Investigators applied single-nuclei transcriptomics alongside spatial transcriptomics and ecological dynamic modelling to chart cell-state changes during inflammatory bowel disease (IBD) progression to cancer. The study team observed distinct shifts in stromal and myeloid compartments as inflammation advanced toward neoplasia. Single-nuclei maps in the source delineated cell clusters and temporal trends. Figures in the article show stepwise increases in specific fibroblast subsets coinciding with rising macrophage infiltration and altered epithelial states.
The authors reported that a CXCL14+ fibroblast population expanded progressively during the inflammation-to-cancer transition. These fibroblasts were identified as the primary local source of collagen in the tumour microenvironment. Data indicated upregulation of COL4A1 and COL4A2 and increased synthesis of collagen IV. Short sentence: collagen accumulation correlated with fibrosis. The source linked fibroblast activation to transcriptional changes, including increased SOX4 activity, which the investigators associated with enhanced COL4A1/A2 expression and collagen IV production.
The study team observed that elevated collagen IV promoted recruitment and differentiation of SPP1+ macrophages. Collagen IV was shown, in the reported analyses, to bind cell-surface receptors ITGB1/CD44 on macrophages. Binding triggered downstream activation of EGR1 and EGR2 transcriptional programs that the authors associated with SPP1+ macrophage differentiation and functional reprogramming. The cohort noted upregulation of S100A10 and ANXA2 in these macrophages, molecules the investigators connected to enhanced recruitment and oncogenic secretory activity.
Investigators reconstructed a molecular axis in which SOX4 activation in CXCL14+ fibroblasts increases COL4A1/A2 expression, driving collagen IV deposition. Collagen IV interaction with ITGB1/CD44 on macrophages then activated EGR1/EGR2, steering macrophages toward an SPP1+ state and elevating S100A10/ANXA2 expression. The authors presented pathway schematics and supporting expression data in figures. Short sentence: the pathway links stroma to myeloid cell fate. The source provided mechanistic connections across cell types but did not, in the abstract, list every intermediate biochemical experiment; full experimental details are contained in the main article.
Data indicated a feedback loop: SPP1+ macrophages secrete TGF-β1, which the investigators reported stimulates CXCL14+ fibroblasts to further increase COL4A1/A2 secretion. This reciprocal signalling, the study team argued, amplifies collagen IV deposition and sustains macrophage recruitment, thereby reinforcing fibrosis and tumour-promoting microenvironmental states. Short sentence: a self-reinforcing stromal–myeloid loop emerges from the data.
The authors analysed spatial patterns and reported that interactions between CXCL14+ fibroblasts and SPP1+ macrophages were more frequent in CMS4-subtyped colorectal tumours. Figures show co-localisation patterns and comparative frequencies across molecular subtypes. The cohort also reported shifts in CXCL9+/SPP1+ macrophage ratios during IBD-related cancer progression. The source linked the stromal–myeloid niche to features commonly attributed to the CMS4 phenotype, including abundant extracellular matrix and macrophage-rich stroma.
The article summarised functional experiments. In vitro co-culture and mechanistic assays supported a role for collagen IV in macrophage recruitment and differentiation; in vivo models of colitis-associated tumourigenesis, as presented in the figures, showed that CXCL14 and collagen IV promote tumour formation. The investigators reported that pharmacologic inhibition targeting DDR1/2 restrained colitis-associated tumourigenesis in the reported model. Short sentence: inhibition reduced tumour burden in the described experiments. The abstract and figures present these experimental findings; the source did not provide dosing details or comprehensive safety data in the abstract itself.
The authors deposited raw sequencing data in the Genome Sequence Archive (GSA-Human: HRA009030) and analysed publicly available datasets from the GDC and GEO. They stated that raw data could be made available upon request and that code is available by contacting the authors. The published preview indicates that full methodological and numerical detail is available in the subscription article. Remaining specific mechanistic questions documented by the source include the precise biochemical intermediates connecting ITGB1/CD44 engagement to EGR1/EGR2 activation in human tissues, and the translational path for DDR1/2 blockade — for example, dosing regimens and clinical safety profiles were not reported in the abstract. For access to full protocols, data tables and raw counts, readers are referred to the deposited datasets and to the full article.