Hypoxia and hypocapnia storage of γ-irradiated red cell concentrates

Abstract

Background

γ-irradiation is used to treat red blood cell (RBC) concentrates (RCCs) transfused to immunosuppressed patients. This treatment damages RBCs and increases storage lesions. Several studies have shown the beneficial effect of reducing O₂ content during RBC storage. The present research work investigated the effect of γ-irradiation on RCCs stored under normal and hypoxia/hypocapnia conditions.

Materials and methods

O₂ concentration (measured as oxyhaemoglobin fraction, sO₂) and ABO-matched RCCs from whole blood donations, leukoreduced and prepared in phosphate, adenine, glucose, guanosine, saline and mannitol (PAGGSM) were pooled and split in two identical RCCs within 24 h post donation. One bag (Hx) was submitted to O₂ and CO₂ adsorption for 3 h on an orbital shaker at 22±2 °C and then transferred to a storage bag impermeable to gas. The other bag (Ctrl) was left as it was. The two bags were then stored at 4 °C. γ-irradiation (25 Gy) was applied at day 2 or 14, and the RCCs were stored until day 43. Different parameters (metabolites, haemolysis, morphology) were measured.

Results

Starting sO₂ values were 63.7±18.4% (n=12) in Ctrl and 20.8±9.8% (n=12) in Hx bags, and reached 90.8±9.1% and 6.6±5.9% at day 43, respectively. As expected, an increase in glycolysis rate was observed after deoxygenation. Extracellular potassium concentrations were identical and reached around 70 mM at expiry with an irradiation-dependent kinetic release. No difference in haemolysis was observed after irradiation on day 2 in either group (<0.40%, p>0.9999). When irradiated at day 14, haemolysis was lower (p=0.033) in RCCs under hypoxia at the end of storage (day 28, 0.67±0.16%) compared to control (1.06±0.33%). Percentages of spherocytes were lower under hypoxia.

Discussion

The storage under hypoxia provided equivalent storage when RCCs were irradiated at day 2 and was advantageous when irradiated at day 14. In summary, O₂-depletion of RCCs enable a better storage of RBCs, particularly when late irradiation is applied.

INTRODUCTION

γ-irradiation is used to treat red blood cell (RBC) concentrates (RCCs) intended for patients who are immunosuppressed and at risk of transfusion-associated graft-vs-host disease. The transfusion of these irradiated products was associated to non-allergic transfusion reactions with an increased effect due to storage¹.

The γ-irradiation treatments are known to damage RBCs and to increase storage lesions^2–6. In particular, the haemolysis rate increases in γ-irradiated RCCs compared to those untreated, independently of additive solutions (e.g., saline, adenine, glucose and mannitol [SAGM]⁴ and phosphate, adenine, glucose, guanosine, saline and mannitol [PAGGSM]³), and irradiated RBCs are primed for stress-induced apoptotic cell death⁵. Storage-induced lesions⁷ in general impair metabolism^8–13, protein structures and functions^9,14–19, as well as cellular morphology^20,21 in a cascade of events stemming from formation of various reversible and irreversible lesions^22–25. Because of such damage, the storage period after irradiation is shortened^26,27. RCCs can be irradiated up to 28 days post donation, and transfused no later than 14 days post irradiation and no later than 28 days post donation²⁷.

One of the causes of storage lesions is the presence of oxygen in RCCs^7,12,13 where total oxygen content and fraction of oxyhaemoglobin (sO₂) increases during the storage due to O₂ influx through the PVC of the storage bag²⁸. Cell damage, such as protein oxidation, during the storage of RBCs are the hallmark of oxygen-related modifications^15,17,29. Several studies have shown the advantages of removing O₂ from RBCs during storage. Yoshida et al. showed an improvement in ATP content and an increase in glucose consumption with O₂-reduced RCCs made by flushing with Ar, then storing in an anaerobic canister³⁰. These improvements were also reported, including a clear reduction in haemolysis, with another system including gas exchange with a membrane oxygenator followed by RCC storage in a canister filled with N2 and an oxygen sorbent, as well as in a bioreactor under argon³¹. Protein degradations were also prevented, as shown by a 2D gel-based proteomics where RCCs were flushed with He³². Moreover, microfluidic devices showed greater RBC deformability under hypoxia. In these experiments, O₂ in RCCs was first adsorbed in a sorbent over a few hours and then the RCCs were stored in a bag which in turn was placed in an oxygen-barrier bag³³. Interestingly, the level of hypoxanthine, a hallmark of storage lesions⁸ that correlates to 24-h post-transfusion recovery, is greatly reduced under hypoxic storage¹⁰. More recently, Cancelas et al. reported strong and significant correlations between post-transfusion recovery and metabolites such as hexose and pentose sugar phosphate or other lipids, and demonstrated an improved recovery when using hypoxic RBCs³⁴.

This study used a new commercial device to produce hypoxic/hypocapnic RCC. Leukoreduced RCCs were stored under hypoxic conditions in the presence of a PAGGSM additive solution. The aim of this study was to investigate the effects of γ-irradiation on RCCs stored under standard condition (i.e., with a bag permeable to gas) and under hypoxia/hypocapnia (i.e., after reduction of O₂ and CO₂, stored in a bag impermeable to gas) in terms of RBC metabolism and cell morphology.

MATERIALS AND METHODS

Blood products and processing

With the signed consent of the donors, whole blood (450±50 mL) was collected in 63 mL CPD in top-top kits (FQE 6240LU, Macopharma, Tourcoing, France). Whole blood was stored overnight at room temperature and processed the day after donation within 24 h. RCCs were prepared after whole blood filtration and stored in PAGGSM additive solution.

Twenty-four RCCs were selected at day 1 (day 0 being the day of donation), following routine procoedures in our Production Unit. Saturations of oxygen (sO₂) were matched (values for pairs of bags as close as possible) in order to cover a large distribution of sO₂ (see Online Supplementary Table SI)^28,35. Two ABO-matched and %sO₂-matched RCC units were pooled and split (Figure 1). One of the pairs was kept as it was at room temperature (Control; Ctrl) and the other (Hypoxic; Hx) was treated with a Hemanext One^® oxygen reduction bag (Hemanext, Lexington MA, USA; submitted for CE mark), for 3 h at room temperature under agitation at 50 rpm (Adolf Kühner AG, Birsfelden, Switzerland; radius of 50 mm) and transferred to a standard PVC bag which was in turn placed in a gas impermeable bag that included O₂/CO₂ sorbent between PVC and overwrap film following the manufacturer’s instructions. The RCCs were then stored at 4 °C until γ-irradiation on either day 2 or day 14.

Figure 1. Experimental design.

Oxyhaemoglobin saturation and ABO-matched red cell concentrates (RCCs) were pooled and split in a conventional storage group (Ctrl) and an oxygen-reduced group (Hx). γ-irradiation treatments were applied at day 2 or day 14. ORB: oxygen reducing bag; HSB: Hemanext storage bag.

The γ-irradiation was applied in our Distribution service according to the Swiss national guidelines (25 Gy, Gammacell 3000 Elan, Theratronics). The issue dates were day 16 in the case of irradiation at day 2 and day 28 for the irradiation at day 14. RCCs were studied up to 43 days.

There were 6 replicates for each condition. Four study arms were, therefore, studied on PAGGSM leukoreduced RCCs: 6 Ctrl and 6 Hx units irradiated on day 2, and 6 Ctrl and 6 Hx units irradiated on day 14.

Additionally, distribution of the age of irradiated RCCs were analysed. Data were collected from 190 RCCs between January 2019 and March 2019 from our Distribution service.

Oximetry and saturation of oxygen

Oxyhaemoglobin saturation was measured non-invasively using Resonance Raman spectroscopy (Pendar Microvascular Oximeter A3U11, Pendar Technologies, Cambridge, USA). The emitter/sensor was placed on a label-free portion of the bag and covered with a black cloth to block ambient light. Levels of oxy- and deoxy-haemoglobin were recorded at 1,380 and 1,355 cm-1 after excitation at 405 nm, respectively, for instantaneous sO₂ readings.

Sample processing and analyses

At each time point, 5 mL of RCCs were taken with a small bag sterilely docked to the storage bag. Haematological parameters were immediately analysed using an automated haematology blood analyzer (KX-21N, Sysmex). RBC deformability was measured on RCC samples diluted at a haematocrit of 40±0.5% with a microfluidic device (MVA), as previously described³³.

In parallel, part of the RCC samples were centrifuged at 2,000 g at 4 °C for 10 minutes and the supernatants were transferred into another tube. The cell pellets were then washed twice in NaCl 0.9%, following the same centrifugation procedure as before.

On the day of sampling, cell morphologies were assessed on washed RBCs (resuspended in HEPA buffer) by digital holographic microscopy and single-cell analysis using CellProfiler and CellAnalystsoftware³⁶, as previously described¹³.

Aliquots of both the pelleted washed RBCs and the supernatants were snap-frozen in liquid nitrogen and saved at −80 °C for further analyses. The Hx samples were collected and processed under argon atmosphere in a mobile disposable gloves box.

Glucose, lactate, urate and K⁺ were quantified at the clinical chemistry laboratory (at the Lausanne University Hospital). CD47 positive microvesicles were quantified in supernatants by flow cytometry (FACScalibur flow cytometer, BD Biosciences), as previously described¹⁷. Haemolysis was measured using the Harboe method by spectrophotometry (NanoDrop 2000c)¹³.

ATP and 2,3-DPG quantifications were carried out on frozen RBCs applying a slightly modified protocol of the ATPlite Luminescence ATP Detection Assay System kit (PerkinElmer, Groningen, the Netherlands), and using the 2,3-DPG kit (Roche, Mannheim, Germany) following the manufacturer’s protocol (see the Online Supplementary Appendix for details).

Data analyses and statistics

GraphPad Prism version 6.07 (GraphPad Software Inc.) were used for data and statistical analyses. Initial glycolytic rates were calculated using a non-linear fitting with an exponential decay as follows:

$$y=(yo-plateau)\hspace{0.17em}{e}^{-kt}+plateau$$ (1)

$$initiale\hspace{0.17em}rate=-k\times yo$$ (2)

where y stands for the glucose concentration in mM, y0 the concentration at day 0, plateau the y value at infinite times (not equal to 0), k the rate constant in 1/day, and t the time in days.

RESULTS

Storage age distribution of γ-irradiated red cell concentrates

According to an established internal rule, RCCs must be irradiated and transfused as follows: for RCC age ≤14 days post donation, the expiry date is 14 days after the irradiation. In practice, RCCs were irradiated between 3 and 13 days; average 7.4±2.3 days (see Online Supplementary Appendix, Online Supplementary Table SII and Online Supplementary Figure S1). Irradiated units are usually distributed and transfused on the same day. Only 2% of the γ-irradiated RCCs were distributed 1–9 days post irradiation (average 6 days).

O₂ reduction and sO₂ evolution

After the incubation process to remove O₂, the sO₂ levels decreased in proportion to the initial values (Figure 2). The average sO₂ was 63.7% (min 38, max 90, n=12) for Ctrl and 20.8% (min 5, max 36, n=12) for Hx. At day 43, sO₂ values increased in conventional bags (Ctrl; normoxia) to 90.8% (min 71, max 99, n=12) and decreased to 6.6% (min 0, max 16, n=12) in Hx bags (hypoxia) as expected²⁸. The γ-irradiation processes had no effect.

Figure 2. Efficiency of oxyhaemoglobin saturation (sO2) reduction and effect of storage time.

Level of reduction (a) and correlation to starting values (b). Evolution of sO₂ in function of storage time for the irradiated red cell concentrates at day 2 (c) and day 14 (d). Empty squares: storage under hypoxia (Hx); black dots: conventional storage (Ctrl). Mean values±standard deviation, n=6.

METABOLISM AND IONS

The effect of hypoxia on the glycolysis was quantified during the storage. The consumption of extracellular glucose and production of lactate was increased under hypoxia as shown in Figure 3 (top panel) and Table I. The initial glycolytic rates were increased 2-fold under hypoxia compared to control. There was a clear correlation between initial sO₂ (day 2) and initial rate of glucose consumption (Pearson r=0.83 and p<0.0001) for Ctrl. In contrast, there was no correlation between initial sO₂ and initial glycolytic rate in the sub-group Hx (Pearson r=0.44 and p=0.39). Despite the increased glycolytic rate, the glucose was not exhausted, and the final levels remained around the physiological concentration. ATP levels were increased in the first week of storage followed by a decrease. The levels were higher in Hx (significant in a few cases) than in Ctrl. 2,3-DPG concentrations were significantly higher in Hx storage compared to Ctrl. These important differences were likely due to CO₂ depletion during the O₂ removal process that produces a hypocapnia and triggers the 2,3-DPG production by increased pH.

Figure 3. Effect of storage conditions and γ-irradiation on red blood cells (RBCs).

(Top panel) Quantification of metabolites in irradiated red cell concentrates (RCCs) at day 2 and day 14. (Bottom panel) Morphological analysis and deformability in irradiated RCCs at day 2 and day 14. Proportion of discocytes (a and e) and spherocytes (b and f) analysed by digital holographic microscopy (see main text for details). Deformability was based on the flow rate (Q) through microchannels (c and g) and ΔQ (d and h) stands for (Q[Hx] – Q[Ctrl]) for each microchip. Blue squares: storage under hypoxia (Hx); red dots: conventional storage (Ctrl). Mean values±standard deviation, n=6. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Table I.

Initial glycolytic rates

	Mean initial rate/mM/day	Standard deviation/mM/day
Ctrl irradiated at day 2	−0.94	0.13
Ctrl irradiated at day 14	−0.90	0.19
Hx irradiated at day 2	−2.23	0.11
Hx irradiated at day 14	−1.95	0.19

Ctrl: Control; Hx: Hemanext; n=6.

The effect of irradiation was not clearly observed on glycolysis. Even though not significant, immediately after irradiation, a slight increase in 2,3-DPG and a decrease in ATP levels were observed in Hx, which was the opposite in Ctrl. As for the glycolysis kinetics, the initial rates were independent of the γ-irradiation (Table I and data in Online Supplementary Appendix), in spite of small differences noted in the Hx condition. The initial rate of irradiated Hx RCCs (irradiated at day 2) was −2.23 mM/day compared to −1.95 mM/day for non-irradiated Hx RCCs (only irradiated at day 14).

The extracellular urate trends for RCCs in all arms were similar compared to conventional RCCs, as reported by Bardyn et al.³⁷. It was slightly higher in Hx than in Ctrl after irradiation.

The extracellular K⁺ levels were equivalent in both cases. The concentrations reached 66±2 mM in Ctrl and 66±3 mM in Hx at day 16 when irradiated at day 2; and 72±3 mM in Ctrl and 72±4 mM in Hx at day 28 when irradiated at day 14. These differences resulted from the γ-irradiation independently of oxygen levels. The irradiation increased K⁺ release kinetics (see the Online Supplementary Appendix and Online Supplementary Figure S3 for a direct comparison). The release was faster after irradiation and the rates were in agreement with those in the literature^2,3.

Haemolysis

Haemolysis in γ-irradiated RCCs increased over time and reached values higher than untreated products (usually <0.4%). No difference was observed after irradiation on day 2 between the two products and haemolysis remained equivalent through the end of storage (Figure 4). At the issue date, i.e. at day 16, the haemolysis was 0.37±0.24% for Ctrl and of 0.34±0.15% for Hx. Haemolysis remained <0.8% (0.62±0.23%) for hypoxic units at day 43 compared to 0.95±0.57% for conventional storage (p=0.083).

Figure 4. Haemolysis and microvesicles counts in function of γ-irradiation and storage.

Haemolysis rates for the irradiated red cell concentrates (RCCs) at day 2 (a) and day 14 (b), respectively (c) and (d) for microvesicles. Empty squares: storage under hypoxia (Hx); black dots: conventional storage (Ctrl). Mean values±standard deviation, n=6. *p<0.05, ***p<0.001, ****p<0.0001. Dashed line stands for the specification where haemolysis has to be <0.8%.

When irradiated at day 14, a sharp increase in haemolysis was observed after the treatment in both storage conditions. Half of the Ctrl RCCs were out-of-specifications immediately after irradiation and two-thirds at 7 days post irradiation, whereas all the Hx RCCs remained within the specification for the same period. At day 28, i.e. 14 days after the irradiation, 5 of 6 Ctrl RCCs and 2 of 6 Hx RCCs were above 0.80% of haemolysis. The difference was more pronounced than when irradiated at day 2. Haemolysis was lower in Hx RCCs at the end of storage (day 28) with a value of 0.67±0.16% compared to Ctrl with a 1.06±0.33% (p=0.033). At day 43, the haemolysis rates were of 1.60±0.44% and of 0.87±0.22% in Ctrl and Hx, respectively (p<0.0001). Storage under hypoxia improved the haemolysis by 0.2–0.4% when RCCs were irradiated at day 14.

Finally, the quantity of microvesicles in supernatant exhibited the same trends as haemolysis (Figure 4).

Cell morphology and deformability

Cell morphologies (Figure 3, bottom panel, a, b, and e, f) were similar in both conditions except for spherocytes that were more abundant in Ctrl compared to Hx bags. Over time, the rate of spherocyte formation was higher in Ctrl compared to Hx. At day 43 of storage, 7.0% and 2.5% of the cells were spherocytes for the irradiation at day 2, and 9.1% and 3.4% for the irradiation at day 14, for Ctrl and Hx, respectively.

Deformability (Figure 3, bottom panel, c, d and g, h) based on flow rates through microchannels decreased during storage as previously reported³³. It was slightly higher under hypoxia/hypocapnia, although these differences did not reach statistical significance.

DISCUSSION

Despite the recommendation from European Directorate for the Quality of Medicines and HealthCare (EDQM)²⁷ and the Swiss national guidelines that permit blood centres to irradiate RCC up to 28 days post donation, the window period was limited to 14 days in our institution for a technical IT reason. The present experiments focused, therefore, on the two extremes: 2 days and 14 days post donation. As expected and reported by others^2,3, γ-irradiation induced the release of around 4–7 mM/day potassium (release of approximately 2–3 mM/day for non-irradiated RCCs). The final values were slightly above that recently observed by de Korte et al. (even in PAGGSM additive solution)⁶ but remained equivalent to that reported by Hauck et al. in PAGGSM³. In contrast to the other parameters measured here, potassium levels were not influenced by storage under hypoxia and hypocapnia. Pre-storage reduction of O₂ content in RCCs was achieved using an adsorption device at room temperature under agitation. In all cases, the sO₂ was lower after the adsorption process but did not reduce the level of O₂ to 0%.

Previous experiments explored the effects of removing O₂ down to < 3% s O₂ level³⁰, rerouting the metabolism away from the pentose phosphate pathway³⁸. Despite the partial O₂ reduction, the lower sO₂ was sufficient to better preserve the RBCs during storage, notably by reducing the oxidative stress⁷. Moreover, the known oxygenation of RBC during the storage due to gas permeable storage bags was blocked in hypoxic storage bag (Figure 2).

One of the main differences stemming from hypoxic RBC storage was seen in the glycolysis rate. A clear increase (slightly more than 2-fold) was observed under hypoxia that favours glycolysis. The reduction in oxygen (and carbon dioxide) in RBCs has been shown to increase glycolytic rate and partially inhibit the shift to the pentose phosphate pathway^{7,9,30,38–41}. Carbon dioxide depletion restores a more neutral pH, which favours glycolytic enzyme activity³⁸. In addition, the regulation of these enzymes is O₂ dependent. The binding of phosphofructokinase, aldolase and glyceraldehyde 3 phosphate dehydrogenase to band 3 is in competition with deoxyhaemoglobin, as shown by Low et al.^40,42. Storage-induced oxidative stress oxidises glyceraldehyde 3 phosphate dehydrogenase and reduce glycolytic flux⁹. Moreover, hypoxia also promotes the consumption of citrate and other carboxylates¹⁰. Consequently, the concentration in ATP was higher in hypoxic RCCs compared to normoxic RCCs. As for 2,3-DPG, the significant increase in Hx condition was mainly due to the removal of CO₂ during the adsorption process (hypocapnia) that releases 2,3-DPG.

Urate was slightly higher in Hx than in Ctrl after irradiation. Because of the anti-oxidant properties of urate, some differences could have been expected between the hypoxia and normal storage. The absence of O₂ should reduce the oxidative stress and preserve the urate level. However, other mechanisms linked to urate and purine metabolism could be involved in the antioxidant defences. Other metabolites should be quantified to explore this.

Marked effects of hypoxia were highlighted by the haemolysis rates and number of microvesicles, especially in the differences between early and late irradiation. Indeed, storage under hypoxia was non-inferior to conventional storage when γ-irradiation was applied at day 2, but better when irradiated at day 14. Since the differences become more apparent with storage time, it can be hypothesised that the RBCs stored under hypoxia are better preserved (as shown by the beneficial impact on energy metabolism) and thus are able to resist stress resulted from 25 Gy of the irradiator. It can be speculated that the same effect would be observed for treatments using X-ray.

From a practical point of view, an additional 4 h (3 h of incubation plus 1 h of processing) is required to treat the RCCs. In spite of this processing, RCCs were processed within the 24 h post donation in agreement with the recommendations. The kit includes all the necessary components and it is used with a sterile connection device. No chemical is added and storage is performed as for conventional RCCs in the original additive solution. PAGGSM was used here instead of SAGM. Both additive solutions are in routine used in our blood bank but the former one was chosen in order to follow the manufacture’s recommendation. PAGGSM additive solution provides improved storage conditions because of the presence of phosphate and guanosine. Haemolysis and percentage of spherocytes were reported to be higher (after 56 days of storage) in SAGM compared to other additive solutions, including PAGGSM⁴³. Recent developments have seen that alkalinization enables superior efficiency in storage⁴⁴. Therefore, the combination of hypoxia and an efficient additive solution is a good option to improve RBC storage. Finally, it has to be noted that the repeated sampling during the study could increase the stress suffered by the RBCs in the bags. A quality control check showed that haemolysis of conventional γ-irradiated RCCs remained <0.8% at day 14 post irradiation. This fact does not alter the conclusions claimed here since both RCCs were processed in parallel.

CONCLUSIONS

Modifications of RBC properties following γ irradiation were assessed on conventional and hypoxic storage with average depletion of sO₂ from 64% to 21%. Hypoxia has a beneficial effect on RBC storage due to a decrease in oxidative stress and improvement of different parameters such as ATP, 2,3-DPG, haemolysis and morphology²⁸. Hypoxic storage provided equivalent storage when RCCs were irradiated at day 2 but showed an advantage when RCCs were irradiated at day 14. The improved storage methods likely provided a higher capacity of RBCs to resist to irradiation-induced lesions. Further in-depth metabolomics and proteomics investigations may provide a mechanistic understanding of this observation.

In summary, storage of γ-irradiated RCCs under hypoxia/ hypocapnia is not inferior to conventional storage when irradiated at day 2, and is superior when γ-irradiation is applied at day 14. Combining irradiation with hypoxia/ hypocapnia retained the improved haemolysis profile of O₂-depleted RBCs. Since late irradiation clearly damaged the RBCs, it is of particular interest to reconsider the storage of irradiated RBCs for transfusion.