Leukotriene B₄ Induces Release of Antimicrobial Peptides in Lungs of Virally Infected Mice¹

Leukotriene B₄ (LTB₄)³ is a potent lipid mediator of inflammation synthesized predominantly by leukocytes of the myeloid lineage such as monocytes/macrophages and neutrophils (1). LTB₄ can bind to two cell surface receptors of the G-protein-coupled receptor family; a high-affinity receptor named BLT1R (2, 3, 4, 5, 6, 7) and a low-affinity receptor named BLT2R (8). BLT1R seems to preferentially be expressed on leukocytes, whereas BLT2R has a more widespread expression (reviewed in Ref. 9).

Complementary to its proinflammatory role, LTB₄ is also important in host defense. The first evidence of antimicrobial defense engendered by leukotriene administration was presented by Demitsu et al. (10) in 1989. This group showed that i.p. injection of LTB₄ led to the resolution of an experimental bacterial peritonitis in mice. Endogenous leukotriene production was also shown to be an important event in innate defense as suggested by Bailie et al. (11), who observed that mice deficient in one important enzyme in the cascade of LTB₄ generation, 5-lipoxygenase, were not able to clear the microorganism Klebsiella pneumoniae from the lungs as opposed to wild-type mice. Moreover, the deficient mice exhibited a higher mortality rate than wild-type mice after infection with K. pneumoniae.

More evidence now indicates that LTB₄ is an important mediator involved in the control of viral infection. Endogenous production of leukotriene has been shown to be protective during the early phases of experimental vesicular stomatitis virus-mediated encephalitis (12). Tafalla et al. (13) observed a potential role for LTB₄ in the antiviral activity of turbot leukocytes against viral hemorrhagic septicemia virus. We also have demonstrated that administration of LTB₄ to healthy human subjects led to the production of anti-HIV mediators such as α-defensins and MIP-1β (14). Recently, we confirmed the antiviral potential of LTB₄ in an animal model of CMV infection. Mice treated with LTB₄ had lower viral burden than untreated control mice. Moreover, LTB₄ was also found to protect immunosuppressed mice from virus reactivation following allogeneic transplantation (15).

Neutrophil granules contain molecules with potential antiviral activities. α- and β-defensins are small cationic peptides contained in the large granule subset of primary/azurophil granules (reviewed in Ref. 16) that exert antiviral activities against different enveloped viruses such as CMV, influenza, HSV-1 and -2, and vesicular stomatitis virus (17). Defensins were also found to inhibit HIV-1 replication in vitro (18, 19, 20). Cathelicidins are other cationic peptides with LL-37 as the only human member and cathelicidin-related antimicrobial peptide (CRAMP) being its murine counterpart (reviewed in Refs. 21 and 22). LL-37 precursor, human cationic antimicrobial protein, is contained in the specific granules of human neutrophils (23) and is cleaved extracellularly by proteinase 3 (24) to release the active antimicrobial LL-37. LL-37 has known selective antiviral activity against vaccinia virus (25) and partial antiviral activity against HSV-1 (26). Eosinophil-derived neurotoxin, a member of the RNase gene family (27), is predominantly expressed in human eosinophils but is also detected in neutrophil cell lysate (28, 29). Moreover, a GPI-anchored form of eosinophil-derived neurotoxin (EDN) has been detected on the surface of human neutrophils (30). EDN has known antiviral activities against respiratory viruses such as respiratory syncytial virus group B (31) and its related rodent counterpart, pneumonia virus of mice (32). EDN also has partial inhibitory activity against HIV-1 (33). In mice, EDN counterparts are RNases of the murine eosinophil-associated RNase family (mEARs), with mEARs possessing antiviral activity against pneumonia virus of mice (34).

In this study, we investigate the role of LTB₄ in the control of infection by respiratory viruses. The viral pathogen chosen was human influenza virus due to its capability to efficiently infect mice. In addition, influenza infection is mostly controlled by neutrophil recruitment to the lung (35, 36), providing an interesting model of innate immunity.

The aim of this study was to determine which types of antimicrobial peptides are released by LTB₄-treated cells leading to a reduction in viral load during in vitro and in vivo infection with influenza A virus and whether such an effect is mediated through BLT1R or BLT2R receptor.

Neutrophils were obtained from peripheral blood of healthy medication-free volunteers after informed consent in accordance with an Internal Review Board-approved protocol. Briefly, peripheral blood leukocytes were enriched by dextran sedimentation, layered over a lymphocyte separation medium cushion, and centrifuged at 1400 rpm for 20 min. Mononuclear leukocytes were collected at the interface, whereas neutrophils were obtained from the pellet as described (37). Cell preparations were depleted of erythrocytes by osmotic shock and then washed and resuspended in HBSS supplemented with calcium until use. Purity of neutrophil preparations was >99%. No significant percentage of other types of granulocyte was detected in cell preparation as evaluated by cytometry analysis. When applicable, cellular preparation could also be depleted of monocytes by adherence of PBMC on autologous serum-treated petri dishes.

Cell cultures were infected with influenza virus at 0.1 multiplicity of infection before a stimulation, 2 min later, with different doses of LTB₄ (0.1, 1 or 10 nM; Johnson Matthey) for 48 h, without cell washing, to investigate direct virion inactivation mediated by LTB₄ treatment. LTB₄ was obtained as an ethanolic solution of the acid form and prepared by dilution of the ethanolic LTB₄ in a saline solution 0.45% w/v NaCl containing 0.25% w/v dextrose. Saline solution without addition of LTB₄ was used as placebo. Cell-free supernatants were then harvested for viral load determination on Madin-Darby canine kidney (MDCK) cells. In some experiments, peripheral blood leukocytes or isolated neutrophils were pretreated or not with the specific BLT1 antagonist U75302 (1 μM; BIOMOL) or BLT2 antagonist LY255283 (1 μM; kindly provided by Eli Lilly) for 60 min before influenza infection, or with anti-α-defensin (HNP 1–3; Alpha Diagnostic Int.), anti-LL-37 (HyCult Biotechnology), or anti-EDN (MBL Corp.) Abs (1 μg/ml) before LTB₄ stimulation.

Influenza A virus (strain A/Puerto Rico/8/34, H1N1) was propagated and isolated from infected MDCK cells and titrated using standard plaque assay (38). The MDCK cell line was grown in MEM supplemented with 10% heat-inactivated FBS and gentamicin sulfate (40 μg/ml).

Female C57BL/6 mice 5–6 wk old were purchased from Charles River Laboratories. BLT1R^−/− mice (on a C57BL/6 background) were generously provided by Dr. Andrew D. Luster (Harvard University, Boston, MA). For in vivo protocols, mice were infected intranasally with 50 PFU of influenza A virus (strain A/Puerto Rico/8/34) and treated i.v. with 1000 ng/kg LTB₄ daily from day 1 up to day 7 postinfection. Ethanolic LTB₄ was diluted in NaCl, 0.45% w/v, containing dextrose, 0.25% w/v, before administration. Placebo treatment was referred to saline solution without LTB₄. Mice were then sacrificed, and lungs were extracted for viral load assessment.

For neutrophil depletion experiments, mice were injected i.p. with 250 μg of anti-Ly6G/Gr-1 Abs (clone RB6-8C5; eBioscience) or isotypic rat Abs as already reported (39). Mice received one injection of Abs 24 h before infection with influenza virus (50 PFU/mouse), followed by a second injection on day 2 postinfection. Mice were treated daily with a placebo or with LTB₄ (1000 ng/kg) starting at day 1 postinfection and were sacrificed at day 5 postinfection. Lungs were extracted for viral load assessment and immunohistochemistry examination as described below. Neutrophil depletion was evaluated by flow cytometry in peripheral blood and lung tissue using PE-conjugated anti-Ly6G/Gr-1 Ab (BD Biosciences).

Wild-type and BLTIR^−/− mice were infected intranasally with 50 PFU of influenza A virus. At 24 h postinfection, mice were treated i.v. or not with LTB₄ (1000 ng/kg) for 1 h and sacrificed for lung extraction. Lung tissues were embedded in OCT compound. Sections of 5 μm were mounted on Superfrost-Plus slides for immunohistochemistry and hematoxylin staining. Immunohistochemistry was performed using an indirect method. Sections were treated with 0.3% hydrogen peroxide for 30 min. Sections were then blocked with 3% FBS for 30 min followed by incubation with isotypic or anti-CRAMP Abs (provided by Dr. R. L. Gallo, University of California, San Diego, CA), anti-β-defensin-3 Abs (Alpha Diagnostic Int.) or anti-mEARs Abs (Drs. Nancy and Jamie Lee, Mayo Clinic, Scottsdale, AZ) for 1 h in a humidified chamber. Sections were washed in PBS, incubated with biotinylated secondary Abs for 1 h at room temperature followed by incubation with HRP-streptavidin complex for 20 min, and visualized using diaminobenzidine substrate followed by counterstaining with hematoxylin.

Lung tissues were prepared as described in Immunohistochemistry. Sections were incubated for 1 h with anti-CRAMP Abs, washed three times, and incubated with anti-rabbit ALEXA 488-conjugated Abs for 1 h. Sections were then stained for 1 h with PE-conjugated anti-Gr-1/Ly6G (BD Biosciences), a specific marker of mature murine neutrophils. Fluorescence was visualized by confocal microscopy.

Neutrophils (2 × 10⁶) were allowed to sediment for 1 h at 37°C in the presence or not of U75302 (1 μM) followed by treatment with cytochalasin B (10 μM) for 30 min at 37°C. Cytochalasin B was used as a priming agent for neutrophil degranulation and has no significant effect on cell activation (40, 41). After cytochalasin treatment, cells were stimulated or not with LTB₄ for 30 min. Next, the cellular suspension was centrifuged at 4°C, and cell-free supernatant was collected and stored at −80°C until assayed for LL-37 (HyCult Biotechnology) or EDN (MBL Corp.) by ELISA following the manufacturer’s instructions.

Neutrophils (2 × 10⁶ cells) were pretreated for 30 min with cytochalasin B (10 μM) followed by incubation with or without LTB₄ (10 nM) for 1 h. Cells were fixed for 15 min at room temperature with 2% paraformaldehyde and then incubated in the presence of anti-EDN Ab for 45 min. After three different washes with HBSS, goat anti-mouse FITC Ab was applied, and incubation for 45 min was performed. Cell surface expression of EDN was analyzed on 10,000 cells per sample using an EPICS XL apparatus (Beckman Coulter).

Data were analyzed by one-tailed ANOVA followed by a Newman-Keuls post hoc test using PRISM3 software. Differences were considered significant at p ≤ 0.05.

We first wanted to investigate the antiviral potential of LTB₄ in a mouse model of influenza infection. Although the BALB/c mouse strain is usually used for studies on viral infection of the lungs due to high viral replication levels and high viral titers on an extended period of time, we chose to use C57BL/6 mice. Infection with influenza virus of C57BL/6 mice leads to high viral titers in lungs of mice early during infection. In addition, we also chose C57BL/6 mice to facilitate comparison and interpretation of data obtained using BLT1R knockout mice generated on a C57BL/6 background. Mice were infected intranasally with influenza A virus (A/PR/8/34 strain) followed by daily i.v. treatment with LTB₄. Whereas C57BL/6 mice naturally control viral infection by reducing viral loads, administration of LTB₄ potentiated viral clearance (82% reduction in viral loads in lungs at day 5 postinfection), as opposed to mice treated with a placebo (Fig. 1). To verify the implication of the high-affinity LTB₄ receptor in such antiviral defense against influenza infection, we performed the same experiment using wild-type or BLT1R-deficient mice. Mice were infected intranasally with influenza virus and were treated daily by i.v. injection starting 24 h postinfection, followed by sacrifice and lung removal on day 5. Wild-type mice treated with LTB₄ showed a significant reduction in viral load as opposed to wild-type mice treated with a placebo (Fig. 2). On the other hand, no viral load reduction was observed in BLT1R^−/− mice treated with LTB₄ compared with knockout mice treated with placebo. These results clearly demonstrate that the antiviral activity of LTB₄ against a respiratory viral pathogen occurs via the activation of the high-affinity BLT1R receptor.

Influenza infection is mostly associated with a massive lymphoid cell recruitment infiltrating the epithelium of lung tissue along with a decrease number of alveolar macrophages (42). Tissue changes are also observed during the course of infection with cellular exudates present in the luminal content of the airways as well as significant epithelial cell changes and necrosis. To evaluate whether LTB₄ treatments could ameliorate tissue integrity in the lungs of influenza-infected mice, we infected mice with influenza followed by daily i.v. LTB₄ treatments from days 1–4 postinfection. Mice were sacrificed at day 5 postinfection, and lungs were harvested and stained using H&E. As shown in Fig. 3, mice treated with LTB₄ showed a significant cellular recruitment only to the lungs, probably of neutrophils to mount a potential immune response against an invading agent. When mice were infected with influenza and left untreated, massive leukocyte recruitment was observed in lung tissue. Moreover, tissue changes were also present in the epithelium as well as thickening of the alveolar wall. When mice were treated with LTB₄ after influenza infection, a more natural histological lung architecture was present, with less leukocyte infiltration as well as less plugging of the airways and thickening of the alveolar wall. These results demonstrate that not only does LTB₄ reduce viral load in lung tissue of infected mice, but it also helps to restore a more normal lung architecture.

We next investigated by which mechanism(s) LTB₄ administration could potentiate viral clearance. We already demonstrated that LTB₄ administration to healthy human subjects lead to secretion of defensins of the α subtype by neutrophils (15). Because mice do not produce α-defensins, we tested for the presence of other antimicrobial peptides with antiviral potential following LTB₄ treatment. Mice were infected or not with influenza virus and injected with placebo or LTB₄ 24 h postinfection. One hour after LTB₄ treatment, mice were sacrificed, and lung tissues were harvested for detection of antimicrobial peptides. As shown in Fig. 4, a slight up-regulation in β-defensin-3 was observed following influenza infection probably due to natural immune activation in the lungs. A slight up-regulation in β-defensin-3 was also observed in mice that were treated with LTB₄ without viral infection (data not shown). However, when wild-type mice were infected with influenza followed by treatment with LTB₄, a strong up-regulation in β-defensin-3 was observed, mostly in epithelial cells of the airways which correlates with the fact that BLTR is detectable in lungs of mice (43, 44). Similarly, up-regulated expression of β-defensin-3 was also accompanied by a reduction in viral loads in lungs of infected mice treated with LTB₄. On the other hand, up-regulation of β-defensin-3 was not observed in lungs of infected BLT1R^−/− mice, demonstrating the requirement for BLT1R activation for such engagement of the innate immune response. A similar pattern of expression was observed for other peptides bearing antiviral potential such as mEARS (Fig. 5) and CRAMP proteins (Fig. 6). CRAMP proteins are mostly produced by neutrophils and to a much lesser extent by epithelial cells. Because immunohistochemistry using anti-CRAMP Abs could not identify with absolute certainty that CRAMP-positive cells were in fact neutrophils, we next performed confocal microscopic examination using the selective marker of murine neutrophil, Gr-1/Ly6G. As shown in Fig. 7, intracellular CRAMP protein colocalizes with Gr-1/Ly6G-positive neutrophils (yellow color on merged panels) that migrated in lung tissue following influenza infection. This observation clarifies the fact that murine neutrophils do produce CRAMP protein after LTB₄ administration during viral infection. It is interesting to note the presence of extracellular staining of CRAMP protein (in green, indicated by arrowheads) in lung tissue that might be the result of neutrophil degranulation following stimulation by LTB₄. This extracellular detection of CRAMP protein seems to be associated with interstitial tissue area where neutrophils migrate after viral threat infection. Moreover, more CRAMP⁺Gr-1/Ly6G⁺ neutrophils were present in lung tissue of infected wild-type mice treated with LTB₄ as opposed to healthy mice or infected mice treated with placebo. All the results presented above demonstrate that LTB₄ administration leads to antimicrobial peptide secretion in lungs of mice, a process that may help control influenza infection.

We previously hypothesized that LTB₄ might target neutrophils to reduce viral presence in infected tissue (15). To investigate the effect of LTB₄ on neutrophil mobilization in mice as well as the role of neutrophils in viral clearance, we injected Abs directed against Ly6G/Gr-1 Ag i.v. into mice to deplete peripheral neutrophils. As shown in Table I, following LTB₄ injection, a substantial reduction in peripheral blood neutrophils was observed, suggesting that neutrophils leave the blood circulation after LTB₄ administration. As expected, no neutrophil could be detected in the bloodstream of mice injected with anti-Ly6G/Gr-1 Abs. Although blood neutrophil counts were not affected by influenza infection, the number of circulating neutrophils increased when mice were infected with influenza virus and treated with LTB₄. This fact is intriguing, given that neither LTB₄ administration nor influenza infection seemed to promote hematopoiesis (E. Gaudreault and J. Gosselin, personal observation).

As expected, daily administration of LTB₄ maintains recruitment of neutrophils in lung tissues even after 4 days of treatment, as compared with the placebo group (Table II). A more substantial mobilization of neutrophils in lungs of mice was also observed at 1 h after LTB₄ administration (data not shown). The number of neutrophils in lungs of mice infected with influenza virus was also found to increase, a process that may reflect a natural host defense against virus infection, as previously suggested by others (45). In line with these results, a massive neutrophil recruitment was also observed in lung tissues of mice infected with influenza virus and treated with LTB₄ as opposed to all other experimental groups of mice (Table II). Overall, this set of experiments suggests that LTB₄ administration to mice infected or not with influenza virus potentiates migration of neutrophils from peripheral blood to the lungs, a mechanism that may contribute to enhance viral clearance.

To further provide an in vivo confirmation of the link between neutrophil recruitment to the lung, influenza viral clearance, and secretion of antimicrobial peptides, we measured influenza titers in the lungs of mice depleted of neutrophils and performed lung immunostaining for the presence of antimicrobial peptides following LTB₄ administration. As shown in Fig. 8,A, when mice were injected with isotypic Abs and treated with LTB₄ after influenza infection, a significant decrease in lung viral titer was observed. On the other hand, when mice were depleted of their neutrophils and treated with a placebo, the lung viral titers were up-regulated by ∼5-fold as opposed to mice treated with isotypic Ab. Moreover, LTB₄ administration was ineffective in reducing lung viral titers, showing the importance of neutrophils in the lung for LTB₄-mediated antiviral activity. As expected, lung tissues of infected mice injected with isotypic Abs were positive for CRAMP, β-defensin-3, and mEARs staining after LTB₄ administration (Fig. 8,B, left). Moreover, nondepleted uninfected mice that were given LTB₄ also showed positive staining for CRAMP, β-defensin-3, and mEARs proteins, albeit to a lesser extent than infected mice treated with LTB₄. In infected mice injected with anti-Ly6G/Gr-1, absence of CRAMP and mEARs as well as a slight reduction in β-defensin-3 staining were observed after LTB₄ administration (Fig. 8 B, right). These results demonstrate that the inability of LTB₄ at reducing lung influenza viral load in absence of neutrophils might be associated with the reduction or absence of antimicrobial peptide secretion.

We next wanted to investigate whether the antiviral potential of LTB₄ during influenza virus infection in mice was also present following in vitro infection of human leukocytes with influenza. For that matter, peripheral blood leukocytes were isolated and infected with influenza virus followed by a treatment with LTB₄. As shown in Fig. 9, LTB₄ administration (1, 10, and 100 nM) leads to a statistically significant reduction of influenza viral load as opposed to cells only infected with the virus. When neutrophils were depleted from the cellular preparation, no significant reduction in viral load was observed after treatment with 10 nM LTB₄. This was not the case when monocytes were depleted from the cellular preparation (data not shown). Such results support, once again, the fact that neutrophils are the major cellular players in LTB₄-mediated antiviral activity. Although under natural conditions influenza virus targets preferentially lung epithelial cells for replication, we chose to use a model of peripheral blood leukocyte infection to clearly demonstrate the implication of neutrophils in LTB₄-mediated antiviral activity. As also shown in Fig. 9, when leukocytes were pretreated with a specific BLT1 antagonist (U75302; Ref. 45), the reduction in viral load was not observed following treatment with LTB₄. Pretreatment with the specific BLT2 antagonist, LY255283, did not succeed in abrogating LTB₄ antiviral activity. These results support the antiviral effect of LTB₄ in vitro against influenza virus infection of human cells via a neutrophil-dependent mechanism which involves BLT1R receptor.

Cathelicidins are a family of antimicrobial proteins found in peroxidase-negative granules of neutrophils. In human neutrophils, hCAP-18 is the only cathelicidin detected. Cleavage of hCAP-18 by proteinase 3 gives rise to LL-37 in the exocytosed material from neutrophils (24). EDN is a member of the RNase A superfamily and was found to be produced by neutrophils. EDN has antiviral activity in vitro, thus making EDN an antimicrobial mediator. Because neutrophil granule secretion seems to be an important strategy in the LTB₄-mediated neutrophil antiviral activity, we next investigated whether LTB₄ treatment might induce the production of LL-37 and EDN during viral infection. First, we observed that LTB₄ induces the release of LL-37 (Fig. 10,A) and EDN (Fig. 10,B) in a dose-dependent manner by neutrophils reaching peak secretion at a concentration of 100 nM LTB₄. Influenza virus itself did not promote neutrophil secretion of LL-37 and EDN in such in vitro setting (data not shown) as opposed to other viruses such as CMV known to induce secretion of such peptides by neutrophils (41). As previously mentioned, β-defensin-3 expression was up-regulated in mouse lung tissue following LTB₄ administration. However, human neutrophils do not seem to produce β-defensin-3. Whereas β-defensins have been isolated from leukocytes and epithelial tissues, human neutrophils express only low levels of β-defensin-4 (HBD-4), which displays a selective spectrum of antibacterial activity (46). We could not detect up-regulation in HBD-4 transcription following stimulation of human neutrophils with LTB₄ or influenza virus (data not shown). This observation suggests that the release of HBD-4 may be regulated by different stimuli and different signal pathways not activated by LTB₄ or viral infection. The release of EDN molecules by neutrophils following LTB₄ is of particular interest because, although the presence of EDN in neutrophils has been previously identified (29), its release from neutrophils following stimulation by LTB₄ or other physiological agonists has not yet been reported. Similarly, the stimulation of neutrophils with LTB₄ for 1 h also led to a significantly higher number of membrane-bound EDN molecules at the cellular surface of neutrophils (Fig. 10 B, inset).

It seems likely that the release of α-defensins, LL-37, and EDN by neutrophils is an important event activated by LTB₄ to counter influenza virus infection. To further support this assumption, we performed another set of experiments. Peripheral blood leukocytes were pretreated with a combination of neutralizing Abs directed against α-defensins, LL-37 and EDN before influenza infection and to treatment with LTB₄ (10 nM). As shown in Fig. 10 C, the effect of LTB₄ in reducing viral loads was significantly affected by the presence of neutralizing Abs in cell cultures, indicating that LTB₄ can mediate its antiviral effect through the release of α-defensins, LL-37, and EDN. However, because incubation of leukocytes with all three types of neutralizing Abs did not completely abrogate the effect of LTB₄ on the viral load, we must consider that other mediators such as superoxide anion and/or the release of primary granules containing myeloperoxidase can also be activated in LTB₄-treated leukocytes to reduce the number of viral particles.

We can now postulate that the secretion of the antimicrobial peptides defensins, EDN, and LL-37 (or their murine counterparts) may represent a mechanism triggered by LTB₄ to counter viral spread.

In this study, we provide biological mechanisms implicated in LTB₄-mediated antiviral activity against in vivo as well as in vitro infection with influenza virus. We also demonstrate the potential implication of neutrophil secretion of antimicrobial peptides such as β-defensins, mEARs and CRAMP in a mouse model of infection and human LL-37 and EDN in such activity. These LTB₄-mediated neutrophil activities also involve the high-affinity LTB₄ receptor, BLT1.

This study clearly shows that administration of leukotriene B₄ in mice infected with influenza virus helps controlling viral infection via the triggering of BLT1R receptor. The antiviral action mediated by LTB₄ seems to involve the secretion of antimicrobial peptides in lung tissue. We demonstrated that neutrophils leave the bloodstream to mobilize in the infected tissue (in this case the lung) following LTB₄ administration. These neutrophils are believed to secrete peptides known to possess antiviral properties such as defensins, CRAMP, and mEARs to control and clear influenza infection.

Degranulation is a major biological activity of human neutrophils that encounter foreign microorganisms. This degranulation can be induced by LTB₄ present at the site of infection (47, 48). Neutrophil granules such as azurophil/primary, secondary, and secretory granules contain different molecules known to possess antiviral properties. Defensins are cationic peptides that possess such antiviral activities. Already in 1986, Daher et al. (17) observed that the human α-defensin HNP-1 was able to directly inactivate different viruses such as CMV, HSV types 1 and 2, vesicular stomatitis virus, and influenza virus strain A/WSN, probably by impairing the ability of such viruses to infect cells. During clinical trials with healthy human subjects, we also demonstrated that neutrophil HNPs are secreted by healthy patients following i.v. LTB₄ administration (14). Also, neutrophils from HIV-positive patients do retain the capability of secreting HNPs after in vitro stimulation with LTB₄. In the present study, because mice do not produce defensins of the α subtypes, we evaluated the secretion of defensins of the β subfamily following LTB₄ administration. In this regard, β-defensin-3 is known to possess antiviral potential against multiple viruses such as HSV-1 (49), vaccinia virus (50), and HIV-1 (31).We observed an up-regulated expression of β-defensin-3 in lung tissue of influenza-infected mice treated with LTB₄. Murine β-defensin-3 is already known to be inducible in epithelia of different organs following Gram-positive bacterial infection (51). Similarly, our results show that production of β-defensin-3 was induced in airways epithelial cells, but during the course of a viral infection. Moreover, β-defensin-3 induction was potentiated following LTB₄ administration. On the other hand, human neutrophils do not produce β-defensin-3 (as reviewed in Ref. 52), and to our knowledge murine β-defensin-3 has not been associated to neutrophils. It is therefore not surprising that the up-regulation in β-defensin-3 in lung tissue of infected mice treated with LTB₄ was also present in mice depleted of their neutrophils. In this case, LTB₄ would seem to act directly on epithelial, endothelial, or fibroblastic cells leading to secretion of such peptides. The fact that BLT1 mRNA is detectable in lung tissue of mice (43) and that human bronchial fibroblast cells (53), endothelial cells (44), and cells of the alveolar wall (54) express BLT1 at their surface further reinforce this assumption.

CRAMP protein staining was also up-regulated in lung tissue of mice following LTB₄ administration, and most intracellular CRAMP was present in neutrophils of the interstitial tissue. In addition, we provide evidence that LTB₄ does stimulate secretion of human LL-37 in a dose-dependent manner in vitro. Given that LL-37 exerts antiviral activities but also acts as a chemotactic agent for neutrophils, monocytes, and T cells (55, 56), we can postulate that LTB₄ administration provides a link between innate and adaptive immunity by stimulating neutrophils to secrete LL-37 that has the ability to recruit and stimulate cells of the adaptive immunity at the site of viral infection. Thus, we must consider that the release of LL-37 by LTB₄-stimulated neutrophils should participate in the elimination of viral particles in vivo. The fact that more CRAMP peptide was present in the lungs of wild-type influenza-infected mice treated with LTB₄ supports such a conclusion. Moreover, no CRAMP was detected in lung tissue when mice were depleted in neutrophils which correlated with an abrogation of LTB₄-mediated and neutrophil-dependent viral clearance.

EDN is a protein of the RNase A family mostly present in eosinophils but also detected in neutrophils (29). It can be synthesized as a secretory molecule present in neutrophil granules or as a GPI-anchored form at the cellular membrane (30). During this study, we observed that LTB₄ stimulation at a physiological concentration leads to neutrophil secretion of EDN in a dose-dependent manner. To our knowledge, this is the first report demonstrating a physiological stimulus leading to neutrophil secretion of EDN. Moreover, an up-regulated concentration of EDN murine counterparts mEARs was also detected in infected mice that were given LTB₄. The level of mEARs in lung tissue matched perfectly with the presence of neutrophils in lung tissue following LTB₄ administration. Because influenza is predominantly a respiratory virus, our results correlate well with the known antiviral activity of EDN directed against respiratory pathogens (57). Moreover, EDN can induce dendritic cell maturation and activation leading to the production of a variety of inflammatory cytokines, chemokines, growth factors, and soluble receptors (58). Therefore, by leading to neutrophil secretion of both LL-37 and EDN, LTB₄ may control the orchestration of the antiviral response by activating actors of both innate and adaptive immunity. The results of the present study also prompt us to postulate that CRAMP secretion is totally dependent on LTB₄ action on neutrophils. On the other hand, we cannot rule out the fact that an indirect action of LTB₄ on neutrophils that would lead to mEARs secretion by epithelial cells could also be involved. Regarding β-defensin-3, our results suggest that LTB₄ must have a direct action on epithelial cells for secretion of such peptide during viral infection.

We have observed that maximal secretion of antimicrobial peptides was detected early after LTB₄ administration (1 h), demonstrating a potentially very early antiviral defense against influenza infection. Secreted peptides in lung tissues were present at a lower but still detectable concentration later after LTB₄ administration (24 h). Because mice were treated daily with LTB₄, we can postulate that this cycle of peptide secretion over a period of 24 h after LTB₄ administration happened daily following drug administration, and this accumulation of antimicrobial peptides in mouse lungs over time culminated in viral clearance. This hypothesis is supported by the fact that LTB₄ treatment seemed more beneficial after four administrations (day 5 postinfection) as opposed to two administrations (day 3 postinfection). At this moment, it is not clear why LTB₄ treatments were less successful at clearing influenza infection 7 days postinfection as opposed to 5 days. However, a probable explanation lies in the fact that granular pools of antimicrobial peptides in lung tissue neutrophils might be exhausted after repeated administrations of LTB₄. This hypothesis is now under investigation. Moreover, because BLT1-mediated neutrophil degranulation involves the activation of the Src family kinase member Yes (59) and PI3K (60), it would be of great interest to further study the intracellular events implicated in BLT1-mediated antiviral neutrophil granule secretion.

To correlate results obtained with a mouse model to humans, we performed in vitro infection of peripheral blood leukocytes. Such population contains neutrophils that seem to be involved in LTB₄-mediated antiviral activity, given that neutrophil depletion abrogated such antiviral mechanism. This report also shows that human neutrophils do, in a dose-dependent manner, secrete LL-37 and EDN following LTB₄ stimulation. Such secretion, accompanied with α-defensin secretion, seem to be an important antiviral mechanism used by LTB₄-stimulated neutrophils in the elimination of influenza virions. This fact was supported by the use of neutralizing Ab directed against such antimicrobial peptides.

In this study, we demonstrate a BLT1-dependent neutrophil antiviral activity directed against influenza virus following administration of LTB₄. This neutrophil-mediated antiviral response implies the secretion of granules containing different peptides bearing antiviral activities such as α-defensin peptides, cathelicidins, and EDN. LTB₄, by orchestrating a very complex antiviral immune response, can now be seen as a potential therapeutic agent for the treatment of a diversity of viral infections.

We thank Mrs. Pierrette Côté for her secretarial assistance and Dr. Eric Philippe from the Department of Anatomy and Physiology, Faculty of Medicine, Laval University (Quebec, Canada), for his expertise in histology.

The authors have no financial conflict of interest.