Integrative bulk and single-cell transcriptomics identify a recurrence-prognostic risk model with exploratory associations with treatment-response phenotypes in colorectal cancer
Aug 2026· Human Cell· Vol 39· 0 citations· 58 references
Medicine
TL;DR
A READ-derived response-associated gene signature for recurrence stratification and exploratory cross-cohort evaluation in additional colorectal cancer cohorts, while further exploring its association with treatment-response phenotypes.
: Backgrounds: Colorectal cancer (CRC) prognosis remains difficult due to molecular heterogeneity and interaction between tumor cells and the immune microenvironment. This study aimed to identify transcriptomic and immune-cell patterns associated with overall survival (OS) and to develop an integrated prognostic model to improve risk stratification. Methods: RNA-sequencing was performed on 131 primary CRC samples and matched normal tissues. Differentially expressed genes (DEGs) were identified and functionally characterized through gene ontology, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, and protein–protein interaction (PPI) network analysis. Immune-cell composition was estimated using CIBERSORTx deconvolution and evaluated for its association with OS. Prognostic DEGs were screened using univariate Cox regression and Least Absolute Shrinkage and Selection Operator (LASSO) analysis to construct a risk score. The model was validated in the Cancer Genome Atlas Database (TCGA)-COAD (colon cancer) and READ (rectal cancer) cohorts. A nomogram integrating molecular and clinicopathological variables were generated. Results: A total of 5589 DEGs were identified between CRC and normal tissues, enriched in pathways related to cell cycle, Tumor Protein P53 (TP53), WNT Family Member (WNT), Janus kinase/signal transducer and activator of transcription 3 (JAK/STAT), calcium signaling, metabolism, and immune regulation. PPI analysis highlighted ten upregulated hub genes involved in mitotic spindle formation and chromosomal stability. Immune infiltration analysys indicated that higher proportions of plasma cells ( p = 6.9 × 10 − 4 ), naïve B cells ( p = 0.019), resting CD4 + memory T cells ( p = 0.02), M0 macrophages ( p = 0.0077), and activated dendritic cells ( p = 1.89 × 10 − 5 ) were associated with improved OS, whereas monocytes ( p = 0.012), neutrophils ( p = 0.041), activated mast cells ( p = 0.0066), and M2 macrophages ( p = 0.014) were linked to poorer survival. A seven-gene signature including aspartate beta-hydroxylase (ASPH), bradykinin receptor B1 (BDKRB1), calcium voltage-gated channel auxiliary subunit beta 1 (CACNB1), C-C motif chemokine receptor 8 (CCR8), cyclic nucleotide gated channel subunit alpha 3 (CNGA3), microtubule associated protein 1A (MAP1A) and oxytocin receptor (OXTR) stratified patients into high-and low-risk groups with significant OS differences. The model demonstrated strong predictive performance (AUC: 0.84 at 1 year) and was validated in TCGA cohorts. Multivariate analysis confirmed the risk score as an independent prognostic factor. The integrated nomogram accurately predicted 1, 3-, and 5-year survival (C-index = 0.757; 95% CI 0.724–0.791). Conclusions: We developed and validated a seven-gene immune-related prognostic signature that, combined with clinicopathological parameters, provides a robust tool for individualized survival prediction and may guide precision management in CRC patients.
Triple-negative breast cancer (TNBC) is an aggressive, heterogeneous form of breast cancer with limited specific therapy options, prevalent metastasis and frequent relapse. Re-analysis of single-cell RNA-sequencing data characterizes the diverse cell subtypes within the tumor and microenvironment of TNBC, supporting a luminal progenitor origin for the cancer and providing clues as to the factors involved in progression of the disease. The relative burdens of these subtypes can be deconvolved from bulk RNA-sequencing data, readily identifying the stem-like, mesenchymal and stromal cell subtypes significantly associated with poor survival and enrichment in metastasis. Importantly, these can be simplified to ten-gene signatures with comparable predictive power, notably in response to different therapeutic strategies, which are linked to relative burdens of different subtypes of stromal fibroblasts. The expression level of these signatures could provide a cheap means for selecting therapy strategies in personalized medicine.
G. Davidson, V. Debien, Tom Sexton· bioRxiv· 0 citations
Background Small cell lung cancer (SCLC) is a highly aggressive malignancy characterized by rapid progression, early metastasis, frequent relapse, and treatment resistance. Although chemoimmunotherapy has improved outcomes in a subset of patients, reliable molecular stratification tools reflecting tumor heterogeneity and survival risk remain limited. Chaperone-mediated autophagy (CMA) is involved in proteostasis, metabolic adaptation, and stress responses, but its cellular heterogeneity and prognostic relevance in SCLC remain unclear. Methods This study integrated single-cell RNA sequencing data, bulk transcriptomic cohorts, and clinical information to characterize CMA-related heterogeneity in SCLC. At the single-cell level, SCLC subtypes were annotated, CMA scores were calculated, and differences among baseline, sensitive, and resistant states were evaluated. At the bulk level, tumor versus normal differential expression analysis and weighted gene co-expression network analysis were performed to identify CMA-related candidate genes. A CMA-related Risk Score was constructed through comprehensive machine learning comparison and evaluated using survival analysis, time-dependent receiver operating characteristic curves, Cox regression, nomogram analysis, functional enrichment, immune microenvironment analysis, and drug sensitivity prediction. ARRDC3 was further validated by in vitro experiments. Results Single-cell analysis revealed marked SCLC subtype heterogeneity and nonuniform CMA activity across cellular subtypes. Significant CMA score differences between sensitive and resistant cells were observed in the SCLC-A_NR0B1/MYCL+ and SCLC_Hypoxia_glycolytic subtypes. Integration of CMA-related co-expression modules with tumor-associated differentially expressed genes identified 53 candidate genes. The machine learning-derived Risk Score effectively stratified patients into high-risk and low-risk groups across the combined cohort, GSE60052 cohort, and cBioPortal cohort, with high-risk patients showing significantly poorer overall survival. The Risk Score remained an independent prognostic factor and was associated with proliferative, cell-cycle, metabolic, immune, and drug sensitivity-related features. In vitro experiments showed that ARRDC3 overexpression suppressed proliferation, colony formation, migration, and invasion in H446 cells. Conclusion This study revealed CMA-related cellular heterogeneity in SCLC and established a robust prognostic Risk Score associated with survival, biological pathway activity, immune microenvironment features, and potential therapeutic responses. Functional validation of ARRDC3 further supports the biological relevance of this model, providing new insights into SCLC molecular risk stratification and CMA-associated resistance states.
Jiahong Han, Shun Xu, Yunfeng Jiang· Frontiers in Cell and Develo...· 0 citations
Background Colorectal cancer (CRC) is characterized by profound molecular heterogeneity and complex tumor–microenvironment interactions, which contribute to invasion, metastasis, and variable clinical outcomes. More effective prognostic biomarkers and therapeutic targets are still needed. Methods We integrated single-cell RNA sequencing, spatial transcriptomics, multi-cohort bulk transcriptomic data, drug response prediction, and in vitro experiments to characterize malignant epithelial states in CRC and identify clinically relevant biomarkers and candidate therapeutics. Stemness, copy number variation, pathway activity, and cell–cell communication were analyzed at single-cell resolution. A prognostic model was established using epithelial marker genes and validated across TCGA and five GEO cohorts through a consensus machine-learning framework. Drug sensitivity was evaluated using CTRP and PRISM datasets, and candidate compounds were further prioritized through a network-based drug repositioning strategy. Spatial transcriptomics was used to define the tissue localization of key genes, and functional assays were performed to assess the role of TRIP6 in CRC cells. Results Malignant epithelial cells exhibited increased stemness, frequent aneuploidy, enhanced glycolysis, along with upregulation of the Wnt/β-catenin and PI3K-AKT-mTOR signaling cascades. Cell–cell communication analysis revealed prominent extracellular matrix-related interactions linking cancer-associated fibroblasts and epithelial cells, suggesting a microenvironmental contribution driving epithelial–mesenchymal transition (EMT). The prognostic model showed stable predictive performance across independent cohorts. Among the model genes, TRIP6 was identified as a risk-associated candidate and was enriched at the tumor–stroma invasive boundary by spatial transcriptomic analysis. In vitro, TRIP6 knockdown suppressed CRC cell proliferation and migration, promoted apoptosis, and partially reversed EMT-associated marker expression, supporting a potential role for TRIP6 in malignant progression. In addition, drug prediction analyses identified several compounds with potential therapeutic relevance for high-risk patients. Conclusion This study provides a multi-omic framework for understanding CRC progression and suggests that TRIP6 may be involved in linking microenvironmental signaling to invasive phenotypes, with potential value for prognostic stratification and therapeutic exploration.
Qiwei Cui, Xiao-Hong Gao, Xiaoqing Wu et al.· Frontiers in Cell and Develo...· 0 citations
A six-gene-fibrosis-based prognostic model based on six genes stratifies survival risk and correlates with immune features and drug sensitivity, but provides a preliminary framework requiring prospective clinical validation.
Yanyan Qiu, Cui Lv, Shubo Ding· Clinical and Translational O...· 0 citations
This study establishes a novel CSC–associated gene signature for diagnosis and prognosis in HCC and nominates belinostat as a repurposing candidate for targeting stemness‐related pathways, offering a promising strategy for personalized therapy.
Yang Zi, Ying Zhang, Jun Wu et al.· Stem Cells International· 0 citations