An oncogenic KRAS-driven secretome involving TNFα promotes niche preparation prior to pancreatic cancer onset
Pancreatic ductal adenocarcinomas (PDACs) are highly lethal and aggressive with oncogenic KRAS being the main oncogenic driver of the disease. PDACs have been extensively profiled at advanced stages, and in advanced disease the tumor microenvironment is a major determinant that critically shapes patient outcomes. Since the molecular events occurring prior to invasive growth remain poorly understood, we aimed to investigate changes in the precancerous epithelium and its surrounding niche. We acquired time-resolved, single-cell transcriptomic (scRNAseq), and accessible-chromatin data from human pluripotent stem cell-derived pancreatic duct-like organoids (PDLO) inducibly expressing KRASG12D and from various niche cells. Analysis of the pure epithelium already revealed key signatures of matrix remodeling and inflammation-related signaling upon few days of KRASG12D expression. Machine learning captured KRASG12D-dependent transcriptomic classifiers with high prediction accuracy and niche preparatory relevance. Various co-culture approaches followed by scRNAseq and functional validation, including T-cell microfluidics, demonstrated that the KRASG12D-induced PDLO-secretome activates pancreatic stellate cells (PaSCs) and protects precancerous organoids from T cell infiltration. Additional, in silico approaches reconstructed a virtual pancreatic (pre)cancerous space to profile cell–cell interactions between PDLOs and niche cells. TNFα emerged as a top-ranked ligand and was functionally validated to mediate T-cell shielding and PaSC activation. Cyst fluid from 80 prospectively sampled Intraductal Papillary Mucinous Neoplasm (IPMNs) –well-known cystic PDAC precursor lesions– showed a stepwise TNFα rise across LGD (low-grade), HGD (high-grade), and IC (invasive cancer). Our study reveals that oncogenic KRAS orchestrates niche-preparatory programs that precede PDAC formation and highlight a T cell exclusion program governed by epithelial-derived TNFα.
Authors
- Markus Breunig (ORCID: https://orcid.org/0000-0002-2817-1974)
- Sandra Heller (ORCID: https://orcid.org/0000-0002-8704-2646)
- Medhanie Mulaw (ORCID: https://orcid.org/0000-0002-2501-6952)
- Élodie Roger (ORCID: https://orcid.org/0000-0002-4930-6292)
- Meike Hohwieler (ORCID: https://orcid.org/0000-0002-0640-0764)
- Sandra Wiedenmann (ORCID: https://orcid.org/0000-0001-8160-0486)
- Peter Loskill (ORCID: https://orcid.org/0000-0002-5000-0581)
- Jessica Lindenmayer (ORCID: https://orcid.org/0009-0001-3673-5995)
- Chantal Allgöwer
- Stefano Crippa
- Benedetta Ferrara (ORCID: https://orcid.org/0000-0002-2569-5115)
- Paul Lopatta (ORCID: https://orcid.org/0000-0001-6604-774X)
- Tengku Ibrahim Maulana (ORCID: https://orcid.org/0000-0002-8661-6871)
- Ivan G. Costa (ORCID: https://orcid.org/0000-0003-2890-8697)
- Anton Lahusen
- Dharini Srinivasan (ORCID: https://orcid.org/0009-0007-7968-9240)
- Claudia Teufel (ORCID: https://orcid.org/0000-0001-8423-8400)
- Eleni Zimmer (ORCID: https://orcid.org/0000-0002-2510-3140)
- Paolo Riccardo Camisa
- J.-Mathias Löhr
- Nadine Therese Gaisa
- Leon Reichardt
- Yuan-Na Lin
- Michael K. Melzer
- Alexander Kleger
- Luisa T. Ferreira
- Stefan Liebau
- Lorenzo Piemonti
- Annika Beck
- Adam Flinders
- Thomas Seufferlein
- Matthias Meier
- Claus Jorgensen
- Elena Anne Ringel
Institutions
- Vita-Salute San Raffaele University (IT)
- Universität Ulm (DE)
- Karolinska Institutet (SE)
- Helmholtz Zentrum München (DE)
- Natural and Medical Sciences Institute (DE)
- Cancer Research UK Manchester Institute (GB)
- IRCCS Ospedale San Raffaele (IT)
- Bernstein Center for Computational Neuroscience Tübingen (DE)
- University Hospital Ulm (DE)
- Pancreas Centre (Canada) (CA)
- University of Tübingen (DE)
- RWTH Aachen University (DE)
- Leipzig University (DE)
Publication Details
- Journal
- Molecular Cancer
- Published
- 2026-02-03
- DOI
- https://doi.org/10.1186/s12943-025-02541-1
- Primary Topic
- Single-cell and spatial transcriptomics
- Type
- article
- Field-Weighted Citation Impact
- 0.00