M. Rashid1 , J.J. Lim1 , E. Chow2 , L. Wedekan3 . 1 KYAN Technologies Pte Ltd, Not Applicable, Singapore, Singapore; 2 National University of Singapore, Cancer Science Institute, Singapore, Singapore; 3 CELLphenomics GmbH, Not Applicable, Berlin, Germany
Background:
Claudin 1 (CLDN1), a membrane tight junction protein, is often overexpressed in multiple solid tumors including breast, colon, stomach, and liver. Non-junctional CLDN1 has been shown to promote exclusion of T cells and resistance to checkpoint inhibition, which could be reversed upon targeting of CLDN1 using monoclonal antibodies. Transpiring from these results, this project aimed to use Optim.AI® to identify effective optimal drug combinations that include a CLDN1 inhibitor (CLDN1i) against colorectal cancer (CRC). We have previously demonstrated applications of a functional precision medicine platform, Optim.AI® , along the drug development pipeline on hematological malignancies, liver, and colorectal cancer. It rationally converges upon optimal drug combinations within a defined drug search space, without any prior knowledge on the background of the models. This allows the identification of unique, novel combinatorial therapies as well as unconventional targets.
Methods:
Optim.AI® was carried out on eight CRC patient-derived organoids (PDOs) from CELLphenomics GmbH. Drug combinations containing 12 inhibitors, including CLDN1i, were treated on each PDO. Post-drug treatment cell viability was used as a phenotypic input for Optim.AI analysis, which maps experimental data points to a second-order quadratic function to predict and rank cell killing efficacies for all permutations of drug combinations within the 12-drug search space.
Results:
Differential responses were observed with CLDN1i across the different PDOs. The results show a distinct subgroup that demonstrate a greater response to CLDN1i in combination with chemotherapy, namely, 5- FU and Gemcitabine. CLDN1i and Gemcitabine was observed to be more sensitive than the standard of care treatment, FOLFOX (Leucovorin + Irinotecan + 5-FU), for five of the eight PDOs. In addition, Encorafenib and Cetuximab were found to be potential drug combination partners with CLDN1i for some PDOs, with similar responses to the standard of care.
Conclusions:
With the identified subgroups demonstrating better responses of CLDN1i with FOLFOX or Gemcitabine, Optim.AI could be used as a companion diagnostic to potentially aid patient stratification on a larger cohort. Paired with other relevant –omics data, additional analysis could further reveal potential biomarkers for these subgroups.
European Journal of Cancer, Volume 211, Supplement 1, October 2024, 114750