NEW YORK – A team from Massachusetts General Hospital, the Broad Institute, and elsewhere has used single-cell spatial transcriptomics and computational approaches to characterize shifts in the tumor microenvironment (TME) that coincide with treatment response in pancreatic ductal adenocarcinoma (PDAC).
"This work provides a broad translational approach that leverages single-cell spatial transcriptomics to better understand baseline multicellular neighborhoods and interactions, and remodeling of these dynamic relationships under perturbative selection pressure," co-senior and co-corresponding authors William Hwang, a researcher affiliated with Mass General, the Broad, and the Massachusetts Institute of Technology, and Martin Hemberg, an investigator with the Broad and Brigham and Women's Hospital, and their colleagues wrote in a paper published in Nature Genetics on Tuesday.
Using a "spatially constrained optimal transport interaction analysis" (SCOTIA) model that combines single-cell spatial transcriptomic profiles with data from preclinical models, the researchers retraced features of the microenvironment that are associated with primary PDAC cases, including half a dozen cases treated with neoadjuvant chemotherapy and seven treatment-naïve cases.
"Cell-cell interactions are known to play an important role in the TME, and we hypothesize that treatment-induced remodeling will involve changes in [ligand-receptor] interactions," the authors explained, adding that "advances in spatial transcriptomics have enabled subcellular resolution while profiling hundreds to thousands of distinct transcripts, offering an unprecedented opportunity to understand cell-cell interactions that underpin the pre- and post-treatment TME in their in situ context."
Based on NanoString Technologies GeoMx digital spatial profiling and single-nucleus RNA sequencing data on nearly 717,500 cells from 13 formalin-fixed paraffin-embedded PDAC samples, together with ligand-receptor interaction scores produced by spatial molecular imaging, the team followed cellular interaction shifts and multicellular tumor microenvironment features linked to treatment response.
The researchers saw more pronounced ligand-receptor interactions between cancer-associated fibroblasts and malignant cells in the TMEs from yet-to-be-treated PDAC cases, where 60 ligand-receptor pairs turned up, compared to the two dozen ligand-receptor pairs detected after chemotherapy.
Along with other chemotherapy-related signaling changes, the investigators uncovered ties between chemotherapy resistance and higher-than-usual levels of interleukin-6 family signaling, affecting interactions between tumor cells and cancer-associated fibroblast cells. They validated these treatment-associated TME shifts subsequently in ex vivo mouse tumoroid co-culture and human cell line experiments.
"[W]e integrated experimental and computational approaches to enable high-resolution, spatially guided discovery of treatment-associated remodeling in the pancreatic microenvironment, which led to the identification and validation of IL-6 signaling between [cancer-associated fibroblasts] and cancer cells as a mechanism of chemoresistance," the authors reported.
Based on these and other findings, they suggested altered TME interactions can offer insights into treatment resistance and argued that similar spatial analyses "can be broadly applied to other contexts."