Identification of distinct basal and luminal subtypes of muscle-invasive bladder cancer with different sensitivities to frontline chemotherapy

Abstract

Muscle-invasive bladder cancers (MIBCs) are biologically heterogeneous and have widely variable clinical outcomes and responses to conventional chemotherapy. We discovered three molecular subtypes of MIBC that resembled established molecular subtypes of breast cancer. Basal MIBCs shared biomarkers with basal breast cancers and were characterized by p63 activation, squamous differentiation, and more aggressive disease at presentation. Luminal MIBCs contained features of active PPARγ and estrogen receptor transcription and were enriched with activating FGFR3 mutations and potential FGFR inhibitor sensitivity. p53-like MIBCs were consistently resistant to neoadjuvant methotrexate, vinblastine, doxorubicin and cisplatin chemotherapy, and all chemoresistant tumors adopted a p53-like phenotype after therapy. Our observations have important implications for prognostication, the future clinical development of targeted agents, and disease management with conventional chemotherapy.

Copyright © 2014 Elsevier Inc. All rights reserved.

Figures

Figure 1. Basal and luminal subtypes of bladder cancer

A. (Left) Whole genome mRNA expression profiling and hierarchical cluster analysis of a cohort of 73 MIBCs. RNA from fresh frozen tumors was analyzed using Illumina arrays. RAS, TP53, RB1, and FGFR3 mutations were detected by sequencing and are indicated in the color bars below the dendogram. Black, mutation; white, wild type; grey, mutation data were unavailable. (Right) Kaplan-Meier plots of overall survival (p = 0.098) and disease-specific survival (p = 0.028) in the 3 tumor subtypes. B. Expression of basal and luminal markers in the 3 subtypes. The heat maps depict relative expression of basal (left) and luminal (right) biomarkers. GSEA analyses (below, left) were used to determine whether basal and luminal markers were enriched in the subtypes. C. Quantitative RT-PCR was used to evaluate the accuracy of the gene expression profiling results. Relative levels of the indicated basal (red shades) and luminal (blue shades) biomarkers measured by RT-PCR were compared to the levels of the same markers measured by gene expression profiling on RNA isolated from macrodissected FFPE sections of the same tumors. Results are presented as relative quantitation (RQ) and the error bars indicate the range of RQ values as defined by 95% confidence level. RT-PCR results are shown on top, DASL gene expression profiling results are shown below. D. Analysis of basal and luminal marker expression by immunohistochemistry. Results from two representative basal (left) and luminal (right) tumors as defined by gene expression profiling are displayed. The scale bars correspond to 100 microns. See also Figure S1.

Figure 2. Characterization of basal and luminal subtypes in other MIBC cohorts

A. Subtype classification of the MD Anderson validation cohort (n = 57). RNA was isolated from macrodissected FFPE tumor sections and whole genome mRNA expression was measured using Illumina’s DASL platform. Kaplan-Meier plots of overall survival (p = 0.011) and disease-specific survival (p=0.004) associated with the 3 subtypes are presented on the right. B. Expression of basal and luminal markers in the molecular subtypes in the MD Anderson validation cohort. The results of GSEA analyses of basal and luminal marker expression in the subtypes are displayed on the left, and heat maps depicting relative basal and luminal marker levels in the subtypes are displayed on the right. C. Subtype classification of the Chungbuk cohort (n = 55). Whole genome mRNA expression profiling (Illumina platform) and clinical data were downloaded from GEO (GSE13507), and the oneNN classifier was used to assign tumors to subtypes. Tumors were assigned to subtypes using the oneNN prediction model (left). Kaplan-Meier plots of overall survival (p = 0.102) and disease-specific survival (p = 0.058) as a function of tumor subtype (right). D. Expression of basal and luminal markers in the molecular subtypes in the Chungbuk cohort. The results of GSEA analyses of basal and luminal marker expression in the subtypes are displayed on the left, and heat maps depicting basal and luminal marker expression are displayed on the right. See also Tables S1 and S2.

Figure 3. Presence of squamous features in the subtypes

A. Tumor squamous feature content in the MD Anderson discovery and validation cohorts. Subtype designations are indicated by the top color bars, and the presence of squamous features (in black) is indicated in the color bars below. B. Relationship between the MD Anderson subtypes and the molecular taxonomy developed by Sjödahl et al (Sjodahl et al., 2012). Whole genome mRNA expression (Illumina platform) and clinical data were downloaded from GEO (GSE32894), and the oneNN classifier was used to assign the Lund tumors to subtypes. Subtype membership is indicated by the top color bars, and FGFR3 and TP53 mutations in the Lund tumors are indicated in color bars below. Black: mutant; white: wild-type; grey (N/A): mutation data were not available. C. Presence of squamous features in the UCSF dataset. Whole genome mRNA expression profiling (in-house platform) and clinical data were downloaded from GEO (GSE1827), and the oneNN classifier was used to assign the UCSF tumors to the subtypes. Subtype memberships for each tumor are indicated in the top color bars, and the presence of squamous features (in black) is indicated in the color bar below. D. Tissue microarray analysis of CK5/6 (basal) and CK20 (luminal) cytokeratin expression. Cytokeratin protein expression was measured by immunohistochemistry and optical image analysis in the MD Anderson Pathology Core on a tissue microarray containing 332 high-grade pT3 tumors. The percentages of positive tumor cells as determined by image analysis are shown. Left panels: mean levels of CK5/6 (top) and CK20 (bottom) in tumors without (TCC) or with (TCC with SD) squamous features. Bars indicate mean values with 95% confidence intervals. Middle panels: representative images of stained cores from tumors that expressed high or low levels of CK5/6 or CK20. The scale bars correspond to 100 microns. Right panel: relationship between CK5/6 and CK20 expression across the cohort. See also Tables S3–5 and Fig. S2.

Figure 4. Subtype-associated gene expression signatures

Signatures were identified using the whole genome mRNA expression profiling data from the MD Anderson discovery cohort and the upstream regulators tool in Ingenuity Pathway Analysis (IPA, Ingenuity® Systems (www.ingenuity.com)). Each heat map displays the expression of the corresponding IPA gene signature as a function of tumor subtype membership; note that genes can be either up- or down-regulated by an active transcription factor. Top left: p63-associated gene expression. Below left: PPARγ-associated gene expression. Right: p53-associated gene expression. See also Tables S6 and S7 and Fig. S3.

Figure 5. Transcriptional control of the basal and luminal gene expression

Whole genome mRNA expression profiling was used to analyze the effects of stable p63 knockdown or rosiglitazone-induced PPARγ activation in human bladder cancer cell lines, and the data were used to generate gene expression signatures characteristic of p63 and PPARγ activation. GSEA was then used to determine whether these signatures were present in the MD Anderson discovery cohort tumor subtypes. A. Effects of p63 or PPARγ modulation on basal and luminal transcriptional signatures. Top panels: significantly activated/inhibited transcriptional pathways after p63 knockdown in UM-UC14 cells (top left), PPARγ activation in UM-UC7 (top middle) or PPARγ activation in UM-UC9 (top right) based on IPA analyses. The heat maps below each graph indicate significant changes in basal and luminal marker expression. B. p63 and PPARγ gene expression signatures in the subtypes of primary MIBCs. Separate results and p values are shown for the signatures derived from the up- and down-regulated genes in each condition. ROSI: rosiglitazone. See also Fig. S4.

Figure 6. Relationship between subtype membership and chemotherapy sensitivity

A. Responses to neoadjuvant chemotherapy in the MD Anderson NAC (n = 34) and MVAC (n = 23) cohorts. Subtype membership was determined using whole genome mRNA expression profiling data obtained from untreated (TURBT) tumors and the oneNN classifier. Pathological response was defined as downstaging to ≤ pT1. B. The IPA-defined p53 gene expression signature from the p53-like primary MIBCs was used to perform unsupervised hierarchical cluster analysis on whole genome expression data from a panel of human bladder cancer cell lines (n = 28). The green boxes on the heat maps indicate expression of the signature in the MD Anderson discovery cohort (left) or the cell lines (right). TP53 mutational status was determined by sequencing and is indicated by the color bar below the heat map (black = mutant, white = wild-type, grey = data were not available). C. Cells were incubated with or without 10 μM cisplatin for 48 h and apoptosis-associated DNA fragmentation was quantified by propidium iodide staining and FACS analysis in 3 independent experiments. The left panel displays a scatter gram comparing the levels within the two subsets of cell lines (mean ± SEM). The right panel displays the mean value of induced apoptosis in each cell line across the entire cohort. See also Tables S8 and S9.

Figure 7. Wild-type p53 gene signatures in tumors before and after treatment with NAC

A. Relationship between subtype membership and response to NAC in the Philadelphia DDMVAC cohort. Subtype membership was determined using pretreatment (TURBT) specimens. Pathological response was defined as downstaging to ≤ pT1. B. Comparison of subtype membership in the chemoresistant Philadelphia tumors before and after NAC. Whole genome mRNA expression profiling was performed on matched tumors before and after NAC, and the oneNN classifier was used to assign tumors to subtypes. “TURBT” refers to the pretreatment tumors and “cystectomy” to the post-treatment tumors. C. Expression of a wild-type p53 gene signature in matched pre- and post-treatment tumors. Left: heat map displaying expression of an active p53 gene signature after NAC (log ratio cystectomy/TURBT of matched tumors). Right: relative expression of the p53 signature in matched pre- and post-treatment tumors arranged according to subtype membership. D. Analysis of an immune infiltration signature in basal tumors. A supervised analysis was performed to compare the differences in gene expression between basal tumors that were either sensitive or resistant to neoadjuvant DDMVAC in the Philadelphia cohort. Left: heat map depicting the relative expression of immune signature genes in basal responders and non-responders. Right: GSEA analyses of immune biomarkers in the basal tumors. See also Tables S10 and S11 and Fig. S5.