Unveiling cellular dynamics: exploring human disease progression and therapeutic potential through organoid models and single-cell technologies

Human disease progression is a highly dynamic process characterized by cellular-level decision-making. It is crucial to employ appropriate technologies, with spatial and temporal resolution, as well as reliable model systems, to study and understand disease progression dynamics. Over the past decade, the emergence of single-cell technologies has facilitated such studies, enabling high-resolution molecular phenotyping of multicellular systems. Moreover, while much of our current knowledge of disease phenotypes has been derived from research on model organisms, organoid cultures have emerged as a viable alternative to bridge the gap to human systems. By recapitulating the cytoarchitecture and cellular complexity of human tissues, organoids offer an opportunity to investigate human-specific traits and obtain more representative outcomes for therapeutics. In this thesis, we leverage state-of-the-art human organoid models and single-cell transcriptome technologies to characterize the morphological and molecular changes associated with pancreatic cancer progression and acute intestinal inflammation. These two areas represent clinically relevant concerns with unmet therapeutic needs, and patient-specific models might bring new inroads into therapy development. We establish novel multi-lineage organoid models for both disease areas and investigate the onset and progression of abnormal cell states over time. Our focus lies particularly on intercellular communication, and we describe dynamic gene regulatory networks that underlie the observed transitions in cell states.
In the first project, we developed a modular stroma-rich tumoroid culture system that models pancreatic ductal adenocarcinoma (PDAC). This system successfully recreates the interactions between cancer, endothelial, and fibroblast cells, mimicking various aspects of primary tumors. Communication between different cell lineages within the cancer microenvironment can enhance cancer cell behavior and influence therapeutic responses. However, generating a complex cancer microenvironment in vitro has been a significant challenge. Our tumoroids consist of interconnected vessels, desmoplastic fibroblasts, and glandular cancer cell phenotypes that develop over time. By employing time-course single-cell transcriptome measurements, we demonstrate that tumoroid formation activates fibroblasts, leading to alterations in the extracellular matrix composition and the induction of specific signal-response signatures and metabolic changes in cancer cells. We identify Syndecan 1 (SDC1) and Peroxisome proliferator-activated receptor gamma (PPARG) as crucial receptor and metabolic nodes involved in cancer cell response to signals from cancer-associated fibroblasts (CAFs), and we show that blocking SDC1 disrupts cancer cell growth within the tumoroid. Analysis of tumoroids from multiple PDAC patients reveals the coexistence of subpopulations associated with classical and basal phenotypes, as well as the presence of migratory cancer cells characterized by a distinct transcriptional signature related to metastasis. This migration signature develops over time, reflecting a stress response mechanism that correlates with a worse clinical outcome.
In the second project, we turn our attention to the intricate relationship between the immune system and the intestinal epithelium. The intestine, a complex mucosal epithelial tissue responsible for food digestion and nutrient absorption, is a highly regenerating yet vulnerable tissue exposed to microbic flora. Perturbations in the delicate balance between epithelial and tissue-resident immune cells can contribute to autoimmune diseases and cancer. However, the dynamics of this relationship have remained elusive due to the lack of suitable experimental protocols for harvesting and cultivating fragile gut-resident immune cells. In this study, we developed a 3D organoid model that combines human intestinal epithelium with autologous intraepithelial lymphocytes (IELs). This model enables us to characterize IEL populations under homeostatic and activated conditions and uncover the underlying processes and interactions involved in inflammatory responses. Our results demonstrate that IELs naturally integrate into the epithelium and dynamically survey both the organoids and the surrounding extracellular matrix. By performing single-cell transcriptome profiling (scRNA-seq), we identify a differential enrichment of cytoskeletal genes in IELs compared to matched peripheral blood mononuclear cells (PBMCs) and provide an explanation for their increased motility and intrinsic ability to inspect the epithelium. Unlike PBMCs, in vitro IELs exhibit rapid responses to cancer-targeting biologics, which are known to raise safety concerns in the intestine. This led to unwanted inflammation against healthy epithelium, a consistent adverse outcome observed clinically. Through time-course experiments and scRNA-seq profiling, we characterize critical IEL populations, uncovering key state trajectories and interactions that drive activation dynamics and result in adverse effects. We propose the antagonization of rho-associated kinases 1 & 2, key cytoskeletal modulators, as well as tumor necrosis factor alpha (TNFa), as a potential strategy to mitigate drug-induced inflammation.
Taken together, our comprehensive analyses and modular developmental systems serve as powerful tools to explore dynamic cell states and interactions, as well as to pave the way for the discovery of personalized therapies. We illustrate how our innovative models, particularly the immune-competent intestinal organoids, serve as better predictors of immunomodulatory drug responses.