Molecular Biology Reports

, Volume 36, Issue 1, pp 37–45

Combined effects of the angiogenic genes polymorphisms on prostate cancer susceptibility and aggressiveness

Authors

    • Department of Molecular Immuno-OncologyFaculty of Medicine
  • Hamadi Saad
    • Department of UrologyEPS Fattouma Bourguiba
  • Faouzi Mosbah
    • Department of UrologyEPS Sahloul
  • Lotfi Chouchane
    • Department of Molecular Immuno-OncologyFaculty of Medicine
Article

DOI: 10.1007/s11033-007-9149-4

Cite this article as:
Sfar, S., Saad, H., Mosbah, F. et al. Mol Biol Rep (2009) 36: 37. doi:10.1007/s11033-007-9149-4
  • 121 Views

Abstract

The single-gene approaches in association studies of polygenic diseases are likely to provide limited value in predicting risk. The combined analysis of genetic variants that interact in the same pathway may amplify the effects of individual polymorphisms and enhance the predictive power. To evaluate higher order gene–gene interaction, we have examined the contribution of four angiogenic gene polymorphisms (VEGF-1154G/A; VEGF-634G/C; MMP9-1562C/T and TSP1-8831A/G) in combination to the risk of prostate cancer. For the combined analysis of VEGF and MMP9 SNPs, we found a significant gene–dosage effect for increasing numbers of potential high-risk genotypes. Compared to referent group (low-risk genotypes), individuals with one (OR = 2.79, P = 0.1), two (OR = 4.57, P = 0.02) and three high-risk genotypes (OR = 7.11, P = 0.01) had increasingly elevated risks of prostate cancer. Similarly, gene–gene interaction of VEGF and TSP1 polymorphisms increased risk of prostate cancer in additive manner (OR = 6.00, P = 0.03), although the TSP1 polymorphism itself was not associated with the risk. In addition, we examined the synergistic effect of these polymorphisms in relation to prostate cancer prognosis according to histopathological grade and clinical stage at diagnosis. Cross-classified analysis revealed potential higher order gene–gene interactions between VEGF and TSP1 polymorphisms in increasing the risk of developing an aggressive phenotype disease. Patients carrying three high-risk genotypes showed a 20-fold increased risk of high-grade tumor (OR = 20.75, P = 0.002). These results suggest that the gene–gene interaction of angiogenic gene polymorphisms’ increased risk of prostate cancer onset and aggressiveness.

Keywords

Angiogenic pathwayCombined analysisCancer prognosisGene–gene interactionsGenetic susceptibilityProstate cancer

Introduction

It is well known that both cancer initiation risk and later neoplastic events (tumor growth, invasion, metastatic spread, response to therapeutic interventions and survival) are strongly affected by factors predetermined by the individual’s genetic background. In many candidate gene approaches of polygenic diseases, including cancer, the effect of each individual SNP is unlikely to be substantial, highlighting the need for a more comprehensive approach for association studies. The low risk conferred by an individual variant is not surprising, given that carcinogenesis is usually a multistep process that occurs through a multifactorial interplay between many genetic and environmental factors. As the modest effects of each single gene may provide a limited practical value in predicting the risk, the pathway-based multigenic approach combining multiple polymorphisms that interact in the same pathway may amplify the effect of individual polymorphisms and enhance the predictive power. In this study, we used prostate cancer as a prototype to examine the combined effects of a panel of polymorphisms in the angiogenic pathway on prostate cancer risk and prognosis and evaluated higher order gene–gene interactions.

Prostate cancer is one of the most common malignancies among men in developed countries [1]. Accumulating evidence supports an important role of genetics in prostate cancer etiology, yet the responsible genes remain largely unidentified. In this context, twin studies suggest that up to 50% of prostate cancer risk may be explained by genetic factors [2]. While rare high-penetrance germline mutations account for a small percentage of all prostate cancers, much of the variation in genetic risk is likely to be explained by combinations of more common low-penetrance variants. So far, case-control studies have generally focused on the investigation of single functional candidate genes to attempt to identify low-penetrance susceptibility genes. However, it is possible that the combination of several polymorphisms in various genes of a biological pathway might significantly influence prostate cancer development and outcome.

Like all solid tumors, the progressive growth and metastasis of prostate cancer are angiogenesis-dependent [3, 4]. Increasing evidence supports the hypothesis that tumor angiogenesis is controlled by an “angiogenic switch,” a physiological mechanism that occurs when the tumor and stroma produce excess of pro-angiogenic molecules over anti-angiogenic factors [5]. Therefore, molecules that are involved in this multistep process, such as the degradation of extracellular matrix or the induction of endothelial cell proliferation, have been described as angiogenic factors. Up to now, many pro-angiogenic molecules have been identified [68]. VEGF was first discovered as a potent endothelial cell-specific mitogen and was described to increase endothelial cell permeability, migration and chemotaxis [9]. VEGF acts in an autocrine and/or paracrine fashion via two high-affinity tyrosine kinase receptors called Flt-1 (VEGFR1) and KDR (VEGFR2) expressed on both endothelial cells and prostate tumor cells [10]. Signaling through such receptors induces the expression of the plasminogen activators (PA), the urokinase receptor (uPAR), and the metalloproteinase interstitial collagenase, facilitating the extracellular matrix degradation and further migration of endothelial cells [9]. MMPs are important modulators of these events because they are responsible for the proteolysis of connective tissue barriers necessary for new vessel formation. It has been reported that the production of MMP9 and other proteases by prostate cancer cells and stromal cells facilitates the degradation of ECM, resulting in tumor invasion and subsequent metastasis [1113]. MMP9 is involved in many steps of cancer development, including decreasing cancer cell apoptotic potential, promoting angiogenesis [14] and regulating immune response to cancer [15]. A direct role for MMP9 in the regulation of angiogenesis via modulation of growth factor availability was initially demonstrated in homozygous mice with a null mutation in MMP9 [16]. Moreover, it has been clearly demonstrated that MMP9 triggers VEGF release from extracellular stores leading to the angiogenic switch [14]. VEGF and MMP9 are two major pro-angiogenic and pro-metastatic molecules that have been associated with negative prognostic features in various types of cancers including prostate tumors. It has been reported that decreased expression of these molecules in vivo directly correlates with decreased neovascularization and lymph node metastasis [17]. In clinical studies, urinary and serum VEGF levels have been promising as prognostic markers, particularly in men with hormone refractory prostate cancer [18].

On the other hand, clinical evidence suggested that prostate carcinomas are able to escape the thrombospondin-1 (TSP1) anti-angiogenic effects. Indeed, TSP1 expression has been inversely correlated with tumorgenicity and metastatic potential of prostate cancer [19]. TSP1 was the first protein to be recognized as a potential endogenous suppressor of capillary morphogenesis in vivo [20]. The normal suppression of angiogenesis by TSP1 involves multiple mechanisms, including inhibiting endothelial cell proliferation [21] and migration [22], inducing endothelial cell apoptosis [2324], inhibiting MMP3-dependent activation of pro-MMP9 [25] and interaction with VEGF.

Nowadays, research is focusing on the relationship between angiogenic genes and tumor susceptibility because the angiogenic switch has been considered to trigger cancer predisposition. Recently, we have showed that individual variants in VEGF and MMP9 genes may be genetic risk factors for prostate cancer onset and severity. However, our finding provided evidence that TSP1 gene may play a limited role in prostate cancer risk and clinical outcome in a Tunisian population [26, 27]. Based on the intertwined and interactive roles that VEGF, MMP9 and TSP1 play at the molecular level in angiogenic pathway and the reported functional relevance of their variants, we further hypothesized a priori that the joint effect of genetic variants in these angiogenesis regulators may increase prostate cancer risk. To our knowledge, this is the first published report on the potential interaction between these genes in any disease.

Materials and methods

Patients and controls

The current study population originates from two previous independent studies of prostate cancer risk in relation to single angiogenic genes SNPs [26, 27]. This case-control study included a total of 175 subjects, consisting of 101 prostate cancer patients and 74 cancer-free controls. Prostate cancer patients were identified and recruited between September 2002 and August 2004 from the two departments of Urology of Monastir and Sousse Hospitals, Tunisia. The diagnosis has been histologically confirmed in 97 patients. The remaining four patients were included in the study according to their radiological evidence of metastasis and their high prostate-specific antigen (PSA) levels. These criteria were designed to ensure that the four patients had advanced disease. The serum PSA values were measured in all cases before treatment. Clinical characteristics, including Gleason grade, TNM stage, age at diagnosis, and family history, were obtained from medical records. Patients with previous cancer or metastasized cancer from other organs were excluded.

Control subjects were randomly selected from the same population living in the central coast of Tunisia, and were matched according to age and geographical origin of patients. The control group consisted of 74 healthy male subjects having no evidence of any personal or family history of cancer. Information was obtained from control participants through standardized questionnaire including age, residence, dietary habits, and family history of cancer. All subjects in both groups provided informed consent to participate in the study and to allow their biological samples to be genetically analyzed. Approval for the study was given by the National Ethical Committee.

Genotyping methods

Details of the study methodology have been previously described [26, 27]. DNA was extracted from peripheral blood samples using the standard method (salting out procedure) [28]. As previously described, the VEGF-1154G/A and VEGF-634G/C genotypes were separately determined using the allele specific PCR (AS-PCR) and the restriction fragment length polymorphism PCR (RFLP-PCR) methods, respectively [26]. Briefly, the AS-PCR reaction was carried out using three specific primers of the VEGF-1154G/A polymorphism and a second internal control primer pair for HGH gene in the same amplification mixture [29]. All products were analyzed into 2% agarose gel. For the VEGF-634G/C genotyping, the RFLP-PCR assay was conducted using the forward primer 5′-TTGCTTGCCATTCCCCACTTGA-3′ and the reverse primer 5′-CCGAAGCGAGAACAGCCCAGAA-3′ yielding a fragment of 470 pb. The PCR products derived from the VEGF-634 G allele were digested by the BsmFI restriction enzyme into two fragments of 196 and 274 pb.

The genotypes of TSP1 (8831A/G) and MMP9 (1562C/T) markers were determined by PCR-RFLP as previously reported [27]. In brief, the TSP1-8831G and the MMP9-1562T created the BsrSI and the SphI restriction enzyme recognition sequences, respectively. The digested fragments were separated by 2% agarose gel electrophoresis and visualized under UV light.

Statistical analysis

The statistical analysis was performed using Epi-info 5.01 and SEM-STATISTQUES programs. The differences in the combined genotype frequencies were analyzed between the study and control groups by 2 × 2 tables using χ2 analysis or Fisher’s exact test if one or more variables in 2 × 2 tables were <5. The effect of combined genotypes on prostate cancer risk was also estimated by calculating odds ratios (ORs) and their 95% confidence intervals (95% CI) using the unconditional logistic regression analysis with the low risk combination as a reference category. The P-values reported in the study are based on a two-sided probability test with a significance level of P < 0.05.

We also performed a logistic regression analysis to determine whether the clinical stage and the pathological grade were associated with the combined genotypes of the angiogenic genes. The clinicopathological parameters were dichotomized as follows: The pathological stage at the time of diagnosis was classified according to TNM system into the localized group (T1-T2N0M0) and the advanced group (T3–T4N0M0 and T1–T4N0–1M1/T1–T4N1M0–1). The histopathological grade was recorded as the Gleason score and was classified into two groups: the low-grade group (Gleason score < 7) and the high-grade group (Gleason score ≥ 7).

Results

The mean age for patients and controls was 70.0 (range: 49–91) and 65.0 (range: 42–80) years, respectively. No significant differences in the mean age of first diagnosis were found between patients and controls (P > 0.05). Genotype distributions of the four analyzed gene SNPs had no deviation from Hardy Weinberg equilibrium.

Gene-gene interaction in modifying prostate cancer risk

Because VEGF functionality interacts with MMP9 and TSP1 in angiogenic processes, we studied the multiloci genetic effects on prostate cancer risk and severity. Based on our previously published results, the VEGF-1154 G, VEGF-634 C and the MMP9 T alleles were designed as risk alleles for susceptibility and poor prognostic outcome of prostate cancer [26, 27]. Table 1 shows the relation between susceptibility ORs and the number of high-risk genotypes for each n SNPs combination. For the primary analysis, the subjects were cross-classified by VEGF-1154G/A and MMP9-1562C/T genotypes, with those having both low-risk genotypes, VEGF AA and MMP9 CC, chosen as a baseline reference category. We defined the GA and GG genotypes of VEGF and the CT and TT genotypes of MMP9 as high-risk genotypes. Relative to the referent genotypes, the presence of one or two putative genotypes showed a significantly increased risk of developing prostate cancer with an OR of 3.35 (P = 0.04) and 4.72 (P = 0.02), respectively. When we evaluated the combined effect of VEGF-634G/C and MMP9-1562C/T genotypes, we found that individuals carrying two high-risk genotypes (GC and CC genotypes of VEGF and the CT and TT genotypes of MMP9) showed a 3.2-fold increase of prostate carcinoma risk using low-risk genotypes (VEGF GG and MMP9 CC) as reference (OR = 3.2, P = 0.03). Interestingly, the estimated relative risks for prostate cancer associated to combined VEGF-1154G/A; VEGF-634G/C and MMP9-1562C/T genotypes showed a positive gradient in the ORs related to the number of high-risk genotypes in a dose-dependent manner. Individuals with three high-risk genotypes showed the highest risk, approximately sevenfold higher risk of prostate cancer (OR = 7.11; P = 0.01) than the reference. Moreover, we examined whether VEGF and TSP1 polymorphisms play an interactive role in prostate cancer risk. Table 2 shows the cross-classified VEGF-TSP1 interaction results. The reference, which is suspected as the lowest risk for cancer, was composed of subjects carrying the VEGF-1154 AA; VEGF-634 GG and TSP1 AA genotypes. We defined the GA and GG genotypes of TSP1 as high-risk genotypes. We noted a significant association with prostate cancer risk only for individuals with two or three high-risk genotypes compared to the reference. The gene–gene interaction of VEGF and TSP1 increased sixfold the risk of prostate cancer (P = 0.03), although the TSP1 polymorphism itself was not associated with prostate cancer susceptibility [27].
Table 1

Interaction and additive effects of VEGF and MMP9 polymorphisms on prostate cancer risk

Number of high-risk genotypes

Controls (N = 70) n (%)

Patients (N = 101) n (%)

OR (95% CI)

P-value

VEGF-1154/ MMP9

  0

9 (12.85)

4 (3.96)

1

 

  1

51 (72.85)

76 (75.24)

3.35 (0.88–13.77)

0.04

  2

10 (14.3)

21 (20.8)

4.72 (1.42–38.03)

0.02

VEGF-634/MMP9

  0

26 (37.14)

23 (22.77)

1

 

  1

38 (54.29)

61 (60.4)

1.81 (0.86–3.85)

0.08

  2

6 (8.57)

17 (16.83)

3.2 (0.97–11.01)

0.03

VEGF-1154/-634/MMP9

  0

8 (11.43)

3 (2.97)

  

  1

21 (30)

22 (21.79)

2.79 (0.56–15.58)

0.1

  2

35 (50)

60 (59.4)

4.57 (1.01–23.48)

0.02

  3

6 (8.57)

16 (15.84)

7.11 (1.12–51.54)

0.01

OR: odds ratio; CI: confidence interval

The high risk genotypes: VEGF-1154 G/A= GG/GA ; VEGF-634 G/C= CC/GC; MMP9-1562 C/T= CT/TT

Table 2

Combined effects of VEGF and TSP1 genetic variants and estimated relative risk for prostate cancer

Number of high-risk genotypes

Controls (N = 74) n (%)

Patients (N = 101) n (%)

OR (95% CI)

P-value

VEGF-1154/TSP1

  0

10 (13.51)

4 (3.96)

1

 

  1

55 (74.33)

83 (82.18)

3.77 (1.02–15.14)

0.02

  2

9 (12.16)

14 (13.86)

3.89 (0.77–21.14)

0.05

VEGF-634/TSP1

  0

29 (39.19)

24 (23.76)

1

 

  1

39 (52.7)

66 (65.35)

2.04 (0.99–4.23)

0.03

  2

6 (8.11)

11 (10.89)

2.22 (0.63–8.01)

0.1

VEGF-1154/-634/TSP1

  0

9 (12.16)

3 (2.97)

  

  1

25 (33.78)

23 (22.77)

2.76 (0.58–14.83)

0.15

  2

35 (47.3)

65 (64.36)

5.57 (1.26–27.98)

0.009

  3

5 (6.76)

10 (9.9)

6.00 (0.86–48.28)

0.03

OR: odds ratio; CI: confidence interval

The high risk genotypes: VEGF-1154 G/A= GG/GA; VEGF-634 G/C= CC/GC; TSP1 8831 G/A= GA/GG

Nevertheless, there was no statistical evidence for the interactive effects of the four gene polymorphisms or other combined genotypes (data not shown).

Stratified analysis of the combined genotypes and prostate cancer prognosis indicators

We performed stratified analysis for the gene–gene interaction between the VEGF, MMP9 and TSP1 variant genotypes by the clinicopathological parameters of prostate cancer (histopathological grade and clinical stage). Table 3 shows the cross-classified VEGF/MMP9 and VEGF/TSP1 interaction results according to the tumor grade. For the combined VEGF and MMP9 genotype distribution, the pooled low-risk genotypes (VEGF-1154 AA/VEGF-634 GG/MMP9 CC) were not detected in any of the high-grade subjects. Therefore, individuals harboring one high-risk genotype were included with the baseline reference category in our subgroup analysis. It was apparent that the VEGF-MMP9 gene–gene interaction was more pronounced in the high-grade subgroup. Compared with the reference group, patients with two putative risk genotypes (OR = 3.15, P = 0.02) and those with three high-risk genotypes had a borderline significantly increased risk of high-grade tumors (OR = 3.75, P = 0.05). Furthermore, the joint analysis of VEGF and TSP1 SNPs revealed a significant trend of increasing high-grade tumor with increasing number of putative high-risk genotypes. Subjects carrying three high-risk genotypes were 20 times more likely to have high-grade tumors relative to reference category (OR = 20.75, P = 0.002).
Table 3

Interaction of VEGF with MMP9 and TSP1 polymorphisms and assessment of tumor aggressiveness

Number of high-risk genotypesa

Low grade N = 42, n (%)

High grade N = 55, n (%)

OR (95% CI)

P-value

VEGF-1154/MMP9

  0–1

34 (80.95)

44 (80)

1.00

 

  2

8 (19.05)

11 (20)

1.06 (0.35–3.29)

0.9

VEGF-634/MMP9

  0–1

36 (85.72)

45 (81.82)

1.00

 

  2

6 (14.28)

10 (18.18)

1.33 (0.39–4.62)

0.6

VEGF-1154/TSP1

  0–1

39 (92.86)

44 (80)

1.00

 

  2

3 (7.14)

11 (11)

3.25 (0.76–15.34)

0.07

VEGF-634/TSP1

  0–1

40 (95.24)

46 (83.64)

1.00

 

  2

2 (4.76)

9 (16.36)

3.91 (0.72–27.97)

0.06

VEGF-1154/-634/MMP9

  0–1

15 (35.71)

8 (14.56)

1.00

 

  2

22 (52.39)

37 (67.27)

3.15 (1.04–9.79)

0.02

  3

5 (11.9)

10 (18.18)

3.75 (0.79–18.98)

0.05

VEGF-1154/-634/TSP1

  0–1

16 (38.09)

7 (12.73)

1.00

 

  2

25 (59.52)

39 (70.91)

3.75 (1.16–14.26)

0.01

  3

1 (2.39)

9 (16.36)

20.75 (19.52–526.04)

0.002

OR: odds ratio; CI: confidence interval

The high risk genotypes: VEGF-1154 G/A= GG/GA; VEGF-634 G/C= CC/GC; MMP9-1562 C/T= CT/TT; TSP1 8831 G/A= GA/GG

Regarding the clinical stage, the combined analysis of all assayed genotypes in the angiogenic pathway did not reveal any significant joint association of these polymorphisms with prostate cancer progression (Table 4).
Table 4

Cross-classified distribution of VEGF, MMP9 and TSP1 genotypes in prostate cancer subgroups according to clinical stage

Number of high-risk genotypes

Localized N = 30, n (%)

Advanced N = 71, n (%)

OR (95% CI)

P-value

VEGF-1154/MMP9

  0–1

22 (73.33)

58 (81.7)

1.00

 

  2

8 (26.67)

13 (18.3)

0.62 (0.2–1.9)

0.34

VEGF-634/MMP9

  0–1

22 (73.33)

62 (87.32)

1.00

 

  2

8 (26.67)

9 (12.68)

0.4 (0.12–1.32)

0.08

VEGF-1154/TSP1

  0–1

26 (86.66)

61 (85.91)

1.00

 

  2

4 (13.34)

10 (14.09)

1.07 (0.27–4.48)

0.5

VEGF-634/TSP1

  0–1

27 (90)

63 (88.73)

1.00

 

  2

3 (10)

8 (11.27)

1.14 (0.25–5.94)

0.5

VEGF-1154/-634/MMP9

  0–1

9 (30)

16 (22.53)

1.00

 

  2

14 (46.66)

46 (64.79)

1.85 (0.6–5.71)

0.2

  3

7 (23.34)

9 (12.68)

0.72 (0.17–3.14)

0.6

VEGF-1154/-634/TSP1

  0–1

8 (26.67)

18 (25.35)

1.00

 

  2

20 (66.66)

45 (63.38)

1.00 (0.33–2.96)

0.8

  3

2 (6.67)

8 (11.27)

1.78 (0.25–15.44)

0.4

OR: odds ratio; CI: confidence interval

The high risk genotypes: VEGF -1154 G/A= GG/GA; VEGF -634 G/C= CC/GC; MMP9-1562 C/T= CT/TT; TSP1 8831 G/A= GA/GG

Discussion

Genetic susceptibility to prostate cancer is consistently observed in a large number of family, twin and case-control studies [30]. Low-penetrance genes are those involved in gene–environment interaction and commonly associated with many sporadic cancers. These genes are known to aggregate with disease and are often seen to interact with other genes to increase the overall risk. However, there are inadequate data to elucidate the genetic mechanism behind the carcinogenesis of prostate cancer. Therefore, the identification of the appropriate molecular marker is worthwhile, as it will revolutionize the prognosis and treatment of prostate cancer.

The importance of angiogenesis for tumor growth and metastasis, originally postulated in 1971 by Folkman, has been confirmed in a variety of systems [31, 32]. In the light of the potential importance of angiogenesis in the development of prostate carcinoma, we have used a multiloci approach to explore the joint effect of the four sequence variants of the VEGF, MMP9 and TSP1 genes on prostate cancer risk and aggressiveness. To our knowledge, no epidemiological studies have addressed the potential gene–gene synergistic effects of these key genes implicated in the angiogenic regulatory pathway.

Our previous single-gene studies showed that SNPs in VEGF (1154G/A and 634G/C) and MMP9 (1562C/T) genes may have statistically significant modest effects on increased prostate tumor susceptibility and aggressiveness [26, 27]. We have demonstrated that the estimated relative risk for prostate cancer associated to a single genetic variant in VEGF-1154G/A, VEGF-634G/C and MMP9-1562C/T was 2.4 for the G allele, 1.95 for the C allele and 2.86 for the T allele, respectively. The present study extends our findings by showing that polymorphisms in these two genes, when considered in combination, confer higher risk for prostate cancer compared to the effect of each individual genetic variant. Men carrying presumed low-risk genotypes of both genes were used as the reference group. We systematically evaluated the joint effect of these genes and predicted the individual susceptibility to prostate cancer based on any combination of two or three variants from all the genotyped polymorphisms. We found a significant trend of increased risk with increasing numbers of high-risk genotypes. The estimated relative risk was higher for individuals carrying combined VEGF and MMP9 risk alleles, and the most significant additive joint effect was observed among individuals with three high-risk genotypes who exhibited a sevenfold higher risk for this malignancy (OR = 7.11, P = 0.01). It follows the hypothesis that genes interact to confer genomic-based susceptibility to prostate cancer. Moreover, our results showed that interaction effects between sequence variants of VEGF and MMP9 genes conferred a borderline significant excess risk of high-grade tumors (OR = 3.75, P = 0.05), although the lack of significance must be interpreted in the context of our limited sample size.

The interaction and additive effects between the VEGF SNPs (1154G/A and 634G/C) and the MMP9-1562C/T polymorphisms may be a direct result of biological interactions between these specific SNPs, or, alternatively, may indirectly reflect other SNPs that are in linkage disequilibrium with these genotyped SNPs. VEGF-1154G/A and VEGF-634G/C were located, respectively, in the promoter and the 5′-untranslated region and may potentially influence transcription of the VEGF gene [3338]. Similarly, Zhang et al. have reported that the MMP9-1562C/T polymorphism influence the binding of a transcription repressor protein and is associated with a higher MMP9 promoter activity [39]. VEGF and MMP9 altered expression directly or indirectly related to these SNPs, may lead to altered responses in the angiogenic signaling pathway that have been implicated in the development of prostate cancer [40]. Our results suggest that an up-regulation of VEGF production could be the mechanism underlying the enhanced growth and vascularization of MMP9 overexpression-tumors. Despite the intriguing genetic findings and the biologically plausible relationship between these genes, replication in other large studies is warranted.

It is well known that progressive tumor growth is associated with multiple changes in the immune compartment of the host. Various studies suggested that T cell infiltrations, regarded as evidence of a protective immune response against tumors, may promote vasculogenesis and microvascular remodeling [41, 42]. In fact, Owen et al. found that MMP9 and VEGF production are up-regulated in the T cells infiltrating tumor-bearing animals, and that VEGF can up-regulate MMP9 expression in tumor bearer’s lymphocytes promoting further protease production by other inflammatory cells within the tumor [43, 44]. Similarly, other reports demonstrated the MMP9 up-regulation by this growth factor in human aortic smooth muscle cells [45], human brain tumor cells [46], and human myelo-monocytic leukemia cell lines [47]. Furthermore, there is evidence that MMP9 promote the accessibility of VEGF to its receptors on endothelial cells, adding another dimension to the pro-angiogenic potential of MMP9 [14]. In agreement with our data, these findings speculate that these two factors may be interacting to promote neoplastic progression. There are no previous reports screening synergistic effects of these genes, and to the best of our knowledge, this is the first report showing combined effect of VEGF and MMP9 genotypes with the risk of prostate cancer.

More interestingly, our findings showed that combined analysis of several SNPs in the same pathway may reveal otherwise undetectable associations between individual SNPs with cancer risk. Although the results of our previous study proved that the single polymorphism in TSP1 (8831A/G) was not significantly associated with prostate cancer risk [27], a significant impact was observed in prostate cancer risk by combining the high-risk genotypes of VEGF (1154G/A and VEGF-634G/C) and TSP1 (8831A/G) SNPs. The presence of three high-risk genotypes enhanced the risk by six times relative to the referent category (OR = 6.00, P = 0.03). These results provide evidence that VEGF and TSP1 may function as tumor-associated angiogenic factors in prostate cancer. Although the functional relevance of the TSP1 N700S polymorphism is not well known, previous biochemical genetics study reported that individuals who were homozygous for the ser700 variant have twofold lower plasma levels of TSP1 and have a ninefold increased risk for myocardial infarction [48]. Moreover, it has been suggested that the ser700 polymorphism decreases the Ca2+ affinity of Ca2+ binding repeats and alters conformation change of full-length TSP1 protein [4951] that may increase susceptibility to proteolytic degradation [52]. Based on this knowledge, we speculated that genetically determined high levels of VEGF product could contribute to the increase in vascular density and the enlarged vascular morphology displayed by the TSP1 deficient cells, a mechanism that may possibly increase the risk of carcinogenesis.

As for the prognostic value, the potential clinical evaluation of the joint effect of these genes indicated that the combined genotypes were specifically associated with the highest Gleason score (aggressiveness). The stratified analysis showed a positive gradient in the ORs related to the number of high-risk genotypes. The risk was more pronounced by combination of the three putative risk genotypes (OR = 20.75, P = 0.002). Our major finding of a 20-fold increased risk for the interaction was supported by only nine high-grade cases and one low-grade patient in the cells carrying the three high-risk genotypes. Although this statistically significant result should be interpreted cautiously, we cannot exclude the possibility that this statistical interaction reflects an underlying biological interaction between these two SNPs. Our results require confirmation in even larger studies. Consistently, some clinical studies have reported that the higher production of VEGF and the lower levels of TSP1 were related to the poor cellular differentiation and to an aggressive behavior of tumor [5354]. Although the mechanisms of these observations remain elusive, these results are in agreement with the notion that MMP9 and TSP1 coordinate with VEGF in the angiogenic signaling pathway. It has been reported that TSP1 inhibits angiogenesis through MMP9 inhibition [55]. Thus, TSP1 inhibits VEGF mobilization from the extracellular matrix by inhibiting active MMP9 [56]. TSP1 may also inhibit the VEGF activity by direct interaction [56]. Nevertheless, we failed to find any statistical evidence for the interactive effects of the four gene polymorphisms, possibly because of the small population sample size and the insufficient statistical power, which suggests that a larger study sample may be required to observe a statistically significant effect.

In summary, our results suggest that individuals with higher numbers of risk alleles in angiogenic genes are at an increased risk of prostate cancer, confirming the importance of taking a pathway-based approach to improve the resolution of the risk assessment process. This study is important not only for prostate cancer, but also for the risk assessment of many complex diseases involving low-penetrance genes.

Acknowledgements

We would like to thank all Tunisian subjects for their participation in this study. This work was supported by the Ministry of Higher Education and Scientific and Technological Research and the Ministry of Public Health of Tunisia. We would like to thank Mr. Adel Rdissi for the English revision.

Copyright information

© Springer Science+Business Media B.V. 2007