Secondary Logo

Journal Logo


Genetic subtypes of invasive bladder cancer

McConkey, David J.; Choi, Woonyoung; Dinney, Colin P.N.

Author Information
doi: 10.1097/MOU.0000000000000200
  • Free



The clinical approaches that are used in the management of muscle-invasive bladder cancers have not changed significantly in decades. In the USA, frontline therapy continues to rely on definitive surgery (cystectomy) with or without perioperative cisplatin-based combination chemotherapy. In Europe, chemoradiation is more broadly employed with the goal of bladder preservation. However, it is currently impossible to prospectively identify candidates for bladder preservation with a high degree of certainty, and it is also impossible to predict who will benefit from cisplatin-based chemotherapy or chemoradiation. In addition, and unlike the situation with almost all other solid tumors, no biologically targeted agents have received approval for invasive bladder cancer treatment.

Fortunately, this disappointing reality appears about to change. Large-scale genomics projects completed over the last 18 months have yielded a wealth of new insights into the biology of invasive bladder cancer, and aggressive efforts are already underway to attempt to translate this information into better tools for personalized cancer management. In addition, a new class of drugs has emerged that can reverse cancer-induced immunosuppression, and they appear to be highly active in a subset of patients. The overall goal of this review is to provide a comprehensive overview of some of the more important recent genomic discoveries and discuss how they might be used to inform personalized approaches to early detection and cancer therapy.

Box 1
Box 1:
no caption available


Lessons from breast cancer

The idea that cancers are heterogeneous has been appreciated for years. However, methods to deeply measure this heterogeneity have only emerged recently with the development of technologies that enable characterization of DNA alterations and RNA expression across the whole genome. The introduction of cDNA microarrays enabled investigators to demonstrate that whole genome mRNA expression profiling could be used to objectively distinguish different hematological malignancies without prior knowledge of associated clinical information [1]. Other studies demonstrated that mRNA expression profiling could also be used to identify distinct molecular subsets of diffuse large cell lymphoma characterized by variable expression of differentiation-associated biomarkers [2]. In what has become a classic study, Perou et al.[3] used an in-house microarray containing 8102 probes to characterize gene expression in a cohort of 65 different breast cancers from 42 patients, extracting the genes that had the highest variation in expression across the tumors and performing unsupervised hierarchical clustering to visualize shared patterns of gene expression within the tumors. They discovered that the cancers formed at least four distinct ‘clusters’, and by identifying the top differentially expressed genes associated with each one, they recognized biomarkers that were expressed by normal human breast epithelial cells at different stages of differentiation [3]. Specifically, one of the clusters (‘basal-like’) was characterized by expression of keratin-5 (KRT5), an intermediate filament protein that is expressed specifically by the cells that line the base of the normal mammary epithelium in contact with the extracellular matrix (ECM), and low levels of markers associated with terminal differentiation [3]. Conversely, another cluster (‘luminal’) was characterized by high-level expression of differentiation-associated biomarkers (including estrogen receptor/ESR1 and its downstream targets) and low-level expression of the basal biomarkers. Tumors in a third cluster (‘HER2-enriched’) expressed very high levels of the epidermal growth factor receptor (EGFR) family member HER-2 (ERBB2) and other genes that are known to be located near ERBB2. The availability of matched specimens obtained before and after neoadjuvant chemotherapy and from primary tumors and matched lymph node metastases enabled the investigators to examine whether the tumors switched subtype based on when and where they were sampled. The results indicated that most of the time they did not, leading the authors to conclude that the subtypes they had identified were truly ‘intrinsic’ properties of the tumors. Since the initial discovery of the breast cancer intrinsic subtypes, another subtype (termed ‘claudin-low’) has been identified, which appears to correspond to a fraction of basal-like cancers that has undergone epithelial-to-mesenchymal transition (EMT) [4,5]. There is considerable (but not perfect) overlap between the basal-like subtype and the ‘triple negative’ subset of tumors identified clinically by immunohistochemistry with antibodies to estrogen receptor, progesterone receptor, and HER2/ERBB2 [6]. In addition, a gene expression-based molecular classifier known as the PAM50 was developed to enable more precise assignment of breast cancers to intrinsic subtypes without having to resort to whole genome mRNA expression profiling [6]. The PAM50 and the intrinsic breast cancer intrinsic subtypes in general have been validated in numerous independent studies, including the TCGA's first breast cancer project [7].

The origins of the intrinsic subtypes of breast cancer remain a topic of intense discussion. One possibility is that they arise from the presence of different DNA alterations (mutations, copy number variations, and so on) in the same normal cell-of-origin (i.e., the normal mammary epithelial stem cell). Alternatively, because differentiation markers define the intrinsic subtypes, it may be that they arise via transformation of different progenitors; for example, mammary epithelial cells at different stages of differentiation [8]. This model could also explain the observations made in diffuse large B-cell lymphomas [2] and other cancers that contain intrinsic subtypes. It is possible to use a technique known as ‘lineage tracing’ to identify the candidate cell(s) of origin in tumors in genetically engineered mouse models [9–11], but it is not clear that the same cells always give rise to the corresponding tumor subtypes in humans.

Knowledge of a breast cancer's intrinsic subtype membership has important implications for prognosis and for predicting therapeutic outcomes [6,12]. Basal-like and HER2-enriched tumors are intrinsically clinically aggressive, and patients with these cancers experience short disease-specific and overall survival in the absence of effective systemic therapy, but they are also sensitive to neoadjuvant chemotherapy [4]. In contrast, although adjuvant therapy with selective estrogen receptor modulators (SERMs) such as tamoxifen or raloxifene has substantial chemopreventative activity and reduces rates of cancer recurrence in women with luminal cancers, it has no impact in women with basal-like or HER2-enriched disease. However, neoadjuvant chemotherapy has much less impact in luminal breast cancers, and it has almost no activity in the luminal A subtype [4]. Finally, clinical benefit from HER2-targeting agents is largely restricted to women with HER2-enriched cancers, and there is the sense that inhibitors of poly(ADP-ribose) polymerase (PARP), which are very active in ovarian cancers that contain inactivating BRCA1 or BRCA2 mutations, may also be active in basal-like breast cancers, particularly if they are combined with other targeted agents [13].

Intrinsic basal and luminal subtypes of bladder cancer

The first studies that employed whole genome mRNA expression profiling to visualize genomic heterogeneity in bladder cancers were performed over a decade ago [14–16]. They demonstrated that low and high-grade cancers formed distinct clusters when they were analyzed by unsupervised methods and suggested that additional subtypes might also be present. Subsequent work integrating whole genome mRNA expression with DNA copy number and focused mutational analyses reinforced the idea that nonmuscle invasive and muscle-invasive cancers were very different from each other at the molecular level, and they implicated genomic instability as one of the major candidate causative factors of these differences [17]. Deeper investigation demonstrated that muscle-invasive cancers contained as many as five distinct molecular subtypes, including one (‘SCC-like’) that shared biomarkers with squamous cell carcinomas of the lung and head and neck and were associated with poor clinical outcomes [18▪▪]. Parallel studies concluded that muscle-invasive bladder cancers could be subdivided based on expression of differentiation stage-specific keratins (KRT5, KRT14, and KRT20), and patients whose tumors expressed basal keratins had the poorest clinical outcomes [19]. Similarly, patients whose tumors expressed high levels of the basal epithelial transcription factor ΔNp63α had shorter disease-specific and overall survival [20,21]. However, the biological significance and potential similarities among the subtypes identified in the different studies were unclear.

Three independent projects completed within the last 18 months provided new insights into the intrinsic subtypes of muscle-invasive bladder cancer [22▪▪,23▪,24▪▪]. Using whole genome mRNA expression profiling data and unsupervised hierarchical clustering, they reported the existence of between two [24▪▪] and four [22▪▪] intrinsic subtypes. What differed between these newer studies and the ones conducted previously was that they established that the bladder cancer subtypes were surprisingly similar to the intrinsic basal and luminal subtypes of breast cancer (Fig. 1) [22▪▪,23▪,24▪▪,25,26]. Like their breast cancer counterparts, basal bladder cancers were intrinsically aggressive; patients with basal bladder cancer who did not receive neoadjuvant chemotherapy tended to have advanced stage and/or metastatic disease at presentation, and they had shorter disease-specific and overall than did patients whose tumors belonged to the luminal subtype(s) [23▪,24▪▪,27▪▪]. Basal cancers were more common in women [23▪,24▪▪], and they were often enriched with squamous and/or sarcomatoid histopathological features [22▪▪,23▪]. Direct comparisons of the subtype calls made by the three groups using the TCGA's RNAseq data revealed remarkable overlap, and similar comparisons of subtype calls in other gene expression profiling datasets confirmed that the basal tumors corresponded to the lethal ‘SCC-like’ clusters identified previously by other groups [14,18▪▪]. Using a gene expression signature from breast cancer, one of the groups identified a ‘claudin-low’ subset of basal cancers [24▪▪,25,26], and subsequent analyses indicated that this claudin-low subtype corresponded well with the TCGA's ‘cluster IV’ basal tumors [22▪▪,25,26]. Luminal tumors also contained two subclusters that were distinguished by levels of stromal cell infiltration and expression of late cell-cycle biomarkers, similar to the differences between luminal A and luminal B breast cancers [18▪▪,22▪▪,23▪,25,26]. These two luminal subtypes corresponded well to two of the subtypes (‘infiltrated’ and ‘genomically unstable’) identified previously by a group from Lund [18▪▪,25]. The fact that there were two basal and two luminal subsets reconciled the different numbers of subtypes identified by the three groups (Fig. 1) [25,26].

Intrinsic basal and luminal subtypes of invasive bladder cancers. Top: Comparisons of subtype calls made using the MD Anderson[23▪], University of North Carolina [24▪▪], or TCGA [22▪▪] genomic classifiers on TCGA's RNAseq dataset. Reproduced with permission [25]. Bottom: Graphic depiction of the relationships among the University of North Carolina, MD Anderson, and TCGA subtypes and the intrinsic subtypes of breast cancer. Reproduced with permission [26].


Molecular characteristics of basal cancers

Aside from basal keratins (KRT5, KRT14), basal bladder cancers were enriched with bladder cancer stem cell biomarkers (including CD44) [23▪,28], ΔNp63α [23▪,29▪], MYC [23▪,24▪▪], and gene expression patterns were consistent with EMT [22▪▪,23▪,24▪▪]. The latter is a physiological process that plays critical roles in development and wound healing, which is also important for cancer metastasis [30]. It is characterized by loss of homotypic cell adhesion and polarity accompanied by increased invasion and migration [30]. Cells that have undergone EMT share biomarkers with cancer stem cells, and in fact enforced expression of EMT-driving transcription factors is sufficient to cause upregulation of these biomarkers and ‘stemness’ as measured in functional assays [31]. As discussed above, the TCGA's ‘cluster IV’ corresponded to a ‘claudin-low’ subset of basal cancers characterized by especially high levels of EMT biomarkers [24▪▪,25,26], consistent with the biological properties of the claudin-low subtype of breast cancer [4,5]. Using the ingenuity pathway analysis (IPA) upstream regulators tool, we searched for transcription factors that were predicted to be activated in basal cancers [23▪]. The results identified active signal transducer and activator of transcription-3 (STAT3), ΔNp63α, NF-κB, and HIF1 in the basal gene expression signature, and all of them had been implicated in the control of EMT and bladder cancer ‘stemness’ [32,33]. Basal bladder cancers were also enriched with EGFR expression and amplification [23▪,27▪▪]. To more directly determine their involvement in the basal gene expression signature, we used RNAi to knockdown expression of STAT3 or ΔNp63α in human bladder cancer cell lines and used whole genome mRNA expression profiling to identify the resulting changes in gene expression. The results confirmed that the STAT3 and ΔNp63α gene signatures we generated were enriched in basal cancers [23▪] (A. Ochoa et al., in preparation). Interestingly, ΔNp63α knockdown also resulted in downregulation of the HIF1 gene expression signature, suggesting that ΔNp63α may act upstream of HIF1 in basal cancers.

Molecular characteristics of luminal tumors

In luminal tumors, ingenuity upstream regulator analyses predicted activation of estrogen receptor (ESR) and TRIM24 [23▪], two transcription factors that also play central roles in luminal breast cancers. (Parallel work by another group also demonstrated that luminal bladder cancers were enriched with luminal breast cancer gene expression signatures [24▪▪].) In addition, the analyses predicted activation of peroxisome proliferator activator receptor (PPARG), a transcription factor that plays critical roles in adipocyte differentiation and fatty acid metabolism [34]. Consistent with these results, the TCGA concluded that PPARG was amplified in a significant fraction of bladder cancers [22▪▪], and PPARG was independently implicated in gene expression and proliferation of luminal tumors by another group [35▪]. Other studies demonstrated that PPARG agonists promote bladder cancer in rodents, and there is also concern that they may also promote bladder cancer in humans [36–38]. We created an active PPARG gene expression signature using human bladder cancer cells exposed to the PPARG agonist rosiglitazone, and we confirmed that this signature was enriched in luminal bladder cancers [23▪]. Therefore, PPARG activation definitely plays a central role in luminal bladder cancer biology, although a detailed understanding of its effects will require further investigation.

Molecular characteristics of ‘p53-like’ tumors

As mentioned previously, the three studies alluded to above [22▪▪,23▪,24▪▪] reported the existence of different numbers of subtypes in muscle-invasive bladder cancers (MIBCs), accounted for by the identification of additional ‘subsets’ with the basal and luminal intrinsic subtypes. One of these subsets, termed ‘cluster II’ by TCGA [22▪▪] and ‘p53-like’ by our group [23▪], consists primarily of luminal tumors characterized by high levels of infiltrating stromal cells, particularly cancer-associated fibroblasts (CAFs) [22▪▪,23▪]. Our decision to name them ‘p53-like’ came from the observation that they expressed gene expression signatures consistent with active wild-type p53 as determined by IPA analyses [23▪]. Direct comparisons of subtype calls also revealed that they were very similar to the Lund group's ‘infiltrated’ tumors [18▪▪]; ‘infiltrated’ also provides an excellent description of their biological properties. Consistent with the Lund group's previous findings [18▪▪], the p53-like and cluster II tumors were found to be enriched with ECM biomarkers [22▪▪,23▪], which are often expressed by CAFs, and they expressed low levels of genes that promote progression through the late phases of the cell cycle [22▪▪,23▪,25]. These biological features are shared by luminal A breast cancers, and when the MD Anderson one NN classifier was used to assign the TCGA's breast cancers to the bladder cancer intrinsic subtypes, the luminal A cancers (as determined by TCGA using the PAM50) were assigned to the p53-like subtype [25]. More recently, we have also found that they share microRNA expression patterns with luminal A breast cancers (A. Ochoa, in preparation). Importantly, the fractions of p53-like tumors that contained mutant p53 were very similar to those observed in the basal or luminal cancers [23▪], indicating that wild-type p53 itself was probably not responsible for the gene expression signatures that characterized the tumors.


As discussed previously, the intrinsic subtypes of breast cancer exhibit different sensitivities to conventional and targeted therapies. Similarly, the intrinsic MIBC subtypes contain features that may also influence their responses to conventional and targeted cancer therapies [39]. As is true in invasive breast cancer, neoadjuvant chemotherapy is considered standard-of-care for high-risk MIBC, and down-staging to no residual MIBC at cystectomy is predictive of excellent long-term clinical outcomes [40]. The FDA has decided to grant fast track approval to agents that produce pathological complete responses in the neoadjuvant setting in breast cancer [41], and adoption of a similar neoadjuvant strategy could greatly accelerate efforts to identify active therapies for MIBC as well [40]. In the following section, we will review some of the features of the intrinsic subtypes of MIBC that could inform efforts to develop clinical tests to predict the outcomes of conventional and targeted therapies.

Intrinsic subtypes and response to NAC

By comparing subtype membership with various clinical parameters in multiple retrospective cohorts, we observed that about half of the basal and half of the luminal MIBCs were down-staged to no residual muscle-invasive disease at cystectomy [23▪]. Given that basal cancers are intrinsically the most aggressive, these observations support the aggressive use of neoadjuvant chemotherapy in patients who have basal tumors, because it has the potential to dramatically (and beneficially) alter clinical outcomes. At first glance these observations may seem to contrast with another high-profile recent report, which demonstrated that chemotherapy causes recruitment and expansion of a basal-like stem cell subpopulation that repopulates tumors after therapy [42▪▪]. However, these observations are actually not at odds with one another; although the biomarkers expressed by basal tumors are shared with bladder cancer stem cells [28], the latter likely represent only a minor subpopulation of KRT5+KRT14+ cells within the bulk tumor that is probably fairly quiescent prior to chemotherapy, and the idea that it is recruited to repopulate the tumor after chemotherapy-induced debulking is consistent with what is known about what happens during epithelial wound healing [42▪▪]. On the contrary, a majority of the p53-like tumors were resistant [23▪]. As discussed previously, the p53-like tumors expressed gene expression signatures characteristic of active, wild-type p53 and low levels of late cell cycle biomarkers, and they appear to be relatively quiescent compared with the tumors in the other subtypes [18▪▪,23▪]. This quiescence probably contributes directly to chemoresistance, because proliferation is associated with sensitivity to apoptosis [43], and a similar link between quiescence and chemoresistance has been observed in luminal A breast cancers [4,44]. Strikingly, comparisons of matched tumors before and after treatment revealed that chemotherapy caused enrichment of the p53-like subtype and induced an active p53-like/quiescent gene expression signature in all tumors [18▪▪,23▪]. Depending on its duration, this quiescence could undermine later attempts to control residual disease with other cytotoxic therapies.

Candidate biological targets in basal cancers

Immune checkpoint inhibitors [45] are poised to become the first targeted agents to receive FDA approval for bladder cancer therapy. Neoadjuvant clinical trials employing the immune checkpoint inhibitor ipilumumab (a blocking anti-CTLA4 antibody) demonstrated immune modulation in muscle-invasive bladder cancer [46], but autoimmune side-effects limited enthusiasm for the approach in the absence of strong evidence of clinical activity. More recent efforts have focused on blocking PD1 or PDL1 [45,47], which have produced less autoimmune toxicity. A recently completed phase I clinical trial of the blocking anti-PDL1 antibody MPDL3280A in patients with metastatic bladder cancer revealed strong clinical activity, with 52% of patients displaying RECIST criteria responses, including complete responses in 7% of them [48▪▪]. Expression of PDL1 as measured by immunohistochemistry correlated with response, but the correlation was not strong, prompting interest in developing genomics-based and other molecular approaches to prospectively identify sensitive tumors. Interestingly, the claudin-low subset of basal bladder cancers (corresponding to TCGA's ‘cluster IV’) expresses particularly high levels of T and B lymphocyte and immune checkpoint biomarkers, whereas luminal tumors express very low levels of them (Fig. 2). Given their high-profile clinical activities, a large number of clinical trials employing PD1 or PDL1 inhibitors are either open or poised to open soon. It will be important to examine whether response to immune checkpoint blockade varies according to tumor intrinsic subtype membership, and in particular, whether the claudin-low/TCGA cluster IV basal tumors are highly sensitive to them.

Claudin-low (TCGA cluster IV) tumors are enriched with lymphocytes. The heat map displays relative expression of an immune signature described in Choiet al. [23▪] in the TCGA RNAseq dataset. Red indicates high relative expression, blue indicates low expression. Reproduced with permission [39].

In addition to immune checkpoint biomarkers, basal cancers are also enriched with high levels of the EGFR [23▪,27▪▪] and high-level EGFR amplification [27▪▪]. Indeed, strong evidence has been obtained from preclinical models supporting the idea that basal MIBCs may be particularly sensitive to EGFR inhibitors [27▪▪]. However, as noted earlier, EMT is a defining feature of basal cancers (and particularly the claudin-low tumors), and EMT causes resistance to EGFR inhibitors in bladder cancer cells [49,50]. The relative impact of this EMT on EGFR inhibitor sensitivity will need to be evaluated directly in clinical trials. It is possible that epigenetic agents (such as histone deacetylase inhibitors) can be used to reverse EMT and promote EGFR inhibitor sensitivity in these cancers [51].

Basal MIBCs are also enriched with gene expression signatures predictive of active STAT3 [39], and preclinical studies have directly implicated STAT3 in basal bladder cancer tumorigenesis [32]. Direct inhibitors of STAT3 are still in preclinical development, and these agents would be predicted to be most active in basal cancers. Basal cancers also appear to contain active HIF1 [23▪], and there are many clinically available targeted agents that can be used to inhibit the HIF1 pathway, including inhibitors of the proteasome [52], mammalian target of rapamycin [53], or vascular endothelial growth factor receptor-2 (VEGFR2) [54]. Although past trials employing single agents targeting VEGF or VEGFR2 in unselected patients yielded disappointing results, recently completed combination trials employing them with conventional chemotherapy produced more interesting clinical activity [55,56]. It may be possible to use genomic assessments of tumor intrinsic subtype membership to increase the impact of these targeted therapies.

Candidate targets in luminal cancers

As emphasized in the TCGA's marker paper [22▪▪], there are a number of attractive biological targets in bladder cancers [39], and many of them are enriched in luminal tumors (Fig. 3). Activating FGFR3 mutations and chromosomal translocations are present in over two-thirds of nonmuscle invasive bladder cancers and a significant fraction of luminal muscle-invasive tumors (Fig. 3) [39], and small molecule and blocking antibody FGFR inhibitors are now being evaluated in patients with these cancers. Activating mutations in ERBB2 and ERBB3 and ERBB2 gene amplification are also enriched in luminal cancers (Fig. 3) [39], and clinical trials employing pan-ERBB inhibitors are in the planning stage. The PI-3 kinase/AKT pathway is also altered in a majority of cancers [22▪▪], and in particular, PIK3CA-activating mutations are enriched in luminal cancers, just as they are luminal breast cancers [39]. There is good evidence that targeting this pathway will produce strong clinical responses in some patients [57].

Clinically actionable mutations in luminal MIBCs. The color bar at the top indicates tumor membership within each TCGA subtype; clusters I and II are luminal, and clusters III and IV are basal. In the color bars below, tumors identified in black contain mutations – inactivating in the case of STAG2, activating for all of the other genes. The results are based on TCGA mutation calls[22▪▪].


Several groups independently identified intrinsic subtypes of bladder cancer [18▪▪,22▪▪,23▪,24▪▪,25,26,27▪▪], and preliminary comparisons of subtype calls have revealed remarkable overlap among them in shared datasets [25]. A bladder cancer intrinsic subtypes consensus meeting was recently held in Madrid, and one of the projects inspired by the meeting was a formal, head-to-head comparison of the different gene expression-based methods to identify intrinsic subtypes to more directly determine the extent of overlap. Another product of the meeting was the design of a collaborative project to develop an immunohistochemical approach to distinguish basal and luminal tumors, akin to the methods currently used to subtype breast cancers. Both projects should yield preliminary results soon.

Most recent genomics efforts have focused on MIBCs, yet more patients present with nonmuscle invasive disease. As discussed previously, past studies concluded that intrinsic subtypes also exist in nonmuscle invasive bladder cancers [15,17,18▪▪], and several groups are currently performing studies to characterize them further and define their biological properties. These studies will also help us to better define the nonmuscle invasive precursor lesions that give rise to MIBCs; at present, we assume that nonmuscle invasive papillary tumors are the precursors for luminal MIBCs, because the latter are enriched with papillary features and genomic alterations that are consistent with them [22▪▪]. However, a corresponding candidate precursor lesion for basal tumors has not yet been identified.

Finally, how urothelial cancers with variant histology (micropapillary, small cell, pure squamous, and so on) and/or upper tract disease relate to the basal and luminal cancers is at present unclear. Specifically, it is possible that they represent completely distinct disease entities (as is widely assumed in the clinical realm), or they could in fact represent unique subsets of the basal and/or luminal tumors. Assumptions about the squamous and sarcomatoid variants have been challenged by the recent studies [22▪▪,23▪], underscoring the need for us to be completely objective as we characterize the other variants. A working group based within the Bladder Cancer Advocacy Network has been assembled to address this question, focusing first on micropapillary cancers [58].


The authors thank all of our collaborators and other members of the bladder cancer research community for their contributions to the work described in this article.

Financial support and sponsorship

This work was supported by the MD Anderson SPORE in Bladder Cancer (National Cancer Institute/National Institutes of Health) and the Baker Foundation.

Conflicts of interest

MD Anderson Cancer Center has filed a patent describing methods for identifying the subtypes of muscle-invasive bladder cancer.


Papers of particular interest, published within the annual period of review, have been highlighted as:

  • ▪ of special interest
  • ▪▪ of outstanding interest


1. Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999; 286:531–537.
2. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000; 403:503–511.
3. Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human breast tumours. Nature 2000; 406:747–752.
4. Prat A, Parker JS, Karginova O, Perou CM. Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast Cancer Res 2010; 12:R68.
5. Taube JH, Herschkowitz JI, Komurov K, et al. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci U S A 2010; 107:15449–15454.
6. Prat A, Ellis MJ, Perou CM. Practical implications of gene-expression-based assays for breast oncologists. Nat Rev Clin Oncol 2012; 9:48–57.
7. Cancer Genome Atlas N. Comprehensive molecular portraits of human breast, tumours. Nature 2012; 490:61–70.
8. Visvader JE. Keeping abreast of the mammary epithelial hierarchy and breast tumorigenesis. Genes Dev 2009; 23:2563–2577.
9. Prater MD, Petit V, Alasdair Russell I, et al. Mammary stem cells have myoepithelial cell properties. Nat Cell Biol 2014; 16:942–950.941-947.
10. Sale S, Lafkas D, Artavanis-Tsakonas S. Notch2 genetic fate mapping reveals two previously unrecognized mammary epithelial lineages. Nat Cell Biol 2013; 15:451–460.
11. Wang D, Cai C, Dong X, et al. Identification of multipotent mammary stem cells by protein C receptor expression. Nature 2015; 517:81–84.
12. Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 2001; 98:10869–10874.
13. Rehman FL, Lord CJ, Ashworth A. The promise of combining inhibition of PI3K and PARP as cancer therapy. Cancer Discov 2012; 2:982–984.
14. Blaveri E, Simko JP, Korkola JE, et al. Bladder cancer outcome and subtype classification by gene expression. Clin Cancer Res 2005; 11:4044–4055.
15. Dyrskjot L, Thykjaer T, Kruhoffer M, et al. Identifying distinct classes of bladder carcinoma using microarrays. Nat Genet 2003; 33:90–96.
16. Sanchez-Carbayo M, Socci ND, Lozano J, et al. Defining molecular profiles of poor outcome in patients with invasive bladder cancer using oligonucleotide microarrays. J Clin Oncol 2006; 24:778–789.
17. Lindgren D, Frigyesi A, Gudjonsson S, 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 2010; 70:3463–3472.
18▪▪. Sjodahl G, Lauss M, Lovgren K, et al. A molecular taxonomy for urothelial carcinoma. Clin Cancer Res 2012; 18:3377–3386.

First identification of the intrinsic subtypes of bladder cancer. The Lund subtypes were largely reproduced in the more recent efforts, which showed that they shared similarities with the intrinsic subtypes of breast cancer.

19. Volkmer JP, Sahoo D, Chin RK, et al. Three differentiation states risk-stratify bladder cancer into distinct subtypes. Proc Natl Acad Sci U S A 2012; 109:2078–2083.
20. Choi W, Shah JB, Tran M, et al. p63 expression defines a lethal subset of muscle-invasive bladder cancers. PLoS One 2012; 7:e30206.
21. Karni-Schmidt O, Castillo-Martin M, Shen TH, et al. Distinct expression profiles of p63 variants during urothelial development and bladder cancer progression. Am J Pathol 2011; 178:1350–1360.
22▪▪. Cancer Genome Atlas Research N. Comprehensive molecular characterization of urothelial bladder carcinoma. Nature 2014; 507:315–322.

One of four recent studies that documented the shared properties of the intrinsic subtypes of muscle-invasive bladder and breast cancers. This project provided the first high-quality, comprehensive inventory of genetic and epigenetic features of muscle-invasive bladder cancers, which has already become a major resource for the research community.

23▪. Choi W, Porten S, Kim S, et al. Identification of distinct basal and luminal subtypes of muscle-invasive bladder cancer with different sensitivities to frontline chemotherapy. Cancer Cell 2014; 25:152–165.

One of four recent studies that documented the shared properties of the intrinsic subtypes of muscle-invasive bladder and breast cancers. This study also showed that tumors in one of the subtypes (‘p53-like’), corresponding to the Lund ‘infiltrated’ tumors and TCGA's cluster II, were resistant to neoadjuvant chemotherapy.

24▪▪. Damrauer JS, Hoadley KA, Chism DD, et al. Intrinsic subtypes of high-grade bladder cancer reflect the hallmarks of breast cancer biology. Proc Natl Acad Sci U S A 2014; 111:3110–3115.

One of four recent studies that documented the shared properties of the intrinsic subtypes of muscle-invasive bladder and breast cancers. The group from UNC also documented the existence of a mesenchymal, ‘claudin-low’ subset of basal cancers.

25. Choi W, Czerniak B, Ochoa A, et al. Intrinsic basal and luminal subtypes of muscle-invasive bladder cancer. Nat Rev Urol 2014; 11:400–410.
26. McConkey DJ, Choi W, Dinney CP. New insights into subtypes of invasive bladder cancer: considerations of the clinician. Eur Urol 2014; 66:609–610.
27▪▪. Rebouissou S, Bernard-Pierrot I, de Reynies A, et al. EGFR as a potential therapeutic target for a subset of muscle-invasive bladder cancers presenting a basal-like phenotype. Sci Transl Med 2014; 6:244ra291.

One of four recent studies that documented the shared properties of the intrinsic subtypes of muscle-invasive bladder and breast cancers. The paper also presented strong preclinical evidence for targeting EGFR in basal cancers.

28. Chan KS, Espinosa I, Chao M, et al. Identification, molecular characterization, clinical prognosis, and therapeutic targeting of human bladder tumor-initiating cells. Proc Natl Acad Sci U S A 2009; 106:14016–14021.
29▪. Hoadley KA, Yau C, Wolf DM, et al. Multiplatform analysis of 12 cancer types reveals molecular classification within and across tissues of origin. Cell 2014; 158:929–944.

The study used TCGA's genomics data from 12 different cancers to document the similarities between basal bladder cancers and basal breast cancers, and revealed that they are even more similar to squamous cancers of the lung and head and neck. The study also confirmed that high level expression of ΔNp63α is a central feature of these cancers.

30. Polyak K, Weinberg RA. Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nat Rev Cancer 2009; 9:265–273.
31. Mani SA, Guo W, Liao MJ, et al. The epithelial-mesenchymal transition generates cells with properties of stem cells. Cell 2008; 133:704–715.
32. Ho PL, Lay EJ, Jian W, et al. Stat3 activation in urothelial stem cells leads to direct progression to invasive bladder cancer. Cancer Res 2012; 72:3135–3142.
33. Ho PL, Kurtova A, Chan KS. Normal and neoplastic urothelial stem cells: getting to the root of the problem. Nat Rev Urol 2012; 9:583–594.
34. Tontonoz P, Spiegelman BM. Fat and beyond: the diverse biology of PPARγ. Annu Rev Biochem 2008; 77:289–312.
35▪. Biton A, Bernard-Pierrot I, Lou Y, et al. Independent component analysis uncovers the landscape of the bladder tumor transcriptome and reveals insights into luminal and basal subtypes. Cell Rep 2014; 9:1235–1245.

Identified the transcriptional mechanisms underlying basal and luminal gene expression and identified PPARG as a candidate therapeutic target in luminal bladder cancers.

36. Dominick MA, White MR, Sanderson TP, et al. Urothelial carcinogenesis in the urinary bladder of male rats treated with muraglitazar, a PPARα/γ agonist: evidence for urolithiasis as the inciting event in the mode of action. Toxicol Pathol 2006; 34:903–920.
37. Egerod FL, Brunner N, Svendsen JE, Oleksiewicz MB. PPARα and PPARγ are co-expressed, functional and show positive interactions in the rat urinary bladder urothelium. J Appl Toxicol 2010; 30:151–162.
38. Faillie JL, Petit P, Montastruc JL, Hillaire-Buys D. Scientific evidence and controversies about pioglitazone and bladder cancer: which lessons can be drawn? Drug Saf 2013; [Epub ahead of print].
39. McConkey DJ, Choi W, Ochoa A, et al. Therapeutic opportunities in the intrinsic subtypes of muscle-invasive bladder cancer. Hematol Oncol Clin North Am 2015; 29:377–394.
40. Dinney CP, Hansel D, McConkey D, et al. Novel neoadjuvant therapy paradigms for bladder cancer: results from the National Cancer Center Institute Forum. Urol Oncol 2014; 32:1108–1115.
41. Prowell TM, Pazdur R. Pathological complete response and accelerated drug approval in early breast cancer. N Engl J Med 2012; 366:2438–2441.
42▪▪. Kurtova AV, Xiao J, Mo Q, et al. Blocking PGE2-induced tumour repopulation abrogates bladder cancer chemoresistance. Nature 2015; 517:209–213.

Provided strong preclinical evidence for a role for basal/stem-like cancer cells in recurrence and for using COX2 inhibitors with chemotherapy.

43. Harrington EA, Fanidi A, Evan GI. Oncogenes and cell death. Curr Opin Genet Dev 1994; 4:120–129.
44. Jackson JG, Pant V, Li Q, et al. p53-mediated senescence impairs the apoptotic response to chemotherapy and clinical outcome in breast cancer. Cancer Cell 2012; 21:793–806.
45. Topalian SL, Drake CG, Pardoll DM. Immune checkpoint blockade: a common denominator approach to cancer therapy. Cancer Cell 2015; 27:450–461.
46. Carthon BC, Wolchok JD, Yuan J, et al. Preoperative CTLA-4 blockade: tolerability and immune monitoring in the setting of a presurgical clinical trial. Clin Cancer Res 2010; 16:2861–2871.
47. Wolchok JD, Chan TA. Cancer: Antitumour immunity gets a boost. Nature 2014; 515:496–498.
48▪▪. Powles T, Eder JP, Fine GD, et al. MPDL3280A (anti-PD-L1) treatment leads to clinical activity in metastatic bladder cancer. Nature 2014; 515:558–562.

Demonstrated that immune checkpoint blockade is clinically active in refractory muscle-invasive bladder cancers. Catalyzed a huge wave of enthusiasm for the further evaluation of checkpoint blockade in both nonmuscle invasive and muscle-invasive bladder cancers.

49. Adam L, Zhong M, Choi W, 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 2009; 15:5060–5072.
50. Black PC, Brown GA, Inamoto T, et al. Sensitivity to epidermal growth factor receptor inhibitor requires E-cadherin expression in urothelial carcinoma cells. Clin Cancer Res 2008; 14:1478–1486.
51. Witta SE, Gemmill RM, Hirsch FR, et al. Restoring E-cadherin expression increases sensitivity to epidermal growth factor receptor inhibitors in lung cancer cell lines. Cancer Res 2006; 66:944–950.
52. Zhu K, Chan W, Heymach J, et al. Control of HIF-1α expression by eIF2 alpha phosphorylation-mediated translational repression. Cancer Res 2009; 69:1836–1843.
53. Semenza GL. HIF-1 mediates metabolic responses to intratumoral hypoxia and oncogenic mutations. J Clin Invest 2013; 123:3664–3671.
54. Kaelin WG Jr. The von Hippel-Lindau tumour suppressor protein: O2 sensing and cancer. Nat Rev Cancer 2008; 8:865–873.
55. Wong YN, Litwin S, Vaughn D, et al. Phase II trial of cetuximab with or without paclitaxel in patients with advanced urothelial tract carcinoma. J Clin Oncol 2012; 30:3545–3551.
56. Hahn NM, Stadler WM, Zon RT, et al. Phase II trial of cisplatin, gemcitabine, and bevacizumab as first-line therapy for metastatic urothelial carcinoma: Hoosier Oncology Group GU 04-75. J Clin Oncol 2011; 29:1525–1530.
57. Iyer G, Hanrahan AJ, Milowsky MI, et al. Genome sequencing identifies a basis for ever olimus sensitivity. Science 2012; 338:221.
58. Willis DL, Flaig TW, Hansel DE, et al. Micropapillary bladder cancer: current treatment patterns and review of the literature. Urol Oncol 2014; 32:826–832.

basal; claudin-low; epithelial-to-mesenchymal transition; luminal; PDL1; TCGA

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.