Predicting PD-L1 expression on human cancer cells using next-generation sequencing information in computational...

Area of Science:

• Oncology
• Computational Biology
• Immunology

Background:

• Programmed death-1 (PD-1) and programmed death-ligand 1 (PD-L1) interaction on T cells and tumor cells causes immunosuppression, hindering anti-tumor immunity.
• PD-L1 expression in tumors is a critical diagnostic and prognostic biomarker for immunotherapy efficacy, often assessed via immunohistochemistry (IHC).
• Variability in IHC results necessitates complementary methods for accurate PD-L1 quantification.

Purpose of the Study:

• To develop and validate computational simulation models for predicting PD-L1 expression in cancer cell lines.
• To establish a computational approach that complements IHC for affirming PD-L1 expression levels.
• To investigate the role of cell genomics in influencing PD-L1 expression.

Main Methods:

• Genomic aberration profiles of multiple myeloma (MM) and oral squamous cell carcinoma (SCC) cell lines (MM.1S, U266B1, SCC4, SCC15, SCC25) were integrated into simulation models.
• Non-transformed cell lines were simulated to establish control baselines for PD-L1 expression.
• Predicted PD-L1 expression was quantitatively verified against experimental data from ELISA, IHC, and flow cytometry.

Main Results:

• Computational models accurately predicted PD-L1 expression in the tested MM and SCC cell lines.
• The simulation results demonstrated a strong correlation between cell genomics and PD-L1 expression levels.
• Genomic data significantly influenced cell signaling pathways affecting downstream PD-L1 expression.

Conclusions:

• Computational modeling provides a reliable method for predicting PD-L1 expression, affirming IHC findings.
• This approach has the potential to enhance the utility of PD-L1 as a biomarker for cancer immunotherapy.
• Extending this predictive model to patient tumors could improve treatment selection and outcomes.