Pancreatic ductal adenocarcinoma (PDAC) remains a treatment-refractory disease and existing molecular subtypes do not inform clinical decisions. Previously-identified bulk transcriptomic subtypes of PDAC were derived exclusively in the treatment-naïve context and several are unintentionally driven by “contaminating” stroma. RNA extraction from pancreatic tissue is difficult and prior single-cell RNA-seq efforts have been limited by suboptimal dissociation/RNA quality and poor performance with the trend towards increased use of neoadjuvant chemoradiotherapy (CRT). To address these limitations, we developed a robust single-nucleus RNA-seq (sNuc-seq) technique compatible with frozen archival PDAC specimens (both naïve and treated) and computational techniques to identify the transcriptomic programs driving intratumoral heterogeneity and therapeutic resistance.
Single nuclei suspensions were extracted from flash frozen primary PDAC specimens and processed with the 10x Genomics Single Cell 3’ v3 pipeline and gene expression libraries were sequenced (Illumina HiSeq 2500). Combined treatment-naïve (n=13) and CRT-treated (n=12) specimens yielded approximately 150,000 high-quality nuclei. In each tumor, distinct clusters with gene expression profiles consistent with acinar, ductal, fibroblast, endothelial, endocrine, lymphocyte, and myeloid populations were identified. Malignant cells were confirmed by inferred copy number variation analysis (InferCNV v3.9) and segregated into several distinct programs by non-negative matrix factorization (NMF) for each individual patient highlighting intratumoral heterogeneity. Previously identified basal-like and classical-like programs were identified in a subset of treatment-naïve cells but not CRT-treated cells. In the treatment-naïve cohort, we identified additional NMF programs that were distinct from the basal-like and classical-like signatures (neuronal, epithelial-mesenchymal transition, anabolic, catabolic, cycling). In CRT-treated specimens, there was enrichment of catabolic and epithelial-mesenchymal transition programs and several unique programs were also uncovered (invasive stem-like, neuronal stem-like, and stress signaling). Deconvolution of clinically-annotated bulk RNA-seq cohorts using our sNuc-seq signatures and characterization of intercellular interactions with receptor-ligand analysis, spatial transcriptomics (Nanostring GeoMx™ Digital Spatial Profiler), and multiplexed immunofluorescence are ongoing.
GeoMx DSP is for Research Use Only and not for use in diagnostic procedures.