Introduction & Objectives
Neoadjuvant chemotherapy (NAC) followed by radical cystectomy (RC) is recommended for muscle invasive bladder cancer (MIBC). However, only ~40% of patients show an objective response. While DNA alterations and RNA classifiers may predict response to NAC in retrospective studies, the proteome has not been evaluated in this context. Here we analyze the MIBC proteome in a NAC-treated cohort to identify markers of response to chemotherapy and to elucidate mechanisms of resistance to NAC.
Materials & Methods
Pre-NAC tissue was included from 107 MIBC patients who received NAC followed by RC. Residual tumor (≥ypT1N0-3M0-1) in the RC specimen was present in 62% of patients after NAC, and was available for 55 of those patients (51%). Multiregional sampling was conducted in 37/107 pre-NAC and 15/55 post-NAC samples. Benign ureter was used as control. SP3-Clinical Tissue Proteomics (SP3-CTP) and bioinformatic analysis using formalin-fixed paraffin-embedded tissue (FFPE) were performed. Immunohistochemistry (IHC) validation was performed on matched tissues.
Results
We quantified 9769 unique proteins across all samples. Unsupervised clustering of pre‑NAC tissue established 4 clusters with different biology and survival outcomes, but no difference in response by pathologic stage. Clusters were confirmed by IHC, and consisted of: Cluster 1 (CC1), with high metabolic activity and a luminal profile; Cluster 2 (CC2) with high nuclear processes activity; Cluster 3 (CC3) with high immune infiltration, and basal characteristics; and Cluster (CC4) with high immune infiltration and stromal signature. CC3 showed worse survival outcomes (p<0.01). Multivariable analysis using pre-NAC tissue identified novel favorable (MAPK9 and MTIF3) and unfavorable (DVL2 and NES) biomarkers. Matched analysis of pre- and post-NAC tissue identified markers and biological pathways indicative of NAC resistance (AZGP1 and ORM1). Analyses of post-NAC (i.e. resistant) tumors identified 2 clusters: Post-NAC Cluster 1 was enriched for nuclear processes and had worse outcomes, whereas Post-NAC Cluster 2 was enriched for immune pathways. Multiregional analysis of pre- and post-NAC matched tumors revealed clonal heterogeneity suggestive of important chemoresistance mechanisms.
Conclusions
Here we describe 4 pre-NAC and 2 post-NAC proteomic clusters with distinct biology and survival outcomes, alongside novel prognostic biomarkers. Future work includes the validation of these clusters by IHC in larger independent MIBC cohorts. A non-NAC cohort using pre-NAC tissue will be used to confirm the prognostic vs. predictive relevance of these findings.