ETD

• chemoresistance
• early diagnosis
• organoids
• pancreatic ductal adenocarcinoma
• preclinical models
• tumour educated platelets

This thesis presents a dual strategy to address the critical challenges of pancreatic ductal adenocarcinoma (PDAC), specifically early diagnosis and therapeutic resistance, by leveraging advanced molecular profiling and novel preclinical models. Recognizing that the high lethality of PDAC stems from late diagnosis and robust chemoresistance, Part I explored a new avenue for minimally invasive detection by analysing Tumour-Educated Platelets (TEPs). By applying a mutation-driven bioinformatic pipeline to a large TEP-RNA sequencing cohort, we successfully identified a mutational signature capable of distinguishing both PDAC and non-malignant pancreatic conditions from healthy donors. This approach, which focuses on genetic variants and their isoforms, was less susceptible to batch variation than standard expression profiling. Key findings included the clustering of mutated genes into two distinct clusters: a "Tumoral” cluster (containing PDAC drivers such as KRAS and PTEN) and an "Activated Platelets” cluster (related to platelet activation). Crucially, the discovery of shared mutational sites between PDAC and non-malignant samples in genes like SCLT1 suggests the potential for detecting early oncogenic events. Furthermore, for the first time in this field, an in vitro co-culture system was employed to validate that healthy platelets actively acquire tumour-derived RNA, strengthening the biological basis for TEPs as a liquid biopsy source. Part II pivoted to therapeutic resistance, focusing on gemcitabine resistance, the mainstay treatment for most advanced PDAC patients. Transcriptomic analysis across multiple PDAC gemcitabine-resistant models revealed a consistent suppression of deoxycytidine kinase (dCK) activity, the rate-limiting enzyme for gemcitabine activation, thus confirming a mechanism for attenuated DNA damage response in GR cells. Concurrently, we identified the consistent upregulation of APOBEC3G in gemcitabine-resistant models, which was subsequently validated to correlate with shorter overall survival in a clinical cohort of PDAC patients, positioning the APOBEC3 pathway as a significant vulnerability and novel therapeutic target against chemoresistance. Leveraging the dCK deficiency, we tested the alternative nucleoside analog RX-3117, as it bypasses the dCK pathway for activation. RX-3117 was successfully shown to overcome gemcitabine resistance in resistant PDAC models, suggesting a viable alternative for non-responsive patients. Finally, Part III focused on establishing an advanced preclinical platform. Recognizing the limitations of 2D culture for clinical translation, an optimized protocol was developed to culture PDOs in the synthetic, xeno-free hydrogel VitroGel. This step creates a more reproducible and controlled microenvironment, addressing the batch variability issues of Matrigel. The established 3D platform, which faithfully mimics tumour architecture, microenvironment, and cellular heterogeneity, will be used in future studies to validate the synergistic potential of TEP-RNA analysis and to test the efficacy of RX-3117 in a complex, clinically relevant setting, thereby advancing personalized therapeutic strategies for PDAC. This work provides two significant contributions: a highly promising diagnostic modality and a therapeutic alternative strategy in case of resistance. These findings, along with their future integration into our 3D model, will facilitate rapid translation research to improve PDAC patient management.