Potential of SERS and proteomics for biomarker detection in cancer cells

Abstract This study investigates the use of quantitative LC-MS/MS-based proteomics and surface-enhanced Raman spectroscopy (SERS) for biomarker detection in classical Hodgkin lymphoma (HL). Two HL cell models with distinct TP53 status were utilized to evaluate the effects of etoposide, a DNA-damaging chemotherapeutic agent, and resveratrol, a polyphenolic compound with known chemosensitizing activity. For SERS, the best performance was achieved by applying logistic regression to classify different treatment conditions and identify discriminative spectral features in the data. Proteomics showed highly reproducible and accurate results with relative standard deviations of below 5% for the sample preparation and about 2% for the measurements. Proteomic profiling revealed a TP53-dependent organization of metabolic and stress-response pathways and demonstrated that cryopreserved aliquots yielded the most consistent proteomic signatures. Treatment-dependent regulation of key biomarker proteins showed direct correspondence to specific SERS features, such as reduced nucleotide/cytochrome-associated signals and enhanced amide and aromatic amino acid signals. Our findings highlight the strength of applying reproducible proteomic profiling with machine learning–guided SERS analysis to improve molecular interpretation and to validate potential biomarkers in cancer research. Future work will focus on refining the analytical workflow and extending it toward integrative multi-omics applications, enabling more comprehensive biomarker detection and mechanistic insight into classical Hodgkin lymphoma. This strategy will be extended to additional model systems—such as metastatic melanoma, melanocytes, and ultimately patient-derived leukemia cells—to help bridge the gap toward clinical translation. Graphical abstract An analytical approach that enables reproducible and complementary biomarker discovery in Hodgkin lymphoma cells by applying robust SERS-based spectral classification with quantitative LC-MS/MS proteomics for molecular validation. It emphasizes reproducibility and data complementarity while identifying treatment-specific molecular signatures.