TL;DR
This study identifies a set of eight long noncoding RNAs (lncRNAs) related to programmed cell death (PCD) that can predict colon cancer (CC) patient prognosis. This prognostic model, validated externally, reveals distinct immune infiltration patterns and therapeutic sensitivities between high- and low-risk groups, highlighting the role of lncRNAs in CC's immune landscape and treatment response.
What Problem Does This Paper Address?
Colon cancer (CC) remains a significant cause of cancer-related deaths, and current clinical management strategies often fall short of achieving satisfactory outcomes, especially for advanced stages. The complex interplay between cancer cells, the tumor microenvironment (TME), and the immune system is crucial for cancer progression and treatment efficacy. Programmed cell death (PCD), encompassing apoptosis, pyroptosis, and ferroptosis, plays a vital role in this dynamic. While long noncoding RNAs (lncRNAs) are known to regulate various cellular processes, including PCD, their specific roles in CC prognosis, immune infiltration, and therapeutic response are not fully understood. This paper aims to bridge this knowledge gap by identifying prognostic PCD-related lncRNAs and developing a model to predict CC patient outcomes and guide treatment strategies.
Methods
The researchers utilized a multi-faceted approach:
- Data Mining and Screening: They began by downloading transcriptome sequencing and clinical data for CC patients from The Cancer Genome Atlas (TCGA) database. They then identified 199 PCD-related genes and subsequently screened for lncRNAs associated with these genes through co-expression analysis.
- Prognostic Model Development: Using univariate Cox regression, they identified 41 prognostic PCD-related lncRNAs. A prognostic model was then constructed using a LASSO regression algorithm, selecting eight key lncRNAs. This model was further validated using an external dataset.
- Subtype Classification: Consensus clustering was employed to classify CC patients into distinct subgroups based on the expression of prognostic PCD-related lncRNAs, allowing for the analysis of their biological characteristics and clinical outcomes.
- Immune Microenvironment Analysis: Various algorithms (TIMER, CIBERSORT, etc.) were used to assess immune cell infiltration, immune scores, and stromal scores within different patient subgroups. Immune checkpoint gene expression was also analyzed.
- Functional Enrichment Analysis: Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to understand the biological functions of differentially expressed genes between subgroups.
- Tumor Mutation Burden (TMB) and Copy Number Variation (CNV) Analysis: TMB was calculated, and CNV analysis was conducted to explore genetic alterations associated with risk groups.
- Therapeutic Response Prediction: The study assessed patient sensitivity to immunotherapy and various chemotherapeutic agents using publicly available databases and tools.
- In Vitro and In Vivo Validation: The role of a specific lncRNA, U62317.4, was investigated through cell culture experiments (proliferation, invasion, migration, apoptosis assays) and tumor xenotransplantation in mice.
Datasets
- The Cancer Genome Atlas (TCGA): This was the primary dataset used for training and developing the prognostic model. It provided transcriptome sequencing and clinical information for CC patients.
- External Dataset: A cohort of 96 CC patients from the Colorectal Cancer Center of Shanghai Tenth People's Hospital was used to validate the prognostic model and examine lncRNA expression.
- Immunotherapy Sensitivity Databases: http://tide.dfci.harvard.edu/ was used to assess immunotherapy sensitivity.
- Chemotherapy Sensitivity Databases: Genomics of Drug Sensitivity in Cancer (GDSC) and the pRRophetic package were used to evaluate sensitivity to chemotherapeutic agents.
Key Insights and Ideas
- PCD-related lncRNAs as Prognostic Biomarkers: The study strongly suggests that specific lncRNAs involved in programmed cell death pathways can serve as powerful prognostic indicators for colon cancer.
- Immune Microenvironment Heterogeneity: Different clusters of CC patients, defined by PCD-related lncRNA expression, exhibit distinct immune cell infiltration patterns and immune function levels. This highlights how lncRNAs can shape the tumor's immune landscape.
- Therapeutic Resistance in High-Risk Groups: Patients classified as high-risk by the prognostic model appear to be more resistant to immunotherapy and certain chemotherapeutic drugs, suggesting a need for tailored treatment approaches.
- U62317.4 as a Carcinogenic Factor: The lncRNA U62317.4 was identified as a key player in the prognostic model and was further validated to promote CC proliferation, invasion, and migration, while inhibiting apoptosis, underscoring its oncogenic role.
- Synergistic Prognostic Value: The developed prognostic model, incorporating clinical characteristics like age and stage, provides a robust tool for predicting patient outcomes, offering a more comprehensive assessment than individual factors alone.
Key Results
- 41 Prognostic PCD-related lncRNAs Identified: Through univariate Cox analysis, 41 lncRNAs were found to be significantly associated with CC prognosis.
- Eight-lncRNA Prognostic Model: A robust prognostic model was developed using eight PCD-related lncRNAs, demonstrating high diagnostic power in both training and validation cohorts.
- Distinct Patient Subgroups: Consensus clustering identified two distinct CC subgroups with significantly different overall survival (OS) rates, with cluster 2 exhibiting suppressed immune function and poorer prognosis.
- Independent Prognostic Factors: Age and the developed risk score were identified as independent prognostic factors for CC patients.
- U62317.4 Inhibition Suppresses CC Growth: Silencing U62317.4 significantly inhibited CC cell proliferation, invasion, and migration, and promoted apoptosis in vitro and in vivo.
- Therapeutic Sensitivity Differences: High-risk patients showed higher TIDE scores (indicating potential immune escape) and were predicted to be more sensitive to certain chemotherapy drugs (imatinib, lapatinib, PHA.665752, MS.275), while low-risk patients were more sensitive to others (metformin, MK.2206, etc.).
Limitations and Future Work
- Multi-omics Data Limitations: The study primarily relied on TCGA data, and while comprehensive, the multi-omics analysis (transcriptome sequencing, TMB, CNV) could not be fully verified due to cost and accessibility. Future work could involve broader validation of these analyses.
- External Data Validation: While an external dataset was used for model validation, further prospective, multicenter studies are recommended to confirm the findings and the clinical utility of the prognostic model.
- TME Heterogeneity: The authors acknowledge that tumor tissues were collected from core areas, which might not fully capture the heterogeneity of the TME across different tumor regions.
- Drug Sensitivity Validation: External data for drug sensitivity results were not available, limiting the ability to fully validate these predictions.
- Future Directions: The study suggests that the identified PCD-related lncRNAs and the prognostic model can provide new insights for improving CC patient OS and facilitating individualized treatment. Further research into the precise molecular mechanisms by which these lncRNAs influence immune responses and therapeutic outcomes is warranted. The development of novel immunotherapies targeting these lncRNAs is also a promising avenue.
Summary updated