Multi-omics study links GPRC5A+ epithelial cells to malignant colorectal cancer traits

Scientists have identified a distinct population of epithelial cells that appears to drive some of the most dangerous features of colorectal cancer, offering a potential new target for treating one of the world’s deadliest malignancies. In a sweeping multi-omics study published in the Journal of Translational Medicine, researchers led by Weichun Tang and Shengli Wang of the Third People’s Hospital of Bengbu, affiliated with Bengbu Medical University in China, combined single-cell sequencing, spatial transcriptomics, proteomics and bulk RNA sequencing to isolate and characterize a malignant cell state marked by the expression of a gene called GPRC5A. Their findings paint a detailed picture of how a small subset of tumor cells may orchestrate aggressive cancer behavior, and they point to an unexpected connection with the tumor microenvironment that could inform future therapeutic strategies.

Colorectal cancer remains one of the most common and lethal cancers worldwide, and its notorious cellular heterogeneity has long frustrated efforts to understand why some tumors progress relentlessly while others respond to treatment. Tumors are not uniform masses of identical cells; they contain diverse populations of cancer cells, immune cells and stromal cells that communicate with one another and collectively shape disease course. Understanding which specific cell types harbor the molecular programs responsible for malignancy is therefore a central question in cancer biology, and answering it requires looking beyond conventional bulk analyses that average signals across thousands of mixed cells.

The research team assembled an extraordinary dataset to tackle this problem. They integrated data from 2,993 colorectal cancer samples spanning four complementary technologies: bulk RNA sequencing from 2,568 samples drawn from two overall survival and recurrence-free survival cohorts; single-cell RNA sequencing capturing 281,961 individual cells from 152 specimens; spatial transcriptomics from six samples, which preserves information about where genes are expressed within intact tissue; and proteomics from 267 samples, which measures the actual proteins produced by tumor cells. This integrated approach allowed the investigators to move from population-level associations down to individual cells and back up to clinically validated signatures, a strategy increasingly seen as the gold standard for dissecting tumor complexity.

Using computational methods to integrate and annotate the single-cell data, the researchers constructed a stage-stratified atlas of colorectal cancer and resolved eleven distinct malignant epithelial subsets within tumors. Among these, one cluster stood out. Designated Epi_4, this subset was enriched in late-stage tumors and carried strong signatures of epithelial-mesenchymal transition, the process by which epithelial cells acquire migratory and invasive properties; hypoxia, reflecting the low-oxygen conditions typical of growing tumors; and inflammatory programs. Critically, patients whose tumors showed high activity in this subset had significantly worse overall survival and recurrence-free survival across the bulk RNA sequencing cohorts.

The defining molecular marker of this aggressive subset proved to be GPRC5A, a gene encoding a G protein-coupled receptor, a class of cell-surface proteins renowned for their roles in cellular signaling and their historical success as drug targets. The researchers designated this population GPRC5A-positive epithelial cells. GPRC5A expression rose steadily from stage I through stage IV disease, tracking with tumor progression, and elevated levels were associated with poor outcomes across multiple independent cohorts. Spatial transcriptomics confirmed that GPRC5A-positive cells occupied specific locations within tumor tissue consistent with the single-cell findings, and proteomic measurements at the protein level corroborated the RNA-based observations, providing a rare degree of concordance across molecular layers.

Correlation alone, however, does not establish function. To test whether GPRC5A actively drives malignant behavior or merely marks it, the team performed CRISPR-based perturbation experiments in colorectal cancer cell lines, altering GPRC5A expression and observing the consequences. Disrupting the gene affected cell proliferation, migration and invasion, the hallmarks of metastatic potential. Immunoblotting revealed corresponding changes in epithelial-mesenchymal transition markers, indicating that GPRC5A influences the molecular machinery that governs cellular plasticity. In mouse xenograft models, manipulating GPRC5A altered tumorigenicity, the capacity of cancer cells to seed and sustain tumors in living tissue. Together, these experiments support the conclusion that GPRC5A is not simply a passive biomarker but a functionally important contributor to malignant phenotypes, at least in the models tested.

The investigators then turned their attention upward along the regulatory hierarchy, asking which molecular master switches control GPRC5A expression. Using SCENIC, a computational framework that infers transcription factor activity from single-cell expression data, combined with analysis of binding motifs in the JASPAR database, they identified FOSL1 as a candidate upstream regulator. FOSL1 belongs to the AP-1 family of transcription factors, well-established players in cancer cell proliferation, invasion and inflammation. Chromatin immunoprecipitation followed by quantitative PCR, a technique that detects whether a specific protein binds to a specific DNA sequence, provided experimental support that FOSL1 physically occupies the GPRC5A promoter region. This finding suggests a concrete regulatory pathway through which malignant epithelial states might be induced and maintained, and it raises the possibility that blocking this axis could suppress the aggressive cell population.

Perhaps the most intriguing dimension of the study concerns the tumor microenvironment, the ecosystem of non-cancerous cells that surrounds and interacts with tumors. Spatial analysis and ligand-receptor mapping, which predicts communication between cell types based on the expression of signaling molecules and their corresponding receptors, revealed a close physical and functional association between GPRC5A-positive epithelial cells and a population of cancer-associated fibroblasts marked by the expression of periostin, designated POSTN-positive fibroblasts. The computational analysis predicted reciprocal signaling between these two cell populations through several ligand-receptor pairs, including COL1A1 interacting with SDC4, COL1A1 and COL1A2 engaging ITGA2 and ITGB1, and PPIA binding BSG. Fibroblasts are known to remodel the extracellular matrix and secrete growth factors that support tumor growth, and this study suggests a potentially reciprocal dialogue in which epithelial cells and fibroblasts reinforce each other’s malignant behaviors. Importantly, patients whose tumors displayed concurrent high signatures of both GPRC5A-positive epithelial cells and POSTN-positive fibroblasts had the worst overall and recurrence-free survival, suggesting that this cellular partnership may be a powerful indicator of aggressive disease.

The translational implications of the work extend to drug response. Using OncoPredict, a computational tool that estimates drug sensitivity from gene expression profiles, the researchers found an association between GPRC5A status and sensitivity to trametinib, an FDA-approved MEK inhibitor used in other cancers. Molecular docking and molecular dynamics simulations produced a computational model of a possible direct interaction between trametinib and the GPRC5A protein, raising the speculative but tantalizing prospect that the drug might act partly through this receptor. The authors are appropriately cautious on this point, emphasizing that the docking model remains experimentally unvalidated and that direct binding studies will be required before any therapeutic conclusion can be drawn. Cell sensitivity assays provided additional exploratory support for the link between GPRC5A and trametinib response, but the researchers stress that this line of investigation is hypothesis-generating rather than definitive.

The study’s conclusions are carefully hedged in ways that reflect both its ambition and its limitations. The authors state that GPRC5A-positive epithelial cells represent a malignancy-associated state in colorectal cancer and that GPRC5A is functionally important for malignant phenotypes in the tested models, conclusions that are well supported by their convergent evidence. However, they explicitly note that the inferred relationships with POSTN-positive fibroblasts and the trametinib findings should be regarded as hypothesis-generating pending functional crosstalk experiments, direct binding validation and therapeutic testing. This level of rigor is notable in a field where single-cell findings are sometimes overinterpreted, and it sets a clear roadmap for follow-up studies: co-culture systems to test epithelial-fibroblast signaling, biophysical assays to confirm or refute trametinib binding to GPRC5A, and ultimately clinical evaluation of GPRC5A as a biomarker for patient stratification.

The work also received approval from institutional ethics committees and was conducted in accordance with the Declaration of Helsinki, with written informed consent obtained from all participants and animal procedures reviewed by the appropriate ethics board. Supported by funding from Anhui Provincial and Bengbu Municipal research programs, the study exemplifies how relatively modest clinical research institutions can now leverage large public datasets and advanced molecular platforms to make contributions of genuine translational significance. If subsequent studies validate the central role of the GPRC5A-positive epithelial state and its interaction with the stromal compartment, the findings could eventually inform diagnostic tests that identify high-risk patients and therapeutic strategies aimed at disrupting the epithelial-fibroblast axis or exploiting the drug sensitivity patterns uncovered here. For now, the study stands as a compelling demonstration of how multi-omics integration can transform a heterogeneous tumor mass into a legible map of malignant cell states, communication networks and therapeutic vulnerabilities.