Influence of vascular endothelial growth factor single nucleotide polymorphisms on non-small cell lung cancer tumor angiogenesis

Vascular endothelial growth factor (VEGF) plays an important role in tumor angiogenesis. Several studies have reported that genomic VEGF polymorphisms may influence VEGF synthesis. To evaluate the role of VEGF single nucleotide polymorphisms (SNPs), we examined the expression of several angiogenesis-related proteins [VEGF, hypoxia-inducible factor-1α (HIF-1α) and delta-like ligand 4 (Dll4)] and the spread of microvessels in resected non-small cell lung cancer (NSCLC). Blood and tumor tissue from 83 patients with NSCLC were examined for VEGF −460T/C (rs833061) and VEGF +405G/C (rs2010963) SNPs using the SNaPshot method. Immunohistochemical staining was performed to measure protein expression and microvessel density (MVD). VEGF −460T/C and +405G/C SNPs showed no association with VEGF or HIF-1α expression and MVD. Patients with VEGF −460TT and the TC genotype had significantly higher MVD compared to those with the CC genotypes. Furthermore, patients with the VEGF −460TT genotype had significantly higher Dll4 expression compared to those with the TC or CC genotypes, while the VEGF +405G/C SNP displayed no association with Dll4 expression and MVD. These findings indicate that the VEGF −460T/C SNP may have a functional influence on tumor angiogenesis in NSCLC. We hypothesize that VEGF SNPs may influence angiogenesis through Dll4.


Introduction
Angiogenesis plays an important role in tumor progression and metastasis, and vascular endothelial growth factor (VEGF) is a key component. Several studies have demonstrated that VEGF mRNA and protein overexpression are associated with tumor progression and prognosis in non-small cell lung cancer (NSCLC) (1)(2)(3).
Several VEGF single nucleotide polymorphisms (SNPs) have been recently described (4). VEGF is located on chromosome 6p21. 3 and is organized into eight exons and seven introns (5,6). The VEGF -460T/C SNP (rs833061) is located in the promoter region and may influence promoter activity (7). Furthermore, the VEGF +405G/C SNP (rs2010963) is located within the 5'-untranslated region and may affect transcription factor binding affinity (7,8). These two SNPs have been investigated in different types of cancers, and the association of various VEGF SNPs with risk or prognosis of several cancers has been examined (9)(10)(11)(12). Recently, VEGF +405 and -460 SNPs have been found to be significantly associated with risk and survival in NSCLC (13)(14)(15). However, the influence of VEGF SNPs on tumor angiogenesis remains unclear. In this study, we examined whether VEGF -460 and +405 SNPs may influence VEGF expression and microvessel density (MVD) in NSCLC.
Tumor angiogenesis is influenced by a number of proteins. Hypoxia occurs early in tumor development and results in stable binding of hypoxia-inducible factor-1α (HIF-1α) to DNA and the activation of other angiogenic genes, such as VEGF (16,17). Delta-like ligand 4 (Dll4) is a ligand for Notch proteins that is expressed by endothelial cells (18,19) and may be induced by VEGF and HIF-1α (20). It plays an important role in tumor vessel maturation and remodeling (21,22). Therefore, we studied whether these VEGF SNPs were associated with the expression of the angiogenesis-related proteins HIF1α and Dll4.

Patients and methods
Study population. Blood and tumor samples were obtained from 83 patients with NCSLC who underwent surgical resection at the Kawasaki Medical School Hospital between October, 2008 and December, 2010. The patients did not receive radio-or chemotherapy before surgery. This study was approved by the Ethics Committee of the Kawasaki Medical School, and informed consent was obtained from all patients for the use of their tissue specimens.
DNA sequencing. Polymorphism analysis was performed using the ABI PRISM ® 310 Genetic Analyzer, and results were evaluated using GeneMapper ® software, ver. 4.1 (all were from Applied Biosystems).
VEGF staining and scoring. To evaluate VEGF expression, the percentage of positively stained cells and staining intensity were scored as follows: grade 0, negative; grade 1, weak; grade 2, moderate; grade 3, high; and grade 4, very high (23). Grade 0 indicated staining intensity equal to the negative control, grade 3 indicated intensity equal to the positive control, and grade 4 indicated intensity higher than the positive control. Stain intensity in the cell cytoplasm was similarly scored (23).
To determine the percentage of cells with the various staining intensities, the number of immunoreactive cells at each intensity was divided by the total number of tumor cells in three fields at x200 magnification (Fig. 1A). The overall VEGF staining score was calculated as follows: VEGF score = 1 x percentage of grade 1 cells + 2 x percentage of grade 2 cells + 3 x percentage of grade 3 cells + 4 x percentage of grade 4 cells. The score was analyzed as a continuous and a dichotomous variable.
HIF-1α staining and scoring. Tumor cells were scored on the intensity and extent of staining as follows: score 1, tumor cells with absent or weak cytoplasmic reactivity and no nuclear reactivity; score 2, tumor cells with moderate/strong cytoplasmic reactivity with a percentage of tumor cells less than their mean percentage and no nuclear reactivity; score 3, tumor cells with moderate/strong cytoplasmic reactivity with a percentage of tumor cells more than their mean percentage; score 4, tumor cells with clear nuclear reactivity (with or without cytoplasmic reactivity regardless of the intensity) (Fig. 1B). Tumors with scores of 1 and 2 were considered to exhibit low HIF-1α expression, whereas those with scores of 2 and 3 were considered to exhibit high HIF-1α expression (24).
Dll4 staining and scoring. Dll4 expression was considered only in endothelial cells, although recent reports have demonstrated its wide cellular distribution beyond vessels (25,26). To evaluate Dll4 staining in tumor cells ( Fig. 1C and D), the intensity of expression was scored on a semiquantitative scale in three x200 magnification fields. Negative cores were scored as 0, cores with weak expression were scored as 1 and those with moderate/strong expression were scored as 2. High Dll4 expression was defined as a score greater than 1.5 (26).
Microvessel staining and counting. MVD was assessed by counting the number of microvessels stained for CD31. Vessels with a clearly defined lumen or well-defined linear vessel shape and no single endothelial cells were selected for counting. Microvessels were counted in the three x200 magnification fields with the highest density (Fig. 1E), and the mean MVD was calculated (1).
Statistical analysis. Vascular scores were presented as the means ± standard deviation and the difference between the groups was analyzed using the unpaired Student's t-test. The association of VEGF SNPs with clinicopathological parameters and immunostaining results was examined using Chi-squared and Fisher's exact tests, respectively. The level of significance was set at P<0.05. All analyses were performed using SPSS software (version 17.0; SPSS, Chicago, IL, USA).

Results
Clinical characteristics. Characteristics of the patients with NSCLC are summarized in Table I Immunohistochemistry of angiogenesis-related proteins. Forty-two patients (50.6%) exhibited a marked increase in VEGF immunoreactivity of tumor cells. The mean VEGF staining score was 2.79±0.67, and the median score of 2.90 was used to distinguish between low and high VEGF staining. VEGF expression was correlated with HIF1α expression (P=0.003), but not with Dll4 expression (P=0.446) (Table II).
VEGF SNPs and clinicopathological characteristics. For the VEGF +405G/C SNP, 50.6% of patients had the GC genotype, 25.3% had CC and 24.1% had GG. For the VEGF -460T/C SNP, 50.6% had the TT genotype, 38.6% had TC and 10.8% had CC. No significant association was observed between VEGF SNPs and clinicopathological characteristics such as gender, pathological stage, lymphatic invasion, vascular invasion, histological type, and smoking status (Table III).

Discussion
Angiogenesis is important for tumor progression and utilizes several factors, with VEGF being the key factor. Recently, several VEGF SNPs have been identified, and their effect has attracted a great deal of attention. An in vivo study by Stevens et al (7) discovered that VEGF -460/+405 SNPs significantly altered VEGF promoter activity in response to phorbol esters. Recent literature has reported the association of VEGF SNPs with risk or prognosis of various types of cancers (9)(10)(11)(12). A large case-control study in Caucasians demonstrated that male patients with NSCLC and the VEGF +405CC+CG genotype had a higher risk of lung adenocarcinoma, while those with the -460T/+405G/936C haplotype had a reduced risk. (14). The C allele of the VEGF +405G/C SNP significantly improved survival in early-stage NSCLC (13), whereas the -460CC genotype decreased overall survival in advanced-stage NSCLC (15). Other studies have suggested a lower survival rate for the VEGF +405CC genotype in gastric and ovarian cancers (27,28). The reason for these conflicting results is currently unclear, and the influence of VEGF SNPs remains uncertain and controversial. However to date, few studies have focused on the association between VEGF SNPs and VEGF expression. Therefore, we conducted a study with NSCLC patients to examine the functional activity of VEGF SNPs and their possible role in VEGF expression and angiogenesis.
The genotype frequencies for VEGF +405G/C (GG, CC, and GC) and VEGF -460T/C (TT, CC and TC) SNPs in this study were equivalent to previous reports involving Japanese patients (4,15). In our current study, there was no association between VEGF SNPs and VEGF expression. Koukourakis et al (29) reported that VEGF SNPs were associated with VEGF expression in NSCLC tumor cells and tumor angiogenic activity. They discovered that the VEGF -2578CC, +405GG (also referred to as -634GG) -1154AA and GA genotypes were associated with low VEGF expression in 36 patients with NSCLC (29). The vascular density of patients with the VEGF -2578CC and +405GG genotypes was also significantly lower compared to that in patiens with other genotypes. This result is not in agreement with our findings, which may be due to variations in genotype function related to racial differences between the patient groups.
We discovered that patients with the VEGF -460TT and TC genotype had significantly higher MVD compared to those with CC genotypes. In general, as in our study, high VEGF expression is associated with high vascular density. However, there was no association between the VEGF -460T/C SNP and VEGF expression in tumors. Furthermore, even in high VEGF expression cases, the -460TT genotype was associated with significantly higher MVD compared to CC genotype. This result suggested that high MVD in -460TT genotype was not caused by VEGF expression. The VEGF -460TT genotype was associated with significantly higher Dll4 protein expression, which demonstrated a significant association with high MVD. From these results, we concluded that Dll4, induced by the VEGF -460TT genotype, influenced the spread of microves- Mean ± SD P-value     sels. Dll4 is generally upregulated by VEGF, which in turn acts as a negative feedback regulator of VEGF. Our results suggest that VEGF SNPs may influence VEGF downstream signaling to Dll4, although potential mechanisms have not been examined in this study. Dll4 is associated with tumor vessel maturation and remodeling (21,22). Thus, high Dll4 expression should theoretically lead to fewer but larger vessels, and Dll4 overexpression or inhibition may consequently impair tumor angiogenesis. However, further study of this visceral function is warranted.
In conclusion, the VEGF -460T/C SNP may have a functional influence on tumor angiogenesis in NSCLC. Although VEGF SNPs were not associated with VEGF expression in tumor cells, they are considered to regulate the response to Dll4 signaling through functional changes in VEGF.