Gene expression profiles classify human osteosarcoma xenografts according to sensitivity to doxorubicin, cisplatin, and ifosfamide.

PURPOSE In osteosarcoma, aggressive preoperative and postoperative multidrug chemotherapy given to all patients has improved patient survival rate to the present level of approximately 60%. However, no tumor marker is available that reliably can identify those patients who will or will not respond to chemotherapy. EXPERIMENTAL DESIGN In an attempt to find leads to such markers, we have obtained microarray gene expression profiles from a panel of 10 different human osteosarcoma xenografts and related the results to their sensitivity to ifosfamide, doxorubicin, and cisplatin. RESULTS The expression data identified genes with highly significant differential expression between poor and good responder xenografts to the three different drugs: 85 genes for doxorubicin, 74 genes for cisplatin, and 118 genes for ifosfamide. Technical validation with quantitative reverse transcription-PCR showed good correlation with the microarray expression data. Gene Ontology-guided analysis suggested that properties of the poorly responsive xenografts were resistance to undergo programmed cell death and, particularly for ifosfamide, a drive toward dedifferentiation and increased tumor aggressiveness. Leads toward metabolic alterations and involvement of mitochondrial pathways for apoptosis and stress response were more prominent for doxorubicin and cisplatin. Finally, small interfering RNA-mediated gene silencing of IER3 and S100A2 sensitized the human osteosarcoma cell line OHS to treatment with 4-hydroperoxyifosfamide. CONCLUSIONS The expression profiles contained several novel biomarker candidates that may help predict the responsiveness of osteosarcoma to doxorubicin, cisplatin, and ifosfamide. The potential of selected candidates will be further validated on clinical specimens from osteosarcoma patients.