TGFB1 and TGFBR1 polymorphic variants in relationship to bladder cancer risk and prognosis

Authors

  • Adela Castillejo,

    1. Grupo de Oncología Molecular, Hospital General Universitario de Elche, Elche, Spain
    Search for more papers by this author
  • Nathaniel Rothman,

    1. Department of Health and Human Services, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
    Search for more papers by this author
  • Cristiane Murta-Nascimento,

    1. Programa de Recerca en Càncer, Institut Municipal d'Investigació Mèdica (IMIM-Hospital del Mar), Barcelona, Spain
    2. Centre for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
    Search for more papers by this author
  • Núria Malats,

    1. Programa de Recerca en Càncer, Institut Municipal d'Investigació Mèdica (IMIM-Hospital del Mar), Barcelona, Spain
    2. Centre for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
    3. Programas de Genética del Cáncer Humano y Patología Molecular, Centro Nacional de Investigaciones Oncológicas, Madrid, Spain
    Search for more papers by this author
  • Montserrat García-Closas,

    1. Department of Health and Human Services, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
    Search for more papers by this author
  • Angeles Gómez-Martínez,

    1. Grupo de Oncología Molecular, Hospital General Universitario de Elche, Elche, Spain
    Search for more papers by this author
  • Josep Lloreta,

    1. Departament de Patologia, Hospital del Mar, Barcelona, Spain
    2. Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Barcelona, Spain
    Search for more papers by this author
  • Adonina Tardón,

    1. Departamento de Epidemiología y Medicina Preventiva, Universidad de Oviedo, Oviedo, Spain
    2. CIBER en Epidemiología y Salud Pública (CIBERESP), Spain
    Search for more papers by this author
  • Consol Serra,

    1. Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Barcelona, Spain
    2. Consorci Hospitalari Parc Taulí, Sabadell, Spain
    Search for more papers by this author
  • Reina García-Closas,

    1. Department of Social Medicine, Unidad de Investigación, Hospital Universitario de Canarias, La Laguna, Spain
    Search for more papers by this author
  • Stephen Chanock,

    1. Core Genotype Facility, Advanced Technology Center, National Cancer Institute, Gaithersberg, MD
    Search for more papers by this author
  • Debra T. Silverman,

    1. Department of Health and Human Services, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
    Search for more papers by this author
  • Mustafa Dosemeci,

    1. Department of Health and Human Services, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, MD
    Search for more papers by this author
  • Manolis Kogevinas,

    1. Programa de Recerca en Càncer, Institut Municipal d'Investigació Mèdica (IMIM-Hospital del Mar), Barcelona, Spain
    2. Centre for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
    3. CIBER en Epidemiología y Salud Pública (CIBERESP), Spain
    4. University of Crete, Heraklion, Greece
    Search for more papers by this author
  • Alfredo Carrato,

    1. Grupo de Oncología Molecular, Hospital General Universitario de Elche, Elche, Spain
    Search for more papers by this author
    • Alfredo Carrato, José Luis Soto and Francisco X. Real have contributed equally to this work and share senior authorship.

  • José Luis Soto,

    1. Grupo de Oncología Molecular, Hospital General Universitario de Elche, Elche, Spain
    Search for more papers by this author
    • Alfredo Carrato, José Luis Soto and Francisco X. Real have contributed equally to this work and share senior authorship.

  • Francisco X. Real

    Corresponding author
    1. Programa de Recerca en Càncer, Institut Municipal d'Investigació Mèdica (IMIM-Hospital del Mar), Barcelona, Spain
    2. Programas de Genética del Cáncer Humano y Patología Molecular, Centro Nacional de Investigaciones Oncológicas, Madrid, Spain
    3. Departament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Barcelona, Spain
    • Centro Nacional de Investigaciones Oncológicas (CNIO), C/ Melchor Fernández Almagro 3, E-28029 Madrid, Spain
    Search for more papers by this author
    • Alfredo Carrato, José Luis Soto and Francisco X. Real have contributed equally to this work and share senior authorship.


Abstract

The transforming growth factor-beta (TGF-β) signalling pathway plays an important role in tumor development and progression. We aimed at analyzing whether 7 different common variants in genes coding for 2 key members of the TGF-β signalling pathway (TGFB1 and TGFBR1) are associated with bladder cancer risk and prognosis. A total of 1,157 cases with urothelial cell carcinoma of the bladder and 1,157 matched controls where genotyped for 3 single nucleotide polymorphisms (SNPs) in TGFB1 (rs1982073, rs1800472, rs1800471) and an additional 3 SNPs and 1 indel polymorphism in TGFBR1 (rs868, rs928180, rs334358 and rs11466445, respectively). In the case-control study, we estimated odds ratios and 95% confidence intervals for each individual genetic variant using unconditional logistic regression adjusting for age, gender, study area and smoking status. Survival analysis was performed using the Kaplan-Meier method and Cox models. The endpoints of interest were tumor relapse, progression and death from bladder cancer. All the SNPs analyzed showed a similar distribution among cases and controls. The distribution of the TGFBR1*6A allele (rs11466445) was also similar among cases and controls, indicating no association with bladder cancer risk. Similarly, none of the haplotypes was significantly associated with bladder cancer risk. Among patients with muscle-invasive tumors, we found a significant association between TGFBR1-rs868 and disease-specific mortality with an allele dosage effect (p-trend = 0.003). In conclusion, the genetic variants analyzed were not associated with an increased risk of bladder cancer. The association of TGFBR1-rs868 with outcome should be validated in independent patient series. © 2008 Wiley-Liss, Inc.

The transforming growth factor-beta (TGF-β) pathway plays an important role in tumor development and progression. There is extensive evidence that TGF-β has antiproliferative effects and can hinder the early steps of tumor progression in several cell types. By contrast, recent evidence has suggested that TGF-β can also play an important role in later steps of tumor progression where it accelerates invasion and metastasis.1, 2

The TGF-β pathway is altered in a wide variety of cancers, including urinary bladder tumors. TGF-β receptors (TGFBR) and other proteins involved in TGF-β signalling, such as Smads, have been shown to be targets of somatic mutations in several cancers including those of the colorectum, pancreas and lung. Furthermore, altered levels of ligand and receptors have been described in the neoplastic tissue.3

More recently, there has been interest in the contribution of genetic variation in this pathway to cancer susceptibility. Several polymorphic variants in genes involved in this pathway have been studied. Single nucleotide polymorphisms (SNPs) in TGFB1 (exon 1: 327 C>T, rs1982073; exon 5: 73 C>T, rs1800472), coding for 1 of the secreted ligands, have been reported to be associated with breast, colorectal, lung and nasopharyngeal carcinomas and non-Hodgkin lymphomas.4–7 A larger number of studies have dealt with genetic variation in TGFBR1, which codes for 1 of the 2 subunits of the heterodimeric membrane receptor kinase.8 The most extensively studied variant allele is TGFBR1*6A, rs11466445, which corresponds to a trinucleotide repeat coding for an Ala repeat. The most common allele among Caucasians contains 9 Ala residues whereas the 6 Ala variant is less common and has been reported to be a susceptibility allele for colorectal and breast cancer.9, 10 Somatic acquisition of this trinucleotide variant has also been reported in colon cancer.11 Very few studies have analyzed its association with bladder cancer risk and their results are not conclusive.12

In this study, we analyzed in a prospectively recruited large case-control population whether 7 different common polymorphisms in genes coding for 2 key members of the TGF-β signalling pathway (TGFB1 and TGFBR1) are associated with bladder cancer risk as well as with patient's prognosis.

Material and methods

Study population

A detailed description of the study population has been reported elsewhere.13 In brief, 1,219 patients with a new diagnosis of bladder cancer, aged 21–80 years (mean = 66 years, SD = 10), were recruited in 18 hospitals in Spain between 1998 and 2001. Sections of paraffin-embedded blocks from the initial tumor were reviewed by a panel of pathologists using the 1999 World Health Organization classification for urothelial lesions14 with modifications as described elsewhere.15 Controls were 1,271 subjects selected from participating hospitals with diagnoses thought to be unrelated to the exposures of interest, individually matched to the cases by age at diagnosis (±5 years), gender, ethnicity and study area. Blood and/or buccal-cell samples were provided by 1,188 cases and 1,173 controls. Exclusions were made to reduce heterogeneity (cases with non-urothelial histology and non-White subjects) or because of low amounts of DNA. The final study population available for the case-control analysis was 1,157 cases and 1,157 controls.16 Mean age of cases and controls was 66 and 65 years, respectively; 13% of subjects were female. Forty-six percent and 47% of cases and controls, respectively, had attained primary education and there were no significant differences among the 2 groups regarding this variable. Fourteen percent and 28% of cases and controls were non-smokers, respectively. Eighty-two percent and 63% of cases and controls were former or current smokers, respectively. The remaining subjects in each group were classified as occasional smokers.

The survival analyses included the 1,105 cases for whom pathological review of the initial tumor was possible, of whom 859 (77.7%) had non-muscle invasive and 246 (22.3%) had muscle-invasive tumors. All study subjects were interviewed during their first hospital admission. Patient's clinical records were reviewed annually in order to obtain information about the outcomes of interest and any treatment change. Furthermore, a telephone interview was performed on an annual basis to complete information on tumor progression and patient's vital status. Last complete follow-up was as of December 31, 2005.15

For the survival analyses, the endpoints of interest were: tumor relapse, defined as reoccurrence of a tumor of any stage or grade after transurethral resection (TUR); tumor progression, defined as development of a muscle-invasive tumor—for patients with non-muscle invasive bladder cancer—or any tumor progression event for patient with muscle-invasive bladder cancer; and disease-specific mortality (DSM), considering the time from TUR to death from bladder cancer. Patients who died because of bladder cancer without having been diagnosed with a tumor progression were considered uncensored at the midpoint period between TUR and death for estimation of time to progression. For patients with non-muscle invasive tumors, recurrence was defined as the reappearance of a non-muscle invasive tumor (pTa or pT1) and progression as the development of a muscle invasive mass (≥pT2). For patients whose tumors were initially classified as muscle-invasive, any tumor reappearance after treatment was considered progression, regardless of its location. The median follow-up time for patients who were free-of-disease at the end of follow-up was 64.4 months (range 0.2–90.2 months).

Genotype assays

A total of 6 SNPs in TGFB1 (rs1982073, rs1800472 and rs1800471) and TGFBR1 (rs868, rs928180 and rs334358) were selected from the SNP500Cancer database (http://snp500cancer.nci.nih.gov) for genotyping. In addition, an insertion/deletion polymorphism (indel) in TGFBR1 (rs11466445) was studied.

SNPs in TGFB1 and TGFBR1 were investigated using germline DNA as previously described.16 The SNP genotype assays were performed at the Core Genotyping Facility at the US National Cancer Institute using TaqMan® assays (Applied Biosystems, Foster City, CA) for TGFB1 Ex1-327C>T (rs1982073) and Ex5-73C>T (rs1800472) and a GoldenGate® assay (Illumina®, San Diego, CA) for TGFB1 Ex1-282C>G (rs1800471), TGFBR1 IVS3-intronic SNPs 2409A>G (rs928180) and IVS8+547G>T (rs334358) and Ex9+195A>G (rs868) in the 3′ untranslated region.

In addition, the TGFBR1*6A (rs11466445) exon 1 polymorphic variant allele, leading to the deletion of 3 Ala residues from a 9-Ala (9A) stretch in the wild-type allele, was genotyped. The length of this GCG repeat was determined by PCR analysis with a 5′ fluorescence-labelled forward primer and capillary electrophoresis. Details on genotyping assays have been reported elsewhere.10, 11

Statistical analysis

Hardy-Weinberg equilibrium was checked among the control population. In the case-control study, we estimated odds ratio (OR) and 95% confidence interval (95% CI) for each individual SNP using unconditional logistic regression adjusting for age, gender, study area and smoking status (non-smoker, occasional, former and current smoker). Interaction effects between SNPs and smoking status were assessed by fitting the model with and without the interaction parameters and conducting a likelihood ratio test. Haplotype frequencies, OR and 95% CI for genes showing blocks of linkage disequilibrium were estimated using SNPStats (http://bioinfo.iconcologia.net/SNPstats).17

Survival curves were estimated using the Kaplan-Meier product limit method and the differences between categories of each variable were assessed using the log-rank and Breslow tests. Hazard ratios (HR) and 95% CI were estimated using Cox models. The following variables were considered for adjustment: geographical area, gender, age, stage, grade, tumor size, tumor location in the bladder, number of tumors, presence of metastases and treatment. For each of the analyses reported, variables used to adjust the final model are specified in the table footnotes. Survival analyses were performed separately for non-muscle invasive and muscle-invasive tumors. Haplotype effects in censored data analysis were estimated using the Thesias 3.1 software (http://genecanvas.ecgene.net).18

To correct for false discovery rate (FDR), the Benjamini-Hochberg test was applied for each of the p-trend values in Tables I, III and IV.19

Table I. Association of TGFB1 and TGFBR1 Polymorphisms with Bladder Cancer Risk
SNPCasesControlsOR195% CIp-LRT2p-trend
  • NA, not applicable.

  • 1

    Adjusted for geographical area, gender, age and smoking status.

  • 2

    p-LRT, p value from the likelihood ratio test.

TGFB1
rs1982073    0.6480.768
Ex1-327C>T
 TT3763771.00Ref.  
 CT5174631.090.90–1.33  
 CC1891751.010.78–1.31  
rs1800472    0.270NA
Ex5-73C>T
 CC10139491.00Ref.  
 CT81711.100.77–1.55  
 TT01NANA  
rs1800471    0.2200.081
Ex1-282C>G
 GG9769141.00Ref.  
 CG1291420.800.61–1.05  
 CC790.700.25–1.94  
TGFBR1
rs868    0.1020.475
Ex9 + 195A>G
 AA6806391.00Ref.  
 AG3513161.070.88–1.30  
 GG41600.660.43–1.02  
rs928180    0.9180.779
IVS3-2409A>G
 AA9058491.00Ref.  
 AG1751790.960.75–1.22  
 GG551.140.29–4.42  
rs334358    0.2390.682
IVS8 + 547G>T
 GG6836501.00Ref.  
 GT3593271.070.88–1.29  
 TT42560.730.47–1.12  
TGFBR1 exon 1    0.7440.704
 *9A/*9A8878121.00Ref.  
 *6A/*9A1991910.990.79–1.25  
 *6A/*6A8110.680.26–1.81  

Results

Genetic variation in TGFB1 and TGBR1 and bladder cancer susceptibility

Genotype distribution in the control population did not deviate significantly from that expected for a population in Hardy-Weinberg equilibrium. The 3 TGFB1 SNPs analyzed showed a similar distribution among cases and controls; a lower, non-significant, risk of bladder cancer was found for the rs1800471 genotype (p for trend = 0.081) (Table I). Regarding the TGFBR1 polymorphisms analyzed, several were in strong positive linkage disequilibrium: rs868 with rs334358 (D′ = 0.99, r2 = 0.99) and rs928180 with rs11466445 (D′ = 0.98, r2 = 0.91). Therefore, only the results for those with the least amount of missing data (rs334358 and rs11466445) were included in the haplotype analysis. All the SNPs analyzed showed a similar distribution among cases and controls (Table I). The distribution of the TGFBR1*6A allele was also similar among cases and controls (Table I) (p = 0.744), indicating no association with bladder cancer risk. Similarly, none of the haplotypes was significantly associated with bladder cancer risk (Table II).

Table II. Association of TGFB1 and TGFBR1 Haplotypes with Bladder Cancer Risk
PolymorphismsHaplotype frequencyOR195% CIp-value2Global p-value
CasesControls
  • 1

    Adjusted for geographical area, gender, age and smoking status.

  • 2

    Simulated maximum score p-value.

TGFB1      
rs1982073rs1800472rs1800471     0.270
 TCG0.590.601.00Ref.  
 CCC0.310.291.070.93–1.230.38 
 CCC0.060.070.830.65–1.050.13 
 CTG0.040.041.060.75–1.500.75 
TGFBR1      
rs928180rs334358     0.730
 AG0.710.691.00Ref.  
 AT0.200.210.950.82–1.110.52 
 GG0.090.090.930.74–1.170.56 

Genetic variation in TGFB1 and TGFBR1 and bladder cancer prognosis

In the survival analyses, the Cox model for TGFB1 and TGFBR1 polymorphisms rendered no significant association with tumor relapse and progression among cases with non-muscle invasive bladder tumors (Table III). By contrast, in patients with muscle-invasive tumors, we found a significant association between TGFBR1 rs868 and DSM. The G allele was an independent predictor of bladder cancer mortality with an allele dosage effect: the HR were 1.85 (95% CI: 1.15–2.97) for heterozygous patients (A/G) and 3.00 (95% CI: 1.15–7.82) for homozygous G/G patients (p-trend = 0.003) (Table IV). SNP rs334358 also showed an association with DSM, as expected because of the linkage disequilibrium with rs868. FDR values for these 2 SNP associations were 0.036 and 0.054, respectively, based on the p-values for trend (1 df) of all variants evaluated in this study. Regarding haplotype analyses, we found a statistically significant increase in risk of death from bladder cancer only for patients with invasive tumors carrying alleles IVS3-2409A (rs928180) and IVS8+547T (rs334358): HR = 1.50 (95% CI: 1.04–2.16) (p = 0.030) (Supp. Info. Table I).

Table III. Cox Model for TGFB1 and TGFBR1 Polymorphisms and Outcome in Patients with Non-Muscle Invasive Bladder Tumors
SNPCasesTumor relapse1Progression2
NHR95% CIp-LRTp-trendNHR95% CIp-LRTp-trend
  • NA, not applicable.

  • 1

    Adjusted for gender, stage-grade, tumor size, number of tumors and treatment.

  • 2

    Adjusted for stage-grade and tumor site.

TGFB1
rs1982073    0.3790.182   0.3640.236
 TT2801101.00Ref.  251.00Ref.  
 CT3751280.930.72–1.20  340.960.57–1.61  
 CC143410.780.54–1.12  100.610.29–1.29  
rs1800472    0.674NA   0.671NA
 CC7432561.00Ref.  621.00Ref.  
 CT62211.100.70–1.73  50.820.33–2.06  
 TT00NANA  0NANA  
rs1800471    0.9160.709   0.643NA
 GG7132471.00Ref.  581.00Ref.  
 CG100361.080.76–1.54  111.240.65–2.39  
 CC621.000.25–4.05  0NANA  
TGFBR1
rs868    0.0880.811   0.1430.799
 AA5091741.00Ref.  391.00Ref.  
 AG259951.190.92–1.54  281.380.84–2.24  
 GG2650.520.21–1.28  10.330.04–2.41  
rs928180    0.9470.889   0.199NA
 AA6692301.00Ref.  541.00Ref.  
 AG131470.990.72–1.36  151.220.68–2.16  
 GG310.730.10–5.30  0NANA  
rs334358         0.1720.860
 GG5081711.00Ref.0.1070.551391.00Ref.  
 GT267991.230.95–1.58  281.330.82–2.17  
 TT2760.630.28–1.43  10.330.05–2.42  
TGFBR1 exon 1         0.232NA
 *9A/*9A6552491.00Ref.0.3630.922511.00Ref.  
 *6A/*9A142531.070.79–1.45  171.310.75–2.30  
 *6A/*6A610.330.05–2.39  0NANA  
Table IV. Cox Model for TGFB1 and TGFBR1 Polymorphisms and Outcome in Patients with Muscle Invasive Bladder Tumors
SNPCasesProgression1Disease-specific mortality2
NHR95% CIp-LRTp-trendNHR95% CIp-LRTp-trend
  • NA, not applicable.

  • 1

    Adjusted for stage, localization, metastasis and treatment.

  • 2

    Adjusted for age, stage, localization and metastasis.

TGFB1
rs1982073    0.9310.796   0.5330.718
 TT78411.00Ref.  331.00Ref.  
 CT116651.090.69–1.71  490.890.56–1.39  
 CC35171.050.56–1.98  151.260.67–2.35  
rs1800472    0.059NA   0.073NA
 CC2181131.00Ref.  881.00Ref.  
 CT16122.021.03–3.97  111.900.99–3.65  
 TT0NANA  0NANA  
rs1800471    0.102NA   0.080NA
 GG2151191.00Ref.  961.00Ref.  
 CG2270.460.20–1.06  40.400.15–1.11  
 CC10NANA  0NANA  
TGFBR1
rs868    0.5310.341   0.0130.003
 AA134671.00Ref.  521.00Ref.  
 AG81441.290.83–2.03  351.851.15–2.97  
 GG1061.160.44–3.03  53.001.15–7.82  
rs334358    0.8010.547   0.0320.009
 GG137701.00Ref.  541.00Ref.  
 GT80421.160.74–1.82  341.671.05–2.68  
 TT1061.130.43–2.95  52.831.09–7.34  
rs928180    0.2590.259   0.813
 AA191991.00Ref.  771.00Ref.  
 AG36191.390.80–2.41  161.070.61–1.87  
 GG00NANA  0NANA  
TGFBR1 exon 1    0.752NA   0.943NA
 *9A/*9A165861.00Ref.  811.00Ref.  
 *6A/*9A41211.090.65–1.81  211.020.61–1.69  
 *6A/*6A00NANA  0NANA  

Discussion

There is extensive evidence that the TGF-β pathway plays an important role in cancer development and progression. Although TGF-β has an antitumor role in early carcinogenesis, there is extensive evidence that it can foster progression at later stages of tumor evolution.1, 2 For these reasons, there is a great interest in analyzing the role of genetic variation in TGFB and the TGFB receptor in cancer susceptibility and disease progression.

In this study, we selected TGFB1 polymorphisms leading to non-synonymous amino acid changes. SNPs rs1982073 (Leu10Pro) and rs1800471 (Arg25Pro), coding for residues located in the signal peptide, have been related with TGF-β1 expression levels in leukocytes.20, 21 In addition, an association between these polymorphisms and the risk of developing breast, colorectal, lung and nasopharyngeal carcinomas has been reported.4, 7 SNP rs1800472 (Thr263Ile) is located at exon 5 and the corresponding residue lies near the site where the latency-associated peptide is cleaved from the active peptide. Therefore, this polymorphism may be related to the activation of TGF-β1, as was previously suggested.22

The TGFBR1 polymorphisms analyzed include 2 intronic SNPs (IVS3-2409A>G and IVS8+54G>T) and 1 in the 3′ untranslated region (Ex9+195A>G). To our knowledge, these polymorphisms have not previously been associated with cancer risk. The fourth polymorphism analyzed in this gene is a coding indel in exon 1 within the leader peptide sequence. The TGFBR1*6A allele functions as a less effective mediator of TGF-β-antiproliferative signals.10 There have been many studies analyzing cancer susceptibility associated with this allele in patients with a variety of tumors. Individuals carrying the TGFBR1*6A allele have been reported to have an increased risk for breast and ovarian cancers.9 Whether this allele confers an increased susceptibility to colorectal cancer is controversial.9–12, 23 Until now, the evidence on the risk of bladder cancer associated with the TGFBR1*6A allele has been inconclusive. Only a few studies have been reported, all of which included a very small number of cases and controls.12 Van Tilborg et al.24 examined 146 patients with transitional carcinoma of the bladder and 183 controls not matched for age, sex or ethnicity and found no association with bladder cancer risk. Our study, the largest of this association reported to date, does not show an association for any of the polymorphisms analyzed and bladder cancer risk. Because of its large size and the largely unbiased nature of the patient population, the results of this study are strongly suggestive of the null effect of the studied polymorphisms in bladder cancer susceptibility. However, we cannot rule out that other genetic variants in these genes may confer a risk for this cancer.

The finding that the TGFBR1 rs868 allele is strongly and independently associated with DSM is potentially relevant. The strong experimental evidence that TGF-β plays an important role fostering progression at late stages of tumor evolution would be consistent with this association.1, 2 This result warrants further investigation in larger series to test possible explanations such as the selective involvement in regional vs. distant metastatic spread and an association with treatment response. Little is known regarding the functional role of the rs868 variant that would support its causal association with DSM; it is also possible that other genetic variants in linkage disequilibrium with rs868 account for the association observed.

Our study was hypothesis-driven and involved a small number of genetic variants. Based on FDR calculations, the associations with DSM are likely not to be due to chance19 although further work is required to confirm this association in independent patient series. The establishment of polymorphic variants of genes in the TGF-β pathway as prognostic markers may contribute to an improved management of bladder cancer patients.

Acknowledgements

We thank the nurses, physicians and patients who contributed to the Spanish Bladder Cancer/EPICURO Study at the participating hospitals and Ms. A. Alfaro and Ms. G. Carretero for technical support.

APPENDIX—PARTICIPATING STUDY CENTERS IN SPAIN

Institut Municipal d'Investigació Mèdica, Universitat Pompeu Fabra, Barcelona – Coordinating Center (M. Sala, G. Castaño, M. Torà, D. Puente, C. Villanueva, C. Murta-Nascimento, J. Fortuny, E. López, S. Hernández, R. Jaramillo); Hospital del Mar, Universitat Autònoma de Barcelona, Barcelona (J. Lloreta, S. Serrano, L. Ferrer, A. Gelabert, J. Carles, O. Bielsa, K. Villadiego), Hospital Germans Trias i Pujol, Badalona, Barcelona (L. Cecchini, J.M. Saladié, L. Ibarz); Hospital de Sant Boi, Sant Boi de Llobregat, Barcelona (M. Céspedes); Consorci Hospitalari Parc Taulí, Sabadell (D. García, J. Pujadas, R. Hernando, A. Cabezuelo, C. Abad, A. Prera, J. Prat); Centre Hospitalari i Cardiològic, Manresa, Barcelona (M. Domènech, J. Badal, J. Malet); Hospital Universitario de Canarias, La Laguna, Tenerife (J. Rodríguez de Vera, A.I. Martín); Hospital Universitario Nuestra Señora de la Candelaria, Tenerife (F.J. Taño, F. Cáceres); Hospital General Universitario de Elche, Universidad Miguel Hernández, Elche, Alicante (F. García-López, M. Ull, A. Teruel, E. Andrada, A. Bustos, A. Castillejo, J.L. Soto); Universidad de Oviedo, Oviedo, Asturias; Hospital San Agustín, Avilés, Asturias (J.L. Guate, J.M. Lanzas, J. Velasco); Hospital Central Covadonga, Oviedo, Asturias (J.M. Fernández, J.J. Rodríguez, A. Herrero), Hospital Central General, Oviedo, Asturias (R. Abascal, C. Manzano, T. Miralles); Hospital de Cabueñes, Gijón, Asturias (M. Rivas, M. Arguelles); Hospital de Jove, Gijón, Asturias (M. Díaz, J. Sánchez, O. González); Hospital de Cruz Roja, Gijón, Asturias (A. Mateos, V. Frade); Hospital Alvarez-Buylla (Mieres, Asturias): P. Muntañola, C. Pravia; Hospital Jarrio, Coaña, Asturias (A.M. Huescar, F. Huergo); Hospital Carmen y Severo Ochoa, Cangas, Asturias (J. Mosquera).

Ancillary