Abstract
Multiphoton fluorescence lifetime imaging microscopy (MPM-FLIM) is extensively proposed as a non-invasive optical method to study tissue metabolism. The approach is based on recording changes in the fluorescence lifetime attributed to metabolic co-enzymes, of which nicotinamide adenine dinucleotide (NADH) is of major importance. However, intrinsic tissue fluorescence is complex. Particularly when utilizing two-photon excitation, as conventionally employed in MPM. This increases the possibility for spectral crosstalk and incorrect assignment of the origin of the FLIM signal. Here we demonstrate that in keratinocytes, proteins such as keratin may interfere with the signal usually assigned to NADH in MPM-FLIM by contributing to the lifetime component at 1.5 ns. This is supported by a change in fluorescence lifetime distribution in KRT5- and KRT14-silenced cells. Altogether, our results suggest that the MPM-FLIM data originating from cellular autofluorescence is far more complex than previously suggested and that the contribution from other tissue constituents should not be neglected-changing the paradigm for data interpretation in this context.