In vitro-transfusional model for red-blood-cell study: the advantage of lowering hematocrit

Abstract

Background

The quality of red blood cells (RBCs) stored in red cell concentrates (RCCs) is influenced by processing, storage and donor characteristics, and can have a clinical impact on transfused patients. To evaluate RBC properties and their potential impact in a transfusion setting, a simple in vitro-transfusional model has been developed.

Materials and methods

Transfusion was simulated by mixing a washed RBC pool from two male-derived RCCs stored at 4°C with a pool of 15 male-derived fresh frozen plasma (FFP) units, representing the recipient, at a hematocrit (HCT) of 30% (“control” setting) or 5% (alternative model). The mixtures were incubated at 37°C, 5% of CO₂ up to 48 h. Different metabolites, hemolysis and microvesicles (MVs) were quantified at several incubation times and RBC-morphology changes and deformability after incubation. For each model, biological triplicates have been investigated with RCCs at storage days 2 and 43.

Results

The 5%-HCT model restored the 2,3-DPG level and maintained the ATP level. Furthermore, glucose consumption and corresponding lactate production were increased in the 5%- vs the 30%-HCT condition. Lower hemolysis was observed with 5%-HCT, but only at day 2. However, morphological analysis by digital holographic microscopy (DHM) revealed a decreased fraction of discocytes at 5% rather than at 30% of HCT at storage day 2 but at day 43, the trend was inverted. Concordantly, RBCs incubated at 5% of HCT were more deformable than at 30% at day 43 (p<0.0001).

Discussion

Higher metabolic activity of RBCs in the 5%-HCT condition was promoted by a higher glucose availability and limited cell-waste accumulation. The conditions of the new proposed model thus enabled rejuvenation of RBCs and maintained them in a physiological-close state in contrast to the 30%-HCT model. It may be used as a first approach to evaluate e.g., the impact of donor and recipient characteristics on RBC properties.

INTRODUCTION

Although the quality of blood products is strictly regulated¹, transfusion-related side effects such as intravascular-hemolytic reactions or transfusion-related lung injury (TRALI) may occur^2–4. These outcomes, amongst others, might be related to individual recipient characteristics but also to the quality of the blood product^2,3,5. The quality of such products is related to the processing methods including centrifugation and separation steps to obtain red cell concentrates (RCCs), blood-donor characteristics and the storage time (usually ranging from up to 42 or 49 days at 4°C) ^3,6–10.

Storage lesions³ are a cascade of events leading to metabolism modifications¹¹, protein alterations^12,13 and morphological changes^3,14–16. For instance, low temperature and waste accumulation, acidifying the pH, decrease metabolic activity. Therefore, glycolysis and thus adenosine-triphosphate (ATP) levels decrease. Furthermore, intact mature RBCs, called discocytes, characterized by a flexible morphology, become spherocytes which are smaller and more rigid^14,15.

In addition, donor characteristics including lifestyle, ethnicity, genetic inheritance, age and sex influence RCC quality^17,18. For example, endurance training correlates with higher RBC deformability¹⁹. Stored RBCs from Hispanic individuals show increased oxidative hemolysis compared to Asian donors¹⁷. Regarding genetic factors, van t’Erve et al.²⁰ identified metabolites, whose levels during storage depend on heritability. As a further example of age- and gender-related differences, RBCs from older donors show decreased oxidative hemolysis and women’s RBCs hemolyse less than male-derived RBCs²¹.

To study the properties of RBCs from fresh blood or RCCs, incubation in classic solutions like NaCl or phosphate-buffered saline (PBS) can be performed^22–24. These conditions, however, are far away from the physiological conditions, which rapidly induce hemolysis. To understand the molecular impact of RCCs in transfused patients, the quality-based molecular consequences in vivo after transfusion need to be considered in addition to the intrinsic ex vivo quality of RCCs. For this purpose, animal models are suitable but do not fully recapitulate human physiology and remain costly^25–28. Bioreactors have also been investigated as alternatives^29,30 but are not always easy-to-use and may be expensive. For these reasons, we developed a simple in vitro-transfusional model based on RBC incubation in plasma, in flasks under a controlled environment, and performed the downstream analyses for metabolites, hematological and morphologic parameters for validation.

The aim of the presented model was to simulate in vivo-transfusion conditions that allow the RBCs to regain their physiological metabolic activity, a process called “rejuvenation”. This may allow to study RBC properties, in physiological-close conditions, dependent on donor (here the RBCs) and recipient (here the plasma) characteristics³¹.

MATERIAL AND METHODS

Blood products

After donor consent, whole blood was processed with the top/bottom method (component filtration) as previously described⁷. Leukoreduced RCCs in saline-adenine-glucose-mannitol (SAGM) solution and fresh frozen plasma (FFP) only from men were chosen to reduce sex variability. RCCs and FFP were stored at 4°C and −30°C, respectively.

In vitro-transfusional model

Transfusions were simulated by mixing RBCs (representing the donor) with a pool of FFP (representing the recipient) (Figure 1).

Figure 1.

Experimental design of in vitro-transfusion simulations

Firstly, 50 mL from RCCs at storage day 2 were centrifuged at 2,000 × g f or 10 min at 4°C (Rotanta 460R, Hettich, Tuttlingen, Germany) and the pellets were washed twice with NaCl 0.9% (Bichsel AG, Interlaken, Switzerland). Pellets from two RCCs were pooled and mixed with a pool of 15 FFP units at a hematocrit (HCT) of 30%, close to the physiological HCT (“control” model), or 5% (alternative model) in a final volume of 140 mL in vented-cap T-flasks. The “control” model refers to the work of Tzounakas et al.³¹ and was adapted for this study. The “reconstituted blood” was then incubated at 37°C (CO₂ Water Jacketed Incubator Series II, Forma Scientific Inc., Marietta, OH, USA) under agitation (STD 3500 Shaker, VWR, Leuven, Belgium) and 5% of CO₂ up to 48 h. Each transfusion model was performed with six RCCs at storage day 2 and repeated at day 43.

Sampling

Samples of 5 mL were taken before incubation (time point “−1”) from RCCs and 15 mL from the flasks at different times of incubation. After centrifugation at 2,000 × g for 10 min at 4°C, the supernatants were again centrifuged whereas the RBC pellets were washed as described above. Supernatants and pellets were snap-frozen in liquid nitrogen and stored at −80°C.

Standard blood parameters

Standard blood parameters including RBC concentration, hemoglobin (Hb) and HCT were measured in fresh RBC pellets and in fresh “reconstituted-blood” samples using a blood analyser (Sysmex KX-21N, Sysmex Digitana AG, Kobe, Japan).

Morphology and deformability analyses

As described elsewhere¹⁵, samples from incubations were diluted in HEPA buffer and analyzed as quintuplicates on coated 96-well plates with a digital holographic microscope (DHM® T1000, Lyncée Tec SA, Lausanne, Switzerland). This is a quantitative-phase-microscopy technique using a low-intensity laser source and allowing a label-free and minimally invasive visualisation of cells. The measured optical-path-difference correlates to the morphology and the refractive index of cells³². Subsequently, the DHM images were treated with the softwares CellProfiler (2.0 r11710, Cimini Lab, Broad Institute of MIT and Harvard, Cambridge, MA, USA) for cell segmentation and feature extraction and CellProfiler Analyst as previously described³³.

The deformability of RBCs is represented by the Elongation Index (EI) that was measured in triplicates in function of an increasing shear stress or in function of a stable shear stress (30 Pa) in combination with an osmolality gradient (Osmoscan) by ektacytometry (Lorrca® MaxSis, RR Mechatronics, Zwaag, the Netherlands)³⁴. To make RBC concentration sufficient for analysis, samples from 48-hour incubations were centrifuged at 2,000 × g for 5 min at 4°C. Subsequently, a part of the supernatant was removed from the 5%-HCT samples to adjust to 30%. According to the manufacturer’s instructions, 25 μL of the sample were mixed with the manufacturer’s Elon ISO solution and 900 μL of this mix were analyzed. For osmolality analysis, 200 μL of the sample were mixed with the Elon ISO solution.

Intracellular ATP quantification

Samples were deproteinized with the Deproteinizing-Sample-Preparation Kit (BioVision, Milpitas, CA, USA) and treated with the ATPlite-Luminescence-ATP-Detection-Assay-System kit (PerkinElmer, Groningen, the Netherlands). Luminescence was measured on duplicates with a microplate reader (Spectramax M3, Molecular Devices, Silicon Valley, CA, USA) at a 1000-ms-integration time.

Intracellular 2,3-DPG quantification

2,3-DPG quantification was carried out on RBC pellets using the 2,3-Diphosphoglycerate (2,3-DPG) kit (Roche, Mannheim, Germany). Absorbance was measured at 340 nm with the Nanodrop 2000C (Thermo Scientifc, Wilmington, DE, USA).

Extracellular glucose and lactate quantification

Glucose and lactate were measured in the supernatants with a Cobas® 8000 (Roche Diagnostics, Rotkreuz, Switzerland) by the “Service de Chimie Clinique” (CHUV, Lausanne, Switzerland). The deltas of glucose and lactate concentrations stand for the consumption and the production, respectively.

Hemolysis rate

The determination of the hemolysis rate was based on Harboe’s method as previously described^15,35.

Quantification of microvesicles

Microvesicles (MVs) were quantified in centrifuged supernatants, which were obtained as described in the Sampling section. In duplicates, 30 μL were mixed with 90 μL of antibody mix, containing 2.2 μg/mL of mouse anti-human CD235a PE (BD Pharmingen™, San Diego, CA, USA) and 10 μL of CD47 FITC (BD Pharmingen™), in NaCl 0.9%, and incubated for 20 min in the dark under gentle agitation. After transfer of 90 μL in Trucount™ Tubes (BD Biosciences, San Jose, CA, USA) containing 400 μL of 0.9% NaCl, measurements were recorded with a flow cytometer (FACSCantoTM II, BD Biosciences) at medium flow until 5.000 counted beads. The resulting MV concentrations were adjusted in function of the MV quantity in the initial FFP pool and corrected by the ratio of RBC quantity in both models.

Data treatment and statistics

Graphs and statistics were carried out with the software GraphPad Prism 6 (GraphPad Software Inc., San Diego, CA, USA). Statistical analyses were based on multiple comparisons with Two-Way ANOVA using a significance threshold of p ≤0.05.

RESULTS

Metabolic activity of RBCs

First, standard blood parameters were measured in the RCCs used for this study (Online Supplementary Content, Table SI), and were shown to align with previously published data^15,36. To determine whether the incubations enable rejuvenation, the glycolytic activity of the RBCs placed in plasma at 5% or 30% of HCT was quantified. Figure 2 shows the quantification of glucose consumption, lactate production and intracellular concentrations of ATP and 2,3-DPG of the incubated RBCs.

Figure 2.

Comparison of different metabolites between the transfusional models at 5% and 30% of HCT

RBCs from RCCs (stored at 4°C) at storage days 2 and 43, were incubated with FFP up to 48 h (37°C, 5% CO₂ under agitation). Metabolites were quantified before (“−1”) (before RBC and FFP mix) or during incubation in the supernatants (extracellular) or in the RBC pellets (intracellular) after washing wit 0.9% NaCl.

N=3; p-values: * <0.05, *** <0.001, **** <0.0001.

Glucose consumption was always higher in RBCs at 5% than at 30% of HCT (p=0.0108 [day 2] and 0.0243 [day 43]) although glucose was not limited in the 30%-condition (Online Supplementary Content, Figure S1A). At the beginning of incubation, the 5%-HCT condition triggered a significantly higher consumption of glucose at storage day 2. At storage day 43, this difference was maintained with a slightly lower glucose consumption in both conditions than at day 2. Of note, the consumption stopped after 48 h of incubation at 30%-HCT.

In accordance with the glucose consumption, the lactate production was always superior at the 5%-HCT condition (p=0.0002 [day 2] and 0.0012 [day 43]). Although the glucose consumption decreased from 0 to 24 h of incubation at storage day 2, the lactate production was stable.

Concerning the ATP level, both conditions showed an increase thus a rejuvenation at 4 h of incubation compared to the RBCs before incubation (point −1). RBCs at 5%-HCT showed highest ATP levels at storage days 2 (p=0.0049) and 43 (p=0.0006). At storage day 43, ATP levels were lower in both conditions compared to day 2 and even depleted at the end of incubation in the 30%-HCT condition.

Unlike ATP, the 2,3-DPG production displayed rejuvenation but only in the case of 5%-HCT, showing especially a high increase (3.6 fold at 4 h) at storage day 43. As for the other metabolites, the 2,3-DPG levels were lower at day 43. Moreover, at the end of incubation at storage day 2 and especially at day 43, the difference between both models was lost.

In brief, the glucose consumption and the lactate production, as well as the intracellular levels of ATP and 2,3-DPG were higher in RBCs diluted at 5% than at 30% of HCT.

Hemolysis and MVs

Hemolysis rates and MVs increased with storage (Online Supplementary Content, Table SI) and incubation time (Figure 3). At storage day 2, the 5%-HCT-transfusional model exhibited a significant lower hemolysis rate despite a higher MV count compared to the 30%-condition. However, at storage day 43, the hemolysis and MV concentration were mainly equivalent in both models.

Figure 3.

Hemolysis rate and MV count during incubation at 5% and 30% of HCT

Hemolysis rates were calculated according to Harboe’s method³⁵. MVs were quantified in supernatants by flow cytometry.

N=3; p-values: * <0.05, ** <0.01.

Morphology of RBCs in RCCs and after transfusion simulations

The morphology was compared between RBCs before (RCCs) and after transfusion simulations by DHM (Figure 4). The RBCs stored in RCCs at 4°C were mostly discocytes with some exceptions at storage day 45 where a few spherocytes appeared, as expected¹⁵ (Figure 4A). In contrast, the 48-hour incubation at 37°C induced an increase of the spherocyte fraction. Whereas the proportion of spherocytic RBCs appeared higher at the 5%- than at the 30%-HCT condition at the storage beginning, the trend was inverted at day 43 (Figure 4A and B). In accordance, RBCs at 5%-HCT showed diminished discocyte and stomatocyte fractions in contrast to the 30%-condition at storage day 2 but echinocyte fractions were always predominant in the 30%-condition. In summary, the morphological damages were higher at 5% than at 30% of HCT at day 2 but lower at 5%-HCT at the end of storage.

Figure 4.

Morphology of RBCs in RCCs and incubated RBCs at 5% or 30% of HCT

Samples from RCCs stored at 4°C or from RBCs after 48 h of incubation were diluted in HEPA buffer and analyzed by DHM. (A) DHM images and (B) proportions of discocytes, stomatocytes, echinocytes I, II/III, spheroechinocytes and spherocytes were performed with CellProfiler® (single cells are shown as example). Images from RCCs were acquired at day 4 (beginning of storage) and day 45 (end of storage). N=3; p-values: ** <0.01, **** <0.0001.

Deformability of RBCs after transfusion simulations

RBCs from RCCs stored at 4°C (Online Supplementary Content, Figure S1B) as well as RBCs after 48-hour incubation were analyzed by ektacytometry to evaluate their deformability capacity (Figure 5). The effect of mechanical shear stress is shown in Figure 5A and of increasing osmolality in Figure 5C. Figure 5B provides interpretation elements of the measurements.

Figure 5.

RBC deformability after incubation at 5% or at 30% of HCT

The deformability (elongation index (EI)) of RBCs was measured at storage days 2 and 43 after 48-hour incubation by ektacytometry (Lorrca®) in function of an increasing shear stress (A) or with a stable shear stress and an increasing osmolality (C). (B) curve interpretation of (C).

N=3; p-values: * <0.05, **** <0.0001.

The deformability of RBCs decreased slightly with storage time at 4°C (Online Supplementary Content, Figure S1B), as expected. After 48-hour incubation, RBCs from the 5%-HCT condition were only slightly more flexible than at 30% of HCT at storage day 2, in spite of an important ATP drop in the 30%-condition. However, the former showed a significantly higher deformability at storage day 43. As well during increasing osmolality, RBCs incubated at a HCT of 5% kept a higher EI.

DISCUSSION

Transfusional-model considerations

Our objective was to develop a simple transfusional model best recapitulating physiological conditions in order to analyse RBC properties in a transfusion setting. The targeted physiological conditions were the temperature, the regain of RBCs’ metabolic activity (rejuvenation), glucose availability and low cell waste concentration. A transfusional model at a low HCT (5%) was chosen allowing high glucose availability and limiting waste accumulation such as lactate. FFP and RCCs from male donors were used to avoid variability attributable to sex and were pooled in order to mitigate inter-individual donor differences.

Other in vitro models have been developed recovering physiological aspects such as glucose 6-phosphate dehydrogenase deficiency³¹, the impact of RBCs on T cell activity^37,38 or the impact of temperature change during transfusion³⁹. They included quantification of ROS³¹, NADP(H)³⁹ and cytokines^37,38. Samples were incubated in tubes³⁹, on plates³⁸, in a fluidized-bed bioreactor³⁰ or in flasks³⁷ up to 40°C³⁹ and the time varied from some minutes⁴⁰ to six days with up to 15% of hemolysis³⁷. The used solution were SAGM³⁹, culture medium^37,38 or plasma³¹. Although plasma would be closer to in vivo conditions, we used FFP for practical reasons and the higher glucose content allows an increased metabolic activity. However, the used FFP was characterized by slightly increased pH values, 7.8 and 7.9, compared to physiological plasma pH around 7.4. During incubations, the pH decreased up to 7.0, especially in the 30%-condition due to a higher cell content producing acidifying lactate. Bicarbonate level, which is also implicated in the acid-base equilibrium, changes the pH. However, this effect is probably less important than the lactate effect, and no drastic pH changes were recorded during the incubations. To confirm the acid-base equilibrium and to quantify the dissolved CO₂, that was set to 5% in the air, bicarbonate levels should have been measured. The pH may have affected metabolic pathways such as glycolysis and especially its pH-dependent phosphofructokinase⁴¹. However, the carbonic anhydrase present in RBCs may have rapidly re-established intracellular acid-base equilibrium⁴².

RBCs are metabolically more active at 5% of HCT

The increased glucose consumption and lactate production in the 5%-HCT condition may result from the increased quantity of available glucose, due to a higher volume of plasma in this condition. In fact, the initial glucose concentration was equivalent in both conditions. However, the glucose concentration decreased rapidly to 6.77±0.40 mM in the 30%-condition, due to a higher concentration of RBCs, whereas the 5%-condition conserved 16.57±0.12 mM of glucose (Online Supplementary Content, Figure S1A). Another advantage of lowering cell concentration is the reduced accumulation of cell waste products such as lactate that may inhibit glycolysis⁴³. Indeed, lactate-induced medium acidification can alter the activity of key glycolytic enzymes such as phosphofructokinase. Interestingly, at storage day 2, lactate production was stable during the first 24 h of incubation although the RBCs from the 5%-HCT condition showed a decreased glucose consumption. In general, the lactate:glucose ratio corresponds to 2 as seen in the 30%-condition at storage day 2. This ratio, however, was of 1.4 at 4 h in the 5%-HCT condition. One explanation might be that the first glucose intake was not completely employed in glycolysis, thus leading to a lower lactate production, or that glycolysis was by-passed. Another explanation may be a preferential entry of pyruvate into the urea cycle followed by glutathione synthesis or purine metabolism, or alternatively, pyruvate accumulation followed by lactate dehydrogenase inhibition. At 24 h, glucose consumption of RBCs in the 5%-condition at day 2 decreased leading to a lactate:glucose ratio close to 2. In fact, glucose intake through GLUT1 is regulated by glucose concentration and temperature, both of which can modulate the availability of glucose for the ATP production⁴⁴. Consequently, a decreased intake might result from a negative feedback regulation of GLUT1 caused by an increased ATP concentration. Interestingly, the 30%-condition showed, after 48 h at day 2, a lactate:glucose ratio >2 as the conditions at day 43 the first 24 h. Indeed, lactate may not only originate from glycolysis but also from citrate or other metabolism pathways⁴⁵.

The increase of ATP and 2,3-DPG levels after 4 h of incubation, especially in the 5%-HCT condition, corresponding to rejuvenation^3,46, may result from an increased glucose consumption. Decreased ATP and 2,3-DPG levels could be explained by an increased lactate accumulation downregulating the glycolytic activity⁴³ (Online Supplementary Content, Figure S1A). The rapid drop down of ATP- and 2,3-DPG-concentrations at storage day 43, especially at 30% of HCT, may originate from an interplay of high lactate levels and oxidative damages accumulated during cold storage rendering RBC less functional.

In conclusion, the 5%-HCT condition allowed an improved metabolic activity resulting in rejuvenation as reported in vivo⁴⁶, when compared to the “control” model. Metabolomic analyses would be required to investigate deeper the intracellular modifications that are influenced by several regulations.

Unscathed RBCs in both in vitro-transfusion conditions

As expected, hemolysis as well as MV counts increased with storage and incubation times but did not always correlate (Figure 3). Indeed, Tzounakas et al.⁴⁷, revealed an inverse correlation between these parameters that may be due to their spatiotemporal mode of apparition during storage. Whereas microvesiculation starts early to discard oxidized and toxic molecules^12,48, hemolysis appears later as a consequence of excessive membrane loss and related morphology alteration. This inverse correlation was only low to moderate and was observable from hemolysis levels higher than 0.17% in the case of CPDA units. Whereas previous publications^9,15,36 did not confirm an inverse correlation, the present study shows this trend but only clearly at the beginning of incubation. The here used RCCs had relatively low hemolysis levels (Online Supplementary Content, Table SI). RCCs that are more hemolytic may induce higher hemolysis and less MVs (inverse correlation) during incubation. Otherwise, centrifugation steps, during RBC-pellet preparation before incubation, could remove apoptotic RBCs so that similar hemolysis and MV count could be observed.

Although the RBCs were metabolically more active at 5% than at 30% of HCT, these data showed mainly no differences between both incubation conditions. These results indicate that an increased metabolism could not counteract hemolysis and microvesiculation suggesting that factors other than the metabolism, such as accumulated oxidative lesions, influence these processes. Nevertheless, both conditions showed limited hemolysis rates indicating that they are suitable to study RBCs.

Unclosed morphological differences between the transfusion conditions

DHM demonstrated an increased fraction of spherocytes for both HCT-conditions when compared to RCCs at 4°C (Figure 4). This may be an effect of the temperature or mechanical stress in flasks due to agitation. It has also been reported that DEHP, a plasticizer of RCC bags, reduces hemolysis⁴⁹. Interestingly, the size of discocytes from incubations was decreased compared to those from RCCs (data not shown). It demonstrates an effect of plasma that should be further investigated.

The images as well as the analyses with CellProfiler indicate that the 5%-condition seems to conserve better RBC morphology at storage day 43 and that, contrary to what was expected, the 30%-condition might be more favourable at storage day 2.

A recent study showed that rejuvenation could, by increasing ATP, revert storage-induced microerythrocytes (SMEs), a hallmark of storage damage⁵⁰, to physiological RBCs. The decreased fraction of spherocytes in the 5%-condition vs the 30%-condition at day 43 could be due to a re-conversion of SMEs to physiologic RBCs. However this explanation does not match with storage day 2 where more spherocytes were observed in the 5%-condition whereas the 30%-condition showed more discocytes despite the absence of ATP after 48 h of incubation. The higher proportion of irreversible shapes at 5%-HCT thus cannot be explained by the ATP level. Additional analyses will be required including SMEs quantification to investigate further this phenomenon.

5%-HCT condition better preserves deformability

The morphological types of RBCs are characterized by a different level of deformability^14,15.

The 5%-condition seems to preserve better the deformability during incubation. Interestingly, the deformability differences between both conditions were higher at day 43 than at day 2 despite low ATP levels for the 30%-condition at both days. Deformability depends on ATP levels, however, the relation between ATP levels and deformability capacity may be a correlation but not necessarily a causality, at least not at hundred percent. Indeed, ektacytometry measures passive deformability that showed only a low correlation with ATP⁵¹. Furthermore, Barshtein et al.⁵² mentioned some publications where deformability of RBCs was not affected by cold storage. ATP may not be the sole factor and there may be a minimum ATP concentration necessary to maintain “normal” deformability. The ATP concentration may still be high enough in the 30%-condition just some hours before the end of incubation allowing having a deformability close to the 5%-condition. In addition, RBCs in an intact state (beginning of storage) may be more able to counteract ATP lack by other factors that may explain that at storage day 2 RBCs from both conditions showed a similar deformability level. As in some publications⁵², long-stored RCCs kept a high deformability similar to fresh RBCs⁵³ (Online Supplementary Content, Figure S1B) because cold storage slows down the apparition of storage lesions.

The deformability in function of an osmolality gradient was also higher at the 5%-condition. Moreover, as shown in Figure 5B and C, the RBCs at 5%-HCT have, on average, a higher cell surface and cell volume that are characteristic features of discocytes. These findings thus corroborate the morphological data at day 43. Furthermore, the shift of the EI peak, in function of osmolality, indicates a difference of the ion-channel function between both incubation conditions (Figure 5B, C). We speculate that the improved deformability at the 5%-condition is related to increased ATP levels which are known to be involved in ion transport and protein phosphorylation⁵⁴. Weed et al.⁵⁵ have shown that decreased ATP levels lead to a loss of deformability and to an increase of intracellular calcium, of RBC-suspension viscosity and a decrease of RBC filterability. The regeneration of ATP reversed some of these abnormalities. Several elements of the cytoskeleton are sensitive to Ca²⁺ concentrations. For instance, adducin interacts with the Ca²⁺-calmodulin complex impacting actin and spectrin properties⁵⁶. It remains, however, unclear why, on a morphological point of view, the 30%-condition seems to be more favourable at storage day 2 despite contradictory deformability data. One explanation could be linked to the higher ATP levels in the 5%-condition, which might improve the deformability of irreversible cell types, which can be found at increased fractions at storage day 2.

Since ATP was depleted after the 48-hour incubation at 30%-HCT, experimental conditions do not recapitulate physiological status where the ATP level remains at a steady-state. A portion of RBC lesions could be delayed after this energy depletion and could remain undetectable in our experiments. Indeed, metabolic dysfunction appears within the first two weeks of storage in RCCs at 4°C however its consequences, e.g., on hemolysis and other morphological defects, appear only during the two last weeks of storage. By analogy, the 48-hour incubation might be too brief to allow the measurement of changes. This model could be adapted in term of incubation time etc. depending on the tested hypothesis. Furthermore, this model represents a closed system, in contrast to physiological conditions, explaining lactate accumulation especially in the 30%-HCT condition. In addition, the glucose concentrations being higher than in vivo may have additional effects.

CONCLUSIONS

Lowering hematocrit at 5% in transfusion simulations (RBCs incubated in plasma) is an alternative to simulate transfusions in nearly physiological conditions in contrast to a 30%-HCT model. The 5%-HCT condition enabled rejuvenation of RBCs showing an increase of ATP and 2,3-DPG levels and the maintenance of energy metabolism as well as the deformability during incubation. Consequently, this model is useful, as a first approach, without a complex configuration to compensate glucose availability and to limit waste accumulation, to study the impact of donor and recipient characteristics or the influence of different treatments on RBC properties in a transfusion context.