Molecular biology of bladder cancer: new insights into pathogenesis and clinical diversity

Key Points

  • Bladder cancer is the fifth most common cancer in men in Western countries (male:female ratio is 3:1), and tobacco smoking is a major risk factor.

  • There are two major groups of patients with distinct prognosis and molecular features. Although local disease recurrence is a major problem for those with low-grade non-muscle-invasive tumours, life expectancy is long and development of invasive disease is infrequent. For those who present with muscle-invasive disease, development of metastatic disease is common, prognosis is dismal and no advances in therapy have been made for decades.

  • Major unmet clinical needs include non-invasive methods for disease surveillance and novel approaches to eliminate both tumour and widespread intraepithelial preneoplasia in patients with non-muscle-invasive disease. New systemic therapeutic approaches are urgently needed for those with muscle-invasive disease.

  • Recent studies reveal important biological features of urothelial metastasis and the epithelial–mesenchymal transition that may contribute to metastatic initiation. A key role for inflammatory processes is evident in the development of metastasis.

  • Heterogeneity in outcome within the two major groups indicates a need for subclassification for more accurate prognostication, prediction of response to current therapies and development of novel therapies.

  • Recent molecular analyses now provide such subclassification with definition of multiple subgroups that are independent of conventional histopathological definitions. This presents major opportunities for personalized patient care.


Urothelial carcinoma of the bladder comprises two long-recognized disease entities with distinct molecular features and clinical outcome. Low-grade non-muscle-invasive tumours recur frequently but rarely progress to muscle invasion, whereas muscle-invasive tumours are usually diagnosed de novo and frequently metastasize. Recent genome-wide expression and sequencing studies identify genes and pathways that are key drivers of urothelial cancer and reveal a more complex picture with multiple molecular subclasses that traverse conventional grade and stage groupings. This improved understanding of molecular features, disease pathogenesis and heterogeneity provides new opportunities for prognostic application, disease monitoring and personalized therapy.


Bladder cancer is the most common cancer of the urinary tract with 380,000 new cases and 150,000 deaths per year worldwide1. It ranks fifth among cancers in men in Western countries. Epidemiological studies identify a range of environmental risk factors, many of which reflect exposure to excreted carcinogenic molecules (Box 1). Recent genome-wide association studies have also identified germline variants that contribute to risk2.

In Europe and North America, >90% of bladder cancers are urothelial carcinoma. These tumours are staged using the Tumour–Node–Metastasis system (TNM system)3, which describes the extent of invasion (Tis–T4), and they are graded according to their cellular characteristics. Two classification systems are in current use4,5. At diagnosis the majority of bladder cancers (60%) are non-muscle-invasive (stage Ta) papillary tumours of low grade (Fig. 1). Stage T1 tumours, which have penetrated the epithelial basement membrane but have not invaded the muscle, are mostly of high grade, as are muscle-invasive bladder cancers (MIBCs; 20% at diagnosis).

Figure 1: Bladder cancer grading and staging.

a | Staging of bladder cancer according to the Tumour–Node–Metastasis (TNM) system is shown3. b | Grading according to the 1973 World Health Organization (WHO) and 2004 WHO/International Society of Urological Pathology (ISUP) criteria is shown4,5. The major difference is in the classification of papillary tumours, which are classified as grades 1, 2 and 3 in the older system and as papillary urothelial malignancy of low malignant potential (PUNLMP; equivalent to grade 1), low-grade papillary urothelial carcinoma or high-grade papillary urothelial carcinoma in the WHO/ISUP 2004 classification.

PowerPoint slide

Non-muscle-invasive bladder cancers (NMIBCs) frequently recur (50–70%) but infrequently progress to invasion (10–15%)6, and five-year survival is 90%. These patients are monitored by cystoscopy and may have multiple resections over many years. Improved monitoring is needed, ideally via urine analysis, which could reduce the morbidity and costs associated with cystoscopy. Although risk tables provide a prognostic tool7, no molecular biomarkers accurately predict disease progression. For these patients, localized therapies to remove residual neoplastic and pre-neoplastic cells post-resection may have major impacts both on quality of life and in health economic terms. MIBCs (of stage T2 and above) have less favourable prognosis with five-year survival <50% and common progression to metastasis (Box 1). Treatment has not advanced for several decades, and new approaches to systemic therapy are needed8.

Improved treatment requires detailed understanding of urothelial carcinoma pathogenesis and molecular biology. A model has evolved, taking into account both histopathological and molecular features. This 'two-pathway' model proposes that papillary NMIBC develops via epithelial hyperplasia and recruitment of a branching vasculature. MIBC is proposed to develop via flat dysplasia and carcinoma in situ (CIS). The molecular characteristics of MIBC and NMIBC are highly distinct (Tables 1,2). While many features of bladder cancer fit well within this model, there is considerable heterogeneity in clinical behaviour. This Review discusses the molecular features of bladder tumours and recent findings that begin to unravel this heterogeneity and pave the way for a step change in personalized patient care.

Table 1 Oncogenes altered in bladder cancer
Table 2 Putative tumour suppressor genes with loss of function in bladder cancer

The molecular landscape

Genomic instability, chromosomal alterations and allelic loss. NMIBCs commonly have near-diploid karyotype and few genomic rearrangements. By contrast, MIBCs are commonly aneuploid with many alterations, including chromothripsis9. Non-homologous end-joining is implicated as a mechanism for error-prone double-strand break repair in MIBC9,10. The identification of mutations in minichromosome maintenance complex component 4 (MCM4), a component of the replication-licensing complex, in some MIBCs has led to the suggestion that failure to rescue stalled replication forks may underlie complex translocation events involving multiple chromosomes9. Inactivating mutations are reported in DNA repair and DNA damage checkpoint genes, including excision repair cross-complementation group 2 (ERCC2), ataxia-telangiectasia mutated (ATM) and Fanconi anaemia complementation group A (FANCA) in MIBC. Frequent mutation of stromal antigen 2 (STAG2), which encodes a component of the cohesin complex, has been identified, and this mutation occurs with higher frequency in NMIBC than in MIBC11,12,13. Cohesin plays a part in chromatid segregation and, in other tumour types, STAG2 inactivation is associated with aneuploidy14. In bladder cancer there is no clear relationship to aneuploidy11,13. As cohesin also plays a part in genomic organization via interaction with CCCTC-binding factor (CTCF)15 and in DNA double-strand break repair by homologous recombination16, these functions may be more important than effects on ploidy. Several other genes involved in sister chromatid cohesion and segregation — STAG1, nipped-B homologue (NIPBL), structural maintenance of chromosomes 1A (SMC1A), SMC1B, SMC3 and extra spindle pole bodies homologue 1 (ESPL1) — show mutation in MIBC11,17.

Such defects lead to a variety of chromosomal alterations. Chromosome 9 deletion is common in both NMIBC and MIBC (>50%). Candidate tumour suppressor genes affected by chromosome 9 deletion are cyclin-dependent kinase inhibitor 2A (CDKN2A; which encodes p16 and p14ARF) and CDKN2B (which encodes p15) at 9p21 (Refs 18,19); patched 1 (PTCH1) at 9q22 (Refs 20,21); deleted in bladder cancer 1 (DBC1; also known as BRINP1) at 9q32–33 (Refs 22,23); and tuberous sclerosis 1 (TSC1) at 9q34 (Refs 24,25). Loss of heterozygosity (LOH) of 9p, homozygous deletion of CDKN2A and loss of expression of p16 in NMIBC are predictors of reduced recurrence-free interval26,27,28. As mouse knockouts and in vitro experiments29,30 suggest that p16 and/or p14ARF are haploinsufficient tumour suppressors, it is plausible that the loss of one allele in 45% of bladder cancers has functional consequences. Importantly, in the small group of MIBC with fibroblast growth factor receptor 3 (FGFR3) mutation (discussed below), a high frequency of CDKN2A homozygous deletion has been reported31. This may identify a progression pathway for FGFR3-mutant NMIBC to muscle invasion. Loss of p16 expression is inversely correlated with RB1 expression32 and, conversely, high-level expression results from negative feedback in tumours with RB1 loss33. Both changes are adverse prognostic biomarkers and are found in >50% of MIBC34.

The best-validated tumour suppressor gene on 9q is TSC1. The TSC1–TSC2 complex negatively regulates the mTOR branch of the PI3K pathway (Fig. 2). A recent study has also identified mutations in NOTCH1 in 18% of tumours35. However, no genes on 9q show biallelic mutational inactivation at frequencies that are compatible with the high frequency of LOH. Indeed, no chromosome 9 genes show significant mutation frequency in exome sequencing studies11,17,36,37. This may indicate haploinsufficiency for one or more genes or implicate epigenetic rather than genetic mechanisms of inactivation. After more than 2 decades, the quest to identify drivers of 9q loss in bladder cancer continues.

Figure 2: The PI3K and MAPK pathways in bladder cancer.

Growth factor-mediated signalling or mutational activation of both PI3K and MAPK pathways is common in bladder cancer. Receptor tyrosine kinases (RTKs), epidermal growth factor receptor (EGFR), ERBB2, ERBB3, fibroblast growth factor receptor 1 (FGFR1) and FGFR3 may be activated by ligand, overexpression and/or mutation in bladder cancer. Through adaptor proteins, these RTKs activate RAS. Signalling via the RAS–RAF–MEK–ERK cascade leads to phosphorylation of many substrates that can have multiple cellular effects depending on the intensity and duration of signalling. In many situations proliferation is induced. Activated RTKs bind to p85 (the regulatory subunit of PI3K) and recruit the enzyme to the membrane, where it phosphorylates phosphatidyinositol-4,5-bisphosphate (PIP2) to generate PIP3. Activated RAS can also directly activate PI3K. PIP3 recruits 3-phosphoinositide-dependent protein kinase 1 (PDK1; also known as PDPK1) and AKT, resulting in activation of AKT by phosphorylation, which leads to both positive and negative regulation of a wide range of target proteins (not all shown). Cyclin D1 (CCND1) and MDM2 are upregulated directly or indirectly, resulting in a positive stimulus via the RB or p53 pathways, respectively. AKT also phosphorylates and inactivates tuberous sclerosis 2 (TSC2), leading to activation of mTOR complex 1 (mTORC1), which controls protein synthesis. TSC1 forms an active complex with TSC2, and loss of function of either protein leads to dysregulated mTOR signalling. AKT phosphorylates and inactivates glycogen synthase kinase 3β (GSK3β), relieving its suppression of β-catenin, which is freed to enter the nucleus and activate gene expression. MYC expression is induced as a consequence of both ERK and AKT signalling. Key genes that are activated in bladder cancer are shown in dark red and those that are inactivated in green. BAD, BCL-2-associated agonist of cell death; FOXO, forkhead box O; RHEB, Ras homologue enriched in brain.

PowerPoint slide

Other copy number changes and allelic loss in bladder cancer have been identified by comparative genomic hybridization (CGH) and LOH analyses, although many target genes remain unknown38,39,40. Regions of deletion associated with aggressive disease are 8p, 2q and 5q41,42,43. In addition to amplicons containing known oncogenes (Table 1), amplicons on 1q21–q24, 3p25, 6p22, 8p12–p11, 11p15, 11q14, 12q24 and 20q12–q13 have been identified by CGH43,44,45 and, apart from homozygous deletion of CDKN2A, homozygous deletion of regions on 9p21.3, 2q36, 11p11, 18p11 and 19q12 have also been reported. Copy number data from low-pass whole-genome sequence and/or single-nucleotide polymorphism (SNP) array analysis of MIBC have also identified amplicons on 19q, 1q22–q23, 8p11 and 20q11, as well as deletions on 2q21, 2q34, 4q22, 5q12, 6p25 and 16p13 (Refs 17,37). Although MIBCs show many chromosomal rearrangements, the only recurrent gene–gene fusion identified is FGFR3TACC3 (transforming, acidic coiled-coil-containing protein 3)37,46.

Mutation frequency and signature. Exome sequence from 294 bladder tumours, mostly MIBC or stage T1 tumours, has been reported11,17,36,37,47. Somatic mutation frequency in MIBC is reported as 300 exonic mutations per sample, with mean and median rates of 7.7 and 5.5 per megabase, respectively37; such a frequency is exceeded only by lung cancer and melanoma, and is dominated by C:G→T:A transitions. A signature of tobacco smoke (polycyclic hydrocarbon) exposure (C:G→A:T transversion) is not apparent. Interestingly, many C:G→T:A mutations are found in the context TpC48 — a pattern that is characteristic of mutations caused by the APOBEC (apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like) family of cytidine deaminases, which normally restrict the propagation of retroviruses and retrotransposons, and which deaminate cytosine to leave this 'tell-tale' signature49. Indeed, APOBEC3B expression is significantly upregulated in human bladder cancer37,50.

FGFR alterations. Up to 80% of stage Ta tumours have activating point mutations in FGFR3 (Fig. 3a), which is associated with favourable outcome51,52,53,54,55,56,57. In stage T1 tumours and MIBC, FGFR3 mutation is less common (10–20% in tumours of stage T2 or above)58,59,60. In cultured normal human urothelial cells (NHUCs), mutant FGFR3 activates the RAS–MAPK pathway and phospholipase Cγ (PLCγ), leading to increased survival and proliferation to high cell density61. This in vitro phenotype suggests that FGFR3 mutation could contribute to early clonal expansion within the urothelium in vivo.

Figure 3: FGFR activation in bladder cancer.

a | A schematic of fibroblast growth factor receptor 3 (FGFR3) protein and corresponding exon positions is shown. Codons showing activating point mutation and relative frequencies as percentage of mutations reported in the literature are indicated. FGFR3 contains the following domains: signal peptide (SP); acid box (AB); immunoglobulin-like domains IgI, IgII and IgIII; transmembrane (TM) region and tyrosine kinase (TK). b | Mechanisms of activation of FGFR1 and FGFR3 in bladder cancer are shown. The receptor-based mechanisms depicted have all been reported in urothelial cancer. There is less information about FGF secretion by urothelial tumour stroma or cancer cells, but FGF-like activity is increased in the urine of patients with bladder tumour.

PowerPoint slide

FGFR3 is also implicated in contributing to the risk of bladder cancer development. A SNP in an intron of TACC3, which is 70 kb from FGFR3, is associated with bladder cancer risk62 and with higher risk of recurrence in stage Ta disease, particularly for FGFR3-mutant NMIBC. The mechanism (or mechanisms) underlying these links is not clear, but one possibility is that altered chromatin structure associated with increased expression of FGFR3 could increase the probability of mutation and/or increase the expression and impact of mutated proteins.

In normal urothelium, FGFR3 is expressed as the IIIb isoform, which mainly binds to FGF1. A splice variant (Δ8–10) encodes a secreted form that lacks the transmembrane domain, which may act as a negative regulator by sequestering FGFs or by binding to full-length receptors63. In bladder cancer cell lines, reduced expression of full-length FGFR3-IIIb and FGFR3-Δ8–10 isoforms, and a switch to the FGFR3-IIIc isoform (which binds to many FGF ligands64) may facilitate autocrine or paracrine signalling63. Furthermore, many bladder cancers, including those without FGFR3 point mutations, show increased expression of FGFR3 (Ref. 60). The microRNAs (miRNAs) miR-99a and miR-100, which negatively regulate the expression of FGFR3, are also commonly downregulated in bladder cancer, particularly in NMIBC65.

Approximately 5% of cell lines and tumours contain chromosomal translocations that generate FGFR3 fusion proteins. These fusion proteins comprise amino acids 1–760 of FGFR3 (IIIb isoform, including the kinase domain) fused in-frame to TACC3 (Refs 17,37,46) or BAI1-associated protein 2-like 1 (BAIAP2L1)46. These are highly activated and transforming oncoproteins. An additional mechanism for the activation of FGFR3 in bladder cancer could include upregulated expression of FGF ligands by tumour cells or the stroma, although this has not been adequately assessed (Fig. 3b).

FGFR1 expression is upregulated in both NMIBC and MIBC66, although no mutations have been reported. An increased ratio of the FGFR1β:FGFR1α splice variants is found in tumours of higher grade and stage66. The β-isoform, which lacks the first extracellular immunoglobulin-like domain, shows increased sensitivity to FGF1 (Ref. 67) (Fig. 3b). FGF2 stimulation of FGFR1β in NHUCs activates the MAPK pathway and PLCγ, leading to increased proliferation and reduced apoptosis66. Similar stimulation in bladder cancer-derived cell lines induced an epithelial–mesenchymal transition (EMT), a major feature of which is PLCγ-mediated upregulation of the cyclooxygenase COX2 (also known as PTGS2)68. Consistent with this, bladder cancer cell lines with the highest FGFR1 expression show a mesenchymal phenotype (indicated by low E-cadherin expression) and upregulated FGF2 expression, and those with epithelial phenotype show higher FGFR3 and E-cadherin expression69.

PI3K pathway alterations. The PI3K pathway is activated by several mechanisms (Fig. 2). Some events are not mutually exclusive, implying non-redundant or non-canonical functions in bladder cancer24. Upstream activators include ERBB receptors. ERBB3 interacts with p110α, the catalytic subunit of PI3K, and conveys signals from ERBB2–ERBB3 heterodimers. Epidermal growth factor receptor (EGFR; also known as ERBB1) induces PI3K activation via RAS activation. Overexpression of EGFR, ERBB2 and/or ERBB3 in subsets of bladder cancer is associated with grade, stage and outcome70,71,72,73, and ERBB2 and ERBB3 are mutated in some MIBCs36,37. ERBB2 amplification or overexpression is more common in metastases than in the corresponding primary tumour, implying a role in the metastatic process74. Activation of the receptor tyrosine kinases, MET and RON (also known as MST1R), also activates the PI3K pathway, and they are upregulated in aggressive bladder cancer75,76,77. It is not yet clear whether FGFR3 is a major activator of the PI3K pathway. Although mutant FGFR3 does not activate the pathway in NHUCs61, higher levels of phosphorylated AKT (which indicates that it is activated) have been reported in FGFR3-mutant tumours than in tumours with wild-type FGFR3 (Ref. 78).

Activating mutations of PIK3CA (which encodes p110α) occur more commonly in the helical domain (E545K and E542K) than in the kinase domain (H1047R) and are found in 25% of NMIBCs and less frequently in MIBCs24,25,78,79,80. p110α-E542K and p110α-E545K require interaction with RAS-GTP but not binding to the PI3K regulatory subunit p85 for activity, whereas p110α-H1047R activity depends on p85 binding but not on RAS-GTP binding81. p110α proteins with mutations in the helical domain possibly cooperate with events that activate RAS in bladder cancer. The expression of mutant PIK3CA confers a proliferative advantage at confluence and stimulates intraepithelial movement in NHUCs, and helical domain mutants show higher levels of activity82.

The lipid and protein phosphatase PTEN negatively regulates PI3K. PTEN commonly shows LOH in MIBC83,84,85, but biallelic inactivation is uncommon. Overall, 46% of bladder cancer cell lines (mostly derived from MIBC) had PTEN alterations24. Downregulated expression of PTEN in MIBC is associated with TP53 (which encodes human p53) alteration and poor outcome24,86. Consistent with this, urothelial Trp53 (which encodes mouse p53) and Pten dual deletion, but not deletion of either gene alone, leads to the development of metastatic bladder cancer in mice86.

The protein phosphatase activity of PTEN influences cell motility. PTEN-G129E, a mutant that is deficient in lipid but not protein phosphatase activity, inhibits the invasive phenotype of the PTEN-mutant T24 bladder cancer cell line87, which implies a more important role for PTEN protein phosphatase activity in invasive bladder cancer. This is consistent with the finding that only PIK3CA mutation, which is predicted to phenocopy loss of PTEN lipid phosphatase activity, occurs in NMIBC.

Other pathway alterations include TSC1 mutation (discussed above) and several mutations reported at low frequency (Tables 1,2).

Activation of the MAPK pathway. The role of MAPK signalling (Fig. 2) and its relationship to key mutations in bladder tumours are not clear. RAS (HRAS or KRAS) and FGFR3 mutations are mutually exclusive in bladder cancer. Mutation of one or the other in >82% of NMIBC88 may suggest that RAS and FGFR3 share a similar function. However, it is likely that there are also non-redundant functions, as RAS mutation is relatively infrequent compared with FGFR3 mutation and, unlike FGFR3 mutation, RAS mutation occurs at similar frequencies in NMIBC and MIBC. Although mutant FGFR3 activates MAPK pathway but not PI3K pathway signalling in NHUCs61, immunohistochemistry for phosphorylated ERK (an indicator of MAPK pathway activation) does not show a strong relationship with FGFR3 mutation or expression in tumour tissues78,89, and at least some bladder cancer cell lines are less sensitive to MEK inhibitors than NHUCs90. Although the finding that both FGFR3 mutation and PIK3CA mutation commonly co-occur in NMIBC78,80 suggests cooperative activation of MAPK and PI3K pathways, the exact role of FGFR3 in activation of the MAPK pathway requires further clarification. The recent finding of inactivating mutations in NOTCH pathway genes implicates this pathway in >40% of bladder cancers35, suggesting that the pathway has a tumour suppressor role in this cellular context. Higher levels of phosphorylated ERK1 and ERK2 were found in tumours with NOTCH pathway alterations than in FGFR3- or RAS-mutant tumours. This was mediated via reduced expression of several dual-specificity protein phosphatases (DUSPs) that target phosphorylated ERK, which are regulated by the intracellular domain of NOTCH1 (N1ic). Exogenous expression of N1ic or the ligand Jagged 1 (JAG1) in bladder cancer cells reduced phosphorylated ERK levels and inhibited proliferation. Overall, current data suggest that the majority of bladder cancers may be highly dependent on ERK.

Hedgehog and WNT signalling. Several components of the canonical WNT signalling pathway are altered in bladder cancer. Low frequencies of mutation in adenomatous polyposis coli (APC) and CTNNB1 (which encodes β-catenin) have been reported, and reduced expression or increased nuclear localization of β-catenin is frequent in MIBC25,91,92,93,94. Epigenetic silencing of the WNT antagonists, secreted frizzled receptor proteins (SFRPs)95 and WNT inhibitory factor 1 (WIF1)96, has also been reported.

The importance of WNT signalling is confirmed in mouse models. Expression of activated β-catenin in suprabasal urothelial cells in conjunction with Pten deletion led to bladder cancer development. Correlation between nuclear β-catenin, reduced PTEN expression and increased levels of phosphorylated AKT in human bladder cancer indicates that there is likely to be cooperation between WNT and PI3K signalling97. Another study that expressed active β-catenin in urothelial basal cells reported development of papillary tumours, which are found more commonly in male mice. Synergy between β-catenin and androgen receptor signalling was demonstrated, which implicates these factors in the observed sexual dimorphism of bladder cancer98. In mice, cooperation between β-catenin and mutant Hras is also reported, with MAPK pathway signalling rather than PI3K pathway signalling activated in the resulting tumours99. It will be important to examine whether such cooperation exists in human bladder cancer, and the relationship to gender.

The importance of hedgehog signalling in MIBC development has been demonstrated in mouse models and human bladder cancer100,101,102. In mice, urothelial regeneration following injury is driven by sonic hedgehog (SHH)-expressing basal cells, which elicit secretion of factors, including the transcription factor GLI1 and WNT pathway proteins by stromal cells, which in turn stimulate proliferation and differentiation of urothelial cells100. These cells were also the progenitors of MIBC in the N-butyl-N-4-hydroxybutyl nitrosamine (BBN)-induced urothelial cancer model101. Although these progenitor cells and normal human urothelium express SHH, expression is lost in BBN-MIBC in mice and in human MIBC cell lines103 and tissues102. Dissection of BBN-induced tumorigenesis has revealed that loss of hedgehog signalling blocks production of stromal factors that induce urothelial differentiation, including bone morphogenetic protein 4 (BMP4) and BMP5, suggesting that CIS progression to MIBC is triggered by loss of hedgehog signalling. Importantly, tumour progression could be blocked by pharmacological activation of the BMP pathway, suggesting a possible therapeutic approach for human NMIBC102. Notably, SHH, BMP4 and BMP5 were shown to be significantly downregulated in mRNA sequencing data from The Cancer Genome Atlas (TCGA) study of MIBC, particularly in the aggressive 'basal' subtype102.

Cell cycle regulation. Almost every MIBC has defects in genes encoding proteins that control the G1 cell cycle checkpoint. Inactivation of TP53, RB1 and CDKN2A is common (Table 2) and has adverse prognostic importance104. Taken together with amplification or overexpression of MDM2, p53 function was predicted to be inactivated in 76% of MIBCs37. Similarly, RB1 loss is common, and amplification and overexpression of E2F3, which is normally repressed by RB1, is associated with RB1 or p16 loss in MIBC105.

Cyclin D1 (CCND1) and CCND3 are implicated in NMIBC. CCND1 (11q13) is amplified in 20% of bladder cancers106. High nuclear expression in 33% of stage Ta and T1 tumours was associated with higher proliferative index and reduced disease-free survival. High expression of CCND3 (in 13% of NMIBCs) was also associated with reduced survival107. Upregulated expression of these cyclins may represent a more specific mechanism of inactivation of the G1 checkpoint in NMIBC.

Epigenetic alterations: chromatin modifiers at centre stage. Extensive DNA methylation changes have been reported in bladder cancer, and many of these show clinicopathological associations108,109,110,111,112,113,114,115,116. A major subtype of MIBC has been identified that has high-level promoter hypermethylation associated with smoking pack-years37. Comparisons of NMIBC and MIBC reveal distinct patterns of hypomethylation in non-CpG islands in NMIBC and widespread CpG island hypermethylation in MIBC114,115. Whereas hypermethylation in promoters is linked to gene silencing, hypomethylation within gene bodies is usually associated with upregulated expression116. In medulloblastoma, such regions of hypomethylation are marked by histone H3 lysine 4 (H3K4) trimethylation, a marker of open chromatin117. Genome-wide analysis of methylation and repressive histone marks in bladder cancer also indicates the importance of both DNA methylation and histone methylation in gene silencing118. Genomic regions showing DNA copy number-independent transcriptional deregulation were found to be associated with H3K9 and H3K27 methylation and H3K9 hypoacetylation rather than with DNA methylation, and this pattern was related to a CIS-associated expression signature119.

Genome sequencing has identified mutations in chromatin regulators that occur at higher frequency in MIBC than in other epithelial cancers11,17,36,37 (Table 2). 89% of MIBCs had a mutation in one or more chromatin-regulating genes37. Frequently mutated genes include lysine-specific demethylase 6A (KDM6A, which encodes a histone demethylase), mixed-lineage leukaemia 2 (MLL2; also known as KMT2D, which encodes a histone methyltransferase) and AT-rich interactive domain 1A (ARID1A, which encodes a component of the SWI/SNF chromatin-remodelling complex). KDM6A demethylates H3K27, leading to a more open chromatin configuration. MLL2 methylates H3K4, which also favours euchromatin formation and suggests transcriptional activation. Thus, loss of function of these genes is predicted to lead to gene silencing. Other genes involved in chromatin modification that are mutated in >5% of MIBC samples include the following: MLL (also known as KMT2A) and MLL3 (also known as KMT2C), which encode histone methyltransferases; E1A-binding protein p300 (EP300) and CREB-binding protein (CREBBP), which encode histone acetyltransferases; nuclear receptor co-repressor 1 (NCOR1), which encodes a histone deacetylase; CHD6 and CHD7, which encode chromodomain helicase DNA-binding proteins; and a CREBBP activator, Snf2-related CREBBP activator protein (SRCAP)17,36. The predominance of inactivating mutations implicates these as tumour suppressor genes. Although several genes are mutated at low frequency, they may be functionally redundant, which potentially reduces complexity to fewer phenotypic subgroups. For example, mutations in KDM6A and MLL2 were reported to be mutually exclusive37.

Molecular biomarkers

A detailed discussion of prognostic biomarkers is beyond the scope of this Review. However, the application of molecular biomarkers for non-invasive monitoring of NMIBC is approaching clinical applicability and merits brief discussion here. As NMIBCs are poorly detected by urine cytology, analysis of FGFR3 mutation combined with other DNA-based biomarkers provides a useful test for disease monitoring120,121,122. Many studies report detection of bladder cancer by analysis of methylation biomarkers in urine. Examples include several panels of biomarkers that are used alone114,123,124,125 or in combination with FGFR3 mutation126. The most common event described in bladder cancer so far is point mutation of the telomerase reverse transcriptase (TERT) promoter in 80% of tumours, regardless of grade and stage. Mutations are mainly at positions −124 bp and −146 bp relative to the transcriptional start site, and such a finding allows development of specific assays that are suitable for detection in urine127,128. Undoubtedly, when used in combination with other urine biomarkers, this will improve sensitivity for the detection of bladder cancer.


Cell of origin. The normal human urothelium comprises a layer of basal cells that are in contact with the basement membrane, several layers of intermediate cells and a single layer of large, superficial 'umbrella' cells with a specialized apical membrane to accommodate bladder expansion and contraction129. Lineage-tracing studies in mice using SHH–Cre have identified SHH+ basal cells that can repopulate the entire urothelium following injury100. Whether this is the case in humans is unknown, but the presence of large monoclonal areas (up to 4.7 mm in diameter) within the urothelium that comprise basal, intermediate and superficial cell layers suggests that the urothelium originates from single basal stem cells130.

Isolation of human bladder cancer stem cells (or tumour-initiating cells) has been achieved using a range of assays (reviewed in Refs 131,132). These show features of basal cells residing at the tumour–stromal interface133. CD44+ cytokeratin 5 (KRT5)+ KRT20 tumour cells had enhanced tumour-initiating ability compared with CD44 KRT5 KRT20+ cells and could give rise to tumours containing both CD44+ and CD44 cells in mice134. As discussed above, lineage-tracing studies identified SHH+ basal cells as tumour-initiating cells in mouse BBN-induced invasive tumours101. Recent evidence suggests that such cancer stem cells may contribute to therapeutic resistance by repopulating residual tumours between chemotherapy cycles135.

In humans, there is evidence for non-basal tumour-initiating cells in more-differentiated bladder cancers. Marker combinations corresponding to different urothelial differentiation states could stratify bladder cancer into clinically relevant subgroups, and tumours with the least differentiated (basal) tumour-initiating cells had the worst outcome136. Similar evidence was reported from a lineage-tracing study in mice, in which cells in the intermediate layer were implicated as the cell of origin for BBN-induced papillary tumours137. Diversity in cancer stem cell phenotype may contribute to the divergent development of NMIBC and MIBC and, in turn, define the genomic events that subsequently participate in tumour development.

Molecular features of pre-neoplastic urothelium. Chromosome 9 LOH is found in 'normal' urothelium and hyperplasia in patients with NMIBC138,139,140,141, and at higher frequency than FGFR3 mutation in flat hyperplasia, implying that chromosome 9 loss is an earlier event than FGFR3 mutation142. Dysplasia and CIS are considered precursors of MIBC and show chromosome 9 LOH, TP53 mutation143,144,145 and multiple other chromosomal alterations146. No FGFR3 mutations have been reported in CIS147. In morphologically 'normal' or dysplastic urothelium adjacent to MIBC, alterations found in the tumour are present, indicating spread of cells from the tumour or tumour development within a field of altered urothelium148,149,150,151. Detailed mapping of entire cystectomy specimens reveals that, in the absence of detectable morphological abnormality, large macroscopically 'normal' urothelial patches show LOH of specific chromosomal regions. Areas of mild, moderate or severe dysplasia show complex patterns of LOH, suggesting that there are sequential evolutionary 'waves' of change associated with the acquisition of growth advantage. Thus, it is envisaged that the development of MIBC involves clonal expansion, within which subclones with additional alterations arise (Fig. 4). Six critical regions of LOH were identified, and biallelic inactivation of integral membrane protein 2B (ITM2B) and lysophosphatidic acid receptor 6 (LPAR6; also known as P2RY5) was demonstrated in one of those regions150,152.

Figure 4: Pre-neoplastic urothelial alterations in the tumour-bearing bladder.

a | A schematic of what the surgeon sees is shown. At presentation the primary tumour and an area of inflamed urothelium subsequently diagnosed as carcinoma in situ (CIS) are resected. During surveillance a second tumour and an associated area of CIS are visually identified and resected. b | The molecular picture is shown. Large areas of urothelium have been replaced with altered cells. Different molecular events are indicated by letters. The primary tumour has developed within a large area of urothelium of normal appearance containing molecular alteration 'c'. Within this 'field' of change a dysplastic area (which also has normal appearance at cystoscopy) represents a subclone with 'c+d' alterations; within this dysplastic area, a patch of CIS representing the immediate precursor to the overt tumour ('c+d+e+f') is recognized as an inflamed patch ('c+d+e'). A second tumour and CIS resected at a later date have also developed within the large altered field containing alteration 'c' but differ in all other alterations (g, h and i). The possibility that tumours with independent initiating events may develop in some patients is indicated by patches 'a' and 'b', both of which may remain morphologically silent for many years. The dotted line indicates a urothelial patch as described by Gaisa et al.124, which is putatively derived from a single stem cell. Despite independent tumour-initiating events, cells with alterations 'a' and 'b' are clonally related. The presence of large and clinically silent areas of pre-neoplastic urothelium emphasizes the need for novel approaches for local therapeutic targeting.

PowerPoint slide

Multifocality, clonality and chronology of events during urothelial carcinoma pathogenesis. The development of multiple bladder cancers in the same patient is common, and this enables examination of clonality and molecular evolution. Two concepts have been proposed to explain the origin of multifocal bladder cancer. As a result of extensive carcinogenic insults, many cells may become altered and give rise to independent tumours. Alternatively, a single clone may spread via intraepithelial migration or implantation. Although a few patients develop more than one apparently independent tumour (that is, oligoclonal disease)153, tumours from the same patient are commonly related154, and there is evidence for subclonal genomic evolution, sometimes complex, in different lesions (reviewed in Ref. 155). Loss of the same allele on chromosome 9 in all related tumours again indicates that this is an early event156,157,158. Construction of phylogenetic trees from multiple tumours suggests that 9q, 9p and 11p occur early in NMIBC, and 8p, 20q+, 17p and 11q are later events. Interestingly, tumours with the highest genomic complexity are not necessarily the last to appear157,159,160, which provides an explanation for the observation that recurrent tumours may be of lower grade than preceding tumours161. These observations provide further evidence for widespread 'field change' in the diseased bladder. The molecular complexity of related tumours suggests that considerable intratumoural heterogeneity may be a feature of individual bladder cancers, although this has not yet been systematically investigated. Analyses so far on bulk tumour samples may not reveal such heterogeneity. Deep sequencing should now allow such complexity and phylogeny to be examined in more detail.

EMT and metastasis. EMT is a reversible process that involves changes in cell morphology, differentiation and motility, facilitating invasion and metastasis. Markers of EMT — loss of expression of E-cadherin and tight-junction proteins, and upregulation of vimentin and fibronectin — are associated with MIBC, resistance to therapeutic agents and poor outcome.

EMT is mediated by zinc-finger E-box binding homeobox 1 (ZEB1), ZEB2, TWIST, SNAIL (also known as SNAI1) and SLUG (also known as SNAI2), which transcriptionally repress epithelial markers. ZEB1 and ZEB2 are regulated by members of the miR-200 family and miR-205 (Ref. 162). miR-200 is downregulated in bladder cancer cell lines with mesenchymal phenotype, and epigenetic silencing is reported in MIBC163. These miRNAs also regulate ERBB receptor feedback inhibitor 1 (ERRFI1), and their silencing confers resistance to EGFR in mesenchymal bladder cancer cells that can be reversed by the expression of miR-200 (Ref. 164).

FGFR1 signalling and subsequent COX2 upregulation can induce EMT in bladder cancer cells68 (discussed above). In an animal model of bladder cancer metastasis using an FGFR1-dependent cell line, FGFR inhibition reduced the development of circulating tumour cells and metastasis but not primary tumour growth69. Interestingly, although FGFR2 shows reduced expression in MIBC165, cells selected for enhanced metastatic capability were found to be dependent on FGFR2 isoform IIIc for this phenotype, which was associated with a mesenchymal–epithelial transition166. Thus, FGFRs are implicated both in the EMT required early in the process of metastasis and in the mesenchymal–epithelial transition required to recapitulate the epithelial phenotype at the metastatic site.

Other factors implicated in bladder cancer cell EMT include the transcription factor inhibitor of DNA binding 1 (ID1)167 and the long non-coding RNA (lncRNA) H19 (which is upregulated in MIBC). H19 promotes WNT–β-catenin activation via association with enhancer of zeste homologue 2 (EZH2) and subsequent downregulation of E-cadherin168. Similarly, the lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), which is upregulated by transforming growth factor-β (TGFβ) in bladder cancer cells, induces EMT169. TP63 (which encodes p63) is expressed in the basal and intermediate cell layers of the normal urothelium. Studies of TP63 isoforms show that ΔNp63 expression is linked to EMT in tumours170, and this is associated with a more 'basal' phenotype with adverse prognosis170,171,172. Recent expression profiling of large numbers of MIBC samples defines a claudin-low basal subtype (discussed below) that is characterized by the enrichment of EMT markers and the expression of low levels of cytokeratins37,173.

Metastasis of NMIBC is rare, but half of all MIBCs metastasize. Cisplatin-based neoadjuvant chemotherapy before cystectomy confers a small survival benefit, and this is thought to be due to the elimination of occult metastases, but survival is poor and improved therapeutic approaches are urgently needed. In addition to genes involved in EMT, preclinical studies suggest that RHO-GDP dissociation inhibitor 2 (RHOGDI2), the activity of which is regulated by SRC, is a metastasis suppressor in MIBC174,175. In contrast to most tumour types, the expression and activation of SRC is highest in NMIBC175,176,177, which is consistent with the activity of RHOGDI2 in these tumours. Further work on the consequences of loss of RHOGDI2 expression implicated endothelin 1 (EDN1) and versican (VCAN) as promoters of an inflammatory environment that involves macrophages and the chemokine (C-C motif) ligand 2 (CCL2)–chemokine (C-C motif) receptor 2 (CCR2) signalling axis that is permissive for metastasis178,179. A similar relationship has been reported between loss of SPARC (secreted protein, acidic, cysteine-rich) and metastasis and regulation of inflammatory response in bladder cancer180. Reduced expression of activating transcription factor 3 (ATF3), which is also a regulator of inflammation, has been implicated in driving metastasis through transcriptional regulation of gelsolin-mediated actin remodelling. Both SPARC and ATF3 are downregulated in MIBC, and ATF3 was shown to reduce metastasis in an in vivo experimental model181. The RAL GTPases RALA and RALB have also been implicated in bladder cancer metastasis182,183. CD24 — a RAL GTPase-regulated glycosyl phosphatidylinositol (GPI)-linked sialoglycoprotein — is required for metastatic bladder cancer colonization in mice and shows increased expression in human bladder cancer metastases184. Interestingly, prognostic value seems to be confined to male patients; in Cd24-null mice, reduced carcinogen-induced tumour incidence was seen only in males185. As CD24 is androgen-responsive, if these data are confirmed, then androgen ablation could have therapeutic impact in males185.

Beyond the two-pathway model

The two tumour groupings that have dominated the literature on bladder cancer cannot provide an explanation for the considerable heterogeneity in clinical behaviour. The two-pathway model has been expanded in Fig. 5 to include potential pathogenic links that are suggested by molecular data and the possibility of multiple molecular routes to each recognized disease state. Previous gene expression profiling studies have reported signatures associated with stage, grade and outcome186,187,188,190,191. Recent studies now begin to unravel the heterogeneity in clinical behaviour, revealing multiple molecular subtypes that traverse grade and stage groupings.

Figure 5: Potential pathways of urothelial carcinoma pathogenesis.

Potential pathogenesis pathways are shown based on histopathological and molecular observations. The blue and purple pathways indicate the two major pathways with distinct histopathological and molecular features that have been recognized for the past two decades. Percentages at diagnosis and key molecular features are indicated. The recent identification of multiple distinct molecular subtypes of non-muscle invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) suggests multiple subpathways within each of the major pathways. It should be noted that the existence of multiple subtypes within all of the histologically recognizable intermediates remains hypothetical. Development of histopathologically recognizable urothelial alterations is preceded by clonal expansion of altered cells within the urothelium. Low-grade papillary tumours may arise via simple hyperplasia and minimal dysplasia, and these are characterized at the molecular level by loss of heterozygosity (LOH) of chromosome 9 and activating mutations of fibroblast growth factor receptor 3 (FGFR3), PIK3CA (which encodes p110α of PI3K) and stromal antigen 2 (STAG2). These non-invasive tumours recur frequently but are genetically stable. Invasive carcinoma is thought to arise via flat dysplasia and carcinoma in situ (CIS), which commonly show TP53 mutation in addition to chromosome 9 deletions but no FGFR3 mutations. Invasive tumours are genetically unstable and accumulate many genomic alterations. Uncertainty about the possible development of high-grade non-invasive papillary tumours from flat dysplasia is indicated by a dashed arrow. The finding of a subtype of invasive carcinomas with FGFR3 mutation and loss of cyclin-dependent kinase inhibitor 2A (CDKN2A) may suggest a route by which low-grade non-invasive papillary tumours can progress to muscle invasion (dashed arrow). ARID1A, AT-rich interactive domain 1A; EMT, epithelial–mesenchymal transition; RHOGDI1, RHO-GDP dissociation inhibitor 2; ZEB1, zinc-finger E-box binding homeobox 1.

PowerPoint slide

An mRNA expression analysis of bladder cancers of all grades and stages by Sjödahl et al. identified five major subtypes termed urobasal A (UroA), UroB, genomically unstable (GU), squamous cell carcinoma-like (SCCL) and 'infiltrated' (which is highly infiltrated with non-tumour cells)192 (Fig. 6). Subsequently, it was suggested that the SCCL group should be termed 'basal', as it shares features with basal-type breast cancers193. UroA and UroB tumours express FGFR3, CCND1 and p63; GU tumours express low levels of these proteins but high levels of ERBB2 and E-cadherin; and SCCL (or basal) tumours express EGFR, P-cadherin, KRT5, KRT6, KRT14 and proteins involved in keratinization. UroA tumours showed good prognosis; GU and infiltrated bladder cancers showed intermediate prognosis; and SCCL (or basal) and UroB tumours had the worst prognosis. Although UroB tumours have TP53 mutation and many were MIBCs, they show epithelial characteristics, including FGFR3 mutation, which may indicate evolution from UroA tumours. UroB tumours also show homozygous deletion of CDKN2A, which could be a mechanism by which FGFR3-mutant NMIBCs progress31 (discussed above).

Figure 6: Molecular subtypes of bladder cancer.

Subtype classification according to four mRNA expression profiling studies is shown. a | Sjödahl et al.192 defined five major subtypes termed urobasal A (UroA), UroB, genomically unstable (GU), squamous cell carcinoma-like (SCCL) and 'infiltrated'. b | Damrauer et al.173 classified bladder cancer into basal and luminal subtypes. c | The Cancer Genome Atlas (TCGA) study37 defined four expression clusters (I–IV). d | Choi et al.194 defined a 'p53-like' luminal subtype. Heat maps show representative expression levels of selected genes that are characteristic of the molecular subtypes defined in each study. Dark red indicates high expression; light blue represents low expression; and pale red indicates intermediate expression, heterogeneous expression or expression only in certain layers of the urothelium. Mutation status, copy number alterations, histological characteristics and information relating to prognosis and treatment response are shown where such information is available. Grey boxes indicate the presence of a mutation, copy number alteration or histological feature. CCND1, cyclin D1; DSS, disease-specific survival; EGFR, epidermal growth factor receptor; FGFR3, fibroblast growth factor receptor 3; FOXA1, forkhead box A1; GATA3, GATA-binding protein 3; KRT, cytokeratin; MI, muscle-invasive; NAC, neoadjuvant chemotherapy; OS, overall survival; PPARG, peroxisome proliferator-activated receptor-γ; TP63 encodes p63; TSC1, tuberous sclerosis 1; UPK, uroplakin.

PowerPoint slide

Three subsequent studies of MIBC have defined transcriptional subtypes37,173,194 with considerable overlap195. There are two major 'basal' and 'luminal' subtypes that show similarities to intrinsic breast cancer subtypes (Fig. 6). Luminal MIBCs are enriched for uroplakins, KRT20, ERBB2 and differentiation markers such as forkhead box A1 (FOXA1; also known as HNF3α), GATA-binding protein 3 (GATA3), tripartite motif-containing protein 24 (TRIM24) and peroxisome proliferator-activated receptor-γ (PPARγ), and these frequently have papillary morphology and FGFR3 upregulation and/or mutation. A TCGA study defined four expression clusters (I–IV), one of which (Cluster I) expressed luminal markers37. Choi et al.194 described a luminal subtype termed 'p53-like' with an activated wild-type p53 signature, low levels of cell cycle and proliferation markers, and enrichment of myofibroblast and extracellular matrix markers, perhaps reflecting stromal and fibroblast infiltration195. The infiltrated subgroup of Sjödahl et al.192 mostly overlapped with this p53-like subgroup. Tumours in TCGA Cluster II shared features with luminal and p53-like subtypes, and some basal and luminal tumours described by Damrauer et al.173 also exhibited characteristics of the p53-like subgroup. Tumours expressing basal markers (KRT5, KRT6, KRT14, CD44 and CDH3) were present in all three studies. Two studies identified claudin-low basal tumours expressing markers that are characteristic of EMT and that have low levels of cytokeratins, which is analogous to some breast cancers37,173. Some infiltrated tumours also showed overlap with this subgroup. In all studies, the SCCL (or basal) MIBCs showed the worst prognosis, and those with papillary architecture and high expression levels of FGFR3, E-cadherin, GATA3, FOXA1 and uroplakins had the best prognosis (Fig. 6). Bioinformatics analyses implicated transcription factors that are active in the basal or stem cell compartment of the normal urothelium — signal transducer and activator of transcription 3 (STAT3), nuclear factor-κΒ (NF-κB), hypoxia-inducible factor 1 (HIF1) and p63 — as potential regulators in basal tumours, and PPARγ and oestrogen receptor pathways were implicated in luminal tumours194. A recent pan-cancer molecular classification also defined a squamous subtype that contains four different tumour types, including a subset of TCGA MIBC samples that show worse overall survival than other bladder cancers196. Network analysis revealed significant TP63 expression that was associated with a high level of expression of the oncogenic ΔNp63 isoform. It is likely that SCCL (or basal) MIBCs expressing this isoform represent the lethal subset of p63-expressing advanced bladder cancers that were reported by others170. The relevance of p63 expression in UroB tumours with poor prognosis remains unresolved, but the expression of different p63 isoforms in NMIBC and MIBC may account for this170. It is currently unclear whether analyses that have assessed only MIBC have revealed all biologically relevant heterogeneity, as some data suggest that transcriptional subtypes are independent of conventional grade and stage groupings192. There is an ongoing debate on how to integrate and explore this complexity197,198.

Recent studies also indicate potential predictive value. Activation of the EGFR pathway has been identified in basal-type MIBC, and it was shown that cell lines with this signature were sensitive to EGFR inhibition199. Additionally, the p53-like subgroup of Choi et al.194 contained tumours that were resistant to cisplatin-based chemotherapy, and it was confirmed in independent samples and in bladder cancer cell lines that this signature had predictive value. Although the signature did not identify all resistant tumours because half of the basal-type tumours were also resistant, this represents an important step towards more rational selection of therapy.

Further work will be needed to generate consensus classifiers and robust assays that are suitable for clinical implementation. A simple classifier based on two histology variables (grade and urothelial differentiation pattern) and the expression of two proteins (KRT5 and CCNB1) has been reported, which reproduced an original genome-wide expression classification192 with an accuracy of 0.88 (Ref. 193). A notable finding was that UroA and UroB tumours retain expression pattern and histology that are reminiscent of the normal urothelium, with expression of several markers, including proliferative biomarkers, in basal cells only, which implies retention of dependence on stromal interactions. SCCL (or basal) tumours had lost this pattern and showed CDH3, KRT5 and EGFR at higher level throughout the tumour parenchyma. This subtype showed a keratin expression profile (KRT14+ KRT5+ KRT20) that characterizes the least differentiated class of tumour-initiating cells described by Chan et al.134 and the class termed basal in the study of Volkmer et al.136.

DNA-based genome-wide analyses also indicate the existence of multiple subclasses of bladder cancer. Analysis of DNA copy number and mutation status has identified multiple genomic subtypes of tumours within the conventional grade and stage groupings43, and MIBCs have been subdivided into three major groups according to copy number alterations and mutation status37,43. The latter groups separate tumours with frequent TP53 mutation from those with FGFR3 mutation and CDKN2A loss, and define a group with an increased frequency of focal amplification and mutations in MLL2 (Ref. 37). DNA methylation profiles also identify bladder cancer subtypes37,115,116,200. Four “epitypes” (Ref. 116) showed broad alignment with previously defined expression subtypes192,201. A key feature of these epitypes was differential expression of homeobox (HOX) genes. Repression of several HOX genes that are also repressed in pluripotent cells was found in the most aggressive epitype (type D).

A major goal is to integrate information from all platforms to provide the best prognostic and predictive biomarkers for clinical application. As DNA and miRNAs are more robust molecules than mRNA in formalin-fixed paraffin-embedded (FFPE) specimens, a combination of these and protein biomarkers may ultimately provide the most useful classifiers.

Conclusions and future perspectives

For many years, molecular understanding of bladder cancer biology has lagged behind that of other solid cancers, and this has represented a major barrier to improving clinical care. Although the long-described disease subtypes (NMIBC and MIBC) are clearly distinct at the molecular level, this has not provided adequate prognostic or predictive information for clinical application. With the advent of large-scale genome-wide profiling studies, the field is now poised to replace the conventional two-pathway model of bladder cancer pathogenesis with a more complex and molecularly credible description of disease pathogenesis and clinical behaviour and to meet the clinical needs of patients in a more personalized way.

Although non-invasive disease is not life-threatening, its recurrent nature makes this more expensive to treat than other cancers, and disease monitoring and treatment are associated with considerable morbidity. Key biomarker panels already hold potential for application in non-invasive urine-based monitoring. A remaining challenge is to develop therapies that can better treat localized disease. Knowledge of the widespread fields of altered cells present in many patients argues for early therapeutic intervention and enhanced efforts to clarify the common and targetable early events in NMIBC. Few NMIBCs have been studied by exome or whole-genome sequencing, and it will be important to establish detailed mutation profiles for this important patient population.

For MIBC, molecular profiling in the context of large clinical trials is required to confirm the ability of molecular signatures to stratify patients before treatment with conventional chemotherapy. For these patients, recent insights should also guide the design of new clinical trials, with many possibilities for therapy with targeted agents. The striking finding of frequent somatic mutation of epigenetic modifiers in bladder cancer may offer completely new treatment avenues, where the potential reversibility of phenotype may hold promise for the treatment of some tumour subtypes. Stage T1 tumours are a particularly difficult group to manage. Progression to invasion is relatively common, but there is a major debate on whether to remove or preserve the bladder. Evidence from studies that include some T1 tumours indicates that, as for MIBC, these may be subclassified to provide objective guidance for treatment. Large collaborative studies are now needed to allow identification of robust prognostic biomarkers and, for all subtypes, robust diagnostic assays are needed.


  1. 1

    Ferlay, J. et al. Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008. Int. J. Cancer 127, 2893–2917 (2010).

    Article  CAS  PubMed  Google Scholar 

  2. 2

    Dudek, A. M., Grotenhuis, A. J., Vermeulen, S. H., Kiemeney, L. A. & Verhaegh, G. W. Urinary bladder cancer susceptibility carkers. What do we know about functional mechanisms? Int. J. Mol. Sci. 14, 12346–12366 (2013). This is an excellent review of findings of genome-wide association studies in bladder cancer and possible functional mechanisms.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. 3

    Union Internationale Contre le Cancer (UICC) in TNM Classification of Malignant Tumors 7th edn 262–265 (UICC, 2009).

  4. 4

    Eble, J. N., Sauter, G., Epstein, J. I. & Sesterhenn, I. A. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Urinary System and Male Genital Organs (IARC Press Lyon, 2004).

    Google Scholar 

  5. 5

    World Health Organization. Histological Typing of Urinary Bladder Tumours. (WHO, 1973).

  6. 6

    Prout, G. R. Jr., Barton, B. A., Griffin, P. P. & Friedell, G. H. Treated history of noninvasive grade 1 transitional cell carcinoma. The National Bladder Cancer Group. J. Urol. 148, 1413–1419 (1992).

    Article  PubMed  Google Scholar 

  7. 7

    Sylvester, R. J. et al. Predicting recurrence and progression in individual patients with stage Ta T1 bladder cancer using EORTC risk tables: a combined analysis of 2596 patients from seven EORTC trials. Eur. Urol. 49, 466–477 (2006).

    Article  PubMed  Google Scholar 

  8. 8

    Soloway, M. S. Bladder cancer: lack of progress in bladder cancer — what are the obstacles? Nature Rev. Urol. 10, 5–6 (2013).

    Article  Google Scholar 

  9. 9

    Morrison, C. D. et al. Whole-genome sequencing identifies genomic heterogeneity at a nucleotide and chromosomal level in bladder cancer. Proc. Natl Acad. Sci. USA 111, E672–E681 (2014).

    Article  CAS  PubMed  Google Scholar 

  10. 10

    Bentley, J., Diggle, C. P., Harnden, P., Knowles, M. A. & Kiltie, A. E. DNA double strand break repair in human bladder cancer is error prone and involves microhomology-associated end-joining. Nucleic Acids Res. 32, 5249–5259 (2004).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. 11

    Balbas-Martinez, C. et al. Recurrent inactivation of STAG2 in bladder cancer is not associated with aneuploidy. Nature Genet. 45, 1464–1469 (2013).

    Article  CAS  PubMed  Google Scholar 

  12. 12

    Solomon, D. A. et al. Frequent truncating mutations of STAG2 in bladder cancer. Nature Genet. 45, 1428–1430 (2013).

    Article  CAS  PubMed  Google Scholar 

  13. 13

    Taylor, C., Platt, F., Hurst, C., Thygensen, H. & Knowles, M. A. Frequent inactivating mutations of STAG2 in bladder cancer are associated with low tumor grade and stage and inversely related to chromosomal copy number changes. Hum. Mol. Genet. 23, 1964–1974 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. 14

    Solomon, D. A. et al. Mutational inactivation of STAG2 causes aneuploidy in human cancer. Science 333, 1039–1043 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. 15

    Baranello, L., Kouzine, F. & Levens, D. CTCF and cohesin cooperate to organize the 3D structure of the mammalian genome. Proc. Natl Acad. Sci. USA 111, 889–890 (2014).

    Article  CAS  PubMed  Google Scholar 

  16. 16

    Kong, X. et al. Distinct functions of human cohesin-SA1 and cohesin-SA2 in double-strand break repair. Mol. Cell. Biol. 34, 685–698 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. 17

    Guo, G. et al. Whole-genome and whole-exome sequencing of bladder cancer identifies frequent alterations in genes involved in sister chromatid cohesion and segregation. Nature Genet. 45, 1459–1463 (2013).

    Article  CAS  PubMed  Google Scholar 

  18. 18

    Cairns, P. et al. Rates of p16 (MTS1) mutations in primary tumors with 9p loss. Science 265, 415–417 (1994).

    Article  CAS  PubMed  Google Scholar 

  19. 19

    Williamson, M. P., Elder, P. A., Shaw, M. E., Devlin, J. & Knowles, M. A. p16 (CDKN2) is a major deletion target at 9p21 in bladder cancer. Hum. Mol. Genet. 4, 1569–1577 (1995).

    Article  CAS  PubMed  Google Scholar 

  20. 20

    Aboulkassim, T. O., LaRue, H., Lemieux, P., Rousseau, F. & Fradet, Y. Alteration of the PATCHED locus in superficial bladder cancer. Oncogene 22, 2967–2971 (2003).

    Article  CAS  PubMed  Google Scholar 

  21. 21

    McGarvey, T. W., Maruta, Y., Tomaszewski, J. E., Linnenbach, A. J. & Malkowicz, S. B. PTCH gene mutations in invasive transitional cell carcinoma of the bladder. Oncogene 17, 1167–1172 (1998).

    Article  CAS  PubMed  Google Scholar 

  22. 22

    Habuchi, T., Luscombe, M., Elder, P. A. & Knowles, M. A. Structure and methylation-based silencing of a gene (DBCCR1) within a candidate bladder cancer tumor suppressor region at 9q32–q33. Genomics 48, 277–288 (1998).

    Article  CAS  PubMed  Google Scholar 

  23. 23

    Nishiyama, H., Hornigold, N., Davies, A. & Knowles, M. A. A sequence-ready 840-kb PAC contig spanning the candidate tumor suppressor locus DBC1 on human chromosome 9q32-q33. Genomics 59, 335–338 (1999).

    Article  CAS  PubMed  Google Scholar 

  24. 24

    Platt, F. M. et al. Spectrum of phosphatidylinositol 3-kinase pathway gene alterations in bladder cancer. Clin. Cancer Res. 15, 6008–6017 (2009).

    Article  CAS  PubMed  Google Scholar 

  25. 25

    Sjodahl, G. et al. A systematic study of gene mutations in urothelial carcinoma; inactivating mutations in TSC2 and PIK3R1. PLoS ONE 6, e18583 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. 26

    Ploussard, G. et al. Prognostic value of loss of heterozygosity at chromosome 9p in non-muscle-invasive bladder cancer. Urology 76, e513–e518 (2010).

    Article  Google Scholar 

  27. 27

    Kruger, S., Mahnken, A., Kausch, I. & Feller, A. C. p16 immunoreactivity is an independent predictor of tumor progression in minimally invasive urothelial bladder carcinoma. Eur. Urol. 47, 463–467 (2005).

    Article  CAS  PubMed  Google Scholar 

  28. 28

    Bartoletti, R. et al. Loss of p16 expression and chromosome 9p21 LOH in predicting outcome of patients affected by superficial bladder cancer. J. Surg. Res. 143, 422–427 (2007).

    Article  CAS  PubMed  Google Scholar 

  29. 29

    Carnero, A., Hudson, J. D., Price, C. M. & Beach, D. H. p16INK4A and p19ARF act in overlapping pathways in cellular immortalization. Nature Cell Biol. 2, 148–155 (2000).

    Article  CAS  PubMed  Google Scholar 

  30. 30

    Serrano, M. The INK4a/ARF locus in murine tumorigenesis. Carcinogenesis 21, 865–869 (2000).

    Article  CAS  PubMed  Google Scholar 

  31. 31

    Rebouissou, S. et al. CDKN2A homozygous deletion is associated with muscle invasion in FGFR3-mutated urothelial bladder carcinoma. J. Pathol. 227, 315–324 (2012). This study suggests that deletion of p16 makes a critical contribution to progression of NMIBCs. Subsequent genome sequencing studies (such as reference 37) define a subgroup of MIBCs with this profile.

    Article  CAS  PubMed  Google Scholar 

  32. 32

    Le Frere-Belda, M. A. et al. Profiles of the 2 INK4a gene products, p16 and p14ARF, in human reference urothelium and bladder carcinomas, according to pRb and p53 protein status. Hum. Pathol. 35, 817–824 (2004).

    Article  CAS  PubMed  Google Scholar 

  33. 33

    Benedict, W. F. et al. Level of retinoblastoma protein expression correlates with p16 (MTS-1/INK4A/CDKN2) status in bladder cancer. Oncogene 18, 1197–1203 (1999).

    Article  CAS  PubMed  Google Scholar 

  34. 34

    Shariat, S. F. et al. p53, 21, pRB, and p16 expression predict clinical outcome in cystectomy with bladder cancer. J. Clin. Oncol. 22, 1014–1024 (2004).

    Article  CAS  PubMed  Google Scholar 

  35. 35

    Rampias, T. et al. A new tumor suppressor role for the Notch pathway in bladder cancer. Nature Med. 20, 1199–1205 (2014).

    Article  CAS  PubMed  Google Scholar 

  36. 36

    Gui, Y. et al. Frequent mutations of chromatin remodeling genes in transitional cell carcinoma of the bladder. Nature Genet. 43, 875–878 (2011). This was the first major whole-exome sequencing study of bladder cancer, which identified frequent mutation of chromatin modifier genes.

    Article  CAS  PubMed  Google Scholar 

  37. 37

    The Cancer Genome Atlas Research Network. Comprehensive molecular characterization of urothelial bladder carcinoma. Nature 507, 315–322 (2014). This is the most comprehensive analysis of MIBC so far, including whole-exome sequencing, RNA sequencing and profiling of microRNAs, methylation and protein expression. See also references 173 and 194.

  38. 38

    Hoglund, M. The bladder cancer genome; chromosomal changes as prognostic makers, opportunities, and obstacles. Urol. Oncol. 30, 533–540 (2012).

    Article  CAS  PubMed  Google Scholar 

  39. 39

    Knowles, M. A. Molecular subtypes of bladder cancer: Jekyll and Hyde or chalk and cheese? Carcinogenesis 27, 361–373 (2006).

    Article  CAS  PubMed  Google Scholar 

  40. 40

    Knowles, M. A. Bladder cancer subtypes defined by genomic alterations. Scand. J. Urol. Nephrol. 218, S116–S130 (2008).

    Article  Google Scholar 

  41. 41

    Lindgren, D. et al. Integrated genomic and gene expression profiling identifies two major genomic circuits in urothelial carcinoma. PLoS ONE 7, e38863 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. 42

    Nishiyama, N. et al. Copy number alterations in urothelial carcinomas: their clinicopathological significance and correlation with DNA methylation alterations. Carcinogenesis 32, 462–469 (2011).

    Article  CAS  PubMed  Google Scholar 

  43. 43

    Hurst, C. D., Platt, F. M., Taylor, C. F. & Knowles, M. A. Novel tumor subgroups of urothelial carcinoma of the bladder defined by integrated genomic analysis. Clin. Cancer Res. 18, 5865–5877 (2012). This study uses a genomic approach to bladder tumour subclassification. It describes potential subclasses of the 'gold-standard' stage and grade groups based on DNA copy number and mutation.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  44. 44

    Nord, H. et al. Focal amplifications are associated with high grade and recurrences in stage Ta bladder carcinoma. Int. J. Cancer 126, 1390–1402 (2010).

    CAS  PubMed  Google Scholar 

  45. 45

    Eriksson, P. et al. Detailed analysis of focal chromosome arm 1q and 6p amplifications in urothelial carcinoma reveals complex genomic events on 1q, and as a possible auxiliary target on 6p. PLoS ONE 8, e67222 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. 46

    Williams, S. V., Hurst, C. D. & Knowles, M. A. Oncogenic FGFR3 gene fusions in bladder cancer. Hum. Mol. Genet. 22, 795–803 (2012). This study was the first to identify oncogenic FGFR3 fusions in bladder cancer. Importantly, cell lines with FGFR3 fusions were those previously reported to show high sensitivity to FGFR-targeted agents.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. 47

    Nordentoft, I. et al. Mutational context and diverse clonal development in early and late bladder cancer. Cell Rep. 7, 1649–1663 (2014).

    Article  CAS  PubMed  Google Scholar 

  48. 48

    Lawrence, M. S. et al. Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature 499, 214–218 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. 49

    Roberts, S. A. et al. An APOBEC cytidine deaminase mutagenesis pattern is widespread in human cancers. Nature Genet. 45, 970–976 (2013).

    Article  CAS  PubMed  Google Scholar 

  50. 50

    Burns, M. B., Temiz, N. A. & Harris, R. S. Evidence for APOBEC3B mutagenesis in multiple human cancers. Nature Genet. 45, 977–983 (2013).

    Article  CAS  PubMed  Google Scholar 

  51. 51

    Cappellen, D. et al. Frequent activating mutations of FGFR3 in human bladder and cervix carcinomas. Nature Genet. 23, 18–20 (1999). This study was the first to describe FGFR3 mutations in bladder cancer.

    Article  CAS  PubMed  Google Scholar 

  52. 52

    Zieger, K. et al. Role of activating fibroblast growth factor receptor 3 mutations in the development of bladder tumors. Clin. Cancer Res. 11, 7709–7719 (2005).

    Article  CAS  PubMed  Google Scholar 

  53. 53

    Hernandez, S. et al. Prospective study of FGFR3 mutations as a prognostic factor in nonmuscle invasive urothelial bladder carcinomas. J. Clin. Oncol. 24, 3664–3671 (2006).

    Article  CAS  PubMed  Google Scholar 

  54. 54

    Kompier, L. C. et al. The development of multiple bladder tumour recurrences in relation to the FGFR3 mutation status of the primary tumour. J. Pathol. 218, 104–112 (2009).

    Article  CAS  PubMed  Google Scholar 

  55. 55

    Barbisan, F. et al. Strong immunohistochemical expression of fibroblast growth factor receptor 3, superficial staining pattern of cytokeratin 20, and low proliferative activity define those papillary urothelial neoplasms of low malignant potential that do not recur. Cancer 112, 636–644 (2008).

    Article  PubMed  Google Scholar 

  56. 56

    Burger, M. et al. Prediction of progression of non-muscle-invasive bladder cancer by WHO 1973 and 2004 grading and by FGFR3 mutation status: a prospective study. Eur. Urol. 54, 835–844 (2007).

    Article  PubMed  Google Scholar 

  57. 57

    van Rhijn, B. W. et al. The FGFR3 mutation is related to favorable pT1 bladder cancer. J. Urol. 187, 310–314 (2012).

    Article  CAS  PubMed  Google Scholar 

  58. 58

    Billerey, C. et al. Frequent FGFR3 mutations in papillary non-invasive bladder (pTa) tumors. Am. J. Pathol. 158, 1955–1959 (2001).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. 59

    Kimura, T. et al. The incidence of thanatophoric dysplasia mutations in FGFR3 gene is higher in low-grade or superficial bladder carcinomas. Cancer 92, 2555–2561 (2001).

    Article  CAS  PubMed  Google Scholar 

  60. 60

    Tomlinson, D. C., Baldo, O., Harnden, P. & Knowles, M. A. FGFR3 protein expression and its relationship to mutation status and prognostic variables in bladder cancer. J. Pathol. 213, 91–98 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  61. 61

    di Martino, E., L' Hôte, C. G., Kennedy, W., Tomlinson, D. C. & Knowles, M. A. Mutant fibroblast growth factor receptor 3 induces intracellular signaling and cellular transformation in a cell type- and mutation-specific manner. Oncogene 28, 4306–4316 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. 62

    Kiemeney, L. A. et al. A sequence variant at 4p16.3 confers susceptibility to urinary bladder cancer. Nature Genet. 42, 425–419 (2010).

    Article  CAS  Google Scholar 

  63. 63

    Tomlinson, D. C., L'Hôte, C. G., Kennedy, W., Pitt, E. & Knowles, M. A. Alternative splicing of fibroblast growth factor receptor 3 produces a secreted isoform that inhibits fibroblast growth factor-induced proliferation and is repressed in urothelial carcinoma cell lines. Cancer Res. 65, 10441–10449 (2005).

    Article  CAS  PubMed  Google Scholar 

  64. 64

    Zhang, X. et al. Receptor specificity of the fibroblast growth factor family. The complete mammalian FGF family. J. Biol. Chem. 281, 15694–15700 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  65. 65

    Catto, J. W. et al. Distinct microRNA alterations characterize high- and low-grade bladder cancer. Cancer Res. 69, 8472–8481 (2009). This study describes widespread alterations in microRNA expression in bladder cancer. A particularly interesting finding in low-grade tumours is loss of miR-99a and miR-100, which leads to upregulation of FGFR3.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  66. 66

    Tomlinson, D. C., Lamont, F. R., Shnyder, S. D. & Knowles, M. A. Fibroblast growth factor receptor 1 promotes proliferation and survival via activation of the mitogen-activated protein kinase pathway in bladder cancer. Cancer Res. 69, 4613–4620 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  67. 67

    Tomlinson, D. C. & Knowles, M. A. Altered splicing of FGFR1 is associated with high tumor grade and stage and leads to increased sensitivity to FGF1 in bladder cancer. Am. J. Pathol. 177, 2379–2386 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  68. 68

    Tomlinson, D. C., Baxter, E. W., Loadman, P. M., Hull, M. A. & Knowles, M. A. FGFR1-induced epithelial to mesenchymal transition through MAPK/PLCγ/COX-2-mediated mechanisms. PLoS ONE 7, e38972 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  69. 69

    Cheng, T. et al. Fibroblast growth factor receptors-1 and -3 play distinct roles in the regulation of bladder cancer growth and metastasis: implications for therapeutic targeting. PLoS ONE 8, e57284 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  70. 70

    Kassouf, W. et al. Distinctive expression pattern of ErbB family receptors signifies an aggressive variant of bladder cancer. J. Urol. 179, 353–358 (2008).

    Article  PubMed  Google Scholar 

  71. 71

    Jimenez, R. E. et al. Her-2/neu overexpression in muscle-invasive urothelial carcinoma of the bladder: prognostic significance and comparative analysis in primary and metastatic tumors. Clin. Cancer Res. 7, 2440–2447 (2001).

    CAS  PubMed  Google Scholar 

  72. 72

    Kruger, S. et al. HER2 overexpression in muscle-invasive urothelial carcinoma of the bladder: prognostic implications. Int. J. Cancer 102, 514–518 (2002).

    Article  CAS  PubMed  Google Scholar 

  73. 73

    Forster, J. A., Paul, A. B., Harnden, P. & Knowles, M. A. Expression of NRG1 and its receptors in human bladder cancer. Br. J. Cancer 104, 1135–1143 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  74. 74

    Fleischmann, A., Rotzer, D., Seiler, R., Studer, U. E. & Thalmann, G. N. Her2 amplification is significantly more frequent in lymph node metastases from urothelial bladder cancer than in the primary tumours. Eur. Urol. 60, 350–357 (2011).

    Article  CAS  PubMed  Google Scholar 

  75. 75

    Cheng, H. L. et al. Co-expression of RON and MET is a prognostic indicator for patients with transitional-cell carcinoma of the bladder. Br. J. Cancer 92, 1906–1914 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  76. 76

    Cheng, H. L. et al. Overexpression of c-met as a prognostic indicator for transitional cell carcinoma of the urinary bladder: a comparison with p53 nuclear accumulation. J. Clin. Oncol. 20, 1544–1550 (2002).

    Article  CAS  PubMed  Google Scholar 

  77. 77

    Sanchez-Carbayo, M., Socci, N. D., Lozano, J. J., Haab, B. B. & Cordon-Cardo, C. Profiling bladder cancer using targeted antibody arrays. Am. J. Pathol. 168, 93–103 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  78. 78

    Juanpere, N. et al. Mutations in FGFR3 and PIK3CA, singly or combined with RAS and AKT1, are associated with AKT but not with MAPK pathway activation in urothelial bladder cancer. Hum. Pathol. 43, 1573–1582 (2012).

    Article  CAS  PubMed  Google Scholar 

  79. 79

    López-Knowles, E. et al. PIK3CA mutations are an early genetic alteration associated with FGFR3 mutations in superficial papillary bladder tumors. Cancer Res. 66, 7401–7404 (2006).

    Article  PubMed  Google Scholar 

  80. 80

    Kompier, L. C. et al. FGFR3, HRAS, KRAS, NRAS and PIK3CA mutations in bladder cancer and their potential as biomarkers for surveillance and therapy. PLoS ONE 5, e13821 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  81. 81

    Zhao, L. & Vogt, P. K. Helical domain and kinase domain mutations in p110α of phosphatidylinositol 3-kinase induce gain of function by different mechanisms. Proc. Natl Acad. Sci. USA 105, 2652–2657 (2008).

    Article  PubMed  Google Scholar 

  82. 82

    Ross, R. L., Askham, J. M. & Knowles, M. A. PIK3CA mutation spectrum in urothelial carcinoma reflects cell context-dependent signaling and phenotypic outputs. Oncogene 32, 768–776 (2012).

    Article  CAS  PubMed  Google Scholar 

  83. 83

    Aveyard, J. S., Skilleter, A., Habuchi, T. & Knowles, M. A. Somatic mutation of PTEN in bladder carcinoma. Br. J. Cancer 80, 904–908 (1999).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  84. 84

    Cappellen, D., Gil Diez de Medina, S., Chopin, D., Thiery, J. P. & Radvanyi, F. Frequent loss of heterozygosity on chromosome 10q in muscle-invasive. Oncogene 14, 3059–3066 (1997).

    Article  CAS  PubMed  Google Scholar 

  85. 85

    Kagan, J. et al. Cluster of allele losses within a 2.5 cM region of chromosome 10 in high-grade invasive bladder cancer. Oncogene 16, 909–913 (1998).

    Article  CAS  PubMed  Google Scholar 

  86. 86

    Puzio-Kuter, A. M. et al. Inactivation of p53 and Pten promotes invasive bladder cancer. Genes Dev. 23, 675–680 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  87. 87

    Gildea, J. J. et al. PTEN can inhibit in vitro organotypic and in vivo orthotopic invasion of human bladder cancer cells even in the absence of its lipid phosphatase activity. Oncogene 23, 6788–6797 (2004).

    Article  CAS  PubMed  Google Scholar 

  88. 88

    Jebar, A. H. et al. FGFR3 and Ras gene mutations are mutually exclusive genetic events in urothelial cell carcinoma. Oncogene 24, 5218–5225 (2005).

    Article  CAS  PubMed  Google Scholar 

  89. 89

    Karlou, M. et al. Activation of extracellular regulated kinases (ERK1/2) predicts poor prognosis in urothelial bladder carcinoma and is not associated with B-Raf gene mutations. Pathol. 41, 327–334 (2009).

    Article  CAS  Google Scholar 

  90. 90

    Swiatkowski, S. et al. Activities of MAP-kinase pathways in normal uroepithelial cells and urothelial carcinoma cell lines. Exp. Cell Res. 282, 48–57 (2003).

    Article  CAS  PubMed  Google Scholar 

  91. 91

    Kastritis, E. et al. Somatic mutations of adenomatous polyposis coli gene and nuclear β-catenin accumulation have prognostic significance in invasive urothelial carcinomas: evidence for Wnt pathway implication. Int. J. Cancer 124, 103–108 (2009).

    Article  CAS  PubMed  Google Scholar 

  92. 92

    Zhu, X., Kanai, Y., Saito, A., Kondo, Y. & Hirohashi, S. Aberrant expression of β-catenin and mutation of exon 3 of the β-catenin gene in renal and urothelial carcinomas. Pathol. Int. 50, 945–952 (2000).

    Article  CAS  PubMed  Google Scholar 

  93. 93

    Kashibuchi, K. et al. The prognostic value of E-cadherin, α-, β- and γ-catenin in bladder cancer patients who underwent radical cystectomy. Int. J. Urol. 14, 789–794 (2007).

    Article  CAS  PubMed  Google Scholar 

  94. 94

    Baumgart, E. et al. Identification and prognostic significance of an epithelial-mesenchymal transition expression profile in human bladder tumors. Clin. Cancer Res. 13, 1685–1694 (2007).

    Article  CAS  PubMed  Google Scholar 

  95. 95

    Marsit, C. J. et al. Epigenetic inactivation of SFRP genes and TP53 alteration act jointly as markers of invasive bladder cancer. Cancer Res. 65, 7081–7085 (2005).

    Article  CAS  PubMed  Google Scholar 

  96. 96

    Urakami, S. et al. Epigenetic inactivation of Wnt inhibitory factor-1 plays an important role in bladder cancer through aberrant canonical Wnt/β-catenin signaling pathway. Clin. Cancer Res. 12, 383–391 (2006).

    Article  CAS  PubMed  Google Scholar 

  97. 97

    Ahmad, I. et al. β-catenin activation synergizes with PTEN loss to cause bladder cancer formation. Oncogene 30, 178–189 (2011).

    Article  CAS  PubMed  Google Scholar 

  98. 98

    Lin, C. et al. Constitutive β-catenin activation induces male-specific tumorigenesis in the bladder urothelium. Cancer Res. 73, 5914–5925 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  99. 99

    Ahmad, I. et al. Ras mutation cooperates with β-catenin activation to drive bladder tumourigenesis. Cell Death Dis. 2, e124 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  100. 100

    Shin, K. et al. Hedgehog/Wnt feedback supports regenerative proliferation of epithelial stem cells in bladder. Nature 472, 110–114 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  101. 101

    Shin, K. et al. Cellular origin of bladder neoplasia and tissue dynamics of its progression to invasive carcinoma. Nature Cell Biol. 16, 469–478 (2014).

    Article  CAS  PubMed  Google Scholar 

  102. 102

    Shin, K. et al. Hedgehog signaling restrains bladder cancer progression by eliciting stromal production of urothelial differentiation factors. Cancer Cell 26, 521–533 (2014). Using the BBN-induced invasive bladder cancer mouse model, this study provides evidence for a tumour suppressor role for hedgehog signalling from urothelium to stroma that in turn influences urothelial cell differentiation. BMP4 and BMP5 were among the SHH-induced stromal factors that stimulated urothelial differentiation. Importantly, pharmacological activation of the BMP pathway was shown to block tumour progression in this system.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  103. 103

    Thievessen, I., Wolter, M., Prior, A., Seifert, H. H. & Schulz, W. A. Hedgehog signaling in normal urothelial cells and in urothelial carcinoma cell lines. J. Cell. Physiol. 203, 372–377 (2005).

    Article  CAS  PubMed  Google Scholar 

  104. 104

    Mitra, A. P., Hansel, D. E. & Cote, R. J. Prognostic value of cell-cycle regulation biomarkers in bladder cancer. Semin. Oncol. 39, 524–533 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  105. 105

    Hurst, C. D., Tomlinson, D. C., Williams, S. V., Platt, F. M. & Knowles, M. A. Inactivation of the Rb pathway and overexpression of both isoforms of E2F3 are obligate events in bladder tumours with 6p22 amplification. Oncogene 27, 2716–2727 (2008).

    Article  CAS  PubMed  Google Scholar 

  106. 106

    Proctor, A., Coombs, L., Cairns, J. & Knowles, M. Amplification at chromosome 11q13 in transitional cell tumours of the bladder. Oncogene 6, 789–795 (1991).

    CAS  PubMed  Google Scholar 

  107. 107

    Lopez-Beltran, A. et al. Prognostic factors in survival of patients with stage Ta and T1 bladder urothelial tumors: the role of G1–S modulators (p53, 21Waf1, 27Kip1, cyclin D1, and cyclin D3), proliferation index, and clinicopathologic parameters. Am. J. Clin. Pathol. 122, 444–452 (2004).

    Article  PubMed  Google Scholar 

  108. 108

    Kim, W. J. & Kim, Y. J. Epigenetic biomarkers in urothelial bladder cancer. Expert Rev. Mol. Diagn. 9, 259–269 (2009).

    Article  CAS  PubMed  Google Scholar 

  109. 109

    Dudziec, E., Goepel, J. R. & Catto, J. W. Global epigenetic profiling in bladder cancer. Epigenomics 3, 35–45 (2011).

    Article  CAS  PubMed  Google Scholar 

  110. 110

    Sanchez-Carbayo, M. Hypermethylation in bladder cancer: biological pathways and translational applications. Tumour Biol. 33, 347–361 (2012). This is a comprehensive review of the extensive literature on DNA hypermethylation in bladder cancer.

    Article  CAS  PubMed  Google Scholar 

  111. 111

    Hoffman, A. M. & Cairns, P. Epigenetics of kidney cancer and bladder cancer. Epigenomics 3, 19–34 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  112. 112

    Kandimalla, R. et al. Genome-wide analysis of CpG island methylation in bladder cancer identified TBX2, TBX3, GATA2, and ZIC4 as pTa-specific prognostic markers. Eur. Urol. 61, 1245–1256 (2012).

    Article  CAS  PubMed  Google Scholar 

  113. 113

    Marsit, C. J. et al. Identification of methylated genes associated with aggressive bladder cancer. PLoS ONE 5, e12334 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  114. 114

    Reinert, T. et al. Comprehensive genome methylation analysis in bladder cancer: identification and validation of novel methylated genes and application of these as urinary tumor markers. Clin. Cancer Res. 17, 5582–5592 (2011).

    Article  CAS  PubMed  Google Scholar 

  115. 115

    Wolff, E. M. et al. Unique DNA methylation patterns distinguish noninvasive and invasive urothelial cancers and establish an epigenetic field defect in premalignant tissue. Cancer Res. 70, 8169–8178 (2010). This genome-wide methylation study of normal urothelium and tumours of all grades and stages found distinct patterns of hypomethylation in NMIBCs and widespread promoter hypermethylation in invasive tumours.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  116. 116

    Lauss, M. et al. DNA methylation analyses of urothelial carcinoma reveal distinct epigenetic subtypes and an association between gene copy number and methylation status. Epigenetics 7, 858–867 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  117. 117

    Hovestadt, V. et al. Decoding the regulatory landscape of medulloblastoma using DNA methylation sequencing. Nature 510, 537–541 (2014).

    Article  CAS  PubMed  Google Scholar 

  118. 118

    Dudziec, E., Gogol-Doring, A., Cookson, V., Chen, W. & Catto, J. Integrated epigenome profiling of repressive histone modifications, DNA methylation and gene expression in normal and malignant urothelial cells. PLoS ONE 7, e32750 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  119. 119

    Vallot, C. et al. A novel epigenetic phenotype associated with the most aggressive pathway of bladder tumor progression. J. Natl Cancer Inst. 103, 47–60 (2011).

    Article  CAS  PubMed  Google Scholar 

  120. 120

    Wild, P. J. et al. Detection of urothelial bladder cancer cells in voided urine can be improved by a combination of cytology and standardized microsatellite analysis. Cancer Epidemiol. Biomarkers Prev. 18, 1798–1806 (2009).

    Article  CAS  PubMed  Google Scholar 

  121. 121

    Zuiverloon, T. C. et al. Combinations of urinary biomarkers for surveillance of patients with incident nonmuscle invasive bladder cancer: the European FP7 UROMOL project. J. Urol. 189, 1945–1951 (2012). This study assesses combinations of urine biomarkers for detection and surveillance of bladder cancer in a large group of patients.

    Article  CAS  PubMed  Google Scholar 

  122. 122

    Zuiverloon, T. C. et al. Optimization of nonmuscle invasive bladder cancer recurrence detection using a urine based FGFR3 mutation assay. J. Urol. 186, 707–712 (2011).

    Article  CAS  PubMed  Google Scholar 

  123. 123

    Chung, W. et al. Detection of bladder cancer using novel DNA methylation biomarkers in urine sediments. Cancer Epidemiol. Biomarkers Prev. 20, 1483–1491 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  124. 124

    Zuiverloon, T. C. et al. A methylation assay for the detection of non-muscle-invasive bladder cancer (NMIBC) recurrences in voided urine. BJU Int. 109, 941–948 (2012).

    Article  CAS  PubMed  Google Scholar 

  125. 125

    Renard, I. et al. Identification and validation of the methylated TWIST1 and NID2 genes through real-time methylation-specific polymerase chain reaction assays for the noninvasive detection of primary bladder cancer in urine samples. Eur. Urol. 58, 96–104 (2010).

    Article  CAS  PubMed  Google Scholar 

  126. 126

    Kandimalla, R. et al. A 3-plex methylation assay combined with the FGFR3 mutation assay sensitively detects recurrent bladder cancer in voided urine. Clin. Cancer Res. 19, 4760–4769 (2013).

    Article  CAS  PubMed  Google Scholar 

  127. 127

    Allory, Y. et al. Telomerase reverse transcriptase promoter mutations in bladder cancer: high frequency across stages, detection in urine, and lack of association with outcome. Eur. Urol. 65, 360–366 (2013). This study identifies TERT promoter mutations at high frequency in bladder tumours of all grades and stages, and reports their use as a urine biomarker. These mutations are the most common event identified in bladder cancer so far.

    Article  CAS  PubMed  Google Scholar 

  128. 128

    Hurst, C. D., Platt, F. M. & Knowles, M. A. Comprehensive mutation analysis of the TERT promoter in bladder cancer and detection of mutations in voided urine. Eur. Urol. 65, 367–369 (2013).

    Article  CAS  PubMed  Google Scholar 

  129. 129

    Apodaca, G. The uroepithelium: not just a passive barrier. Traffic 5, 117–128 (2004).

    Article  CAS  PubMed  Google Scholar 

  130. 130

    Gaisa, N. T. et al. The human urothelium consists of multiple clonal units, each maintained by a stem cell. J. Pathol. 225, 163–171 (2011).

    Article  CAS  PubMed  Google Scholar 

  131. 131

    Ho, P. L., Kurtova, A. & Chan, K. S. Normal and neoplastic urothelial stem cells: getting to the root of the problem. Nature Rev. Urol. 9, 583–594 (2012).

    Article  CAS  Google Scholar 

  132. 132

    van der Horst, G., Bos, L. & van der Pluijm, G. Epithelial plasticity, cancer stem cells, and the tumor-supportive stroma in bladder carcinoma. Mol. Cancer Res. 10, 995–1009 (2012).

    Article  CAS  PubMed  Google Scholar 

  133. 133

    He, X. et al. Differentiation of a highly tumorigenic basal cell compartment in urothelial carcinoma. Stem Cells 27, 1487–1495 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  134. 134

    Chan, K. S. et al. Identification, molecular characterization, clinical prognosis, and therapeutic targeting of human bladder tumor-initiating cells. Proc. Natl Acad. Sci. USA 106, 14016–14021 (2009).

    Article  PubMed  Google Scholar 

  135. 135

    Kurtova, A. et al. Blocking PGE2-induced tumor repopulation abrogates bladder cancer chemoresistance. Nature (2014).

  136. 136

    Volkmer, J. P. et al. Three differentiation states risk-stratify bladder cancer into distinct subtypes. Proc. Natl Acad. Sci. USA 109, 2078–2083 (2012). This study identifies tumour-initiating cells from bladder tumours of different grades and stages. A key finding is that tumour-initiating cells from aggressive tumours showed a basal phenotype, whereas those from non-invasive tumours had features of more differentiated cells, suggesting a non-basal derivation.

    Article  PubMed  Google Scholar 

  137. 137

    Van Batavia, J. et al. Bladder cancers arise from distinct urothelial sub-populations. Nature Cell Biol. 16, 982–991 (2014).

    Article  CAS  PubMed  Google Scholar 

  138. 138

    Junker, K. et al. Analysis of genetic alterations in normal bladder urothelium. Urology 62, 1134–1138 (2003).

    Article  PubMed  Google Scholar 

  139. 139

    Obermann, E. C. et al. Fluorescence in situ hybridization detects frequent chromosome 9 deletions and aneuploidy in histologically normal urothelium of bladder cancer patients. Oncol. Rep. 11, 745–751 (2004).

    CAS  PubMed  Google Scholar 

  140. 140

    Chow, N. H. et al. Papillary urothelial hyperplasia is a clonal precursor to papillary transitional cell bladder cancer. Int. J. Cancer 89, 514–518 (2000).

    Article  CAS  PubMed  Google Scholar 

  141. 141

    Hartmann, A. et al. Frequent genetic alterations in simple urothelial hyperplasias of the bladder in patients with papillary urothelial carcinoma. Am. J. Pathol. 154, 721–727 (1999).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  142. 142

    van Oers, J. M. M. et al. Chromosome 9 deletions are more frequent than FGFR3 mutations in flat urothelial hyperplasias of the bladder. Int. J. Cancer 119, 1212–1215 (2006).

    Article  CAS  PubMed  Google Scholar 

  143. 143

    Hartmann, A. et al. Occurrence of chromosome 9 and p53 alterations in multifocal dysplasia and carcinoma in situ of human urinary bladder. Cancer Res. 62, 809–818 (2002).

    CAS  PubMed  Google Scholar 

  144. 144

    Spruck, C. H. 3rd. et al. Two molecular pathways to transitional cell carcinoma of the bladder. Cancer Res. 54, 784–788 (1994).

    CAS  PubMed  Google Scholar 

  145. 145

    Hopman, A. H. et al. Identification of chromosome 9 alterations and p53 accumulation in isolated carcinoma in situ of the urinary bladder versus carcinoma in situ associated with carcinoma. Am. J. Pathol. 161, 1119–1125 (2002).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  146. 146

    Rosin, M. P., Cairns, P., Epstein, J. I., Schoenberg, M. P. & Sidransky, D. Partial allelotype of carcinoma in situ of the human bladder. Cancer Res. 55, 5213–5216 (1995).

    CAS  PubMed  Google Scholar 

  147. 147

    Zieger, K., Marcussen, N., Borre, M., Ørntoft, T. F. & Dyrskjøt, L. Consistent genomic alterations in carcinoma in situ of the urinary bladder confirm the presence of two major pathways in bladder cancer development. Int. J. Cancer 125, 2095–2103 (2009).

    Article  CAS  PubMed  Google Scholar 

  148. 148

    Czerniak, B. et al. Genetic modeling of human urinary bladder carcinogenesis. Genes Chromosomes Cancer 27, 392–402 (2000).

    Article  CAS  PubMed  Google Scholar 

  149. 149

    Gallucci, M. et al. Status of the p53, p16, RB1, and HER-2 genes and chromosomes 3, 7, 9, and 17 in advanced bladder cancer: correlation with adjacent mucosa and pathological parameters. J. Clin. Pathol. 58, 367–371 (2005).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  150. 150

    Lee, S. et al. Forerunner genes contiguous to RB1 contribute to the development of in situ neoplasia. Proc. Natl Acad. Sci. USA 104, 13732–13737 (2007).

    Article  CAS  PubMed  Google Scholar 

  151. 151

    Stoehr, R. et al. Histologic-genetic mapping by allele-specific PCR reveals intraurothelial spread of p53 mutant tumor clones. Lab Invest. 82, 1553–1561 (2002).

    Article  CAS  PubMed  Google Scholar 

  152. 152

    Majewski, T. et al. Understanding the development of human bladder cancer by using a whole-organ genomic mapping strategy. Lab Invest. 88, 694–721 (2008).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  153. 153

    Hafner, C. et al. Evidence for oligoclonality and tumor spread by intraluminal seeding in multifocal urothelial carcinomas of the upper and lower urinary tract. Oncogene 20, 4910–4915 (2001).

    Article  CAS  PubMed  Google Scholar 

  154. 154

    Cheng, L. et al. Precise microdissection of human bladder carcinomas reveals divergent tumor subclones in the same tumor. Cancer 94, 104–110 (2002).

    Article  PubMed  Google Scholar 

  155. 155

    Habuchi, T. Origin of multifocal carcinomas of the bladder and upper urinary tract: molecular analysis and clinical implications. Int. J. Urol. 12, 709–716 (2005).

    Article  CAS  PubMed  Google Scholar 

  156. 156

    Takahashi, T. et al. Clonal and chronological genetic analysis of multifocal cancers of the bladder and upper urinary tract. Cancer Res. 58, 5835–5841 (1998).

    CAS  PubMed  Google Scholar 

  157. 157

    Kawanishi, H. et al. High throughput comparative genomic hybridization array analysis of multifocal urothelial cancers. Cancer Sci. 97, 746–752 (2006).

    Article  CAS  PubMed  Google Scholar 

  158. 158

    Kawanishi, H. et al. Genetic analysis of multifocal superficial urothelial cancers by array-based comparative genomic hybridisation. Br. J. Cancer 97, 260–266 (2007).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  159. 159

    van Tilborg, A. A. et al. Molecular evolution of multiple recurrent cancers of the bladder. Hum. Mol. Genet. 9, 2973–2980 (2000).

    Article  CAS  PubMed  Google Scholar 

  160. 160

    Letouzé, E., Allory, Y., Bollet, M. A., Radvanyi, F. & Guyon, F. Analysis of the copy number profiles of several tumor samples from the same patient reveals the successive steps in tumorigenesis. Genome Biol. 11, R76 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  161. 161

    Borhan, A. et al. Grade progression and regression in recurrent urothelial cancer. J. Urol. 169, 2106–2109 (2003).

    Article  CAS  PubMed  Google Scholar 

  162. 162

    Lamouille, S., Subramanyam, D., Blelloch, R. & Derynck, R. Regulation of epithelial–mesenchymal and mesenchymal–epithelial transitions by microRNAs. Curr. Opin. Cell Biol. 25, 200–207 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  163. 163

    Wiklund, E. D. et al. Coordinated epigenetic repression of the miR-200 family and miR-205 in invasive bladder cancer. Int. J. Cancer 128, 1327–1334 (2011).

    Article  CAS  PubMed  Google Scholar 

  164. 164

    Adam, L. et al. miR-200 expression regulates epithelial-to-mesenchymal transition in bladder cancer cells and reverses resistance to epidermal growth factor receptor therapy. Clin. Cancer Res. 15, 5060–5072 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  165. 165

    De Medina, S. G. et al. Relationship between E-cadherin and fibroblast growth factor receptor 2b expression in bladder carcinomas. Oncogene 18, 5722–5726 (1999).

    Article  CAS  PubMed  Google Scholar 

  166. 166

    Chaffer, C. L. et al. Mesenchymal-to-epithelial transition facilitates bladder cancer metastasis: role of fibroblast growth factor receptor-2. Cancer Res. 66, 11271–11278 (2006).

    Article  CAS  PubMed  Google Scholar 

  167. 167

    Hu, H. et al. A novel role of Id-1 in regulation of epithelial-to-mesenchymal transition in bladder cancer. Urol. Oncol. 31, 1242–1253 (2012).

    Article  CAS  PubMed  Google Scholar 

  168. 168

    Luo, M. et al. Long non-coding RNA H19 increases bladder cancer metastasis by associating with EZH2 and inhibiting E-cadherin expression. Cancer Lett. 333, 213–221 (2013).

    Article  CAS  PubMed  Google Scholar 

  169. 169

    Fan, Y. et al. TGF-β -induced upregulation of malat1 promotes bladder cancer metastasis by associating with suz12. Clin. Cancer Res. 20, 1531–1541 (2014).

    Article  CAS  PubMed  Google Scholar 

  170. 170

    Karni-Schmidt, O. et al. Distinct expression profiles of p63 variants during urothelial development and bladder cancer progression. Am. J. Pathol. 178, 1350–1360 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  171. 171

    Tran, M. N. et al. The p63 protein isoform ΔNp63α inhibits epithelial-mesenchymal transition in human bladder cancer cells: role of miR-205. J. Biol. Chem. 288, 3275–3288 (2013).

    Article  CAS  PubMed  Google Scholar 

  172. 172

    Choi, W. et al. p63 expression defines a lethal subset of muscle-invasive bladder cancers. PLoS ONE 7, e30206 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  173. 173

    Damrauer, J. S. et al. Intrinsic subtypes of high-grade bladder cancer reflect the hallmarks of breast cancer biology. Proc. Natl Acad. Sci. USA 111, 3110–3115 (2014). This is one of three recent gene expression profiling studies of MIBC that define major basal and luminal expression subtypes. (See also references 37 and 194).

    Article  CAS  PubMed  Google Scholar 

  174. 174

    Griner, E. M. & Theodorescu, D. The faces and friends of RhoGDI2. Cancer Metastasis Rev. 31, 519–528 (2012).

    Article  CAS  PubMed  Google Scholar 

  175. 175

    Wu, Y. et al. Src phosphorylation of RhoGDI2 regulates its metastasis suppressor function. Proc. Natl Acad. Sci. 106, 5807–5812 (2009).

    Article  PubMed  Google Scholar 

  176. 176

    Fanning, P. et al. Elevated expression of pp60c–src in low grade human bladder carcinoma. Cancer Res. 52, 1457–1462 (1992).

    CAS  PubMed  Google Scholar 

  177. 177

    Qayyum, T. et al. The interrelationships between Src, Cav-1 and RhoGD12 in transitional cell carcinoma of the bladder. Br. J. Cancer 106, 1187–1195 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  178. 178

    Said, N., Sanchez-Carbayo, M., Smith, S. C. & Theodorescu, D. RhoGDI2 suppresses lung metastasis in mice by reducing tumor versican expression and macrophage infiltration. J. Clin. Invest. 122, 1503–1518 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  179. 179

    Said, N. & Theodorescu, D. RhoGDI2 suppresses bladder cancer metastasis via reduction of inflammation in the tumor microenvironment. Oncoimmunology 1, 1175–1177 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  180. 180

    Said, N., Frierson, H. F., Sanchez-Carbayo, M., Brekken, R. A. & Theodorescu, D. Loss of SPARC in bladder cancer enhances carcinogenesis and progression. J. Clin. Invest. 123, 751–766 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  181. 181

    Yuan, X. et al. ATF3 suppresses metastasis of bladder cancer by regulating gelsolin-mediated remodeling of the actin cytoskeleton. Cancer Res. 73, 3625–3637 (2013).

    Article  CAS  PubMed  Google Scholar 

  182. 182

    Smith, S. C. & Theodorescu, D. The Ral GTPase pathway in metastatic bladder cancer: key mediator and therapeutic target. Urol. Oncol. 27, 42–47 (2009).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  183. 183

    Saito, R. et al. Downregulation of Ral GTPase-activating protein promotes tumor invasion and metastasis of bladder cancer. Oncogene 32, 894–902 (2013).

    Article  CAS  PubMed  Google Scholar 

  184. 184

    Overdevest, J. B. et al. CD24 offers a therapeutic target for control of bladder cancer metastasis based on a requirement for lung colonization. Cancer Res. 71, 3802–3811 (2011).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  185. 185

    Overdevest, J. B. et al. CD24 expression is important in male urothelial tumorigenesis and metastasis in mice and is androgen regulated. Proc. Natl Acad. Sci. USA 109, E3588–E3596 (2012).

    Article  PubMed  Google Scholar 

  186. 186

    Blaveri, E. et al. Bladder cancer outcome and subtype classification by gene expression. Clin. Cancer Res. 11, 4044–4055 (2005).

    Article  CAS  PubMed  Google Scholar 

  187. 187

    Dyrskjot, L. et al. Analysis of molecular intra-patient variation and delineation of a prognostic 12-gene signature in non-muscle invasive bladder cancer; technology transfer from microarrays to PCR. Br. J. Cancer 107, 1392–1398 (2012).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  188. 188

    Dyrskjot, L. et al. Identifying distinct classes of bladder carcinoma using microarrays. Nature Genet. 33, 90–96 (2003).

    Article  CAS  PubMed  Google Scholar 

  189. 189

    Dyrskjot, L. et al. Gene expression signatures predict outcome in non-muscle-invasive bladder carcinoma: a multicenter validation study. Clin. Cancer Res. 13, 3545–3551 (2007).

    Article  CAS  PubMed  Google Scholar 

  190. 190

    Sanchez-Carbayo, M., Socci, N. D., Lozano, J., Saint, F. & Cordon-Cardo, C. Defining molecular profiles of poor outcome in patients with invasive bladder cancer using oligonucleotide microarrays. J. Clin. Oncol. 24, 778–789 (2006).

    Article  CAS  PubMed  Google Scholar 

  191. 191

    Kim, W.-J. et al. Predictive value of progression-related gene classifier in primary non-muscle invasive bladder cancer. Mol. Cancer 9, 3 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  192. 192

    Sjödahl, G. et al. A molecular taxonomy for urothelial carcinoma. Clin. Cancer Res. 18, 3377–3386 (2012). This study was the first to describe a molecular taxonomy for bladder cancer based on gene expression profiling of tumours of all stages and grades. Molecular subtypes traversing pathological classification and with distinct clinical outcomes are reported.

    Article  CAS  PubMed  Google Scholar 

  193. 193

    Sjödahl, G. et al. Toward a molecular pathologic classification of urothelial carcinoma. Am. J. Pathol. 183, 681–691 (2013).

    Article  CAS  PubMed  Google Scholar 

  194. 194

    Choi, W. et al. Identification of distinct basal and luminal subtypes of muscle-invasive bladder cancer with different sensitivities to frontline chemotherapy. Cancer Cell 25, 152–165 (2014). This is one of three recent gene expression profiling studies of MIBC that define major basal and luminal expression subtypes. See also references 37 and 173. This study defines two luminal subtypes, one of which contains tumours that show resistance to cisplatin-based therapy.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  195. 195

    Choi, W. et al. Intrinsic basal and luminal subtypes of muscle-invasive bladder cancer. Nature Rev. Urol. 11, 400–410 (2014). This is a useful review of recent expression profiling studies that identifies overlap of clusters defined by different groups and highlights specific features of the subtypes defined.

    Article  CAS  Google Scholar 

  196. 196

    Hoadley, K. A. et al. Multiplatform analysis of 12 cancer types reveals molecular classification within and across tissues of origin. Cell 158, 929–944 (2014). This excellent pan-cancer study includes data on MIBC from a TCGA study and shows alignment of bladder tumours with squamous features with aggressive breast and ovarian cancers.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  197. 197

    Aine, M., Liedberg, F., Sjodahl, G. & Hoglund, M. Re: David J., McConkey, Woonyoung Choi, Colin, P. N. Dinney. New insights into subtypes of invasive bladder cancer: considerations of the clinician. Eur. Urol. (2014).

  198. 198

    McConkey, D., Choi, W. & Dinney, C. Reply to Mattias Aine, Fredrik Liedberg, Gottfrid Sjodahl & Mattias Hoglund's Letter to the Editor re: David, J. McConkey, Woonyoung Choi, Colin, P. N. Dinney. New insights into subtypes of invasive bladder cancer: considerations of the clinician. Eur. Urol. (2014).

  199. 199

    Rebouissou, S. et al. EGFR as a potential therapeutic target for a subset of muscle-invasive bladder cancers presenting a basal-like phenotype. Sci. Transl Med. 6, 244ra291 (2014).

    Article  CAS  Google Scholar 

  200. 200

    Wilhelm-Benartzi, C. S. et al. DNA methylation profiles delineate etiologic heterogeneity and clinically important subgroups of bladder cancer. Carcinogenesis 31, 1972–1976 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  201. 201

    Lindgren, D. et al. Combined gene expression and genomic profiling define two intrinsic molecular subtypes of urothelial carcinoma and gene signatures for molecular grading and outcome. Cancer Res. 70, 3463–3472 (2010).

    Article  CAS  PubMed  Google Scholar 

Download references


The authors thank Cancer Research UK (C6228/A5433; C6228/A12512; C37059/A11941) and Yorkshire Cancer Research (L346, L362, L367, L372, L376PA) for past and current funding of their work. They also thank M. Höglund and K. S. Chan for discussions during the writing of this Review.

Author information



Corresponding author

Correspondence to Margaret A. Knowles.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

PowerPoint slides


Tumour–Node–Metastasis system

(TNM system). A classification system used to describe the stage of a tumour. T describes the extent of local invasion; N describes whether the tumour has spread to local lymph nodes; and M describes distant metastatic spread.


Chromosomal shattering. A phenomenon whereby cancer cells acquire many clustered chromosomal rearrangements as a single catastrophic event during tumour development.

Non-homologous end-joining

An error-prone mechanism of DNA repair in which broken DNA ends are joined without the guidance of a large homologous template. Short homologous DNA sequences termed microhomologies are used to guide repair.

Replication-licensing complex

A complex of proteins that assemble at origins of DNA replication during late G1 phase of the cell cycle to ensure precise and timely DNA replication.

Homologous recombination

A high-fidelity mechanism of DNA repair that uses recombination with an intact homologous template to repair double-strand breaks.

Bacillus Calmette–Guerin

(BCG). A vaccine used to induce immunity to tuberculosis. It is also used to treat high-risk localized bladder cancer and carcinoma in situ. Instillation of BCG into the bladder induces a localized immune response that is able to eliminate cancer cells. For bladder cancer treatment, a course of 6 weekly treatments is usual.

Smoking pack-years

A measure of exposure to cigarette smoke that is calculated by multiplying the number of packs of cigarettes smoked per day by the number of years the person has smoked. For example, 1 pack-year is equal to smoking 20 cigarettes (1 pack) per day for 1 year or 40 cigarettes per day for half a year.


Less densely packed or 'open' chromatin that is often associated with active transcription.

Field change

A process by which molecular alterations are accumulated within a large tissue area in response to carcinogenic stimuli.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Knowles, M., Hurst, C. Molecular biology of bladder cancer: new insights into pathogenesis and clinical diversity. Nat Rev Cancer 15, 25–41 (2015).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing