Renal-cell carcinomas (RCC) are often resistant to conventional cytotoxic agents. Xenograft models are used for in vivo preclinical studies and drug development. The validity of these studies is highly dependent on the phenotypic and genotypic stability of the models. Here we assessed the stability of six aggressive human RCC xenografted in nude/NMRI mice. We compared the initial samples (P0), first (P1) and fifth (P5) passages for the following criteria: histopathology, immunohistochemistry for CK7, CD10, vimentin and p53, DNA allelic profiles using 10 microsatellites and CGH-array. Next we evaluated the response to sunitinib in primary RCC and corresponding xenografted RCC. We observed a good overall stability between primary RCC and corresponding xenografted RCC at P1 and P5 regarding histopathology and immunohistochemistry except for cytokeratin 7 (one case) and p53 (one case) expression. Out of 44 groups with fully available microsatellite data (at P0, P1 and P5), 66% (29 groups) showed no difference from P0 to P5 while 34% (15 groups) showed new or lost alleles. Using CGH-array, overall genomic alterations at P5 were not different from those of initial RCC. The xenografted RCC had identical response to sunitinib therapy compared to the initial human RCC from which they derive. These xenograft models of aggressive human RCC are clinically relevant, showing a good histological and molecular stability and are suitable for studies of basic biology and response to therapy.
Keywords: Xenograft; preclinical model; renal cell carcinoma; tumor genotype; tumor phenotype.