Raman spectroscopy for assessing normal and cancer response in vitro following radiation exposure.
Devpura S, Brown S, Zhang Y, Barton K, Siddiqui F, Sethi S, Divine G, Pandya A, and Chetty I. Raman spectroscopy for assessing normal and cancer response in vitro following radiation exposure. Med Phys 2018; 45(6):e578.
Purpose: To characterize the radiation-induced biochemical responses in normal and cancer cells spectroscopically using established lung and prostate cells. Methods: Nine cells lines were investigated, including 4 human lung cells lines including fibroblasts (CCL186) and epithelial, squamous or adeno-squamous cancer cells (A549, CRL5928, and HTB178), 4 human prostate cells lines (two normal prostate epithelial (PCS440010 and CRL11610) and two prostate cancer (CRL2505 and DU145)) and 1 murine cell line (TRAMPC2). Cells were irradiated with 2, 5, 10, 15 and 30 Gy single fraction using a Cs-137 irradiator. A total of 1492 Raman spectra were acquired as a function of time post-radiation (24, 48 or 72 h). Principal component analysis (PCA) and discriminant function analysis (DFA) were used to analyze the Raman spectral features. Results: PCA captured 95-97% of the variance in the Raman spectral features for all the cell lines. Overall, PCA/DFA analysis showed unirradiated cells were correctly classified in the range of 72-100% depending on the type of cells and the time point. All irradiated lung and prostate cells (normal or cancer) were classified with 79.0 ± 13.6% accuracy. Unirradiated prostate normal cells were correctly identified in the range of 95-100% and lung fibroblast cells were correctly identified in the range of 84-100%. Unirradiated human prostate carcinoma cells and murine prostate cells were classified with an accuracy of >88% and 93%, respectively. Unirradiated lung cancer cells were classified in the range of 72-100%. Multiple Raman bands assignable to DNA/RNA, protein and lipids showed prominent contributions in cell differentiation between unirradiated and irradiated with different doses. Conclusion: 2 Gy is sufficient to characterize radiation-induced Raman features per PCA/DFA analysis for the majority of cell lines studied with a high degree of accuracy. Accuracy improved at the later time points, 48 and 72 h relative to 24 h in 4 cell lines.