Epstein-Barr Virus Infects and Induces Apoptosis in Human Neutrophils

Neutrophils were isolated from venous blood obtained from normal healthy volunteers using Ficoll-Hypaque density centrifugation (Pharmacia-Biotech Inc, Baie d'Urfé, Canada) as previously described.18 All neutrophil preparations contained fewer than 1% monocytes as determined by monoesterase staining and less than 0.3% of B and T lymphocytes, as evaluated by cytometry using anti-CD2, anti-CD3, and anti-CD19 monoclonal antibodies (MoAbs) (Becton Dickinson, San Jose, CA). Viability, estimated by the trypan blue-dye exclusion procedure, was greater than 99% in all preparations.

Neutrophils were incubated in Hank's balanced saline solution (HBSS) (GIBCO-BRL, Burlington, Ontario, Canada) supplemented with 1% of heat-inactivated autologous plasma and 10 mmol/L of HEPES buffer for all experiments, except for studies of apoptotic cells which were performed in RPMI-1640 supplemented with 10% of heat-inactivated fetal bovine serum (FBS). Culture medium tested for the presence of endotoxin by the limulus amebocyte assays (Sigma, St Louis, MO) contained less than 10 pg/mL of contaminating endotoxins. Neutrophils were incubated at specified cells densities with either infectious EBV or 3 nmol/L recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF; provided by Genetics Institute, Boston, MA) for the indicated times.

Viral preparations of EBV strain B95-8 were produced as previously described.18 Briefly, B95-8 cells (which were mycoplasma-free tested) were grown in RPMI-1640 medium supplemented with 10% heat-inactivated FBS. When the viability of cell cultures reached 20% or less, as determined by trypan blue-dye exclusion, cell-free culture supernatants were obtained and filtered through a 0.45-μm pore size filter, and viral particles were purified by differential ultracentrifugation. Virus stocks were resuspended in RPMI-1640, aliquoted, and stored at −80°C until use. Viral titers were measured and evaluated at 10⁷ transforming units (TFU/mL).

Neutrophils were resuspended in cold HBSS and incubated with EBV at 4°C for 5 to 15 minutes to allow binding to the cell surface. Virus-bound cells were then cultured at 37°C for varying time periods and processed for electron microscopy examination as described.19 20 Briefly, cells were fixed for 1 hour in 1.5% glutaraldehyde in 0.1 mol/L cacodylate buffer pH 7.3. After fixation, cells were pelleted and pre-embedded in 20% bovine serum albumin (BSA), polymerized with 25% glutaraldehyde, and cut in one millimeter³. The blocks were rinsed with buffer, postfixed in 1% buffered osmium tetroxyde, treated with 0.1% buffered tannic acid, and stained in bloc with 2% ethanolic uranyl acetate. The neutrophils were dehydrated through a graded series of ethanol and propylene oxide, and embedded in Epon 812 (JBem Services Inc, Quebec, Canada). Ultrathin sections were cut with a diamond knife on an Ultracut S ultramicrotome (Leica Canada Inc, Montreal, Canada) and mounted on 200 mesh copper grids. The sections were stained with 2% uranyl acetate and 0.5% lead citrate, and examined with a Jeol 1010 electron microscope (Jeol Canada Inc, St Hubert, Canada).

Neutrophils were incubated for various periods of time in the presence or absence of EBV and then washed three times with HBSS to remove residual virus. For Southern blotting analysis, DNA was extracted as described.21 RNA was removed by RNASE ONE (Promega, Madison, WI) treatment and DNA was digested with BamHI. Ten micrograms of DNA cells were loaded onto a 0.6% agarose gel and size-fractionated by electrophoresis. Transfer onto HYBOND-N membrane (Amersham Canada Limited, Oakville, Ontario, Canada) was performed by capillary diffusion in 10× sodium saline citrate (SSC) overnight. After prehybridization the membranes were hybridized with random-primed ³²P-labeled probes in 50% formamide overnight at 42°C. The membranes were then washed and exposed to Kodak X-OMAT films (Eastman Kodak, Rochester, NY) with an intensifying screen at −70°C. The BamHI W probe was a 400-bp polymerase chain reaction (PCR) amplicon located in the BamHI W region of the EBV genome (primer 1: 5′ GCAGTAACAGGTAATCTCTG 3′, positions 20124-20143 and primer 2: 5′ ACCAGAAATAGCTGCAGGAC 3′, positions 20523-20504.22 The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) cDNA probe was used as control to demonstrate equal loading of DNA in each lane. For each sample the amount of viral DNA was quantitated relatively to their respective levels of GAPDH. The given ratio was compared with that obtained in unstimulated cells, using this formula: (optical density [OD] of DNA of the sample/OD of GAPDH of the sample)/(OD of DNA of unstimulated cells/OD of GAPDH of unstimulated cells).

Neutrophils (10 × 10⁶ cells/mL) were pretreated with cytochalasin B (an inhibitor of phagocytosis) (Sigma, Oakville, Ontario, Canada) at 10 μmol/L and infected with EBV. After 10 hours of culture, cells were washed and resuspended in ice-cold buffer containing sucrose 0.25 mol/L, HEPES 10 mmol/L, EGTA 1 mmol/L, and protease inhibitors, phenylmethylsulfonyl fluoride (PMSF) 1 mmol/L, aprotinin and leupeptin 100 μg/mL. Nuclei were then extracted as previously described.23 Briefly, neutrophils were sonicated on ice (3 × 20 seconds, at a power setting of 2 and 60% duty cycle) in a Branson Ultrasonic Processor (VWR/Canlab, Montreal, Quebec, Canada). Sonicates were centrifuged at 12,000g, 10 minutes, 4°C. The corresponding pellets referred to nuclei of neutrophils and the supernatants referred to cellular membranes and cytosols. Nuclei were resuspended in HBSS in DNA was extracted as described above. The presence of EBV genome was evaluated by PCR using theBamHI W primers detailed in the previous section.

Neutrophils from IM patients were isolated and DNA was extracted as described above. Neutrophils obtained from healthy EBV-seropositive donors were also used as negative controls. The presence of EBV genome was evaluated by PCR analysis using the BamHI-W primers detailed in the previous section. PCR was performed in Perkin Elmer Gene Amp 9600 (Cetus, Emeryville, CA) for 35 cycles (denaturation: 60 seconds at 95°C, ramp 60 seconds; annealing: 30 seconds at 55°C , ramp 30 seconds; and extension: 30 seconds at 72°C, ramp 60 seconds). Amplified product was detected by hybridization to an³²P-end-labeled oligonucleotide probe: 5′-TATCTTTAGAGGGG AAAAGAGGAATAAG-3, positions 20313-20340.

Unstimulated cells and EBV-treated neutrophils (50 × 10⁶ cells/mL) were cultured for various time periods before RNA extraction. Total RNA was extracted by the TRIzol Reagent (GIBCO-BRL) method according to the manufacturer. First-strand cDNA was made using Moloney-murine leukemia virus (M-MLV) reverse transcriptase (RT) using downstream primer. For EBNA-2 detection, the downstream primer was 5′ TGACGGGTTTCCAAGACTATCC 3′, exon Y3/P, positions 48583 to 48562 and the upstream primer was 5′ AGAGGAGGTGGTAAGCGGTTC 3′, exon W2, positions 14802 to 14822. The PCR fragments were separated on a 2% agarose gel and transferred to a nylon membrane (Hybond N; Amersham) using 10× SSC. Amplified EBNA-2 products were detected by hybridization to an end-labeled oligonucleotide probe (5′ GAGAGTGGCTGCTACGCATT 3′, Y2 exon, positions 47885-47904). cDNA from Raji cell lines was used as positive controls. For BZLF-1 detection, the downstream primer used was 5′-GGCAGCAGCCA CCTCACGGT-3′, exon 2/3 splice, positions 102330-102341/102426-102433, and the upstream primer was 5′-TTCCACAGCCTGCACCAGTG-3′, exon 1, positions 102719-102700. Amplified BZLF-1 products were detected by hybridization to an end-labeled oligonucleotide probe (5′-CTTAAACTTGGCCCGGCATT-3′, exon 2, positions 102450-102469). cDNA from B95-8 cell lines were used as positive controls. The sequences of EBV PCR primers and probes used to detect BZLF-1 and EBNA-2 transcript have been validated in another study.24 The β-actin cDNA was used as internal control to demonstrate equal concentration of RNA in each sample. Sequences of primers used have been previously reported.25 

After EBV treatment, we identified living, apoptotic, and necrotic cells based on the differences in their stainability with propidium iodide (PI) and Hoechst (HO) 33342. Analyses were done by flow cytometry based on previously described procedures.26 27Briefly, human neutrophils (2 × 10⁶/mL) were cultured in RPMI-1640 containing 10% FBS in the presence or absence of EBV and obtained at specified times. Cells were washed in phosphate-buffered saline (PBS) (pH 7.4), and 50 μL of PI (20 μg/mL) was added to the pellet. Tubes were mixed and kept on ice for 30 minutes. After this incubation time, 950 μL of 25% ethanol and 25 μL of HO 33342 (112 μg/mL) were added to each tube. Samples were vortexed and kept at 4°C in the dark for 4 hours. Cells were analyzed by cytofluorometry (EPICS ELITE ESP; Coulter, Hialeah, FL) with the following settings: the fluorescence emission was first filtered through a 488-nm dichroic filter with the <488-nm fluorescence filtered through a 450-nm long-pass filter (HO33342 fluorescence). The >488-nm fluorescence was filtered through a 515-nm long-pass filter and a 560-nm dichroic filter with the >560-nm fluorescence filtered through 610-nm long-pass filters (PI fluorescence). Analyses were performed from the samples of 10,000 cells. When indicated, cells were pretreated with GM-CSF (3 nmol/L) for 3 hours before EBV infection.

Neutrophils (3 × 10⁶ cells/mL) were cultured in the presence or absence of EBV during 20 hours and obtained for immunofluorescence staining. Cell-surface expression of Fas and Fas L was assayed by flow cytometry using murine MoAbs anti-human Fas UB2 (Immunotech, Burlington, Ontario, Canada) and anti-human Fas L NOK-1 (PharMingen, Mississauga, Ontario, Canada) for primary straining and revealed with fluorescein isothiocyanate (FITC)-conjugated purified F(ab′)₂ goat anti-mouse IgG (Cappel, Durham, NC) for secondary staining. Negative control staining was performed with irrelevant murine IgG. Cells were stained with specific antibodies for 30 minutes at 4°C, washed with PBS, and fixed with 0.5% paraformaldehyde in PBS before cytometric analysis.

Cell-free supernatants from unstimulated and EBV-treated neutrophils (10 × 10⁶ cells/mL) were obtained at indicated times and tested for the release of soluble Fas L using an enzyme-linked immunosorbent assay (ELISA) kit (Medicorp, Montreal, Canada).

Statistical analyses were performed using Student's paired (two-tailed) t-test, and significance was attained atP < .05.

We have previously reported that EBV binds to human neutrophils (approximately 30%) and recognizes a receptor distinct from the CD21 antigen.17 Such interactions result in cellular aggregation and in de novo protein synthesis by neutrophils. Here we have examined the association between neutrophils and EBV by electron microscopy. Viral particles were incubated with neutrophils at 4°C to allow binding and then at 37°C for various periods of time. EBV adsorption to the outer cell membrane of neutrophils could be observed (Fig 1A and insert). Fully mature virions, with electron dense core and bilayer membrane, were found associated to the cells. In Fig 1B, a virion is engaged in internalization, as seen by the fusion of the viral envelope with the cellular membrane. Nucleocapsids of EBV were observed later in the cytoplasma and in the nuclei of neutrophils (Fig 1C and D). Chromatin was condensed following EBV localization to the nucleus and this process was observed between 5 to 15 minutes of incubation. Some EBV particles were also observed in cellular vacuoles, showing that viral particles were also phagocytized by neutrophils (data not shown). Virions within the neutrophils were identical to those seen in the cytoplasma of B95-8 cell line. The percentage of EBV-infected neutrophils evaluated by electron microscopy was similar to the one observed in binding assay.17 

To further confirm the results obtained by electron microscopy, we first examined the presence of EBV genome in neutrophils by Southern blotting analysis (Fig 2A). We used aBamHI W fragment from the internal repeats of EBV genome. Neutrophils were treated with EBV for different periods of time (from 15 minutes to 24 hours). Expression of EBV fragments was detectable at 2 hours postinfection and increased at 24 hours postinfection, suggesting that the number of copies of viral genome or the number of infected neutrophils is increasing with time. Because viral internalization can occur by phagocytosis, we performed an additional experiment to confirm the presence of EBV in the nuclei of neutrophils. Cells were first pretreated with cytochalasin B, an inhibitor of phagocytosis, then treated with EBV. At 10 hours postinfection, DNA was extracted from isolated nuclei and tested for EBV genomic DNA by PCR. As shown in Fig 2C, EBV genome was detected in isolated nuclei from neutrophils treated with EBV, suggesting that EBV can penetrate neutrophils independently of phagocytosis. As negative control, when P815 cells (murine mastocytoma exerting phagocytosis) were pretreated with cytochalasin B before EBV treatment, no specific amplification of EBV DNA was detected in isolated nuclei under the same experimental conditions (data not shown).

We next investigated the expression of two viral transcripts in neutrophils. First, we amplified by PCR the early lytic BZLF-1 gene that encodes the Zebra protein and second, the EBNA-2 transcript that is associated with EBV immortalization. In both cases, no transcripts were detected in EBV-infected neutrophils in contrast to the positive controls B95-8 and Raji cells (Fig 3).

Because EBV was found to infect human neutrophils in vitro, we tested for the presence of EBV genome in neutrophils from IM patients. This was performed by PCR amplification of a specific region of the EBV genome using DNA isolated from neutrophils of EBV patients. Interestingly, EBV genome was detected in neutrophils from EBV patients in contrast to healthy seropositive donors (Fig 4).

Because EBV can induce or inhibit apoptosis in mononuclear cells28 29 and the fact that neutropenia is observed in the majority of IM patients, we next investigated the effects of EBV on neutrophil viability. This was performed by flow cytometric analysis using the dye HO33342 and DNA intercalating dye propidium iodine. This technique allows the discrimination between necrotic, apoptotic, and viable cells. Viable cells can be divided into three subpopulations, V₁, V₂, and V₃, based on their degree of permeability to PI. Cells in V₁ are less permeable to PI than cells in V₂, and cells in V₂ are less permeable to PI than cells in V₃. Permeability is inversely proportional to cellular viability. The more permeable cells, found in the V₃ area, are thus closer to death than cells located in the V₁ or V₂regions. In one representative experiment, after 20 hours of culture, neutrophils treated with EBV showed 77% apoptotic cells as compared to 22% for unstimulated cells (Fig 5). Necrotic neutrophils were not detected at 20 hours postinfection, suggesting that EBV causes neutrophil death by apoptosis rather than necrosis. In addition, cells in V₃ area were only detected in EBV-treated cultures, indicating a decrease in cellular viability. These results were consistently observed throughout several experiments. The apoptotic process induced by EBV in neutrophils was confirmed by two other techniques, eg, condensation of chromatin (percent of apoptotic cells: control, 31% ± 6%; EBV-treated cells, 59% ± 10%; P < .05; n = 3) and DNA fragmentation studies (data not shown).

Human neutrophils were incubated with or without EBV for increasing period of time and apoptosis was assessed by flow cytometric method (Fig 6). The number of apoptotic cells significantly increased after 20 hours in EBV-infected cell cultures as compared with unstimulated neutrophils.

GM-CSF is known to prolong survival of human neutrophils in vitro.30 To determine if this effect was also protective on EBV-treated neutrophils, we incubated unstimulated and EBV-infected neutrophils with or without GM-CSF. After 20 hours of incubation, the percentage of apoptotic cells was evaluated by flow cytometry. As shown in Table 1, treatment with GM-CSF strongly suppressed the ability of EBV to induce neutrophil apoptosis. In fact, the number of apoptotic cells in EBV-treated cultures was similar to that from unstimulated cell cultures.

The Fas/Fas L system is an important cellular pathway mediating apoptosis in neutrophils.31 We therefore evaluated the effects of EBV on the expression of Fas/Fas L antigens on neutrophils. As shown in Fig 7, the expression of Fas and Fas L was significantly increased on neutrophils infected by EBV (93% and 19%, respectively). The time course of appearance of Fas and Fas L antigens was similar to the kinetics of neutrophil death induced by EBV. Furthermore, the release of soluble Fas L was also increased in culture supernatants of EBV-treated neutrophils (Fig 8). These results suggest that EBV may induce apoptosis in neutrophils via the Fas/Fas L system and that soluble Fas L may function in an autocrine/paracrine pathway to mediate cell death.

As a key element in the nonspecific defense mechanism, neutrophils are likely to encounter invading agents such as viruses. Two of the major functions of neutrophils are phagocytosis and the release of inflammatory mediators. Growing evidence shows the ability of neutrophils to induce inflammatory response by releasing several immunoregulatory cytokines upon stimulation.1,2 In this study we show that EBV infects and induces apoptosis of human neutrophils, which could disrupt the initial defense against EBV infection. Electron microscopy studies show contact between EBV and neutrophils, as well as fusion between cellular and viral membranes. These results support a previous study17 showing that EBV binds to a surface membrane antigen on neutrophil via a receptor different from CD21. This is also in agreement with other studies using T-cell lines.11,16 The increased expression of EBV fragments in neutrophils over time either may suggest that EBV can replicate in these cells or that more neutrophils get infected over time. However, we cannot overlook that a number of viral particles can also enter into neutrophils by phagocytosis. In this regard, the results obtained with isolated nuclei from EBV-infected neutrophils are of particular interest. In fact, the presence of EBV genome in these nuclei indicates that EBV can penetrate into neutrophils without being phagocytosed because all cell cultures were treated with cytochalasin B, an inhibitor of phagocytosis. Taken together, the results indicate that although some EBV particles are phagocytosed, others penetrate into neutrophils via an alternate route. Since EBV can penetrate neutrophils, we looked for the synthesis of viral transcripts EBNA-2 and ZEBRA. EBNA-2, readily detectable in the first 24 hours of infection,32 is required for the initiation of lymphocyte immortalization and is an essential transactivator of viral gene expression.33,34 ZEBRA is a key immediate-early protein essential for lytic cell induction and is synthesized during the first hours postinfection.35 36 None of these two genes were found to be expressed in EBV-infected neutrophils, suggesting an abortive type of infection. However, because EBV can encode several other genes, further studies are required to identify viral genes expressed in EBV-infected neutrophils.

Death of neutrophils may represent a strategy of host cells to abort the infectious process. Indeed, after 20 hours of culture, neutrophils treated with EBV showed more than 70% of apoptotic cells which may significantly affect the establishment of productive infection. Such a cellular defense, named apoptosis or programmed cell death, has been reported with Sinbis virus37 and with influenza A and B virus.38-40 It was previously shown that the expression of the EBV latent membrane protein 1 (LMP-1) protects infected B cells from apoptosis and that this effect is mediated, at least in part, by the upregulation of the bcl-2 proto-oncogene.28,41EBNA-2, another EBV latent protein, can increase the effect of LMP-1 onbcl-2 expression, thus improving the protection against apoptosis. This mechanism does not seem to be used by EBV with regard to neutrophils, especially as EBNA-2 protein was not detected in neutrophils; moreover, bcl-2 and bcl-x expression is likely to be absent in this cell type.42 43 However, we evaluated the regulation of the human bfl-1 gene, abcl-2 homologue, in EBV-treated neutrophils. Unfortunately, no significant difference was observed between unstimulated and EBV-infected neutrophils after 20 hours of culture.

It was recently reported that apoptosis in neutrophils may be regulated, at least in part, by the coexpression of cell-surface Fas and Fas L. In the present study our results suggest that the Fas/Fas L system may be involved in neutrophil death induced by EBV. The presence of EBV in cell cultures significantly increased the expression of Fas and Fas L on the membrane surface of neutrophils, as compared with unstimulated cells. Similarly, EBV was found to induce the release of soluble Fas L by neutrophils. However, treatment of freshly isolated neutrophils with supernatants obtained from EBV-treated neutrophils did not significantly modify the rate of spontaneous apoptosis observed in unstimulated neutrophils. At present we cannot exclude the involvement of the Fas/Fas L system in the apoptotic process induced by EBV. Other soluble mediators induced by EBV in addition to virally encoded proteins (still unidentified) may act in synergy with the Fas/Fas L to induce apoptosis of neutrophils. EBV-induced apoptosis of neutrophils is a complex phenomenon necessitating multiple events. In fact, viral entry is necessary but not sufficient to induce apoptosis. This statement is supported by the fact that UV-irradiated particles do not cause apoptosis of neutrophils even though viruses enter the cells (data not shown). The fact that EBV infects and induces apoptosis in neutrophils in vitro and that we detect EBV genome in neutrophils from IM patients are in accordance with some clinical observations made in IM and immunocompromised patients. In 60% to 90% of IM patients, an absolute neutropenia has been detected between the third and fourth week of illness.44-48 This was also observed in severe chronic EBV infection where the number of circulating neutrophils was found to decrease. Moreover, antineutrophil antibodies have been detected in sera of a high proportion of patients with active IM,49 50 a process possibly associated with death of neutrophils.

In conclusion, we showed that EBV penetrates and causes apoptosis of neutrophils. Interactions of EBV with phagocytes may alter the primary immune response and favor the spread of EBV infection. Further studies on the mechanisms of EBV-induced apoptosis in neutrophils should give us new insights on the interactions between EBV and this cell type.

We thank Pierrette Côté for her excellent secretarial assistance. We also thank Dr Robert Delage, who provided blood samples from IM patients.