Label-Free Hyperspectral Imaging of Intracellular Hemoglobin in Human Erythrocytes

Abstract

A broadband hyperspectral imaging (HSI) technique, originally developed for remote sensing and imaging in geophysical applications, recently has been employed to enable noninvasive label-free optical imaging of biological specimens. Here we report our results on hyperspectral microscopy of non-labeled erythrocytes to study the distribution of hemoglobin. We have chosen to focus on cellular hemoglobin because it is an endogenous chromorphore that is present in high concentrations, which makes it a good candidate for establishing proof of principle. In our imaging technique, hyperspectral data cubes (x, y, λ) are collected by either a wide-field or confocal microscope equipped with a broadband light source, such as a white light laser or spectral light engine in which narrow-band selection and wavelength scanning is achieved by fast optoelectronic devices. The obtained data cubes then are analyzed by algorithms based on a spectral angle mapper, yielding a set of unique absorption spectral signatures (i.e. endmembers) corresponding to subforms of hemoglobin, such as oxyhemoglobin and methemoglobin. Wavelength-dependent scattering signatures of cell membranes are also resolved. Unique endmembers of specific hemoglobin subforms are identified and used to build a map of intracellular hemoglobin distribution by estimating the abundance of each specific endmembers from the hyperspectral data cube. Our ongoing efforts include technology development for the determination of the local thickness of a single erythrocyte, to enable single cell volumetric measurement. The technique may be further developed for label-free molecular and chemical imaging of a broad range of endogenous biomarkers in cells and tissues, and ultimately for in vivo molecular imaging as well.