During hypothermic pre-transfusion storage, part of the RBC metabolism shifts from glycolysis to the pentose phosphate pathway after 10–14 days of storage, leading to a progressive decrease of the intracellular ATP level (D’Alessandro et al., 2015; Bordbar et al., 2016). The fragilized RBC is thus less capable of coping with the oxidative stress generated by storage, further affecting RBC integrity (Reisz et al., 2016). These metabolic and oxidative stresses contribute to the progressive modifications of RBC properties (Yoshida et al., 2019). Among those, RBC morphology is a well-documented and key RBC property altered during storage.

A number of techniques are used to explore RBC morphology. The pioneer work of Bessis (1972, 1974) using scanning electron microscopy defined RBC morphological categories. This technique is still used to assess RBC morphology during storage (Berezina et al., 2002; Zehnder et al., 2008; Antonelou et al., 2012; Blasi et al., 2012; D’Alessandro et al., 2012; Roussel et al., 2021).

Digital holographic microscopy was recently used to explore RBC morphology along storage (Bardyn et al., 2017). Light microscopy can also be used to observe fixed (Högman et al., 1985; Tchir et al., 2013; Reinhart et al., 2015) or unfixed cells in a physiological medium (Roussel et al., 2017; Lu and Shevkoplyas, 2020). RBC morphology is indeed sensitive to RBC intrinsic properties and to the suspension medium (Bessis, 1972). Imaging flow cytometry (IFC) rapidly acquires tens of thousands of images of unfixed cells in a physiological medium and identifies a well-demarcated subpopulation of morphologically altered and smaller RBC, comprising type III echinocytes, spheroechinocytes, and spherocytes (Roussel et al., 2017). These small RBC, defined as storage-induced micro-erythrocytes (SME), accumulate during storage, reaching a mean proportion of 24% of the entire RBC population at day 42 and their proportion negatively correlates with post-transfusion recovery in healthy volunteers (Roussel et al., 2021). Furthermore, the capacity of IFC to discriminate fluorescently stained RBC allowed to observe their clearance from circulation upon ex vivo perfusion in human spleens and in vivo transfusion in mice (Roussel et al., 2021). The proportion of SME is thus a relevant storage quality marker that identifies the erythrocyte subpopulation targeted for rapid post-transfusion clearance. This study aimed to develop an alternative method of SME quantification and sorting using standard flow cytometry. Obtaining pure subpopulations of SME or morphologically normal erythrocytes would indeed greatly facilitate the exploration of the intrinsic alterations of RBC inducing their clearance.

We describe here a simple staining flow cytometry-based protocol to quantify SME that accumulate during pretransfusion storage. In addition, pure subpopulations of SME and long-stored morphologically normal erythrocytes can now be obtained using flow cytometry-based sorting.

We observed that CFDA-SE stained long-stored RBC show a bimodality in fluorescence intensity after an incubation at 37°C. This bimodality was not observed immediately after staining nor after a 24 h incubation at 4°C suggesting that it could be due to an enzymatic process. Non-fluorescent CFDA-SE diffuses passively into cells and is rapidly processed by cellular esterases resulting in high fluorescent CFSE. The resulting molecule is less permeant to cell membrane and highly reactive to free cellular amines, which leads to its spontaneous covalent binding to cellular components, mostly intracellular proteins (Banks et al., 2013). Thus, the persistence of CFSE labeling in cells is expected to depend predominantly on protein turn-over and its natural fluorescence decay, making it a useful and widely used cell division tracker (Lyons et al., 2013).

In our study, the decrease in CFSE fluorescence intensity was not evenly distributed across all long-stored RBC, revealing instead two well-demarcated RBC subpopulations. IFC images showed that RBC with stable and thus higher CFSE fluorescence display marked morphological alterations comprising mostly SME, while RBC with decreased CFSE fluorescence are morphologically normal. This suggests that an enzymatic and/or metabolic process is lost in SME and preserved in morphologically normal erythrocytes. Even though the nature of the enzymatic and/or metabolic process remains to be identified, CFDA-SE reactivity can now be considered a new marker of the storage lesion. The segregation of this defect in the morphologically altered RBC strongly suggests that only a subpopulation of RBC is severely altered during storage, possibly corresponding to the older RBC present in the concentrate at the beginning of storage, as previously suggested by others (Tuo et al., 2014; Mykhailova et al., 2020).

A classic approach to study the accumulation of storage lesions is to compare groups of different storage durations (Paglia et al., 2016). Another possible approach is by using a property that evolves during storage and that allows the implementation of an enrichment strategy. Cell density is known to increase during storage (Diaz et al., 1996) and permits to study the evolution of cellular properties in each cell fraction along storage (Owusu et al., 2013; Ciana et al., 2017) or even to first separate different subpopulations and to evaluate the evolution of storage lesion markers in each fraction (Mykhailova et al., 2020). Microfluidics can also be used to enrich in RBC based on their margination capacity in circulation (Chen et al., 2018).

Imaging flow cytometry and scanning electron microscopy show that the CFDA-SE staining protocol presented here offers the possibility to obtain >95% pure subpopulations of SME (CFSE^high) by flow cytometry-based sorting. Pure subpopulations of SME or morphologically normal erythrocytes would greatly facilitate the exploration of the intrinsic alterations of RBC inducing their clearance.

In our study, CFSE^high erythrocytes were slightly more numerous than SME. This difference may stem from the presence of a few additional altered erythrocytes appearing upon incubation for staining, not detected when experiments are performed on RBC sampled directly from refrigerated blood bag. These erythrocytes, unable to maintain their cellular homeostasis at 37°C, would become SME and remain highly fluorescent due to their enzymatic and/or metabolic lesions. Indeed, an overnight incubation step at 37°C was shown to expose some RBC lesions induced by hypothermic storage (Burger et al., 2013).

Roussel et al. (2021) recently showed that SME accumulate during storage, reaching a mean proportion of 24% of the entire RBC population at day 42 of storage. The proportion of SME in circulation decreased rapidly when perfusing long-stored red cell concentrates in an ex vivo model of human spleen. Similarly, a mouse model showed decreased transfusion recovery of SME when compared to morphologically normal long-stored RBC. Furthermore, the proportion of SME in red cell concentrates negatively correlated with transfusion recovery in healthy volunteers using ⁵¹Chr-labeling procedure, confirming that SME quantification is a physiologically relevant marker of RBC storage quality, predictive of transfusion recovery.

Even though transfusing short-stored, as compared with standard-issue, red cell concentrates does not reduce mortality (Lelubre and Vincent, 2013; McQuilten et al., 2018, 2020), storage duration is associated with decreased transfusion recovery (Luten et al., 2008; Rapido et al., 2018). Suboptimal transfusion recovery is expected to increase the number of red blood cell concentrates needed to restore hemoglobin concentration. Along this line, transfusion-induced hemoglobin increments are significantly decreased when red cell concentrates are stored for longer durations (Hunsicker et al., 2018; Roubinian et al., 2019; Rydén et al., 2019) and may be associated with donor genetic factors (Roubinian et al., 2022). Pretransfusion control of SME, and elimination of lower-quality products, would be particularly beneficial to chronically transfused patients with sickle cell disease, thalassemia or myelodysplastic syndromes. Transfusion-related iron overload is indeed a major cause of morbidity and mortality in these patients and the need to provide more transfusions could have significant adverse consequences.

Quantification of SME using IFC is a label-free, operator-independent, quantitative, reproducible and simple method that can be achieved in a short time. Classical microscopy techniques are effective and reliable to quantify the evolution of RBC morphology but often require long experiments and analysis by qualified technicians. Automatic morphology classification approaches by machine-learning is an active field of research that can be used, for example, to analyze images acquired with a confocal microscope or IFC (Doan et al., 2020; Simionato et al., 2021). The use of IFC and/or artificial-intelligence driven analysis promises a rapid evolution of accurate and fast methods of morphological characterization in the research field. They, however, necessitate specialized equipment, which are not yet present in a majority of clinical hematology laboratories.

Quantification of CFSE^high RBC is also operator-independent and reproducible. It does not directly provide information regarding RBC morphology, but accumulation of CFSE^high erythrocytes was strongly correlated to the proportion of SME detected using IFC. Quantification of CFSE^high RBC only requires a standard flow cytometer: a technology available in both clinical and research laboratories. The staining protocol is fairly simple and therefore has the potential to be adapted and standardized for use in the quality control of red cell concentrates.

Carboxyfluorescein Diacetate Succinimidyl Ester staining reactivity is a new marker of storage lesion that allows quantification and sorting of morphologically altered stored RBC. Future research using this method will explore the SME cellular properties and should help better understand the storage lesion and RBC senescence.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

MM, SP, YH, SG, and MD performed the experiments. MM, SP, YH, SG, MD, CR, OH, PR, PB, and PA analyzed the data. MM, CR, OH, PB, and PA designed the research. MM, SP, and PA wrote the manuscript. PB edited the manuscript. All authors read and approved the manuscript.

PA is funded in part by New Health Science. MM, SP, MD, CR, PB, and PA declare that the European patent application EP21306765 was filed on December 12th, 2021. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.