Innate Immune Defense against Pneumococcal Pneumonia Requires Pulmonary Complement Component C3
http://www.100md.com
感染与免疫杂志 2005年第7期
Division of Infection and Immunity, IBLS, University of Glasgow, Glasgow G12 8QQ, United Kingdom
ABSTRACT
Complement is known to be involved in protection against systemic infection with Streptococcus pneumoniae. However, less is known about effects of complement within the lungs during pneumococcal pneumonia. By intranasally infecting transgenic mice unable to express complement C3, we investigated the role of complement in pulmonary defenses against S. pneumoniae. It was demonstrated that within the lungs, there is a requirement for C3 during the initial hours of infection. It was found that within 1 h of infection, bacterial loads decreased within lung airways of control mice as C3 protein increased. The lack of C3 resulted in the inability to control growth of wild-type or attenuated pneumococci within the lungs and bloodstream, resulting in an overwhelming inflammatory response and shorter survival times. Our results show that during the initial hours of infection with S. pneumoniae, C3 is protective within the lungs and subsequently plays an important role systemically.
INTRODUCTION
Streptococcus pneumoniae (the pneumococcus) is an important pathogen of humans. The organism causes diseases such as pneumonia, meningitis, bacteremia, and otitis media. Several problems are associated with the control of pneumococci, including a high morbidity rate despite the availability of antibiotics; increasing levels of antibiotic resistance; and limited efficacy of vaccines (23 valent polysaccharide vaccine), high cost, and limited serotype coverage (conjugate vaccines). These facts emphasize the requirement for greater understanding of a protective immune response against the pneumococcus.
A requirement for complement in protection against S. pneumoniae is highlighted by the fact that humans with a deficiency in any of the three complement pathways display heightened sensitivity to pneumococcal infection (3, 13, 25, 29).
S. pneumoniae inhibits the complement pathway in several ways. As a gram-positive bacterium, the pneumococcus is resistant to the bactericidal and lytic activities of complement (19). Pneumolysin can stimulate the classical complement pathway in the absence of pneumolysin-specific antibody, diverting the inflammatory response from intact pneumococci (27). Pneumococcal surface protein A (PspA) prevents the deposition of C3b onto the surface of pneumococci and interferes with both the classical and alternate complement pathways (28, 37). PspC and the homologous protein Hic can both bind complement factor H (11, 18), which inhibits the alternative pathway C3 convertase and protects host cells from complement attack. Factor H bound to pneumococcal PspC should retain the ability to bind C3b and inhibit complement activation (12). The pneumococcal surface protein PhtB (also known as PhpA and BVH-11) has the ability to degrade C3 (4, 42).
The liver is thought to be the major source of complement, with most complement components in the lung being derived from plasma during inflammation (32). It has therefore been presumed that complement does not play a major role in antipneumococcal defense within the lungs during the early stages of pneumonia but rather becomes involved when the host is bacteremic. However, production of complement also occurs within the lungs. Both pulmonary macrophages and epithelial cells can synthesize and secrete C2, C4, C3, C5, and factor B in vitro (10, 33).
We have investigated the importance of complement within the lungs in vivo by inducing pneumococcal pneumonia in transgenic mice unable to express C3.
MATERIALS AND METHODS
Mice. Male and female C57BL/6 mice (20 to 30g; Harlan Olac, Bicester, United Kingdom), C3-deficient mice (C3– [39]), and C5a receptor-deficient mice (C5ar– [15]) (both bred in-house and 20 to 30g) were used in these studies. All experiments were carried out in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986.
Bacteria. The pneumococcal strains used were serotype 2 (D39, NCTC 7466; Central Public Health Laboratory, London, United Kingdom), D39 unable to express pneumolysin (PLY– [see below]), acapsular serotype 2 (GalU– [see below]), and serotype 4 (TIGR4, ATCC BAA-334).
Pneumococci were cultured on blood agar base number 2 (Oxoid, Basingstoke, United Kingdom) plus 5% (vol/vol) horse blood (E&O Laboratories, Bonnybridge, United Kingdom). Strain validation involved the Quellung reaction, sensitivity to optochin, and multilocus sequence typing (performed at the Scottish Meningococcus and Pneumococcus Reference Laboratory, Stobhill Hospital, Glasgow, United Kingdom).
For in vivo studies, pneumococci were passaged intraperitoneally through mice and stored at –80°C. GalU– pneumococci are avirulent in wild-type animals, so a large inoculum (4 x 108 CFU) was passaged through a C3– animal. Bacteria for infections were thawed rapidly, harvested by centrifugation, and resuspended in sterile phosphate-buffered saline (PBS) (Oxoid).
Production of PLY– and GalU– pneumococci. PLY– pneumococci were constructed using plasmid pLC1 that contains the pneumolysin gene in PCR-Script vector (Invitrogen, Paisley, United Kingdom). This plasmid was used for inverse PCR with primers AGGCGCGCCTTACAGCTTACAGAAACGGAGATT and GACGGCGCGCCTGATGCTACTTATCCAAC, which correspond to positions 189 to 201 and 1064 to 1087, respectively, of the nucleotide sequence of the pneumolysin gene. Sites for the restriction enzyme AscI were introduced on the primers (underlined). The product from inverse PCR (containing the 5' and 3' regions of the pneumolysin gene and the pPCR-Script vector) was purified, cut with AscI, and self-ligated, and recombinant plasmid was isolated. This plasmid was linearized with AscI, and an erythromycin resistance cassette with AscI ends was ligated into the construct. The resulting plasmid (pLC2) contained the erythromycin resistance gene flanked on either side by approximately 300 bp of 5' and 3' sequence from the pneumolysin gene and was used to transform D39 as previously described (17). Transformants were selected on 1 μg/ml erythromycin.
GalU (UTP:glucose-1-phosphate uridylyltransferase) is a pneumococcal protein essential for capsule biosynthesis (23). An allelic replacement galU (spr1903) knockout mutant was created using the construct employed previously by Marra and Brigham (22). Transformants in D39 were selected on 500 μg/ml spectinomycin.
PCR and DNA sequencing confirmed disruption of the pneumolysin or galU gene.
Infection of mice. Prior to intranasal infection, mice were anesthetized with halothane (Concord Pharmaceuticals Ltd., Dunmow, United Kingdom). A total of 106 CFU of pneumococci were then administered via the nostrils in a 50-μl volume. For intravenous infection, 2.5 x 103 CFU of D39 were injected via a tail vein.
Signs of disease were monitored frequently until mice were deemed to have irreversibly succumbed to the infection (36), at which point they were humanely sacrificed. Mice displaying no signs of illness for 336 h were considered to have survived the infection.
Sample collection and processing. Following cervical dislocation, cardiac blood was sampled. A 16-gauge nonpyrogenic angiocath (F. Baker Scientific, Runcorn, United Kingdom) was used to lavage lungs with 2 ml PBS. Bronchoalveolar lavage fluid (BALF) for bacteriological investigation was placed on ice, and samples for cytokine analysis were snap-frozen by immersion in liquid nitrogen. Lungs for bacteriological investigation were homogenized in 5 ml PBS using a glass hand-held tissue homogenizer (Jencons, Leighton Buzzard, United Kingdom). Samples for cytokine analysis were snap-frozen and processed as described previously (20).
Viable bacteria in blood, BALF, and lung tissue were counted by plating out serial dilutions on blood agar base number 2 plus 5% (vol/vol) horse blood.
Measurement of C3. C3 was measured, by enzyme-linked immunosorbent assay (ELISA), using a modification of the method described previously by Circolo et al. (9). Goat antiserum to mouse C3 was used as capture antibody with horseradish peroxidase-conjugated goat anti-mouse C3 immunoglobulin G fraction as a detection antibody (both from MP Biomedicals, Aurora, OH). C3 used as the standard was purified from fresh frozen pooled serum from C57BL/6 mice using anion exchange chromatography as described previously by Van den Berg (38).
Measurement of immune modulators. Tumor necrosis factor (TNF) activity was measured by bioassay and interleukin-6 (IL-6) was measured by ELISA as described previously (20). Total protein levels were measured using Bradford reagent (Sigma-Aldrich, Poole, United Kingdom) with bovine serum albumin as the standard.
Histology. Lungs were inflated with 1 ml 4% formaldehyde in PBS and fixed for 24 h. Samples were then embedded in paraffin and blocked, and 5-μm-lung sections were stained with hematoxylin and eosin (BDH Laboratory Supplies, Poole, United Kingdom).
Lung pathology was scored blind on the following criteria (adapted from reference 5): edema, cell influx, tissue disruption, and hemorrhage. A score of 0, 1, or 2 was given depending on whether the pathology was absent, mild, or severe, respectively.
Statistical analysis. Bacteriology results are expressed as mean ± 1 standard error of the mean. Bacterial load data from time course experiments were compared using one-way analysis of variance with Scheffe's post hoc test (not all time points were gathered from an individual experiment). Where samples contained fewer CFU/ml than the lower detection limit for the viable counting assay (log 1.92 per ml blood and log 0.92 per ml BALF or lung homogenate), they were ascribed a value just below the detection limit (log 1.91 or 0.91). Comparisons of bacterial loads between mouse strains or treatments were made with unpaired Student's t tests. Cytokine, C3, histology scores, and protein levels are expressed as median ± median absolute deviation (20). Survival times, cytokine, C3, and protein levels were analyzed using nonparametric Mann Whitney U analysis with Bonferroni correction carried out with cytokine and histology analyses.
Statistical analyses were carried out using StatView, version 4.1 (Abacus Concepts, Berkeley, CA), with P < 0.05 considered statistically significant for all analyses. Symbols used in figures and tables denote statistical significance between C3– and C57BL/6 groups. Differences from 0 h within C3– and C57BL/6 groups are not denoted with symbols but are described in Results.
RESULTS
C3– mice are significantly more susceptible to systemic and pulmonary infection with S. pneumoniae. We monitored outcome of infection to investigate whether C3 provided protection against S. pneumoniae. C3– mice were highly susceptible to intranasal challenge with D39 (Table 1), with 100% mortality recorded by 36 h. In contrast, only five out of eight wild-type C57BL/6 mice succumbed (between 49 and 77 h; P < 0.01 longer than that for C3– mice).
Intravenous infection of C3– mice with D39 pneumococci resulted in survival times of between 24 h and 73 h. Again, these were significantly shorter (P < 0.05) than the survival times of C57BL/6 mice (median, 336 h; 3/5 surviving).
C3– mice have significantly higher bacterial loads within lungs and circulation following intranasal infection with D39 pneumococci. We monitored bacterial loads following intranasal infection to investigate whether C3 was involved in control of pneumococcal viability. Pneumococcus viable counts within the lung airways of C3– mice increased immediately following intranasal infection (Fig. 1A), with levels significantly higher than those at 0 h from 3 h until 12 h (P < 0.05). Counts then decreased between 12 h and 24 h in C3– BALF. There was no significant increase in S. pneumoniae viability within the airways of C57BL/6 mice following intranasal infection. At 6 h and 12 h postchallenge, C57BL/6 bacterial loads were significantly lower than those in C3– group samples (P < 0.01).
Between 0 h and 3 h, pneumococcal counts in lung tissue increased 10-fold in both C3– and C57BL/6 samples (Fig. 1B). Bacterial loads in C3– lung tissue plateaued between 3 h and 12 h postinfection before increasing again during the last 12 h of the experiment. Between 3 h and 6 h, levels of S. pneumoniae decreased in C57BL/6 lung tissue such that they were significantly lower than those in C3– lung tissue (P < 0.01).
Bacteremia was found in C3– mice 3 h postinfection and was significant by 12 h (Fig. 1C) (P < 0.01 higher than that at 0 h). Only at 24 h were significant numbers of pneumococci found in C57BL/6 blood samples (P < 0.01 higher than that at 0 h). At 12 h and 24 h, levels of bacteremia in C3– mice were significantly higher than those in C57BL/6 mice (P < 0.01).
Significant increases in C3 levels within BALF correlate with decreased pneumococcal viability. We used an ELISA to measure levels of C3 in order to investigate the relationship between C3 levels and bacterial loads within the lungs and blood during infection.
C3 was undetectable in BALF, lung homogenate, or serum from C3– mice (P < 0.01 lower than C57BL/6 samples at all times).
Immediately following infection, around 3,400 ng/ml C3 was found within BALF from C57BL/6 mice (Fig. 2A). Levels of C3 were significantly higher than baseline at 1 h (P < 0.05 higher than that at 0 h) and from 12 h until 24 h (P < 0.01 higher than that at 0 h).
Levels of C3 within C57BL/6 lung tissue and in serum (2,400 μg/ml and 4,000 μg/ml, respectively, at 0 h) (Fig. 2B and C) did not alter significantly during infection.
Lungs from C3– mice displayed an elevated inflammatory response during pneumococcal pneumonia. To investigate whether a lack of C3 altered the inflammatory response during pneumococcal pneumonia, we measured TNF activity and IL-6 production.
The airways of C57BL/6 mice did not contain significant levels of TNF activity or IL-6 during the infection (Table 2). C3– airways had significantly elevated levels of TNF activity at 12 h and 24 h compared to that at 0 h (P < 0.05) and to those of C57BL/6 samples (P < 0.05 at 12 h and P < 0.01 at 24 h).
Production of IL-6 within airways of C3– mice was significantly increased from 3 h until 24 h (Table 2) (P < 0.01 higher than that at 0 h) with levels between 10- and 100-fold higher than those in C57BL/6 samples.
Total protein levels were measured to examine disruption of lung integrity during infection (31). Similar amounts of protein were recovered in C3– and C57BL/6 airways immediately after infection (Fig. 3). At 1 h and 3 h, protein levels were significantly increased in both strains of mice (P < 0.01 higher than that at 0 h). Protein levels within C57BL/6 BALF returned to baseline by 6 h and stayed unchanged until the end of the experiment. Disruption to C3– lungs remained throughout the infection (P < 0.05 higher than that at 0 h, 6 h, and 12 h and P < 0.01 higher at 24 h). At 24 h, the amount of protein in C3– BALF was significantly higher than that in C57BL/6 samples (P < 0.05).
As with lung airways, lung homogenates from C57BL/6 mice did not contain significant levels of TNF activity or IL-6 protein at any time postinfection (Table 2). At both 3 h and 24 h, lung homogenates from C3– mice contained levels of TNF activity that were significantly higher than baseline, and at 24 h, levels were higher than those of C57BL/6 samples (P < 0.05).
Low levels of IL-6 were found in C3– lung tissue until 24 h postinfection, when 5,350 pg/ml was detected. This was significantly higher than the 0 pg/ml found in C3– samples at 0 h and C57BL/6 samples at 24 h (P < 0.05).
Complement deficiency does not alter cellular recruitment during pneumococcal pneumonia. We studied lung histology to investigate whether C3 was required for inflammatory cell recruitment during pneumococcal pneumonia.
Sections of lung from both C3– and C57BL/6 mice displayed areas of perivascular neutrophil (as judged by nuclear morphology) recruitment at 1 h and 3 h postchallenge. These areas remained around blood vessels until 12 h. Between 12 h and 24 h, more inflammatory cells were recruited, resulting in perivascular lesions covering large proportions of the affected lobe. Pathology scores were not significantly different between C3– and C57BL/6 mice at any time (Table 3).
To further investigate this apparent lack of requirement for complement in cell recruitment during pneumococcal pneumonia, we infected mice deficient in the receptor for C5a. Lung sections from C5ar– mice were studied at 6 h, 12 h, and 24 h postinfection. As with C3 deficiency, a lack of C5a receptor did not prevent cell influx to the lungs during pneumococcal pneumonia. Perivascular areas of inflammatory cells were evident at all times studied (data not shown).
Requirement for C3 is not limited to serotype 2 pneumococci. Infections with serotype 4 pneumococci were done to investigate whether the necessity for C3 in pulmonary protection against S. pneumoniae is limited to serotype 2 organisms.
C3– mice were significantly more susceptible to intranasal infection with TIGR4 than C57BL/6 mice (median survival time, 28.5 h versus 113.5 h; P < 0.01 shorter for C3–) (Table 1).
Investigation of bacterial loads revealed more viable pneumococci in C3– BALF, lung tissue, and blood at 24 h postinfection than in C57BL/6 samples (Fig. 4) (P < 0.01 higher for C3– in each).
Significantly higher levels of TNF activity and IL-6 were found in BALF from the C3– mice 24 h postinfection than from C57BL/6 mice (TNF, 70 U/ml versus 0 U/ml, respectively [P < 0.05]; IL-6, 2,288 pg/ml versus 26 pg/ml, respectively [P < 0.01]). C3– BALF contained significantly elevated total protein levels at 24 h compared to those at 0 h (289 μg/ml versus 146 μg/ml, respectively [P < 0.05]). Despite higher levels of total protein in BALF from C3– mice than from C57BL/6 mice at 24 h (C57BL/6 samples contained 214 μg/ml), this difference was not significant.
C3 is required for protection against pneumolysin-deficient pneumococci. To better define the interaction of pneumolysin with complement in vivo, we infected C3– mice with pneumolysin-deficient pneumococci.
Despite being avirulent in C57BL/6 mice (all five mice survived intranasal challenge), PLY– pneumococci remained virulent in C3– mice (median survival time, 101 h; P < 0.01 shorter than for C57BL/6 mice) (Table 1).
Numbers of PLY– pneumococci within the airways of both strains of mice were significantly reduced from baseline at 24 h and 48 h (Fig. 1D) (P < 0.01). From 6 h onwards, numbers of PLY– pneumococci were significantly higher within C3– BALF than in C57BL/6 samples (P < 0.01 at 6 h until 24 h; P < 0.05 at 48 h).
Counts of PLY– pneumococci within C3– lung tissue increased following infection, such that by 48 h, significantly more viable PLY– pneumococci were found within lung tissue from C3– mice than was found originally (Fig. 1E) (P < 0.05). From 6 h until 48 h, there were significantly higher bacterial loads in lung tissue from C3– mice than from C57BL/6 mice (P < 0.01).
PLY– pneumococci retained the ability to cause bacteremia in C3– mice but not in C57BL/6 mice (Fig. 1F). At 12 h, 24 h, and 36 h, significantly higher numbers of viable PLY– pneumococci were found in the circulation of C3– mice than at 0 h (P < 0.01). From 12 h until 36 h, these numbers were significantly higher than those found in C57BL/6 samples (P < 0.01).
Pulmonary clearance of avirulent pneumococci is attenuated in C3– mice. To define the effects of a pulmonary deficiency in C3– mice in isolation from the systemic immune response, pneumococci unable to produce the polysaccharide capsule (23) were used.
GalU– pneumococci were avirulent in both C3– and C57BL/6 mice following intranasal infection: all mice survived and none developed bacteremia (Table 1).
GalU– pneumococci were rapidly cleared from lung airways of both mouse strains; this was significant in C57BL/6 mouse airways by 6 h (Fig. 1G) (P < 0.01 lower than that at 0 h) but not until 12 h in C3– airways (P < 0.05 lower than that at 0 h). From 3 h onwards, GalU– pneumococcus viability was significantly higher within C3– BALF than in C57BL/6 (P < 0.05 at 3 h and 12 h; P < 0.01 at 6 h).
Clearance of GalU– pneumococci was quicker in C57BL/6 lung tissue, with a significant reduction in counts by 6 h (Fig. 1H) (P < 0.05), whereas a significant reduction in counts in C3– tissue was not seen until 12 h (P < 0.01 compared to that at 0 h). Bacterial loads were significantly higher in C3– lung tissue than in C57BL/6 samples at 3 h and 6 h (Fig. 1H) (P < 0.05 and P < 0.01, respectively).
Pulmonary clearance of avirulent pneumococci is associated with lung disruption and elevated C3 levels. Protein levels in BALF from GalU– pneumococcus-infected mice were measured to relate lung integrity and C3 levels within the lungs during pneumonia without associated bacteremia. Significant disruption of lung integrity was evident within GalU– pneumococcus-infected C57BL/6 mice by 3 h postinfection (Fig. 3) (P < 0.05 higher than that at 0 h). At this time, around 200 μg/ml protein was measured, compared to 56 μg/ml at 0 h. Levels of protein then decreased but remained significantly higher at 12 h than those at baseline. Lung integrity in C3– mice was not affected until 12 h postinfection (P < 0.05 higher than that at 0 h). At both 1 h and 3 h, significantly higher levels of total protein were recovered from C57BL/6 BALF than from C3– samples (P < 0.01).
C3 levels in BALF from GalU– pneumococcus-infected C57BL/6 mice were detected at around 2,700 ng/ml of C3 immediately following infection. This level was significantly elevated at both 1 h and 12 h (Fig. 2D) (P < 0.01 higher than that at 0 h). Levels of C3 in C3– BALF were significantly lower than those of C57BL/6 samples at all times (P < 0.01).
DISCUSSION
It is well known that complement plays a protective role during pneumococcal bacteremia (8, 16, 40). In this report, we focused on pneumococcal pneumonia and show a requirement for C3 within the lungs during the initial hours following intranasal inoculation of S. pneumoniae.
Mice with a genetic deficiency in C3 were significantly more susceptible to intranasal or intravenous challenge with S. pneumoniae (Table 1), confirming previous results from our laboratory and those of others (6, 7). This effect was not limited to serotype 2 pneumococci but was also found with serotype 4 organisms. It is possible that infection with other serotypes would not result in this effect, although a similar requirement for complement during early infection has been demonstrated by others (24).
Shorter survival times in C3– mice were associated with significantly elevated bacterial loads in the lung airways, lung tissue, and bloodstream within 6 h of infection (Fig. 1). D39 viability was controlled between 3 h and 6 h within C57BL/6 lung tissue but not in C3– lung tissue. As neutrophil recruitment is not different between the two mouse strains, it is likely that C3 is acting as an opsonin for resident alveolar macrophages in C57BL/6 mice during this time.
Previously, cobra venom factor-treated mice were found to have elevated pulmonary bacterial loads within hours of infection (14). Elevated bacteremia in C3– mice may be due to this pulmonary deficiency allowing increased numbers of pneumococci to enter the circulation or may be due to the bacteria being able to grow more quickly in the bloodstream of C3– mice.
Complement activation by pneumolysin was previously found to be involved in pulmonary pneumococcal growth (30) and to promote pneumococcal sepsis in cirrhotic rats (where complement levels are limiting [1]). We used pneumolysin-deficient bacteria (PLY–) to investigate whether complement activation by pneumolysin is the main action of the toxin during pneumonia. No difference between D39 and PLY– pneumococcus counts in C3– mice would indicate that complement activation by pneumolysin is the toxin's major role in virulence during pneumococcal pneumonia (complement activation by pneumolysin is already redundant in C3– mice, so infection with PLY– pneumococci should have no effect). This was not found to be the case. Infection with PLY– pneumococci resulted in longer survival times and lower bacterial loads than with D39 (Table 1 and Fig. 1). Thus, another function of pneumolysin is involved in virulence in our model.
Acapsular GalU-deficient S. pneumoniae isolates were utilized to focus on the importance of C3 within the lungs in the absence of bacteremia, as GalU– pneumococci were undetectable within the circulation of both C3– and C57BL/6 mice following intranasal inoculation. GalU– pneumococcus viability within lung airways declined immediately in C57BL/6 mice but reached significantly higher counts in C3– mice (Fig. 1G). This suggests that even when pneumococci lack a capsule, they can survive more successfully in the lungs of C3– mice. Therefore, during pneumonia without associated bacteremia, pulmonary C3 is crucial for control of the bacteria.
One hour following infection with either D39 or GalU– pneumococci, C3 levels in BALF, as measured by ELISA, increased (Fig. 2A and D), corresponding to a decrease in pneumococcal viability (Fig. 1A and G). C3 levels were then transiently depleted. It was not until bacterial loads were brought under control at later time points that an excess of C3 was present. We believe that these experiments are the first to report an association between C3 levels and bacterial loads during pneumococcal pneumonia. That serum C3 levels did not alter significantly during the infection agrees with previous data following intravenous infection with S. pneumoniae (1) but contrasts with those reported previously by Tu et al. (37). Discrepancies may reflect differences in the bacterial and mouse strains used.
C3 may be directly antibacterial (26), or the elevated C3 levels may be a marker of other antibacterial serum proteins leaking into the lungs.
As previously reported (41), C3 levels are lower in C57BL BALF than in serum (around 1,000-fold). The ability of pneumococci to activate the complement pathway (27) and deplete complement levels (2) should be more effective against the lower pulmonary level of C3 than the higher systemic levels. This may explain why C3 levels correlate with bacterial loads in the lungs but not in the circulation.
We measured total protein levels to discover whether C3 within BALF was locally produced or a serum component had leaked into the lungs (31) (Fig. 3). In C57BL/6 mice, C3 levels increased significantly within 1 h of intranasal infection with either D39 or GalU– pneumococci. At the same time, total protein levels also increased. Such kinetics indicate that the C3 within C57BL/6 airways is not locally produced but comes from the circulation.
We also found that a deficiency in C3 resulted in an elevated inflammatory response during infection (Table 2). Significantly elevated levels of TNF activity and IL-6 protein were present within C3– BALF and lung tissue during late infection.
There are two explanations for this. The elevated bacterial loads within C3– samples may act as a greater inflammatory stimulus. However, at 24 h, bacterial counts in the lungs of the two strains of mice were similar, while cytokine levels were significantly higher in C3– samples (Fig. 1A and B and Table 2). The alternative explanation is that C3 plays an anti-inflammatory role during pneumococcal infection. C3 has previously been found to decrease TNF- production by human peripheral blood mononuclear cells stimulated in vitro with lipopolysaccharide (34). In addition, C3a receptor-deficient mice produce more IL-1 following intravenous lipopolysaccharide administration than wild-type mice (21).
To investigate which cell types act with C3 in the lungs, we looked at pulmonary histology. Neutrophil recruitment to lungs was found within 1 h of intranasal infection with S. pneumoniae, with numbers increasing with time. Inflammatory cell recruitment was not attenuated in C3– mouse lungs (Table 3). By infecting C5ar– mice, we confirmed that complement was not required for phagocyte recruitment. We found that the large areas of perivascular inflammatory cell recruitment seen in wild-type and C3– mice were still evident in the absence of a C5a receptor. In a previous investigation, significant neutrophil recruitment was shown to occur in the lungs of C5-deficient mice (35). As neither C3 nor C5 is required for cell recruitment during pneumococcal pneumonia, the main role of complement during pneumococcal pneumonia must be in opsonophagocytosis (40) or exertion of direct antibacterial effects (26).
Overall, we have shown that C3 is essential within the lungs for an optimal immune response against S. pneumoniae, even when mice are not bacteremic. Without it, pneumococci reach higher bacterial loads, induce an elevated inflammatory response, and result in shorter survival times.
ACKNOWLEDGMENTS
Thanks go to Michael C. Carroll (Harvard Medical School, Boston, Mass.) for permission to use C3– mice and to James E. Marsh (Guys Hospital, London, United Kingdom) for providing them. Thanks to Craig Gerard (Harvard Medical School, Boston, Mass.) for supplying the C5ar– mice. Thanks to Lyndsey Hall for production of the pneumolysin-deficient bacteria.
The Wellcome Trust funded this work.
REFERENCES
1. Alcantara, R. B., L. C. Preheim, and M. J. Gentry. 1999. Role of pneumolysin's complement-activating activity during pneumococcal bacteremia in cirrhotic rats. Infect. Immun. 67:2862-2866.
2. Alcantara, R. B., L. C. Preheim, and M. J. Gentry-Nielsen. 2001. Pneumolysin-induced complement depletion during experimental pneumococcal bacteremia. Infect. Immun. 69:3569-3575.
3. Alper, C. A., N. Abramson, R. B. J. Johnston, J. H. Jandl, and F. S. Rosen. 1970. Increased susceptibility to infection associated with abnormalities of complement-mediated functions and of the third component of complement (C3). N. Engl. J. Med. 282:350-354.
4. Angel, C. S., M. Ruzek, and M. K. Hostetter. 1994. Degradation of C3 by Streptococcus pneumoniae. J. Infect. Dis. 170:600-608.
5. Bergeron, Y., N. Ouellet, A.-M. Deslauriers, M. Simard, M. Olivier, and M. G. Bergeron. 1998. Cytokine kinetics and other host factors in response to pneumococcal pulmonary infection in mice. Infect. Immun. 66:912-922.
6. Blue, C. E., G. K. Paterson, A. R. Kerr, M. Berge, J.-P. Claverys, and T. J. Mitchell. 2003. ZmpB, a novel virulence factor of Streptococcus pneumoniae that induces tumor necrosis factor alpha production in the respiratory tract. Infect. Immun. 71:4925-4935.
7. Brown, J. S., T. Hussell, S. M. Gilliland, D. W. Holden, J. C. Paton, M. R. Ehrenstein, M. J. Walport, and M. Botto. 2002. The classical pathway is the dominant complement pathway required for innate immunity to Streptococcus pneumoniae infection in mice. Proc. Natl. Acad. Sci. USA 99:16969-16974.
8. Circolo, A., G. Garnier, W. Fukuda, X. Wang, T. Hidvegi, A. J. Szalai, D. E. Briles, J. E. Volanakis, R. A. Wetsel, and H. R. Colten. 1999. Genetic disruption of the murine complement C3 promoter region generates deficient mice with extrahepatic expression of C3 mRNA. Immunopharmacology 42:135-149.
9. Circolo, A., T. B. Nutter III, and R. C. Strunk. 1997. Biosynthesis of complement components, p. 199-221. In A. W. Dodds and R. B. Sim (ed.), Complement: a practical approach. Oxford University Press, Oxford, United Kingdom.
10. Cole, F. S., W. J. Matthews, Jr., T. H. Rossing, D. J. Gash, N. A. Lichtenberg, and J. E. Pennington. 1983. Complement biosynthesis by human bronchoalveolar macrophages. Clin. Immunol. Immunopathol. 27:153-159.
11. Dave, S., A. Brooks-Walter, M. K. Pangburn, and L. S. McDaniel. 2001. PspC, a pneumococcal surface protein, binds human factor H. Infect. Immun. 69:3435-3437.
12. Dave, S., S. Carmicle, S. Hammerschmidt, M. K. Pangburn, and L. S. McDaniel. 2004. Dual roles of PspC, a surface protein of Streptococcus pneumoniae, in binding human secretory IgA and factor H. J. Immunol. 173:471-477.
13. Figueroa, J. E., and P. Densen. 1991. Infectious diseases associated with complement deficiencies. Clin. Microbiol. Rev. 4:359-395.
14. Gross, G. N., S. R. Rehm, and A. K. Pierce. 1978. The effect of complement depletion on lung clearance of bacteria. J. Clin. Investig. 62:373-378.
15. Hopken, U. E., B. Lu, N. P. Gerard, and C. Gerard. 1996. The C5a chemoattractant receptor mediates mucosal defence to infection. Nature 383:86-89.
16. Hosea, S. W., E. J. Brown, and M. M. Frank. 1980. The critical role of complement in experimental pneumococcal sepsis. J. Infect. Dis. 142:903-909.
17. Ibrahim, Y. M., A. R. Kerr, J. McCluskey, and T. J. Mitchell. 2004. Role of HtrA in the virulence and competence of Streptococcus pneumoniae. Infect. Immun. 72:3584-3591.
18. Janulczyk, R., F. Iannelli, A. G. Sjoholm, G. Pozzi, and L. Bjork. 2000. Hic, a novel surface protein of Streptococcus pneumoniae that interferes with complement function. J. Biol. Chem. 275:37257-37263.
19. Kemper, C. A., and S. C. Deresinki. 1994. The immunology of pneumococcal pneumonia, p. 29-49. In H. Chmel, M. Bendinelli, and H. Friedman (ed.), Pulmonary infections and immunity. Plenum Press, New York, N.Y.
20. Kerr, A. R., J. J. Irvine, J. J. Search, N. A. Gingles, A. Kadioglu, P. W. Andrew, W. L. McPheat, C. G. Booth, and T. J. Mitchell. 2002. The role of inflammatory mediators in resistance and susceptibility to pneumococcal infection. Infect. Immun. 70:1547-1557.
21. Kildsgaard, J., T. J. Hollman, K. W. Mathews, K. Bian, F. Murad, and R. A. Wetsel. 2000. Targeted disruption of the C3a receptor gene demonstrates a novel protective anti-inflammatory role for C3a in endotoxin-shock. J. Immunol. 165:5406-5409.
22. Marra, A., and D. Brigham. 2001. Streptococcus pneumoniae causes experimental meningitis following intranasal and otitis media infections via a nonhematogenous route. Infect. Immun. 69:7318-7325.
23. Mollerach, M., R. Lopez, and E. García. 1998. Characterization of the galU gene of Streptococcus pneumoniae encoding a uridine diphosphoglucose pyrophosphorylase: a gene essential for capsular polysaccharide biosynthesis. J. Exp. Med. 188:2047-2056.
24. Nakajima, R., K. Namba, Y. Ishida, T. Une, and Y. Osada. 1990. Protective role of complement in the development of experimental pneumococcal pneumonia in mice. Chemotherapy 36:287-293.
25. Newman, S. L., L. B. Vogler, R. D. Feigin, and R. B. J. Johnston. 1978. Recurrent septicemia associated with congenital deficiency of C2 and partial deficiency of factor B and the alternative complement pathway. N. Engl. J. Med. 299:290-292.
26. Nordahl, E. M., V. Rydengrd, P. Nyberg, D. P. Nitsche, M. Mrgelin, M. Malmsten, L. Bjrk, and A. Schmidtchen. 2004. Activation of the complement system generates antibacterial peptides. Proc. Natl. Acad. Sci. USA 101:16879-16884.
27. Paton, J. C., B. Rowan-Kelly, and A. Ferrante. 1984. Activation of human complement by the pneumococcal toxin pneumolysin. Infect. Immun. 43:1085-1087.
28. Ren, B., A. J. Szalai, S. K. Hollingshead, and D. E. Briles. 2004. Effects of PspA and antibodies to PspA on activation and deposition of complement on the pneumococcal surface. Infect. Immun. 72:114-122.
29. Roy, S., K. Knox, S. Segal, D. Griffiths, C. E. Moore, K. I. Welsh, A. Smarason, N. P. Day, W. L. McPheat, D. W. Crook, and A. V. S. Hill. 2002. MBL genotype and risk of invasive pneumococcal disease: a case-control study. Lancet 359:1569-1573.
30. Rubins, J. B., D. Charboneau, C. Fasching, A. M. Berry, J. C. Paton, J. E. Alexander, and P. W. Andrew. 1996. Distinct roles for pneumolysin's cytotoxic and complement activities in the pathogenesis of pneumococcal pneumonia. Am. J. Respir. Crit. Care Med. 153:1339-1346.
31. Rubins, J. B., D. Charboneau, J. C. Paton, T. J. Mitchell, P. W. Andrew, and E. N. Janoff. 1995. Dual function of pneumolysin in the early pathogenesis of murine pneumococcal pneumonia. J. Clin. Investig. 95:142-150.
32. Stockley, R. A. 1997. Soluble proteins in lung defence, p. 18-38. In R. A. Stockley (ed.), Pulmonary defences. John Wiley & Sons Ltd., Chichester, United Kingdom.
33. Strunk, R. C., D. M. Eidlen, and R. J. Mason. 1988. Pulmonary alveolar type II epithelial cells synthesize and secrete proteins of the classical and alternative complement pathways. J. Clin. Investig. 81:1419-1426.
34. Takabayashi, T., E. Vannier, B. D. Clark, N. H. Margolis, C. A. Dinarello, J. F. Burke, and J. A. Gelfand. 1996. A new biologic role for C3a and C3a desArg. J. Immunol. 156:3455-3460.
35. Toews, G. B., and W. C. Vial. 1984. The role of C5 in polymorphonuclear leukocyte recruitment in response to Streptococcus pneumoniae. Am. Rev. Respir. Dis. 129:82-86.
36. Toth, L. A. 1997. The moribund state as an experimental endpoint. Contemp. Top. Lab. Anim. Sci. 36:44-48.
37. Tu, A. T., R. L. Fulgham, M. A. McCrory, D. E. Briles, and A. J. Szalai. 1999. Pneumococcal surface protein A inhibits complement activation by Streptococcus pneumoniae. Infect. Immun. 67:4720-4724.
38. Van den Berg, C. W., H. Van Dijk, and P. J. A. Capel. 1989. Rapid isolation and characterisation of native mouse complement components C3 and C5. J. Immunol. Methods 122:73-78.
39. Wessels, M. R., P. Butko, M. Ma, H. B. Warren, A. L. Lage, and M. C. Carroll. 1995. Studies of group B streptococcal infection in mice deficient in complement component C3 or C4 demonstrate an essential role for complement in both innate and acquired immunity. Proc. Natl. Acad. Sci. USA 92:11490-11494.
40. Winkelstein, J., M. Smith, and H. Shin. 1975. The role of C3 as an opsonin in the early stages of infection. Proc. Soc. Exp. Biol. Med. 149:397-401.
41. Younger, J. G., S. Shankar-Sinha, M. Mickiewicz, A. S. Brinkman, G. A. Valencia, J. V. Sarma, E. M. Younkin, T. J. Standiford, F. S. Zetoune, and P. A. Ward. 2003. Murine complement interactions with Pseudomonas aeruginosa and their consequences during pneumonia. Am. J. Respir. Cell Mol. Biol. 29:432-438.
42. Zhang, Y., A. W. Masi, V. Barniak, K. Mountzouros, M. K. Hostetter, and B. A. Green. 2001. Recombinant PhpA protein, a unique histidine motif-containing protein from Streptococcus pneumoniae, protects mice against intranasal pneumococcal challenge. Infect. Immun. 69:3827-3836.(Alison R. Kerr, Gavin K. )
ABSTRACT
Complement is known to be involved in protection against systemic infection with Streptococcus pneumoniae. However, less is known about effects of complement within the lungs during pneumococcal pneumonia. By intranasally infecting transgenic mice unable to express complement C3, we investigated the role of complement in pulmonary defenses against S. pneumoniae. It was demonstrated that within the lungs, there is a requirement for C3 during the initial hours of infection. It was found that within 1 h of infection, bacterial loads decreased within lung airways of control mice as C3 protein increased. The lack of C3 resulted in the inability to control growth of wild-type or attenuated pneumococci within the lungs and bloodstream, resulting in an overwhelming inflammatory response and shorter survival times. Our results show that during the initial hours of infection with S. pneumoniae, C3 is protective within the lungs and subsequently plays an important role systemically.
INTRODUCTION
Streptococcus pneumoniae (the pneumococcus) is an important pathogen of humans. The organism causes diseases such as pneumonia, meningitis, bacteremia, and otitis media. Several problems are associated with the control of pneumococci, including a high morbidity rate despite the availability of antibiotics; increasing levels of antibiotic resistance; and limited efficacy of vaccines (23 valent polysaccharide vaccine), high cost, and limited serotype coverage (conjugate vaccines). These facts emphasize the requirement for greater understanding of a protective immune response against the pneumococcus.
A requirement for complement in protection against S. pneumoniae is highlighted by the fact that humans with a deficiency in any of the three complement pathways display heightened sensitivity to pneumococcal infection (3, 13, 25, 29).
S. pneumoniae inhibits the complement pathway in several ways. As a gram-positive bacterium, the pneumococcus is resistant to the bactericidal and lytic activities of complement (19). Pneumolysin can stimulate the classical complement pathway in the absence of pneumolysin-specific antibody, diverting the inflammatory response from intact pneumococci (27). Pneumococcal surface protein A (PspA) prevents the deposition of C3b onto the surface of pneumococci and interferes with both the classical and alternate complement pathways (28, 37). PspC and the homologous protein Hic can both bind complement factor H (11, 18), which inhibits the alternative pathway C3 convertase and protects host cells from complement attack. Factor H bound to pneumococcal PspC should retain the ability to bind C3b and inhibit complement activation (12). The pneumococcal surface protein PhtB (also known as PhpA and BVH-11) has the ability to degrade C3 (4, 42).
The liver is thought to be the major source of complement, with most complement components in the lung being derived from plasma during inflammation (32). It has therefore been presumed that complement does not play a major role in antipneumococcal defense within the lungs during the early stages of pneumonia but rather becomes involved when the host is bacteremic. However, production of complement also occurs within the lungs. Both pulmonary macrophages and epithelial cells can synthesize and secrete C2, C4, C3, C5, and factor B in vitro (10, 33).
We have investigated the importance of complement within the lungs in vivo by inducing pneumococcal pneumonia in transgenic mice unable to express C3.
MATERIALS AND METHODS
Mice. Male and female C57BL/6 mice (20 to 30g; Harlan Olac, Bicester, United Kingdom), C3-deficient mice (C3– [39]), and C5a receptor-deficient mice (C5ar– [15]) (both bred in-house and 20 to 30g) were used in these studies. All experiments were carried out in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986.
Bacteria. The pneumococcal strains used were serotype 2 (D39, NCTC 7466; Central Public Health Laboratory, London, United Kingdom), D39 unable to express pneumolysin (PLY– [see below]), acapsular serotype 2 (GalU– [see below]), and serotype 4 (TIGR4, ATCC BAA-334).
Pneumococci were cultured on blood agar base number 2 (Oxoid, Basingstoke, United Kingdom) plus 5% (vol/vol) horse blood (E&O Laboratories, Bonnybridge, United Kingdom). Strain validation involved the Quellung reaction, sensitivity to optochin, and multilocus sequence typing (performed at the Scottish Meningococcus and Pneumococcus Reference Laboratory, Stobhill Hospital, Glasgow, United Kingdom).
For in vivo studies, pneumococci were passaged intraperitoneally through mice and stored at –80°C. GalU– pneumococci are avirulent in wild-type animals, so a large inoculum (4 x 108 CFU) was passaged through a C3– animal. Bacteria for infections were thawed rapidly, harvested by centrifugation, and resuspended in sterile phosphate-buffered saline (PBS) (Oxoid).
Production of PLY– and GalU– pneumococci. PLY– pneumococci were constructed using plasmid pLC1 that contains the pneumolysin gene in PCR-Script vector (Invitrogen, Paisley, United Kingdom). This plasmid was used for inverse PCR with primers AGGCGCGCCTTACAGCTTACAGAAACGGAGATT and GACGGCGCGCCTGATGCTACTTATCCAAC, which correspond to positions 189 to 201 and 1064 to 1087, respectively, of the nucleotide sequence of the pneumolysin gene. Sites for the restriction enzyme AscI were introduced on the primers (underlined). The product from inverse PCR (containing the 5' and 3' regions of the pneumolysin gene and the pPCR-Script vector) was purified, cut with AscI, and self-ligated, and recombinant plasmid was isolated. This plasmid was linearized with AscI, and an erythromycin resistance cassette with AscI ends was ligated into the construct. The resulting plasmid (pLC2) contained the erythromycin resistance gene flanked on either side by approximately 300 bp of 5' and 3' sequence from the pneumolysin gene and was used to transform D39 as previously described (17). Transformants were selected on 1 μg/ml erythromycin.
GalU (UTP:glucose-1-phosphate uridylyltransferase) is a pneumococcal protein essential for capsule biosynthesis (23). An allelic replacement galU (spr1903) knockout mutant was created using the construct employed previously by Marra and Brigham (22). Transformants in D39 were selected on 500 μg/ml spectinomycin.
PCR and DNA sequencing confirmed disruption of the pneumolysin or galU gene.
Infection of mice. Prior to intranasal infection, mice were anesthetized with halothane (Concord Pharmaceuticals Ltd., Dunmow, United Kingdom). A total of 106 CFU of pneumococci were then administered via the nostrils in a 50-μl volume. For intravenous infection, 2.5 x 103 CFU of D39 were injected via a tail vein.
Signs of disease were monitored frequently until mice were deemed to have irreversibly succumbed to the infection (36), at which point they were humanely sacrificed. Mice displaying no signs of illness for 336 h were considered to have survived the infection.
Sample collection and processing. Following cervical dislocation, cardiac blood was sampled. A 16-gauge nonpyrogenic angiocath (F. Baker Scientific, Runcorn, United Kingdom) was used to lavage lungs with 2 ml PBS. Bronchoalveolar lavage fluid (BALF) for bacteriological investigation was placed on ice, and samples for cytokine analysis were snap-frozen by immersion in liquid nitrogen. Lungs for bacteriological investigation were homogenized in 5 ml PBS using a glass hand-held tissue homogenizer (Jencons, Leighton Buzzard, United Kingdom). Samples for cytokine analysis were snap-frozen and processed as described previously (20).
Viable bacteria in blood, BALF, and lung tissue were counted by plating out serial dilutions on blood agar base number 2 plus 5% (vol/vol) horse blood.
Measurement of C3. C3 was measured, by enzyme-linked immunosorbent assay (ELISA), using a modification of the method described previously by Circolo et al. (9). Goat antiserum to mouse C3 was used as capture antibody with horseradish peroxidase-conjugated goat anti-mouse C3 immunoglobulin G fraction as a detection antibody (both from MP Biomedicals, Aurora, OH). C3 used as the standard was purified from fresh frozen pooled serum from C57BL/6 mice using anion exchange chromatography as described previously by Van den Berg (38).
Measurement of immune modulators. Tumor necrosis factor (TNF) activity was measured by bioassay and interleukin-6 (IL-6) was measured by ELISA as described previously (20). Total protein levels were measured using Bradford reagent (Sigma-Aldrich, Poole, United Kingdom) with bovine serum albumin as the standard.
Histology. Lungs were inflated with 1 ml 4% formaldehyde in PBS and fixed for 24 h. Samples were then embedded in paraffin and blocked, and 5-μm-lung sections were stained with hematoxylin and eosin (BDH Laboratory Supplies, Poole, United Kingdom).
Lung pathology was scored blind on the following criteria (adapted from reference 5): edema, cell influx, tissue disruption, and hemorrhage. A score of 0, 1, or 2 was given depending on whether the pathology was absent, mild, or severe, respectively.
Statistical analysis. Bacteriology results are expressed as mean ± 1 standard error of the mean. Bacterial load data from time course experiments were compared using one-way analysis of variance with Scheffe's post hoc test (not all time points were gathered from an individual experiment). Where samples contained fewer CFU/ml than the lower detection limit for the viable counting assay (log 1.92 per ml blood and log 0.92 per ml BALF or lung homogenate), they were ascribed a value just below the detection limit (log 1.91 or 0.91). Comparisons of bacterial loads between mouse strains or treatments were made with unpaired Student's t tests. Cytokine, C3, histology scores, and protein levels are expressed as median ± median absolute deviation (20). Survival times, cytokine, C3, and protein levels were analyzed using nonparametric Mann Whitney U analysis with Bonferroni correction carried out with cytokine and histology analyses.
Statistical analyses were carried out using StatView, version 4.1 (Abacus Concepts, Berkeley, CA), with P < 0.05 considered statistically significant for all analyses. Symbols used in figures and tables denote statistical significance between C3– and C57BL/6 groups. Differences from 0 h within C3– and C57BL/6 groups are not denoted with symbols but are described in Results.
RESULTS
C3– mice are significantly more susceptible to systemic and pulmonary infection with S. pneumoniae. We monitored outcome of infection to investigate whether C3 provided protection against S. pneumoniae. C3– mice were highly susceptible to intranasal challenge with D39 (Table 1), with 100% mortality recorded by 36 h. In contrast, only five out of eight wild-type C57BL/6 mice succumbed (between 49 and 77 h; P < 0.01 longer than that for C3– mice).
Intravenous infection of C3– mice with D39 pneumococci resulted in survival times of between 24 h and 73 h. Again, these were significantly shorter (P < 0.05) than the survival times of C57BL/6 mice (median, 336 h; 3/5 surviving).
C3– mice have significantly higher bacterial loads within lungs and circulation following intranasal infection with D39 pneumococci. We monitored bacterial loads following intranasal infection to investigate whether C3 was involved in control of pneumococcal viability. Pneumococcus viable counts within the lung airways of C3– mice increased immediately following intranasal infection (Fig. 1A), with levels significantly higher than those at 0 h from 3 h until 12 h (P < 0.05). Counts then decreased between 12 h and 24 h in C3– BALF. There was no significant increase in S. pneumoniae viability within the airways of C57BL/6 mice following intranasal infection. At 6 h and 12 h postchallenge, C57BL/6 bacterial loads were significantly lower than those in C3– group samples (P < 0.01).
Between 0 h and 3 h, pneumococcal counts in lung tissue increased 10-fold in both C3– and C57BL/6 samples (Fig. 1B). Bacterial loads in C3– lung tissue plateaued between 3 h and 12 h postinfection before increasing again during the last 12 h of the experiment. Between 3 h and 6 h, levels of S. pneumoniae decreased in C57BL/6 lung tissue such that they were significantly lower than those in C3– lung tissue (P < 0.01).
Bacteremia was found in C3– mice 3 h postinfection and was significant by 12 h (Fig. 1C) (P < 0.01 higher than that at 0 h). Only at 24 h were significant numbers of pneumococci found in C57BL/6 blood samples (P < 0.01 higher than that at 0 h). At 12 h and 24 h, levels of bacteremia in C3– mice were significantly higher than those in C57BL/6 mice (P < 0.01).
Significant increases in C3 levels within BALF correlate with decreased pneumococcal viability. We used an ELISA to measure levels of C3 in order to investigate the relationship between C3 levels and bacterial loads within the lungs and blood during infection.
C3 was undetectable in BALF, lung homogenate, or serum from C3– mice (P < 0.01 lower than C57BL/6 samples at all times).
Immediately following infection, around 3,400 ng/ml C3 was found within BALF from C57BL/6 mice (Fig. 2A). Levels of C3 were significantly higher than baseline at 1 h (P < 0.05 higher than that at 0 h) and from 12 h until 24 h (P < 0.01 higher than that at 0 h).
Levels of C3 within C57BL/6 lung tissue and in serum (2,400 μg/ml and 4,000 μg/ml, respectively, at 0 h) (Fig. 2B and C) did not alter significantly during infection.
Lungs from C3– mice displayed an elevated inflammatory response during pneumococcal pneumonia. To investigate whether a lack of C3 altered the inflammatory response during pneumococcal pneumonia, we measured TNF activity and IL-6 production.
The airways of C57BL/6 mice did not contain significant levels of TNF activity or IL-6 during the infection (Table 2). C3– airways had significantly elevated levels of TNF activity at 12 h and 24 h compared to that at 0 h (P < 0.05) and to those of C57BL/6 samples (P < 0.05 at 12 h and P < 0.01 at 24 h).
Production of IL-6 within airways of C3– mice was significantly increased from 3 h until 24 h (Table 2) (P < 0.01 higher than that at 0 h) with levels between 10- and 100-fold higher than those in C57BL/6 samples.
Total protein levels were measured to examine disruption of lung integrity during infection (31). Similar amounts of protein were recovered in C3– and C57BL/6 airways immediately after infection (Fig. 3). At 1 h and 3 h, protein levels were significantly increased in both strains of mice (P < 0.01 higher than that at 0 h). Protein levels within C57BL/6 BALF returned to baseline by 6 h and stayed unchanged until the end of the experiment. Disruption to C3– lungs remained throughout the infection (P < 0.05 higher than that at 0 h, 6 h, and 12 h and P < 0.01 higher at 24 h). At 24 h, the amount of protein in C3– BALF was significantly higher than that in C57BL/6 samples (P < 0.05).
As with lung airways, lung homogenates from C57BL/6 mice did not contain significant levels of TNF activity or IL-6 protein at any time postinfection (Table 2). At both 3 h and 24 h, lung homogenates from C3– mice contained levels of TNF activity that were significantly higher than baseline, and at 24 h, levels were higher than those of C57BL/6 samples (P < 0.05).
Low levels of IL-6 were found in C3– lung tissue until 24 h postinfection, when 5,350 pg/ml was detected. This was significantly higher than the 0 pg/ml found in C3– samples at 0 h and C57BL/6 samples at 24 h (P < 0.05).
Complement deficiency does not alter cellular recruitment during pneumococcal pneumonia. We studied lung histology to investigate whether C3 was required for inflammatory cell recruitment during pneumococcal pneumonia.
Sections of lung from both C3– and C57BL/6 mice displayed areas of perivascular neutrophil (as judged by nuclear morphology) recruitment at 1 h and 3 h postchallenge. These areas remained around blood vessels until 12 h. Between 12 h and 24 h, more inflammatory cells were recruited, resulting in perivascular lesions covering large proportions of the affected lobe. Pathology scores were not significantly different between C3– and C57BL/6 mice at any time (Table 3).
To further investigate this apparent lack of requirement for complement in cell recruitment during pneumococcal pneumonia, we infected mice deficient in the receptor for C5a. Lung sections from C5ar– mice were studied at 6 h, 12 h, and 24 h postinfection. As with C3 deficiency, a lack of C5a receptor did not prevent cell influx to the lungs during pneumococcal pneumonia. Perivascular areas of inflammatory cells were evident at all times studied (data not shown).
Requirement for C3 is not limited to serotype 2 pneumococci. Infections with serotype 4 pneumococci were done to investigate whether the necessity for C3 in pulmonary protection against S. pneumoniae is limited to serotype 2 organisms.
C3– mice were significantly more susceptible to intranasal infection with TIGR4 than C57BL/6 mice (median survival time, 28.5 h versus 113.5 h; P < 0.01 shorter for C3–) (Table 1).
Investigation of bacterial loads revealed more viable pneumococci in C3– BALF, lung tissue, and blood at 24 h postinfection than in C57BL/6 samples (Fig. 4) (P < 0.01 higher for C3– in each).
Significantly higher levels of TNF activity and IL-6 were found in BALF from the C3– mice 24 h postinfection than from C57BL/6 mice (TNF, 70 U/ml versus 0 U/ml, respectively [P < 0.05]; IL-6, 2,288 pg/ml versus 26 pg/ml, respectively [P < 0.01]). C3– BALF contained significantly elevated total protein levels at 24 h compared to those at 0 h (289 μg/ml versus 146 μg/ml, respectively [P < 0.05]). Despite higher levels of total protein in BALF from C3– mice than from C57BL/6 mice at 24 h (C57BL/6 samples contained 214 μg/ml), this difference was not significant.
C3 is required for protection against pneumolysin-deficient pneumococci. To better define the interaction of pneumolysin with complement in vivo, we infected C3– mice with pneumolysin-deficient pneumococci.
Despite being avirulent in C57BL/6 mice (all five mice survived intranasal challenge), PLY– pneumococci remained virulent in C3– mice (median survival time, 101 h; P < 0.01 shorter than for C57BL/6 mice) (Table 1).
Numbers of PLY– pneumococci within the airways of both strains of mice were significantly reduced from baseline at 24 h and 48 h (Fig. 1D) (P < 0.01). From 6 h onwards, numbers of PLY– pneumococci were significantly higher within C3– BALF than in C57BL/6 samples (P < 0.01 at 6 h until 24 h; P < 0.05 at 48 h).
Counts of PLY– pneumococci within C3– lung tissue increased following infection, such that by 48 h, significantly more viable PLY– pneumococci were found within lung tissue from C3– mice than was found originally (Fig. 1E) (P < 0.05). From 6 h until 48 h, there were significantly higher bacterial loads in lung tissue from C3– mice than from C57BL/6 mice (P < 0.01).
PLY– pneumococci retained the ability to cause bacteremia in C3– mice but not in C57BL/6 mice (Fig. 1F). At 12 h, 24 h, and 36 h, significantly higher numbers of viable PLY– pneumococci were found in the circulation of C3– mice than at 0 h (P < 0.01). From 12 h until 36 h, these numbers were significantly higher than those found in C57BL/6 samples (P < 0.01).
Pulmonary clearance of avirulent pneumococci is attenuated in C3– mice. To define the effects of a pulmonary deficiency in C3– mice in isolation from the systemic immune response, pneumococci unable to produce the polysaccharide capsule (23) were used.
GalU– pneumococci were avirulent in both C3– and C57BL/6 mice following intranasal infection: all mice survived and none developed bacteremia (Table 1).
GalU– pneumococci were rapidly cleared from lung airways of both mouse strains; this was significant in C57BL/6 mouse airways by 6 h (Fig. 1G) (P < 0.01 lower than that at 0 h) but not until 12 h in C3– airways (P < 0.05 lower than that at 0 h). From 3 h onwards, GalU– pneumococcus viability was significantly higher within C3– BALF than in C57BL/6 (P < 0.05 at 3 h and 12 h; P < 0.01 at 6 h).
Clearance of GalU– pneumococci was quicker in C57BL/6 lung tissue, with a significant reduction in counts by 6 h (Fig. 1H) (P < 0.05), whereas a significant reduction in counts in C3– tissue was not seen until 12 h (P < 0.01 compared to that at 0 h). Bacterial loads were significantly higher in C3– lung tissue than in C57BL/6 samples at 3 h and 6 h (Fig. 1H) (P < 0.05 and P < 0.01, respectively).
Pulmonary clearance of avirulent pneumococci is associated with lung disruption and elevated C3 levels. Protein levels in BALF from GalU– pneumococcus-infected mice were measured to relate lung integrity and C3 levels within the lungs during pneumonia without associated bacteremia. Significant disruption of lung integrity was evident within GalU– pneumococcus-infected C57BL/6 mice by 3 h postinfection (Fig. 3) (P < 0.05 higher than that at 0 h). At this time, around 200 μg/ml protein was measured, compared to 56 μg/ml at 0 h. Levels of protein then decreased but remained significantly higher at 12 h than those at baseline. Lung integrity in C3– mice was not affected until 12 h postinfection (P < 0.05 higher than that at 0 h). At both 1 h and 3 h, significantly higher levels of total protein were recovered from C57BL/6 BALF than from C3– samples (P < 0.01).
C3 levels in BALF from GalU– pneumococcus-infected C57BL/6 mice were detected at around 2,700 ng/ml of C3 immediately following infection. This level was significantly elevated at both 1 h and 12 h (Fig. 2D) (P < 0.01 higher than that at 0 h). Levels of C3 in C3– BALF were significantly lower than those of C57BL/6 samples at all times (P < 0.01).
DISCUSSION
It is well known that complement plays a protective role during pneumococcal bacteremia (8, 16, 40). In this report, we focused on pneumococcal pneumonia and show a requirement for C3 within the lungs during the initial hours following intranasal inoculation of S. pneumoniae.
Mice with a genetic deficiency in C3 were significantly more susceptible to intranasal or intravenous challenge with S. pneumoniae (Table 1), confirming previous results from our laboratory and those of others (6, 7). This effect was not limited to serotype 2 pneumococci but was also found with serotype 4 organisms. It is possible that infection with other serotypes would not result in this effect, although a similar requirement for complement during early infection has been demonstrated by others (24).
Shorter survival times in C3– mice were associated with significantly elevated bacterial loads in the lung airways, lung tissue, and bloodstream within 6 h of infection (Fig. 1). D39 viability was controlled between 3 h and 6 h within C57BL/6 lung tissue but not in C3– lung tissue. As neutrophil recruitment is not different between the two mouse strains, it is likely that C3 is acting as an opsonin for resident alveolar macrophages in C57BL/6 mice during this time.
Previously, cobra venom factor-treated mice were found to have elevated pulmonary bacterial loads within hours of infection (14). Elevated bacteremia in C3– mice may be due to this pulmonary deficiency allowing increased numbers of pneumococci to enter the circulation or may be due to the bacteria being able to grow more quickly in the bloodstream of C3– mice.
Complement activation by pneumolysin was previously found to be involved in pulmonary pneumococcal growth (30) and to promote pneumococcal sepsis in cirrhotic rats (where complement levels are limiting [1]). We used pneumolysin-deficient bacteria (PLY–) to investigate whether complement activation by pneumolysin is the main action of the toxin during pneumonia. No difference between D39 and PLY– pneumococcus counts in C3– mice would indicate that complement activation by pneumolysin is the toxin's major role in virulence during pneumococcal pneumonia (complement activation by pneumolysin is already redundant in C3– mice, so infection with PLY– pneumococci should have no effect). This was not found to be the case. Infection with PLY– pneumococci resulted in longer survival times and lower bacterial loads than with D39 (Table 1 and Fig. 1). Thus, another function of pneumolysin is involved in virulence in our model.
Acapsular GalU-deficient S. pneumoniae isolates were utilized to focus on the importance of C3 within the lungs in the absence of bacteremia, as GalU– pneumococci were undetectable within the circulation of both C3– and C57BL/6 mice following intranasal inoculation. GalU– pneumococcus viability within lung airways declined immediately in C57BL/6 mice but reached significantly higher counts in C3– mice (Fig. 1G). This suggests that even when pneumococci lack a capsule, they can survive more successfully in the lungs of C3– mice. Therefore, during pneumonia without associated bacteremia, pulmonary C3 is crucial for control of the bacteria.
One hour following infection with either D39 or GalU– pneumococci, C3 levels in BALF, as measured by ELISA, increased (Fig. 2A and D), corresponding to a decrease in pneumococcal viability (Fig. 1A and G). C3 levels were then transiently depleted. It was not until bacterial loads were brought under control at later time points that an excess of C3 was present. We believe that these experiments are the first to report an association between C3 levels and bacterial loads during pneumococcal pneumonia. That serum C3 levels did not alter significantly during the infection agrees with previous data following intravenous infection with S. pneumoniae (1) but contrasts with those reported previously by Tu et al. (37). Discrepancies may reflect differences in the bacterial and mouse strains used.
C3 may be directly antibacterial (26), or the elevated C3 levels may be a marker of other antibacterial serum proteins leaking into the lungs.
As previously reported (41), C3 levels are lower in C57BL BALF than in serum (around 1,000-fold). The ability of pneumococci to activate the complement pathway (27) and deplete complement levels (2) should be more effective against the lower pulmonary level of C3 than the higher systemic levels. This may explain why C3 levels correlate with bacterial loads in the lungs but not in the circulation.
We measured total protein levels to discover whether C3 within BALF was locally produced or a serum component had leaked into the lungs (31) (Fig. 3). In C57BL/6 mice, C3 levels increased significantly within 1 h of intranasal infection with either D39 or GalU– pneumococci. At the same time, total protein levels also increased. Such kinetics indicate that the C3 within C57BL/6 airways is not locally produced but comes from the circulation.
We also found that a deficiency in C3 resulted in an elevated inflammatory response during infection (Table 2). Significantly elevated levels of TNF activity and IL-6 protein were present within C3– BALF and lung tissue during late infection.
There are two explanations for this. The elevated bacterial loads within C3– samples may act as a greater inflammatory stimulus. However, at 24 h, bacterial counts in the lungs of the two strains of mice were similar, while cytokine levels were significantly higher in C3– samples (Fig. 1A and B and Table 2). The alternative explanation is that C3 plays an anti-inflammatory role during pneumococcal infection. C3 has previously been found to decrease TNF- production by human peripheral blood mononuclear cells stimulated in vitro with lipopolysaccharide (34). In addition, C3a receptor-deficient mice produce more IL-1 following intravenous lipopolysaccharide administration than wild-type mice (21).
To investigate which cell types act with C3 in the lungs, we looked at pulmonary histology. Neutrophil recruitment to lungs was found within 1 h of intranasal infection with S. pneumoniae, with numbers increasing with time. Inflammatory cell recruitment was not attenuated in C3– mouse lungs (Table 3). By infecting C5ar– mice, we confirmed that complement was not required for phagocyte recruitment. We found that the large areas of perivascular inflammatory cell recruitment seen in wild-type and C3– mice were still evident in the absence of a C5a receptor. In a previous investigation, significant neutrophil recruitment was shown to occur in the lungs of C5-deficient mice (35). As neither C3 nor C5 is required for cell recruitment during pneumococcal pneumonia, the main role of complement during pneumococcal pneumonia must be in opsonophagocytosis (40) or exertion of direct antibacterial effects (26).
Overall, we have shown that C3 is essential within the lungs for an optimal immune response against S. pneumoniae, even when mice are not bacteremic. Without it, pneumococci reach higher bacterial loads, induce an elevated inflammatory response, and result in shorter survival times.
ACKNOWLEDGMENTS
Thanks go to Michael C. Carroll (Harvard Medical School, Boston, Mass.) for permission to use C3– mice and to James E. Marsh (Guys Hospital, London, United Kingdom) for providing them. Thanks to Craig Gerard (Harvard Medical School, Boston, Mass.) for supplying the C5ar– mice. Thanks to Lyndsey Hall for production of the pneumolysin-deficient bacteria.
The Wellcome Trust funded this work.
REFERENCES
1. Alcantara, R. B., L. C. Preheim, and M. J. Gentry. 1999. Role of pneumolysin's complement-activating activity during pneumococcal bacteremia in cirrhotic rats. Infect. Immun. 67:2862-2866.
2. Alcantara, R. B., L. C. Preheim, and M. J. Gentry-Nielsen. 2001. Pneumolysin-induced complement depletion during experimental pneumococcal bacteremia. Infect. Immun. 69:3569-3575.
3. Alper, C. A., N. Abramson, R. B. J. Johnston, J. H. Jandl, and F. S. Rosen. 1970. Increased susceptibility to infection associated with abnormalities of complement-mediated functions and of the third component of complement (C3). N. Engl. J. Med. 282:350-354.
4. Angel, C. S., M. Ruzek, and M. K. Hostetter. 1994. Degradation of C3 by Streptococcus pneumoniae. J. Infect. Dis. 170:600-608.
5. Bergeron, Y., N. Ouellet, A.-M. Deslauriers, M. Simard, M. Olivier, and M. G. Bergeron. 1998. Cytokine kinetics and other host factors in response to pneumococcal pulmonary infection in mice. Infect. Immun. 66:912-922.
6. Blue, C. E., G. K. Paterson, A. R. Kerr, M. Berge, J.-P. Claverys, and T. J. Mitchell. 2003. ZmpB, a novel virulence factor of Streptococcus pneumoniae that induces tumor necrosis factor alpha production in the respiratory tract. Infect. Immun. 71:4925-4935.
7. Brown, J. S., T. Hussell, S. M. Gilliland, D. W. Holden, J. C. Paton, M. R. Ehrenstein, M. J. Walport, and M. Botto. 2002. The classical pathway is the dominant complement pathway required for innate immunity to Streptococcus pneumoniae infection in mice. Proc. Natl. Acad. Sci. USA 99:16969-16974.
8. Circolo, A., G. Garnier, W. Fukuda, X. Wang, T. Hidvegi, A. J. Szalai, D. E. Briles, J. E. Volanakis, R. A. Wetsel, and H. R. Colten. 1999. Genetic disruption of the murine complement C3 promoter region generates deficient mice with extrahepatic expression of C3 mRNA. Immunopharmacology 42:135-149.
9. Circolo, A., T. B. Nutter III, and R. C. Strunk. 1997. Biosynthesis of complement components, p. 199-221. In A. W. Dodds and R. B. Sim (ed.), Complement: a practical approach. Oxford University Press, Oxford, United Kingdom.
10. Cole, F. S., W. J. Matthews, Jr., T. H. Rossing, D. J. Gash, N. A. Lichtenberg, and J. E. Pennington. 1983. Complement biosynthesis by human bronchoalveolar macrophages. Clin. Immunol. Immunopathol. 27:153-159.
11. Dave, S., A. Brooks-Walter, M. K. Pangburn, and L. S. McDaniel. 2001. PspC, a pneumococcal surface protein, binds human factor H. Infect. Immun. 69:3435-3437.
12. Dave, S., S. Carmicle, S. Hammerschmidt, M. K. Pangburn, and L. S. McDaniel. 2004. Dual roles of PspC, a surface protein of Streptococcus pneumoniae, in binding human secretory IgA and factor H. J. Immunol. 173:471-477.
13. Figueroa, J. E., and P. Densen. 1991. Infectious diseases associated with complement deficiencies. Clin. Microbiol. Rev. 4:359-395.
14. Gross, G. N., S. R. Rehm, and A. K. Pierce. 1978. The effect of complement depletion on lung clearance of bacteria. J. Clin. Investig. 62:373-378.
15. Hopken, U. E., B. Lu, N. P. Gerard, and C. Gerard. 1996. The C5a chemoattractant receptor mediates mucosal defence to infection. Nature 383:86-89.
16. Hosea, S. W., E. J. Brown, and M. M. Frank. 1980. The critical role of complement in experimental pneumococcal sepsis. J. Infect. Dis. 142:903-909.
17. Ibrahim, Y. M., A. R. Kerr, J. McCluskey, and T. J. Mitchell. 2004. Role of HtrA in the virulence and competence of Streptococcus pneumoniae. Infect. Immun. 72:3584-3591.
18. Janulczyk, R., F. Iannelli, A. G. Sjoholm, G. Pozzi, and L. Bjork. 2000. Hic, a novel surface protein of Streptococcus pneumoniae that interferes with complement function. J. Biol. Chem. 275:37257-37263.
19. Kemper, C. A., and S. C. Deresinki. 1994. The immunology of pneumococcal pneumonia, p. 29-49. In H. Chmel, M. Bendinelli, and H. Friedman (ed.), Pulmonary infections and immunity. Plenum Press, New York, N.Y.
20. Kerr, A. R., J. J. Irvine, J. J. Search, N. A. Gingles, A. Kadioglu, P. W. Andrew, W. L. McPheat, C. G. Booth, and T. J. Mitchell. 2002. The role of inflammatory mediators in resistance and susceptibility to pneumococcal infection. Infect. Immun. 70:1547-1557.
21. Kildsgaard, J., T. J. Hollman, K. W. Mathews, K. Bian, F. Murad, and R. A. Wetsel. 2000. Targeted disruption of the C3a receptor gene demonstrates a novel protective anti-inflammatory role for C3a in endotoxin-shock. J. Immunol. 165:5406-5409.
22. Marra, A., and D. Brigham. 2001. Streptococcus pneumoniae causes experimental meningitis following intranasal and otitis media infections via a nonhematogenous route. Infect. Immun. 69:7318-7325.
23. Mollerach, M., R. Lopez, and E. García. 1998. Characterization of the galU gene of Streptococcus pneumoniae encoding a uridine diphosphoglucose pyrophosphorylase: a gene essential for capsular polysaccharide biosynthesis. J. Exp. Med. 188:2047-2056.
24. Nakajima, R., K. Namba, Y. Ishida, T. Une, and Y. Osada. 1990. Protective role of complement in the development of experimental pneumococcal pneumonia in mice. Chemotherapy 36:287-293.
25. Newman, S. L., L. B. Vogler, R. D. Feigin, and R. B. J. Johnston. 1978. Recurrent septicemia associated with congenital deficiency of C2 and partial deficiency of factor B and the alternative complement pathway. N. Engl. J. Med. 299:290-292.
26. Nordahl, E. M., V. Rydengrd, P. Nyberg, D. P. Nitsche, M. Mrgelin, M. Malmsten, L. Bjrk, and A. Schmidtchen. 2004. Activation of the complement system generates antibacterial peptides. Proc. Natl. Acad. Sci. USA 101:16879-16884.
27. Paton, J. C., B. Rowan-Kelly, and A. Ferrante. 1984. Activation of human complement by the pneumococcal toxin pneumolysin. Infect. Immun. 43:1085-1087.
28. Ren, B., A. J. Szalai, S. K. Hollingshead, and D. E. Briles. 2004. Effects of PspA and antibodies to PspA on activation and deposition of complement on the pneumococcal surface. Infect. Immun. 72:114-122.
29. Roy, S., K. Knox, S. Segal, D. Griffiths, C. E. Moore, K. I. Welsh, A. Smarason, N. P. Day, W. L. McPheat, D. W. Crook, and A. V. S. Hill. 2002. MBL genotype and risk of invasive pneumococcal disease: a case-control study. Lancet 359:1569-1573.
30. Rubins, J. B., D. Charboneau, C. Fasching, A. M. Berry, J. C. Paton, J. E. Alexander, and P. W. Andrew. 1996. Distinct roles for pneumolysin's cytotoxic and complement activities in the pathogenesis of pneumococcal pneumonia. Am. J. Respir. Crit. Care Med. 153:1339-1346.
31. Rubins, J. B., D. Charboneau, J. C. Paton, T. J. Mitchell, P. W. Andrew, and E. N. Janoff. 1995. Dual function of pneumolysin in the early pathogenesis of murine pneumococcal pneumonia. J. Clin. Investig. 95:142-150.
32. Stockley, R. A. 1997. Soluble proteins in lung defence, p. 18-38. In R. A. Stockley (ed.), Pulmonary defences. John Wiley & Sons Ltd., Chichester, United Kingdom.
33. Strunk, R. C., D. M. Eidlen, and R. J. Mason. 1988. Pulmonary alveolar type II epithelial cells synthesize and secrete proteins of the classical and alternative complement pathways. J. Clin. Investig. 81:1419-1426.
34. Takabayashi, T., E. Vannier, B. D. Clark, N. H. Margolis, C. A. Dinarello, J. F. Burke, and J. A. Gelfand. 1996. A new biologic role for C3a and C3a desArg. J. Immunol. 156:3455-3460.
35. Toews, G. B., and W. C. Vial. 1984. The role of C5 in polymorphonuclear leukocyte recruitment in response to Streptococcus pneumoniae. Am. Rev. Respir. Dis. 129:82-86.
36. Toth, L. A. 1997. The moribund state as an experimental endpoint. Contemp. Top. Lab. Anim. Sci. 36:44-48.
37. Tu, A. T., R. L. Fulgham, M. A. McCrory, D. E. Briles, and A. J. Szalai. 1999. Pneumococcal surface protein A inhibits complement activation by Streptococcus pneumoniae. Infect. Immun. 67:4720-4724.
38. Van den Berg, C. W., H. Van Dijk, and P. J. A. Capel. 1989. Rapid isolation and characterisation of native mouse complement components C3 and C5. J. Immunol. Methods 122:73-78.
39. Wessels, M. R., P. Butko, M. Ma, H. B. Warren, A. L. Lage, and M. C. Carroll. 1995. Studies of group B streptococcal infection in mice deficient in complement component C3 or C4 demonstrate an essential role for complement in both innate and acquired immunity. Proc. Natl. Acad. Sci. USA 92:11490-11494.
40. Winkelstein, J., M. Smith, and H. Shin. 1975. The role of C3 as an opsonin in the early stages of infection. Proc. Soc. Exp. Biol. Med. 149:397-401.
41. Younger, J. G., S. Shankar-Sinha, M. Mickiewicz, A. S. Brinkman, G. A. Valencia, J. V. Sarma, E. M. Younkin, T. J. Standiford, F. S. Zetoune, and P. A. Ward. 2003. Murine complement interactions with Pseudomonas aeruginosa and their consequences during pneumonia. Am. J. Respir. Cell Mol. Biol. 29:432-438.
42. Zhang, Y., A. W. Masi, V. Barniak, K. Mountzouros, M. K. Hostetter, and B. A. Green. 2001. Recombinant PhpA protein, a unique histidine motif-containing protein from Streptococcus pneumoniae, protects mice against intranasal pneumococcal challenge. Infect. Immun. 69:3827-3836.(Alison R. Kerr, Gavin K. )