Proof of concept: hypoxanthine from stored red blood cells induces neutrophil activation

Abstract

Background

Red blood cell (RBC) units may contain a variety of molecules that can activate the neutrophil cascade turning neutrophils into targets for immunomodulatory molecules. Our metabolomics profiling of RBC units revealed a significant increase of hypoxanthine concentration during storage. Hypoxanthine catabolism in vivo ends with the production of uric acid through a reaction catalysed by xanthine oxidase during which reactive oxygen species are generated. Some authors have described in vitro neutrophil activation after treatment with stored RBC medium. However, the response of neutrophils to the action of xanthine oxidase upon hypoxanthine accumulation in the supernatant of RBC units has never been investigated.

Materials and methods

Neutrophils were isolated from peripheral whole blood and cultured at 37 °C in a humidified incubator with 5% CO₂. Hypoxanthine and RBC supernatants were tested to verify neutrophil stimulation. To prove the involvement of hypoxanthine in neutrophil activation, xanthine oxidase was pre-incubated with or without allopurinol before addition to the neutrophil cultures. Intracellular expression of tumour necrosis factor-α (TNF-α) and interleukin-8 (IL-8) was assessed by a cytofluorimetric assay and early-stage release of IL-8 was detected by a Luminex^® assay.

Results

In the presence of xanthine oxidase, hypoxanthine, alone and in combination with RBC supernatants, caused increases of TNF-α- and IL-8-positive cells after 5 hours of treatment. Moreover, IL-8 was quickly released, 30 min after stimulation.

Discussion

Here we show, for the first time, that neutrophil activation by stored RBC depends, in part, on the presence of hypoxanthine contained in the RBC units. Our results add hypoxanthine to the already known mediators of inflammation present in RBC units, supporting the evidence that medium from stored RBC may concur to boost inflammatory processes in transfusion recipients, potentially leading to negative post-transfusion outcomes.

INTRODUCTION

During storage, red blood cells (RBC) undergo biological and biochemical changes collectively referred to as “storage lesions”^1,2. Despite observational studies suggesting a relationship between RBC storage lesions and post-transfusion complications^3–7, this matter is still controversial and is further complicated by pre-existing comorbidity as well as genetic variability in recipients’ responses to transfusion². The type of additive solution, the presence of white blood cells, the duration of storage and the number of transfused RBC units are all expected to affect the onset of negative outcomes (i.e. mortality, infections, and longer time spent in hospital)³.

In previous studies, we found that the consumption of adenine, present in the conservation medium SAGM (sodium chloride, adenine, glucose and mannitol), by stored RBC is associated with a concomitant increase of its metabolic intermediate, hypoxanthine (HX). After 42 days of storage, the concentration of HX in the units could reach 1 mM^8–10, while plasma levels usually range between 1 and 8 μM¹¹. Storage-derived accumulation of HX negatively correlates with recovery following RBC transfusion¹², and seems to be involved in the oxidative damage observed in several pathological conditions including acute respiratory distress syndrome and ischaemia^13,14, although recent evidence indicates alternative scenarios¹⁵.

HX is metabolised by xanthine oxidase (XO) to xanthine, and then to uric acid, with contemporary production of reactive oxygen species (ROS: O₂^•- and H₂O₂). Physiological levels of plasma XO are low, but increase significantly in the presence of damage to intestine, liver or endothelial cells of the microvasculature¹⁶. In these patients, the transfusion of 42-day old RBC could result in increased production of ROS and, eventually, to transfusion-related endothelial dysfunction and tissue injury, such as transfusion-related acute lung injury (TRALI)^10,17. This life-threatening transfusion effect, in particular, seems to be related to priming of neutrophils and their sequestration in the lungs^18–20.

Neutrophils, besides being the first responders against pathogens, are the first type of cell that migrates to sites of injury, and play a role in the acute injury repair process. They act as phagocytic cells, produce high amounts of reactive oxygen intermediates with antimicrobial activity and generate a number of inflammatory factors including tumour necrosis factor-α (TNF-α) and interleukin-8 (IL-8)^20,21.

Some authors have described the in vitro activation of neutrophils after treatment with RBC storage medium^22,23. In particular, the supernatant from leukoreduced RBC units induces neutrophil priming, which might be attributed to some-not yet fully identified-bioactive substances, such as arachidonic acid and its derivatives^24,25.

Since both intracellular and extracellular ROS play a role in neutrophil activation^26,27, we hypothesised that the increase of HX due to multiple transfusions of old RBC might trigger neutrophil activation in patients with high levels of XO¹⁰. The present study aims to verify whether HX could be an additional mediator of neutrophil activation induced by the medium of stored RBC²².

MATERIALS AND METHODS

Study population

The study was approved by the Reggio Emilia Ethics Committee on 21/01/2013. Written informed consent, in accordance with the Declaration of Helsinki, was obtained from five volunteer blood donors.

Preparation of red blood cell supernatants and quantification of hypoxanthine

The exact amount of HX in each supernatant, from leukodepleted and leukoreduced RBC units, was measured by proton nuclear magnetic spectroscopy (¹H-NMR). The results of the quantification were published recently^8–10.

Neutrophils isolation

Neutrophils were isolated from human EDTA-anticoagulated whole blood by using the manufacturer’s instructions (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany)²⁸. The efficiency of neutrophils separation was evaluated by flow cytometry, by analysing the expression of neutrophil-specific cell surface markers. Briefly, 50 μL of enriched neutrophils, collected in duplicate, were incubated with 5 μL of phycoerythrin-conjugated anti-CD15 and allophycocynanin-conjugated anti-CD16 (Miltenyi Biotec GmbH) human monoclonal antibodies for 10 min in the dark at room temperature. Next, 2 mL of Lysing Solution (Stem-Kit Reagents, Beckman Coulter, Brea, CA, USA) were added to avoid interference from residual RBC. After 10 min of incubation at room temperature, 50 μL of calibration beads (LeukoSure Fluorospheres, Beckman Coulter) were added and CD15⁺/CD16⁺ samples were acquired by flow cytometry (see Online Supplementary Figure S1 as an example).

Neutrophil culture and stimulation

Isolated neutrophils were cultured at a final concentration of 2×10⁶ cells/mL in RPMI1640 medium (Euroclone, Pero, Italy) supplemented with 0.5% foetal bovine serum (Euroclone) for further treatments. In order to verify the involvement of HX oxidation in neutrophil activation, a pathological concentration²⁹ (10 mU/mL) of xanthine oxidase (XO, Sigma Aldrich, St. Louis, MO, USA) was pre-incubated with or without allopurinol (100 μM, Sigma Aldrich), a specific XO inhibitor, 4 hours before addition to the neutrophil cultures. HX was used at a concentration similar to that reached after transfusion of three 42-day old RBC units, i.e. 200 μM. To simulate the transfusion of three old RBC units, we used 42 day-old RBC supernatants obtained by centrifugation of RBC units collected in a previous study (approved by the Reggio Emilia Ethics Committee on 21/01/2013), and stored at −80 °C. Isolated neutrophils were cultured at a final concentration of 2×10⁶ cells/mL and treated as described in Table I.

Table I.

Scheme of the treatments performed on isolated neutrophils cultured at the concentration of 2×10⁶ cells/mL

Sample	Treatment
	LPS 1μg/mL	Hypoxanthine 200 μM	Xanthine oxidase 10 mU/mL	42-day old RBC supernatant	Allopurinol 100 μM
Negative control
CTRL	✓
HX		✓
HX + XO		✓	✓
HX + XO + ALL		✓	✓		✓
SUP				✓
SUP+XO			✓	✓
SUP+XO +ALL			✓	✓	✓

CTRL: control; HX: hypoxanthine; XO: xanthine oxidase; ALL: allopurinol; SUP: supernatant.

Neutrophil cultures were incubated at 37 °C in 5% CO₂ and left to incubate with stimulation medium for 30 min or 5 hours.

Intracellular expression of tumour necrosis factor-α and interleukin-8

Neutrophilcultures were incubated at 37 °Cin 5% CO₂. After 5 hours of incubation in the different conditions described previously and reported in Table I, cells were collected and centrifuged at 350 g for 5 min at room temperature and pellets were collected for the flow cytometric evaluation of the intracellular expression of TNF-α and IL-8.

Neutrophils were fixed and permeabilised using a FIX & PERM Cell Fixation & Cell Permeabilization Kit, according to the manufacturer’s instructions (Thermofisher Scientific, Waltham, MA, USA). Subsequently, 1×10⁶ permeabilised and fixed cells were re-suspended in 100 μL of permeabilisation solution and incubated for 30 min in the dark at room temperature with 1:20 fluorescein isothiocyanate-conjugated anti-TNF-α (Affymetrix, eBioscience, Thermofisher Scientific) and phycoerythrin-conjugated anti-IL-8 (Miltenyi Biotec GmbH) monoclonal antibodies. After incubation, cells were washed and re-suspended in 500 μL phosphate-buffered saline (Euroclone). Flow cytometry was carried out using a CX-500 flow cytometer (Beckman Coulter, Pasadena, CA, USA), and the plots were analysed by CXP cytometry software. Lipopolysaccharide was used as a positive control.

Luminex^® assay

At the end of the 30 min incubation time, cells were collected and centrifuged at 350 g for 5 min at room temperature. and supernatants were stored at −20 °C until undergoing a Luminex^® assay. The levels of IL-8 and TNF-α, released by the cells into the supernatants, were evaluated by commercially available Luminex® kits according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN, USA).

Statistical analysis

Results are expressed as the mean ± standard deviation. An analysis of variance multiple comparisons test and a non-parametric Student’s τ-test were applied to compare the means of the conditions tested. Differences were considered statistically significant for p values <0.05 (*p<0.05, **p<0.01 and *** p<0.001).

RESULTS

In the present study we examined the direct effects of HX contained in 42-day old RBC supernatants on isolated human neutrophils in the presence of XO, mimicking a pathological condition³⁰.

Neutrophils were isolated from peripheral whole blood of five healthy donors by using a specific isolation kit enabling fast and untouched isolation of human neutrophils that guarantees high purity and high recovery of enriched neutrophils without affecting the function of the cells²⁸. The assay of the neutrophil-enriched fractions is reported in Online Supplementary Figure S1.

The activation of human neutrophils was assessed cytofluorimetrically by measuring TNF-α and IL-8 expression and verified using lipopolysaccharide, a well-known stimulus for neutrophil activation²⁵ (see Online Supplementary Figure S2). Indeed, human neutrophils display an elevated auto-fluorescence that can be erroneously confused with a positive cytokine signal³¹.

To evidence the role of HX, neutrophils were also pre-treated with allopurinol, a well-known XO inhibitor. Exposure of peripheral blood neutrophils to XO and HX (HX+XO) resulted in an increase in the amount of TNF-α (Figure 1A) and IL-8 (Figure 1B) positive cells (Figure 1).

Figure 1.

Flow-cytometry evaluation of TNF-α and IL-8-neutrophils intracellular expression after 5 hours of treatment with hypoxanthine

Data are presented as mean ± standard deviation of percentages of positive cells. Gating was detected from lipopolysaccharide analysis (not shown). Analysis of variance multiple comparisons test comparing the mean of each condition to the mean of control cells *p<0.05. TNF-α: tumour necrosis factor-α; IL-8: interleukin-8; CTRL: control; HX: hypoxanthine; XO: xanthine oxidase; ALL: allopurinol.

Next, we assessed the effect of RBC supernatants on TNF-α and IL-8 production by human neutrophils. As illustrated in Figure 2, co-treatment of neutrophils with 42-day old RBC supernatants and XO resulted in significant increases of TNF-α and IL-8 positive cells. Again, pre-treatment with allopurinol reduced the percentage of positive cells. It is worth noting that the percentages of both TNF-α and IL-8 positive cells were significantly higher when neutrophils were treated with stored RBC supernatant (SUP) than with HX, as shown in Figure 1: HX+XO 40.51%±20.04 vs SUP+XO 63.76%±13.79 (p=0.043). IL-8 showed a similar trend: 19.98%±5.00 for HX+XO and 44.62%±16.50 for SUP+XO (p=0.0128), as shown in Figure 2A non-parametric t-test was performed to evaluate the statistical significance of differences between the two groups.

Figure 2.

Flow-cytometry evaluation of TNF-α and IL-8-neutrophils intracellular expression after 5 hours of treatment with 42-days old supernatant

Data are presented as mean ± standard deviation of percentages of positive cells. Gating was detected from lipopolysaccharide analysis (not shown). Analysis of variance multiple comparisons test comparing the mean of each condition to the mean of control cells. *p<0.05, ***p<0.001. TNF-α: tumour necrosis factor-α; IL-8: interleukin-8; CTRL: control; SUP: supernatant; XO: xanthine oxidase; ALL: allopurinol.

We also tested early cytokine release (30 min) by incubating neutrophils with culture medium containing XO. The cell culture supernatants were then assayed for TNF-α and IL-8. In accordance with its chemoattractant role, IL-8 was released, already after 30 min, by stimulation with HX+XO. It is noteworthy that the release of IL-8 after 30 min of treatment with RBC supernatant was much greater than that after incubation in control conditions or with HX+XO. The presence of allopurinol resulted in a decrease of IL-8 release, confirming the pro-inflammatory effect of the HX-XO combination. Even 42-day old RBC supernatants alone induced IL-8 release, confirming the already described in vitro activating effect of RBC supernatant^23,24 (Figure 3). As expected²⁹, we did not observe early release of TNF-α (data not shown).

Figure 3.

Quantification of IL-8 release in the cell culture medium after stimulation for 30 min

Data are reported as mean (pg/mL) ± standard deviation. *p<0.05, **p<0.01 and ***p<0.001. IL-8: interleukin-8; CTRL: control; HX: hypoxanthine; XO: xanthine oxidase; SUP: supernatant; ALL: allopurinol.

Indeed, the post-transcriptional regulatory mechanism of TNF-α production in human neutrophils implies de novo synthesis, which occurs at a later stage³² (Figure 3).

DISCUSSION

RBC stored for long periods undergo significant biochemical changes such as accumulation of pro-inflammatory lipids and cytokines in the RBC supernatant⁷. Neutrophils are targets for these RBC-derived immunomodulatory molecules, and the exposure of these cells to a variety of extracellular stimuli results in the production of pro-inflammatory mediators^29,32–34. Human neutrophils have been shown to express and produce many pro- and anti-inflammatory mediators, both in vitro and in vivo, particularly upon appropriate stimulation³⁵.

The present study aimed to verify whether the accumulation of HX, derived from the adenine additive used to store RBC units, could be an additional mediator of neutrophil activation induced by stored RBC medium. To confirm our hypothesis, we preliminary evaluated whether 200 μM HX could induce neutrophil activation in the presence of a high concentration of XO. Since pre-treatment with allopurinol significantly reduced the percentage of TNF-α and IL-8 positive cells (Figure 1), our results indicate an involvement of the ROS produced by the reaction catalysed by XO in neutrophil-dependent TNF-α and IL-8 secretion during an acute inflammatory reaction.

Allopurinol did not seem capable of totally inhibiting the effects of the HX-XO system. We hypothesise that this incomplete inhibition was due to the fact that the inflammatory cascade was triggered, albeit weakly, by a small amount of free enzyme or by the binding kinetics in vitro.

The same approach was used to test the effect of the stored RBC supernatants on neutrophils in vitro. As shown by the results reported in Figures 2 and 3, in the presence of XO, supernatants boost neutrophil activation. Indeed, according to previous studies, reporting that RBC supernatants may contain a variety of molecules able to activate the neutrophil cascade, enhance the oxidative burst and induce the expression of adhesion molecules (e.g. CD11 and CD16)³⁵, the percentage of positive cells after treatment with stored RBC supernatants (SUP+XO) was higher than that following treatment with HX+XO, evidencing the role played by HX in association with lipid mediators³⁴. We hypothesise a synergistic effect of the lipid molecules plus the HX-XO system, with lipid-mediated priming that, in turn, enhances the effect of ROS produced by HX-XO.

Our data demonstrate that, aside from lipid mediators, in vitro neutrophil activation by stored RBC supernatants also depends on the presence of HX.

The potential clinical implication of our findings lies on the fact that storage lesions and especially RBC by-products, are dependent on, among other factors, the additive solutions used. Recently, rejuvenation methods, consisting in the incubation of RBC with a solution containing pyruvate, inosine, phosphate and adenine, are being increasingly proposed^36–38. If, on the one hand, rejuvenation can improve the quality of stored RBC by replacing antioxidant metabolites and restoring RBC mechanical properties, on the other hand, it dramatically increases HX levels in the RBC unit due to inosine metabolism^36–38.

Our results indicate the need to better characterise the biochemical and inflammatory effects of purine metabolism in recipients of RBC units. Indeed, the recipients could have a varying degree of neutrophil activation, which might be partly affected by their plasma XO levels but also by HX levels in the transfused units. This aspect must be taken into account in particular when stored RBC are transfused into patients with sickle cell disease or thalassaemia, who may receive blood transfusions almost every 2 weeks³⁶. Accumulation of HX in these patients might expose them to a high risk of developing transfusion-related immunomodulation.

One possible way of specifically reducing HX accumulation in RBC units, and consequently reduce neutrophil activation, is to use alkaline solutions. Compared to non-alkaline preservatives such as SAGM (the storage solution currently used in Europe), alkalising preservatives, such as SOLX, ESOL-5 and PAG3M, have lower HX levels, without altering the accumulation of oxidized products of fatty acids³⁹.

Finally, the storage condition (normoxia vs hypoxia) plays a crucial role in RBC metabolism, thereby affecting post-transfusion-related events. In rodent models of shock given transfusion of conventionally vs anaerobically stored units, Williams and colleagues demonstrated that anaerobically stored RBC improved systemic metabolic parameters associated with haemorrhagic shock and prevented the release of pro-inflammatory cytokines, suggesting that anaerobic storage of RBC could result in better outcomes for recipients⁴⁰.

In this complex scenario, our study represents a connection point between the collection, production and storage of RBC and the pathophysiological status of the recipient, paving the way for a personalised approach.

CONCLUSIONS

In agreement with previous observations²⁹, we here confirm that reaction of HX+XO is responsible for neutrophil activation, both in the early stage (after 30 min) and after 5 hours of treatment. The catalytic activity of XO is the key regulator of pro-inflammatory effects of HX+XO since addition of an XO inhibitor (allopurinol) resulted in reductions in the percentage of activated cells and IL-8 release.

In conclusion, postulating the presence of XO in transfusion recipients, we here show, for the first time, that neutrophil activation by stored RBC may depend not only on lipid mediators but also on the presence of HX contained in the RBC units. Future studies focused on tracing the changes in neutrophil metabolic pathways that may take place during in vitro stimulation with 42-day old RBC, with different additive solutions and storage methods, could enable the identification of biomarkers of activation and/or of negative side effects in vivo.

The clinical implications of our study comprise the adverse effects of RBC transfusion, such as TRALI. Indeed, our data indicate that the potential risk of TRALI after multiple RBC transfusions might be exacerbated by old stored RBC with high concentrations of HX. Indeed, accumulation of neutrophils in the airways is associated with high concentrations of neutrophil-derived mediators, in particular pro-inflammatory cytokines such as IL-8 and TNF-α²⁰.