Abstract B2: Polyubiquitin heterogeneity in colorectal cancer and its liver metastatic cancers: implications for TAK-243 (MLN7243) pharmacodynamic marker sampling and interpretation

Ubiquitin is a small protein (8.5KDa) involved in post-translational modification of substrate proteins. The ubiquitin tag can be poly- or mono-ubiquitin and determines the tagged proteins9 fate in many cellular functions including protein degradation, induce receptor internalization, DNA damage repair or protein-protein interactions. Drugs targeting the ubiquitin-proteasome system have been recognized as viable anti-tumor options. In both hematological malignancies and solid tumors, clonal evolution due to tumor environmental or therapeutic pressures may contribute to the molecular diversity of the tumor cells within one tumor mass. This diversity of cells with a tumor mass is known as tumor heterogeneity. It is important to understand the potential impact of tumor heterogeneity on the interpretation of pharmacodynamic (PD) biomarkers in order to differentiate whether pre vs. post dose biomarker changes are due to a sampling or drug effect. Global polyubiquitin levels are being assessed in clinical samples as a PD readout for inhibition of the investigational ubiquitin activating enzyme 1 inhibitor, TAK-243(MLN7243). We investigated and report here ways to minimize false positive/negative interpretation of PD biomarker due to tumor heterogeneity. Polyubiquitin expression was processed by immunohistochemistry (IHC) in colorectal, and colorectal cancer liver metastasis from purchased human resected tumors and images were acquired from the whole sections. Images were then quantitatively measured on a per-cell basis using Definiens image analysis software. A statistical simulation was performed to generate 5000 virtual pairs of biopsies from the same resected tumor sample images. These images were then used to calculate differences in the above biomarker expression between the pair of biopsies. The distribution of difference between pairs of biopsies in polyubiquitin expression provides a quantitative range of biomarker variation due to non-drug related biological variation. Statistical simulations were performed for the polyubiquitin biomarker to calculate the tolerance boundaries to identify potential false positive/negative cutoffs. Statistical simulations revealed that biopsy variability depends on both intrinsic intra-tumor heterogeneity and the number of viable cells for biopsies. The more viable cells from the virtual biopsy samples, the less heterogeneous the data became. Analysis also revealed that the tolerance intervals for mean intensity are narrower than those for H-score. The data suggests, a higher threshold (95% tolerance interval) will reduce the false positive PD effect readouts and using a lower threshold (such as an 80% tolerance interval) will reduce false negative PD effect calls. This approach may offer a method to distinguish PD marker change from biological differences due to intra- tumor heterogeneity and drug effect. Although operational alternatives may be implemented as risk mitigation tactics for tumor heterogeneity, this approach offers a novel intervention to maximize the clinical PD biomarker data from a paired biopsy to establish a meaningful PD biomarker positivity threshold with minimal operational burden. Citation Format: Yuko Ishii. Polyubiquitin heterogeneity in colorectal cancer and its liver metastatic cancers: implications for TAK-243 (MLN7243) pharmacodynamic marker sampling and interpretation. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr B2.