Protective immunity to Bordetella pertussis requires both B cells and CD4(+) T cells for key functions other than specific antibody production.

To investigate the fundamental nature of protective immunity to Bordetella pertussis, we studied intranasal immunization of adult mice with formalin-fixed B. pertussis (FFBP), followed by aerosol B. pertussis challenge. Mice given two doses of FFBP intranasally completely cleared a subsequent pertussis aerosol challenge from tracheae and lungs (defined as protection), but there was no correlation between levels of specific antibody and clearance of bacteria. Further, transfer of immune serum before aerosol challenge had minimal effects on bacterial burdens. However, pertussis-specific T cells producing interferon gamma but not interleukin 4 or interleukin 10 were detected in draining lymph nodes of FFBP-immunized mice. Significantly, repeated immunization of B cell knockout (BKO) mice resulted in partial protection, and complete protection was reconstituted by transfer of pertussis-immune B cells; reconstituted BKO mice had little if any detectable antipertussis antibodies. Immunization of mice lacking all T cells or lacking CD4(+) T cells did not lead to protection; in contrast, CD8(-) mice were protected. Mice depleted of CD4(+) T cells after immunization but before aerosol challenge, which thus had normal amounts of specific antibodies, were not optimally protected. Taken together, these data indicate that protective immunity to pertussis is dependent on both CD4(+) T cells and B cells, and both cell types provide significant functions other than specific antibody production.

Role of gamma interferon in natural clearance of Bordetella pertussis infection.

Using a mouse model of Bordetella pertussis infection, we have analyzed the role of gamma interferon (IFN-gamma) in bacterial clearance from the respiratory tract. Adult BALB/c mice began to clear a respiratory infection within 3 weeks postinfection, with complete resolution of infection 6 to 8 weeks postinfection. In contrast, neither adult SCID mice (which lack mature B and T lymphocytes) nor adult nude mice (which lack mature T lymphocytes) controlled B. pertussis infection, and both strains died within 3 to 5 weeks postinfection. Short-term T-cell lines generated from the draining lymph nodes of the lungs of infected BALB/c mice were found to be CD4+ and produced IFN-gamma but no detectable interleukin-4. Analyses of IFN-gamma mRNA induction in the lungs of mice following B. pertussis infection showed that in both BALB/c and C57BL/6 mice, IFN-gamma mRNA levels increased sharply by 1 week postinfection and then subsequently declined. Further exploration of a potential role for IFN-gamma demonstrated that infection of adult BALB/c mice depleted of IFN-gamma in vivo with anti-IFN-gamma monoclonal antibodies resulted in greater numbers of bacteria recovered from the lungs than in infected, control BALB/c mice, although IFN-gamma-depleted mice could subsequently clear the infection. Infection of mice which have a disrupted IFN-gamma gene resulted in bacterial clearance with a time course similar to those seen with IFN-gamma-depleted mice. These results indicate that IFN-gamma plays a role in controlling B. pertussis infection.

Pertussis antigens that abrogate bacterial adherence and elicit immunity.

Infectious disease processes follow the initial steps of adherence of the organism to host tissues and subsequent colonization of the target tissues that can occur through specific adhesion-receptor systems. Bordetella pertussis, the human pathogen that causes whooping cough, has evolved a genetically controlled system whereby adhesins are expressed when they enter the human host. Two adhesins, filamentous hemagglutinin (FHA) and pertactin, mediate the adherence of the bacterium to eukaryotic cells through varied attachment mechanisms, including lectin-like binding sites that interact with sulfated sugars on cell surface glycoconjugates and the ARG-GLY-ASP binding sequence, which recognizes a family of integrins found on the cell surface. The differential expression of relevant receptors by various eukaryotic cells likely plays a role in the pathogenesis and immune response to the bacterium by the host, directing the organism to specific cell types and to specific tissue sites. Substantial evidence exists that the B. pertussis adhesins, FHA and pertactin, elicit immune responses that are protective in animal models for the disease, including serum antibody production and local immune responses in the respiratory tract following nasal administration of encapsulated antigens. Both of these adhesins are components of new acellular pertussis vaccines that have proven safe and highly effective for prevention of serious disease in infants.

Analysis of protective and nonprotective monoclonal antibodies specific for Bordetella pertussis lipooligosaccharide.

In this study, it has been determined that immunoglobulin G1 (IgG1) and IgG3 monoclonal antibodies directed to the lipooligosaccharide A of Bordetella pertussis were able to protect mice from fatal aerosol infection. No correlation was found between the bactericidal activity in vitro in the presence of complement and the protection in mice, since a bactericidal IgG3 did not elicit protection. In addition, no significant difference in protective capacity was observed with bactericidal and nonbactericidal IgG1 antibodies, indicating that bactericidal activity is not a requirement for protection mediated by certain anti-lipooligosaccharide A antibodies. A reduction in protection in C5-deficient mice was observed, suggesting a significant role for complement in certain host defense mechanisms against B. pertussis infection.

Long-lived respiratory immune response to filamentous hemagglutinin following Bordetella pertussis infection.

Systemic and mucosal B-cell-mediated immune responses to purified filamentous hemagglutinin (FHA) in mice were analyzed at different times following a single respiratory infection with Bordetella pertussis. Serum immunoglobulin G (IgG) anti-FHA and respiratory IgG and IgA anti-FHA antibodies were first detected at 3 weeks postinfection, reached high levels by 8 weeks postinfection, and remained at high levels 12 to 32 weeks postinfection. FHA-specific B lymphocytes isolated from the spleens or lungs of uninfected control mice or mice convalescing from B. pertussis respiratory infection were analyzed in limiting-dilution cultures. Analysis of culture supernatants for the production of antibodies to FHA revealed an increased frequency of FHA-specific B cells of both the IgG- and the IgA-secreting classes in the lungs and tracheas of aerosol-challenged mice; these levels remained high as late as 25 weeks postinfection, compared with those in uninfected controls. No corresponding increase in the frequency of FHA-specific B cells in the spleens of aerosol-infected mice was observed. This long-lasting response observed in cultured cells was radiation resistant, a result suggesting that this response was due to B cells already activated in vivo. Polymerase chain reaction analysis revealed low but detectable levels of B. pertussis chromosomal DNA in 75% of mice tested at 8 weeks postinfection and 37.5% of mice tested at 26 weeks postinfection, at which times high levels of anti-FHA antibody were detected. One explanation for these data may be that, in this animal model, a major adhesin of B. pertussis can persist and interact with components of the immune system to stimulate the production of specific antibody in the respiratory tract many weeks after a single B. pertussis infection.

Mucosal immunization with filamentous hemagglutinin protects against Bordetella pertussis respiratory infection.

Mucosal immunization of mice with purified Bordetella pertussis filamentous hemagglutinin (FHA), by either the respiratory or the gut route, was found to protect against B. pertussis infection of the trachea and lungs. Intranasal immunization of BALB/c and (C57BL/6 x C3H/HeN)F1 adult female mice with FHA prior to B. pertussis aerosol challenge resulted in a 2 to 3 log reduction in number of bacteria recovered from the lungs and the tracheas of immunized mice in comparison to unimmunized controls. Intraduodenal immunization of adult mice with FHA before infection also resulted in approximately a 2 log reduction in the recovery of bacteria from the lungs and the tracheas of immunized mice in comparison to unimmunized controls. Immunoglobulin A and immunoglobulin G anti-FHA were both detected in bronchoalveolar lavage fluids of mucosally immunized mice. Limiting dilution analysis revealed a 60-fold increase in the frequency of FHA-specific B cells isolated from the lungs of mice immunized intranasally with FHA in comparison to unimmunized control mice. These data suggest that both gut and respiratory mucosal immunization with a major adhesin of B. pertussis generates a specific immune response in the respiratory tract that may serve as one means of mitigating subsequent B. pertussis respiratory infection.

Contribution of the B oligomer to the protective activity of genetically attenuated pertussis toxin.

An enzymatically deficient recombinant S1 subunit, in which Arg-9 was replaced by Lys, was combined with native B oligomer to form a mutant holotoxin molecule. This molecule exhibited decreased leukocytosis-promoting and histamine-sensitizing activities compared with those of the native toxin, supporting the view that the B oligomer is not responsible for these activities. The protective activity of this genetically attenuated pertussis toxin was compared with that of B oligomer alone. The mutant pertussis toxin and B oligomer were similarly capable of protecting mice against a respiratory infection with Bordetella pertussis, suggesting that the B oligomer makes a significant contribution to the protection afforded by the genetically attenuated holotoxin.

The molecular engineering of pertussis toxoid.

The demand for a safer pertussis vaccine has led to the development of acellular vaccine products. We have sought to manufacture a component vaccine based upon the genetic inactivation of pertussis toxin derived by recombinant DNA technology and protein engineering. Rational site-directed mutagenesis of the S1 subunit of pertussis toxin has resulted in an enzymatically-deactivated polypeptide which retains its immunogenic potential. Mutagenic analysis of the other subunits of this toxin has permitted a delineation of the structural determinants involved in its recognition of cellular receptors. The in vitro assembly of holotoxin species possessing selectively engineered subunits may facilitate the production of a molecularly-defined genetic toxoid for pertussis prophylaxis.

Mechanism of pertussis toxin B oligomer-mediated protection against Bordetella pertussis respiratory infection.

Immunization with the B oligomer of pertussis toxin protected neonatal mice from a lethal respiratory challenge with Bordetella pertussis. All mice immunized with 8 micrograms of B oligomer survived aerosol challenge and had peripheral leukocyte counts and weight gains similar to those of mice immunized with pertussis toxoid before challenge and to those of control mice that were not challenged. Unprotected mice challenged with an aerosol of B. pertussis had an increase in peripheral leukocyte count, failed to gain weight, and died within 21 days of challenge. Protection appeared to be dose dependent, since a dose of 1 microgram of B oligomer per mouse prevented death in 100% of the mice challenged with B. pertussis, whereas 0.4 micrograms of B oligomer protected 50% of the challenged mice. Mice immunized with the B oligomer had increases in immunoglobulin G (IgG) anti-B oligomer in sera and in IgG and IgA anti-B oligomer in bronchoalveolar lavage fluids 1 to 3 weeks after respiratory challenge. Specific anti-B oligomer antibodies could not be detected in unimmunized, infected mice at the same time after challenge. Intravenous administration of the monoclonal antibody 170C4, which binds to the S3 subunit of the B oligomer, protected neonatal mice from B. pertussis respiratory challenge, while administration of an IgG1 anti-tetanus toxin monoclonal antibody, 18.1.7, was not protective. We conclude that anti-B-oligomer-mediated neutralization of pertussis toxin is one mechanism of protection in the mouse model of B. pertussis aerosol challenge.

Characterization of the protective capacity and immunogenicity of the 69-kD outer membrane protein of Bordetella pertussis.

Immunization with the 69-kD outer membrane protein (OMP) of Bordetella pertussis protected neonatal mice against lethal respiratory challenge with B. pertussis 18323. Active immunization elicited a serum IgG anti-69-kD OMP response at the time of challenge, with IgG anti-69-kD OMP antibodies detected in bronchoalveolar lavage fluid after challenge. Intravenous administration of BPE8, a monoclonal IgG1 anti-69-kD OMP, also protected young mice against B. pertussis challenge. Intravenously injected BPE8 was detected in the lungs of mice at the time of aerosol challenge, suggesting that the presence of specific antibody in the lungs may mediate protection. Thus the 69-kD OMP of B. pertussis is a protective antigen in mice that elicits specific serum antibody that can transude to the lung. The 69-kD OMP was detected in a preparation of a Takeda acellular vaccine by immunoblot analysis and a serum antibody response to the 69-kD OMP was observed in 18-mo-old children boosted with this preparation of Japanese acellular vaccine. Our results demonstrate that the B. pertussis 69-kD OMP is a protective antigen in animals, is immunogenic in humans, and is present in a preparation of acellular pertussis vaccine that is widely used in Japan. These findings indicate that the 69-kD OMP should be seriously considered as a candidate for inclusion in new formulations of antigenically defined acellular pertussis vaccines.

Structural and functional properties of a 69-kilodalton outer membrane protein of Bordetella pertussis.

A-69-kDa outer membrane protein present on virulent Bordetella pertussis cells is recognized by the agglutinating monoclonal antibodies BPE3, BPD8, and BPE8. The amino acid composition of this protein, purified from heat extracts of B. pertussis BP353 cells, is different from that of the two major fimbrial antigens of B. pertussis, which is consistent with its being a nonfimbrial protein based on other criteria. Western blot analysis using the monoclonal antibody BPE3 demonstrated that a slightly larger but antigenically cross-reactive protein is also expressed by Bordetella bronchiseptica and Bordetella parapertussis. In addition, a large molecular weight species of about 180-kDa is found in outer membrane extracts of B. bronchiseptica which may represent a precursor form of the protein or indicate that the protein can exist as an oligomer. The monoclonal antibody BPD8 directed against the 69-kDa protein almost completely inhibited the enzymatic activity of adenylate cyclase purified from B. pertussis and also inhibited the intoxication of mammalian cells by this enzyme. Since little enzymatic activity was found associated with the purified 69-kDa protein, these data suggest a role for the 69-kDa protein in regulating the adenylate cyclase toxin of B. pertussis. An additional monoclonal antibody directed against the 69-kDa protein, BPE8, decreases lymphocytosis and delays death in mice receiving a respiratory challenge of virulent B. pertussis cells. These studies suggest that further investigation into the role of this protein as a protective antigen and vaccine candidate is warranted.

Neutrophil recruitment and bacterial clearance correlated with LPS responsiveness in local gram-negative infection.

The inflammatory response to Gram-negative infection was studied in LPS responder and nonresponder C3H mice. Twenty-four hours after ascending E. coli urinary tract infection, an influx of neutrophils into the urine was observed in C3H/HeN mice (Lpsn,Lpsn); no significant neutrophil influx occurred in C3H/HeJ mice (Lpsd,Lpsd) at this time. A second peak of urinary neutrophil excretion was observed in both strains of mice approximately 6 days post-infection. The first, but not the second peak was inducible by inoculation with formalin-killed E. coli but not by Gram-positive bacteria. This finding suggested that the first peak is triggered by LPS, whereas the second peak emanates from other bacterial components which activate both LPS responder and nonresponder mice. The first peak of the inflammatory response was inversely related to bacterial clearance. C3H/HeJ mice (Lpsd,Lpsd) retained about 2000-fold more E. coli in the kidneys than C3H/HeN mice (Lpsn,Lpsn). The infection persisted despite the late-occurring influx of neutrophils in C3H/HeJ mice. These results suggest that an inflammatory response to LPS is required for the elimination of a local Gram-negative infection.

CH isotype switching in B cell differentiation.

We have functionally defined a number of B cell subsets that likely represent B cells at different stages of development, based on the pattern of CH isotypes expressed by their clones in splenic fragment or microcultures and on those factors necessary in culture to support the growth of a clone displaying a particular isotype or set of isotypes. Our observations are consistent with isotype switching being a stochastic process which results in the occurrence of progressive isotype restriction in members of a diversifying clone. The surface marker best predictive of the pattern of isotypes a clone may secrete is the sIg isotype of its B cell precursor. Those B cells that have switched to the expression of non-IgM isotypes in vivo can be stimulated in vitro in splenic fragments to give an antibody-secreting clonal culture but so far cannot be stimulated in a microculture of dispersed cells that supports clones secreting IgM alone or with other isotypes.

Rise in inulin-sensitive B cells during ontogeny can be prematurely stimulated by thymus-dependent and thymus-independent antigens.

We have shown that the delayed acquisition of competence in expression of an anti-inulin (IN) response by neonatal BALB/c mice is preceded by a natural increase in the frequency of IN-sensitive B cells between 3 to 5 weeks of life. Up until 3 weeks of age, BALB/c mice resemble adult germfree mice in their low frequency of IN-sensitive B cells (1 to 2/10(6) B cells). Thereafter, the population of IN-reactive B cells rises about 10-fold to the young adult level, without deliberate immunization. This naturally expanded population of IN-reactive cells has an isotype profile resembling populations arising after intentional priming with other antigens in that it contained a large proportion of cells which generated clones expressing immunoglobulin G and immunoglobulin A isotypes, often without detectable immunoglobulin M. The late rise in frequency of IN-sensitive precursors could be induced prematurely by deliberate priming at 3 days of age with either the thymus-dependent antigen IN-hemocyanin or the thymus-independent bacterial levan. However, no circulating anti-IN could be detected following this early administration of antigen. Thus, the delayed expression of anti-IN does not reflect an absence of B cells of the appropriate specificity which can be stimulated to divide by antigen but rather other events occurring subsequent to perinatal generation of antigen-sensitive cells. Our observations support a role for the beta2 --> 1 fructosyl group as an environmental determinant which selectively expands out preexisting antigen-sensitive B cells at 3 to 5 weeks of age and these include clonotypes which express the predominant anti-IN idiotypes.