78P - High resolution in situ analysis reveals phenotypic and functional diversification within single pancreatic tumour ducts

Background

Transcriptomic studies have identified two major subtypes of pancreatic ductal adenocarcinoma (PDAC), i.e. a ‘classical’ and a ‘basal-like’ subtype. These subtypes have differential expression of GATA6, with prognostic and potentially predictive value, and the basal-like subtype has reportedly increased gene dosage of mutant KRAS. Here, we project these findings to the tissue level by high resolution spatial analysis in human PDAC and derived organoids to gain additional biological insights.

Methods

We use formalin-fixed paraffin embedded (FFPE) human PDAC cell lines, tumour organoids and surgical samples that were subjected to hot spot mutation analysis and transcriptomic subtyping by DNA and RNA sequencing. The Basescope assay was optimized for specific in situ detection of KRASpoint mutations in above FFPE samples. BaseScope and RNAScope were combined with multiplex immunostainings and whole tissue sections were analyzed with HALO (Indica™) software.

Results

Apart from inter- and intra-patient heterogeneity, we demonstrate phenotypic diversity of tumour cells within single ducts, revealing spatial phenotypes with varying mRNA expression levels of GATA6 and KRAS^G12D. Novel gene signature-informed mRNA marker panels underscore the identification of co-existing classical and basal-like zones, as well as co-expressor cells, within single tumour duct. These zones were related to functional (proliferation and epithelial to mesenchymal transition) diversity. PDAC organoids recapitulate the single duct phenotypic diversification that can be shifted experimentally by co-culturing with CAFs.

Conclusions

We successfully establish in situ profiling of transcriptomic subtypes and single point mutations and refine our insights into pancreatic tumour cell plasticity. We reveal extensive intra-tumour diversity that will put extra challenges to novel therapeutic approaches.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

VUB and InnoSer Belgie.

Funding

InnoSer Belgie.

Disclosure

All authors have declared no conflicts of interest.