Molecular subtypes of cancer-associated fibroblasts and their regulation of the tumour microenvironment in pancreatic ductal adenocarcinoma.

Abstract

Pancreatic Ductal Adenocarcinoma (PDAC) remains a leading cause of cancer-related deaths, with a 5-year survival rate of less than 10 % in Western countries. Transcriptomic analysis of PDAC in the last decade has unearthed molecular classifiers that has shaped our understanding of inter-tumoural heterogeneity that drives the abysmally poor survival rates associated with this disease. However, expectations that these classifiers would be able to identify targeted therapies based on molecular signatures of surgically relevant candidates unfortunately remains limited due to less than reliable biomarkers for these subtypes. Further, the presence of the highly desmoplastic stroma enriched in cancer-associated fibroblasts (CAFs) and immune cells among others, aggravates treatment options. CAFs make up a significant proportion of the stroma and were once believed to represent a homogeneous population. In the last five years, this misconception has largely been removed by the identification of CAF subtypes through partitioning around medoid (PAM) centroids and single-cell RNA sequencing (scRNA-seq). CAFs are an integral population in the pancreatic tumour microenvironment (TME) and are involved in intricate crosstalk with neoplastic and immune cells. Therefore, they serve as an attractive therapeutic target to devise strategies against PDAC. In my first results chapter, largely through scRNA-seq analysis, I investigate CAF heterogeneity, evolution, and prognosis, and provide mechanistic insight into the role CAF subtypes play in the TME. Specifically, I demonstrate the presence of a pericyte-like CAF subtype present in adjacent normal tissue that is antagonised by myCAFs and provides a physical barrier to prevent invasion and metastasis. I also highlight the presence of three to four main CAF communities through the integration of multiple published CAF transcriptomic signatures and demonstrate the prognostic power of these CAF signatures in over 600 patient samples. In my second results chapter, to identify hetero-cellular relationships, I uncover the variable association of CAF subtypes to cancer subtypes. Further, I demonstrate the suppression of CD8 T cells by myCAFs, the latter of which also demonstrated positive correlation with anti-inflammatory macrophages. Additionally, I evaluate the CAF subtypes in the context of hypoxia and inflammatory cell death, additional hallmarks of PDAC that our lab has shown precedence in in other cancers such as pancreatic neuroendocrine tumours. Importantly, this analysis led to the identification of myCAF subsets through both bulk sequencing data and scRNA-seq analysis. Notably, I imply that there is a myofibroblast and inflammatory spectrum on which myCAFs and iCAFs lie, and only CAFs on the extreme myofibroblast end of this spectrum display immunosuppressive features, suggesting the presence of plasticity or cell states, as opposed to terminally differentiated cell types. I dedicate my final chapter to understanding the modulation of CAF subtypes to therapy. Specifically, I demonstrate myCAFs confer resistance to gemcitabine, which forms part of the standard of care for patients with PDAC, through the suppression of iCAFs and the cytotoxic ability of T and NKT cells. This is also associated with reduction in pro-inflammatory cDC1 cells, a dendritic cell subset. In an alternative treatment strategy, through scRNA-seq analysis and in vivo stimulation of an immune dormant murine orthotopic model using engineered cytokine therapy, I demonstrate the suppression of endothelial cells by myCAFs, implicating this subtype in reduced tumour vasculature. Given a wide range of interactions between CAF subtypes and other TME-based cell types demonstrated in this chapter and results chapter 2, I perform a comprehensive scRNA-seq ligand-receptor analysis in PDAC and validate these findings across different cancer types to identify key targets and druggable candidates.