Sensitive detection and identification method of erythrocyte-like cells upon doxorubicin induced differentiation with vibrational techniques

Altered differentiation of blood cell precursors and their clonal expansion occurs in various types of leukaemia. One treatment strategy is to induce differentiation into the mature form, which is capable of undergoing apoptosis. It has been found that low doses of doxorubicin (DOX) induce phenotypic and morphological changes in malignant erythroid precursors to erythrocyte-like cells. These are usually studied by time-consuming examination of haemoglobinisation by benzidine staining or glycophorin A expression by Western blot or flow cytometry. As an alternative, we propose an in-depth investigation of DOX-induced erythroid differentiation using Raman spectroscopy and stimulated Raman scattering microscopy and fast and sensitive classification using the probe for mitochondrial imaging (MitoBADY). Machine learning methods, including Orthogonal Partial Least Squares, Principal Component Analysis, and Multivariate Curve Resolution-Alternating Least Squares, etc., were implemented to extract spectroscopic markers of differentiation from both single spectra and the hyperspectral images.