G-CSF-treated donors: effect of stimulant, apheresis and storage.
Andréa Murru1,2,3, Marie-Ève Allard3, Guillaume Paré1,2, Myriam Vaillancourt1,2, Lucie Boyer3, Marie-Pierre Cayer3, Julien Vitry1,2, Patricia Landry3, Marie-Michèle Labrecque1,2, Nancy Robitaille4, Donald Branch5, Mélissa Girard3, Maria J. Fernandes1,2
1Infectious and Immune Diseases Division, Laval University, CHU de Québec Université Laval research centre, Québec, Canada, G1V 4G2
2Department of Microbiology-Infectious Diseases and Immunology, Faculty of Medicine, Laval University, CHU de Québec Université Laval research center, Québec, Canada, G1V 4G2
3Medical Affairs and Innovation, Héma-Québec, Quebec, Canada, G1V 5C3
4Transfusion medicine, Héma-Québec, Canada, G1V 5C3
5Department of Medicine, University of Toronto, Centre for Innovation, Canadian Blood Services, Toronto, Canada
Correspondence: Dr. Maria Fernandes, Laval University, CHU de Québec Université Laval research centre, Room T1-49, 2705 Boulevard Laurier, Québec, QC, Canada, G1V 4G2, Tel. 418-656-4141, ext. 46106 E-mail: maria.fernandes@crchudequebec.ulaval.ca
Keywords: granulocyte transfusion, neutrophils, leukapheresis, prednisone, G-CSF, cell surface marker, calcium mobilization, chemotaxis, phagocytosis, cytokines, reactive oxygen species, viability
2.1 Résumé
Contexte : La transfusion de concentré de granulocytes (GC) constitue une thérapie alternative pour enrayer les infections incurables des patients neutropéniques lorsque les traitements conventionnels sont inefficaces. Les GC obtenus des donneurs stimulés avec un corticostéroïde, telle la prednisone utilisée au Canada, et/ou du G-CSF par leucaphérèse sont certifiés selon le nombre absolu de neutrophiles (ANC) récolté bien que ce critère ne reflète pas la capacité fonctionnelle des cellules.
Objectifs : Comparer la viabilité et les fonctions des neutrophiles des GC obtenus par prednisone (GC-pred) de ceux obtenus par G-CSF (GC-G-CSF), principalement utilisé aux États-Unis. Nous avons évalué comment l’agent de stimulation, la leucaphérèse et l’entreposage affecte les neutrophiles des GC.
Méthodes : Les GC ont été préparé par leucaphérèse de dix donneurs stimulés à la prednisone puis six mois après au G-CSF. La caractérisation des neutrophiles inclut l’analyse de la viabilité, l’expression des marqueurs de surface, la mobilisation du calcium intracellulaire, la chimiotaxie, la phagocytose, la production d’espèces réactives de l’oxygène (ROS) et d’interleukine-8. L’analyse métabolique inclut le suivi du pH, du glucose et du lactate pendant l’entreposage. Les tests ont été réalisés sur les neutrophiles de la circulation et des GC jusqu’à 48 hr d’entreposage.
Résultats : L’ANC des GC-G-CSF était deux fois supérieure à celui des GC-pred. Les deux types de GC contenaient des lymphocytes et des monocytes et les GC-G-CSF contenaient 40 % de neutrophiles immatures CD10low alors que la prednisone mobilisa uniquement des neutrophiles matures. La phagocytose des neutrophiles des GC-pred était augmentée alors que les neutrophiles des GC-G-CSF présentaient une altération de l’activité chimiotactique et une augmentation de la sécrétion d’interleukine-8. L’entreposage affecta négativement la viabilité et les fonctions des neutrophiles des deux types de GC ainsi que l’activité métabolique provoquant la déplétion du glucose et la baisse du pH après 24 hr. La leukaphérèse eu un effet minime sur la mobilisation du calcium et sur la production intracellulaire de ROS des neutrophiles des GC-pred.
Conclusion : Nos résultats révèlent des différences significatives entre les GC-pred et les GC-G-CSF.
La viabilité, le phénotype et les fonctions des neutrophiles dépendent de l’agent de stimulation utilisé alors que l’entreposage est le paramètre le plus critique indépendamment du type de GC. Le rôle des neutrophiles immatures et des autres leucocytes ainsi que de nouvelles conditions d’entreposage préservant les neutrophiles doivent être évalués dans de prochaines études.
2.2 Abstract
Background and objectives: Transfusion of granulocyte concentrates (GTX) is an alternative therapy to eradicate life-threatening infections in neutropenic patients when conventional antimicrobial treatments are ineffective. Granulocyte concentrates (GCs) obtained by stimulation of healthy donors with a glucocorticoid, such as prednisone used in Canada, and/or G-CSF are certified based on the absolute neutrophils count (ANC) harvested which does not reflect the cell functional capacity.
Objective: To characterize prednisone-derived GC neutrophils viability and function and compared them to G-CSF-derived GC neutrophils, mostly used in the US. We determined how pre-treatment, leukapheresis and storage affect GC neutrophils.
Materials and methods: GCs were prepared by leukapheresis from 10 healthy donors pre-treated with prednisone and then with G-CSF after a 6-month washout period. Neutrophil characterization included the analysis of viability, surface marker expression, intracellular calcium mobilization, chemotaxis, phagocytosis, production of reactive oxygen species (ROS) and interleukin-8 (IL-8).
Metabolic analysis included the monitoring of pH, lactate and glucose during storage. Assays were performed on circulating neutrophils and GC neutrophils up to 48 hrs of storage.
Results: The ANC in G-CSF-derived GCs was 2-fold higher than in prednisone-derived GCs. Both GCs type contained lymphocytes, monocytes and G-CSF-derived GCs contained approximately 40 % of CD10low immature neutrophils while prednisone mobilized only mature neutrophils. Neither mobilizing agent affected neutrophil viability. Prednisone-derived GC neutrophils had increased phagocytic capacity whereas G-CSF-derived GC neutrophils had altered chemotactic activity and increased IL-8 secretion. Storage duration had a global negative effect on neutrophils viability and function for both GCs type and on metabolic activity causing extreme glucose depletion and pH drop as early as 24 hrs. Leukapheresis procedure had minimal effect on calcium mobilization and intracellular ROS production of prednisone-derived GC neutrophils only.
Conclusion: Our findings revealed significant differences between prednisone and G-CSF GCs.
Neutrophils viability, phenotype and functions differed depending on the type of stimulating agent whereas storage was the most critical parameter regardless of the type of GC. The role of immature neutrophils and other leukocytes as well as new storage conditions that preserve neutrophils need to be address in further studies.
2.3 Introduction
Life-threatening infections are a major health concern due to the growing resistance to antimicrobial and antifungal therapies (1). Rapid and effective protection against microbial infections is provided by the most abundant circulating granulocyte in circulation, the neutrophil (2, 3). Patients with a critically low number of neutrophils, below 1 x109 L-1 (4), are thus highly susceptible to these types of infections (5). The prevalence of neutropenia is on the rise due to the growing use of aggressive chemotherapy and hematopoietic stem cell transplants increasing the number of patients with untreatable bacterial or fungal infections (6). The only potential lifesaving therapy for these severe neutropenic patients is the transfusion of granulocytes.
Granulocyte transfusions (GTXs) temporarily increase neutrophil count until the bone marrow restores granulopoiesis (7). To obtain the minimal number of granulocytes for GTX (1010/transfusion), healthy donors are stimulated with either G-CSF and / or a corticosteroid to generate granulocyte concentrates (GCs). GCs are irradiated prior to transfusion to avoid transfusion-associated graft versus host disease (8). While in Canada all GCs are prepared from prednisone-stimulated donors by leukapheresis, GC donors are stimulated with G-CSF in the US (9). In Europe, buffy-coat derived GCs are more routinely used (10).
The key neutrophil functions required for the effective eradication of an infection include chemotaxis, phagocytosis of the pathogen and its destruction by antimicrobial peptides and reactive oxygen species (ROS) produced upon neutrophil activation. Neutrophils also bind and sequester pathogens for destruction by releasing extracellular traps (NETs) mostly composed of DNA and intracellular proteins (11, 12). Cytokines also play a role by further activating neutrophils and neighbouring cells and promoting the recruitment of additional circulating leukocytes including neutrophils and monocytes (13-15). Without these defences, survival from an infection is limited to a few days in persons with absolute neutropenia (16-18).
A key challenge in optimizing GTX is the lack of a functional test to certify this blood product for clinical use. The ANC is the only granulocyte associated parameter used to certify GCs for transfusion. (19).
This is problematic as neutrophil functional responses are known to vary between individuals. Some studies reported similar chemotactic activity (20, 21) and ROS production (21) in G-CSF-derived GC neutrophils compared to neutrophils of non-stimulated healthy donors. Other studies observed, however, that phagocytosis was impaired in these GC neutrophils (22) and the level of pro-inflammatory cytokines in GC supernatant was elevated (23, 24). We mention only a few studies as a comprehensive comparison of all reports is challenging due to the different experimental approaches used. A better characterization of this blood product is thus needed to optimize its preparation, certification and administration. Since prednisone-derived GCs are less well characterized than G-CSF GCs, the aim of this study was to compare the viability and antimicrobial functions of neutrophils in these two types of GCs. To minimize the number of confounding variables, prednisone and G-CSF GCs were prepared from the same cohort of healthy donors and all GCs were prepared by one blood centre, Héma-Québec (Montréal, Canada). We also assessed the effects of leukapheresis and storage on GC neutrophils.
2.4 Materials and Methods
2.4.1 Donor recruitment and GTX collection
Ten healthy donors were recruited by Héma-Québec for two GC donations. A wash out period of at least 6 months was established between the two donations in order to ensure that the first mobilizing agent used did not influence data obtained from the second donation. Informed consent from the donors was obtained in accordance with our approved institutional review board protocol. All donors were stimulated with 50 mg prednisone per os or 300 mcg G-CSF intravenously 12-18 hrs or 24 hrs prior to leukapheresis, respectively. Leukapheresis was performed with SpectraOptia (Terumo) according to Héma-Québec guidelines. A GC volume of 350 ml was harvested for all donations in blood collection bags containing 30 ml of sodium citrate (46.7 %) and 500 ml of hydroxyethyl starch (HESPAN® 6 % B. Braun Medical Inc.). This anticoagulant solution is used to enhance red blood cell sedimentation to a ratio of 13:1 (product:anticoagulant solution). GC composition was determined directly after leukapheresis with a cell counter (Beckman Coulter ACP 5diff). GCs were transported from the Globule Laval collection centre to Québec City Héma-Québec establishment for irradiation at 25 Gy prior to delivery to our laboratory for analysis on the day of collection (D0) as well as 24 hrs (D1) and 48 hrs (D2) post-collection. GCs were stored at room temperature without agitation. A sample of whole blood from each donor was also drawn prior to leukapheresis in acid-citrate-dextrose (ACD) (BD Vacutainer) tubes for analysis of circulating neutrophils. Whole blood composition was done at Héma-Québec collection centre with a cell counter (Beckman Coulter ACP 5diff). From the 10 GC donors recruited, one donation was excluded from the functional analysis because of a change in the sedimentation agent during the first procedure with prednisone pre-treatment. For comparison, unstimulated healthy donors, matched for sex and age, were recruited by the Clinical Research Platform at the CHU de Québec-Laval University and were used as control.
2.4.2 Gases and metabolites analysis
Gases and metabolites were measured in an undiluted sample of GC with an ABL90 flex blood gas analyser (Radiometer).
2.4.3 Neutrophils and peripheral blood mononuclear cells isolation
All leukocyte isolation procedures were conducted at room temperature under sterile conditions.
Neutrophils were isolated from non-stimulated, healthy donors and from stimulated donor blood prior to leukapheresis as previously described in (25). Briefly, red blood cells were sedimented in 2 % dextran and neutrophils isolated by centrifugation at 600 x g for 20 min on lymphocyte separation medium (Wisent) cushions. After hypotonic lysis of erythrocytes in the pellet, neutrophils were resuspended in Mg2+-free HBSS (Hank balanced salt solution, Wisent) containing 1.6 mM CaCl2 at 10 x106 ml-1. Neutrophils were isolated from GCs by density gradient as described above without prior dextran sedimentation as GCs already contain hydroxyethyl starch that has sedimenting properties, and leukapheresis reduces the number of erythrocytes. To identify neutrophils in the mononuclear cell layer of the gradient (PBMC layer), PBMCs were washed and resuspended in PBS (Phosphate buffered saline, Wisent) at 10 x106 ml-1.
2.4.4 Flow cytometry analysis
Cells were stained with the following antibodies at room temperature in the dark for 20 min: anti-CD45 (2D1), anti-CD3 (UCHT1), anti-CD19 (SJ25C1), anti-CD56 (NCAM16.2), anti-CD14 (M5E2), anti-CD66b (G10F5), anti-CD10 (Hl10a), anti-CD16 (3G8), anti-CCR3 (5E8), all from Becton Dickinson and anti-CLEC12A (50C1) from BioSciences. Cells were also stained with a viability marker, FVS605, and washed with HBSS 0.5% BSA (Sigma Aldrich) prior to flow cytometry. The specificity of this antibody panel was confirmed with a full minus one (FMO) control for each fluorochrome and
Compensation Bead Kit (ThermoFisher) for FVS605. Ultra Rainbow calibration particles (Spherotech) were used for routine calibration control. Flow cytometry data were acquired using BD LSRII flow cytometer and analysed with Diva software.
2.4.5 Viability assay
Neutrophils were stained with Annexin-V (BD) / 7-AAD (Invitrogen) in Ca2+ -free HBSS at room temperature in the dark for 15 min. Cells were kept on ice until analysis by flow cytometry with a BD Canto II flow cytometer.
2.4.6 Intracellular Calcium Mobilization
Neutrophils were pre-incubated with 1 μM Fura-2-AM (Invitrogen) for 30 min at 37 °C, washed in HBSS and resuspended at 5 x106 ml-1 prior to stimulation with 10-7 fMLF. Fluorescence was monitored with a fluorescence spectrophotometer (Fluorolog-SPEX from Jobin Yvon Inc., Edison, NJ) at two excitation wavelengths, 340 and 380 nm, and an emission wavelength of 510 nm. Levels of cytoplasmic calcium were calculated as the ratio of fluorescence at 340 and 380 nm.
2.4.7 Chemotaxis
Chemotaxis towards fMLF was assessed as described in (25). The results are expressed as the percentage of migrated cells.
2.4.8 Intracellular ROS production
Neutrophils were incubated with 1µM CM-H2DCFDA (C6827, ThermoFisher) for 15 min at 37oC prior to being stimulated for 15 min at 37oC with 10-7 fMLF or 10-7 PMA (positive control) or diluent (DMSO, the negative control). Oxidation of CM-H2DCFDA was quantitated measuring the mean fluorescence intensity (MFI) of the probe with a BD Canto II flow cytometer.
2.4.9 Superoxide Anion Measurement
Superoxide production was measured using the reduction of cytochrome c assay as previously described in (26). Neutrophils were stimulated with 10-7 fMLF, 10-7 PMA (positive control) or diluent (DMSO, the negative control) for 10 min at 37 °C with agitation in the presence of 125 μM cytochrome c (Sigma Aldrich). The OD was measured at 550 and 540 nm in cell-free supernatants after centrifugation. To calculate the quantity of superoxide anion in nmol O2-/5 ×106 ml-1, the following formula was used: (OD 550nm – OD 540nm) x the millimolar extinction coefficient of 47.4.
2.4.10 Phagocytosis
Non-opsonized phagocytosis was quantified after incubating neutrophils with 15µg of pHRodoTM Red zymosan A Bioparticles® Conjugated (Molecular probes, Life technologies). Briefly, cells were centrifuged at 2000 rpm for 15 sec to synchronize phagocytosis prior to a 30 min incubation at 37oC or 4oC for comparative purposes. Phagocytosis was stopped by adding cold HBSS Ca2+. Cells were kept on ice until analysis by flow cytometry with a BD LSRII flow cytometer to determine the proportion of phagocytic neutrophils and the acidification of the phagosomal compartment with the MFI.
2.4.11 Cytospin
Neutrophils and PBMCs were deposited on slides and centrifuged at 3200 rpm for 5 min before staining with Hemacolor® Rapid staining of blood smear (Sigma Aldrich) as per the manufacturers’
guidelines. Slides were analyzed by microscopy (OLYMPUS BX51) and stored at 4oC.
2.4.12 Cytokine production
Neutrophils were stimulated with 10 ng.ml-1 LPS or incubated with the same volume of filtrated MiliQ water as a negative control for 24 hrs at 37oC in RPMI 1640 (Wisent) containing 5 % of decomplemented FBS (Wisent). Cell-free supernatants were prepared by two successive rounds of centrifugation, the first at 3000 rpm for 3 min and the second at 13 000 rpm for 10 min. Supernatants were aliquoted and stored at -80oC until analyzed by ELISA for IL-8 production (BD Biosciences). Each sample was measured in duplicate.
2.4.13 Statistical analysis
Statistical non-parametric analysis is realized with Prism Graph Pad and significance is defined with an alpha of 5%. The pre-treatment and the leukapheresis effect is determined with Wilcoxon rank-signed test for paired samples, with Mann-Whitney test when pairing was not possible, or with 1-way ANOVA with Kruskal-Wallis test multiple comparison of prednisone and G-CSF compared with control group. For ROS and IL-8 production analysis, we used 2-way ANOVA with Dunnett's multiple comparisons test with non-stimulated group as reference. The data at D2 and D3 compared to D1 were analyzed with one-sample t-test using a theoretical mean.
2.5 Results
2.5.1 Cohort characteristics and GC collection
GCs were prepared from 10 healthy donors that had not previously donated blood for leukapheresis. To diminish inter-donor variability, all GC donors, 50 % male and 50 % female donated GCs with the two stimulating agents. Exclusion and inclusion criteria are summarized in Table 1.
Briefly, prednisone was administered to all donors for the first GC donation followed by a wash out period of 6 months prior to the intravenous administration of G-CSF for the second GC donation. The average blood volume treated during leukapheresis was 4964 ml (4211-6001 ml) during prednisone procedure and 4831 ml (3611-6039 ml) during G-CSF procedure (Table 2). Ten healthy donors frequency-matched for age and sex were also recruited for comparative purposes.
2.5.2 Comparison of neutrophil counts in prednisone and G-CSF GCs
Since G-CSF mobilizes neutrophils more efficiently than glucocorticoids, we analyzed the leukocyte content in both types of GCs. Prednisone stimulation consistently generated GCs of a similar or significantly lower leukocyte concentration compared to G-CSF (62.9 x109/Land 109.5 x109/L white blood cells (WBC), respectively; Table 2 and Fig. 1A. Consequently, the minimal ANC required per transfusion of 1010 neutrophils was observed in 6/9 prednisone GCs. In contrast all G-CSF GCs contained the minimal ANC or higher numbers of neutrophils (Fig. 1B). Moreover, neutrophils comprise 56.8 % of all leukocytes in prednisone GCs and 75.5 % in G-CSF GCs. Neutrophils were 1.5 to 4-fold more abundant in G-CSF-derived than prednisone-derived GCs.
While neutrophils are the most abundant cells in GCs, other leukocytes are also present in this cell therapy product including monocytes, basophils, eosinophils and lymphocytes. Notably, in
prednisone-derived GCs, lymphocytes represent 30 % of all leukocytes and 16 % in G-CSF GCs (Table 2). A substantial number of monocytes are also present in these GCs, 11% and 7% of all leukocytes in prednisone and CSF GCs respectively. Other leukocyte populations present in prednisone and G-CSF GCs include basophils and eosinophils. There were no significant differences in the number of platelets but G-CSF GCs had significantly lower haemoglobin levels and haematocrit compared to prednisone GCs (Table 2). Together, these findings indicate that G-CSF is more efficient at mobilizing sufficient neutrophils for granulocyte collection than prednisone, and that GCs contain a significant proportion of lymphocytes and monocytes in addition to neutrophils.
2.5.3 Viability and surface marker expression of prednisone and G-CSF mobilized neutrophils
To characterize the effect of prednisone and G-CSF stimulation on neutrophils in vivo, we compared their viability, maturity and cell-surface marker expression (Supplementary table 1) prior to leukapheresis. Viability as assessed by spontaneous apoptosis and necrosis, was not significantly altered in circulating neutrophils of donors administered one dose of prednisone or G-CSF compared to neutrophils of non-stimulated healthy donors (Fig. 2A).
Since G-CSF is known to induce the release of immature neutrophils from the bone marrow (27), we stained the mobilized circulating neutrophils with the maturity marker CD10. We observed that G-CSF induced the release of both mature (60 % of CD10high/CD16high) and immature (40 % of CD10low/CD16low) neutrophils into the circulation (Fig.2C and D). Neutrophil maturity was confirmed by microscopy and revealed the presence of neutrophils with a mature or immature nuclear morphology (Fig.2B). Contrary to immature neutrophils of low-density whose numbers increase in some disease states, these immature neutrophils are of a similar density as mature neutrophils as they pelleted to the bottom of a density gradient. In contrast, prednisone only mobilized CD10high neutrophils with multilobular nuclei, a typical feature of mature neutrophils (Fig. 2B and C) Moreover, they express significantly lower levels of CD16 (CD16low) compared to their mature counterparts that are CD16high.
While GC immature neutrophils from some donors express higher levels of CD66b and CD15 compared to CD10high G-CSF-derived neutrophils, this trend did not reach significance (Fig. 2D).
Similarly, the expression of the other activation or inhibition cell-surface markers tested in prednisone and G-CSF-derived GC neutrophils was not altered in mobilized neutrophils. Together, these data indicate that prednisone and G-CSF GC neutrophils differ in neutrophil count and maturity.
2.5.4 Antimicrobial functions of prednisone and G-CSF-derived, GC neutrophils
We assayed the key neutrophil functions required to fight infections of GC neutrophils including
We assayed the key neutrophil functions required to fight infections of GC neutrophils including