|Home | About | Journals | Submit | Contact Us | Français|
Despite the availability of efficacious vaccines, the incidence of whooping cough is still high in many countries and is even increasing in countries with high vaccine coverage. Most severe and life-threatening pertussis cases occur in infants who are too young to be sufficiently protected by current vaccine regimens. As a potential solution to this problem, we have developed an attenuated live Bordetella pertussis vaccine strain, named BPZE1. Here, we show that after a single administration, BPZE1 induces dose-dependent protection against challenge with virulent B. pertussis in low-dose and in high-dose intranasal mouse lung colonization models. In addition, we observed BPZE1 dose-dependent antibody titers to B. pertussis antigens, as well as cell-mediated immunity, evidenced by the amounts of gamma interferon (IFN-γ) released from spleen cells upon stimulation with B. pertussis antigens. These two parameters may perhaps be used as readouts in clinical trials in humans that are currently being planned.
Pertussis, or whooping cough, caused by the exclusively human pathogen Bordetella pertussis, is a highly contagious acute disease of the respiratory tract and is responsible for approximately 300,000 deaths in children worldwide every year. Despite the availability and intensive use of efficacious vaccines for several decades, pertussis has not been eliminated in any country. In fact, the incidence of the disease is increasing in many countries with high vaccine coverage, and whooping cough remains globally among the top 10 causes of childhood death (4). Although by far most pertussis-linked deaths occur in young infants, adolescent and adult pertussis is also a growing and often underestimated problem in countries with high vaccine coverage (33). In contrast to infant infections, many B. pertussis infections in adolescents and adults are mild or subclinical and are usually not life-threatening (13). However, infected adults and adolescents are now considered an important reservoir for the whooping cough agent, able to transmit the infection to infants before they are sufficiently protected by vaccination. This epidemiological change is likely to be due to progressive waning of vaccine-mediated immunity during adolescence. However, other epidemiological features may also potentially contribute to the increasing pertussis burden in areas of high vaccine coverage, such as adaptation of B. pertussis strains in response to vaccine-induced immunity (3, 9, 10, 26).
Several strategies to solve this problem can be envisaged. As one of the potential solutions, it has been proposed to provide regular booster doses to adolescents and adults (7). However, repeated administrations of current pertussis vaccines are sometimes associated with local adverse effects, such as large swellings that may involve the entire limb (15). In addition, compliance of adolescents and adults to receiving booster doses is usually low for any vaccine (32). Maternal immunization has also been described as a potential approach to protect newborns (for a recent review, see reference 25). Nevertheless, the vaccination schedule will have to be carefully defined, both safety and efficacy of such a strategy still need to be assessed in clinical trials, and the acceptance among mothers may constitute an important hurdle. As an alternative, infant vaccination occurring as early as possible, preferably at birth, has been proposed in order to protect children during their most vulnerable period (2, 28). However, early protection by vaccination is hampered by the relative immaturity of the neonatal and infant immune system, especially of the cell-mediated immune arm (31), known to be important for protection against B. pertussis (23). In addition, optimal protection requires at least three doses of the current vaccines (6), usually given at 1- or 2-months intervals. Therefore, acceptable protection would not be achieved before 3 to 4 months, even if vaccination was started at birth.
In contrast to vaccination, infection with B. pertussis is able to quickly induce a strong Th1-type immune response in very young children, characterized by the production of high levels of B. pertussis antigen-specific gamma interferon (IFN-γ) (18). Furthermore, studies of nonhuman primates have led to the conclusion that “ultimate protection against whooping cough probably best follows a live B. pertussis inoculation” (14). This has prompted us to construct an attenuated B. pertussis strain as a live vaccine candidate by genetically altering or removing three B. pertussis toxins, pertussis toxin (PTX), tracheal cytotoxin (TCT), and dermonecrotic toxin (DNT). Briefly, this strain, named BPZE1, expresses an enzymatically inactive PTX by altering two key amino acids for the enzymatic activity of the toxin, shows a 100-fold reduction in TCT activity by the replacement of the B. pertussis ampG gene with that of Escherichia coli, and does not produce DNT by the deletion of its structural gene. We showed that BPZE1 still colonizes the mouse respiratory tract and is able to provide protection against B. pertussis challenge after a single nasal administration in a mouse model (22).
In this study, we investigated the dose response of a single nasal administration of BPZE1 in mice to identify the protective doses needed against challenge infection with virulent B. pertussis.
Virulent B. pertussis BPSM, a streptomycin-resistant Tohama I derivative (20), and the attenuated vaccine strain BPZE1 (22) were grown on Bordet-Gengou (BG) agar (Difco, Detroit, MI) supplemented with 1% glycerol, 20% defibrinated sheep blood, and 100 μg/ml streptomycin (Sigma, St. Louis, MO). The bacteria were then harvested and suspended in sterile phosphate-buffered saline (PBS).
Groups of 8-week-old or 3-week-old (for the weight gain testing) female BALB/c mice (Iffa Credo, L'Arbresle, France) were kept under specific-pathogen-free conditions, and all experiments were carried out under the guidelines of the Institut Pasteur de Lille animal study board. Mice were intranasally infected with the indicated quantities of bacteria in 20 μl PBS, and the lung colonization was evaluated as previously described (21). For challenge infection, mice were intranasally infected 2 months after vaccination with the indicated doses of BPSM in 20 μl PBS. Lung colonization was determined 3 h and 7 days later. For intraperitoneal vaccination with acellular pertussis vaccine, one-fifth of a human dose of Tetravac (Aventis Pasteur, France) vaccine containing filamentous hemagglutinin (FHA) and pertussis toxoid was used as described previously (22). The limit of detection for bacterial colonization was 10 viable bacteria per individual lung.
Human pulmonary epithelial (A549; ATCC CCL-185) and monocytic (THP-1; ATCC TIB 202) cell lines were cultured at 37°C in 5% CO2 in K-12 and RPMI 1640 media, respectively, supplemented with sodium penicillin G (1,000 U/ml) and streptomycin (50 μg/ml) (Gibco), 2 mM l-glutamine (Gibco), and 10% heat-inactivated fetal calf serum (FCS; Gibco). Before infection, cells were scrapped, washed with PBS to remove antibiotics, and plated at the required concentration in 1% FCS medium in 24-well plates. Cells were infected with 100 μl of bacterial suspension containing BPZE1 or Escherichia coli K-12 (multiplicity of infection [MOI], 0, 1, 10, or 100) for 5 h at 37°C in 5% CO2 and then extensively washed with PBS to remove extracellular bacteria. Infected cells were reincubated with fresh complete medium supplemented with 200 μg/ml of amikacin (Merck, France) to kill remaining extracellular bacteria for different time periods. The percentage of viable cells was evaluated by trypan blue exclusion.
Mice were sacrificed, and sera and bronchoalveolar lavage fluids were collected for assessment of antibody responses by enzyme-linked immunosorbent assays (ELISAs) as previously described (21). Briefly, 96-well plates (MaxiSorp; Nunc) were coated overnight at 4°C with 50 μl of FHA (10 μg/ml) purified from PTX-deficient B. pertussis BPRA (1), as previously described (19), or with 10 μg/ml PTX purified from FHA-deficient B. pertussis BPGR4 (17), as previously described (29). Samples were then added in 2-fold serial dilutions and incubated overnight at 4°C. Goat anti-mouse IgG(H+L) or IgA horseradish peroxidase-conjugated antibodies (Southern Biotechnologies Associates, Inc., Birmingham, AL) were then added for 2 h at 37°C. The ELISAs were developed using tetramethylbenzidine and hydrogen peroxide (Interchim, Montluçon, France), according to the manufacturer's specifications. The results are expressed in titers, defined as the reciprocal of the dilution giving an optical density at 492 nm of three times that of the conjugate control.
Spleen cells from individual mice were harvested 6 weeks after BPZE1 administration and stimulated at 107 cells/ml with heat-killed B. pertussis BPSM (106 cells/ml) and 5.0 μg/ml of heat-inactivated PTX or FHA. Supernatants were harvested from triplicate cultures after 72 h of incubation at 37°C in the presence of 5% CO2, and IFN-γ concentrations were determined by immunoassays (BD OptEIA set; Pharmingen, San Diego, CA).
The results were analyzed using the unpaired Student t test and the Kruskal-Wallis test, followed by Dunn's posttest (GraphPad Prism program) when appropriate. Differences were considered significant at P values of ≤0.05.
In vivo and in vitro tests were used to assess the safety of BPZE1. Infant (3-week-old) mice were inoculated with 1 × 106 BPZE1, and weight changes were monitored over a period of 2 weeks. No clinical symptoms were observed following BPZE1 administration in infant mice, and their weight gain was similar to that of nonimmunized mice (Fig. (Fig.1a).1a). Histopathological analysis of the lungs of mice at 5 days after intranasal administration of BPZE1 showed minimal cell infiltration, similar to that observed in lungs from mice inoculated with sterile PBS (Fig. (Fig.1b).1b). In contrast, nasal infection with the virulent B. pertussis BPSM strain induced strong peribronchiovascular infiltrates and recruitment of inflammatory cells in the lungs of infected mice.
In vitro assays were used to measure the potential cell toxicity of BPZE1. The kinetics of the viability of human pulmonary epithelial A549 cells and monocytic THP-1 cells were measured following infection with different MOIs of live B. pertussis BPZE1 or live E. coli K-12 as a control. As shown in Fig. Fig.1c,1c, BPZE1 expresses no toxic effect on pulmonary epithelial A549 cells at any MOI tested. A slight reduction of the monocyte THP-1 viability was observed 6 days after interaction with BPZE1 after incubation at the highest bacterial load (MOI ≥ 10). However, this reduction was similar to that observed after infection with nonpathogenic E. coli, indicating that BPZE1 is not more toxic than nonpathogenic E. coli.
We have previously shown that BPZE1 is able to colonize and to persist in the lungs of mice as long as virulent B. pertussis (22). Bacterial persistence may be important for prolonged immune stimulation and thus for induction of efficient long-term immunity. Therefore, we first determined the colonization profile of different doses of BPZE1 administered nasally. Groups of BALB/c mice were intranasally inoculated with 10-fold increasing doses of BPZE1. Initial inoculum sizes reaching mouse lungs were determined 3 h after nasal administration of BPZE1 and were shown to rank from 5 × 102 to 1 × 106 CFU. Bacterial loads in the lungs were subsequently assessed every week for each group of mice. As shown in Fig. Fig.22 and consistent with our previous data (22), inoculation with the highest dose of BPZE1 (106 CFU) resulted in persistent colonization of up to 3 to 4 weeks. Persistence was not strongly affected by decreasing the inoculum size down to 103 CFU, although the bacterial loads at each analyzed time point depended on the initial inoculum size, showing a good dose-dependent colonization profile. However, when the inoculum size was further decreased down to 5 × 102 CFU, the bacteria did not persist, and no evidence for colonization was observed. These results indicate that the minimal nasal dose of BPZE1 to establish lung colonization in BALB/c mice was 103 CFU.
Both antibody and T-cell responses to B. pertussis antigens, such as PTX and FHA, contribute to immunity against pertussis (23). The kinetics of serum antibody responses were first measured after immunization of mice with the highest dose (1 × 106 CFU) of BPZE1. As shown in Fig. Fig.3a,3a, the serum IgG response to heat-killed B. pertussis extracts was already detectable 14 days after nasal immunization and increased during the first month. Sera were then collected from the different mice inoculated with the various doses of BPZE1 at 2 months after BPZE1 administration, and antibody titers were estimated by ELISA. The results showed that administration of BPZE1 induced serum IgG responses against total heat-killed B. pertussis extract and against both PTX and FHA and that the antibody titers increased proportionally to the dose of bacteria administered (Fig. 3b to d). No serum IgA response against whole-cell antigens was detected even after administration of the highest dose (106 CFU) of BPZE1 (data not shown). Since BPZE1 is administered nasally, we analyzed local antibody responses to B. pertussis extracts after administration of 1 × 106 CFU of BPZE1. Significant levels of IgA and IgG responses against B. pertussis whole-cell antigens were measured in bronchoalveolar lavage fluid samples from BPZE1-vaccinated mice (Fig. (Fig.4).4). In comparison, acellular pertussis vaccine, which is administered systemically, induced IgG, but there were no IgA responses in the respiratory tract.
In parallel, the IFN-γ production in response to stimulation with the B. pertussis antigens was assessed as a readout to measure T-cell responses. Spleen cells from individual mice were therefore harvested 6 weeks after BPZE1 administration and stimulated with heat-killed B. pertussis BPSM, PTX, or FHA, and the IFN-γ concentrations in the culture supernatants were measured. Consistent with a previous study (22), we found that by using the highest bacterial dose (1 × 106 CFU), BPZE1 immunization induced a Th1-type immune response, with high levels of IFN-γ produced by stimulated spleen cells. This IFN-γ response was significantly higher than that induced by stimulated spleen cells from mice immunized with acellular pertussis vaccine (Fig. (Fig.5a).5a). Next, we showed that increasing BPZE1 inoculum sizes resulted in increasing IFN-γ responses to all three antigenic stimuli (Fig. 5b and c). A linear dose response was observed for IFN-γ production upon stimulation with PTX and with FHA, starting with significant IFN-γ production after administration of 104 CFU (Fig. (Fig.5b),5b), whereas the threshold of 104 CFU resulted in a nearly maximal IFN-γ response to total heat-killed B. pertussis cell extract. Increasing the BPZE1 doses above 104 CFU did not result in a further increase in the level of total heat-killed B. pertussis cell extract-induced IFN-γ (Fig. (Fig.5c5c).
To evaluate the protective immunity induced by nasal vaccination with the different doses of BPZE1, the mice were challenged 2 months after administration of BPZE1 by intranasal infection with 5 × 104 CFU of virulent B. pertussis BPSM, and the colony counts in the lungs were determined 3 h and 7 days later. In the control unvaccinated mice, the colony counts increased by approximately 10-fold between 3 h and 7 days postchallenge. In contrast, mice immunized with a single nasal dose of BPZE1 showed strong protection, as evidenced by the low to undetectable CFU counts 7 days after challenge infection (Fig. (Fig.6a).6a). With a vaccine dose as low as 104 CFU of BPZE1, full protection was observed, as evidenced by the total bacterial clearance at day 7 after challenge. Vaccination with 103 CFU of BPZE1 resulted in an approximately 100,000-fold CFU reduction at day 7 after challenge, compared to that of the nonvaccinated mice.
When the mice were challenged with a 100-fold-higher dose of virulent B. pertussis BPSM, a correlation was found between the vaccine dose and the protective effect (r = 0.99). Although total bacterial clearance was not observed at day 7 after challenge for any vaccine group, vaccination with 106 CFU of BPZE1 resulted in an approximately 1,000,000-fold reduction in bacterial counts at day 7 postchallenge, compared to that of the nonvaccinated mice, and vaccination with 103 CFU of BPZE1 resulted in an approximately 1,000-fold reduction (Fig. (Fig.6b6b).
To investigate whether nasal vaccination with BPZE1 induces a B-cell memory response, antibody titers to FHA and PTX were determined 7 days after challenge and compared to the titers measured before challenge. At all vaccine doses, anamnestic responses to either of the antigens were detected after challenge, with maximal antibody titers to FHA and PTX reached at the threshold of 105 CFU of BPZE1 (Fig. (Fig.77).
Attenuated live vaccines against bordetellosis have been developed to protect dogs and pigs against Bordetella bronchiseptica infections (5, 30). Furthermore, experiments on nonhuman primates carried out more than 40 years ago (14) have suggested that infection by B. pertussis may be the best way to protect against subsequent infection. Although for vaccination against whooping cough this strategy has so far received little attention to the benefit of the development of acellular vaccines composed of defined B. pertussis protein antigens, attempts to genetically attenuate B. pertussis for that purpose have been described several years ago (27). They consisted of deleting the aroA gene, which indeed resulted in strong attenuation but rather poor immunogenicity, probably due to the failure of B. pertussis aroA mutants to colonize the respiratory tract, and several administrations of very high doses were required to protect mice against subsequent challenge with virulent B. pertussis. More recently, we have developed the attenuated BPZE1 vaccine strain, able to persist in the mouse respiratory tract and to induce strong protection after a single nasal administration in a murine respiratory infection model (22). In addition, BPZE1 was shown to protect against Bordetella parapertussis, the second most frequent cause of whooping cough-like illness in humans, whereas the acellular vaccine did not, suggesting that BPZE1 may protect against a wide variety of B. pertussis isolates. This is particularly interesting in light of the potential vaccine escape by the relatively recent strain adaptation observed for B. pertussis, especially in areas of intensive vaccination (3, 9, 10, 26).
Although humans are the only natural reservoir for B. pertussis and mice are comparatively more difficult to infect than humans, as substantially more bacteria are needed to establish infection in mice than in humans, mouse models are nevertheless widely accepted models to preclinically test pertussis vaccines. In particular, the intranasal murine lung infection model has been validated as a good assay to discriminate between the efficacies of different pertussis vaccines (8, 24), whereas the classical intracerebral challenge model developed by Kendrick et al. (16) and used in the European pharmacopeia is much less discriminatory. We therefore used this model in the dose-response studies with the BPZE1 vaccine strain and found a strong correlation between the vaccine dose of BPZE1 and protection against subsequent challenge with virulent B. pertussis. Both high-dose and low-dose challenge models were used, and in all cases, a single BPZE1 administration significantly protected against challenge infection. In the low-dose challenge model, 104 CFU of BPZE1 provided full protection, and 103 CFU of BPZE1 already reduced the bacterial counts by 100,000 compared to those of the nonvaccinated mice. In the high-dose challenge model, total bacterial clearance was not observed at any vaccine dose, but a strong correlation of protection with the vaccine dose was observed. In addition, a strong correlation of the vaccine dose with antibody titers against the major B. pertussis antigens FHA and PTX, as well as with IFN-γ production by spleen cells upon stimulation with these antigens, was observed, and a vaccine dose-dependent anamnestic response was observed upon challenge infection with virulent B. pertussis. Since both antibodies and cell-mediated immunity appear to contribute to protection against whooping cough, these parameters may be valid assays to use for the monitoring of immune responses induced by BPZE1 in future clinical trials. In addition, administration of BPZE1 resulted in the appearance of anti-B. pertussis antibodies in bronchoalveolar lavage fluids, both of the IgG and of the IgA isotypes. Although local anti-B. pertussis IgG was likely to be exudated from sera, it is possible that the IgA had been locally induced, since no anti-B. pertussis was found in the sera. In contrast to vaccination with BPZE1, after systemic vaccination with a commercial acellular vaccine, no anti-B. pertussis IgA production was detected in the bronchoalveolar lavage fluids. Since B. pertussis causes a strictly mucosal respiratory infection, antigen-specific IgA responses might represent important effectors for immunity against pertussis (12).
Importantly, even after administration of the highest vaccine dose (106 CFU), no significant lung pathology was observed, although infection of mice with 106 CFU of the virulent strain resulted in strong lung infiltration of inflammatory cells, attesting to the safety of BPZE1, even at the highest dose tested here. Likewise, infection of infant mice with 106 CFU of BPZE1 did not affect their weight gain, at least over 15 days. In vitro, infection of pneumocyte cultures with MOIs of up to 100 did not cause any cytotoxicity, and the cytotoxicity of BPZE1 infection on a monocyte cell line was minimal, even at an MOI of 100, similar to that of infection with a nonpathogenic E. coli K-12 laboratory strain. These results have contributed to downgrading BPZE1 from a biosafety level 2 organism to a biosafety level 1 organism in France.
Although the human dose of acellular pertussis vaccines has been validated in this intranasal mouse challenge model, it is difficult to extrapolate the optimal BPZE1 vaccine dose in this model to a human dose. It is likely that colonization of the respiratory tract is an important factor for the induction of infection-induced immune protection. B. pertussis is a strictly human pathogen, and it is highly contagious to humans. Only a few bacteria are necessary to establish an infection in humans (11). B. pertussis is much less infectious to mice. Therefore, the optimal human doses of BPZE1 may actually be lower than the optimal doses for mice. A minimum of 103 CFU of BPZE1 is required to establish a persistent colonization in mice, but it is possible that for humans that much less bacteria are needed for colonization to achieve protective immunity. This is an important consideration for the development of BPZE1-based vaccines to protect humans against whooping cough and will be the major issue of the first clinical trials in humans that are currently being planned.
We are grateful to M. Loyens for technical assistance during this study and to C. Creuzi for her expertise in histological analysis.
This work was partly supported by the Agence Nationale de la Recherche (project ANR-05-MIME-BPVAC, project no. A05137ES) and by the European Commission (FP7, grant agreement 201502; Child-Innovac).
Published ahead of print on 27 January 2010.