UNDERSTANDING CANCER DRUG RESPONSES BY COMPUTATIONAL ANALYSES OF GENE EXPRESSION SIGNATURES

Introduction: Technological advancements have introduced next-generation sequencing (NGS) platforms in cancer research, enabling large-scale genomic analysis of tumors. NGS allows for comprehensive DNA and RNA analysis, identifying gene mutations, splicing variations, and changes in gene expression. Specifically, RNA sequencing (RNA-Seq) helps detect gene expression and epigenetic changes. These technologies can be applied to various types of cancer. This thesis focuses on rare neuroendocrine tumors, specifically pheochromocytoma (PCC) and paraganglioma (PGL), together PPGLs, and medullary thyroid carcinoma (MTC). Surgical treatment remains the best option for aggressive PPGL and MTC, but it does not provide a definitive cure. For advanced or metastatic tumors, systemic treatment is necessary, but conventional chemotherapy has limited effectiveness. Targeted therapies have shown some improvement in progression-free survival, but they have not significantly increased overall survival, and often associate with severe side effects. This highlights the urgent need for new therapeutic targets that can lead to more effective and less toxic treatments as well as address therapy resistance. Materials and Methods: We employed an in-silico approach, starting with RNA-seq data analysis, to identify new drug combinations for treating PPGLs and MTC. For PPGLs, we used a PI3K inhibitor, a CDK4/6 inhibitor, and a combination of the two to treat rat PC12 cells. RNA was extracted from treated cells. For MTC, two cell lines, MZ and TT, representative of the most common genetic subgroups were treated with cabozantinib (a multi-target kinase inhibitor, mTKI), mon (a compound not previously used in human cancer treatment), and a combination of both. Moreover, to validate in vitro data obtained from PPGLs experiments and from our in-silico analyses, we applied the same bioinformatics approaches to human metastatic PPGLs samples, obtained from Calsina et.al, 2023 and from the public TCGA PCPG database. Following treatment, we conducted differential expression analysis (using DESeq2), pathway enrichment analysis (using GSEA and GO), and a transcription factor enrichment analysis (TFEA). Additionally, we evaluated transcription factor activity using the DecoupleR tool. Results: For PPGL, in silico analysis using treated PC12 cells demonstrated that the combination of PI3K and CDK4/6 inhibition led to significant transcriptional changes, particularly disrupting pathways related to cell cycle regulation and mitotic spindle assembly. Concordantly, in vitro functional assays confirmed that the combination treatment effectively induced cell cycle arrest and impaired cancer cell proliferation. In human metastatic PPGL samples, we observed upregulation of mitotic spindle-related genes, including ASPM, CEP55, and KIF14, which correlated with tumor aggressiveness. The drug combination specifically targeted these upregulated pathways, suggesting its potential efficacy in halting metastatic progression. For MTC, this study investigated the combined effects of cabozantinib and a novel compound, mon, on human cell lines. The combination treatment significantly altered gene expression, especially in MZ cells, showing a pronounced impact on cell cycle pathways. GSEA revealed that the combination therapy downregulated cell proliferation pathways, including the G2M checkpoint and E2F targets, while upregulating stress-related processes like the unfolded protein response. Conclusion: These findings underscore the potential of these drug combinations to enhance treatment efficacy through complementary mechanisms, offering new avenues for targeting aggressive tumors and improving patient outcomes.