Xenografts of primary human gynecological tumors grown under the renal capsule of NOD/SCID mice show genetic stability during serial transplantation and respond to cytotoxic chemotherapy

Objectives. Human cancer tissue xenograft models may provide a more accurate reflection of tumor biology than cell lines. This study evaluates the genetic and phenotypic stability of primary human gynecological tumors grown as serially transplanted xenografts. The response to conventional chemotherapy and novel molecular targeted chemotherapy is assessed in one of the transplantable xenograft lines. Methods. Fresh tumor was transplanted beneath the renal capsule of NOD/SCID mice. Transplantable tumor lines were derived from 5 tumors (4 ovarian carcinomas and 1 uterine sarcoma), and serially transplanted for 2-6 generations. Comparisons were made between primary tumor and corresponding transplantable xenografts by CGH array, immunohistochemistry, and BRCA mutation analysis. Transplantable xenografts created from known BRCA1 germline mutation carriers were analyzed for histopathologic response (tumor volume, apoptotic and mitotic indices) to combination carboplatin/paclitaxel and to PARP inhibitor (PJ34). Results. Unsupervised hierarchical cluster analysis applied to a 287 feature CGH array demonstrated a low degree of intratumoral genetic variation in 4/5 cases, with greater degree of variation in the fifth case (clear cell ovarian carcinoma derived from an omental sample). Assessment of proliferation using MIB-1 staining was concordant between primary tumor and transplantable xenograft in all ovarian cancer cases. BRCA mutation analysis identified germline BRCA1 mutation for further testing and this xenograft showed a significant response to carboplatin/paclitaxel chemotherapy, including a decrease in tumor volume and proliferation but did not demonstrate a response to the poly (ADP-ribose) polymerase-1 inhibitor PJ34. Conclusions. Xenografts derived from gynecologic tumors can be serially transplanted and grown under renal capsule of NOD/SCID mice with minimal genetic change. This model may be used to study progression of tumors, identify therapeutic targets, and test treatment modalities in tumors with well-characterized abnormalities in genes of fundamental importance in ovarian carcinogenesis, such as loss of BRCA1. (c) 2008 Elsevier Inc. All rights reserved.