Search

Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy with a 5-year survival rate of 11%. PDAC has limited response to current cytotoxic, targeted and immunotherapies. Thus, it is crucial to explore the molecular underpinnings of this disease in order to identify expression patterns and intercellular dialogue that are limiting the effectiveness of current therapies. Here, we dissected the single-cell spatial transcriptomic landscape of untreated (n=7) and treated (chemoradiation, n=7) primary resected human PDAC tumors using spatial molecular imaging with a custom 990-plex panel. We recovered 719,965 high-quality single cells of 20 distinct types, including epithelial, immune, endothelial, endocrine, and diverse stromal cells. We further dissected the malignant, cancer-associated fibroblasts (CAF), and endothelial cells by applying single-nucleus RNA sequencing (snRNA-seq)-derived signatures. We explored how treatment altered the intraglandular heterogeneity of malignant programs and intratumoral composition of multicellular neighborhoods, for which we developed a Gaussian-based computational model. In concordance with previous studies, we found that myofibroblastic CAFs spatially localized with malignant cells. Treatment-enriched neural-like progenitor malignant cells spatially localized with CD8+ T cells, which combined with other findings suggests that different malignant subtypes may employ distinct mechanisms of immune evasion. We further investigated receptor-ligand interactions associated with therapeutic resistance at subcellular spatial resolution. To this end, we developed an optimal transport-based computational method to infer cell-cell communication at the cancer-stromal interface. We functionally validated a subset of predicted interactions using a murine stromal tumoroid model, in which we co-cultured transformed pancreatic ductal cells and CAFs in a solubilized extracellular matrix. We evaluated the strength of predicted interactions in response to ex vivo chemotherapy using snRNA-seq. Moreover, we modulated the expression of select ligand-receptor pairs in the malignant-CAF co-cultures using CRISPR-mediated genetic engineering. We tested the effects of these perturbations on therapeutic resistance using cell viability assays, including live-cell label-free phase nanoscopy. In conclusion, we integrated single-cell spatial transcriptomics and novel computational methods to identify mechanisms of therapeutic resistance in PDAC at both cellular and molecular levels. We functionally validated a subset of cancer-stromal receptor-ligand interactions predicted to associate with treatment resistance using genetically-engineered stromal tumoroids, demonstrating the translational potential of spatial biology for guiding therapeutic development. Citation Format: Carina Shiau, Jingyi Cao, Mark Gregory, Tae Kyung Kim, Youngmi Kim, Jason W. Reeves, Peter L. Wang, Steven Wang, Nicole A. Lester, Jennifer Su, Jimmy Guo, Joseph M. Beechem, David T. Ting, Martin Hemberg, William L. Hwang. Mechanisms of therapeutic resistance in pancreatic cancer identified by single-cell spatial transcriptomics and genetically-engineered stromal tumoroids [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1165.

Pancreatic ductal adenocarcinoma (PDAC) features abundant perineural invasion (PNI). Intra-tumoral nerves play critical roles in cancer initiation, progression, recurrence, treatment-resistance, metastasis, and mortality for many malignancies but the diverse molecular mechanisms underlying tumor-nerve crosstalk remain largely unknown—hindering the development of therapies targeting this key pathological process. To address this gap, we performed whole transcriptome digital spatial profiling on twelve custom tissue microarrays (n=288 cores) derived from intratumorally-matched regions with and without PNI in primary PDAC specimens (n=31 patients) and independently measured gene expression from cancer cells, fibroblasts, and nerves. We undertook a differential gene expression analysis comparing malignant cells in PNI-present and -absent regions. We quantified the growth kinetics of dorsal root ganglia (DRG) sensory neurons cultured with exogenous candidate proteins, which validated that some candidates augment (e.g., Lgals1) and others inhibit (e.g., Sema3b) neurite outgrowth. Next, we mapped our previously discovered malignant cell programs onto the epithelial segments and observed a significant enrichment of the mesenchymal, basal-like, and neural-like progenitor (NRP) programs versus depletion of the classical program in PNI-present regions. To determine the effects of malignant subtype on nerve outgrowth, we engineered isogenic KrasG12D/+;Trp53FL/FL;Rosa26-dCas9-VPR (KP;dCas9-VPR) organoids to overexpress the master transcription factors (TFs) for each malignant subtype (e.g., Gata6 for classical, Glis3 for NRP). We then performed the DRG neuronal outgrowth assay using conditioned media from each subtype-specific organoid line versus an off-target control and observed that the classical line suppressed neurite outgrowth, the mesenchymal and basal-like lines were neutral, and the NRP line enhanced neurite outgrowth dynamics comparable to the Ngf positive control. Taken together, our findings suggest that the mechanisms underlying nerve recruitment/outgrowth and perineural invasion may be partly decoupled. To further test this hypothesis, we are performing transwell invasion assays comparing KP;dCas9-VPR cancer cell lines that overexpress each of the candidate PNI-associated genes and malignant subtype TFs. We anticipate that this study will transform our understanding of how cancer cells and the peripheral nervous system collaborate and guide prioritization for therapeutic intervention in the burgeoning cancer neuroscience field. Citation Format: William L. Hwang, Jennifer Su, Carina Shiau, Peter L. Wang, Jimmy A. Guo, Nicole A. Lester, Jaimie L. Barth, Hannah I. Hoffman, Prajan Divakar, Jason W. Reeves, Eric Miller, Joseph M. Beechem, Andrew J. Aguirre, Lei Zheng, David T. Ting, Mari Mino-Kenudson, Tyler Jacks. Distinct cancer-intrinsic mechanisms mediate nerve recruitment/outgrowth versus perineural invasion [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2513.