TL;DR
This study maps genome-wide DNA methylation patterns in oral pre-malignant lesions (leukoplakia) and gingivobuccal complex oral cancers (GBC-OSCC), revealing distinct epigenetic signatures that evolve with disease progression. By integrating methylation data with genomic copy number and gene expression, the researchers identified novel candidate biomarkers for early detection and prognosis, offering potential for improved patient risk stratification and management.
What Problem Does This Paper Address?
Oral squamous cell carcinoma (OSCC), particularly in the gingivobuccal complex (GBC), is an aggressive cancer with high mortality, often preceded by pre-malignant lesions like leukoplakia. While genomic alterations in OSCC are known, the role of DNA methylation patterns across different stages of oral carcinogenesis remains less understood. Current diagnostic methods, primarily histopathology, are insufficient for predicting malignant transformation or disease progression. This knowledge gap highlights the urgent need for robust biomarkers for early detection of high-risk lesions, predicting treatment response, and improving prognosis in GBC-OSCC.
Methods
The study employed a multi-omics approach, integrating genome-wide DNA methylation, copy number alteration (CNA), and gene expression (GE) data.
- DNA Methylation Profiling: Genome-wide DNA methylation was assessed using Illumina's Infinium Methylation EPIC (850K) BeadChips. Data processing and analysis were performed using the RnBeads pipeline, including quality control, preprocessing, normalization, and differential methylation analysis.
- Integrative Analysis: The R-package CNAmet was utilized to integrate methylation, CNA, and GE data. This tool identifies genes synergistically regulated by these molecular changes, prioritizing candidates based on a synergistic score and FDR-corrected P-value.
- Survival Analysis: Regularized Cox regression with L1 penalty (LASSO) was applied to the integrated data to identify prognostic markers associated with patient survival.
- Validation: Candidate genes were validated using bisulfite pyrosequencing for promoter methylation and TaqMan real-time PCR for CNAs. Survival analysis was further performed using Kaplan-Meier curves and univariate/multivariable Cox regression.
Datasets
- Training Set:
- 22 normal oral tissue samples
- 22 leukoplakia tissue samples
- 74 gingivobuccal complex oral cancer (GBC-OSCC) tissue samples
- Validation Set:
- 127 GBC-OSCC tissue samples
- 20 normal oral tissue samples (used for validation of specific methylation markers)
- External Data: TCGA-HNSC dataset (MethCNA database) was used for validation of prognostic markers.
The study did not introduce a new benchmark dataset but rather leveraged existing datasets and public repositories for validation.
Key Insights and Ideas
- Epigenetic Drift with Disease Progression: DNA methylation profiles clearly distinguish normal tissues from leukoplakia and GBC-OSCC, with aberrant methylation increasing as the disease progresses from pre-malignant lesions to cancer. This suggests a stepwise epigenetic alteration during oral carcinogenesis.
- Shared and Distinct Methylation Signatures: While leukoplakia and GBC-OSCC have distinct methylation profiles compared to normal tissue, a significant fraction of differentially methylated promoters (DMPs) are shared between these two stages, hinting at early epigenetic events that persist or evolve.
- Integrative Multi-Omics Approach is Crucial: Combining DNA methylation, CNA, and GE data proved essential for identifying genes with prognostic significance, as synergistic regulation by these molecular layers points to key drivers of cancer progression.
- Novel Prognostic Biomarkers Identified: The study identified several candidate genes (e.g., FAT1, GLDC, HOXB13 for methylation; CYB5A, CASP4, ISG15 for CNA) that show promise as biomarkers for risk stratification and prognosis in GBC-OSCC.
- ISG15 as a Potential Positive Prognostic Marker: Notably, a gain in ISG15 was associated with better recurrence-free, disease-specific, and overall survival, a finding that warrants further investigation as a potential indicator of better outcomes.
Key Results
- DMPs Identified: 846 DMPs in leukoplakia and 5111 DMPs in GBC-OSCC were identified.
- Shared DMPs: 45 hypermethylated promoters were common between leukoplakia and GBC-OSCC.
- Prognostic Markers: 32 candidate prognostic markers were identified through integrative analysis and survival analysis.
- Methylation Validation: Bisulfite pyrosequencing confirmed differential promoter methylation for HOXB13 (hypermethylation), GLDC (hypermethylation), and FAT1 (hypomethylation) in GBC-OSCC compared to normal controls.
- CNA Validation: Copy number alterations were validated for several genes, with losses of CYB5A and CASP4 associated with poor recurrence-free survival, and a gain of ISG15 associated with better survival outcomes.
- Survival Association: ISG15 gain showed significant association with better RFS, DSS, and OS, with an AUC of 0.712 for RFS.
Limitations and Future Work
- Limitations:
- The study acknowledges that the CNAmet software might miss genes not regulated synergistically, co-expressed, or co-methylated.
- Clinical outcome data for leukoplakia patients were not available, limiting their inclusion in survival validation.
- The choice of |Δβ| threshold for differential methylation analysis can influence the number of identified DMPs, presenting a trade-off between sensitivity and specificity.
- The study did not observe strong correlations between methylation/CNA targets and clinical outcomes in all cases, suggesting a need for further investigation.
- Future Work:
- Functional validation of the identified candidate genes is crucial.
- Further investigation into the role of novel TSGs like HOXB13 and GLDC in oral carcinogenesis.
- Exploring the potential of ISG15 as a prognostic marker in larger, independent cohorts.
- Quantitative studies to optimize the |Δβ| threshold for improved DMP discovery.
- Investigating the interplay between CNA, GE, and methylation for a more comprehensive understanding of oral carcinogenesis.
Summary updated