Annals of Oncology Advance Access originally published online on September 5, 2007
Annals of Oncology 2007 18(12):1990-1994; doi:10.1093/annonc/mdm361
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2007 European Society for Medical Oncology
gastrointestinal tumors |
Polymorphisms in the genes of the urokinase plasminogen activation system in relation to colorectal cancer
1 Division of Molecular Genetic Epidemiology, German Cancer Research Center, Heidelberg, Germany
2 Center for Family and Community Medicine, Karolinska Institute, Huddinge
3 Department of Oncology, Norrlands University Hospital
4 Department of Public Health and Clinical Medicine/Nutritional Research, Umeå University
5 Department of Medical Biosciences, Umeå University, Umeå, Sweden
* Correspondence to: Asta Försti, German Cancer Research Center, Division of Molecular Genetic Epidemiology C050, Im Neuenheimer Feld 580, D-69120 Heidelberg, Germany. Tel: +49-6221-421803; Fax: +49-6221-421810; E-mail: a.foersti{at}dkfz.de
 |
Abstract |
|---|
 Top Abstract introductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Background: Extracellular matrix degradation, mediated by the urokinase plasminogen activation (uPA) system, is a critical step in tumor invasion and metastasis. High tumor levels of uPA and its inhibitor PAI-1 have been correlated with poor cancer prognosis. We examined four single nucleotide polymorphisms (SNPs) with a potential effect on expression of genes in the uPA system for their role in colorectal cancer susceptibility and prognosis.
Patients and methods: We genotyped the SNPs in 308 Swedish incident colorectal cancer patients with up to 16 years of follow-up and in 585 age- and sex-matched controls. We evaluated the associations between genotypes and colorectal cancer and Dukes' stage. Survival probabilities were compared between different subgroups.
Results: Patients with PAI-1 –675 5G/5G genotype had better survival than patients with 4G/4G or 4G/5G genotypes when they had Dukes’ stage A or B tumors (P = 0.023 and P = 0.015, respectively). No statistically significant association was observed between the SNPS and the risk of colorectal cancer or Dukes’ stage.
Conclusions: Our results suggest a role for the PAI-1 genotype in colorectal cancer prognosis, but further studies are needed to evaluate the impact of our finding in the clinic.
Key words: colorectal cancer, PAI-1, prognostic biomarker, single nucleotide polymorphism, uPA system
|
introduction |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Extracellular matrix (ECM) degradation mediated by the urokinase plasminogen activation (uPA) system is an important mechanism involved in both physiological and pathological tissue remodeling [1–3]. Increased activity of the uPA system has been linked to tumor invasion and metastasis as well as to poor cancer prognosis [1, 3–5]. In addition to the ligand uPA, the uPA sytem includes the receptor uPAR, and two inhibitors, PAI-1 and PAI-2. uPA/uPAR promotes ECM degradation by converting plasminogen to plasmin, a proteinase that can activate certain matrix metalloproteinases and degrade most components of ECM [1–3]. The uPA system can also regulate cell adhesion and migration through the interaction with vitronectin (VTN) and adhesion receptors of the integrin family. Finally, uPA/uPAR can initiate signal transduction and induce chemotaxis. The effect of uPA/uPAR on plasminogen activation is regulated by PAI-1 and PAI-2. PAI-1 can also modulate the interaction between uPAR and VTN and some integrins.
Tumor tissue levels of the components of the uPA system have been implicated as promising prognostic markers in a wide range of malignancies, including colorectal cancer (CRC) [1, 4–7]. Expression levels of the uPA system components have been shown to be very low in normal colorectal mucosa, but high in invasive carcinomas [1]. Additionally, high plasma uPAR and PAI-1 levels have been associated with worse survival in patients with CRC [8, 9]. However, these two studies did not measure the protein levels in the tumor tissues and other studies have found no correlation between the plasma and the tumor tissue levels of the components of the uPA system in colorectal and breast cancer patients [10, 11].
Expression of the components of the uPA system can be upregulated by several hormones, growth factors and cytokines [1, 5]. Since single nucleotide polymorphisms (SNPs) located within the promoter or other regulatory regions of the genes in the uPA system may affect the expression of the gene or the protein, genotyping of these SNPs could provide a simple method to predict the risk and the prognosis of cancer. We selected a subset of SNPs from the genes in the uPA system and their substrates and examined whether they confer any CRC risk or if they could serve as prognostic biomarkers. We genotyped these SNPs in a Swedish incident CRC series, from which data of Dukes’ stage and long-term survival were available.
|
material and methods |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
population
The analyses were performed on genomic DNA from 308 Swedish CRC cases together with 585 controls. The cases, with the age- and gender-matched controls, were drawn from the population-based Västerbotten intervention project and the mammary screening project, which contain blood samples collected between October 1987 and April 2002 from an ethnically homogenous population living in a geographically defined region in North Sweden [12]. Prospective cases were identified from the cohorts by record linkage to the regional cancer registry. The controls were selected from the same cohort as the corresponding case. They were matched with the case by age at baseline (+/– 6 months) and the time of sampling (+/– 2 months). The controls had to be alive at the time of diagnosis of the corresponding case and without any previous cancer diagnosis, except carcinoma in situ of cervix uteri. The blood samples were stored at –80 °C until the time of sample selection and DNA isolation for genotyping analyses. Clinical data for the CRC cases were retrieved from the registry managed by the Northern Sweden Collaborative Group for Colorectal Diseases (Table 1). Information about the date of death was collected from the Swedish population register, with a follow-up until September 2006. The study was approved by the ethical committee of Karolinska Institute Syd and Umeå University.
|
selection of SNPs
The SNPs were selected for the genotyping analyses based on functional considerations and our previous study on their effect on breast cancer susceptibility and prognosis (manuscript submitted). Functional data were available for the PAI-1 (official HUGO name: SERPINE) rs1799889 –675 4G/5G SNP [13–16]. Additionally, the Transcription Element Search System (TESS) [17] and TRANSFAC [18] databases were used to predict the effect of the promoter SNPs on transcription factor binding sites. In the case of coding SNPs, the PolyPhen website tool [19] was used to predict the impact of the amino acid substitution on the structure and function of the protein. In addition to the PAI-1 SNP, we selected for our analyses the following SNPs: uPA (PLAU) rs2227564 + 54 C/T in exon 6 (amino acid change Leu/Pro); uPAR (PLAUR) rs344781 –516 A/G; and VTN rs704 + 220 C/T in exon 7 (Thr/Met).
genotyping
The selected SNPs were genotyped by Taqman assay by design; primer and probe sequences are available upon request. Taqman probes were ordered from Applied Biosystems (CA, USA). The results were read in ABI PRISM 7900 HT Sequence Detection System and the genotypes were assigned automatically by the software SDS2.1.
statistical analyses
The observed genotype frequencies in the controls were tested for Hardy–Weinberg equilibrium (HWE) and the difference between the observed and expected frequencies was tested for significance using the 
2 test. Statistical significance for the differences in the genotype frequencies between the CRC cases and controls was determined by using the 2 test. Odds ratios (ORs) and 95% confidence intervals (CIs) for associations between genotypes and CRC and tumor characteristics were calculated by conditional logistic regression. For a polymorphism with a risk genotype frequency between 10% and 20%, our study had greater than 80% power to detect an OR of 1.8 to 1.6. The Kaplan–Meier method was used to estimate the survival probabilities, and the log-rank test was used to test differences between subgroups. Statistical analyses were performed using SPSS, Version 14.
|
results |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
no association with CRC susceptibility or Dukes’ stage
Genotype distribution in controls followed HWE for all SNPs genotyped in this study. As shown in Table 2, none of the SNPs conferred an increased risk of developing CRC. The uPAR –516 GG genotype seemed to protect against CRC, even though the association was not statistically significant (OR 0.56, 95% CI 0.30–1.01). Nor did we observe any significant association between the genotypes and Dukes’ stage (Table 2). However, the PAI-1 5G/5G genotype was less common in patients with Dukes’ stage C and D tumors than in patients with Dukes’ stage A and B tumors (16.4% vs 22.6%; OR 0.67, 95% CI 0.35–1.28). Patients with Dukes’ stage C and D tumors were also more often uPA TT genotype carriers than patients with Dukes’ stage A and B tumors (9.2% vs 6.8%; OR 1.32, 95% CI 0.55–3.18).
|
association with survival
The overall survival after 100 months follow-up time was
84% for patients with Dukes' stage A tumors, decreasing with the severity of the disease (Dukes' stage B, 64%; C, 50%; D, 8%) (Figure 1). Stratification of the cases by the PAI-1 genotype showed a better survival for 5G/5G genotype carriers than for carriers of the other genotypes (5G/5G vs 4G/4G, P = 0.058; 5G/5G vs 4G/5G, P = 0.025; see Figure 2A), especially for patients with Dukes’ stage A or B tumors (5G/5G vs 4G/4G, P = 0.023; 5G/5G vs 4G/5G, P = 0.015; see Figure 2B). The difference between the heterozygous 4G/5G genotype carriers and the homozygous 4G/4G genotype carriers among all patients or among patients with Dukes’ stage A or B tumors was not statistically significant (P = 0.638 and P = 0.611, respectively). The genotype did not affect survival among patients with Dukes’ stage C and D tumors (Figure 2C). None of the other SNPs had any effect on survival.
|
|
|
discussion |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
Accumulating evidence suggest that increased activity of the uPA system associates with increased tumor invasion, metastasis and bad prognosis for cancer patients, also with CRC [1, 3–5]. So far, only a few, rather small studies focusing on uPA and PAI-1 SNPs have been published on CRC [20, 21], although SNPs in the regulatory regions of the genes may affect gene expression and further protein expression [13–16]. None of the previous studies have contained data of survival. Our aim was to evaluate the impact of SNPs in several genes of the uPA system on the susceptibility to and prognosis of CRC. The SNPs in the uPA, uPAR, PAI-1 and VTN genes genotyped by us were not associated with the risk of CRC. However, the PAI-1 5G/5G genotype carriers had better survival than 5G/4G or 4G/4G carriers when the patients had Dukes’ stage A or B tumors.
The present study was particularly suitable for survival analysis, because all patients were sampled before the diagnosis of CRC, independently of their genotype. The follow-up time was up to 16 years, allowing analysis of long-term survival. In Sweden, there have been no large changes in the clinical management of CRC since 1990, thus excluding the effects of genotype-specific responses to different treatments on survival. Additionally, the genotype distribution was similar over the follow-up time. Because data of Dukes’ stage of the tumors were available, subdivision of the patients according to the severity of the disease based on the traditional prognostic marker was also possible. In our study more patients with Dukes’ stage A tumors was observed than expected. The participation rate in the Västerbotten Intervention Project was 60–70% of the population. It is possible that there was a selection bias in the recruited population, as individuals aware of their health may have been more prone to participate.
As in all case–control association studies, analysis of multiple SNPs represents a statistical multiple comparison problem. Although there are several methods for handling multiple comparisons in molecular epidemiological studies, no standard approach has been universally adopted [22]. It is also unclear how the likely biological relevance of an SNP should be taken into account. In our study, the statistical significance of the observed association between the PAI-1 4G/5G SNP and survival was not overwhelming. However, there is strong evidence that this SNP is biologically significant [13–16, 23].
The better survival of the PAI-1 5G/5G genotype carriers compared to the 4G allele carriers may be explained by the reported effect of the 4G/5G SNP on free plasma PAI-1 levels [13–15]. We observed the better survival only in patients with Dukes’ stage A or B tumors, possible because in patients with more severe disease many other factors contributing to poor survival may mask the positive effect of an individual SNP. Both the 4G and the 5G alleles bind a transcriptional activator, whereas the 5G allele also binds a repressor protein [13, 14]. As a result, the 4G allele shows an enhanced transcription, and the plasma PAI-1 levels in 4G/4G homozygotes are about 25% higher than in 5G/5G homozygotes [14]. Recently, the 4G/5G polymorphism has been shown to be located within the functional binding elements for the transcription factors Smad and NF-
B, which mediate the effects of the two known inducers of the PAI-1 expression, TGFβ and TNF
, respectively [16, 23].
The relationship of the SNPs in the genes of the uPA system and the prognosis of CRC has been studied only for the PAI-1 4G/5G SNP [20]. In that study, the 4G/4G genotype was associated with more advanced tumors (Dukes’ stage C and D) and the 5G/5G genotype with Dukes’ stage A and B tumors. Also in our study the 4G/4G genotype was more common in patients with Dukes’ stage C and D tumors than the 5G/5G genotype. In patients with Dukes’ stage A and B tumors, 5G/5G was the more common one. However, the association was not statistically significant. Nor were there any statistically significant associations between the other SNPs studied and Dukes’ stage.
In our survival analysis, the PAI-1 –675 5G/5G genotype was associated with a good survival in patients with Dukes’ stage A and B tumors. It has been discussed that new prognostic markers would be needed, especially to identify patients with Dukes’ stage B tumors with poor survival, who would benefit most from adjuvant therapy [9]. Both uPA and uPAR levels have been suggested to predict survival in patients with Dukes’ stage B tumors [9, 24]. Our study and that conducted by Loktionov et al [20] add PAI-1 to the list of putative survival markers for patients with Dukes’ stage B tumors.
Our result of finding no association of the studied polymorphisms with the risk of CRC is concordant with the few previously published studies [20, 21]. A small study in 40 CRC patients did not find any association between uPA +54 C/T SNP and CRC risk [21], although the CC genotype was associated with higher levels of uPA antigens in the CRC samples [25]. For PAI-1, the 4G/5G SNP was not associated with the risk of CRC in a study of 206 CRC patients and 355 healthy controls [20]. The uPAR –516 A/G SNP, located within one of the PEA3/ets transcriptional silencing elements in the uPAR promoter, mediates the effects of integrin β3 on uPAR expression [26, 27]. This may explain the small, non-significant protective effect of the SNP on CRC risk. Taken together, these results suggest that the SNPs in the genes of the uPA system may not be important in the early stage of cancer development. The contributory effect of PAI-1 4G/5G SNP on plasma PAI-1 levels has been evaluated to be 3%, but only among women [28]. With our sample size of 308 CRC patients and 585 controls and the risk genotype frequency of 20%, we had a power of 80% to detect an OR of 1.6. Thus, only major effects of the studied SNPs on the CRC risk could be excluded.
In conclusion, genetic variation may affect prognosis of CRC patients [29]. Our results suggest a better survival for the carriers of the PAI-1 –675 5G/5G genotype than for the carriers of the other genotypes, especially for patients with Dukes’ stage A and B tumors. To evaluate the significance of our findings in the clinic, further independent studies are needed.
|
Acknowledgements |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
We thank 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 Collaborative Group for Colorectal Diseases is appreciated for providing the clinical data. The project was partially funded by Wallenberg Consortium North, Sweden. The study was supported by a grant from the EU (LSHC-CT-2004-503465).
|
Footnotes |
|---|
Current address: Department of Cell Biology, Harvard Medical School, Boston, USA 
Received for publication March 30, 2007. Revision received June 18, 2007. Accepted for publication June 19, 2007.
|
References |
|---|
|
TopAbstractintroductionmaterial and methodsresultsdiscussionAcknowledgementsReferences |
|---|
1. Berger DH. Plasmin/plasminogen system in colorectal cancer. World J Surg (2002) 26:767–771.[CrossRef][Web of Science][Medline]
2. Blasi F, Carmeliet P. uPAR: a versatile signalling orchestrator. Nat Rev Mol Cell Biol (2002) 3:932–943.[CrossRef][Web of Science][Medline]
3. Sidenius N, Blasi F. The urokinase plasminogen activator system in cancer: recent advances and implication for prognosis and therapy. Cancer Metastasis Rev (2003) 22:205–222.[CrossRef][Web of Science][Medline]
4. Duffy MJ, Maguire TM, McDermott EW, O'Higgins N. Urokinase plasminogen activator: a prognostic marker in multiple types of cancer. J Surg Oncol (1999) 71:130–135.[CrossRef][Web of Science][Medline]
5. Andreasen PA, Kjoller L, Christensen L, Duffy MJ. The urokinase-type plasminogen activator system in cancer metastasis: a review. Int J Cancer (1997) 72:1–22.[CrossRef][Web of Science][Medline]
6. Sakakibara T, Hibi K, Koike M, et al. Plasminogen activator inhibitor-1 as a potential marker for the malignancy of colorectal cancer. Br J Cancer (2005) 93:799–803.[CrossRef][Web of Science][Medline]
7. Yang JL, Seetoo D, Wang Y, et al. Urokinase-type plasminogen activator and its receptor in colorectal cancer: independent prognostic factors of metastasis and cancer-specific survival and potential therapeutic targets. Int J Cancer (2000) 89:431–439.[CrossRef][Web of Science][Medline]
8. Nielsen HJ, Pappot H, Christensen IJ, et al. Association between plasma concentrations of plasminogen activator inhibitor-1 and survival in patients with colorectal cancer. Br Med J (1998) 316:829–830.
9. Stephens RW, Nielsen HJ, Christensen IJ, et al. Plasma urokinase receptor levels in patients with colorectal cancer: relationship to prognosis. J Natl Cancer Inst (1999) 91:869–874.
10. Abe J, Urano T, Konno H, et al. Larger and more invasive colorectal carcinoma contains larger amounts of plasminogen activator inhibitor type 1 and its relative ratio over urokinase receptor correlates well with tumor size. Cancer (1999) 86:2602–2611.[CrossRef][Web of Science][Medline]
11. Grebenchtchikov N, Maguire TM, Riisbro R, et al. Measurement of plasminogen activator system components in plasma and tumor tissue extracts obtained from patients with breast cancer: an EORTC Receptor and Biomarker Group collaboration. Oncol Rep (2005) 14:235–239.[Web of Science][Medline]
12. 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) 13:307–316.[CrossRef][Web of Science][Medline]
13. Dawson SJ, Wiman B, Hamsten A, et al. The two allele sequences of a common polymorphism in the promoter of the plasminogen activator inhibitor-1 (PAI-1) gene respond differently to interleukin-1 in HepG2 cells. J Biol Chem (1993) 268:10739–10745.
14. Eriksson P, Kallin B, van 't Hooft FM, et al. Allele-specific increase in basal transcription of the plasminogen-activator inhibitor 1 gene is associated with myocardial infarction. Proc Natl Acad Sci USA (1995) 92:1851–1855.
15. Francis CW. Plasminogen activator inhibitor-1 levels and polymorphisms. Arch Pathol Lab Med (2002) 126:1401–1404.[Web of Science][Medline]
16. Swiatkowska M, Szemraj J, Cierniewski CS. Induction of PAI-1 expression by tumor necrosis factor alpha in endothelial cells is mediated by its responsive element located in the 4G/5G site. Febs J (2005) 272:5821–5831.[CrossRef][Medline]
17. TESS. http://www.cbil.upenn.edu/tess.
18. TRANSFAC. http://thr.cit.nih.gov/molbio/signal/.
19. PolyPhen. http://tux.embl-heidelberg.de/ramensky/polyphen.cgi.
20. Loktionov A, Watson MA, Stebbings WS, et al. Plasminogen activator inhibitor-1 gene polymorphism and colorectal cancer risk and prognosis. Cancer Lett (2003) 189:189–196.[CrossRef][Web of Science][Medline]
21. Przybylowska K, Smolarczyk K, Blasiak J, et al. A C/T polymorphism in the urokinase-type plasminogen activator gene in colorectal cancer. J Exp Clin Cancer Res (2001) 20:569–572.[Web of Science][Medline]
22. Thomas DC, Clayton DG. Betting odds and genetic associations. J Natl Cancer Inst (2004) 96:421–423.
23. Chen YQ, Sloan-Lancaster J, Berg DT, et al. Differential mechanisms of plasminogen activator inhibitor-1 gene activation by transforming growth factor-beta and tumor necrosis factor-alpha in endothelial cells. Thromb Haemost (2001) 86:1563–1572.[Web of Science][Medline]
24. Mulcahy HE, Duffy MJ, Gibbons D, et al. Urokinase-type plasminogen activator and outcome in Dukes’ B colorectal cancer. Lancet (1994) 344:583–584.[CrossRef][Web of Science][Medline]
25. Przybylowska K, Smolarczyk K, Kulig A, et al. Antigen levels of the urokinase-type plasminogen activator and its gene polymorphisms in colorectal cancer. Cancer Lett (2002) 181:23–30.[CrossRef][Web of Science][Medline]
26. Hapke S, Gawaz M, Dehne K, et al. Beta(3)A-integrin downregulates the urokinase-type plasminogen activator receptor (u-PAR) through a PEA3/ets transcriptional silencing element in the u-PAR promoter. Mol Cell Biol (2001) 21:2118–2132.
27. Soravia E, Grebe A, De Luca P, et al. A conserved TATA-less proximal promoter drives basal transcription from the urokinase-type plasminogen activator receptor gene. Blood (1995) 86:624–635.
28. Henry M, Tregouet DA, Alessi MC, et al. Metabolic determinants are much more important than genetic polymorphisms in determining the PAI-1 activity and antigen plasma concentrations: a family study with part of the Stanislas Cohort. Arterioscler Thromb Vasc Biol (1998) 18:84–91.
29. Hunter KW, Crawford NP. Germ line polymorphism in metastatic progression. Cancer Res (2006) 66:1251–1254.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||