(1088-A) Optimization of DNA methyltransferase inhibitors-based combination through quadratic phenotypic optimization platform (QPOP) to target the epigenome of gastric cancer

Abstract: Aberrant DNA methylation is one of the critical epigenetic modifications that has become a widespread phenotype in solid tumours. Similar to haematological malignancies, many studies have shown that the DNA methylation landscape of solid tumours exhibits a pattern of either hypermethylation of the promoter regions of tumour-suppressive genes in advanced cancer or global hypomethylation in the early stages of carcinogenesis. Besides that, dysregulated DNA methylation seems to remodel the tumour microenvironment and affect immune cells' response to immune checkpoint inhibitors. In gastric cancer, comprehensive DNA methylation analyses have demonstrated that the majority of the gastric tumours carry the CpG island methylation phenotype (CIMP), resulting in extended molecular-based subgroups such as the extremely high-methylation epigenotype (E-HME), HME, and low-methylation epigenotype (LME). Furthermore, the pathogenic causes of gastric cancer, H.pylori and Epstein-Barr Virus are also known to induce aberrant DNA methylation. Therefore, current standard-of-care (SOC) treatments are insufficient to address the multifaceted epigenome of gastric cancer. This shows the therapeutic potential of using DNA hypomethylating drugs like DNA methyltransferase inhibitors (DNMTi) for solid tumours with high methylation epigenotypes such as gastric cancer.
In our study, we have carefully curated a 10-drug set list from a pool of FDA-approved drugs that are commonly used for gastric cancer, including the DNTMi, Decitabine and 5-Azacytidine. Thereafter, we exploited the experimental-computational hybrid platform, Quadratic Phenotypic Optimization Platform (QPOP) to identify top-ranking DNMTi-based drug combinations across gastric cancer cell lines (GC-CL) and patient-derived organoids [(GC-PDO), generated from patients in ClinicalTrials.gov: NCT04611035]. Based on QPOP analysis, we showed that Docetaxel/5-Azacytidine (Doc/5-Aza), a non-SOC drug combination is frequently top-ranking. In addition, we also found that the non-CIMP GC-CL is more sensitive towards monotherapy of Docetaxel compared to CIMP-high GC-CL. Interestingly, our findings suggest that the CIMP-high gastric cancer cells are more susceptible to DNMTi possibly through the specific modulation of DNMT1 levels or DNMT-targeted/related proteins to reverse certain genes that are caused by abnormal DNA methylation epigenotype. This clearly reflects the specificity of DNMTi towards gastric cancer cells that possess high methylation epigenotype. As we further validate the effects of Doc/5-Aza by comparing CIMP-high and non-CIMP GC-CL via the Infinium methylation array, we observed robust gene methylation changes in a series of genes that are commonly involved in non-canonical Wnt/ROR Signaling and EMT such as DKK1, FZD5, SOX17, AXIN1 and etc. Our early findings provide preliminary mechanistic insights into using DNMTi to sensitize CIMP-high gastric tumours towards Taxanes by acting on the non-canonical Wnt pathway and EMT activators. Collectively, our work continually emphasized the importance and the power of the QPOP to uncover epigenetic targeted-based drug combinations towards specific disease indications such as gastric cancer in a clinically relevant manner.