Anoxia and glucose supplementation preserve neutrophil viability and function

Human polymorphonuclear neutrophils (neutrophils) are commonly described as “short-lived cells,” as compared with lymphocytes, mast cells, or other phagocytic cells such as macrophages. Neutrophil life span was measured upon neutrophil labeling ex vivo prior to transfusion in vivo and was estimated to be between 1.5 and 10 hours in mice and humans.¹  A recent study based on oral administration of deuterated water concluded that human circulating neutrophil life span could reach 5.4 days² ; however, this appealing result remains discussed by other groups.³ 

In vitro, it is well known that purified neutrophil life span is limited, barely exceeding 8 hours under atmospheric conditions containing 21% O₂ (pO₂ = 160 mmHg). Consequently, these cells remain difficult to culture and manipulate; indeed, neutrophils stand as the last white blood cell population that cannot be efficiently manipulated genetically despite previous attempts.^4,5  In addition, neutrophil transfusion efficiency has not been demonstrated despite its urgent need in many life-threatening multiresistant bacterial or fungal infections in neutropenic patients.^6,7  It is well known that neutrophil transfusion efficiency relies on dose and viability of the transfused cells.^8,9  In conclusion, increasing neutrophil viability in vitro represents a major challenge to allow neutrophil physiology and antimicrobial function studies but also to promote their clinical use.

Walmsley et al¹⁰  have previously shown that the survival of neutrophils was extended in low oxygen environments and that this beneficial effect, mediated at least in part by HIF-1α, was higher under anoxia (pO₂ = 0 mmHg) than hypoxia (pO₂ = 3 kPa or 22.5 mmHg). HIF-1α subunit abundance decreases under normoxic conditions because of hydroxylation by prolyl hydroxylases and subsequent targeting to proteasomal degradation.^11,12  Prolyl hydroxylase 3 was shown to be essential for HIF-1α degradation in neutrophils¹³  and was inhibited by the pan-hydroxylase inhibitor dimethyloxalylglycine (DMOG), which decreased neutrophil apoptosis.¹³  Importantly, the authors demonstrated that neutrophils were transcriptionally active (g3pdh, mif, nfκb) under these conditions.¹⁰ 

Here, we demonstrate that neutrophil life span is increased in vitro when conditioned for 20 hours at 37°C in anoxic culture medium supplemented with 3 mM glucose and 32 μg/mL DMOG. Conversely, neutrophils exposed 20 hours at 37°C under atmospheric conditions (+O₂) in RPMI medium only are named unconditioned neutrophils. Conditioned neutrophils remained viable, and we show that their morphology, degranulation, reactive oxygen species (ROS) production, chemotaxis, and bactericidal properties are similar to freshly purified neutrophils. Furthermore, we demonstrate that conditioning neutrophils in these conditions allows efficient DNA transfection in vitro and efficient transfusion in vivo.

All participants gave written informed consent in accordance with the Declaration of Helsinki principles. Peripheral human blood was collected from healthy patients at the Investigation Clinique et Accès aux Ressources Biologiques service of the Pasteur Institut (authorization DC No.2008-68) and from healthy patients stimulated 5 days with granulocyte colony-stimulating factor (G-CSF) at the Gustave Roussy Cancer Campus (Villejuif, France). Human blood was collected from the antecubital vein into tubes containing sodium citrate (3.8% final) as anticoagulant molecules. When specified, sodium heparinate (1% final concentration) was alternatively used.

Neutrophils were isolated from total blood samples immediately after blood collection with citrate or heparin as anticoagulant molecules.

Blood samples (50 mL) were centrifuged at 450g for 15 minutes. Platelet-rich plasma was collected and centrifuged at 2500g for 20 minutes to form platelet-poor plasma (PPP). Collected blood cells were resuspended in NaCl 0.9% and dextran sulfate (0.72%). After 30 minutes sedimentation, the upper layer, containing neutrophils, was centrifuged at 240g for 10 minutes. The pellet was resuspended in 1 mL of PPP, and cells were then separated on a Percoll (GE Healthcare) gradient (lower phase, 51% Percoll/49% PPP; upper phase, 42% Percoll/58% PPP) by centrifugation at 240g for 20 minutes. Neutrophils were collected and washed in RPMI medium 1640 (Sigma-Aldrich).

In order to increase the purity of the neutrophil-enriched fraction, remaining dead cells were removed using Dead Cell Removal MicroBead (negative selection; Miltenyi Biotec, Auburn, CA), and red blood cells were removed using CD235a (glycophorin) microbeads (negative selection; Miltenyi Biotec).

Neutrophil fractions with >95% purity were resuspended at 5 × 10⁶ cells/mL in RPMI 1640 with 2 mM l-glutamine, 25 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES; Gibco, Life Technologies), 10% heat-inactivated fetal bovine serum (Invitrogen, Gibco), the indicated concentrations of glucose (Sigma-Aldrich), and DMOG (Cayman Chemical Company).

For comparative experiments aimed at characterizing the role of oxygen on neutrophil viability, purified neutrophils were split into 2 identical fractions: one that was exposed to atmospheric conditions at 37°C with 5% CO₂ (21% O₂ condition, +O₂) and a second one that was conditioned at 37°C in an anoxic chamber (Minimacs, Don Withley) under an atmosphere containing 90% N_2, 5% CO₂, and 5% H₂ (0% O₂ condition, −O₂), during the indicated time. H₂ is used by a palladium catalyzer to remove oxygen traces in the chamber.

The optimal storage condition described in this study, 20 hours postpurification was obtained under anoxic conditions in RMPI supplemented with 3 mM glucose and 32 μg/mL DMOG. In this medium, the absolute cell count was assessed using a Malassez cell, and the neutrophil survival rate was determined by flow cytometry.

For neutrophil viability analysis, human neutrophils (10⁶ cells) were washed twice in phosphate-buffered saline and resuspended in 1 mL of Annexin V binding buffer; 100 µL sample (1 × 10⁵ cells) was stained with 5 µL of allophycocyanin–Annexin V and 5 µL of propidium iodide (PI; 0.5 µg/mL final concentration). At least 10⁴ events were acquired for each condition on a FACS-Calibur flow cytometer (BD Biosciences), and data were analyzed using CellQuest Pro Software (BD Biosciences). Viable cells were defined through Annexin V negative (Annexin V⁻)/PI negative (PI⁻) quantifications.

Neutrophils (2 × 10⁶) purified with citrate were infected with S. flexneri at a multiplicity of infection (MOI) 20 in RMPI + HEPES (10 mM). After 10 minutes centrifugation at 300g, infected neutrophils were incubated during 30 minutes at 37°C. Cell-surface markers exposure was quantified by flow cytometry as follows: 2.5 × 10⁵ purified cells or 1 × 10⁹ whole blood cells were collected and incubated with conjugated antibodies (see supplemental Methods, available on the Blood Web site). Cell labeling was analyzed using a FACS-Calibur flow cytometer (BD Biosciences), and data were analyzed using CellQuest Pro Software (BD Biosciences). Results are representative of 3 independent experiments.

Plasmid or short interfering RNA (siRNA) transfections were performed by nucleofection (Lonza) on neutrophils purified with citrate (2 × 10⁶ cells per transfection). Neutrophils were preincubated 2 hours in 500 µL of supplemented culture media (RPMI + 10% fetal calf serum + 3 mM glucose + 32 µg/mL DMOG) at 37°C under anoxic conditions. Cells were collected by centrifugation (300g, 10 minutes at room temperature) and washed in RPMI. After centrifugation, the pellets were suspended in 100 µL supplemented Mouse T Cell Nucleofector solution (Lonza); 2 µg of pmaxGFP plasmid (Lonza), 1 μM interleukin-8 (IL-8) siRNA (SMARTpool: Accell IL8 siRNA; Thermo Fischer Scientific), proliferating cell nuclear antigen (PCNA) siRNA (Applied Biosystems), Bax siRNA (Santa Cruz), or negative control siRNA (Qiagen) were added before transferring the mixture into nucleofection cuvettes. DNA transfer was performed by electroporation using Nucleofector Program Y-01 (Amaxa Nucleofactor, Lonza). Transfected cells were analyzed 20 hours postnucleofection by flow cytometry, fluorescence imaging (plasmid), quantitative polymerase chain reaction, enzyme-linked immunosorbent assay, or western blot (siRNA) (supplemental Methods).

Neutrophil transfusions were performed in neutropenic Dunkin-Hartley guinea pigs (150 g). Neutropenia was generated by 2 intraperitoneal injections of cyclophosphamide (100 mg/kg on day 7 and 50 mg/kg on day 1). When indicated, guinea pigs were infected intrarectally with S. flexneri (10⁹ colony-forming unit [CFUs]). When indicated, neutrophils purified from guinea pig or human were transfused (5 × 10⁶ or 10⁷, respectively), either freshly purified, conditioned for 20 hours, or not conditioned, to neutropenic guinea pigs by intracardiac injection 1 hour postinfection. After conditioning and prior transfusion, neutrophils are kept away from atmospheric exposure to avoid reoxygenation. Neutrophil transfusion efficiency was assessed 8 hours postinfection by quantifying colonic S. flexneri CFU and myeloperoxidase (MPO) activity (supplemental Methods). Experiments were performed on 3 animals per condition.

Transfused human neutrophil recovery was calculated by transfusing 10⁷ human neutrophils (freshly purified, conditioned, or transfected) in naïve neutropenic guinea pigs. Eight hours posttransfusion, human neutrophils were purified from 2 mL citrated blood and labeled with CD66b-PE and PI; viable (PI⁻) neutrophils were counted by flow cytometry using Trucount tubes (BD Biosciences). The percentage of CD15⁺/PI⁻ recovery rate was calculated using 6 mL total blood volume. Experiments were performed on 4 animals for each condition.

Data were analyzed using Prism 5.0 software (GraphPad) using Student t test. Significance was accepted when P < .05.

Neutrophil viability was evaluated by flow cytometry, by quantifying the proportion of viable (PI⁻) and nonapoptotic (Annexin V⁻) cells. Neutrophils purified from citrated blood were separated from remaining red blood cells with CD235a Microbeads (supplemental Figure 1). We showed that neutrophil viability at 37°C in RPMI medium was significantly increased under anoxia 20 hours or 48 hours postpurification, compared with atmospheric conditions (supplemental Figure 2A). In the absence of oxygen, cellular energy production relies on anaerobic glycolysis, using glucose as a substrate.^14,15  Accordingly, neutrophil viability was further increased 20 hours and 48 hours postpurification by supplementing the anoxic culture medium with 3 mM glucose, whereas no beneficial effect of glucose supplementation was observed under atmospheric conditions (supplemental Figure 2A).

Walmsley et al previously demonstrated that neutrophil apoptosis was reduced when stabilizing HIF-1α in the presence of DMOG.¹⁰  Accordingly, in a glucose-supplemented anoxic culture medium, the presence of DMOG (32 μg/mL) significantly increased neutrophil viability compared with atmospheric conditions without supplementation (supplemental Figure 2B). Under atmospheric conditions, a beneficial effect of DMOG was observed to a lower extent, consistent with the antiapoptotic role of HIF-1α previously described.¹⁰  When heparinized blood was used instead of citrated blood, neutrophil viability was also maintained 20 hours and 48 hours postpurification under anoxic conditions, in the presence of glucose and DMOG, however to a lower extent (supplemental Figure 3A-C). This might be because of higher levels of neutrophil activation generated by heparin compared with citrate.¹⁶ 

In summary, the optimal conditions for neutrophil viability were anoxic conditions, at 37°C in RPMI containing 10 mM HEPES, supplemented with 3 mM glucose and 32 mg/mL DMOG (Figure 1A). After 20 hours conditioning in this medium, the absolute neutrophil count was unchanged, and the survival rate reached 74 ± 2% (P < .05) 20 hours postpurification (vs 38 ± 14% under atmospheric conditions without supplementation (Figure 1A). Prior to analysis or functional characterization, viable conditioned neutrophils (hereafter named “conditioned neutrophils”) were purified by negative selection to obtain a pure Annexin V⁻/PI⁻ population (supplemental Figure 4).

The morphology of freshly purified or conditioned neutrophils was analyzed by electron microscopy. No difference was observed regarding the nucleus multilobular shape or the granular content (Figure 1B). G-CSF was previously shown to prolong neutrophil survival,^17,18  in particular by inhibiting the increase in intracellular calcium levels associated with apoptosis.^19,20  It was also described that G-CSF induced pseudo-Pelger-Huët anomalies, similar to the ones observed in hereditary neutrophils morphological modifications.²¹  Consistently, symmetric bilobed nuclei in neutrophils purified from G-CSF–stimulated volunteers were observed (Figure 1B), which is a typical pseudo-Pelger-Huët anomaly. No comparable anomaly was observed in freshly purified or conditioned neutrophils (Figure 1B). In addition, upon N-formyl-methionyl-leucyl-phenylalanine (fMLF) stimulation, the increase in intracellular calcium concentration in freshly purified neutrophils was comparable to the one in conditioned neutrophils, whereas it was reduced in neutrophils from G-CSF–stimulated volunteers (Figure 1C), as previously reported²⁰  (G-CSF treatment leading to increased concentrations of banded neutrophils and progenitors with altered reactivity for fMLF).

The abundance of physiological cell-surface markers such as the neutrophil-specific epitope CD66²²  (data not shown), CD63 (primary granules), CD66b (secondary granules), the C3b receptor (CD35),^2,8  the chemokine receptor CCR6 (CD196),²³  the adhesion molecule Icam-1 (CD54),²⁴  or the l-selectin CD62L were not significantly different on whole blood or conditioned neutrophils (Figure 2A). Whereas conditioned neutrophil cell-surface abundance of CD35, CD63, CD66b, and CD54 was significantly increased upon S. flexneri infection (Figure 2A). Moreover, the l-selectin CD62L level was increased in conditioned neutrophils as compared with whole blood neutrophils, and its level was reduced upon S. flexneri infection (Figure 2A).

We demonstrated that S. flexneri was efficiently phagocytized and killed, as previously reported²⁵  (Figure 2B). The PMA-induced ROS production (Figure 2C) and the IL-8-dependent chemotaxis (Figure 2D) properties of conditioned neutrophils were not significantly different from freshly purified neutrophils, further confirming that the conditioning process does not alter neutrophil bactericidal functions and motility.

The extended viability of conditioned neutrophils was hypothesized to allow an increase in DNA transfer efficiency. Indeed, so far, plasmid transfer in neutrophils by nucleofection was inefficient 2 hours posttransfection (∼5%).^4,5  The extended viability of conditioned neutrophils allowed longer transfection experiments, up to 20 hours. We demonstrated that conditioned neutrophils were efficiently transfected with pmaxGFP plasmid (39 ± 4% of total cells) while analyzed 20 hours posttransfection (Figure 3A-B); without altering neutrophil viability, only 10% of transfected cells were PI⁺ upon nucleofection (supplemental Figure 5), consistent with previous reports.^4,5  The absolute cell count was unchanged upon transfection (Figure 3C). Conditioned neutrophils were efficiently transfected with IL-8-siRNA by nucleofection in lipopolysaccharide (LPS)-stimulated neutrophils^26,27  (Figure 3D). IL-8 expression was significantly reduced in neutrophils transfected with IL-8 siRNA, 20 hours postnucleofection, but not in negative control siRNA-transfected cells (Figure 3D). These results were further confirmed by quantifying the corresponding LPS-dependent IL-8 release; LPS-dependent p65 abundance was used as a positive control²⁸  (Figure 3E).

Key antiapoptotic proteins involved in neutrophil viability regulation, such as Mcl-1²⁹  and PCNA,³⁰  were stabilized in conditioned neutrophils, together with pro-caspase-3 (Figure 4A), when the reoxygenation of conditioned neutrophils reduced their survival and led to a destabilization of Mcl-1, PCNA, and pro-caspase-3 (Figure 4B). PCNA siRNA nucleofection of conditioned neutrophils was efficient (reduction of PCNA abundance) and increased neutrophil apoptosis 20 hours posttransfection (Figure 4C), confirming its antiapoptotic function.³⁰  Consistently, Bax siRNA nucleofection in conditioned neutrophils reduced Bax abundance (Figure 4D) and decreased neutrophil apoptosis 48 hours posttransfection, further extending conditioned neutrophil viability (Figure 4E).

Neutrophil transfusion efficiency depends on the number and the viability of transfused cells.^8,9  We aimed to further confirm that conditioned neutrophils were functional in vivo, in a preclinical study. Guinea pig (Figures 5A-C) or human (Figures 5D,G) conditioned neutrophils were transfused to neutropenic guinea pigs (naïve or infected with S. flexneri). Considering that neutrophil reoxygenation alters the conditioning benefits on their survival maintenance (Figure 4B), neutrophils were manipulated without atmospheric exposure prior to transfusion. Compared with unconditioned neutrophils, bacterial clearance (Figure 5B) and MPO activity (Figure 5C) in infected colonic tissue were increased upon transfusion of guinea pig conditioned neutrophils. Similar results were obtained upon the transfusion of human conditioned neutrophils, with lower activities (Figures 5D-E).

The throughout recovery of human neutrophils in the guinea pig blood flow was evaluated 8 hours posttransfusion. Upon transfusion of freshly purified or conditioned human neutrophils in naïve guinea pigs, similar viable neutrophil recovery rates were observed (23 ± 11% vs 19 ± 7%) 8 hours posttransfusion (Figure 5F). These throughout recovery rates were higher than upon unconditioned neutrophil transfusion (5 ± 3%) (Figure 5F). The recovery of transfused neutrophils in this model was decreased upon PCNA siRNA nucleofection and increased upon Bax siRNA nucleofection (Figure 5G), confirming in vivo the importance of these 2 proteins in the conditioned neutrophil viability modulation characterized in vitro (Figure 4C-E).

Neutrophils have long been termed “short-lived cells.”³¹  Our results suggest that this assumption is largely dependent on the conditions of their manipulation in vitro rather than on their actual life span in vivo, which remains under discussion.^2,3  In this report, we showed that limiting the exposure of neutrophils to atmospheric levels of oxygen increased their viability (Figure 1A-B), consistent with previous reports Walmsley et al.^10,13  In anoxic conditions, neutrophil morphological (Figure 1B), physiological, and bactericidal functions (Figures 1C and 2A-D) remained constant. The extended viability of neutrophils under anoxia in the presence of glucose and DMOG allowed efficient transfection with plasmid or siRNA (Figures 3 and 4). In particular, the genetic manipulation of key pro- and antiapoptotic factors (Bax, PCNA) was shown to modulate conditioned neutrophil viability (Figure 4). The reoxygenation of conditioned neutrophils decreased their life span (Figure 4B), justifying the necessity to keep transfused neutrophils away from atmospheric air exposure (Figure 5). This observation is particularly important regarding neutrophil manipulation for transfusion and will have to be considered in the future to improve their therapeutical use. Considering that oxygen has a negative impact on neutrophil viability, further experiments will be required to evaluate the potential beneficial effect of neutrophil collection, purification, and maintenance under anoxic conditions on their viability in vitro and clinical use.

Our results confirm the adaptation of neutrophils in low oxygen conditions and are probably related to the fact that neutrophil energy production mainly relies on glycolysis and not on mitochondrial activity. This assumption is supported by the low abundance of mitochondria in neutrophil cytosol; their activity was found to be restricted to apoptosis induction, not energy production,^15,32  but also chemotaxis modulation.³³  This physiological characteristic represents a fundamental difference between neutrophils and platelets or lymphocytes.³⁴ 

In this study, the functional characterization of conditioned neutrophils was achieved under atmospheric conditions, in order to compare the results with standard assays. Further experiments will be required to better understand the role of oxygen on the modulation of neutrophil function, particularly in relation to mitochondria abundance and activity, as previously reported.³³ 

Up to now, studies addressing the adaptation of neutrophils to hypoxia or anoxia were restricted to inflammatory or infectious diseases.^13,35,36  However, under physiological conditions, neutrophils naturally evolve in low oxygen environments. Neutrophil maturation from hematopoietic stem cells occurs in the bone marrow, which is a hypoxic environment.^37,38  In this context, maturating cells produce energy through glycolysis rather than respiration in an HIF-1–dependent manner.^39,40  Once released into the blood circulation through sinusoids, neutrophils will circulate in the blood plasma fraction. The blood circulation is characterized as an oxygenated compartment. However, the blood plasma contains low levels of dissolved oxygen, estimated to be ∼2% of the total oxygen transported by red blood cells.⁴¹  Considering that the arterial pO₂ equals 75 to 100 mmHg and the venous pO₂ equals 30 to 50 mmHg,⁴²  the plasmatic pO₂ equals 0.3 to 1 mmHg (0.04% to 0.13% O₂), which is quasi-anoxic.

In conclusion, neutrophils are very sensitive to oxygen, the abundance of which is limited in vivo under physiological conditions. This fundamental aspect will have to be further integrated into in vitro models to improve the characterization of neutrophil physiology.

The authors thank Valérie Lapierre and Dominique Tramalloni (Laboratoire de Thérapie Cellulaire, Gustave Roussy Cancer Campus, Villejuif, France) for supplying blood samples from G-CSF–stimulated patients.

This work was supported by grants from the European Union’s European Research Council (HOMEOPITH, 272398) and the European Union’s Seventh Framework Programme (TORNADO, 222720) (B.S.M. and P.S.J.), and by a principal investigator award from Science Foundation Ireland (C.T.T.).

Contribution: V.M., M.-C.P., C.C.-L., and B.S.M. performed the research and together with M.-N.U. provided healthy donors’ blood samples; C.C., C.T.T., V.W.-S., and P.J.S. interpreted data; and B.S.M. designed the research and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Benoit S. Marteyn, Institut Pasteur, Unité PMM, 28 rue du Dr Roux, 75724 Paris Cedex 15, France; e-mail: marteyn@pasteur.fr.