Biomarker Strategies (United States)

Abstract Introduction: Molecularly targeted agents, such as the BRAF inhibitor vemurafenib, may produce short-term responses in some patients; however, most patients are intrinsically resistant, or develop resistance through restructuring of signal transduction pathways. SnapPath™ is a live-cell-processing platform that utilizes ex vivo signal transduction modulation of live tumor samples to produce Functional Signaling Profiles (FSPs). Application of this technology to Champions TumorGraft models may provide novel insights to guide oncology drug development as these models preserve the biological properties of the original human tumor. Methods: Fresh melanoma tumor specimens were collected from patients and implanted into immunodeficient mice. Fine needle aspiration biopsies were performed on each melanoma TumorGraft model and processed on the SnapPath™ platform (BioMarker Strategies) to modulate tumor cell signal transduction networks through brief ex vivo exposure to the vemurafenib tool compound PLX-4720. Cell lysates were then analyzed using a multiplexed immunoassay to assess the inhibition of the downstream MAPK markers pMEK1 and pERK-1/2. FSPs were then created for each TumorGraft model based on baseline and modulated levels of each phosphoprotein. In parallel, the in vivo sensitivity to vemurafenib and BRAF mutation status was evaluated in each Champions TumorGraft model. FSPs were then compared with in vivo efficacy, gene expression and genotype data. Results: Functional profiling stratified the TumorGraft models into two distinct groups upon ex vivo exposure to a BRAF inhibitor: 1) MAPK markers suppressed and 2) MAPK markers not suppressed. As anticipated, TumorGraft models that showed resistance to ex vivo BRAF inhibition demonstrated vemurafenib resistance in vivo and were BRAF wild type. There were other models that displayed MAPK suppression with ex vivo BRAF inhibition and vemurafenib sensitivity in vivo or MAPK suppression with ex vivo BRAF inhibition but demonstrated vemurafenib resistance in vivo. One of these TumorGrafts contained a BRAF V600E mutation, suggesting the activation of an alternate pathway that conferred resistance. The other TumorGraft contained a novel BRAF insertion. The functional profiling suggests that this insertion may activate BRAF and is susceptible to vemurafenib inhibition, but the tumor may contain an alternate pathway that confers resistance. Analysis of gene expression data demonstrated hierarchical clustering of BRAF mutated TumorGraft models. Conclusions: These results demonstrate the capability of the SnapPath™ platform to generate FSPs from FNAs of Champions melanoma TumorGraft models. Overall, the combination of Champions TumorGraft models with functional profiling represents a powerful tool for pharmacodynamic assessment of targeted therapeutics in clinically relevant models and has the potential to guide oncology therapy. Citation Format: Elizabeth M. Bruckheimer, Adam Schayowitz, Kala Barnes, Greg Bertenshaw, Tin Khor, James Cotton, Jay Friedman, Dhanrajan Tiruchinapalli, Douglas P. Clark. Functional profiling of Champions TumorGraft™ models from metastatic melanoma patients. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3498. doi:10.1158/1538-7445.AM2013-3498

The pace of technical and scientific advancement for genomics-based technologies has outstripped the ability of the US regulatory bodies to keep abreast of the emerging paradigms, resulting in regulatory pronouncements that often appear dated and payment systems that are misaligned. Over burdensome evidentiary requirements, encroachment of federal regulators into the practice of laboratory medicine, and failure to align public health objectives with pay or valuation sufficient to support professional pathology services and necessary capital investment threaten to stifle continued innovation in genomic-based diagnostic tools. Nevertheless, the US FDA is committed to improving regulatory science and to increased stakeholder participation in policy-making, and serious efforts to address coding and test value are underway. Moreover, federal legislation will be debated in the coming months that, if enacted, could clarify authorities and institute meaningful regulatory and reimbursement paradigms better suited to molecular testing. This review explores these dynamic policy issues and their implications for genomic pathology as the foundation of personalized medicine.

Abstract Background: Abnormal signal transduction networks are frequent targets of existing and emerging molecularly targeted agents (MTAs). Unfortunately, most predictive biomarkers to guide therapeutic selection are based on indirect assessment of signal transduction through DNA mutations or transcriptional profiles rather than dynamic assessment of signal transduction proteins themselves. Classification of breast cancer based on functional signaling profiles derived from a set of signaling phosphoproteins induced upon growth factor stimulation of live breast cancer cells is likely to provide a more accurate system for MTA selection than indirect methods utilizing fixed or frozen tissue. Objective: This study aimed to demonstrate stratification of multiple breast cancer model systems based on functional signaling profiles elicited from live tumor cells in response to ex vivo stimuli. Methods: Breast cancer cell lines (MCF-7, HCC-1937, MDA-MB-231, BT474, and SKBR3) were exposed to either vehicle (control) or stimulated with 200 ng/ml epidermal growth factor (EGF) for 5 minutes then lysed and proteins extracted. Mean Fluorescence Intensity (MFI) levels of six phosphoproteins (pEGFR, pErk, pAkt, pP70S6K, pGSK3β, and pSTAT3) were determined in sextuplet using a multiplexed bead-immunoassay (BioPlex, BioRad) and a modulation score (MS), defined as the log 2 (MFI stimulated /MFI contrd), calculated for each. Scores were ranked by percentile relative to the median (0.66) and inter-quartile range (IQR) (154). Moderate responders were classified as those with MS between the 75th percentile (2.20) and the 75th percentile plus the IQR (3.74). High responders were those MS > 3.74. Low responders were those MS falling between the IQR and 75th percentile (1.54-2.20) whereas non responders were classified as MS < 1.54 Results: EGF stimulation resulted in high levels of EGFR-phosphorylation in all cells except BT474, which responded moderately (2.57). MS for pErk were high in MCF-7 cells (3.92), moderate in HCC -1937 (2.89) and none for the other lines tested. Moderate STAT-3 phosphorylation was observed in only MCF-7cells (2.34) whereas low pAkt MS were observed in only SKBR3 (1.78). All other markers across the five cell lines tested were non responders (< 1.54), with pGSK3β and pP70S6K yielding MS < 1.0 for all five cell lines. Interestingly, the relative MS rank order of all six proteins differed across each cell line suggesting further opportunity for stratification. Conclusions: Our data demonstrate that different breast cancer cell lines display unique functional phosphoprotein signaling profiles, thereby providing a mechanism for stratifying tumors based on individual signal transduction pathway activation. Further studies are underway to correlate the functional signaling profiles identified here to sensitivity and resistance to specific MTA treatment. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-04-06.