Hypoxic fibroblasts produce cytokines that drive ductal cell epithelial-mesenchymal transition in pancreas cancer

Introduction: The pancreatic ductal adenocarcinoma (PDAC) tumor microenvironment is poorly vascularized (creating regions of hypoxia) and densely populated by cancer-associated fibroblasts (CAFs). While hypoxia and CAFs have independently been shown to promote PDAC progression [1], the possibility of their cooperative effects remains largely unexplored. One potential route for cooperation is through the ability of hypoxia, via a yet unknown pathway, to promote the inflammatory CAF (iCAF) phenotype [2]. CAFs and hypoxia independently support ductal cell (i.e., cancer cell) epithelial-mesenchymal transition (EMT) [3,4], a process that increases chemoresistance and invasiveness. We hypothesize that hypoxia promotes an iCAF-like phenotype and stimulates CAFs to secrete growth factors capable of inducing EMT in epithelial-derived cancer cells through signaling pathways that can be pharmacologically targeted. Testing this hypothesis and identifying the specific pathways involved may reveal new therapeutic vulnerabilities in PDAC.

Materials and

Methods: Publicly available single-cell RNA sequencing (scRNA-seq) data from human PDAC patients [5] were used to predict CAF-ductal cell ligand-receptor interactions that promote ductal cell EMT using CellPhoneDB and NicheNet. A high-throughput screen of 430 inhibitors was used to nominate EMT-driving growth factors secreted by hypoxic CAFs, with inferences from the screen tested in ductal cells using exogenous growth factors or CAF-conditioned medium and receptor kinase inhibitors. The hypothesized hypoxia-regulated mechanism in CAFs was investigated using iterative indirect immunofluorescence imaging (4i) to measure signaling pathway activation, epigenetic modifications, and growth factor expression simultaneously. shRNA-mediated knockdown of pathway intermediates was performed to validate the proposed mechanism. Tumors from the KPCY (KrasG12D, Pdx1-Cre, Tp53loxP/-, RosaYFP) autochthonous mouse model of PDAC were stained with antibodies to measure signaling markers in CAFs present in hypoxic regions. CAFs engineered with a hypoxia fate-mapping reporter were subcutaneously implanted with PDAC cells in mice, and the resulting tumors were stained using an adapted 4i approach to generate spatially resolved data that were used for cell type proximity analyses.

Results, Conclusions, and Discussions: Patient tumor scRNA-seq data revealed that iCAFs are enriched in hypoxia-regulated genes and express specific growth factors predicted to promote ductal cell mesenchymal gene expression. Conditioned medium from multiple iCAF lines consistently increased expression of the mesenchymal marker vimentin in PDAC cell lines. CAFs cultured in 1% O2 exhibited an enhanced ability to drive ductal cell EMT compared to those cultured in 21% O2 (Figure 1A). Based on scRNA-seq analysis and the inhibitor screen, ligands for two receptor tyrosine kinases were identified as EMT-promoting growth factors secreted by hypoxic CAFs. Expression of those cytokines by hypoxic CAFs depends on the coincident activation of NF-κB signaling in response to reactive oxygen species accumulation and a histone methylation mark. Spatial analysis of highly multiplexed images of tumors from an autochthonous mouse model of PDAC revealed that hypoxic cancer cells proximal to hypoxic CAFs displayed increased vimentin expression, suggesting that hypoxic CAFs actively support EMT over a length scale consistent with paracrine signaling in vivo (Figure 1B, C). Tumors generated from hypoxia fate-mapped CAFs and PDAC cells corroborated these findings and further suggested that the EMT-supportive effects of hypoxia on CAFs are reversible after once-hypoxic CAFs no longer occupy hypoxic regions of tumor tissues. Our findings indicate that hypoxia is a key regulator of CAF heterogeneity in PDAC and modulates paracrine signaling between CAFs and cancer cells to promote EMT. This work highlights the opportunity to identify new drug targets in PDAC (and other cancers) by studying interactions between tumor microenvironmental factors and transformed cells.

Acknowledgements (Optional): This work was supported by the NCI U01 CA243007 (MJL), NCI 5T32CA0091909, and NCI P30 CA044579.

References: [1] Bulle & Lim, Signal Transduction & Targeted Therapy, 2020. [2] Schworer et al., Cancer Research, 2023. [3] Chen et al., Nature Communications, 2022. [4] Brown et al. Cancer Research, 2024. [5] Elyada et al., Cancer Discovery, 2019.