Towards personalized medicine: Evaluation of blood and tissue biomarkers for drug sensitivity in a newly established human xenograft model

B14 The goal of personalized medicine is to enhance our understanding of the disease at the molecular level and optimize drug treatment to best meet the needs of an individual. In addition to genomic biomarkers for drug response, the mitochondrial genome (mtDNA) and its heteroplasmic and homoplasmic sequence variants are gaining more recognition as important pharmacodynamic markers. Emerging proteomic assay platforms, multiplexed immunoassays and mass spectral blood and tissue analyses have yielded new putative biomarkers and protein signature patterns that are moving into clinical validation.To develop an experimental strategy for identification of biomarkers for drug sensitivity in tumors from individual donors, we performed a feasibility study using a recently established transplantable mesenchymal chondrosarcoma (MCS) model. For model development, the tumor was removed from a patient and fragments xenografted into athymic nude mice. Tumors were propagated and amplified until growth was stable. Following establishment, the MCS model was evaluated for sensitivity towards a number of single agent and combination therapies including the investigational agent perifosine and the recently approved VEGF inhibitor sorafenib (Nexavar ® ), both alone and in combination with bevacizumab (Avastin ® ). Both perifosine and sorafenib treatment resulted in significant tumor growth inhibition which was enhanced by bevacizumab. Moreover, perifosine as a monotherapy induced a partial tumor regression and all tumors regressed in the combination group.Blood serum and tumor tissue specimens were obtained in early phases of model establishment, and from the vehicle and drug treatment groups in the established model. We have evaluated phenotypic stability of tumors in consecutive passages through the analysis of hypervariable regions of mitochondrial DNA using the multiplexing Luminex xMAP platform. Assessment of changes in expression and circulation of blood cytokines and growth factors and analysis of potential proteomic biomarkers for drug response in serum and tumor xenograft tissue were carried out with a combination of multiplexed immunoassays and tryptic peptide fingerprinting by liquid chromatography/mass spectrometry (LC/MS).To date, examination of a limited sampling of early passage tumors revealed no significant changes in mtDNA expression. Tumors treated with sorafenib alone or along with bevacizumab demonstrated decreased serum levels of several examined cytokines and growth factors including IL-2, IL-6, interferon-γ, and TNF-α. LC/MS peptide fingerprinting underscored the presence of distinct peptide patterns that could distinguish between treated versus control tumors despite clear evidence of large heterogeneity of tumor proteomes within the control and drug treatment groups. Further elucidation of amino acid sequences of select peptides is under way. This pilot study in a human sarcoma model indicates feasibility of extensive studies which, in conjunction with in vivo tumor growth inhibition and chemosensitivity testing, will lead to identification of biomarkers for drug sensitivity, provide more insight into drug mechanism of action in responding models, and facilitate selection of the most effective therapies for patients with cancer.