Genetic variants of vascular endothelial growth factor predict risk and survival of gliomas

Abstract

The vascular endothelial growth factor regulates angiogenesis that is increased in glioma. VEGF polymorphisms are thought to modulate vascular endothelial growth factor plasma levels and therefore may be implicated in glioma risk. We aimed to clarify the role of VEGF and von Willebrand factor polymorphisms in glioma susceptibility and prognosis. A case–control study of 126 glioma patients and 180 cancer-free controls was performed. Using Sequenom MassARRAY platform, 11 VEGF and 1 VWF polymorphisms were genotyped. Unconditional multivariate logistic regression models were used to calculate odds ratios and 95% confidence intervals. The associations between polymorphisms and survival were evaluated using a Cox regression model. Bonferroni’s adjustment was used to correct for multiple testing. The VEGF polymorphism rs833061 was strongly associated with increased risk for glioma (odds ratio = 164.85) and glioblastoma (odds ratio = 155.66), confirmed after Bonferroni correction. Also, the VEGF polymorphisms rs3024994, rs2010963, and particularly the homozygous carriers of rs1005230 were associated with a worse prognosis for glioma and glioblastoma. Our data support a role of VEGF and VWF polymorphisms as glioma biomarkers, with additional potential relevance for molecular stratification of patients for anti-angiogenic therapies.

Keywords

Vascular endothelial growth factor single-nucleotide polymorphisms biomarker risk factor prognostic factor

Introduction

Gliomas are the most frequent primary malignant central nervous system tumors.^1–3 These tumors present a multifactorial etiology, being the result of interactions between environmental exposure and genetic susceptibility.⁴ The only proven exogenous environmental cause of glioma is ionizing radiation.^5,6 Perhaps other environmental causes are involved but remain to be proven. Accordingly, the genetic polymorphisms most frequently studied in the context of brain tumors are located in genes that might influence susceptibility to these tumors in concert with environmental exposures, such as those involved in DNA repair, cell cycle regulation, or immune response.^7–12 In addition, few genome-wide association studies have focused on gliomas identifying five risk loci for glioma mostly related to telomerase.^13–17

Common pathways in glioma biology include growth factor receptor tyrosine kinases and their signaling pathways. Polymorphisms on the epidermal growth factor (EGF) and its receptor, EGFR, have been identified as risk factors for glioma.^17–19 In addition, vascular endothelial growth factor (VEGF) is the principal angiogenic factor in both embryonic development and tumor growth and is highly up-regulated in glioblastoma, and its receptors are over-expressed in tumor vessels.²⁰ Importantly, the anti-angiogenic drug bevacizumab, a humanized monoclonal immunoglobulin (Ig) G1 antibody that binds to and inhibits the biological activity of human VEGFA, is currently being explored in different clinical trials involving glioma and has been used in combination with chemotherapy for the treatment of recurrent glioblastoma,^20–24 which resulted in improved response rates and extended progression-free survival (PFS). Recent studies have shown that some single nucleotide polymorphisms (SNPs) of VEGF/VEGF-related genes could contribute to the modulation of serum VEGF levels, possibly influencing risk and prognosis of solid tumors, including glioma.^25–27 Moreover, VEGF and VEGFR2 genetic polymorphisms could predict not only patient outcome, as shown by Sjostrom et al.,²⁸ but also the response to bevacizumab in glioblastoma patients.²⁹

The von Willebrand factor (VWF) is an essential component of hemostasis at sites of vascular injury.³⁰ The VWF gene is highly polymorphic, and SNPs have been found to influence the levels of VWF in the circulation.^31,32 Moreover, VWF SNPs were also associated with lung cancer risk; however, susceptibility was related with patients’ origin.^33,34

Given the important role of VEGF signaling in gliomas and the influence tumor risk and prognosis of particular VEGF/VEGFR polymorphisms in different tumor types, we investigated the relevance of 11 genetic polymorphisms in the 6p12 (VEGFA) and 6p21 (VWF) loci on glioma risk and patient survival.

Materials and methods

Study population

In this case–control study, we enrolled 126 glioma samples from Hospital S. João, Porto, and Hospital de Braga, Braga, Portugal. Tumors were classified according to the World Health Organization (WHO) criteria (Table 1).¹ All patients above 18 years of age, with diagnosis of glioma, treated between 2009 and 2010, and that by informed consent accepted to participate in the study, were included. The patients with different diagnosis or that refused to sign the informed consent were excluded. Complete treatment data were retrieved for 90 patients. Of those, 80 patients underwent surgery, 9 stereotactic biopsy, 77 radiotherapy, 71 chemotherapy (65 underwent Stupp protocol,³⁵ 7 surgery only, 8 surgery+radiotherapy, 5 surgery+radiotherapy+chemotherapy (other than temozolomide), 2 surgery+chemotherapy, and 2 biopsy only), and one additional patient refused treatment. The 180 control samples were selected from cancer-free blood donors at Hospital de Braga, to match for gender and adjust for age in the statistical analysis (Table 1). All patients and controls had a Caucasian background and come from the same region in the north of Portugal. The follow-up information was obtained by consulting clinical records. The control group had no follow-up information. Signed informed consent was obtained from each participant. Local institutional ethics committees of Hospital S. João and Hospital de Braga approved the procedures followed in this study.

Table 1.

Clinico-pathological features of gliomas and controls.

Groups (WHO grade)	N	Age, years (median ± SD)	p value^a	M/F ratio	p value^a	PFS, months (median)	OS, months (median)
Controls	180	35–85 (53.0 ± 10.8)	p = 0.003	1.69	p = 0.768	n.a.	n.a.
Gliomas (2–4)	126	20–83 (58.0 ± 12.7)	p = 0.003	1.63	p = 0.768	9 (1–48)	16.5 (1–52)
Astrocytic tumors (2–4)	101	22–83 (59.0 ± 11.9)		1.97		9 (1–40)	16 (1–52)
Diffuse astrocytoma (2)	5	35–67 (41.0 ± 13.5)		0.67		15.5 (9–22)	15.5 (9–22)
Anaplastic astrocytoma (3)	6	32–62 (47.0 ± 12.2)		1.0		36	48
Glioblastoma (4)	86	22–83 (59.5 ± 11.6)		2.07		9 (1–40)	15.5 (1–52)
Gliosarcoma (4)	4	50–72 (61.5 ± 9.1)		All M		5.5 (3–10)	22.5 (7–31)
Oligodendroglioma (2–3)	25	20–78 (50.0 ± 14.1)		0.78		9 (3–48)	23 (6–48)
Oligodendroglioma (2)	7	20–58 (45 ± 13.4)		0.4		48	40.5 (33–48)
Anaplastic oligodendroglioma (3)	18	23–78 (53 ± 14.0)		1.0		8.5 (3–48)	22 (6–48)

WHO: World Health Organization; n.a.: not applicable; SD: standard deviation; M: male; F: female; OS: overall survival; PFS: progression-free survival.

p value for independent samples t-test, for testing differences in age and M/F ratio between controls and glioma.

Genotyping

DNA was extracted from the leukocytes of blood samples using the commercially available Citogene® Blood Kit according to the manufacturer’s recommendations. Genotyping of the allele-specific primer extension products, which were generated from amplified DNA sequences, was performed using the Sequenom MassARRAY iPLEX Gold platform (Sequenom, San Diego, CA) at the Instituto Gulbenkian de Ciência (IGC), Lisbon, Portugal. Primers were designed using MassARRAY Assay Design 3.1 software (Sequenom), and genotyping was performed in a blinded manner regarding sample status (i.e. case or control subjects). The genotyping quality was assessed by duplicate analysis of 10% of the samples, which demonstrated a 100% agreement rate.

Statistical analysis

Independent samples t-test, χ² test, and nonparametric Wilcoxon–Mann–Whitney test were used to compare the frequency distribution of age, sex, and genotypes of VEGF and VWF and the allele distribution among the cases and controls. Moreover, the χ² test was used to verify whether the observed allele distribution in the control group was in Hardy–Weinberg equilibrium (HWE). The odds ratio (OR) and 95% confidence intervals (95% CIs) for the effect of the polymorphic variants on the risk for glioma were estimated by multivariate logistic regression analysis, adjusted for gender and age as a continuous variable. Bonferroni’s adjustment was used to correct for multiple testing. Associations between polymorphic variants and patient survival were assessed using a multivariate Cox regression model adjusted for patient gender and age. All statistical tests were two-sided, and significance was considered for p < 0.05. Data analysis was performed using IBM^® SPSS Statistics, version 20.0.

Results

Our cohort of samples included 126 glioma samples (WHO grade II–IV) and 180 cancer-free controls (Table 1). All studied variants were in HWE, apart from the polymorphism rs3025035, for which it was not possible to calculate due to the absence of the homozygous recessive genotype.

Glioma risk assessment

In order to explore the associations between VEGF and VWF variants and risk for glioma, we used an unconditional multivariate logistic regression analysis adjusted for potential confounding variables (SNP genotypes, patient age, and sex).

We could successfully include on the multivariate logistic analysis 107 cases and 143 controls. The frequencies of each genotype in controls and cases are presented in Table 2 for VEGF and VWF. The most frequent allele was used to establish the reference genotype. The control group was compared with all glioma cases (WHO grades 2–4), and also with glioblastoma (WHO grade 4), and oligodendroglioma (WHO grades 2–3; Table 2), since these were the most frequent subtypes in our series (Table 1).

Table 2.

Multivariate logistic regression analysis of associations between VEGF/VWF polymorphisms loci and risk for glioma groups.

Polymorphism	Control	Glioma (grades 2–4)	OR (95% CI)	p value^a	p value^b	Glioblastoma (grade 4)	OR (95% CI)	p value^a	p value^b	Oligodendroglioma (grades 2–3)	OR (95% CI)	p value^a	p value^b
rs3025035 (VEGF int7 C/T)
CC	116	91	–			63	–			16	–
CT	27	16	1.52 (0.61–3.77)	0.369	–	10	1.23 (0.38–3.98)	0.727	–	4	2.04 (0.47–8.89)	0.340	–
TT	0	0	–			0	–			0	–
CT+TT	27	16	1.23 (0.52–2.93)	0.637	–	10	1.08 (0.36–3.24)	0.895	–	4	1.45 (0.36–5.76)	0.599	–
rs3025039 (VEGF +936 C/T)
CC	111	82	–			55	–			15	–
CT	29	21	–			16	–			4	–
TT	3	4	2.96 (0.30–29.6)	0.357	–	2	2.73 (0.20–36.87)	0.450	–	1	3.06 (0.16–60.4)	0.462	–
CT+TT	32	25	–			18	–			5	–
rs3025040 (VEGF 3′UTR T/C)
TT	112	81	–			54	–			15	–
TC	28	22	–			17	–			4	–
CC	3	4	–			2	–			1	–
TC+CC	31	26	–			19	–			5	–
rs1005230 (VEGF −2489 C/T)
CC	46	29	–			21	–			2	–
CT	73	51	1.09 (0.06−21.44)	0.957	–	34	2.32 (0.10–53.64)	0.600	–	11	–
TT	24	27	–			18	–			7	–
CT+TT	97	78	0.94 (0.05–16.71)	0.965	–	52	1.80 (0.10–33.75)	0.695	–	18	–
rs699947 (VEGFA –2578 C/A)
CC	45	27	–			19	–			2	–
CA	74	53	0.03 (0.001–1.54)	0.082	0.984	36	0.02 (0.00–1.72)	0.087	–	11	–
AA	24	27	–			18	–			7	–
CA+AA	98	80	0.04 (0.001–1.52)	0.084	–	54	0.03 (0.001–1.36)	0.072	0.864	18	–
rs833061 (VEGF –460 T/C)
TT	41	3	–			3	–			0	–
TC	74	77	164.85 (24.72–1099.54)	<0.001	<0.012	52	155.66 (17.29–1401.69)	<0.001	<0.012	13	–
CC	28	27	–			18	–			7	–
TC+CC	102	104	116.54 (20.59–659.69)	<0.001	<0.012	70	87.73 (12.96–593.67)	<0.001	<0.012	20	–
rs833070 (VEGF int2 G/A)
GG	46	28	–			20	–			2	–
GA	73	52	4.01 (0.12–130.16)	0.435	–	35	3.49 (0.08–156.58)	0.520	–	11	–
AA	24	27	–			18	–			7	–
GA+AA	97	79	3.55 (0.13–97.15)	0.454	–	53	2.82 (0.09–90.97)	0.559	–	18	–
rs3025010 (VEGF int5 T/C)
TT	60	42	–			31	–			3	–
TC	72	48	0.57 (0.23–1.41)	0.224	–	31	0.49 (0.17–1.39)	0.181	–	12	0.97 (0.16–5.88)	0.970	–
CC	11	17	1.67 (0.37–7.63)	0.508	–	11	2.97 (0.49–17.88)	0.236	–	5	2.58 (0.18–36.7)	0.484	–
TC+CC	83	65	0.59 (0.25–1.43)	0.243	–	42	0.54 (0.20–1.46)	0.222	–	17	1.23 (0.21–7.35)	0.818	–
rs3024994 (VEGF int2 C/T)
CC	131	93	–			63	–			18	–
CT	11	14	4.86 (1.29–18.35)	0.020	0.24	10	6.36 (1.31–30.83)	0.022	0.264	2	1.46 (0.13–16.4)	0.757	–
TT	1	0	–			0	–			0	–
CT+TT	12	14	2.90 (0.87–9.66)	0.083	0.996	10	3.06 (0.80–11.70)	0.102	–	2	0.98 (0.14–6.98)	0.987	–
rs25648 (VEGF –7 C/T)
CC	107	73	–			55	–			8	–
CT	35	31	1.12 (0.52–2.45)	0.770	–	16	0.64 (0.24–1.70)	0.373	–	11	2.82 (0.81–9.81)	0.103	–
TT	1	3	–			2	–			1	–
CT+TT	36	34	1.63 (0.79–3.37)	0.191	–	18	1.01 (0.42–2.43)	0.987	–	12	3.69 (1.11–12.3)	0.033	0.396
rs2010963 (VEGF−634 G/C)
GG	58	53	–			36	–			11	–
GC	71	43	0.96 (0.41–2.25)	0.928	–	28	0.86 (0.31–2.41)	0.771	–	9	1.23 (0.26–5.86)	0.799	–
CC	14	11	1.82 (0.23–14.37)	0.572	–	9	4.09 (0.39–43.41)	0.243	–	0	–
GC+CC	85	54	0.73 (0.38–1.39)	0.338		37	0.70 (0.33–1.50)	0.357	–	9	0.69 (0.23–2.09)	0.514	–
rs9295740 (VWF G/A)
GG	94	65	–			41	–			15	–
GA	40	41	1.62 (0.82–3.22)	0.166	–	31	2.24 (1.00–4.98)	0.049	0.588	5	0.73 (0.18–2.92)	0.652	–
AA	9	1	0.05 (0.002–1.50)	0.084	–	1	0.06 (0.00–2.92)	0.153	–	0	–
GA+AA	49	42	1.51 (0.79–2.88)	0.213	–	32	1.88 (0.89–3.96)	0.097	–	5	0.92 (0.27–3.22)	0.900	–

VEGF: vascular endothelial growth factor; OR (95% CI): odds ratio with 95% confidence interval.

Bold-faced values indicate significant difference at 5% level.

p value without correction.

p value after Bonferroni correction for multiple comparisons.

We found a significant association between the VEGF SNPs heterozygous rs833061 and risk for glioma and glioblastoma (OR = 164.85, 95% CI = 24.72–1099.54 and OR = 155.66, 95% CI = 17.29–1401.69, respectively, p < 0.001 and p < 0.012 after Bonferroni correction, Table 2), as well as the variant carriers of this SNP had an increased glioma and glioblastoma risk (OR = 116.54, 95% CI = 20.59–659.69 and OR = 87.73, 95% CI = 12.96–593.67, respectively, p < 0.001 and p < 0.012 after Bonferroni correction, TT vs TC+CC). Also, the heterozygous rs3024994 was associated with increased risk for glioma (OR = 4.86, 95% CI = 1.29–18.35, p = 0.020, Table 2) and glioblastoma (OR = 6.36, 95% CI = 1.31–30.83, p = 0.022, Table 2), and the rs25648 variant carriers demonstrated an increased risk for oligodendroglioma (OR = 3.69, 95% CI = 1.11–12.26, p = 0.033, Table 2); however, the significance of these results was not confirmed after Bonferroni correction (p > 0.05).

Regarding the VWF rs9295740 SNP, we observed a significant association between the heterozygous genotype and risk for glioblastoma (OR = 2.24, 95% CI = 1.00–4.98, p = 0.049, Table 2); however, the significance of this result was not confirmed after Bonferroni correction (p > 0.05).

No statistically significant associations were found between the VEGF SNPs rs3025035, rs3025039, rs3025040, rs1005230, rs699947, rs833070, rs3025010, and rs2010963 variants and risk for glioma, glioblastoma, or oligodendroglioma (Table 2; all p values > 0.05).

Glioma prognostic assessment

In a subset of glioma patients, follow-up data were also available (n = 57). The PFS of the samples was 9 months (1–48 months), and the overall survival (OS) was 16.5 months (1–52 months; Table 1). We interrogate the associations between each VEGF variant, VWF, and PFS or OS by a multivariate Cox proportional hazard model, adjusted for patient age and sex (Table 3). The heterozygous rs3025035 was significantly associated with improved PFS of glioma and glioblastoma (hazard ratio (HR) = 0.15, 95% CI = 0.03–0.73, p = 0.019 and HR = 0.11, 95% CI = 0.02–0.72, p = 0.021, respectively, Table 3). Improved PFS for glioma was also seen for the heterozygous rs3025010 and for the variant carriers of this SNP (HR = 0.23, 95% CI = 0.08–0.70, p = 0.009 and HR = 0.29, 95% CI = 0.11–0.78, p = 0.015, respectively, Table 3). On the other hand, glioma patients’ carriers of rs833061 variant (HR = 12.59, 95% CI = 1.19–38.55, p = 0.035, Table 3) and heterozygous or variant carriers of rs3024994 (HR = 6.52, 95% CI = 1.38–30.82, p = 0.018 and HR = 5.08, 95% CI = 1.30–19.81, p = 0.019, respectively, for OS and HR = 16.33, 95% CI = 2.79–95.43, p = 0.002 and HR = 6.17, 95% CI = 1.51–25.15, p = 0.011, respectively, for PFS, Table 3) had a decreased survival. However, after Bonferroni correction, only the heterozygosity of rs3024994 was found to be significantly associated with decreased survival (p = 0.024, after correction).

Table 3.

Multivariate Cox proportional hazard model analysis of associations between VEGF/VWF polymorphisms and glioma patient’s survival (adjusted for patient’s age and sex).

Polymorphism	Glioma (grades 2–4)								Glioblastoma (grade 4)
Polymorphism	N	OS HR (95% CI)	p value^a	p value^b	N	PFS HR (95% CI)	p value^a	p value^b	N	OS HR (95% CI)	p value^a	p value^b	N	PFS HR (95% CI)	p value^a	p value^b
rs3025035 (VEGF int7 C/T)
CC	47	–			45	–			33	–			31	–
CT	10	0.61 (0.17–2.20)	0.448	–	7	0.15 (0.03–0.73)	0.019	0.228	9	0.61 (0.14–2.71)	0.517	–	7	0.11 (0.02–0.72)	0.021	0.252
TT	0	–			0	–			0	–			0	–
CT+TT	10	0.91 (0.30–2.74)	0.862	–	7	0.43 (0.12–1.64)	0.219	–	9	0.90 (0.29–2.83)	0.855	–	7	0.48 (0.12–1.99)	0.314	–
rs3025039 (VEGF +936 C/T)
CC	46	–			42	–			33	–			30	–
CT	9	0.61 (0.21–1.78)	0.369	–	8	0.61 (0.21–1.73)	0.352	–	8	0.78 (0.24–2.56)	0.682	–	7	0.40 (0.11–1.54)	0.183	–
TT	2	0.60 (0.11–3.46)	0.570	–	2	1.99 (0.32–12.20)	0.459	–	1	15.94 (0.72–353.16)	0.080	0.960	1	0.67 (0.02–22.32)	0.822	–
CT+TT	11	0.63 (0.29–1.39)	0.257	–	10	0.82 (0.35–1.93)	0.651	–	9	0.54 (0.21–1.39)	0.198	–	8	0.54 (0.19–1.52)	0.242	–
rs3025040 (VEGF 3′UTR T/C)
TT	46	–			42	–			33	–			30	–
TC	9	–			8	–			8	–			7	–
CC	2	–			2	–			1	–			1	–
TC+CC	11	–			10	–			9	–			8	–
rs1005230 (VEGF –2489 C/T)
CC	19	–			18	–			14	–			13	–
CT	23	0.22 (0.01–4.78)	0.334	–	20	0.29 (0.01–6.44)	0.436	–	16	0.02 (0.00–0.75)	0.035	0.420	14	0.21 (0.01–8.07)	0.397	–
TT	15	85.07 (4.38–1654.35)	0.003	0.036	14	65.72 (3.00–1439.81)	0.008	0.096	12	249.48 (8.84–7040.50)	<0.001	<0.012	11	672.46 (12.37–36,562.04)	0.001	<0.012
CT+TT	38	0.78 (0.06–11.06)	0.856	–	34	0.74 (0.05–12.03)	0.834	–	28	0.35 (0.02–5.94)	0.464	–	25	0.63 (0.03–13.60)	0.770	–
rs699947 (VEGFA –2578 C/A)
CC	18	–			17	–			13	–			12	–
CA	24	9.77 (0.41–234.92)	0.160	–	21	15.81 (0.67–374.43)	0.087	–	17	40.31 (1.07–1515.77)	0.046	0.552	15	74.74 (1.55–3594.07)	0.029	0.348
AA	15	–			14	–			12	–			11	–
CA+AA	39	2.80 (0.20–38.55)	0.442	–	35	2.84 (0.20–39.43)	0.437	–	29	5.44 (0.32–92.51)	0.241	–	26	4.74 (0.28–81.66)	0.284	–
rs833061 (VEGF –460 T/C)
TT	3	–			2	–			3	–			2	–
TC	39	9.24 (0.88–96.76)	0.063	0.756	36	7.56 (0.79–72.60)	0.080	0.960	27	11.94 (1.04–137.51)	0.047	0.564	25	16.83 (1.13–250.61)	0.040	0.480
CC	15	–			14	–			12	–			11	–
TC+CC	54	12.59 (1.19–38.55)	0.035	0.420	50	5.41 (0.59–49.66)	0.136	–	39	16.08 (1.35–191.51)	0.028	0.336	36	7.85 (0.68–90.43)	0.098	–
rs833070 (VEGF int2 G/A)
GG	18	–			17	–			13	–			12	–
GA	24	–			21	–			17	–			15	–
AA	15	–			14	–			12	–			11	–
GA+AA	39	–			35	–			29	–			26	–
rs3025010 (VEGF int5 T/C)
TT	24	–			22	–			18	–			16	–
TC	24	0.48 (0.15–1.60)	0.233	–	21	0.23 (0.08–0.70)	0.009	0.108	17	0.95 (0.26–3.45)	0.936	–	15	0.26 (0.06–1.06)	0.060	0.720
CC	9	0.22 (0.03–1.78)	0.156	–	9	0.32 (0.03–3.55)	0.350	–	7	0.04 (0.002–0.63)	0.022	0.264	7	0.45 (0.01–14.38)	0.653	–
TC+CC	33	0.53 (0.17–1.64)	0.268	–	30	0.29 (0.11–0.78)	0.015	0.180	24	0.91 (0.26–3.17)	0.880	–	22	0.40 (0.13–1.23)	0.108	–
rs3024994 (VEGF int2 C/T
CC	51	–			47	–			37	–			34	–
CT	6	6.52 (1.38–30.82)	0.018	0.216	5	16.33 (2.79–95.43)	0.002	0.024	5	7.89 (1.22–51.19)	0.030	0.360	4	176.32 (8.58–3622.74)	<0.001	< 0.012
TT	0	–			0	–			0	–			0	–
CT+TT	6	5.08 (1.30–19.81)	0.019	0.228	5	6.17 (1.51–25.15)	0.011	0.132	5	6.05 (1.08–33.94)	0.041	0.492	4	16.84 (2.60–109.26)	0.003	0.036
rs25648 (VEGF –7 C/T)
CC	43	–			41	–			32	–			30	–
CT	12	1.94 (0.48–7.85)	0.351	–	9	0.68 (0.13–3.55)	0.649	–	8	5.36 (1.00–28.89)	0.051	0.612	6	0.22 (0.01–3.88)	0.298	–
TT	2	0.94 (0.12–7.70)	0.955	–	2	0.76 (0.07–8.31)	0.820	–	2	6.25 (0.41–95.90)	0.188	–	2	0.48 (0.02–15.70)	0.680	–
CT+TT	14	0.96 (0.33–2.81)	0.941	–	11	0.89 (0.30–2.69)	0.842	–	10	0.88 (0.25–3.06)	0.844	–	8	0.54 (0.13–2.20)	0.388	–
rs2010963 (VEGF−634 G/C)
GG	25	–			24	–			20	–			19	–
GC	25	2.12 (0.64–6.99)	0.216	–	21	0.77 (0.21–2.79)	0.686	–	16	11.16 (2.12–58.79)	0.004	0.048	13	0.67 (0.08–5.59)	0.709	–
CC	7	2.77 (0.49–15.68)	0.248	–	7	5.35 (0.89–32.22)	0.067	0.804	6	7.65 (0.82–71.88)	0.075	0.900	6	20.10 (1.45–279.11)	0.025	0.300
GC+CC	32	1.03 (0.48–2.22)	0.932	–	28	0.74 (0.32–1.72)	0.483	–	22	1.87 (0.72–4.82)	0.197	–	19	0.84 (0.29–2.43)	0.749	–
rs9295740 (VWF G/A)
GG	33	–			28	–			24	–			20	–
GA	24	0.68 (0.30–1.60)	0.363	–	24	1.16 (0.50–2.69)	0.735	–	18	0.62 (0.23–1.68)	0.348	–	18	1.41 (0.46–4.26)	0.546	–
AA	0	–			0	–			0	–			0	–
GA+AA	24	0.81 (0.38–1.73)	0.592	–	24	0.96 (0.45–2.04)	0.923	–	18	0.79 (0.32–1.95)	0.605	–	18	0.96 (0.39–2.35)	0.930	–

HR (95% CI): hazard ratio with 95% confidence interval; OS: overall survival; PFS: progression-free survival.

Bold-faced values indicate significant difference at 5% level.

p value without correction.

p value after Bonferroni correction for multiple comparisons.

The rs1005230 homozygous carriers presented a decreased OS and PFS both for glioma (OS: HR = 85.07, 95% CI = 4.38–1654.35, p = 0.003 and p = 0.036 after Bonferroni correction and PFS: HR = 65.72, 95% CI = 3.00–1439.81, p = 0.008 and p = 0.096 after Bonferroni correction, Table 3) and glioblastoma patients (OS: HR = 249.48, 95% CI = 8.84–7040.50, Figure 1(a) and PFS: HR = 672.46, 95% CI = 12.37–36,562.04, p < 0.001 and p < 0.012 after Bonferroni correction, Table 3). The heterozygous genotype, has shown to confer an increased OS for glioblastoma patients (HR = 0.02, 95% CI = 0.00–0.75, p = 0.035, Table 3); however, not confirmed after Bonferroni correction.

Figure 1.

Survival graphs of significant associations between overall survival of glioblastoma patients and VEGF SNPs. Multivariate Cox proportional hazard model regression with respect to glioblastoma patient survival of (a) rs1005230, hazard ratio (HR) = 249.48; 95% confidence interval (CI) = 8.84–7040.50, p < 0.001, for TT genotype and (b) rs2010963, HR = 11.16, 95% CI = 2.12–58.79, p = 0.004.

In addition, when only glioblastoma patients were analyzed, we further observed a decreased survival for rs699947 heterozygous (OS: HR = 40.31, 95% CI = 1.07–1515.77, p = 0.046 and PFS: HR = 74.74, 95% CI = 1.55–3594.07, p = 0.029, Table 3), rs833061 heterozygous or variant carriers (OS: HR = 11.94, 95% CI = 1.04–137.51, p = 0.047 and HR = 16.08, 95% CI = 1.35–191.51, p = 0.028, respectively, and PFS: HR = 16.83, 95% CI = 1.13–250.61, p = 0.040, Table 3); however, the significance of these results was not confirmed after Bonferroni correction (p > 0.05). The rs3024994 heterozygous or variant carriers (OS: HR = 7.89, 95% CI = 1.22–51.19, p = 0.03 and HR = 6.05, 95% CI = 1.08–33.94, p = 0.041, respectively, and PFS: HR = 176.32, 95% CI = 8.58–3622.74, p < 0.001 and p < 0.012 after Bonferroni correction, and HR = 16.84, 95% CI = 2.60–109.26, p = 0.003 and p = 0.036 after Bonferroni correction, Table 3) and rs2010963 heterozygous and variant homozygous (OS: HR = 11.16, 95% CI = 2.12–58.79, p = 0.004 and p = 0.048 after Bonferroni correction, Figure 1(b) and PFS: HR = 20.10, 95% CI = 1.45–279.11, p = 0.025 and p = 0.300 after Bonferroni correction, respectively, Table 3). Finally, rs3025010 variant homozygous carriers had an improved OS for glioblastoma (HR = 0.04, 95% CI = 0.002–0.63, p = 0.022 and p = 0.264 after Bonferroni correction, Table 3).

Discussion

VEGF is the major proangiogenic factor involved in tumor growth and progression, including glioblastoma, being over-expressed in tumor vessels.²⁰ VEGF–VEGFR signaling pathway stimulates not only angiogenic signals but also the secretion of the VWF from endothelial cells, and VWF was described as having a potential role on angiogenesis.^36–38 We have studied the contribution of 11 VEGF and 1 VWF polymorphisms to the risk and outcome of glioma and glioblastoma in the Portuguese population. These polymorphisms were selected based on some of the several studies that reported that some SNPs in the VEGF and VWF loci may conditionate serum VEGF expression modulation and therefore confer tumor risk.^{25–27,34,39} These polymorphisms were reported to modify VEGFA serum levels, in particular rs3025039 and rs2010963, and to modulate risk, aggressiveness prognosis of different solid tumors, specifically rs3025035, rs3025040, rs699947, rs833061, rs833070, rs3025010, and rs3024994.^34,40–44 Moreover, anti-angiogenic therapies are currently in clinical use for glioma patients, and allelic variations of VEGF and VEGFR genes have being associated with glioma patient responses to those therapies.²⁸

We found that, in our population, the VEGF rs833061 SNP conferred an increased risk for glioma and glioblastoma. Strikingly, rs833061 T/C heterozygotes were over 150 times as likely to develop glioblastoma (OR = 155.66, 95% CI = 17.29–1401.69) or glioma (WHO grade 2–4; OR = 164.85, 95% CI = 24.72–1099.54). Although the physiological or pathological function of the SNP rs833061 in glioblastoma is still unknown, this SNP has already shown to be related to increased PFS in response to bevacizumab in combination with sorafenib in recurrent glioblastoma patients.²⁹

The rs3024994 C/T genotype presented an OR = 4.86 (95%CI = 1.29–18.35) for glioma and OR = 6.36 (95% CI = 1.31–30.83) for glioblastoma and also the VWF rs9295749 SNP was associated with an increased risk for glioblastoma; however, the significance of this result was not confirmed after Bonferroni correction (p > 0.05). Li et al.²⁷ have reported that rs3024994 variant was associated with a decreased risk, whereas rs2010963 variant carriers and rs3025039 variant homozygous had an increased risk for glioma. Also Jiang et al.⁴⁰ described rs3025039 and rs2010963 variant homozygosity to be associated with an increased risk for glioma. These discrepancies could be related to different background of the populations studied. While the previous works have studied Chinese patients, our samples are from Caucasian origin. Although no significant differences in genotype or allele frequencies are observed between different population origins,⁴⁵ it was already reported for other diseases that these polymorphisms differentially related to risk in distinctive populations.^46,47

Importantly, VEGF polymorphisms have been related with patients’ response to anti-angiogenic drugs, particularly bevacizumab.^29,48 Few studies have described an association between VEGF SNPs, specifically rs2010963, rs699947, rs833061, and rs1005230 and response to bevacizumab in glioblastoma patients. In particular, the rs2010963, rs699947, and rs833061 variant homozygous genotypes were associated with an increase in PFS, whereas the rs1005230 variant homozygous genotype was associated with a decrease in PFS of glioblastoma patients following bevacizumab treatment. A major limitation of our study is the small number of patients treated with bevacizumab (n = 8), which did not allow any statistical analysis to access the value of the studied polymorphisms in the context of the bevacizumab treatment. Therefore, we have also analyzed the effect of these polymorphisms on patients’ survival, independently of the treatment regimen applied.

The rs1005230 variant homozygous consistently presented a decreased OS and PFS both for glioma and glioblastoma patients. Similarly, the rs3024994 variant heterozygous consistently presented a decreased PFS both for glioma and glioblastoma patients. Also the rs2010963 heterozygous glioblastoma patients had a reduced OS.

In conclusion, our data strongly suggest that different VEGF SNPs highly increase the risk to develop gliomas, and specifically glioblastoma, particularly the presence of the rs833061 variant. Moreover, several VEGF SNPs seem to be prognostic markers of glioma and glioblastoma patients’ survival. In particular, rs3024994 heterozygous and the homozygous carriers of rs1005230 confer a worse prognosis for these patients. Although these data need confirmation by independent studies, they demonstrate that VEGF SNPs are potential glioblastoma biomarkers.

Footnotes

Acknowledgements

Paulo Linhares and Marta Viana-Pereira contributed equally to this work.

Compliance with ethical standards

This study was approved by “Comissão de tica para a Sade,” stating that it was unanimously approved, with nothing to object. Signed informed consent was obtained from each participant. Local institutional ethics committees of Hospital São João, Portugal and Hospital de Braga, Portugal, approved the procedures followed in this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by Fundação para a Ciência e Tecnologia (FCT), Portugal (; grant PTDC/SAU-ONC/115513/2009_FCOMP-01-0124-FEDER-015949 to R.M.R.). This article has been developed under the scope of the project NORTE-01-0145-FEDER-000013, supported by the Northern Portugal Regional Operational Programme (NORTE 2020), under the Portugal 2020 Partnership Agreement, through the European Regional Development Fund (FEDER). M.V.-P. is recipient of an FCT Post-Doctoral fellowship (ref: SFRH/BPD/104290/2014).

References

Louis

Ohgaki

Wiestler

et al . The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol 2007; 114(2): 97–109.

Tanaka

Louis

Curry

et al . Diagnostic and therapeutic avenues for glioblastoma: no longer a dead end? Nat Rev Clin Oncol 2013; 10(1): 14–26.

Weller

van den Bent

Hopkins

et al . EANO guideline for the diagnosis and treatment of anaplastic gliomas and glioblastoma. Lancet Oncol 2014; 15(9): e395–e403.

Ostrom

Gittleman

Stetson

et al . Epidemiology of gliomas. Cancer Treat Res 2015; 163: 1–14.

Ohgaki

Epidemiology of brain tumors. Methods Mol Biol 2009; 472: 323–342.

Wrensch

Fisher

Schwartzbaum

et al . The molecular epidemiology of gliomas in adults. Neurosurg Focus 2005; 19(5): E5.

Adel Fahmideh

Schwartzbaum

Frumento

et al . Association between DNA repair gene polymorphisms and risk of glioma: a systematic review and meta-analysis. Neuro Oncol 2014; 16(6): 807–814.

Liu

Scheurer

El-Zein

et al . Association and interactions between DNA repair gene polymorphisms and adult glioma. Cancer Epidemiol Biomarkers Prev 2009; 18(1): 204–214.

McKean-Cowdin

Barnholtz-Sloan

Inskip

et al . Associations between polymorphisms in DNA repair genes and glioblastoma. Cancer Epidemiol Biomarkers Prev 2009; 18(4): 1118–1126.

10.

Zawlik

Kita

Vaccarella

et al . Common polymorphisms in the MDM2 and TP53 genes and the relationship between TP53 mutations and patient outcomes in glioblastomas. Brain Pathol 2009; 19(2): 188–194.

11.

Pandey

Kaur

Costa

et al . Immunoglobulin genes implicated in glioma risk. Oncoimmunology 2014; 3: e28609.

12.

Vieira de Castro

Goncalves

Costa

et al . Impact of TGF-β1 -509C/T and 869T/C polymorphisms on glioma risk and patient prognosis. Tumour Biol 2015; 36(8): 6525–6532.

13.

Shete

Hosking

Robertson

et al . Genome-wide association study identifies five susceptibility loci for glioma. Nat Genet 2009; 41(8): 899–904.

14.

Walsh

Codd

Smirnov

et al . Variants near TERT and TERC influencing telomere length are associated with high-grade glioma risk. Nat Genet 2014; 46(7): 731–735.

15.

Walsh

Rice

Decker

et al . Genetic variants in telomerase-related genes are associated with an older age at diagnosis in glioma patients: evidence for distinct pathways of gliomagenesis. Neuro Oncol 2013; 15(8): 1041–1047.

16.

Yang

Kon

Hung

et al . Combinations of newly confirmed glioma-associated loci link regions on chromosomes 1 and 9 to increased disease risk. BMC Med Genomics 2011; 4: 63.

17.

Rajaraman

Melin

Wang

et al . Genome-wide association study of glioma and meta-analysis. Hum Genet 2012; 131(12): 1877–1888.

18.

Costa

Ferreira

Costa

et al . Association between functional EGF+61 polymorphism and glioma risk. Clin Cancer Res 2007; 13(9): 2621–2626.

19.

Costa

Viana-Pereira

Fernandes

et al . Impact of EGFR genetic variants on glioma risk and patient outcome. Cancer Epidemiol Biomarkers Prev 2011; 20(12): 2610–2617.

20.

Plate

Scholz

Dumont

DJ.

Tumor angiogenesis and anti-angiogenic therapy in malignant gliomas revisited. Acta Neuropathol 2012; 124(6): 763–775.

21.

Curry

Dahiya

Alva Venur

et al . Bevacizumab in high-grade gliomas: past, present, and future. Expert Rev Anticancer Ther 2015; 15(4): 387–397.

22.

Ghiaseddin

Peters

KB.

Use of bevacizumab in recurrent glioblastoma. CNS Oncol 2015; 4(3): 157–69.

23.

Mesti

Moltara

Boc

et al . Bevacizumab and irinotecan in recurrent malignant glioma, a single institution experience. Radiol Oncol 2015; 49(1): 80–85.

24.

Carvalho

BFA

Almeida

Sampaio

et al . Second-line chemotherapy in recurrent glioblastoma: a 2-cohort study. Oncol Res Treat 2015; 38(7–8): 348–354.

25.

Jain

Vargo

Danesi

et al . The role of vascular endothelial growth factor SNPs as predictive and prognostic markers for major solid tumors. Mol Cancer Ther 2009; 8(9): 2496–2508.

26.

Watson

Webb

Bottomley

et al . Identification of polymorphisms within the vascular endothelial growth factor (VEGF) gene: correlation with variation in VEGF protein production. Cytokine 2000; 12(8): 1232–1235.

27.

Zhao

Fan

et al . Possible association between polymorphisms of human vascular endothelial growth factor A gene and susceptibility to glioma in a Chinese population. Int J Cancer 2011; 128(1): 166–175.

28.

Sjostrom

Wibom

Andersson

et al . Genetic variations in VEGF and VEGFR2 and glioblastoma outcome. J Neurooncol 2011; 104(2): 523–527.

29.

Galanis

Anderson

Lafky

et al . Phase II study of bevacizumab in combination with sorafenib in recurrent glioblastoma (N0776): a north central cancer treatment group trial. Clin Cancer Res 2013; 19(17): 4816–4823.

30.

James

Goodeve

AC.

von Willebrand disease. Genet Med 2011; 13(5): 365–376.

31.

Mancuso

Tuley

Westfield

et al . Structure of the gene for human von Willebrand factor. J Biol Chem 1989; 264(33): 19514–19527.

32.

Sadler

JE.

Biochemistry and genetics of von Willebrand factor. Annu Rev Biochem 1998; 67: 395–424.

33.

TanTai

Shen

Zhao

Quantitative assessment of the influence of common variations on 6p21 and lung cancer risk. Tumour Biol 2014; 35(1): 689–694.

34.

De Mello

Ferreira

Soares-Pires

et al . The impact of polymorphic variations in the 5p15, 6p12, 6p21 and 15q25 Loci on the risk and prognosis of Portuguese patients with non-small cell lung cancer. PLoS ONE 2013; 8(9): e72373.

35.

Hegi

Diserens

Godard

et al . Clinical trial substantiates the predictive value of O-6-methylguanine-DNA methyltransferase promoter methylation in glioblastoma patients treated with temozolomide. Clin Cancer Res 2004; 10(6): 1871–1874.

36.

Xiong

Huo

Chen

et al . Vascular endothelial growth factor (VEGF) receptor-2 tyrosine 1175 signaling controls VEGF-induced von Willebrand factor release from endothelial cells via phospholipase C-gamma 1- and protein kinase A-dependent pathways. J Biol Chem 2009; 284(35): 23217–23224.

37.

Yang

Sun

et al . Gastric cancer-associated enhancement of von Willebrand factor is regulated by vascular endothelial growth factor and related to disease severity. BMC Cancer 2015; 15: 80.

38.

Starke

Ferraro

Paschalaki

et al . Endothelial von Willebrand factor regulates angiogenesis. Blood 2011; 117(3): 1071–1080.

39.

Zhao

Chen

et al . Impact of VEGFA polymorphisms on glioma risk in Chinese. Oncotarget 2017; 8(48): 83712–83722.

40.

Jiang

Lian

Xie

et al . Three single nucleotide polymorphisms of the vascular endothelial growth factor (VEGF) gene and glioma risk in a Chinese population. J Int Med Res 2013; 41(5): 1484–1494.

41.

Jeon

Kim

Park

et al . Interplay between 3’-UTR polymorphisms in the vascular endothelial growth factor (VEGF) gene and metabolic syndrome in determining the risk of colorectal cancer in Koreans. BMC Cancer 2014; 14: 881.

42.

Wang

Zhao

et al . rs833061 and rs699947 on promoter gene of vascular endothelial growth factor (VEGF) and associated lung cancer susceptibility and survival: a meta-analysis. Med Sci Monit 2014; 20: 2520–2526.

43.

Beeghly-Fadiel

Shu

et al . Genetic variation in VEGF family genes and breast cancer risk: a report from the Shanghai Breast Cancer Genetics Study. Cancer Epidemiol Biomarkers Prev 2011; 20(1): 33–41.

44.

Ruggiero

Dalmasso

Nutile

et al . Genetics of VEGF serum variation in human isolated populations of Cilento: importance of VEGF polymorphisms. PLoS ONE 2011; 6(2): e16982.

45.

Abecasis

Auton

Brooks

et al . An integrated map of genetic variation from 1,092 human genomes. Nature 2012; 491(7422): 56–65.

46.

Shi

et al . Two polymorphisms (rs699947, rs2010963) in the VEGFA gene and diabetic retinopathy: an updated meta-analysis. BMC Ophthalmol 2013; 13: 56.

47.

et al . Association of VEGF genetic polymorphisms with recurrent spontaneous abortion risk: a systematic review and meta-analysis. PLoS ONE 2015; 10(4): e0123696.

48.

Di Stefano

Labussiere

Lombardi

et al . VEGFA SNP rs2010963 is associated with vascular toxicity in recurrent glioblastomas and longer response to bevacizumab. J Neurooncol 2015; 121(3): 499–504.