The research team, led by Haonan Zhang, Yulu Shi and Qiuyu Song under the supervision of Mingsong Li at the Third Affiliated Hospital of Guangzhou Medical University, combined single-nuclei transcriptomics, spatial transcriptomics, ecological dynamic modelling, multiplex immunohistochemistry, in vitro co-culture experiments and in vivo animal studies. This comprehensive approach allowed them to track how individual cell subtypes change and interact as inflammatory bowel disease progresses toward cancer, addressing a gap that previous bulk and single-cell studies had left open, particularly because few investigations had been conducted at single-nuclei level.

At the centre of the finding are CXCL14+ fibroblasts, a subset of stromal cells defined by their production of the chemokine CXCL14. The analysis revealed that these fibroblasts are the primary source of collagen within the gut microenvironment and that their abundance increases progressively during the transition from inflammation to cancer. Crucially, they secrete elevated levels of collagen IV, encoded by the COL4A1 and COL4A2 genes. Collagen IV is best known as a structural component of basement membranes, but the new data show it acts as a signalling molecule that reshapes the immune landscape of the inflamed gut.

The technical pathway the researchers traced begins with the transcription factor SOX4. During the inflammation-cancer transition, activation of SOX4 in CXCL14+ fibroblasts leads to increased COL4A1/A2 secretion, enhancing collagen IV synthesis. The secreted collagen IV then binds to ITGB1 and CD44 receptors on the surface of SPP1+ macrophages, a subset of immune cells named for their production of the glycoprotein osteopontin, or SPP1. This binding activates the transcription factors EGR1 and EGR2, promoting differentiation of these macrophages while upregulating the proteins S100A10 and ANXA2.

Activation of S100A10 and ANXA2 promotes recruitment of SPP1+ macrophages into the tissue, and this influx contributes to immune escape and the secretion of oncogenic factors, further promoting fibrosis and tumorigenesis. The study also found that the ratio of CXCL9+ to SPP1+ macrophages decreases as IBD-related cancer progresses, a shift consistent with recent work in Science showing that CXCL9:SPP1 macrophage polarity identifies cellular programs that control human cancers. In the inflamed gut, the balance tips toward the SPP1+ state, which favours scarring and tumour growth rather than antitumour immunity.

Perhaps the most striking discovery is that the interaction is not one-directional. SPP1+ macrophages stimulate CXCL14+ fibroblasts to secrete more COL4A1/A2 via TGF-β1, forming a feedback loop that amplifies the crosstalk within the tumour microenvironment. Each cycle of this loop deposits more collagen, recruits more macrophages, and pushes the tissue further along the road from chronic inflammation to fibrosis and cancer. Ecological dynamic modelling showed that these interactions become more frequent in the CMS4 molecular subtype of colorectal cancer, which is characterised by a prominent stromal component.

The in vivo experiments provided direct causal evidence. CXCL14 and collagen IV promoted colitis-associated tumorigenesis in animal models, confirming that the fibroblast-macrophage axis is not merely a correlate of disease progression but an active driver of it. When the researchers inhibited collagen IV secretion, the results were twofold: the intervention not only prevented fibrosis but also hindered the development of the inflammation-cancer transition. This dual effect suggests that a single molecular target could address both of the major long-term complications of inflammatory bowel disease.

The therapeutic implications extend to a class of drugs already under development. Treatment with a discoidin domain receptor 1/2 inhibitor, which blocks collagen signalling receptors, restrained colitis-associated tumorigenesis in the animal studies. DDR1/2 inhibitors such as VU6015929 have been explored as antifibrotic agents in other contexts, and the new findings position them as candidates for preventing both stricture formation and malignancy in patients with long-standing inflammatory bowel disease. Because tumour-associated fibrosis is known to impair the response to immunotherapy, dismantling this stromal barrier could also improve the effectiveness of existing cancer treatments in patients who do develop IBD-associated malignancies.

The study addresses a clinically urgent problem. Population-based cohort studies with decades of follow-up have established that patients with inflammatory bowel disease face elevated risks of colorectal cancer, small bowel cancer and other gastrointestinal malignancies, and systematic reviews show that IBD-associated colorectal cancer carries distinct clinicopathological features and survival outcomes. Fibrosis leading to strictures is itself a risk factor for high-grade dysplasia and cancer in ulcerative colitis, and serological biomarkers of type I, III and IV collagen turnover have been associated with the presence and future progression of stricturing and penetrating Crohn’s disease. The new work provides a mechanistic framework that connects these epidemiological observations to specific cell types and signalling molecules.

By resolving the dynamics of cell subtypes at single-nuclei level and validating the findings through spatial mapping, co-culture and animal models, the researchers have identified CXCL14+ fibroblasts and their collagen IV secretion as a pivotal node in the inflammation-cancer transition. The identification of the SOX4-COL4A1/A2-ITGB1/CD44-EGR1/2-S100A10/ANXA2 pathway, together with the TGF-β1-mediated feedback loop, offers a detailed molecular map for developing innovative therapeutic strategies. The raw sequence data have been deposited in the Genome Sequence Archive under accession HRA009030, and the authors report no competing interests. The work was supported by the National Natural Science Funds of China and regional research funds from Guangdong and Guangzhou.

Subject of Research: Cellular and Molecular Biology

Article Title: CXCL14+ fibroblasts promote development of IBD-associated fibrosis and cancer by interacting with SPP1+ macrophages via collagen IV secretion

Article References: Zhang, H., Shi, Y., Song, Q., Li, N., Lin, Z., Chen, Q., Yang, Y., Chen, Y., Liu, Q., Tan, Y., Zheng, Y., Wang, S., Sun, Y., Li, X., He, C., Zhu, L., & Li, M. (2026). CXCL14+ fibroblasts promote development of IBD-associated fibrosis and cancer by interacting with SPP1+ macrophages via collagen IV secretion. British Journal of Cancer. https://doi.org/10.1038/s41416-026-03636-2

Image Credits: AI Generated

DOI: 10.1038/s41416-026-03636-2

Keywords: inflammatory bowel disease, colorectal cancer, fibrosis, cancer-associated fibroblasts, macrophages, collagen IV, single-nuclei transcriptomics, tumour microenvironment, DDR1/2 inhibitors, spatial transcriptomics, CXCL14, fibroblasts