Pancreatic cancer (PC) is the deadliest among gastrointestinal cancers due to the lack of symptoms till an advanced stage and no screening options. However, PC seems to be preventable to a certain degree. One of the defense mechanisms that prevent carcinogenesis is the Nrf2 pathway. It induces an antioxidant response and is involved in the cytoprotection of cells. Nevertheless, studies suggest that Nrf2 overactivation can promote tumor growth. Hence, the consequences of the Nrf2 modulation are context-specific and need to be better understood. We investigated three natural compounds found in the human diet, resveratrol, genistein, and isoliquiritigenin, in pancreatic ductal adenocarcinoma (PDAC) cell lines (CFPAC-1, MIA PaCa-2, and BxPC-3) in terms of their anticancer and Nrf2 modulatory activities, but none of the compounds showed significant effects. Moreover, the Nrf2 pathway was initially activated in CFPAC-1 and MIA PaCa-2 cells, suggesting that the inhibition and not activation of Nrf2 could be a strategy against PC. The second part of the thesis focused on the deadliest malignancy — lung cancer. There is growing evidence that the circadian rhythm is implicated in its development. Therefore, we investigated this interaction to look for new diagnostic and therapeutic options. The gene expression analysis of 132 paired tissue and benign human lung samples representing five histotypes (adenocarcinoma, squamous cell carcinoma, large cell neuroendocrine carcinoma, carcinoid, and small cell lung cancer) demonstrated that core clock genes were vastly dysregulated in lung tumors in a histotype-specific manner. However, the study of human lung cancer cell lines representing adenocarcinoma (A549 and A427), squamous cell carcinoma (H1264), and carcinoid (H727) subtypes, as well as primary adenocarcinoma tumor and non-tumor cells showed that they all exhibit self-sustained circadian oscillations of two clock gene reporters, Bmal1-luc and Per2-luc. Yet, the circadian parameters differed between tumor and non-tumor primary cells suggesting that the circadian clock is disturbed in lung cancer. Importantly, CLOCK silencing inhibited cell growth and increased cell death in A549 cells. Altogether, these results suggest that clock genes may be used as diagnostic markers for identifying lung cancer subtypes as well as serve as therapeutic targets. Following the cytotoxic effect of clock disruption through CLOCK silencing, we searched for clock-modulatory compounds that could display a similar cytotoxic effect in lung cancer cells. Chir99021, GO289, KL001, KS15, longdaysin, nobiletin, and SR9011 were screened in A549 cells. Nevertheless, most of the selected compounds showed relatively low cytotoxicity. Therefore, further compounds should be evaluated. Finally, we developed two resveratrol-based dry powder formulations intended for pulmonary administration using an inhaler to assess its chemopreventive activity against lung cancer. The formulations will be first tested in mice. The best formulation will be afterward tested in phase I clinical trial to assess its toxicity in healthy participants. Eventually, such a formulation could be used as a chemopreventive agent by people at risk or as an adjunct to standard cancer treatments. Altogether, this work contributed to the expansion of knowledge on the molecular biology of pancreatic and lung cancers. Further studies will be carried out to translate the findings into clinical practice.