Establishment of patient-derived renal cell carcinoma (RCC) models based on orthotopic xenografts (PDX) and cancer stem cell (CSC) isolation to provide prognostic and predictive information.

Abstract

e16055 Background: The introduction of VEGF/VEGFR- and mTOR, and immune checkpoint-targeted agents has radically changed metastatic RCC treatment; however, predictive biomarkers are still lacking and the potential for curative impact of such treatments in earlier disease stages is hotly debated. To identify potential diagnostic, prognostic, and predictive biomarkers, we developed patient-derived preclinical RCC models based on PDX obtained from freshly dissociated cancer tissues and combining isolation of cell populations endowed with self renewal and multi-lineage differentiation abilities (CSC). Methods: 182 surgical RCC specimens (clear cell, n = 110; papillary type I/II, n = 23, chromophobe/oncocytoma, n = 26, other, n = 23; stage I-III, n = 124, stage IV, n = 28, n/a n = 30) were collected; tumor spheroids were obtained and characterized in 57 cases, using a stem-cell isolating medium supplemented with EGF/b-FGF; 30 cancer samples were orthotopically injected in immunocompromised mice and were able to engraft in 67% of cases (G3-G4 cases more frequently than G1-G2 cases, p = 0.04). Results: PDX tumors recapitulated histological appearance, sarcomatoid features when they were present, and tumor heterogeneity of their human counterpart; interestingly, PDX engraftment was significantly correlated with shorter DFS in the corresponding patient population (p = 0.12). We conducted proteomic analysis by Reverse Phase Protein Array (RPPA) in 21 patient-derived CSC models in vitro: a panel of established (HIF, VEGFR, mTOR) and novel (EGFR(Y1148)), AKT, PI3K) oncogenic signals were deregulated in our isolated cell populations; molecular endpoints were correlated with grading and with angiogenesis and mTOR pathways (p < 0.05). Data analysis is ongoing to ascertain whether specific RPPA-signatures can potentially complement clinical prognostic information and suggest candidate biomarkers and potential targets for therapy. Conclusions: In the era of personalized therapy, the combined use of PDX and RPPA from tumor specimens may be very useful for drug testing and patient stratification in RCC.