|Home | About | Journals | Submit | Contact Us | Français|
The threat of a pandemic spread of highly virulent influenza A viruses currently represents a top global public health problem. Mass vaccination remains the most effective way to combat influenza virus. However, current vaccination strategies face the challenge to meet the demands in a pandemic situation. In a mouse model of severe influenza virus-induced pneumonitis, we observed that prior nasal administration of an attenuated strain of Bordetella pertussis (BPZE1) provided effective and sustained protection against lethal challenge with two different influenza A virus subtypes. In contrast to most cross-protective effects reported so far, the protective window offered upon nasal treatment with BPZE1 lasted up to at least 12 weeks, suggesting a unique mechanism(s) involved in the protection. No significant differences in viral loads were observed between BPZE1-treated and control mice, indicating that the cross-protective mechanism(s) does not directly target the viral particles and/or infected cells. This was further confirmed by the absence of cross-reactive antibodies and T cells in serum transfer and in vitro restimulation experiments, respectively. Instead, compared to infected control mice, BPZE1-treated animals displayed markedly reduced lung inflammation and tissue damage, decreased neutrophil infiltration, and strong suppression of the production of major proinflammatory mediators in their bronchoalveolar fluids (BALFs). Our findings thus indicate that protection against influenza virus-induced severe pneumonitis can be achieved through attenuation of exaggerated cytokine-mediated inflammation. Furthermore, nasal treatment with live attenuated B. pertussis offers a potential alternative to conventional approaches in the fight against one of the most frightening current global public health threats.
Influenza virus pandemics are unpredictable but recurring events that can have severe consequences on societies worldwide. In the 20th century, three novel influenza virus strains emerged, causing the 1918, 1957, and 1968 pandemics, the most devastating being the 1918 Spanish flu that led to an estimated 50 million deaths (47). The recent spread of highly pathogenic avian influenza (HPAI) H5N1 virus across parts of Asia, Europe, and the Middle East, with an overall fatality rate of over 60% for humans, as well as the rapid pandemic spread of a novel influenza A virus of the H1N1 subtype, has caused worldwide concern about a potential remake of the 1918 disaster (8).
Severe complications arising from pandemic influenza or HPAI H5N1 viruses are associated with rapid, massive inflammatory cell infiltration, resulting in acute respiratory distress, and reactive hemophagocytosis with multiple organ involvement. Both the 1918 Spanish influenza virus and HPAI H5N1 induce a cytokine storm characterized by an exaggerated production of inflammatory cytokines and chemokines in the serum and lungs caused by uncontrolled activation of the host's innate immune system. This triggers massive pulmonary edema, primary and/or secondary pneumonia, and alveolar hemorrhage with acute bronchopneumonia (4, 12, 24, 27, 37, 40, 43, 44).
The relationship between mortality, viral load, and immunopathology during influenza virus infection remains elusive and somewhat controversial. Some studies suggest that severe lung immunopathology is a direct consequence of a high viral load that the host is unable to resolve (12, 13), whereas others have reported that influenza virus-induced mortality is not a direct function of viral burden but a consequence of immune-mediated pathology (9, 11). Moreover, the picture is further complicated by the fact that different highly virulent influenza A viruses may induce distinct pathological signatures and lead to different courses of acute respiratory distress syndrome, refuting the hypothesis of a single, universal cytokine storm underlying all fatal influenza virus diseases (16).
Currently, vaccination remains the cornerstone of influenza virus prevention. However, due to constant antigenic drift and shift of the two major viral surface proteins hemagglutinin (HA) and neuraminidase (NA) (7), influenza virus vaccines must be reformulated each year in order to match the circulating subtypes (41). The potential emergence of an influenza virus pandemic at any time, combined with limited vaccine supplies, has rendered global vaccination strategies difficult. Therefore, a universal influenza virus vaccine that can provide protection against different variants or strains and thus not require frequent updates is highly desirable.
Here, we report that nasal administration of a recently developed live attenuated Bordetella pertussis vaccine strain, named BPZE1 (35), provides effective and sustained protection against lethal challenge with mouse-adapted H3N2 or H1N1 (A/PR/8/34) influenza A viruses. We demonstrate that the protective mechanism(s) does not target the viral particles or the infected host cells but controls the influenza virus-mediated inflammation by dampening the cytokine storm. As BPZE1 has recently entered phase I safety trials with humans (http://www.child-innovac.org), our observations support the potential application of this vaccine strain as a universal prophylactic treatment against highly pathogenic influenza A viruses.
B. pertussis BPZE1 is a streptomycin-resistant Tohama I derivative deleted of the dermonecrotic (DNT)-encoding gene, producing inactivated pertussis toxin (PT) and background levels of tracheal cytotoxin (TCT) (35). BPZE1 bacteria were grown at 37°C for 72 h on Bordet-Gengou (BG) agar (Difco, Detroit, MI) supplemented with 1% glycerol, 10% defibrinated sheep blood, and 100 μg/ml streptomycin (Sigma Chemical, St. Louis, MO). Liquid cultures were performed as described previously (33) with Stainer-Scholte (SS) medium containing 1 g/liter heptakis (2,6-di-o-methyl) β-cyclodextrin (Sigma). When appropriate, heat inactivation of the bacteria was performed at 95°C for 1 h.
Mouse-adapted A/Aichi/2/68 (H3N2) virus (passage 10) was obtained as described previously (36). H1N1 A/PR/8/34 virus was purchased from the ATCC (no. VR-95) and amplified in egg following ATCC's recommendations. Where indicated, heat inactivation of H3N2 virus was performed at 56°C for 30 min.
Mouse lungs were harvested and homogenized using mechanical disruption (Omni homogenizer) and tested for the presence of viable virus by a 50% tissue culture infectious dose (TCID50) assay using a modified method reported by the WHO (46). Briefly, 90% confluent Madin-Darby canine kidney (MDCK) cells in 96-well plates were inoculated with 100 μl of 10-fold serially diluted lung homogenates. Plates were incubated at 35°C in a humidified incubator (5% CO2) for 3 days. The TCID50 was determined by a reduction in cytopathic effect (CPE) of 50%, and the log TCID50/lung was derived. Five mice per group per time point were individually assessed.
MDCK cells (5 × 104) were seeded in 96-well flat-bottom plates and incubated at 37°C in a 5% CO2 atmosphere for 24 h. Twenty-five microliters of 2-fold serial dilutions of heat-treated (56°C, 30 min) sera or neat bronchoalveolar lavage fluids (BALFs) was mixed with equal volumes of 102 TCID50s of virus and incubated for 1 h at 37°C in 5% CO2. The antibody-virus mixtures were then transferred to a MDCK cell monolayer that had been washed twice with serum-free Dulbecco modified Eagle medium (DMEM) (Gibco) and incubated for 1 h at 37°C in 5% CO2. Three days later, the cells were observed for CPE, and the neutralizing antibody titer was read as the highest dilution of serum that inhibited virus growth and prevented CPE. Each dilution was assayed in six to eight replicates, and each neutralization assay was repeated twice independently.
All the animal experiments were carried out under the guidelines of the Institutional Animal Care and Use Committee, National University of Singapore. Six- to eight-week-old female BALB/c mice were kept under specific-pathogen-free conditions in individual ventilated cages. For BPZE1 treatment, sedated mice were intranasally (i.n.) administered once or twice (as indicated) with 5 × 105, 5 × 106, or 5 × 107 CFU (as indicated) of live or dead (heat-inactivated) BPZE1 bacteria in 20 μl sterile phosphate-buffered saline (PBS) supplemented with 0.05% Tween 80 (PBST) (Sigma) as previously described (20). For influenza virus infection, sedated mice were i.n. administered with 2 50% lethal doses (LD50) of mouse-adapted H3N2, and 2, 4, or 10 LD50 (as indicated) of H1N1 A/PR/8/34 in sterile PBS supplemented with penicillin and streptomycin. Ten mice per group were used to determine the survival rates based on body weight loss, and the mice were euthanized when body weight loss exceeded 20% of the original body weight.
Four adult female BALB/c mice were administered i.n. with 5 × 105, 5 × 106, or 5 × 107 CFU of live BPZE1. At the indicated time points, four animals per group and per time point were euthanized; their lungs were individually harvested and homogenized as described previously (20). Appropriate dilutions were plated for colony counting.
High-titer anti-B. pertussis immune sera were generated in 10 adult BALB/c mice nasally infected twice at a 4-week interval with live BPZE1 bacteria. Another group of 10 adult naïve BALB/c mice were injected intraperitoneally (i.p.) with 105.5 TCID50s of heat-inactivated human A/Aichi/2/68 (H3N2) virus (HI-H3N2) in complete Freund's adjuvant and boosted with the same amount of HI-H3N2 virus in incomplete Freund's adjuvant 2 weeks later. Immune serum samples from each mouse group were collected 2 weeks after the boost and pooled, and the anti-B. pertussis and anti-influenza virus antibody titers were measured by enzyme-linked immunosorbent assay (ELISA). In addition, HI-H3N2 serum was tested for the presence of neutralizing antibodies by neutralization assay. The immune serum samples were filter sterilized, heat treated at 56°C for 30 min, and stored at −80°C until further use. Serum samples from naïve control mice were also collected.
Six- to 8-week-old recipient BALB/c mice were i.p. injected with 200 μl of naïve, anti-BPZE1 or anti-H3N2 immune serum 1 day prior to lethal challenge with mouse-adapted H3N2 virus. Body weight losses were monitored to determine the survival rates. Ten mice per group were assayed.
The mouse lungs were harvested and fixed in 10% formalin in PBS, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E). Observations were made using an inverted light microscope at 10× and 40× objectives.
Individual BALF samples were recovered by injecting 1 ml of sterile PBS into the lungs of sacrificed animals and performing one lavage step. BALFs were then centrifuged, and the supernatant was removed and stored at −80°C for cytokine detection. Cells were resuspended, spotted onto a glass slide using a Cytospin device (Thermo Shandon), and stained using a modified Wright staining procedure (3). Results were expressed as the percentage of each cell type in the total cell population. A total of 500 cells were considered per slide. Four mice per group were individually assessed.
Single-cell suspensions were prepared by digesting the mouse lungs at 37°C for 15 min in 2 ml digestion buffer containing 0.5 mg/ml Liberase (Roche) in RPMI with 1% fetal calf serum (FCS) and 2 U/ml DNase I (Qiagen) and centrifuging them on Ficoll-Paque Plus (GE) for 20 min at 600 × g at room temperature. Cells were collected and washed twice with sterile fluorescence-activated cell sorter (FACS) buffer (2% FCS, 5 mM EDTA in PBS). Cells (106) were stained with fluorescein isothiocyanate (FITC)-labeled anti-mouse CD3 antibody (eBioscience) and analyzed on a CyAn ADP cytometer (Dako). Five mice per group per time point were individually assessed.
Cytokine and chemokine levels in the BALF supernatants were measured using a multiplex cytokine detection kit (Bioplex; Bio-Rad) according to the manufacturer's instructions. The samples were analyzed using a Bio-Plex instrument (Bio-Rad). Granulocyte-macrophage colony-stimulating factor (GM-CSF), KC (interleukin-8 [IL-8]), IL-1β, IL-6, IL-12p70, gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), monocyte chemoattractant protein 1 (MCP-1), and IL-10 were assayed. In addition, transforming growth factor β (TGF-β) levels were measured using a human/mouse TGF-β1 ELISA kit (eBioscience) according to the manufacturer's instructions.
The presence of antibodies in the serum and BALF samples were measured by ELISA. Ninety-six-well microtiter plates (Costar; Corning) were coated overnight at 4°C with 100 μl of 0.1 M carbonate buffer (pH 9.6) containing 5 μg/ml of heat-inactivated H3N2 viral particles. After blocking with 2% bovine serum albumin (BSA) in PBS containing 0.1% Tween 20, 100 μl of serum diluted at 1:40 (anti-BPZE serum) or 1:1,000 (anti-H3N2 serum) or 50 μl of undiluted BALFs was added to the wells. The plates were incubated at 37°C for 1 h, rinsed in PBS-0.1% Tween 20, and incubated at 37°C for 1 h with 50 μl of horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (H + L) or anti-mouse IgA secondary antibodies (Sigma) diluted 1:3,000 or 1:2,000, respectively. The reaction was then developed using o-phenylenediamine dihydrochloride substrate (Sigma) at room temperature for 30 min in the dark and stopped by the addition of 1 M sulfuric acid. The absorbance at 490 nm was measured by an ELISA plate reader (Tecan Sunrise).
Lymphocyte proliferation was measured by incorporation of tritiated [3H]thymidine as described previously (25). Briefly, spleens from naïve, HI-H3N2- and BPZE1-immunized mice (6 mice per group) were pooled, and single-cell suspensions were prepared and centrifuged on Ficoll-Paque Plus (GE) for 20 min at 600 × g at room temperature. The splenocytes were seeded in 96-well round-bottom plates (Nunc) at a density of 2 × 105 cells/well in 100 μl medium (RPMI 640 supplemented with 10% FCS, 5 × 10−5 M β-mercaptoethanol, 2 mM l-glutamine, 10 mM HEPES, 200 U/ml penicillin, 200 μg/ml streptomycin). Medium (100 μl) containing 20 μg/ml of BPZE1 whole-cell lysate or 105 TCID50s of HI-H3N2 (test antigen) was added to the splenocytes. Noninfected egg amniotic fluid (100 μl) and medium containing 5 μg/ml concanavalin A (conA) (100 μl) were used as mock and vitality controls, respectively. After 3 days of incubation at 37°C in a 5% CO2 atmosphere, the cells were pulsed with 0.4 μCi [3H]thymidine in 20 μl RPMI complete medium. After incubation for 18 h, cells were harvested and washed and the incorporated radioactivity was measured in a TopCount NXT microplate scintillation and luminescence counter (PerkinElmer). Results are expressed as the stimulation index (SI) corresponding to the ratio between the mean of [3H]thymidine uptake in the presence of test antigen and the mean of [3H]thymidine uptake in the absence of test antigen. An SI value of >2 was considered positive. Each sample was assayed in quadruplicate.
The frequency of antigen-specific IFN-γ-producing splenocytes was determined by an enzyme-linked immunospot (ELISPOT) assay using a mouse ELISPOT set (BD PharMingen) according to the manufacturer's instructions. Briefly, single-cell suspensions of individual spleens from naïve and BPZE1-treated mice were plated in 96-well microplates (Millipore, Bedford, MA) precoated with 100 μl of 5 μg/ml anti-IFN-γ antibody in sterile PBS overnight at 4°C, washed three times, and blocked for 2 h at room temperature with RPMI 1640 containing 10% FCS. Cells were then incubated with 20 μg/ml of BPZE1 whole-cell lysate, with 105 TCID50s HI-H3N2, or with 5 μg/ml conA for 12 to 20 h at 37°C in a 5% CO2 atmosphere. The plates were then washed, and biotin-conjugated anti-mouse IFN-γ antibody was added for 2 h at room temperature. After the wells were washed, streptavidin-HRP conjugate was added and incubated at room temperature for 1 h. The wells were washed again and developed with a 3-amino-9-ethyl-carbazole (AEC) substrate solution until spots were visible. After drying, spot-forming cell numbers were counted by Bioreader 4000 (Biosystem). Six animals per group were individually assayed.
Unless otherwise stated, in the figures, bars represent means ± standard deviations (SD) and averages were compared using a bidirectional unpaired Student t test with a 5% significance level (*, P ≤ 0.05; **, P ≤ 0.01; and ***, P ≤ 0.001).
Adult BALB/c mice were nasally inoculated once with 5 × 106 CFU of live B. pertussis BPZE1 and challenged 3 or 6 weeks later with 2 LD50 of mouse-adapted H3N2 virus, obtained after 10 successive lung-to-lung passages of the A/Aichi/2/68 (H3N2) virus (36). Survival rates (Fig. (Fig.1A)1A) and body weight changes (Fig. (Fig.1B)1B) indicated that the mice challenged 6 weeks after nasal BPZE1 treatment were significantly protected (60% protection rate), whereas mice challenged at 3 weeks post-BPZE1 treatment were not protected. A protection rate of 60% was still observed when the viral challenge was performed 12 weeks post-BPZE1 treatment (Fig. (Fig.1C).1C). In contrast, nasal administration of the same amount of heat-killed bacteria did not provide any significant protection against H3N2 lethal challenge (Fig. (Fig.1D1D).
To further explore the protective efficacy of live BPZE1 bacteria, 10-fold serially diluted BPZE1 suspensions were administered nasally to the mice prior to lethal challenge with H3N2 virus performed 6 weeks later. The mouse group which received 5 × 105 CFU of BPZE1 was not significantly protected, whereas mice administered with 5 × 106 CFU and 5 × 107 CFU were significantly protected, with 50% and 62.5% survival rates, respectively (Fig. (Fig.2A).2A). Interestingly, the animals immunized with the highest bacterial dose (5 × 107 CFU) displayed significantly smaller body weight losses than the animal group administered with 5 × 106 CFU (Fig. (Fig.2B).2B). Concurrently, the lung colonization profiles of BPZE1 were determined for each bacterial dose and showed that the bacterial load present in the lungs during the colonization period was directly dependent on the initial bacterial dose administered (Fig. (Fig.2C2C).
Taken together, these results indicate that the protective efficacy of BPZE1 against lethal challenge with H3N2 virus depends on the bacterial dose administered and that the ability of the bacteria to effectively colonize the mouse respiratory tract is necessary to prime the protective mechanisms.
The effect of a second nasal BPZE1 treatment was addressed. Live BPZE1 bacteria were nasally administered twice at a 4-week interval prior to lethal challenge with mouse-adapted H3N2 virus performed 2 weeks after the last BPZE1 treatment. A survival rate of 100% was obtained (Fig. (Fig.3A),3A), with minimal body weight changes, in contrast to the infected control mice (Fig. (Fig.3B).3B). These data demonstrate that a second nasal administration of BPZE1 not only enhances the protective efficacy but also shortens the time necessary to trigger the protective mechanism(s).
The protective potential of BPZE1 against a different influenza A virus subtype was explored. Whereas a single nasal administration of 5 × 106 CFU of live BPZE1 bacteria did not confer any significant protection against lethal challenge with H1N1 A/PR/8/34 virus (data not shown), three consecutive administrations of live BPZE1 provided about 50% protection (Fig. (Fig.4A).4A). These observations indicate that BPZE1 is able to protect against different influenza A virus subtypes, although with variable efficacy. In this experiment, an H1N1 dose of 4 LD50 was administered to the mice, whereas H3N2 challenge was performed with a dose of 2 LD50. One could thus argue that the differential protective efficacy of BPZE1 against the two virus subtypes may be due to the different lethal doses used for the challenge. However, we found that the survival rates, times of death, and body weight losses of naïve mice infected with either 2 LD50 or 10 LD50 of H1N1 A/PR8/34 virus were comparable (Fig. 4B&C), indicating that a higher initial viral dose does not worsen the disease severity, based on survival rate, time of death, and body weight loss. Therefore, the differential ability of nasal BPZE1 treatment to protect against H1N1 or H3N2 subtypes is likely not attributable to the different lethal doses used for the viral challenge but rather suggests differences in the virulence mechanisms developed by both virus subtypes, as recently reported (16).
Lung immunopathology was examined by histology of lung sections from infected BPZE1-treated and nontreated animals. As expected and as described previously (36), the infected nontreated mice displayed signs of severe bronchopneumonia and interstitial pneumonitis, characterized by necrosis of the bronchiolar epithelium and the presence of necrotic debris in the bronchioles and alveoli, as well as significant pulmonary emphysema and moderate edema (Fig. (Fig.5A).5A). In contrast, only mild inflammation, minimal airway and alveolar damage, and mild perivascular or peribronchiolar damage, together with minimal edema, were observed with the lungs of the protected BPZE1-treated mice.
The cell populations present in the BALF samples recovered from protected and nonprotected animals were also examined. Whereas the total numbers of cells present in the BALF specimens from both animal groups were comparable (11.8 × 105 versus 16.1 × 105), a significantly higher number of macrophages but a lower number of neutrophils were found in the infected BPZE1-treated mice than in the nontreated mice 3 days after viral challenge (Fig. (Fig.5B5B).
Lymphocyte depletion has been reported for mice infected with highly pathogenic H1N1 (1918) and H5N1 influenza viruses (24, 29, 44), as well as for mice infected with the mouse-adapted H3N2 virus strain used in this study (36). The lymphocyte populations present in the lungs of protected and nonprotected mice were thus compared. At 3 days post-viral challenge, the percentages of CD3+ cells in the infected control and BPZE1-treated mice were comparable to those found in the animals before challenge (Fig. (Fig.5C).5C). However, significant CD3+ T-cell depletion was observed with the infected control animals at 5 days post-viral challenge. In contrast, the T-cell population remained constant before and after challenge in the protected BPZE1-treated animals, indicating that nasal treatment with live BPZE1 prevented influenza virus-induced lymphocyte depletion.
The presence of cross-reactive and virus-neutralizing antibodies and T cells in the BPZE1-treated animals was examined. A BLAST search failed to identify any matching epitopes between B. pertussis and H3N2 or H1N1 influenza A viruses (data not shown).
Moreover, the serum and BALF samples from BPZE1-treated mice did not react with whole H3N2 viral antigens in an ELISA (Fig. 6A and B) and did not neutralize virus infectivity in vitro (Fig. (Fig.6C).6C). Consistently, high-titer anti-BPZE1 immune sera did not confer any significant protection against H3N2 lethal challenge in an in vivo passive transfer experiment, whereas immune sera raised against heat-inactivated H3N2 virus provided 100% protection (Fig. (Fig.7A).7A). Finally, proliferation and IFN-γ ELISPOT assays revealed that splenocytes from BPZE1-treated mice did not proliferate and did not produce IFN-γ upon stimulation with H3N2 viral particles, whereas strong stimulation and IFN-γ production were seen upon stimulation with BPZE1 extracts (Fig. 7B and C). All together, these data indicated that B. pertussis-specific adaptive immunity is unlikely to mediate the cross-protection against influenza A viruses.
To further characterize BPZE1-induced protection against influenza A viruses, the viral loads in the lungs of mice either untreated or treated with BPZE1 were quantified and lethally challenged with mouse-adapted H3N2 virus. No significant differences in the viral loads were observed between the two groups at 3 and 5 days postchallenge (Fig. (Fig.7D).7D). This result indicates that the BPZE1-induced protective mechanism does not directly target the virus particles and/or infected host cells and demonstrates that effective protection against influenza A virus-induced pathology and mortality can be achieved without affecting the virus titer in the lungs.
The levels of 8 major proinflammatory cytokines and chemokines as well as TGF-β and IL-10 in the BALF samples from BPZE1-treated and untreated mice were measured before and 3 days after lethal challenge with H3N2. A significant upregulation of 6 out of the 8 proinflammatory cytokines/chemokines was observed with the untreated infected mice compared to that in the levels measured for noninfected animals (Fig. (Fig.8A).8A). IFN-γ and IL-12p70 production was unchanged in the infected mice compared to that in the control group. Interestingly, the level of TGF-β was markedly reduced in the infected mice, whereas IL-10 levels were higher than those in the controls (Fig. (Fig.8A).8A). In mice treated once with BPZE1, the levels of 3 out of 8 proinflammatory cytokines and chemokines, namely, IL-1β, IL-6, and GM-CSF, were significantly lower than those measured in the untreated infected controls (Fig. (Fig.8B).8B). In mice treated twice with BPZE1, the production of all 8 proinflammatory cytokines and chemokines was either strongly reduced or completely suppressed compared to that in the nontreated infected control group (Fig. (Fig.8B).8B). Thus, nasal treatment with live BPZE1 prior to viral challenge suppresses the influenza virus-induced cytokine storm, strongly suggesting a correlation between protection and reduced lung inflammation.
Interestingly, no significant difference in the levels of TGF-β between BPZE1-treated mice and the infected controls (Fig. (Fig.8B)8B) was detected, whereas IL-10 production was reduced in the BPZE1-treated animals.
Severe respiratory disease and immunopathology resulting in high case fatality rates are hallmarks of highly pathogenic influenza virus infections in humans as well as in other mammals. The cytokine storm, characterized by excessive levels of chemokines and cytokines in the serum and lungs, has been linked to fatal outcome in humans (4, 12, 37, 44) and in experimental animals infected with reconstructed 1918 H1N1 and H5N1 influenza viruses (24, 27, 40, 43). Furthermore, histological and pathological indicators strongly suggest a key role for an excessive host response in mediating at least some of the extreme pathology associated with highly pathogenic influenza viruses. Thus, although cytokine production may be important for viral clearance, cytokine inflammatory properties may also lead to tissue damage (28). In this study, we found that administration of the live attenuated B. pertussis BPZE1 strain protects mice against lethal challenge with influenza A virus by strongly decreasing lung immunopathology and reducing the production of major proinflammatory cytokines and chemokines, without affecting the viral load. These findings thus indicate a direct link between protective efficacy and reduction of the cytokine-mediated inflammation and strongly argue for an important role of the overproduction of proinflammatory cytokines in influenza virus disease severity and mortality.
These conclusions are in line with those of a recent study in which the intratracheal administration of the sphingosine analog AAL-R significantly reduced lung immunopathology caused by influenza virus by decreasing the release of cytokines and chemokines known to contribute to the cytokine storm effect, while no change in lung viral titers was observed (31). However, other studies have suggested that the severe immunopathology observed with infected animals is a consequence of the inability to resolve a high viral load in the respiratory tract (12, 13). Moreover, reduced inflammatory cell infiltration and pulmonary damage but delayed viral clearance were observed with macrophage inflammatory protein 1α (MIP-1α) gene knockout mice (9). Likewise, CCR2 (primary receptor for MCP-1)-deficient mice displayed reduced mortality, with decreased pulmonary cell infiltration and tissue damage, but with a significantly elevated viral burden compared to control mice (11). The relationship between viral load, immunopathology, and disease severity appears rather complex and involves a variety of host factors which play different roles during influenza virus infection and whose disruption or inactivation leads to different outcomes.
The reduced production of proinflammatory cytokines and chemokines in the respiratory tract of the protected BPZE1-treated animals likely impacted on cellular infiltration and immune cell activation. Lower neutrophil counts were observed with BALF samples from the protected animals, consistent with lower levels of KC and TNF-α, two cytokines that are involved in the recruitment and activation of neutrophils in the infected tissues (17, 26, 28). Neutrophils can elicit numerous responses in the presence of IFN-γ, including increased oxidative burst and the induction of antigen presentation, and chemokine production (14, 15). Therefore, although increased numbers of neutrophils in the lungs contribute to the inhibition of virus replication as part of the host innate immune response, they may also play a role in the enhanced immunopathology induced upon infection with highly pathogenic influenza viruses.
Likewise, the complete suppression of IL-12 and GM-CSF observed with the protected BPZE1-treated mice likely impaired activation, differentiation, and recruitment of various immune cells, including macrophages, dendritic cells (DC), cytotoxic CD8+ T cells, and natural killer (NK) cells whose activation may be involved in immunopathology upon release of inflammatory mediators (28).
Interestingly, a higher number of macrophages were observed with the BALF samples from protected BPZE1-treated mice. Alveolar macrophages (AMs) constitute the predominant macrophage population recovered in BALFs (23) and display suppressive effects on the inflammatory reaction by regulating T-cell function (42) and by suppressing dendritic cell maturation (5, 21) and migration to the mesenteric lymph nodes (23). Whether the increased numbers of AMs present in the BPZE1-treated animals upon influenza virus challenge contributes to the control of lung inflammation warrants further investigation.
In addition, treatment with BPZE1 prevented influenza virus-induced CD3+ T-cell depletion. Lymphocyte depletion during highly pathogenic influenza virus infection has previously been reported (24, 29, 36, 44), and apoptosis has been proposed as a potential mechanism (19, 29, 44). Since we observed no difference in viral loads between protected and nonprotected animals, virus-induced lymphocyte apoptosis does not likely result from a direct cytotoxic effect of the virus itself. Instead, consistent with previous studies of H5N1-infected humans and mice (30, 44), lymphocyte apoptosis may be attributed to cytokine dysregulation and overactivation of the host immune response. In particular, TNF-α and related TNF-superfamily members, including TNF-related apoptosis-inducing ligand (TRAIL), are known to induce T-cell apoptosis (40, 45). Consistently, lower levels of TNF-α were measured in the BALF samples from protected BPZE1-treated mice upon influenza virus challenge, thus possibly translating into diminished T-cell apoptosis.
Nonspecific protection against influenza viruses has previously been reported. Nasal administration of Autographa californica nuclear polyhedrosis baculovirus (AcNPV) protected mice from lethal H1N1 A/PR/8/34 influenza virus challenge (1). Reduced immunopathology, together with lower levels of IL-6 production but also with reduced viral loads, was observed with the protected mice compared to the control mice. In another report, prophylactic intranasal treatment with chitin microparticles enhanced the local accumulation of NK cells and suppressed hyperinduction of cytokines, resulting in protection against infection with HPAI virus (22). However, in both studies, the prophylactic treatment needed to be performed less than a few days prior to the virus challenge in order to be protective, resulting in a transient, short-term downregulation of the host inflammatory response. In contrast, BPZE1-induced protection takes more than 3 weeks to be effective and lasts up to at least 12 weeks posttreatment. The protective mechanism(s) primed upon nasal treatment with BPZE1 thus appears to be rather unique and likely does not involve fast-induced short-lived innate immune cells such as NK, DC, or macrophages.
The lack of T- and B-cell cross-reactivity between B. pertussis and influenza A virus, as well as the inability to confer protection by passive transfer of antiserum from BPZE1-treated mice, further demonstrates that B. pertussis-specific adaptive immunity is not involved. In addition, the observation that live but not heat-killed BPZE1 bacteria induce protection indicates that bacterial lung colonization, i.e., a prolonged exposure to the host immune system, is necessary. This hypothesis is further supported by the fact that the protection rate is bacterial dose dependent, which directly correlates with lung colonization efficacy. Moreover, we showed that a second BPZE1 treatment shortened the time necessary to induce protection and enhanced the protection rate, implying that some memory has been triggered upon first exposure to BPZE1 and can be boosted by a second encounter with BPZE1 bacteria.
Many activities of B. pertussis virulence factors are dedicated to immunomodulation in order to suppress, subvert, and evade the host defense system (6). The immune response to B. pertussis is initiated and controlled through Toll-like receptor 4 (TLR-4) signaling, inducing the production of the anti-inflammatory cytokine IL-10 to inhibit inflammatory responses and limit pathology in the airways (18). Filamentous hemagglutinin (FHA), the major B. pertussis adhesin, can stimulate IL-10 production and inhibit TLR-induced IL-12 production, resulting in the development of IL-10-secreting Tr1 cells (32). However, we found no difference in the production levels of TGF-β and reduced levels of IL-10 in the BPZE1-treated mice compared to those in the infected control animals, thus excluding the potential involvement of FHA-mediated induction of Tr1 cells as a mechanism of BPZE1-mediated protection against influenza A virus. The involvement of other types of regulatory mechanisms is thus likely at play and is being investigated.
In conclusion, our study demonstrates the anti-inflammatory properties of the B. pertussis BPZE1 strain and supports the potential use of this bacterial agent as a novel, highly effective prophylactic agent, with long-lasting effects, against severe and lethal pneumonitis induced by H3N2 and H1N1 influenza A viruses. In addition, attenuated B. pertussis is particularly well adapted for the nasal delivery of heterologous vaccine antigens and represents an attractive mucosal vaccine delivery system (2, 10, 20, 34, 38, 39). Constructing BPZE1 derivatives producing viral antigens, thereby combining anti-inflammatory properties with the ability to induce adaptive anti-influenza virus immune responses, may thus constitute an interesting strategy for future development.
We gratefully thank K. Tan (Department of Microbiology, National University of Singapore) for his critical reading of and useful comments on the manuscript.
This work was supported by the National Medical Research Council (Individual Research Grants no. NMRC/0962/2005 and NMRC/1135/2007) and the Department of Microbiology and Immunology Programme, National University of Singapore (start-up grant no. R-182-000-122-731).
We declare that we have no conflicts of interest.
Published ahead of print on 5 May 2010.