Multimodal monitoring of cellular metabolic states by SRS and TPE imaging

Sammanfattning

Cancer cells display clear metabolic differences between different cell lines that indicate proliferation, survival and disease progression. Characterising such metabolic differences is vital for understanding cell behaviour and for determining potential biomarkers. In this study, a multimodal opticalimaging approach combining stimulated Raman scattering (SRS) microscopy, two-photon excitation optical redox ratio (TPE-ORR) imaging, and two-photon excitation fluorescence lifetime imaging microscopy (TPE-FLIM) was used to investigate metabolic differences between the breast cancer cell lines MCF7 and MDA-MB-231. SRS microscopy was utilised to quantify the prevalence and size of lipid droplets, while TPE-ORR and TPE-FLIM measurements were used to assess cellular redox state and fluorescence lifetime characteristics of NADH and FAD, including phasor-based analysis of lifetime distributions. Measurements were acquired across multiple experimental dates, and differences between cell lines were evaluated using both parametric and non-parametric statistical methods. MDA-MB-231 cells exhibited more than a 8-fold increase in lipid droplet prevalence than MCF7 cells, corresponding to more than 700% increase. In contrast, no robust difference in lipid droplet size was identified due to significant variability between biological replicates and experimental sessions. TPE-ORR measurements consistently distinguished the two cell lines, with pooled analyses showing increases of more than 200% and 100% in pixel-wise and cell-wise optical redox ratios, respectively, for MDA-MB-231 relative to MCF7. TPE-FLIM measurements revealed additional metabolic differences, particularly in FAD-related parameters and phasor centroid positions. However, several FLIM-derived metrics exhibited greater sensitivity to signal-to-noise limitations, fitting assumptions, and experimental variability, reducing their robustness as discriminative biomarkers. Overall, the results demonstrate that multimodal optical imaging can distinguish metabolic phenotypes between MCF7 and MDA-MB-231 cells. Among the investigated techniques, optical redox ratio measurements and SRS-derived lipid droplet prevalence provided the most robust and reproducible discrimination, while FLIM offered complementary metabolic information that may benefit from further methodological optimisation. Collectively, these findings demonstrate that individualparameters capture different and complementary aspects of cellular metabolism, and that a multiparametric approach is therefore better suited to robust metabolic phenotyping than reliance on any single modality alone.