Multispectral imaging of hematopoietic cells: where flow meets morphology.

ABSTRACT

Normal and abnormal blood cells are typically analyzed by either histologic or flow cytometric approaches. Histology allows morphological examination of complex visual traits but with relatively limited numbers of cells. Flow cytometry can quantify multiple fluorescent parameters on millions of cells, but lacks morphological or sub-cellular spatial detail. In this review we present how a new flow technology, the ImageStream (Amnis Corporation, Seattle, WA), blends morphology and flow cytometry and can be used to analyze cell populations in ways not possible by standard histology or flow cytometry alone. The ImageStream captures brightfield, darkfield and multiple fluorescent images of individual cells in flow. The images can then be analyzed for levels of fluorescence intensity in multiple ways (i.e. maximum, minimum, or mean) as well as the shape and size of the area of fluorescence. Combinatorial measurements can also be defined to compare levels and spatial associations for multiple fluorescent channels. We demonstrate an application of this technology to distinguish six stages of erythroid maturation which have been classically defined by morphological criteria, by measuring changes in Ter119 mean intensity and area, DNA (DRAQ5 stain) mean intensity and area, and RNA content (thiazole orange stain). Using this approach, we find that other characteristics of erythroid maturation, such as marker expression and nuclear offset, vary appropriately within the defined cell subsets. Finally, we show that additional measurements of cell characteristics not classically analyzed in cytometry, including surface unevenness and unusually high contrast in brightfield images combined with fluorescent markers allow complex discriminations of rare populations of cells.