Studies for improved understanding of lipid distributions in human skin by combining stimulated and spontaneous Raman microscopy

Abstract

Advanced Raman techniques, such as stimulated Raman spectroscopy (SRS), have become a valuable tool for investigations of distributions of substances in biological samples. However, these techniques lack spectral information and are therefore highly affected by cross-sensitivities, which are due to blended Raman bands. One typical example is the symmetric CH2 stretching vibration of lipids, which is blended with the more intense Raman band of proteins. We report in this work an approach to reduce such cross-sensitivities by a factor of 8 in human skin samples. This is accomplished by careful spectral deconvolutions revealing the neat spectra of skin lipids. Extensive Raman studies combining the complementary advantages of fast mapping and scanning, i.e. SRS, as well as spectral information provided by spontaneous Raman spectroscopy, were performed on the same skin regions. In addition, an approach for correcting artifacts is reported, which are due to transmission and reflection geometries in Raman microscopy as well as scattering of radiation from rough and highly structured skin samples. As a result, these developments offer improved results obtained from label-free spectromicroscopy provided by Raman techniques. These yield substance specific information from spectral regimes in which blended bands dominate. This improvement is illustrated by studies on the asymmetric CH2 stretching vibration of lipids, which was previously difficult to identify due to the strong background signal from proteins. The advantage of the correction procedures is demonstrated by higher spatial resolution permitting to perform more detailed investigations on lipids and their composition in skin.