PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Eur Urol. Author manuscript; available in PMC Nov 18, 2011.
Published in final edited form as:
PMCID: PMC3220186
NIHMSID: NIHMS282864
Genetic Susceptibility to Distinct Bladder Cancer Subphenotypes
Lin T. Guey,ae Montserrat García-Closas,b Cristiane Murta-Nascimento,e Josep Lloreta,cd Laia Palencia,e Manolis Kogevinas,efgh Nathaniel Rothman,b Gemma Vellalta,e Malu Calle,i Gaëlle Marenne,a Adonina Tardón,j Alfredo Carrato,k Reina García-Closas,l Consol Serra,m Debra T Silverman,b Stephen Chanock,b Francisco X. Real,ade and Núria Malatsae*, for the EPICURO/Spanish Bladder Cancer Study investigatorsn
aSpanish National Cancer Research Centre (CNIO), Madrid, Spain
bDivision of Cancer Epidemiology and Genetics, National Cancer Institute, Department of Health and Human Services, Bethesda, MD, USA
cDepartament de Patologia, Hospital del Mar – IMAS, Barcelona, Spain
dDepartament de Ciències Experimentals i de la Salut, Universitat Pompeu Fabra, Barcelona, Spain
eInstitut Municipal d’Investigació Mèdica – Hospital del Mar, Barcelona, Spain
fCentre for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
gUniversity of Crete, Heraklion, Greece
hCIBER en Epidemiología y Salud Pública (CIBERESP), Barcelona, Spain
iUniversity of Vic, Vic, Spain
jUniversidad de Oviedo, Oviedo, Spain
kHospital Universitario de Elche, Elche, Spain
lUnidad de Investigación, Hospital Universitario de Canarias, La Laguna, Spain
mConsorci Hospitalari Parc Taulí, Sabadell, Spain
*Corresponding author Spanish National Cancer Research Centre (CNIO), Melchor Fernández Almagro, 3, E-28029, Madrid, Spain, nuria/at/cnio.es
nEPICURO/Spanish Bladder Cancer Study Investigators in Annex 1
Background
Clinical, pathologic, and molecular evidence indicate that bladder cancer is heterogeneous with pathologic/molecular features that define distinct subphenotypes with different prognoses. It is conceivable that specific patterns of genetic susceptibility are associated with particular subphenotypes.
Objective
To examine evidence for the contribution of germline genetic variation to bladder cancer heterogeneity.
Design, setting, and participants
The Spanish Bladder Cancer/EPICURO Study is a case-control study based in 18 hospitals located in five areas in Spain. Cases were patients with a newly diagnosed, histologically confirmed, urothelial cell carcinoma of the bladder from 1998 to 2001. Case diagnoses were reviewed and uniformly classified by pathologists following the World Health Organisation/International Society of Urological Pathology 1999 criteria. Controls were hospital-matched patients (n = 1149).
Measurements
A total of 1526 candidate variants in 423 candidate genes were analysed. Three distinct subphenotypes were defined according to stage and grade: low-grade nonmuscle invasive (n = 586), high-grade nonmuscle invasive (n = 219), and muscle invasive (n = 246). The association between each variant and subphenotype was assessed by polytomous risk models adjusting for potential confounders. Heterogeneity in genetic susceptibility among subphenotypes was also tested.
Results and limitations
Two established bladder cancer susceptibility genotypes, NAT2 slow-acetylation and GSTM1-null, exhibited similar associations among the subphenotypes, as did VEGF-rs25648, which was previously identified in our study. Other variants conferred risks for specific tumour subphenotypes such as PMS2-rs6463524 and CD4-rs3213427 (respective heterogeneity p values of 0.006 and 0.004), which were associated with muscle-invasive tumours (per-allele odds ratios [95% confidence interval] of 0.56 [0.41–0.77] and 0.71 [0.57–0.88], respectively) but not with non–muscle-invasive tumours. Heterogeneity p values were not robust in multiple testing according to their false-discovery rate.
Conclusions
These exploratory analyses suggest that genetic susceptibility loci might be related to the molecular/pathologic diversity of bladder cancer. Validation through large-scale replication studies and the study of additional genes and single nucleotide polymorphisms are required.
Keywords: Urinary bladder cancer, Genetic polymorphism, Heterogeneity, Tumour subphenotypes, Pathologic characteristics
Urothelial cell carcinoma of the bladder (UCCB) is the fifth most common cancer in men; it occurs with a male-to-female ratio of approximately 3:1. Spain has one of the highest incidence rates among men (55 per 100 000) yet one of the lowest among women (7.4 per 100 000), with a male-to-female ratio of 7:1 [1]. Because of its commonly indolent course, UCCB is one of the most expensive cancers in terms of medical care costs per patient [2].
Pathologic and molecular evidence suggest that UCCB is not a unique phenotype. Based on stage, tumours are classified as non–muscle invasive and muscle invasive, the former representing approximately 80% of cases at presentation. Non–muscle-invasive tumours are subclassified into low and high grade. Tumours of these subphenotypes have different biologic behaviour and are associated with distinct patient prognosis [35]. This classification has been supported by the identification of molecular alterations that identify two major genetic pathways leading to low-grade, genomic stable and high-grade, genomic unstable tumours, the latter including a proportion of non–muscle-invasive tumours and the vast majority of muscle-invasive tumours. Activating mutations in FGFR3 and PI3KCA are associated with low-grade non–muscle-invasive tumours, whereas p53 and RB pathway alterations are critical for high-grade nonmuscle-invasive and muscle-invasive tumours (Fig. 1) [6,7]. However, the precise mechanisms through which tumours evolve need to be more precisely defined.
Fig. 1
Fig. 1
A model for bladder cancer development and progression. Two pathways are defined according to distinctive bladder cancer subphenotypes: The first pathway is composed of low-grade non–muscle-invasive (Ta G1/Ta G2)—genomic stable—tumours; (more ...)
The aetiology of UCCB is multifactorial, with both environmental and genetic factors identified. The best established exogenous risk factors are tobacco and occupational exposure to aniline dyes and aromatic amines [8,9]. Their influence on UCCB risk is similar for tumours of varying stages and grades [10]; thus exogenous risk factors cannot currently explain disease heterogeneity. UCCB is also one of the tumours for which a role of low-penetrance genetic variants, such as GSTM1-null and NAT2-slow, has been best established [11]. Furthermore, strong evidence indicates an interaction between the NAT2-slow variant and smoking [11]. Other genetic polymorphisms have been reported to be associated with UCCB risk, but replication is necessary to establish risk conclusively [12,13]. One explanation for this lack of success is that polymorphisms may exhibit a small to moderate effect only within particular subphenotypes of the disease [6,14]. Thus polymorphisms that may be important for a particular subphenotype could be overlooked in an analysis that pools the subphenotypes due to an attenuation of effects. The aim of this work is to investigate whether UCCB subphenotypes defined according to clinicopathologic characteristics have different associations with germline polymorphisms in the Spanish Bladder Cancer (SBC)/EPICURO Study. Although this approach has been recently applied to breast [14] and prostate cancer [15], this report is the first to do so for UCCB. A total of 1526 polymorphisms in 423 cancer candidate genes were examined with respect to three distinct UCCB subphenotype groups.
2.1. Study population
The details of the study population have been described previously [8,11,16]. Briefly, the SBC/EPICURO Study is a case-control study based in 18 hospitals located in five areas in Spain. Cases were patients with a newly diagnosed, histologically confirmed UCCB from 1998 to 2001. A panel of expert pathologists reviewed slides to confirm the diagnosis and ensure uniformity of the classification criteria according to the 1998 system of the World Health Organisation and the International Society of Urological Pathology [3]. The TNM classification was applied for stage assignment [17]. Controls were patients admitted to participating hospitals for diagnoses believed to be unrelated to the exposures of interest and were individually matched to cases on age, gender, ethnic origin, and region. Information was obtained during the first hospital admission. A total of 1219 cases and 1271 controls agreed to participate in the study and were interviewed. Of them, 1188 cases (97%) and 1173 controls (92%) provided a blood or buccal cell sample. Exclusions were due to insufficient DNA and to reduce heterogeneity (nonwhite, lack of diagnostic slides, nontransitional histology, and urothelial neoplasms of low malignant potential). The final study population consisted of 1051 cases and 1149 controls. Subsequently, cases were classified in three subphenotypes according to tumour stage (T) and histologic grade (G): low-grade non–muscle invasive (Ta G1/Ta G2, n = 586), high-grade non–muscle invasive (Ta G3/T1 G2/T1 G3, n = 219), and muscle invasive (T≥2, n = 246).
2.2. Genotyping
Genotyping was performed as described using leukocyte (1011 cases, 1032 controls) or mouthwash (40 cases, 117 controls) DNA and the GoldenGate (Illumina, San Diego, CA, USA) and/or TaqMan (Applied Biosystems) assays [11,12]. The GoldenGate assay was comprised of single nucleotide polymorphisms (SNPs) from the SNP500Cancer project (http://snp500cancer.nci.nih.gov) belonging to candidate genes related to the carcinogenic process; an attempt was made to select for nonsynonymous variants or those with evidence for functional significance [12]. All polymorphisms were SNPs, with the exception of NAT1, which was categorised into NAT1*4 and NAT1*10, the “at-risk” allele [11]; NAT2, which was categorised into slow-acetylator and intermediate/rapid-acetylator phenotypes; and GSTM1 and GSTT1 genotypes, which were defined as “null” (−/−) and “present” (+/− or +/+) [11]. SNPs are reported as the gene name followed by the reference SNP (rs) number, according to the international coding nomenclature. Details regarding all 1526 polymorphisms can be found in Supplementary Table 1.
Genotype concordance in the duplicate quality-control samples was ≥99%. For the 63 SNPs genotyped by both platforms, the concordance rate was ≥98%. For those SNPs, information from the Illumina platform was first applied, followed by TaqMan data in the event of missing information. Approximately 6% (96 of 1526) of the gene variants showed significant departures from Hardy-Weinberg equilibrium (p < 0.05) in the control population, consistent with what would be expected by chance (Supplementary Table 1). Polymorphisms were excluded if they had a low minor allele frequency (<0.05) in the control population or if they were highly correlated with other variants (r2 ≥ 0.85). A total of 451 variants were excluded (Supplementary Table 1).
2.3. Statistical methods
Demographic characteristics were compared among individuals in the three UCCB subphenotypes and in the control group using the analysis of variance for continuous variables and the χ2 test of independence for categorical variables. Additionally, these comparisons were made among the three subphenotypes only.
Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated for the three UCCB subphenotypes using polytomous logistic regression models for each variant considering a codominant and additive mode of inheritance. The codominant model makes no assumptions regarding the contribution of the alleles in the three genotypes, whereas the additive model assumes a dose response or trend with every increase in allele count (0, 1, 2). The model compared each tumour group to the control group, adjusting for the following potential confounders: age at interview, gender, region and smoking status (never, occasional, former, and current) [8]. Thus an OR >1 indicates an increased risk and an OR <1 indicates a decreased risk for a particular subphenotype compared with the controls. The global association between variants and risk of any UCCB was tested using a likelihood ratio test (LRT) comparing models with and without all variant coefficients for both modes of inheritance. Similarly, the association between variants and risk of distinct subphenotypes was tested with a 2-df LRT and a per-allele trend test for each subphenotype.
Heterogeneity of genotype ORs (risk estimates) was tested using a LRT comparing a model in which the variant coefficients were constrained to be equal across the subphenotypes with an unconstrained model for both modes of inheritance. These LRTs are analogous to homogeneity of OR tests. Post hoc pairwise comparisons were performed if the heterogeneity p value was <0.01. The p values <0.01 were considered statistically significant. The robustness of the results for nonredundant and common polymorphisms was evaluated using the false discovery rate (FDR) [18]. Rather than applying an arbitrary FDR cut-off threshold, the most significant findings are presented so that the reader may assess the robustness of the results. All analyses were performed in R v.2.6.0 (http://www.r-project.org).
Patients with tumours of the three subphenotypes were similar with respect to all demographic characteristics except for age (Table 1). Cases with low-grade non–muscle-invasive tumours were slightly younger than those with high-grade non–muscle-invasive (mean age: 65.2 vs 67.3 yr) and muscle-invasive (mean age: 67.1 yr) tumours. Most patients were men (approximately 87%) and either former smokers (39.2%) or current smokers (43.0%). Patients with muscle-invasive tumours were more likely to be current smokers (47.2%) than those with low-grade (42.3%) and high-grade (40.2%) non–muscle-invasive tumours, although not significantly so (p = 0.7).
Table 1
Table 1
Characteristics of the Spanish Bladder Cancer/EPICURO study population*
Controls were similar to cases with respect to gender and region, as expected. They were slightly younger (mean age: 64.7 yr) than cases in the high-grade non–muscle-invasive and muscle-invasive groups, although, again, these differences were quite small and there was no overall age difference between pooled cases and controls. As expected, controls were less likely to be current smokers than cases.
3.1. Heterogeneity in genetic susceptibility according to tumour subphenotypes
There was evidence of heterogeneity in genetic susceptibility (p < 0.01) to UCCB according to subphenotypes in 13 SNPs belonging to 12 genes under the codominant or the additive mode of inheritance (Table 2a and 2b, respectively). Four of them (PMS2-rs6463524, DNAJC18-rs4315920, BRCA2-rs1801406, and XRCC5-rs828702) showed significant heterogeneity p values under both modes of inheritance. PMS2-rs6463524 and DNAJC18-rs4315920 were only associated with the muscle-invasive group and provided the strongest evidence of heterogeneity in the genotype ORs under the codominant (respective p values, 0.002 and 0.005) and additive (respective p values, 0.006 and 0.001) modes of inheritance. These findings were not robust to multiple testing because all FDR values for the heterogeneity p values, based on the 1075 polymorphisms examined in the analyses, were >0.75.
Table 2
Table 2
Heterogeneity in single nucleotide polymorphism (SNP)* risk estimates among bladder cancer subphenotypes in the Spanish Bladder Cancer Study (p ≤ 0.01)
Pairwise comparisons of the SNPs with significantly different risk estimates showed that the more extreme subphenotypes (low-grade non–muscle-invasive vs muscle-invasive) were significantly different (p < 0.01) under the additive model. This was not the case for the codominant model, where only three of the eight SNPs showed different risk estimates between the extreme groups. Pairwise comparisons also showed that risk estimates differed between the high-grade non–muscle-invasive and muscle-invasive groups under both modes of inheritance. In contrast, risk estimates between the low-grade and high-grade non–muscle-invasive groups were similar for all of these SNPs except for two (MBL2-rs5030737 and MMP1-rs10488) under the codominant model.
Interestingly, these 13 SNPs were all nonsignificant (p > 0.01) in a logistic regression analysis when patients in the three subphenotype groups were pooled. This is due to an attenuation of the risk estimates (Table 3), although the trend tests for the MMP1-rs10488 and PMS2-rs6463524 SNPs were marginally significant (p = 0.04).
Table 3
Table 3
Risk estimates for pooled tumour subphenotypes for those single nucleotide polymorphisms (SNPs)* that exhibited genetic heterogeneity
Figure 2 displays the risk estimates and 95% CIs for the three subphenotypes separately and jointly under the additive pattern of inheritance for the following four SNPs: CD4-rs3213427, PMS2-rs6463524, MMP1-rs10488, and ALAD-rs1139488. These SNPs conferred increased (or decreased) risks for a particular subphenotype and neutral risks for the remaining tumour groups. Estimates for the CD4-rs3213427 and PMS2-rs6463524 SNPs were only significant for the muscle-invasive group, whereas estimates for the MMP1-rs10488 and ALAD-rs1139488 SNPs were significant only for the high-grade and low-grade non–muscle-invasive groups, respectively. The figure also shows the attenuation in risk estimates for the pooled tumour groups. Supplementary Table 2 displays all genetic association results for the three subphenotypes, including the genotype frequencies for each polymorphism.
Fig. 2
Fig. 2
Per-allele risk estimates and 95% confidence intervals for bladder cancer subphenotypes and pooled cases. Per-allele risk estimates for bladder cancer subphenotypes are represented by boxes that are proportional to sample size.
3.2. Genetic variants associated with bladder cancer: GSTM1, nAT2
GSTM1-null and NAT2-slow genotypes provide the best evidence of genetic susceptibility to UCCB [11] and conferred similar risks among the three subphenotypes (Fig. 3a). The GSTM1-null genotype was associated (p < 0.001) with all three subphenotypes. The NAT2-slow genotype was associated with the non–muscle-invasive groups with p values < 0.009 and with the muscle-invasive group with p = 0.05.
Fig. 3
Fig. 3
Fig. 3
(a) Risk estimates and 95% confidence intervals (CIs) for bladder cancer subphenotypes for GSTM1-null genotype and NAT2 slow acetylation. (b) Risk estimates and 95% CIs for bladder cancer subphenotypes for the VEGF-rs25648 single nucleotide polymorphism (more ...)
3.3. Previously identified variants in the Spanish Bladder Cancer Study
Our previous reports suggested novel SNPs to be associated with UCCB risk [12]. We analysed their association with the tumour subphenotypes and focused on SNPs from the 13 top candidates reported (Table 1 of Garcia-Closas et al [12]). The most robust association with overall UCCB was for VEGF-rs25648 [12]; this variant was associated with all three subphenotypes: tumour-specific 2-df LRT p < 0.008 (Fig. 3b and Supplementary Table 3). The remaining SNPs in genes previously reported to be associated with UCCB also showed similar risks among the three subphenotypes (p > 0.01), but for the most part they were significantly associated with only one or two of the tumour groups because of a statistical power issue (Supplementary Table 3).
We investigated evidence of heterogeneity of genetic susceptibility to three well-established UCCB subphenotypes with different biologic behaviours and prognoses, leading to the identification of novel tumour subgroup-specific associations. These findings require replication in future studies. Our aim was to identify gene variants that differentially predispose bladder tumours to papillary low-grade, genomically stable (α), high-grade non–muscle-invasive (β1), or muscle-invasive (β2) subphenotypes (Fig. 1). We previously found that an inherited FGFR3 variant was differentially associated with low-grade versus high-grade non–muscle-invasive tumours [16].
Risk estimates for the two best established UCCB susceptibility genes, GSTM1 and NAT2, were similar across subphenotypes. Similarly, VEGF-rs25648 and other polymorphisms previously identified to be associated with UCCB risk in our study [12] conferred similar risks among subphenotypes.
Among the patients with non–muscle-invasive tumours, risk estimates were similar for the SNPs analysed, except for MBL2-rs5030737 and MMP1-rs10488, which conferred significantly decreased risk for heterozygotes in high-grade tumours. Mannose-binding lectin 2 (MBL2) participates in innate immunity and rs5030737 (R52C) is associated with impaired function [20]. Other MBL2 variants have been associated with risk of autoimmune and infectious diseases [20] and cancer [21,22]. Recently, Pine et al reported the prognostic value of MBL2 polymorphisms, including rs5030737, for lung cancer survival [23]. MMP1 is involved in tissue remodelling, tumour invasion, and metastasis. Although associations found only for the heterozygous genotype are often disregarded because they are believed to be biologically implausible, there is precedent that some of them are detected consistently across studies [24,25]. Also, it is likely that these SNPs follow other modes of inheritance (eg, dominant or recessive) but, due to the exploratory nature of these analyses, only the codominant and additive modes of inheritance were considered. Another caveat is that, when stratified by subphenotypes, the genotype frequencies can be considerably low and thus add noise to the OR estimates (Supplementary Table 2c).
The most notable differences in risk estimates occurred when comparing muscle-invasive and non–muscle-invasive tumours. Irrespective of the mode of inheritance, heterogeneity was observed for DNAJC18-rs4315920, coding for a heat shock protein with methyl-transferase activity and SNPs in DNA repair genes, PMS2-rs6463524, BRCA2-rs1801406, and XRCC5-rs828702. There is some evidence on the role of genetic variation in double-strand break DNA repair mechanisms in UCCB susceptibility [13].
CD4-rs3213427 and PMS2-rs6463524 showed neutral risk estimates (OR: ≈1) for the non–muscle-invasive groups and decreased risk for the muscle-invasive group (Fig. 2 and Table 2). Other SNPs conferred risk estimates among subphenotypes in opposite directions. Although we cannot discard the possibility that this may be due to random variation or chance, it is possible that variants confer contrasting risks to different subgroups due to the complex mechanisms involved in the development/progression of tumours. Because the effect of each genetic variant is probably modulated by other genetic factors, both inherited and somatic, as well as by micro- and macroenvironmental exposures, such contrasting risks remain biologically plausible. Specific somatic alterations are involved in the different subphenotypes, possibly leading to distinct molecular profiles that may explain inherited factors in opposite directions. In fact, subgroup analyses in other studies have observed such crossover interactions [26,27], with some underlying debate about their biologic plausibility [28].
An additional aspect regarding the interpretation of the results relates to the functionality of the variants differently associated with bladder cancer subphenotypes. Some of them are placed in coding regions leading to a change of amino acid and for whom an altered protein function has been proved (23). Other significant SNPs, although placed in coding regions, do not lead to an amino acid change. This fact does not rule out a functional effect because it has been reported that SNPs can be an important mechanism affecting splicing regulation (29,30). Furthermore, it is noteworthy that most of the very reproducible “hits” in genome-wide association studies identified over the past 2–3 yr indeed are placed in gene deserts or in areas of unknown function; the findings remain highly relevant (A Catalog of Published Genome-Wide Association Studies; available at: http://www.genome.gov/26525384). Among them, that reported by Kiemeney et al was highly associated with bladder cancer risk [31].
The study’s strengths include a large sample size, high participation rates, high quality of information on exposures and genotyping, and the uniform pathologic assessment of tumours. Furthermore, this study was prospectively designed to assess the role of genetic, environmental, and clinical factors in UCCB risk. Nonetheless, the results reported here should be interpreted with caution in light of several limitations. The genes/SNPs analysed were not selected to assess subphenotype heterogeneity. Some subjectivity in the pathologic evaluation of tumours cannot be ruled out, although it was reduced by ensuring a uniform tumour classification by trained pathologists. By considering molecular alterations occurring early in urothelial carcinogenesis, a more accurate subphenotyping classification may be applied in future studies. Furthermore, the study was based on a tumour development/progression model that is likely to be an oversimplification. Although rare, a few patients with low-grade tumours eventually invade the muscle; furthermore, confluence of tumours can also occur in the context of a multicentric neoplasm such as UCCB. Although we observed some evidence of heterogeneity in genetic susceptibility, the results were not robust to multiple testing due to limited statistical power. Although this is one of the largest UCCB studies carried out so far, it was not designed to have sufficient power to investigate heterogeneity by subphenotypes. However, the study provides valuable novel information on the heterogeneity in genetic susceptibility of UCCB that, jointly with the molecular and pathologic heterogeneity, will help to disentangle the complexity of this disease. This work should stimulate replication of our findings in independent UCCB case series.
To our knowledge, this is the first report to investigate heterogeneity of genetic susceptibility to subphenotypes defined by clinicopathologic characteristics in a large epidemiologic study. Importantly, the well-established variants associated with UCCB displayed similar risks for all subphenotypes. Furthermore, we identified new polymorphisms potentially associated with particular subphenotypes that did not show overall associations with UCCB risk. These exploratory analyses suggest that inherited factors may also contribute to the molecular/pathologic diversity of UCCB. However, validation through large-scale replication studies is required. Ongoing genome-wide association studies should contribute to these aims [31].
Although this research may not have immediate consequences at the clinical level, these results help to disentangle the molecular and pathologic diversity of UCCB. Furthermore, novel genetic associations with particular subphenotypes could lead to the discovery of novel genes involved in disease progression and to novel therapeutic targets. The relevance of tumour subphenotypes in cancer treatment is becoming obvious in areas where there is more knowledge, such as is the case for breast cancer (ie, luminal vs basal) or colorectal cancer (ie, mismatch repair vs nonmismatch repair). UCCB offers an opportunity in this sense.
Take-home message
This report is the first to examine genetic susceptibility to distinct bladder cancer subphenotypes. We provide evidence of heterogeneity in variant risk estimates among bladder subphenotypes. Replication will be required for conclusive evidence.
01: Supplementary Table 1
Detailed information of polymorphisms examined in the case-case-case-control analysis within the Spanish Bladder Cancer/EPICURO Study.
02: Supplementary Table 2
(a) Associations between bladder cancer subphenotypes and all 1075 polymorphisms examined in this report under the codominant model; (b) associations between bladder cancer subphenotypes and all 1075 polymorphisms examined in this report under the additive model; (c) sample size and genotype distribution among bladder cancer subphenotypes and controls for all 1075 polymorphisms examined in this report.
03: Supplementary Table 3
(a) Associations between bladder cancer subphenotypes and genes previously identified in the Spanish Bladder Cancer Study under the codominant model; (b) associations between bladder cancer subphenotypes and genes previously identified in the Spanish Bladder Cancer Study under the additive model.
Acknowledgments
Acknowledgement statement: The authors acknowledge Drs. C. Samanic, M. Sala, G. Castaño-Vinyals, and M. Torà for their work in data and sample collection; A. Amorós and F. Fernández for their support in data management; I. Molina, I. Roquer, and G. J. Andres for their secretarial help; and physicians, field workers, and study participants for their continuous support during conduct of the study.
Funding/Support and role of the sponsor: This work was partially supported by the Fondo de Investigación Sanitaria, Spain (G03/174, PI051436, PI061614); Fundació la Marató de TV3; Red Temática de Investigación Cooperativa en Cáncer (RTICC), Instituto de Salud Carlos III, Ministry of Health, Spain; the Intramural Research Program of the Division of Cancer Epidemiology and Genetics, National Cancer Institute, USA; and by the EU (HEALTH-STREP - 2006-037739 and HEALTH-F2-2008-201663). The sponsors were involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation of the manuscript.
Footnotes
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Author contributions: Núria Malats had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Malats, Real, Kogevinas, Rothman, Silverman, Dosemeci, García-Closas, Murta-Nascimento, Chanock.
Acquisition of data: Malats, Real, Kogevinas, Rothman, Silverman, Dosemeci, García-Closas, Tardón, Serra, Carrato, García-Closas, Murta-Nascimento, Chanock, Lloreta.
Analysis and interpretation of data: Guey, Malats, Real, Kogevinas, Rothman, Silverman, Dosemeci, García-Closas.
Drafting of the manuscript: Guey, Malats, Real.
Critical revision of the manuscript for important intellectual content: Guey, Malats, Real, Kogevinas, Rothman, Silverman, Dosemeci, García-Closas, Tardón, Serra, Carrato, García-Closas, Murta-Nascimento, Chanock, Calle.
Statistical analysis: Guey, Malats, Palencia, Vellalta, Marenne, Calle.
Obtaining funding: Malats, Real, Kogevinas, Rothman, Silverman, Dosemeci.
Administrative, technical, or material support: None.
Supervision: Malats.
Other (specify): None.
Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/ affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None.
1. Ferlay JBF, Pisani P, Parkin DM. GLOBOCAN 2002. IARC Cancer Base No. 5, Version 2.0. Lyon, France: IARC Press; 2004. Cancer incidence, mortality and prevalence worldwide.
2. Lotan Y, Svatek RS, Malats N. Screening for bladder cancer: a perspective. World J Urol. 2008;26:13–18. [PubMed]
3. Mostofi FK, Davis CJ, Sesterhenn I. Histological typing of urinary bladder tumours. 2nd ed. Berlin, Germany: Springer; 1999.
4. Babjuk M, Oosterlinck W, Sylvester R, Kaasinen E, Bohle A, Palou-Redorta J. EAU guidelines on non-muscle-invasive urothelial carcinoma of the bladder. Eur Urol. 2008;54:303–314. [PubMed]
5. Pasin E, Josephson DY, Mitra AP, Cote RJ, Stein JP. Superficial bladder cancer: an update on etiology, molecular development, classification, and natural history. Rev Urol. 2008;10:31–43. [PubMed]
6. Wu XR. Urothelial tumourigenesis: a tale of divergent pathways. Nat Rev Cancer. 2005;5:713–725. [PubMed]
7. Lopez-Knowles E, Hernandez S, Malats N, et al. PIK3CA mutations are an early genetic alteration associated with FGFR3 mutations in superficial papillary bladder tumours. Cancer Res. 2006;66:7401–7404. [PubMed]
8. Samanic C, Kogevinas M, Dosemeci M, et al. Smoking and bladder cancer in Spain: effects of tobacco type, timing, environmental tobacco smoke, and gender. Cancer Epidemiol Biomarkers Prev. 2006;15:1348–1354. [PubMed]
9. Silverman DT, Devesa SS, Moore LE, Rothman N. Bladder cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. New York, NY: Oxford University Press; 2006. pp. 1101–1127.
10. Sturgeon SR, Hartge P, Silverman DT, et al. Associations between bladder cancer risk factors and tumour stage and grade at diagnosis. Epidemiology. 1994;5:218–225. [PubMed]
11. Garcia-Closas M, Malats N, Silverman D, et al. NAT2 slow acetylation, GSTM1 null genotype, and risk of bladder cancer: results from the Spanish Bladder Cancer Study and meta-analyses. Lancet. 2005;366:649–659. [PMC free article] [PubMed]
12. Garcia-Closas M, Malats N, Real FX, et al. Large-scale evaluation of candidate genes identifies associations between VEGF polymorphisms and bladder cancer risk. PLoS Genet. 2007;3:e29. [PMC free article] [PubMed]
13. Malats N. Genetic epidemiology of bladder cancer: scaling up in the identification of low penetrance genetic markers of bladder cancer risk and progression. WHO Bladder Cancer report. Scand J Urol Nephrol Suppl. 2008;218:131–140. [PubMed]
14. Garcia-Closas M, Hall P, Nevanlinna H, et al. Heterogeneity of breast cancer associations with five susceptibility loci by clinical and pathological characteristics. PLoS Genet. 2008;4 e1000054. [PMC free article] [PubMed]
15. Thomas G, Jacobs KB, Yeager M, et al. Multiple loci identified in a genome-wide association study of prostate cancer. Nat Genet. 2008;40:310–315. [PubMed]
16. Hernandez S, Lopez-Knowles E, Lloreta J, et al. Prospective study of FGFR3 mutations as a prognostic factor in nonmuscle invasive urothelial bladder carcinomas. J Clin Oncol. 2006;24:3664–3671. [PubMed]
17. Greene FL. American Joint Committee on Cancer; American Cancer Society. AJCC cancer staging manual. 6th ed. New York, NY: Springer; 2002.
18. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. Controlling the false discovery rate in behavior genetics research. Behav Brain Res. 2001;125:279–284. [PubMed]
19. Knowles MA. Molecular subtypes of bladder cancer: Jekyll and Hyde or chalk and cheese? Carcinogenesis. 2006;27:361–373. [PubMed]
20. Garred P, Larsen F, Seyfarth J, Fujita R, Madsen HO. Mannose-binding lectin and its genetic variants. Genes Immun. 2006;7:85–94. [PubMed]
21. Baccarelli A, Hou L, Chen J, et al. Mannose-binding lectin-2 genetic variation and stomach cancer risk. Int J Cancer. 2006;119:1970–1975. [PubMed]
22. Bernig T, Boersma BJ, Howe TM, et al. The mannose-binding lectin (MBL2) haplotype and breast cancer: an association study in African-American and Caucasian women. Carcinogenesis. 2007;28:828–836. [PubMed]
23. Pine SR, Mechanic LE, Ambs S, et al. Lung cancer survival and functional polymorphisms in MBL2, an innate-immunity gene. J Natl Cancer Inst. 2007;99:1401–1409. [PubMed]
24. Dossou-Yovo OP, Lapoumeroulie C, Hauchecorne M, et al. Variants of the mannose-binding lectin gene in the Benin population: heterozygosity for the p.G57E allele may confer a selective advantage. Hum Biol. 2007;79:687–697. [PubMed]
25. Hraber P, Kuiken C, Yusim K. Evidence for human leukocyte antigen heterozygote advantage against hepatitis C virus infection. Hepatology. 2007;46:1713–1721. [PubMed]
26. Landi MT, Baccarelli A, Tarone RE, et al. DNA repair, dysplastic nevi, and sunlight sensitivity in the development of cutaneous malignant melanoma. J Natl Cancer Inst. 2002;94:94–101. [PubMed]
27. Schneider HJ, Saller B, Klotsche J, et al. Opposite associations of age-dependent insulin-like growth factor-I standard deviation scores with nutritional state in normal weight and obese subjects. Eur J Endocrinol. 2006;154:699–706. [PubMed]
28. Weiss NS. Subgroup-specific associations in the face of overall null results: should we rush in or fear to tread? Cancer Epidemiol Biomarkers Prev. 2008;17:1297–1299. [PubMed]
29. Kwan T, Benovoy D, Dias C, et al. Heritability of alternative splicing in the human genome. Genome Res. 2007;17:1210–1218. [PubMed]
30. Hull J, Campino S, Rowlands K, et al. Identification of common genetic variation that modulates alternative splicing. PLoS Genet. 2007;3:e99. [PubMed]
31. Kiemeney LA, Thorlacius S, Sulem P, et al. Sequence variant on 8q24 confers susceptibility to urinary bladder cancer. Nat Genet. 2008;40:1307–1312. [PubMed]