The first two authors contributed the majority of the writing (E.J.P.) and the experimentation (G.S.K.).
Antibody-mediated bacterial clearance from the lower respiratory tract of mice requires complement component C3
Version of Record online: 15 DEC 2003
Copyright © 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 34, Issue 1, pages 184–193, January 2004
How to Cite
Pishko, Elizabeth J., Kirimanjeswara, Girish S., Pilione, Mylisa R., Gopinathan, L., Kennett, Mary J. and Harvill, Eric T. (2004), Antibody-mediated bacterial clearance from the lower respiratory tract of mice requires complement component C3. Eur. J. Immunol., 34: 184–193. doi: 10.1002/eji.200324234
- Issue online: 19 DEC 2003
- Version of Record online: 15 DEC 2003
- Manuscript Accepted: 7 NOV 2003
- Manuscript Revised: 20 OCT 2003
- Manuscript Received: 4 JUN 2003
- Animal model;
To assess the contribution of complement to respiratory immunity in the context of a natural bacterial infection, we used mice genetically deficient in complement components and the murine pathogen Bordetella bronchiseptica. Complement component C3 was not required for the control of bacterial infection or for the generation of infection-induced protective immunity. However, C3-deficient (C3–/–) mice were severely defective, compared to wild type, in vaccine-induced protective immunity. Adoptively transferred immune serum from convalescent wild-type or C3–/– animals rapidly cleared B. bronchiseptica from the lungs of wild-type mice but did not affect its growth in C3–/– mice, indicating that the defect is not in the generation of protective immunity, but in its function. Immune serum was effective in C5-deficient mice but had little effect in the lungs of mice lacking either Fcγ receptors (FcγR) or CR3, suggesting bacterial clearance is not via direct complement-mediated lysis. Together, these data indicate that complement is required for antibody-mediated clearance of Bordetella and suggest the mechanism involves C3 opsonization of bacteria for phagocytosis that is both CR3- and FcγR-dependent.
Lower respiratory tract
Membrane attack complex
The complement system is a critical aspect of innate immunity that also contributes to the generation of adaptive immunity 1–3. Bacterial pathogens can initiate the complement cascade through several major activation pathways leading to cleavage of C3, which generates C3a, a powerful anaphylatoxin involved in signaling and inflammatory cell recruitment, and the opsonin C3b 4, 5. C3b is also a primary component of the alternative pathway C3 convertase and both C5 convertases. Complement concentrations are relatively low in healthy respiratory tract secretions but can increase rapidly in response to infection 6. Recent work with C3-deficient (C3–/–) mice has revealed important roles for complement within the respiratory tract 1, 7–13. C3–/– mice exhibit decreased airway hyper-responsiveness and lung eosinophilia in an allergic asthma model 14 and are more susceptible to influenza virus 11. However, no such role has been established for complement in respiratory immunity to bacterial pathogens. Indeed, although Mycobacterium avium is sensitive to complement in vitro, C3–/– mice showed no defect in their resistance to this organism 13.
The genus Bordetella comprises a group of closely related gram-negative bacterial respiratory pathogens with a broad host range. B. pertussis and B. parapertussis are the causative agents of whooping cough in humans while B. bronchiseptica infects a number of non-human mammals, causing atrophic rhinitis in pigs, snuffles in rabbits and kennel cough in dogs. B. pertussis and B. parapertussis have recently been reclassified as subspecies of B. bronchiseptica, due to their extremely close relatedness 15.The two human pathogens appear to have made recent independent host range jumps from a B. bronchiseptica-like progenitor to emerge as pathogens that efficiently infect and spread amongst humans 15. However, both B. pertussis and B. parapertussis are more closely related to B. bronchiseptica than they are to each other 15. Since B. bronchiseptica is genetically manipulatable and naturally infects mice, it provides a powerful model system to examine the molecular interactions between a bacterial pathogen and the host immune response within the respiratory tract.
We have shown that antibodies are necessary and sufficient to clear B. bronchiseptica from the lower respiratory tract (LRT) of mice 16, and B. bronchiseptica has been shown to activate compliment in vitro17. Based on the known roles of complement in innate immunity as well as in the generation and function of adaptive immunity, we hypothesized that complement would be required for some aspect of the in vivo immune response to B. bronchiseptica. Interestingly, C3–/– mice showed no defect in their ability to control and clear B. bronchiseptica from the respiratory tract and generated effective protective immunity that rapidly cleared a subsequent infection. However, vaccination of C3–/– mice with heat-killed B. bronchiseptica, while resulting in detectable antibody titers, did not confer any protection against bacterial challenge.
Vaccinated C3–/– mice generate antibodies that are effective in rapidly clearing B. bronchiseptica from wild-type mice. However, immune serum from either wild-type or C3–/– mice had no effect on bacterial numbers in C3–/– mice. Immune serum rapidly cleared B. bronchiseptica from C5-deficient (C5–/–) mice but was less effective in FcγR-deficient (FcγR–/–) or CR3–/– mice, indicating that the primary function of antibody-mediated complement activation is not the direct lysis of bacteria but facilitation of their removal by phagocytic cells. Together these data indicate that complement can contribute to antibody-mediated clearance of bacterial pathogens within the respiratory tract, and reveal a potentially important mechanistic difference between vaccine-induced and infection-induced immunity.
2.1 Normal immune responses of C3-deficient mice to B. bronchiseptica infection
In order to investigate the contribution of complement to the immune response of mice during an initial B. bronchiseptica infection, we compared the ability of wild-type C57BL/6 and congenic C3–/– mice to clear bacterial infection from three regions of the respiratory tract. Mice were inoculated with 5×105 CFU in a volume of 50 μl PBS applied intranasally (i.n.), a method shown to distribute bacteria throughout the respiratory tract 18. On days 0, 7, 14, 28, 49, 70 and 105 post-inoculation bacterial colonization levels were determined in the nasal cavity, trachea and lungs. No significant difference was observed in the levels of B. bronchiseptica colonization between wild-type and C3–/– mice on any of the indicated days post-inoculation (Fig. 1), indicating that C3 is not required for effective clearance of B. bronchiseptica from the respiratory tracts of mice.
To determine whether complement is required for the control of a secondary infection with B. bronchiseptica after a primary infection has been cleared from the LRT, C57BL/6 and C3–/– mice were inoculated with 5×105 CFU in a volume of 50 μl, as described above, and allowed to clear the LRT. These mice were then challenged with a second inoculum of B. bronchiseptica on day 70 post-inoculation, as described above, and bacterial numbers in the nasal cavity, trachea and lungs were determined 3 days after the second inoculation. In both mouse strains bacterial numbers were only a fraction of those found 3 days after primary infection (Fig. 1); approximately 1/100 in the nasal cavity, 1/1,000 in the trachea and 1/3,500 in the lungs (Fig. 2). Since no significant difference was observed between wild-type and C3–/– strains, mice were apparently capable of mounting an efficient anamnestic response to B. bronchiseptica in the absence of a functional complement cascade.
2.2 Vaccine-mediated clearance of B. bronchiseptica from the lung
Although substantial differences have been noted between the protective immunity generated in response to infection and vaccination, there is little understanding of the mechanisms involved in these differences. In both clinical studies and mouse experiments infection-induced immunity has been shown to be more protective than vaccination-induced immunity (19, 20 and unpublished data). To examine whether differences in the involvement of complement could be responsible for these observations, we examined the role of C3 in effective vaccine-induced protection against B. bronchiseptica. Groups of C57BL/6 and C3–/– mice were vaccinated i.p. with 108 CFU heat-killed B. bronchiseptica, followed by a second dose 2 weeks later. Two weeks after the second vaccination, the mice were challenged i.n. with 5×105 CFU of B. bronchiseptica as described above. Animals were sacrificed on day 3 post-challenge and respiratory tract colonization levels were determined as described above.
Vaccination had little effect on bacterial numbers in the nasal cavity and a moderate effect in the trachea that was not complement-dependent (Fig. 3A). In the lungs of wild-type animals, vaccination reduced bacterial numbers to 1/60,000 that of naive animals on day 3 post-inoculation. However, vaccinated C3–/– mice exhibited only a 30-fold reduction in CFU in the lungs and had more than 2,000 times the number of bacteria present in the lungs of vaccinated wild-type mice (Fig. 3A). Since complement can contribute to the generation of adaptive immunity, we compared the anti-B. bronchiseptica titers of various antibody isotypes in the serum of vaccinated wild-type and C3–/– mice. Although C3–/– mice had IgM titers similar to that of wild-type mice, their titers of each IgG isotype were substantially lower (Fig. 3B).
2.3 Role of complement in antibody-mediated clearance of B. bronchiseptica from the LRT
Since complement is involved in both the generation and function of antibodies, the failure of vaccinated C3–/– mice to clear B. bronchiseptica could be due to a defect in either antibody generation or function. To discriminate between these possibilities we performed reciprocal adoptive transfer of immune serum between wild-type and C3–/– mice followed by challenge with B. bronchiseptica. Heat-inactivated serum (usually 0.2 ml) from wild-type or C3–/– mice that were naive (naive serum) or convalescent from B. bronchiseptica infection 28 days post-inoculation (immune serum) was adoptively transferred i.p. into wild-type or C3–/– mice. We then immediately challenged these mice with 5×105 CFU of B. bronchiseptica i.n. as described above.
Naive serum had no effect on bacterial numbers in either mouse strain. Wild-type immune serum rapidly cleared B. bronchiseptica from the LRT of wild-type mice from day 3 post-inoculation/transfer onward, but had no effect in C3–/– mice (Fig. 4). As little as 0.02 ml of immune serum was able to lower bacterial numbers significantly, with 0.1 ml and higher able to reduce bacterial numbers in wild-type mice to 1/1,000 that of untreated wild-type mice on day 3 post-inoculation/transfer, whereas up to 1 ml of that same serum had no effect on bacterial numbers in C3–/– mice (data not shown). Immune serum from C3–/– mice was as effective as serum from wild-type mice in clearing B. bronchiseptica from the LRT of wild-type mice (data not shown). Similarly, C3–/– vaccine serum was effective in clearing B. bronchiseptica from the LRT of wild-type mice, indicating that functional impairment of serum antibodies rather than their generation is responsible for the defect observed in vaccinated C3–/– mice. Therefore, independent of any effect on antibody effect on antibody generation, C3–/– mice are severely defective in antibody functions involved in mediating bacterial clearance.
There are several ways in which complement could be acting in conjunction with antibodies to clear bacterial infections in the lung. Complement may directly lyse bacteria via formation of the membrane attack complex (MAC), it may attract and/or activate phagocytic cells, and it may facilitate phagocytosis and bacterial killing by phagocytic cells. In order to determine whether direct complement-mediated lysis was occurring, we adoptively transferred serum into wild-type and C5–/– mice as described above. Antibodies effectively reduced the numbers of B. bronchiseptica in the tracheas and lungs of both strains of mice on day 3 post-inoculation (Fig. 5). Since C5–/– mice lack C5b, a critical part of the MAC, and these mice reduced bacterial numbers to the same extent as wild-type mice after adoptive transfer of antibodies, it appears that direct MAC-mediated bacterial lysis is not required for bacterial clearance. FcγR–/– mice, which are defective in FcγR-mediated phagocytic functions but have a normal complement system, failed to clear B. bronchiseptica from the lungs after adoptive transfer of antibodies (data not shown). Together these data suggest that antibody-mediated bacterial clearance requires complement not to directly lyse bacteria, but to facilitate phagocytosis via FcγR.
2.4 Recruitment of inflammatory cells to the lungs during B. bronchiseptica infection
Activation of the complement cascade results in the release of numerous factors that can attract and activate phagocytic cells. Although effective antibody-mediated clearance of B. bronchiseptica from C5–/– mice indicates that C5a is not required, other factors, including C3a, are not dependent on C5 and may affect recruitment and activation of phagocytes. We therefore examined inflammatory cell recruitment in the presence and absence of C3. Lungs were excised from B. bronchiseptica-infected wild-type and C3–/– mice that had been adoptively transferred naive or immune serum, as described above. Formalin-fixed, and hematoxylin and eosin (H&E)-stained sections were examined by a trained mouse pathologist blinded as to the treatment.
On day 1 post-transfer and -inoculation there was no detectable difference in cell recruitment or tissue damage between the wild-type or C3–/– mice regardless of the serum used in the transfer (Fig. 6A). Lesions included perivascular and peribronchiolar lymphoid cuffing, suppurative alveolar pneumonia, and mixed inflammatory cell infiltrates. Lungs from wild-type mice given naive serum, and C3–/– mice given naive, or immune serum, exhibited similar lesions and inflammatory cell infiltrates on day 3 post-transfer and -inoculation, and contained similar high numbers of bacteria (Fig. 4). Wild-type mice given immune serum had the highest overall lesion scores on day 3, although they had cleared bacteria by this time (Fig. 6A). C5–/– mice showed no significant difference in histopathology on day 3 post-inoculation and -adoptive transfer when compared to wild-type controls treated similarly (data not shown). The lack of either C3a or C5a appears to have little if any effect on pathological lung inflammation.
In order to quantify inflammatory cell trafficking to the lungs in response to B. bronchiseptica infection in the presence and absence of C3, we determined the number of total leukocytes, macrophages, neutrophils and lymphocytes in the lungs 12 h post-inoculation. No significant differences were observed in the total number of cells (Fig. 6B) or in any of the individual cell subpopulations (Fig. 6C) in the lungs between wild-type and C3–/– mice, suggesting mice can mount a normal inflammatory response to B. bronchiseptica in the absence of complement. Since antibodies can accentuate the inflammatory response by activating the classical complement pathway, we investigated inflammatory cell recruitment to the lungs in the presence of adoptively transferred antibodies in wild-type and C3–/– mice. Serum antibodies significantly increased the numbers of total cells and neutrophils recovered from both wild-type and C3–/– mice (Fig. 6B, C), indicating that C3 is not essential for cell recruitment to this organ during a B. bronchiseptica bacterial infection.
2.5 Role of CR3 in antibody-mediated clearance of B. bronchiseptica from the LRT
To distinguish between the effects of C3a on recruitment and activation of phagocytes from the effects of C3b opsonization on phagocytosis, we used mice lacking the α subunit (CD11b) of CR3, the primary receptor involved in recognizing and phagocytosing iC3b-coated bacteria. CR3–/– mice were adoptively transferred naive or immune serum just prior to inoculation with 5×105 CFU of B. bronchiseptica as described above. CR3–/– mice treated with immune serum were substantially deficient in their ability to clear B. bronchiseptica from the LRT on day 3 post-inoculation compared to wild-type mice (Fig. 7). Colonization levels were reduced to approximately 1/6 and 1/500 the untreated levels in the trachea and lung of CR3–/– mice, respectively. In contrast, antibody treatment reduced bacterial numbers to 1/4,000 and 1/100,000 of untreated levels in wild-type mice. Although there was a small CR3-independent antibody effect, CR3 appears to be required for serum antibodies to fully and efficiently clear B. bronchiseptica from the LRT of mice.
Even in relatively simple in vitro experiments, the numerous serum proteins that comprise the complement system are known to contribute a complex set of signals and activities that are affected by the concentrations, proximity and orientation of many activators and inhibitors 2, 21, 22. In addition, the effects of complement activation are highly dependent on the nature of the target, and bacterial pathogens have evolved multiple mechanisms to evade these effects 23–27. It would therefore be difficult to extrapolate from even the most rigorous in vitro studies to the in vivo role of complement in respiratory tract immunity. Conditions that affect such assays are likely to differ dramatically from the basal aqueous sol to the surfactant- and mucus-rich air interface of the respiratory mucosa. Here we examined the role of complement in respiratory immunity by measuring its contribution to the control and clearance of a respiratory pathogen in a natural host infection model. While our initial experiments in C3–/– mice indicate that complement is not required to control B. bronchiseptica infection and limit it to the upper respiratory tract, it is certainly required for effective vaccine-induced immunity and for the actions of adoptively transferred antibodies in the LRT.
There are multiple mechanisms by which complement could contribute to the elimination of bacteria from the LRT. The most direct is by MAC formation, which leads to lysis of bacterial cells through membrane disruption. We have previously shown that antibodies very efficiently activate complement to directly lyse B. bronchiseptica by this mechanism in vitro, even when serum is diluted in PBS to concentrations as low as 1% 17. However, this mechanism is apparently not sufficient to eliminate B. bronchiseptica from the lungs of mice, since antibodies did not clear the bacteria from FcγR–/– mice, which have an intact complement system, but did clear the bacteria from C5–/– mice, which lack the ability to form the MAC.
The other two likely mechanisms of complement-mediated bacterial clearance involve effector cells. Phagocytic cells can be recruited and activated by the diffusible molecules that are released by the catalytic breakdown of complement components. Alternatively, molecules that are deposited on bacterial membranes can act as ligands for receptors, such as CR3, that mediate phagocytosis and killing of bacterial cells 28, 29. Since antibodies are also efficient opsonins, and the FcR that they bind are required for efficient antibody-mediated clearance, we hypothesized that complement may be acting primarily to recruit and activate FcγR-bearing phagocytes. However, we observed no significant difference between wild-type and C3–/– mice in their recruitment of inflammatory cells (Fig. 6). Furthermore, antibodies were effective in C5–/– mice, where C3b opsonization occurs but the highly active chemoattractant C5a is missing. In addition, antibodies had little effect in CR3–/– mice, indicating that complement-mediated opsonization is required for efficient antibody-mediated clearance. These experiments indicate that C5a is not required for efficient antibody-mediated clearance; and since CR3 is required it is likely that both MAC formation and complement-dependent recruitment are not sufficient for clearance in the absence of complement-mediated opsonization.
Although not the focus of this study, these experiments revealed substantial, and potentially important, differences between infection-induced and vaccination-induced protective immunity. Both vaccination and adoptively transferred serum conferred protective immunity that was dependent on complement. We have also observed that vaccination does not confer protection in B cell-deficient MuMT mice but adoptive transfer of vaccine-induced serum does (unpublished data). Together these data suggest that vaccination induces primarily serum antibodies which require complement to effectively clear B. bronchiseptica infection from the LRT. In contrast, infection-induced immunity conferred protection that was independent of complement, suggesting mucosal infection induces some immune mechanism independent of serum antibodies. B. bronchiseptica is not believed to invade tissues or cells, and there is no obvious mechanism for clearance by T cells. In addition, MuMT mice do not generate protective immunity 17, indicating that antibodies are required. Since systemic vaccination does not result in efficient generation of mucosal IgA, we propose that infection-induced immunity results in the production of IgA, which clears B. bronchiseptica by complement-independent mechanisms. The failure to produce substantial secretory IgA may explain certain shortcoming of vaccines to the Bordetellae, as well as to other mucosal pathogens, and suggests strategies to overcome these shortcomings.
Together, these data support a detailed model for antibody-mediated clearance of B. bronchiseptica from the LRT of mice. We propose that antibodies activate complement to opsonize bacteria which are then phagocytosed via a combination of both CR3 and FcγR. This is consistent with the phosphatidylinositol 3-kinase-dependent model proposed by Jones et al. for activation of phagocytosis by human neutrophils 29. In this model, ligation of antibody-bound FcγR on the neutrophil surface trigger activation of CR3 through the phosphatidylinositol 3-kinase pathway, which leads to increased avidity of CR3 for bacterial surface-bound iC3b 29, 30. We are currently investigating the contributions of individual FcγR as well as the role of CR3 in serum antibody-mediated clearance of the Bordetellae. It will also be important to determine whether this mechanism of clearance is common to vaccine-induced protection from other respiratory pathogens.
4 Materials and methods
B. bronchiseptica RB50 (wild type) was obtained after a single passage from an original rabbit isolate 31. Bacteria were maintained on Bordet-Gengou agar (Difco) supplemented with 7.5% defribrinated sheep blood and 20 μg/ml streptomycin. Bacteria were grown in Stainer-Scholte broth with supplements and 20 μg/ml streptomycin to mid-log phase (optical densities of approximately 0.3 at 600 nm) at 37°C on a roller drum for experiments.
C57BL/6 mice were obtained from Jackson Laboratories (Bar Harbor, ME), and C3–/– mice, back-crossed extensively onto a C57BL/6 background, have been described elsewhere and were kind gifts of Dr. Rick Wetsel 8. C5–/– mice (B10.D2-H2dH2-T18cHc0/nSnJ) and their parent strain (B10.D2-H2dH2-T18cHc1/nSnJ) were also obtained from Jackson Laboratories. CR3–/– mice (C57BL/6 background) were the kind gift of Dr. Christie M. Ballantyne and have also been previously described 32.
4.3 Inoculation and immunization protocols
Mice were lightly sedated with isoflurane, and inoculated i.n. by pipetting 5×105 CFU of bacteria in 50 μl onto the tip of the external nares. For the time-course experiment groups of four animals were sacrificed on days 0, 3, 5, 7, 14, 28, 49, 70 and 105 post-inoculation. Colonization levels were determined by homogenizing the indicated organ in 1× PBS and plating aliquots for colony counts. The nasal cavity and trachea were homogenized in 500 μl of PBS and lungs in 1 ml PBS. The homogenates and necessary dilutions were plated in 50-μl volumes onto Bordet-Gengou agar with streptomycin. Colonies were counted after 2 days incubation at 37 °C. For re-infection, animals were infected as described above and then 70 days post-infection challenged with a second i.n. dose of 5×105 CFU of B. bronchiseptica. On day 3 post-challenge mice were killed and colonization levels determined as described above.
For vaccination, animals were immunized i.p. twice at 2-week intervals with 108 CFU of heat-killed B. bronchiseptica. Heat-killed bacteria were prepared by incubating bacteria grown to mid-log phase at 80°C for 30 min. At 2 weeks after administration of the second vaccination, animals were challenged i.n. as described above and killed on day 3 post-challenge for determination of colonization levels as described. For passive transfer of antibodies, wild-type mice were inoculated as described and serum was collected on day 28 post-inoculation. Titers of anti-B. bronchiseptica antibodies in the convalescent sera were determined by ELISA as previously described 33. Injection of 200 μl of this immune serum i.p. into naive mice was immediately followed by i.n. inoculation as described above. Animals were killed on the indicated day post-transfer and inoculation and colonization determined as described. Animals were handled in accordance with institutional guidelines, in keeping with full accreditation from the Association for Assessment and Accreditation of Laboratory Animal Care International.
4.4 Histological evaluation
The lungs from treated mice were removed, inflated and fixed with 10% neutral-buffered formalin, embedded in paraffin, and sectioned. Slides were routinely prepared and stained with H&E. An assessment of microscopic lesions was made by one of the authors (M.J.K.), experienced in rodent pathology and blinded to experimental treatment. Descriptive evaluations of the lesions were recorded, and lung lesions were graded using a scale of 0 to 4. Sections with no lesions and no inflammation were given a score of 0, a score of 1 indicated slight inflammation with few or scattered lesions and <10% of lung fields affected, 2 indicated mild lesions with 10–20% of lung fields affected, 3 indicated moderate lesions with 20–30% of the lung fields affected, and those given grade 4 werecharacterized by extensive lesions, marked inflammation, and >30% of the lung affected.
4.5 Preparation of lung leukocytes
Lung leukocytes were isolated as previously described 34. Briefly, the lungs were perfused with 2 ml of PBS, excised, rinsed in PBS, minced, and enzymatically digested for 2 h at 37°C using 3 ml of digestion buffer consisting of Dulbecco's modified Eagle's medium, 10% fetal calf serum, 1% penicillin-streptomycin solution (10,000 μg/ml), 1 mg/ml collagenase and 10 μg/ml DNase. Lung leukocytes were isolated using density gradient centrifugation and then counted on a hemacytometer. Cell differentials (neutrophils, macrophages, and lymphocytes) werevisually counted from Wright-Giemsa-stained samples of lung cell suspensions cytospun onto glass slides (Shandon Cytospin, Pittsburgh, PA). A total of 300 cells were counted from randomly chosen high-powered-microscope fields for each sample. The absolute number of a leukocyte subset was calculated by multiplying the percentage of each subset in an individual sample by the total number of lung leukocytes in that mouse. Statistical significance of data points was determined using a Student's unpaired t-test.
This work was funded by a grant from Neose Corp, Pennsylvania Department of Agriculture grant ME440678 and USDA grant 2002–35204–11684.
- 6Complement-mediated host defense in the lung. Am. J. Physiol. Lung Cell. Mol. Physiol. 2000. 279: 790–798.