Analysis of single nuclear chromatin accessibility reveals unique myeloid populations in human pancreatic ductal adenocarcinoma

Abstract Background A better understanding of the pancreatic ductal adenocarcinoma (PDAC) immune microenvironment is critical to developing new treatments and improving outcomes. Myeloid cells are of particular importance for PDAC progression; however, the presence of heterogenous subsets with different ontogeny and impact, along with some fluidity between them, (infiltrating monocytes vs. tissue‐resident macrophages; M1 vs. M2) makes characterisation of myeloid populations challenging. Recent advances in single cell sequencing technology provide tools for characterisation of immune cell infiltrates, and open chromatin provides source and function data for myeloid cells to assist in more comprehensive characterisation. Thus, we explore single nuclear assay for transposase accessible chromatin (ATAC) sequencing (snATAC‐Seq), a method to analyse open gene promoters and transcription factor binding, as an important means for discerning the myeloid composition in human PDAC tumours. Methods Frozen pancreatic tissues (benign or PDAC) were prepared for snATAC‐Seq using 10× Chromium technology. Signac was used for preliminary analysis, clustering and differentially accessible chromatin region identification. The genes annotated in promoter regions were used for Gene Ontology (GO) enrichment and cell type annotation. Gene signatures were used for survival analysis with The Cancer Genome Atlas (TCGA)‐pancreatic adenocarcinoma (PAAD) dataset. Results Myeloid cell transcription factor activities were higher in tumour than benign pancreatic samples, enabling us to further stratify tumour myeloid populations. Subcluster analysis revealed eight distinct myeloid populations. GO enrichment demonstrated unique functions for myeloid populations, including interleukin‐1b signalling (recruited monocytes) and intracellular protein transport (dendritic cells). The identified gene signature for dendritic cells influenced survival (hazard ratio = .63, p = .03) in the TCGA‐PAAD dataset, which was unique to PDAC. Conclusions These data suggest snATAC‐Seq as a method for analysis of frozen human pancreatic tissues to distinguish myeloid populations. An improved understanding of myeloid cell heterogeneity and function is important for developing new treatment targets in PDAC.