Contribution of Imaging Flow Cytometry to Storage Lesion Assessment: Identification of a Sub-Population of Morphologically Altered Erythrocytes

The storage lesion encompasses a series of biochemical and molecular modifications that alter erythrocytes during hypothermic storage, reducing transfusion yield. Indeed, in humans, up to 25% of transfused erythrocytes are cleared from the circulation in a few hours (Luten et al., 2008). The mechanisms underlying this rapid clearance are not fully elucidated but it is reasonable to assume that these erythrocyte alterations are sensed by the spleen, resulting in retention. Among those, membrane shedding may have a major impact on post-transfusion clearance of erythrocytes since it causes a progressive decrease in the surface-volume ratio of the cell, leading to the loss of its flexible biconcave shape. The proportion of "damaged" erythrocytes cleared in the hours following transfusion increases with the duration of storage while the deformability measured by ektacytometry (Frank et al., 2013) progressively decreases during this period. To characterize the morphological alteration of stored erythrocytes, we used imaging flow cytometry (Imagestream X Mark II, AMNIS°). This technology allows a simultaneous high-speed multispectral imaging of cells in brightfield, darkfield, and 9 channels of fluorescence. It combines the ability of conventional flow cytometry to analyze a very high number of events with a powerful exploration of cell morphology.

We analyzed the morphological, biochemical, metabolic, and bio-mechanical characteristics of erythrocytes stored in optimal blood bank conditions for 6 donors, at Day 3, 21, 28, 35 and 42 of the storage period. This longitudinal study of parameters such as mean corpuscular volume, intracellular ATP level, hemolysis, osmotic fragility, deformability and the plasma levels of ions and metabolites has highlighted a great inter-donor variability in the storage lesion process. Moreover, Imagestream analysis of front views of sharp, single cells revealed a subpopulation of small erythrocytes. The "projected surface area" distribution on normalized frequency plots was bimodal in 5 of 6 concentrates, showing a well-demarcated subpopulation of less than 62 µm2. The proportion of this sub-population increased with storage from 0.5-3.4% at D3 to 4-23.5% at D42 (p<0.05). These cells displayed a low fluorescence staining in the EMA-binding test, a diagnostic test for hereditary spherocytosis. With a more detailed morphological analysis we could determine that this sub-population corresponds to a mix of echinocytes III, spherocytes and sphero-echinocytes (Bessis classification). These results have been confirmed by differential interference phase contrast microscopy (DIC) observations, carried out in parallel, as a gold standard of our imaging flow cytometry study. Indeed, we found a very good correlation between the proportion of small erythrocytes detected in imaging flow cytometry and the echinocytes III, sphero- and spheroechino-cytes detected by DIC (correlation coefficient=0.84).

These morphological alterations have been considered irreversible (Berezina et al., 2002) and are reminiscent of those associated with mechanical clearance of erythrocytes in the spleen of patients with hereditary spherocytosis (Mohandas et al., 2008). We hypothesize that these storage-induced small erythrocytes correspond to the subpopulation of "damaged" erythrocytes that are rapidly cleared after transfusion. Confirmation of these findings using ex-vivo perfusion of human spleens and observational studies in transfused patients would support the use of imaging flow cytometry to predict transfusion yield.