Novel Therapies in Urothelial Carcinoma

A Biomarker-Driven Approach

G. Iyer; J. E. Rosenberg


Ann Oncol. 2018;29(12):2302-2312. 

In This Article

Novel Therapies and Therapeutic Strategies in Urothelial Malignancies

Identification of prognostic biomarkers (to provide information on overall cancer outcome in patients and/or to facilitate cancer diagnosis) and predictive biomarkers (to inform treatment decisions) are integral to a personalized oncology approach in cancer. A number of strategies have been used to identify clinically useful biomarkers for UC. The Cancer Genome Atlas (TCGA) analyzed 412 muscle-invasive, high-grade bladder cancers using multiple platforms, including whole exome sequencing, RNA sequencing, DNA methylation profiling, and others. The total mutation burden was high, similar to that of melanoma and non-small-cell lung cancers; moreover, most mutations were clonal and were detected within the context of an APOBEC mutation signature (Figure 2).[32] RNA expression subtyping delineated five distinct molecular subtypes of bladder cancer, including luminal, luminal-papillary, luminal-infiltrated, basal-squamous, and neuronal. The neuronal subtype was associated with the worst survival of the five subtypes and was enriched for TP53 and RB1 alterations while the luminal infiltrated subtype has been correlated with response to clinical trials of immune checkpoint blockade in UC.

Figure 2.

Characterization of muscle-invasive bladder cancer [32]. Alteration landscape for 412 primary tumors. Top to bottom: synonymous and nonsynonymous somatic mutation rates, with one ultra-mutated sample with a POLE signature. Mutational signature (MSig) cluster, APOBEC mutation load, and neoantigen load by quartile. Normalized activity of four mutational signatures. Combined tumor stage (T1, 2 versus T3, 4) and node status, papillary histology, gender, and squamous histology. Somatic mutations for significantly mutated genes (SMGs) with frequency ≥7%. Copy number alterations for selected genes and FGFR3 and PPARG gene fusions.

While several groups have identified similar molecular subtypes based upon expression profiling of UC, the predictive and prognostic characteristics of these subtypes are as yet unclear and require prospective validation. A novel gene expression algorithm, Co-eXpression ExtrapolatioN (COXEN), is an informatics-based approach to predict sensitivity of independent cell line panels and patient responses to therapeutic agents.[34,35] The COXEN model combines in vitro and in vivo molecular profiling of cancer cell lines and drug responsiveness using the NCI-60 and BLA-40 cancer cell line panels (Figure 3). For the purposes of prospective validation, the COXEN algorithm is being incorporated into the SWOG S1314 study in which muscle-invasive bladder cancer patients are randomized to two different neoadjuvant chemotherapy (NAC) regimens. The results of COXEN will be correlated with response to chemotherapy at cystectomy to gauge the capability of the algorithm to predict for chemo-sensitivity. Additional correlative analysis of pre-treatment specimens and post-chemotherapy residual tumors is planned, including whole transcriptome profiling to facilitate correlation of the published molecular subtypes with response and survival.

Figure 3.

Application and performance characteristics of the COXEN algorithm for prediction of drug sensitivity in the BLA-40 human urothelial cancer cell lines [36]. Step 1: Experimentally determine the drug's pattern of activity in cells of set 1. Step 2: experimentally measure molecular characteristics of the cells in set 1. Step 3: select a subset of those molecular characteristics that most accurately predicts the drug's activity in cell set 1 ('Chemosensitivity signature' selection). Step 4: experimentally measure the same molecular characteristics of the cells in set 2. Step 5: among the molecular characteristics selected in step 3, identify a subset that shows a strong pattern of co-expression extrapolation between cell sets 1 and 2. Step 6: use a multivariate algorithm to predict the drug's activity in set 2 cells on the basis of the drug's activity pattern in set 1 and the molecular characteristics of set 2 selected in step 5. The output of the multivariate analysis is a COXEN score.

Next-generation sequencing has enabled a more detailed analysis of the underlying genetic changes characterizing UC and identification of potential new therapeutic targets. Although most tumors harbor potentially tractable genomic alterations that may predict for response to target-selective agents, substantial genomic heterogeneity has been identified in urothelial bladder cancer.[37] Additionally, subclonal mutational heterogeneity, which may contribute to chemotherapy resistance,[38] poses an obstacle to the effective application of targeted agents.

Among the promising therapeutic targets, FGFR3 is of particular interest. It encodes for a transmembrane protein, fibroblast growth factor receptor-3, involved in regulating cell proliferation and angiogenesis; putative downstream signaling pathways include the PI3K-AKT, PKC, and Ras/MAPK pathways (Figure 4). The most common FGFR3 alterations in UC are activating point mutations. Dysregulation of FGFR3 signaling occurs in ~80% of noninvasive and 50% of invasive bladder cancers, through mutation, overexpression, or both, and over-expression is also common in tumors with no detectable mutations, including in muscle-invasive tumors.[40] FGFR3 alterations have been reported in up to 21% of patients with high-grade and advanced stage bladder UC. Activating FGFR3 fusions have also been described in UC.[41] Urothelial cell lines harboring these fusions are highly sensitive to FGFR-selective agents and responses to FGFR-directed inhibitors have been observed in patients with FGFR3 fusion positive UC. Thus, these alterations may serve as predictive biomarkers that could aid patient selection for FGFR-targeted therapy.[42,43]

Figure 4.

FGFR pathway and dysregulation in cancer [39].

Although dovitinib, a multi-targeted inhibitor of tyrosine kinases including FGFR3, has not shown significant activity in UC (due in part to challenges with appropriate patient selection using the available sequencing platforms at the time), other agents targeting FGFR3-mutant UC remain under investigation.[44] BGJ398 is an FGFR1–3 inhibitor that showed promising antitumor activity in patients with advanced solid tumors, including those with UC.[45] These data led to evaluation of drug efficacy in an extended cohort of patients with advanced/metastatic UC and activating FGFR3 mutations/fusions after platinum-based chemotherapy.[46] Seventy-six percent of patients had received two or more prior treatments. The ORR was 36%, including eight partial responses, suggesting clinical benefit with FGFR3 signaling blockade in this patient population. Clinical activity has also been observed with erdafitinib, a pan-FGFR inhibitor, in patients with metastatic UC and this drug recently received FDA breakthrough designation for this disease. In a phase II study of erdafitinib in patients with metastatic UC (BLC 2001) and pre-specified FGFR alterations, the ORR was 42% (3% CR rate).[47] Notably, the response rate was 59% in the subset of patients who had received prior immune checkpoint blockade.

Inhibition of mechanistic target of rapamycin (mTOR) is also a target for therapy in UC under investigation. Everolimus, which blocks mTOR signaling, was tested in two phase II studies in patients with advanced UC.[48,49] The study by Milowsky et al. did not reach the primary end point of an improvement in 2-month progression-free survival compared with historical rates. In the study by Seront et al., a disease control rate of 27% was seen, with two partial responses. Everolimus in combination with paclitaxel was not effective as a second-line treatment of UC in the German AUO Trial AB 35/09[50] showing a 13% ORR. In addition, a regimen combining everolimus with gemcitabine and split-dose cisplatin in advanced UC was not feasible due to dose-limiting hematologic toxicities.[51] Despite these outcomes, evidence to support the use of everolimus in specific molecular contexts exists, with some patients harboring specific mutations in the mTOR pathway experiencing exceptional responses to everolimus-containing therapy. The activating mutations MTOR E2419K and MTOR E2014K were found in a patient tumor that was exquisitely sensitive to mTOR inhibition; the patient had platinum- and taxane-refractory UC and experienced a complete radiologic response that lasted for 14 months on pazopanib plus everolimus.[52] Loss-of-function mutations in TSC1 and NF2, both regulators of mTOR pathway activation, have been correlated with durable everolimus sensitivity in an exceptional responder that has been ongoing for greater than 6 years;[53] additionally, NF2 mutations have been identified as conferring significant response to everolimus-containing regimens in patients with urothelial malignancies.[54] Studies to investigate the clinical benefit of everolimus targeted to cancer patients with inactivating TSC1 or TSC2 mutations or activating MTORmutations (NCT02201212) and NF2 mutations (NCT02352844) are on-going.

Analysis of TCGA has also shown that epidermal growth factor receptor (EGFR) signaling pathways are upregulated in selected advanced UC tumors, with EGFRamplification in 11% of UC and ERBB2 amplification and mutation in 9%.[55] Lapatinib is a dual tyrosine kinase inhibitor of EGFR and HER2 that showed antitumor activity in a phase II study in a subset of 34 patients with EGFR/HER2-overexpressing tumors.[56] Conversely, in a study of 232 patients with metastatic bladder cancer who had clinical benefit from first-line chemotherapy and HER1/HER2-positive status confirmed by centralized immunohistochemistry (IHC), no improvement in progression-free survival was observed with maintenance lapatinib compared with placebo.[57] Targeting mutant/amplified tumors may still be a useful strategy, however, and negative results may reflect a failure of proper patient selection due to the assay used to determine gene amplification or protein overexpression.

Afatinib, another dual EGFR and HER2 inhibitor, has demonstrated clinical activity in patients with platinum-refractory UC with ERBB2 or ERBB3 genetic alterations detected by quantitative polymerase chain reaction or fluorescent in situhybridization.[58] Five of the 6 patients with ERBB2 and/or ERBB3 alterations achieved the primary end point of 3-month progression-free survival compared with none of the 15 patients without such alterations (P < 0.001); median time to progression/therapy discontinuation was 6.6 months in patients with ERBB2/ERBB3alterations versus 1.4 months in those without alterations (P < 0.001). This study suggests that selection by genomic alterations may be more relevant to the management of UC than protein expression as detected by IHC.

Overall mutational load is also emerging as an additional biomarker of response in advanced UC in patients treated with PD-1-blocking agents (Table 1). The phase II studies of atezolizumab in patients with locally advanced and metastatic UC who have progressed following treatment with platinum-based chemotherapy[13] and as first-line treatment in cisplatin-ineligible patients with locally advanced and metastatic UC[59] showed that higher mutation load was consistently associated with improved response to atezolizumab therapy.

Combining various immunotherapies may also yield benefits surpassing those of monotherapy, and several combinations are currently being examined, including combination immune checkpoint blockade plus chemotherapy as well as with novel agents (both targeted therapies and immune modulating agents). A number of trials are examining the utility of combining platinum-based chemotherapy with checkpoint blockade in the first-line setting for metastatic UC in an attempt to leverage the response rates observed with platinum chemotherapy in this disease with the durability seen with checkpoint blockade. A study of durvalumab as first-line monotherapy or in combination with tremelimumab, an anti-CTLA-4 antibody, in patients with unresectable stage IV urothelial bladder cancer (the phase III DANUBE study) is recruiting patients,[63] with primary outcome results expected in 2018. The BISCAY trial, an open-label, randomized, multi-drug, biomarker-directed, phase Ib study in patients with bladder cancer who have failed at least one prior platinum regimen, will assess immunotherapy with durvalumab as monotherapy and in combination with select targeted therapies. Treatment assignment is based upon the tumor biomarker profile, and includes: olaparib, a PARP inhibitor; vistusertib, an mTOR inhibitor; AZD4547, an FGFR inhibitor; and AZD1775, a WEE1 inhibitor. The incorporation of prospective molecular analysis of tumors to inform assignment to a specific therapy could optimize the efficacy of small molecule inhibitors and possibly enhance responses to checkpoint blockade through the generation of neoantigens in this setting. Multiple checkpoint inhibitors are currently being investigated in combination with novel immunotherapies to enhance or overcome resistance to PD-1 blockade in patients with advanced solid tumors. Some of these include combinations with GSK3174998, an agonist of OX40, a potent costimulatory tumor necrosis factor receptor expressed on activated CD4+ and CD8+ T cells; MK-4280, targeting the lymphocyte-activation gene 3 (LAG3) receptor; NKTR-214, a CD122 agonist; TSR-022, an anti-T-cell immunoglobulin and mucin containing protein-3 (TIM-3) antibody; BMS-98625, an IDO1 inhibitor; and CDX-1127, targeting CD27, a lymphocyte-specific member of the tumor necrosis factor receptor.