with VEGF expression by stimulated pbm cells.
Hum Immunol
2003
;
64
:
S125
.
6
Shahbazi M, Fryer AA, Pravica V, et al. Vascular endothelial growth factor gene polymorphisms are associated with acute renal allograft rejection.
J Am Soc Nephrol
2002
;
13
:
260
–4.
7
Awata T, Inoue K, Kurihara S, et al. A common polymorphism in the 5′-untranslated region of the VEGF gene is associated with diabetic retinopathy in type 2 diabetes.
Diabetes
2002
;
51
:
1635
–9.
8
Renner W, Kotschan S, Hoffmann C, Obermayer-Pietsch B, Pilger E. A common 936 C/T mutation in the gene for vascular endothelial growth factor is associated with vascular endothelial growth factor plasma levels.
J Vasc Res
2000
;
37
:
443
–8.
9
Krippl P, Langsenlehner U, Renner W, et al. A common 936 C/T gene polymorphism of vascular endothelial growth factor is associated with decreased breast cancer risk.
Int J Cancer
2003
;
106
:
468
–71.
10
Watson CJ, Webb NJ, Bottomley MJ, Brenchley PE. Identification of polymorphisms within the vascular endothelial growth factor (VEGF) gene: correlation with variation in VEGF protein production.
Cytokine
2000
;
12
:
1232
–5.
11
Howell WM, Bateman AC, Turner SJ, Collins A, Theaker JM. Influence of vascular endothelial growth factor single nucleotide polymorphisms on tumour development in cutaneous malignant melanoma.
Genes Immun
2002
;
3
:
229
–32.
12
McCarron SL, Edwards S, Evans PR, et al. Influence of cytokine gene polymorphisms on the development of prostate cancer.
Cancer Res
2002
;
62
:
3369
–72.
13
Linderholm B, Grankvist K, Wilking N, Johansson M, Tavelin B, Henriksson R. Correlation of vascular endothelial growth factor content with recurrences, survival, and first relapse site in primary node-positive breast carcinoma after adjuvant treatment.
J Clin Oncol
2000
;
18
:
1423
–31.
14
Jin Q, Hemminki K, Grzybowska E, et al. Polymorphisms and haplotype structures in genes for transforming growth factor b1 and its receptors in familial and unselected breast cancers.
Int J Cancer
2004
;
112
:
94
–9.
15
Kaaks R, Lundin E, Rinaldi S, et al. Prospective study of IGF-I, IGF-binding proteins, and breast cancer risk, in northern and southern Sweden.
Cancer Causes Control
2002
;
13
:
307
–16.
16
Forsti A, Jin Q, Grzybowska E, et al. Sex hormone-binding globulin polymorphisms in familial and sporadic breast cancer.
Carcinogenesis
2002
;
23
:
1315
–20.
17
Forsti A, Jin Q, Sundqvist L, Soderberg M, Hemminki K. Use of monozygotic twins in search for breast cancer susceptibility loci.
Twin Res
2001
;
4
:
251
–9.
18
Brogan IJ, Khan N, Isaac K, Hutchinson JA, Pravica V, Hutchinson IV. Novel polymorphisms in the promoter and 5′ UTR regions of the human vascular endothelial growth factor gene.
Hum Immunol
1999
;
60
:
1245
–9.
19
Lambrechts D, Storkebaum E, Morimoto M, et al. VEGF is a modifier of amyotrophic lateral sclerosis in mice and humans and protects motoneurons against ischemic death.
Nat Genet
2003
;
34
:
383
–94.
20
Stevens A, Soden J, Brenchley PE, Ralph S, Ray DW. Haplotype analysis of the polymorphic human vascular endothelial growth factor gene promoter.
Cancer Res
2003
;
63
:
812
–6.
21
Balasubramanian SP, Brown NJ, Reed MW. Role of genetic polymorphisms in tumour angiogenesis.
Br J Cancer
2002
;
87
:
1057
–65.
22
Houlston RS, Peto J. The future of association studies of common cancers.
Hum Genet
2003
;
112
:
434
–5.
23
Antoniou AC, Easton DF. Polygenic inheritance of breast cancer: Implications for design of association studies.
Genet Epidemiol
2003
;
25
:
190
–202.