Abstract A31: Development of an immunohistochemistry-based pharmacodynamic biomarker program for phase I trials of MLN4924, an inhibitor of the NEDD8 activating enzyme

Abstract

Abstract The novel small molecule MLN4924 is an inhibitor of the NEDD8 activating enzyme (NAE) and is currently in Phase I trials. As part of the ubiquitin-proteasome system, NAE is required for activity of cullin ring ligases (CRLs), a category of ubiquitin E3 ligases. When NAE is inhibited, levels of CRL substrates increase due to impaired ubiquitination and proteasomal degradation; CRL substrates can therefore serve as pharmacodynamic (PD) biomarkers of NAE pathway inhibition. We describe the development of Immunohistochemistry (IHC) assays for two CRL substrates: NRF2 (a transcription factor involved in the oxidative stress response) and CDT1 (a protein that regulates DNA replication). CDT1 and NRF2 IHC assays were initially evaluated in PD studies of xenograft tumors grown in immunocompromised mice and showed dose-dependent regulation. The IHC assays were then adapted for use in clinical samples, including skin and tumor biopsies from patients in Phase I trials of MLN4924. Sample quality and morphology were assessed by pathologist review of hematoxylin and eosin pathology stain. A set of formalin-fixed, paraffin-embedded cell pellets was evaluated with each IHC run to detect technical variability and establish pass/fail criteria for each experimental IHC run. The QC samples were generated from the HCT116 cell line treated with various concentrations of MLN4924. The QC controls were assayed multiple times to generate an expected range as a percent positive area above a threshold for each concentration of MLN4924. The ranges were then applied to each experimental run to pass or fail the experiment. In cases of assay failure, the clinical samples were re-stained. Skin and melanoma tumors are inherently challenging for IHC analysis due to the population variability of melanin pigment, which can be indistinguishable from brown DAB chromagen. Adaption of the assays to a blue chromagen (NBT/BCIP) enabled a contrast between the chromagen stain and endogenous pigments. A region of interest (ROI) for skin was manually drawn to select the epidermal area. Semi-automated image analysis was used to measure percent positive area for CDT1 and NRF2 within the total epidermal region. To determine the baseline levels for CDT1 and NRF2 signal, 10 benign skin samples were studied. Staining of six sections at 10-micron increments was performed on each sample to investigate both the inter- and intra-patient variability. Overall, the percent positive area of both substrates was below 1% in benign skin and variation from section to section was minimal. To survey the baseline levels of NRF2 and CDT1 across a variety of tumor types, we evaluated archival tumors of various types including melanoma, head & neck squamous cell carcinoma, prostate, lung, and breast cancer. A manual ROI was identified by pathologist review to select for tumor area, and semi-automated image analysis was used to measure the percent positive area for NRF2 and CDT1 within the total tumor area. Unlike skin, the tumor samples showed variable levels of NRF2 and CDT1. However, most samples showed a percent positive area less than 40%, indicating the feasibility of detecting a PD effect by comparing pre- and post-dose tumor biopsies. In summary, reliable IHC PD biomarker assays have been established for the NAE inhibitor MLN4924. The assays are currently being used to measure PD effects of NAE inhibition in skin and solid tumors from Phase I trials of MLN4924. Citation Information: Clin Cancer Res 2010;16(7 Suppl):A31