henhao
Sign in
Cancer BiologyCell
Peer-reviewed journal article

Oncogenic and tumor-suppressive forces converge on a progenitor niche at the benign-to-malignant transition

Jose ReyesiD, Isabella S. Del PrioreiD, Andrea C. ChaikovskyiD, Nikhita Pasnuri, Ahmed M. ElhossinyiD, Jin Park, Philipp WeileriD, Tobias Krause, Andrew R Moorman, Catherine Snopkowski, Meril T. TakizawaiD, Cassandra BurdziakiD, Nalin RatnayekeiD, Ignas MasilionisiD, Yu-Jui HoiD, Ronan ChalignéiD, Paul Bernard RomesseriD, Aveline A. FillioliD, Tal NawyiD, John P. MorrisiD, Zhen Zhao, Marina Pasca di MaglianoiD, Direna Alonso‐CurbeloiD, Dana Pe’eriD, Scott W. LoweiD

Abstract

The benign-to-malignant transition is a defining step in cancer progression. To investigate when and how malignancy initiation occurs and tissue reorganization proceeds, we combine single-cell and spatial transcriptomic profiling in mouse models of pancreatic ductal adenocarcinoma (PDAC) that capture spontaneous p53 loss. Among Kras-mutant cells, we find that oncogenic and tumor-suppressive programs, including those controlled by p53, CDKN2A, and SMAD4, are co-activated in a discrete progenitor-like population, engaging senescence-like responses. Using a framework we develop for spatial analysis, we show that a niche centered on these cells undergoes stepwise remodeling during tumor progression, mirroring invasive PDAC. Transient KRAS inhibition depletes progenitor-like cells and dismantles their niche, delaying malignancy initiation. Conversely, p53 suppression enables progenitor cell expansion, epithelial-mesenchymal reprogramming, and immune-privileged niche formation. These findings position the progenitor-like state at the convergence of cancer-driving mutations, plasticity and tissue remodeling, revealing a critical window for intercepting malignancy.

PDFDOICode
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
11
FWCI
29.61
Field percentile
top 1%
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (15 Apr 2026) · CC BY 4.0

In mouse pancreatic cancer models, p53, CDKN2A and SMAD4 programs converge on a rare Kras-mutant progenitor-like state whose niche KRAS inhibition collapses and p53 loss expands.

How genetic lesions, cell-state changes and microenvironment remodeling jointly trigger the benign-to-malignant switch is poorly understood, partly because early events are hard to capture. This work uses spontaneous p53-loss reporter mice with single-cell and spatial transcriptomics to localize tumor-suppressor action to one progenitor-like population and its niche.

Key results

  1. ShownAmong premalignant Kras-mutant epithelial states, progenitor-like cells uniquely showed strong p53-dependent target expression, higher Cdkn2a (both p19ARF and p16INK4A isoforms), elevated TGFβ/SMAD4-dependent programs and senescence signatures, alongside RAS, glycolysis and EMT programs.

    Progenitor-like cells are rare (1–2% of premalignant cells). p53 targets (Cdkn1a, Ccng1, Bbc3, Bax) fell after p53 loss in this state while most other states changed little; transcriptionally closest to PDAC by diffusion distance. No effect sizes are given in the text. · Capturing cell states after spontaneous p53 loss; Rare progenitor-like cells exhibit peak…; Dimensionality reduction, batch correction and…

  2. ShownA spatial niche framework ordering epithelial-anchored neighborhoods by gastric-to-progenitor pseudotime showed stepwise remodeling: Itgax+ macrophages, Tnc/Postn myofibroblasts and immunosuppressive myeloid programs (Spp1, Arg1, Il1b) accumulate as progenitor-like cells dominate.

    Xenium 480-gene panel on KC mice: 10 mice at 1–2 days and 5 mice at 3 weeks after caerulein, over 3.5 million cells. Niches are cells within a 60-μm radius of epithelial anchors; trends were robust across 30 μm to 200 μm radii. · Adoption of progenitor-like identity is coupled…; Niche pseudotime reconstructs assembly of a…

  3. ShownTransient KRAS^G12D inhibition (MRTX1133) preferentially depleted progenitor-like cells and collapsed their niche; a brief 72-hour pulse in early neoplasia significantly delayed PDAC onset.

    scRNA-seq: 24-fold depletion of progenitor-like cells vs 10-fold (cycling) and 8-fold (gastric pit-like). Cleaved caspase 3 in 44% of progenitor-like cells vs less than 6% of other types. Xenium 2,686,667 cells, MRTX1133 n = 4 vs vehicle n = 2. · Oncogenic KRAS inhibition dismantles the…; Quantification of cleaved caspase 3 after…

  4. ShownInducible p53 knockdown prevented resolution of the injury-induced progenitor-like state, driving a more mesenchymal 'progenitor 2' state and large, immune-privileged niches enriched in Itgax+/Cd274-high macrophages.

    KC^shp (9 samples) vs KC^shCtrl (5), 4,611,972 Xenium cells. Multiome (n = 9,081 cells) plus integration gave 138 genes with p53-motif accessibility and p53-dependent expression; 84 were upregulated on p53 restoration in vitro and 54 were progenitor-specific. · p53 enforces resolution of progenitor-like…; p53 naturally collapses the progenitor niche

  5. SuggestedA rare KRT17+ basal-like epithelial population in human pancreatitis and tumor-adjacent tissue sits in a layered niche resembling the mouse progenitor niche, with senescence and ARG1/TNC markers and apparent T-cell exclusion.

    67 tissue cores (27 non-neoplastic, 40 PDAC tumor-adjacent) by multiplexed IF; KRT17+ neighborhoods enriched for p53, p16INK4A, CRYAB, HMGA1, TNC, ARG1, pERK, KI67 and depleted of CD8a and GZMB. No effect sizes in the text. · Rare communities in human pancreatitis resemble…; Local neighborhood composition around KRT17+ or…; Distance-based quantification of marker gene…

Methods, data and reproducibility

scRNA-seq, Xenium spatial transcriptomics (480-gene panel) and multiome profiling of KP^LOH and KC^shp mouse pancreas, with an epithelial-anchored niche pseudotime framework, plus KRAS inhibition, p53 knockdown and macrophage depletion perturbations.

  • KP^LOH scRNA-seq of pre-tumor p53-proficient, pre-tumor p53-deficient and tumor p53-deficient epithelial cells · Mus musculus · n = 4 pre-tumor p53-proficient, 6 tumor p53-deficient samples (DE subset) · GSE315243
  • Xenium spatial profiling of KC mouse pancreas 1–2 days and 3 weeks after caerulein pancreatitis · Mus musculus · n = 10 mice (1–2 days), 5 mice (3 weeks); over 3.5 million cells · GSE315243
  • Xenium of premalignant pancreas after MRTX1133 or vehicle (48 h) · Mus musculus · n = 4 MRTX1133 and 2 vehicle mice; 2,686,667 cells · GSE315243
  • Xenium of KC^shp and KC^shCtrl pancreas 3 weeks after pancreatitis · Mus musculus · n = 9 and 5 samples; 4,611,972 cells · GSE315243
  • Single-cell RNA+ATAC multiome of p53-proficient Kras-mutant epithelium one day after injury · Mus musculus · n = 9,081 cells · GSE315243
  • Multiplexed IF of human pancreatitis tissue array (mild, acute, chronic; non-neoplastic and tumor-adjacent) · Homo sapiens · n = 67 tissue cores
  • Reanalyzed scRNA-seq of pancreatic epithelium from cancer-free warm-autopsy donors · Homo sapiens · GSE226829
  • Published injury-induced tumorigenesis scRNA-seq · Mus musculus · GSE207920
Methods & tools
scanpy, PhenoGraph, batchelor fastMNN, custom inferCNV, diffxpy, DESeq2, miloR, SEACells, ArchR, HOMER, Cellpose-SAM, Calligraphy
Code
https://github.com/dpeerlab/p53-niche-dynamics
Reagents & environment
scanpy v1.9.1, DESeq2 v1.42.1, diffxpy v0.7.4, miloR, SEACells v0.3.3, ArchR v1.0.3, Cellpose-SAM v4.0.7, Cell Ranger 9.0.0, lifelines v0.30.0.
Reproducibility gaps
Raw and processed imaging data, plus engineered mESCs and organoid lines, are available only on request from the Lead Contact.
Limitations (6)
  • Wild-type p53 loss was modeled by knockout or knockdown; whether missense mutant p53 proteins with gain-of-function activity further modulate progenitor states and niches was not tested. · stated by the authors
  • The data show that p53 eliminates the progenitor-like state but do not distinguish apoptosis, senescence followed by immune clearance, or differentiation; fate mapping is needed. · stated by the authors
  • KRAS inhibition and anti-CSF1R treatment also affect epithelial populations outside the progenitor niche, so the perturbations are not specific to this state. · stated by the authors
  • Experiments used mostly male mouse cohorts from male mESCs, so sex effects on the in vivo results are unknown. · stated by the authors
  • The Xenium KRAS-inhibition comparison relies on only two vehicle mice, and IF tests used groups as small as n = 3, limiting statistical power. · noted in this summary
  • The human evidence is cross-sectional, correlative tissue-core staining using KRT17 as a surrogate anchor, with KRAS status unknown, so niche causality is untested in human tissue. · noted in this summary

Who should read:Read this if you study early pancreatic tumorigenesis, spatial transcriptomics niche analysis, or how p53 and KRAS shape epithelial plasticity and tissue remodeling.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISTonelli '21Hur '23Twardowski '24Zhang '26DelGiorno '19English '25
73% semantic matchCancer Biology
C. Tonelli, A. Deschenes, M. A. Yao+2 more·bioRxiv·Dec 31, 2021