Abstract

   Redox imbalance and oxidative stress have emerged as generative causes
   of the structural and functional degradation of red blood cells (RBC)
   that happens during their hypothermic storage at blood banks. The aim
   of the present study was to examine whether the antioxidant enhancement
   of stored RBC units following uric (UA) and/or ascorbic acid (AA)
   supplementation can improve their storability as well as
   post-transfusion phenotypes and recovery by using in vitro and animal
   models, respectively. For this purpose, 34 leukoreduced CPD/SAGM RBC
   units were aseptically split in 4 satellite units each. UA, AA or their
   mixture were added in the three of them, while the fourth was used as
   control. Hemolysis as well as redox and metabolic parameters were
   studied in RBC units throughout storage. The addition of antioxidants
   maintained the quality parameters of stored RBCs, (e.g., hemolysis,
   calcium homeostasis) and furthermore, shielded them against oxidative
   defects by boosting extracellular and intracellular (e.g., reduced
   glutathione; GSH) antioxidant powers. Higher levels of GSH seemed to be
   obtained through distinct metabolic rewiring in the modified units:
   methionine-cysteine metabolism in UA samples and glutamine production
   in the other two groups. Oxidatively-induced hemolysis, reactive oxygen
   species accumulation and membrane lipid peroxidation were lower in all
   modifications compared to controls. Moreover, denatured/oxidized Hb
   binding to the membrane was minor, especially in the AA and mix
   treatments during middle storage. The treated RBC were able to cope
   against pro-oxidant triggers when found in a recipient mimicking
   environment in vitro, and retain control levels of 24h recovery in mice
   circulation. The currently presented study provides (a) a detailed
   picture of the effect of UA/AA administration upon stored RBCs and (b)
   insight into the differential metabolic rewiring when distinct
   antioxidant “enhancers” are used.

   Keywords: RBC storage lesion, Glutathione metabolism, Oxidative stress,
   Antioxidant supplementation, Uric acid, Ascorbic acid

Graphical abstract

   [51]Image 1
   [52]Open in a new tab

Highlights

     * •
       Uric and ascorbic acids protect stored red blood cells from
       oxidative damage.
     * •
       GSH synthesis is differentially regulated after uric and ascorbic
       acid addition.
     * •
       The antioxidant advantage is sustained in recipient-mimicking
       conditions.

1. Introduction

   Red blood cells (RBCs) stored under blood bank conditions present an
   array of biochemical and physiological alterations, collectively named
   storage lesions, known to affect their post-transfusion efficacy. The
   root of this structural and functional deterioration, which is mainly
   fueled by the lack of plasma components and hypothermic conditions, is
   the redox imbalance [[53]1,[54]2]. Indeed, the presence of oxygen in
   stored RBCs leads to the formation of free radicals, which promote
   damages in structural proteins and enzymes, including energy-related,
   antioxidant and proteostatic proteins, a phenomenon that further
   aggravates the redox imbalance [[55][3], [56][4], [57][5]]. This
   dysregulation in redox and energy metabolism can lead to elevated
   hemolysis post-transfusion [[58]6,[59]7], a complication which can in
   turn begin a cascade of adverse events [[60]8,[61]9].

   In the last years, there is augmenting research activity in the field
   of transfusion medicine regarding the effect of specific metabolites
   and antioxidants upon stored RBCs in the context of ameliorating the
   detrimental impact of oxidative stress on cell integrity, metabolism
   and functionality. Recently, through a high-throughput metabolomics
   platform, the addition of several metabolites was applied to test their
   influence upon energy and redox metabolism. For example,
   supplementation of stored RBCs with the antioxidant N-acetylcysteine
   seemed to beneficially impact the pentose phosphate pathway (PPP),
   while the addition of glutamine did not alter glutathione (GSH)
   synthesis, with both results being supported by other studies
   [[62][10], [63][11], [64][12]]. Incubation of stored RBCs in fruit and
   vegetable extracts showed that the bio-flavonoid naringin protected
   them from lysis and externalization of removal signals [[65]13],
   whereas polyphenol quercetin had a minor impact [[66]14]. Accordingly,
   use of nanoparticles containing antioxidant enzymes resulted in the
   improvement of the storability profile of RBCs [[67]15], as also
   observed in the case of GSH loading [[68]16].

   Uric acid (UA), a radical species scavenger, represents the main
   non-enzymatic antioxidant in human plasma, with beneficial effects upon
   circulating RBCs [[69]17]. Notably, the extracellular UA appears to be
   rapidly increased in the first week of storage [[70]18], while stored
   RBCs donated by individuals with elevated serum UA have proven to cope
   better with storage stress [[71]19,[72]20]. Nonetheless, to our
   knowledge, the enhancement of RBC units exclusively with UA has not
   been studied yet. In the same context, plasma ascorbic acid (AA) is an
   exceptionally effective antioxidant, able to protect plasma lipids
   against peroxidation and maintain RBC hemoglobin (Hb) in a reduced
   state [[73]21]. Its addition to stored RBCs has been found protective
   against RBC fragility [[74]22] and 2,3-bisphosphoglycerate depletion
   [[75]23]. In most cases, AA supplementation was examined in combination
   with other antioxidants, resulting in improved GSH preservation and
   reduced hemolysis [[76]24,[77]25]. Interestingly, a recent study
   demonstrated that restoration of normal plasma levels of UA and AA in
   RBC units achieved a partial prevention of the detrimental metabolic
   shift observed during RBC storage [[78]26]. Having all of the above in
   mind, and the fact that UA and AA can act synergistically [[79]27], the
   aim of the present study was to evaluate the effects of
   higher-than-physiological concentrations of UA and AA separately, as
   well as in combination, upon (a) an array of RBC storability
   parameters, including hemolysis, membrane vesiculation, proteostasis,
   redox profile, metabolism and protein recruitment to the membrane, and
   (b) the −unknown to date− post-transfusion RBC physiology and recovery,
   by using elegant in vitro and animal models of transfusion.

2. Materials and methods

2.1. Biological samples and blood unit preparation

   Thirty-four leukoreduced RBC units containing
   citrate-phosphate-dextrose (CPD)/saline-adenine-glucose-mannitol (SAGM)
   (final volume: 338 ± 27 mL) were prepared from healthy individuals
   (female/male ratio: 10/24). Each RBC unit was then split under aseptic
   conditions into four sub-units (resulting in n = 136 smaller units) of
   equal volume (79 ± 5 mL each). One sub-unit from each donor was used as
   (untreated) control, while the other three were supplemented with UA
   (in-bag concentration: 8 mg/dL for male and 7.2 mg/dL for female
   donors; diluted in 0.01 M NaOH), AA (in-bag concentration: 2.3 mg/dL;
   diluted in 0.9% NaCl) or their mixture. In this paired analysis, unit
   splitting contributed to absorbing the well-known donor variation
   effect upon several redox and physiological parameters of stored RBCs
   [[80][28], [81][29], [82][30]]. The different levels of UA
   supplementation in RBC units from male and female donors were chosen on
   the basis of the known intersex differences in the normal range of
   serum UA concentration in vivo [[83]31], to avoid excessive amounts of
   uric acid that could potentially harm the stored RBCs from female
   donors. Since previous studies of our team [[84]19,[85]20] revealed
   that RBCs from donors with increased serum UA levels demonstrate
   superior storability profile, we performed preliminary experiments in
   freshly drawn RBCs to select the optimum concentrations (higher than
   the normal-range in vivo values) with respect to hemolysis, pH and ROS
   accumulation. In order to prevent immediate contact of the
   supplementation stock solution with the cells (and thus unwanted
   effects due to exposure to high solution concentrations) the blood
   units were left overnight for RBC settling, and the antioxidants were
   added in the supernatant. All units were stored for 42 days at 4 °C and
   sampling was aseptically performed after gentle agitation in early (day
   7), middle (day 21) and late (day 42) storage. No RBC settling was
   observed in any of the preparations during the 42-day storage period.
   All experiments were performed in n = 34 per group, unless otherwise
   stated. Regarding the experiments with lower number of tested samples,
   the selection was completely random and not based on the outcome of
   other experimental procedures. The study was approved by the Ethics
   Committee of the Department of Biology, School of Science, NKUA.
   Investigations were carried out upon donor consent, in accordance with
   the principles of the Declaration of Helsinki.

2.2. Hemolysis Parameters

   The levels of spontaneous hemolysis were calculated as extracellular Hb
   using the spectrophotometric method firstly shown by Harboe [[86]32],
   along with Allen's correction. In order to evaluate the resistance of
   RBCs to osmotic lysis, the samples were subjected to ascending NaCl
   concentrations to finally calculate their mean corpuscular fragility
   (MCF; %NaCl at 50% hemolysis). Accordingly, to assess the mechanical
   fragility, the samples were rocked with stainless steel beads for 1h,
   while non-rocked aliquots were used to subtract spontaneous Hb release.
   After centrifugation, the levels of extracellular Hb were calculated
   through Harboe's method [[87]33].

2.3. Redox parameters

   The antioxidant capacity of the supernatant (namely, the extracellular
   compartment of the RBC unit that contains the
   anticoagulant/preservative solutions and a small volume of the donor
   plasma) was analyzed by the ferric reducing antioxidant power (FRAP)
   assay [[88]34], with and without prior treatment with uricase.
   Oxidative hemolysis was calculated post exposure of the samples to
   phenylhydrazine (PHZ; 17 mm) for 1h at 37 °C. Afterwards, a
   centrifugation was performed to measure the released Hb (Harboe's
   method). Intracellular reactive oxygen species (ROS) accumulation was
   evaluated by fluorometry (VersaFluor™ Fluorometer System, BIORAD
   Hercules, CA, United States) at 490 nm excitation and 520 nm emission
   wavelengths. Small aliquots (∼1% hematocrit) of RBCs were treated with
   5-(and-6)-chloromethyl-2′,7′-dichloro-dihydrofluoresceindiacetate,
   acetyl ester (CM-H[2]DCFDA; 10 μmol/L for 30 min at room temperature),
   a molecule that emits fluorescence when oxidized by ROS. This was also
   performed for RBCs previously incubated with tert-butyl hydroperoxide
   (tBHP; 100 μM for 45 min at 37 °C). Three quick washes with PBS of
   310mOsm at 1000×g were performed to remove the excess of tBHP, while
   after the incubation with DCFDA, the aliquots were washed once and were
   then incubated for a short recovery time of 10–15 min at room
   temperature to render the dye responsive to oxidation. The fluorescence
   units were thereafter normalized to protein concentration (Bradford;
   Bio-Rad, Hercules, CA). Hemichromes (HMC) were detected
   spectrophotometrically on isolated RBC membranes by measuring heme
   absorbance through the following equation: HMC = −133xAbs[577]
   -114xAbs[630] +233xAbs[560] [[89]35]. Evaluation of membrane lipid
   peroxidation was based on the formation of a chromogenic complex
   between malondialdehyde (MDA), a biomarker of lipid peroxidation, and
   thiobarbituric acid (TBA). Briefly, the lipid part of RBCs was
   retrieved after subjection to trichloroacetic acid (TCA; 20%) and was
   then treated with 0.67% TBA. The chromogenic complex was measured at
   532 nm [[90]36].

2.4. Membrane and vesicle isolation and immunoblotting

   Hypotonic lysis was performed to isolate RBC membranes (n = 6 per
   group), as previously extensively described [[91]37]. Briefly, packed
   RBCs (∼1.6 mL after centrifugation at 1000×g/10 min and supernatant
   disposal) were exposed for 45 min (1:20 ratio for membrane isolation;
   1:10 for cytosol isolation) to hypotonic sodium phosphate buffer
   (5 mmol/L, pH 8.0) supplemented with protease inhibitors (0.3 mM
   phenyl-methyl-sulfonyl fluoride; PMSF). A centrifugation was then
   performed at 19,000×g for 20 min, after which cytosols were collected
   and stored, while the precipitated membranes were washed under the same
   conditions to remove the excess of Hb, until full discoloration of the
   pellet. Samples of supernatant from late-stored RBCs (starting volume
   of 8 mL) were ultra-centrifuged at 37,000×g for 1 h, after passing
   through sterile 0.8 μm nitrocellulose filters (Millipore,
   Carrigtwohill, County Cork, Ireland), to isolate extracellular vesicles
   (EV; n = 6 per group). The EV pellet was resuspended in PBS and was
   ultracentrifuged twice under the same conditions. Finally, the isolated
   EVs were resuspended in PBS along with protease and phosphatase
   inhibitors (1 mM PMSF, 1/100 protease inhibitor cocktail, 2 μL/mg
   phosphatase inhibitor cocktail; Sigma, St. Louis, MO) and the protein
   concentration was determined (by the Bradford method) to calculate
   total vesicular protein per RBC unit volume [[92]38].

   Equal amount of isolated membranes (30 μg) or EVs (40 μg) were loaded
   in 10% Laemmli gels and transferred to nitrocellulose membranes for
   immunoblotting. Primary antibodies against 4.1R (kindly provided by
   Prof. J. Delaunay, Laboratoire d’ Hématologie, d’ Immunologie et de
   Cytogénétique, Hopital de Bicetre, Le Kremlin-Bicetre, France), Band 3
   (Sigma-Aldrich, Munich, Germany), Hb (Europa Bioproducts, Wicken, UK),
   stomatin (kindly provided by Prof. R. Prohaska, Institute of Medical
   Biochemistry, University of Vienna, Austria), Caspase-3, DJ-1, CD9
   (Cell Signaling Technology, Danvers, MA, USA), calpain-1, heat shock
   protein 70 (HSP70; Santa Cruz Biotechnology, Santa Cruz, CA, USA),
   peroxiredoxin-2 (prdx2; Acris, Luzern, Switzerland) and human IgGs
   (Sigma Aldrich, St. Louis, MO, United States) were used. The
   immunoblots were developed through chemiluminescence, thus
   species-specific HRP-conjugated secondary antibodies were also used.
   The evaluation of the developed bands was assessed by scanning
   densitometry (Gel Analyzer v.1.0, Athens, Greece). In order to estimate
   the carbonylation of membrane proteins, the Oxyblot kit was used, as
   per manufacturer's specifications (Oxyblot, Millipore, Chemicon,
   Temecula, CA, USA).

2.5. Intracellular calcium, procoagulant EVs and RBC morphology

   The accumulation of Ca^2+ in the cytosol was analyzed through
   fluorometry using Fluo-4 AM (2 μmol/L final concentration for 40 min at
   37 °C; Invitrogen, Molecular Probes, Eugene, OR), an ester that emits
   fluorescence when interacting with calcium cations [[93]39]. The
   measured fluorescence units were normalized to protein concentration
   (Bradford assay). The procoagulant activity of EVs was evaluated using
   a functional ELISA kit (Zymuphen MPactivity, Hyphen BioMed,
   Neuvillesur-Oise, France). The principal on which this assay is based
   is the conversion of prothrombin to thrombin in the presence of
   phosphatidylserine on EVs, a phenomenon that is captured by using a
   chromogenic thrombin substrate (absorbance at 405 nm). Morphological
   evaluation of the late-stored RBCs (n = 3) was performed by confocal
   laser scanning microscopy (CSLM; Digital Eclipse C1, Nikon, NY)
   following labeling with the lipophilic dye D-383
   (1,1′-Didodecyl-,3,3′,3′-Tetramethylindocarbocyanine Perchlorate), as
   per manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). The
   fluorescence excitation and emission maxima for membrane-bound dye were
   549 nm and 565 nm, respectively. The percentage of discocytes in each
   sample was evaluated in duplicate assays after counting at least ∼600
   RBCs per group in randomly chosen fields.

2.6. Proteasome activity

   The activity of the proteasome machinery was measured in both cytosol
   and membrane fractions (n = 6 per group). To determine the three
   distinct proteasome activities, the samples were incubated with the
   fluorogenic substrates Suc-Leu-Leu-Val-Tyr-aminomethylcoumarin (AMC)
   for chymotrypsin-like (CH-like), z-Leu-Leu-Glu-AMC for caspase-like
   (CASP-like), and Boc-Leu-Arg-Arg-AMC for trypsin-like (TR-like)
   activity for 1.30 h (CH-like) or 3 h (CASP- and TR-like) at 37 °C
   [[94]30]. The same procedure was followed in the presence of inhibitors
   (10–20 μM bortezomib for the CH- and CASP-like activities, 200 μM
   MG-132 for the CH-like activity, and 100 μM lactacystin for the TR-like
   activity) showing high levels of inhibition (93–98%). All substrates
   and inhibitors were procured from Enzo Life Sciences (New York, NY,
   United States). Fluorescent units (after subtraction of the unspecific
   levels) were normalized to protein concentration (Bradford assay).

2.7. Metabolomics analysis

   Each sample (n = 6 per group) was added to 1000 μl of a
   chloroform/methanol/water (1:3:1 ratio) solvent mixture stored at
   −20 °C. The tubes were mixed for 30 min and subsequently centrifuged at
   1000×g for 1 min at 4 °C, before being transferred to −20 °C for 2–8 h.
   The solutions were then centrifuged for 15 min at 15,000×g and dried to
   obtain visible pellets. Finally, the dried samples were resuspended in
   0.1 mL of water, 5% formic acid and transferred to glass autosampler
   vials for LC/MS analysis. Twenty-microliter of extracted supernatant
   samples was injected into an ultra-high-performance liquid
   chromatography (UHPLC) system (Ultimate 3000, Thermo) and run in
   positive ion mode. A Reprosil C18 column (2.0 mm × 150 mm, 2.5 μm - Dr
   Maisch, Germany) was used for metabolite separation. Chromatographic
   separations were achieved at a column temperature of 30 °C and flow
   rate of 0.2 mL/min. A 0–100% linear gradient of solvent A (ddH[2]O,
   0.1% formic acid) to B (acetonitrile, 0.1% formic acid) was employed
   over 20 min, returning to 100% A in 2 min and a 1-min post-time solvent
   A hold. The UHPLC system was coupled online with a mass spectrometer
   Q-Exactive (Thermo) scanning in full MS mode (2 μscans) at 70,000
   resolution in the 60–1000 m/z range, target of 1 × 106 ions and a
   maximum ion injection time (IT) of 35 ms. Source ionization parameters
   were: spray voltage, 3.8 kV; capillary temperature, 300 °C; sheath gas,
   40; auxiliary gas, 25; S-Lens level, 45. Calibration was performed
   before each analysis against positive ion mode calibration mixes
   (Piercenet, Thermo Fisher, Rockford, IL) to ensure sub ppm error of the
   intact mass. Data files were processed by MAVEN.8.1
   ([95]http://genomics-pubs.princeton.edu/mzroll/) upon conversion of raw
   files into mzXML format through MassMatrix (Cleveland, OH).

2.8. Invitro Model of Transfusion

   An in vitro model of transfusion was used to evaluate the impact of
   recipient plasma and body temperature on stored RBCs, as previously
   extensively described [[96]40]. For this purpose, fresh blood was drawn
   by ten healthy potential recipients into citrate vacutainer tubes.
   Then, packed RBCs of early and late storage (n = 10 per group) were
   mixed with these plasma samples. A portion of the stored unit's
   supernatant was added in the fresh recipient plasma to reach a ratio
   that mimics the administration of two blood units per transfusion
   recipient, resulting in 32–34% final hematocrit. The samples were
   incubated at body temperature (simulation of the temperature
   environment of the recipient) for 24h (the standard time point of
   transfusion evaluation). Hemolysis and oxidative stress-related
   parameters were then assessed.

2.9. Animal model of transfusion

   All the experiments regarding the xenobiotic model of transfusion were
   performed in the Biomedical Research Foundation of the Academy of
   Athens (BRFAA), and the study protocol was approved by the Department
   of Agriculture and Veterinary Service of the Prefecture of Athens
   (Permit Number: 534915, July 23, 2020). For the evaluation of 24 h
   post-transfusion recovery, a total of 24 wild type C57BL/6J male mice,
   8–12 weeks old, were used, as previously described [[97]40]. Briefly,
   RBCs of early and late storage (n = 6 per group) were labeled with the
   lipophilic dye D-383, as per manufacturer's instructions, washed 3
   times, and then infused into recipient-mice by intravenous injection in
   the tail vein. The volume infused (200 μL at ∼55% hematocrit with
   sterile PBS of 310 mOsm) is respective to the transfusion of two blood
   units. To assess the 24h recovery of the transfused RBC through flow
   cytometry (FACSAria IIu/Diva software, BD Pharmingen, San Jose, CA,
   USA), blood sampling was performed (∼50–70 μL) via the facial vein
   after 20 min (100% recovery) and 24 h post-transfusion. For the flow
   cytometry analysis, the blood samples were diluted (1:200) in a PBS 310
   mOsm buffer containing 5% glucose and 1% BSA, and were then filtered
   for debris removal. The fluorescence excitation and emission maxima for
   membrane-bound dye were 549 nm and 565 nm, respectively. A minimum of
   1,000,000 events per sample was counted and the recovery was calculated
   as the % ratio of the fluorescence levels of the 24h time point to the
   respective levels of the 20 min measurement. The animals were evaluated
   weekly regarding their well-being (weight, mobility, food and water
   consumption, social behavior).

2.10. Statistical analysis

   All experiments were performed in duplicate. The statistical package
   SPSS Version 22.0 (IBM Hellas, Athens, Greece, administered by NKUA)
   was used for the statistical evaluation of the results. All parameters
   were tested for normal distribution and the presence of outliers
   (Shapiro-Wilk test and detrended normal Q–Q plots). In order to present
   as reliable as possible results regarding the differences between the
   selected treatments the following strategy was followed: in the
   presence of extreme outliers the respective values were excluded from
   the analysis and the statistical test of choice was repeated. If the
   statistical outcome was not affected by the presence of the outlier,
   its value was considered “valid” and was re-entered to the data set. In
   the opposite case, the outlier was removed. Two-way repeated measures
   ANOVA with Bonferroni-like adjustment for multiple comparisons was used
   for the determination of the between groups differences throughout the
   storage period. Significance was accepted at p < 0.05. In the case of
   metabolomics analyses, and in order to identify which metabolic
   pathways were mostly affected in the treated samples compared to the
   control groups, the MetPA (Metabolomic Pathway Analysis) bioinformatic
   tool included in the MetaboAnalyst 5.0 Software package, which combines
   the results from the pathway enrichment analysis with the pathway
   topology analysis, was employed. False discovery rate (FDR) was used
   for controlling multiple testing (q-value threshold was ≤0.05).
   Accepted metabolites were verified manually using HMDB, KEGG, and
   PubChem databases. A Human library was used for pathway analysis
   (KEGG). Global test was the selected pathway enrichment analysis
   method, whereas the node importance measure for topological analysis
   was the relative betweenness centrality.

3. Results

3.1. Physiological, redox and proteostasis parameters

   The addition of UA, AA or their combination did not alter the levels of
   released Hb, while the presence of AA slightly increased the RBC
   osmotic fragility during storage (e.g., ∼0.5% elevation during early
   storage), but decreased the mechanical fragility of long-stored cells
   ([98]Fig. 1A). The supplementations tested did not affect extracellular
   pH ([99]Fig. 1B), nor the intracellular accumulation of calcium and the
   recruitment of the calcium-dependent protease calpain-1 to the membrane
   ([100]Fig. 1C). The number of discocytes in the modified units of late
   storage was also similar to the control ([101]Supplementary Fig. 1). On
   the contrary and as expected, the total antioxidant capacity (TAC) of
   the supernatant was increased in all cases in comparison to control,
   while its UA-dependent and UA-independent aspects were mostly elevated
   in the presence of UA or AA, respectively ([102]Fig. 1D).

Fig. 1.

   [103]Fig. 1
   [104]Open in a new tab

   Antioxidants supplementation effects upon qualitative parameters of
   stored red blood cells. Hemolysis parameters (A), pH (B), calcium
   accumulation along with calpain-1 recruitment to the membrane (C) and
   extracellular antioxidant capacity (D) during storage of red blood
   cells under standard conditions or upon supplementation with uric acid
   (UA) and/or ascorbic acid (AA). Representative immunoblots are shown (n
   = 6). 4.1R protein was used as internal loading control. Data are
   presented as mean ± SD. C: control samples. (For interpretation of the
   references to color in this figure legend, the reader is referred to