|Home | About | Journals | Submit | Contact Us | Français|
Staphylococcus aureus is responsible for a wide range of infections, including soft tissue infections and potentially fatal bacteremias. The primary niche for S. aureus in humans is the nares, and nasal carriage is a documented risk factor for staphylococcal infection. Previous studies with rodent models of nasal colonization have implicated capsule and teichoic acid as staphylococcal surface factors that promote colonization. In this study, a mouse model of nasal colonization was utilized to demonstrate that S. aureus mutants that lack clumping factor A, collagen binding protein, fibronectin binding proteins A and B, polysaccharide intercellular adhesin, or the accessory gene regulator colonized as well as wild-type strains colonized. In contrast, mutants deficient in sortase A or clumping factor B (ClfB) showed reduced nasal colonization. Mice immunized intranasally with killed S. aureus cells showed reduced nasal colonization compared with control animals. Likewise, mice that were immunized systemically or intranasally with a recombinant vaccine composed of domain A of ClfB exhibited lower levels of colonization than control animals exhibited. A ClfB monoclonal antibody (MAb) inhibited S. aureus binding to mouse cytokeratin 10. Passive immunization of mice with this MAb resulted in reduced nasal colonization compared with the colonization observed after immunization with an isotype-matched control antibody. The mouse immunization studies demonstrate that ClfB is an attractive component for inclusion in a vaccine to reduce S. aureus nasal colonization in humans, which in turn may diminish the risk of staphylococcal infection. As targets for vaccine development and antimicrobial intervention are assessed, rodent nasal colonization models may be invaluable.
Staphylococcus aureus causes a diverse spectrum of severe infections in humans, including bacteremia, endocarditis, and osteomyelitis, as well as skin and soft tissue infections. Notorious for decades as a major source of nosocomial infections, S. aureus has recently taken on a new role in causing an escalating number of community-acquired infections. To offset the problems of antibiotic-resistant S. aureus strains, preventive measures (e.g., immunization) should be explored as a complement to existing therapeutic approaches aimed at controlling this bacterial pathogen.
Humans are a reservoir for S. aureus, and the nose is the principal site of staphylococcal colonization. Approximately 20% of people persistently carry S. aureus in the anterior nares, ~60% are intermittent carriers, and ~20% are noncarriers (19). Nasal carriage is a known risk factor for staphylococcal infection in a number of clinical settings (51). Certain patient populations that show higher rates of S. aureus nasal colonization have an increased risk of staphylococcal infection. These populations include patients with diabetes, eczema, and human immunodeficiency virus infection, individuals receiving continuous ambulatory peritoneal dialysis or hemodialysis, and injection drug users (19). Moreover, patients in hospitals or individuals living in crowded conditions often show higher-than-normal rates of S. aureus nasal colonization. The source of ~80% of S. aureus bacteremias is endogenous since infecting bacteria have been shown by genotypic analysis to be identical to organisms recovered from the nasal mucosa (48, 53). These observations support an approach in which systemic S. aureus infections are prevented by eliminating or reducing nasal carriage.
One approach commonly used to reduce S. aureus carriage in individuals at risk for staphylococcal infection involves topical treatment with a nasal ointment containing the antibiotic mupirocin. Eradication of nasal carriage with topical mupirocin has been correlated with a reduction in the incidence of S. aureus infection in some patient populations (20, 45), but not in others (40, 54). Whereas mupirocin is effective in decolonizing nasal carriers, recolonization often occurs from extranasal carriage sites (52). Of further concern is the emergence of mupirocin resistance in S. aureus (31, 46). The utility of more recent experimental strategies to decrease colonization, including nasal application of tea tree oil (8), lysostaphin (22), or mersacidin (24), remains to be seen. Hence, nonantimicrobial approaches to combat S. aureus nasal carriage, including approaches that target staphylococcal adhesins that promote colonization, merit investigation.
S. aureus adheres to host extracellular matrix components, such as collagen, fibronectin, and fibrinogen, via surface protein adhesins called microbial surface components recognizing adhesive matrix molecules (MSCRAMMs). Clumping factor B (ClfB) is an S. aureus MSCRAMM that binds to fibrinogen (33, 35). O'Brien et al. (36) reported that ClfB also binds to the type 1 cytokeratin molecule K10 on the surface of desquamated human nasal epithelial cells and to both recombinant human and murine cytokeratin 10 (36, 49). Mutants lacking ClfB were poorly adherent to cytokeratin 10 and showed reduced adherence to human nasal epithelial cells (36). When ClfB was expressed on the surface of the heterologous host Lactococcus lactis, a significant increase in the binding of the L. lactis to squamous epithelial cells was observed compared with the binding of L. lactis expressing clumping factor A or L. lactis carrying the empty vector. These findings suggest that ClfB may be an important determinant of staphylococcal nasal colonization.
In this investigation, we examined the abilities of a variety of S. aureus surface components to promote colonization; these components included protein adhesins and the polysaccharide intercellular adhesin that has been implicated in staphylococcal biofilm formation. Whereas in previous studies researchers have evaluated potential immunogens to determine their abilities to prevent systemic S. aureus infection, in this report we provide evidence that immunization with killed S. aureus or recombinant ClfB (rClfB) can reduce S. aureus nasal colonization in a mouse model.
The S. aureus strains used for nasal colonization experiments are listed in Table Table1.1. Spontaneous streptomycin-resistant (Smr) mutants of wild-type S. aureus strains Newman (18), 8325-4 (34), and 502A (1) were selected on tryptic soy agar (TSA) plates containing 0.5 mg/ml streptomycin. The Smr mutants probably had single mutations in the rpsL gene, but they were phenotypically identical to the parental strains in terms of the growth rate, hemolysis on sheep blood agar plates, and the metabolic profile on API Staph test strips (Biomerieux, Inc., Durham, NC). Mutations in genes encoding surface components, including fnbA and fnbB (13, 30), clfA (30), clfB (28), cna (37), sdrCDE (35), srtA (26), and agr (23), were transduced from the original mutant strains into Smr wild-type strains with phage 85 or 80α, with selection for the appropriate antibiotic resistance marker. S. aureus strain Phillips (37) and its isogenic cna mutant PH100 were both Smr, so it was not necessary to further modify them. All of the strains used for the colonization experiments were Smr.
Mutants of strain 502A that had deletions in srtA or ica were constructed by allelic replacement mutagenesis (17), and the authenticity of the mutants was confirmed by PCR and Southern blot analysis. The temperature-sensitive plasmid pSC57 (6) was used to replace the wild-type ica locus; the resultant mutant, 502AΔica, showed minimal biofilm formation in a microtiter plate assay (6). Similarly, the temperature-sensitive plasmid pSrtA-KO (26) was used to replace the native srtA gene in strain 502A.
S. aureus expresses fibronectin binding proteins and ClfB predominantly during exponential growth (29, 33). To study the role of these proteins in nasal colonization, S. aureus strains were grown in tryptic soy broth (TSB) to an A650 of 0.34, harvested by centrifugation, and suspended in normal saline. Otherwise, S. aureus strains were cultivated on Columbia agar with 2% NaCl (CSA) at 37°C for 24 h as described previously (18). The agr mutant and wild-type strains were harvested from cultures grown on Columbia agar with 2% NaCl or from exponential-phase TSB cultures for evaluation in separate nasal colonization experiments. The viable count of each inoculum was determined by plating serial dilutions onto TSA.
ICR mice that were 4 to 6 weeks old were purchased from Harlan Sprague Dawley, Inc. (Indianapolis, Ind.) and given food and water ad libitum. The animals were housed four per cage in a modified barrier facility under viral antibody-free conditions. Animal care was in accordance with the institutional guidelines set forth by Brigham and Women's Hospital and Harvard Medical School.
Mice were given drinking water containing streptomycin sulfate (0.5 to 1.0 g/liter; Sigma Chemical Co., St. Louis, Mo.) 1 day prior to bacterial inoculation and for the course of the experiment. The drinking water and cages were changed twice a week. Mice were inoculated by the intranasal (i.n.) route with 10 μl of an S. aureus suspension as previously described (18), except that the mice were not anesthetized. Each S. aureus isogenic mutant pair was tested in parallel by using inocula of 108, 107, and 106 CFU per mouse. Colonization was evaluated by using quantitative cultures of the nasal tissues from separate groups of mice that were euthanized 7 or 14 days after bacterial inoculation. The area around the nasal region was wiped with 70% isopropyl alcohol, and the nose was excised and homogenized for 20 s in 400 μl TSB. The tissue homogenate was plated onto TSA with 5% sheep blood to assess the total nasal flora and onto TSA with 0.5 mg/ml streptomycin to determine the number of Smr S. aureus CFU per nose. All results described below are combined results from a minimum of two independent experiments.
Nasal colonization by strain Newman and the clfB mutant was also evaluated with Wistar rats that were bred at the Channing Laboratory. Male or female rats that were 7 weeks old were inoculated i.n. with 1.4 × 109 CFU S. aureus in a 10-μl suspension as described above. Colonization was evaluated by using quantitative cultures of the excised nasal tissues from rats that were euthanized 7 days after bacterial inoculation.
Mice were immunized with an acapsular mutant of S. aureus Reynolds designated JLO22 (7) that was cultivated in TSB to the logarithmic phase of growth (A650, 0.34). The bacteria were pelleted, suspended in phosphate-buffered saline at a concentration of 108 CFU/ml in an open petri dish, and exposed to a UV light source that was 8 cm away for 10 min on a rotator in the dark. The bacteria were concentrated by centrifugation, and 10 μl of the UV-killed bacterial suspension containing 108 CFU S. aureus with or without 5 μg of cholera toxin B (CTB) (List Biological Laboratories, Inc., Campbell, Calif.) was applied to each mouse nose on days 0, 5, and 10. CTB was omitted from the third immunization to reduce nonspecific protection. Two weeks after the third immunization, the mice were inoculated with 108 CFU S. aureus strain Newman cultivated in TSB to the logarithmic phase (A650, 0.34). Colonization was evaluated after 14 days.
The ligand binding activity of ClfB resides in the A or binding domain of the molecule (amino acids 44 to 542). The ClfB A region is composed of three subdomains designated N1, N2, and N3. Recombinant forms of the subdomains and the full-length A-region N123 were prepared and purified as described previously (39). Mice were immunized by the subcutaneous (s.c.) route with 30 μg of recombinant ClfB (binding domain N123) or bovine serum albumin (BSA) mixed with Freund's complete adjuvant (Mycobacterium butyricum; Difco Laboratories, Detroit, MI). In one experiment mice received a second immunization on day 10 by using the immunogen mixed with incomplete Freund's adjuvant. The last immunizing dose was given 7 days prior to bacterial inoculation with either 108 or 109 CFU S. aureus Smr Newman, and the mice were euthanized for quantitative culture 7 days later.
For mucosal immunization with ClfB, mice were anesthetized with ketamine (100 mg/kg) and xylazine (10 mg/kg), and a 17.5-μl dose containing 15 μg of ClfB and 5 μg of CTB was delivered i.n. Three immunizations were performed at 1-week intervals, and CTB was omitted from the third immunization. The last immunizing dose was given 10 days prior to bacterial inoculation with 109 CFU S. aureus Smr Newman, and the mice were euthanized for nasal culture 7 days later.
Monoclonal antibodies (MAbs) against the N2N3 subdomains (amino acids 197 to 542) of region A of ClfB were produced essentially as described by Kohler and Milstein (21), with minor modifications (44). Positive hybridomas designated 3D6 and 6C5 were grown to a high density, and MAbs were purified from supernatants of the hybridomas by ammonium sulfate precipitation, followed by affinity chromatography on a protein G-Sepharose column according to the recommendations of the manufacturer (Amersham Biosciences, Europe GmbH). Isotyping of the MAbs was performed using a Mouse-Typer subisotyping kit (Bio-Rad, Hercules, CA).
Binding inhibition studies were performed in microtiter plates coated with recombinant mouse cytokeratin 10 (1 μg/well) expressed and purified as previously described (49). A suspension containing 108 CFU S. aureus Newman was incubated for 1 h at 22°C with increasing amounts of MAbs before the bacteria were added to the coated microtiter wells. After incubation for 1 h at 37°C, each microtiter plate was washed, and adherent bacteria were detected by incubation with rabbit antibodies to S. aureus, followed by incubation with a peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG).
The N2N3 protein was biotinylated with N-hydroxysuccinimide biotin (Sigma) as reported previously (47). Binding inhibition studies were performed by incubating biotinylated N2N3 protein with increasing amounts of MAbs before the mixture was added to coated microtiter wells. After washing, binding of the protein was detected by addition of peroxidase-conjugated avidin.
Passive immunization experiments were performed to determine whether MAbs to ClfB could prevent or reduce S. aureus nasal colonization. Mice received 300 μg of MAb by the intraperitoneal (i.p.) route 10 h prior to inoculation with 109 CFU S. aureus Smr Newman, and quantitative culture analyses were performed with the mouse noses after 7 days.
Blood was collected from mice by tail nicking, and saliva production in mice was induced by i.p. injection of 2.5 to 5 μg carbamylcholine chloride (carbachol; Sigma). Saliva was collected from the oral cavity of each mouse with a pipette tip and centrifuged for 5 min at 21,000 × g prior to testing. Enzyme-linked immunosorbent assay (ELISA) plates (Immulon 2; Dynatech Laboratories, Chantilly, VA) were coated overnight at 4°C with 5 μg/ml rClfB and then blocked for 1 h with 0.5% skim milk. Mouse sera were diluted 1:100 in phosphate-buffered saline with 0.05% Tween 20, and saliva samples were diluted 1:20 before testing. The ELISA was performed as described previously (25), except that the saliva samples were incubated in the rClfB-coated coated ELISA plates overnight at 4°C. The relevant alkaline phosphatase-labeled conjugates included goat anti-mouse IgG and IgM (heavy and light chains; Kirkegaard & Perry Laboratories, Gaithersburg, MD) and goat anti-mouse IgA (α-chain specific; Sigma) for serum and saliva, respectively. Antibody levels were expressed as ELISA indices (A405 of the experimental sample [serum or saliva]/A405 of a high-titer control sample), as described previously (25).
Significant (P < 0.05) differences between quantitative culture results for different mouse groups were determined by the Mann-Whitney test (InStat; GraphPad Software, San Diego, CA).
Previous studies indicated that S. aureus mutants defective in production of capsular polysaccharide or ribitol teichoic acid, but not the penicillin-binding protein PBP2a, showed reduced nasal colonization in rodent models of staphylococcal nasal carriage (12, 18, 50). Because most facets of staphylococcal virulence are multifactorial, in this study we investigated the impact of other S. aureus surface components and one global regulator on nasal carriage. S. aureus isogenic mutants defective in the major surface-associated MSCRAMMs, the polysaccharide intercellular adhesin, and the agr global regulator were compared with the parental strain by using the murine nasal colonization model. Most of the mutations were constructed in the strain Newman genetic background (Table (Table1).1). However, S. aureus Newman lacks the cna gene that encodes the collagen binding protein, and so strain Phillips was used for some experiments. Strain Newman is also a poor biofilm producer (3, 6), so in colonization experiments in which we evaluated the role of the polysaccharide intercellular adhesin we utilized strain 502A.
Colonization by S. aureus mutants defective in clumping factor A, collagen binding protein, and SdrC, SdrD, and SdrE (surface-associated proteins with unknown ligands), as well as fibronectin binding proteins A and B, was not significantly reduced at days 7 or 14 compared with colonization by the wild-type S. aureus strains (data not shown). Grundmeier et al. (14) recently reported that the fibronectin binding proteins of wild-type strain Newman are truncated and not anchored to the cell wall. Therefore, we also evaluated the nasal colonization of mice inoculated with S. aureus strain 8325-4 and an isogenic mutant deficient in fibronectin binding proteins A and B (13). The fibronectin binding protein mutant in either strain background colonized the nares of mice at a level similar to that of the parental strain. Even a triple mutant of strain Newman defective in fibronectin binding proteins A and B and clumping factor A colonized mice as well as wild-type S. aureus strain Newman colonized mice. Likewise, the staphylococcal polysaccharide intercellular adhesin and the global regulator encoded by agr (tested in the S. aureus 8325-4 genetic background) failed to influence colonization, since bacterial strains with mutations in these loci also showed no reduction in nasal colonization (not shown).
Many surface proteins of S. aureus are linked to the cell wall by sortase, a transpeptidase that cleaves bacterial polypeptides at a conserved LPXTG motif and covalently links them to the peptidoglycan glycine cross bridges (27). To investigate whether any of the cell wall-anchored proteins might play a role in nasal colonization, we evaluated nasal colonization by sortase mutants of S. aureus. Sortase mutants of strains Newman and 502A both had growth curves similar to those of the parental strains, but they were clumping factor negative and they failed to bind to IgG or human fibrinogen immobilized on microtiter plates. Mice inoculated with 107 CFU of S. aureus Newman had significantly (P = 0.001) greater CFU/nose after 14 days than mice inoculated with the srtA mutant (Fig. (Fig.1).1). A similar reduction in colonization was observed when the sortase mutant of S. aureus strain 502A was compared to the parental strain (Fig. (Fig.1).1). The defect in nasal colonization of both sortase mutants was overcome when the inoculum was increased to 108 CFU per mouse (data not shown). With an inoculum of 106 CFU, there was little staphylococcal colonization after 14 days in either group of animals.
Previous studies demonstrated that S. aureus adheres to desquamated nasal epithelial cells and that ClfB promotes this interaction by binding to cytokeratin 10 (36, 49). Because ClfB is covalently linked to the staphylococcal peptidoglycan layer by sortase and is not regulated by agr (28), we sought to determine whether an S. aureus mutant defective in ClfB exhibited reduced nasal colonization in mice. Similar to the parental Newman clfB mutant, the Smr Newman clfB mutant exhibited reduced adherence to murine cytokeratin 10 immobilized on microtiter plates. Mice challenged with 107 or 106 CFU of the clfB mutant showed reduced colonization on day 14 compared with mice challenged with wild-type strain Newman (Fig. (Fig.2).2). Similar to our findings with the sortase mutant, the colonization defect was abrogated when the inoculum was increased to 108 CFU per mouse. Compared to mice colonized with the wild-type strain, mice inoculated with the sortase or clfB mutants had only slightly reduced carriage rates, but they carried significantly fewer S. aureus CFU/nose.
The S. aureus SdrC and SdrD proteins have also been shown to promote the adherence of S. aureus to desquamated nasal epithelial cells (T. Foster and E. Walsh, Clin. Microbiol. Infect. 11[Suppl. 2]:43, abstr. S182, 2005). Therefore, we inoculated mice with parental strain Newman or a mutant defective in ClfB and the SdrCDE proteins. Mice inoculated with 108 CFU of strain Newman had a median of 520 CFU S. aureus/nose (range, 1 to 96,000 CFU/nose) after 14 days. In contrast, mice inoculated with the Newman clfB sdrCDE mutant had a median of 2 CFU S. aureus/nose (range, 0 to 105 CFU/nose) (P = 0.0379). The nasal carriage for a mutant defective only in the SdrCDE proteins was similar to that for parental strain Newman (not shown), indicating that the SdrC, SdrD, and SdrE proteins are not involved in colonization of the mouse nares.
Other investigators have suggested that more consistent and higher levels of S. aureus nasal colonization can be obtained with rats than with mice (22). To determine whether ClfB was critical for nasal colonization in rats, we inoculated Wistar rats with 109 CFU S. aureus Newman or its isogenic clfB mutant (eight rats/group). Seven days after bacterial challenge, rats inoculated with Smr Newman had a median of 358 CFU S. aureus/nose (range, 11 to 3,036 CFU/nose), whereas rats inoculated with the ClfB mutant had a median of 17 CFU S. aureus/nose (range, 0 to 120 CFU/nose) (P = 0.0281). These results indicate that the clfB mutant was impaired for nasal colonization of both mice and rats.
One strategy with the potential to reduce or prevent nasal colonization in carriers is mucosal immunization with killed S. aureus cells or purified antigens. Results of our pilot experiments indicated that immunization with killed stationary-phase staphylococci (expressing abundant capsule but no ClfB) did not influence S. aureus nasal colonization of mice (not shown). Therefore, we immunized mice with a capsule-negative strain of S. aureus (JLO22) that was cultivated in TSB to the logarithmic phase of growth. Under these growth conditions, clfB is maximally expressed (33). Mice were given three doses of UV-killed bacteria (with or without the mucosal adjuvant CTB) and challenged with 108 CFU of the heterologous S. aureus strain Newman 2 weeks after the last dose. As shown in Table Table2,2, intranasal immunization with killed S. aureus resulted in only slightly reduced carriage rates, but the immunized mice had significantly (P < 0.05) fewer S. aureus CFU in their nares than the control animals had.
ClfB was chosen for immunization since reduced nasal colonization of mice was observed with clfB mutants of S. aureus. Furthermore, ClfB is detected on staphylococcal cells only during the logarithmic phase of growth (33), and mice immunized i.n. with logarithmic-phase S. aureus cells showed reduced nasal colonization. Mice immunized twice by the s.c. route with 30 μg rClfB mixed with Freund's complete adjuvant showed a rapid serum antibody response to rClfB (Fig. (Fig.3).3). Animals immunized with rClfB or BSA were inoculated with 108 CFU S. aureus Newman, and their noses were cultured quantitatively 7 days later. Mice immunized with rClfB had a median of 6 CFU S. aureus per nose (range, 0 to 81 CFU/nose), whereas mice immunized with BSA had a median of 369 CFU/nose (range, 0 to 2,300 CFU/nose). Subsequent experiments were performed with mice that were immunized once with rClfB or BSA mixed with complete Freund's adjuvant and were inoculated 7 days later with 109 CFU S. aureus Newman. As shown in Fig. Fig.4,4, the median number of CFU/nose for mice immunized once with rClfB was significantly reduced (P = 0.0473) compared with the value for mice immunized with BSA, although the carriage rates were similar for the two mouse groups.
To determine whether mucosal administration of rClfB could reduce S. aureus nasal carriage, we immunized mice i.n. with three doses consisting of 15 μg rClfB mixed with 5 μg CTB. Control animals received BSA mixed with CTB. A third group received lactated Ringer's solution to ensure that administration of CTB did not reduce colonization nonspecifically. As shown in Table Table3,3, mucosal immunization with rClfB resulted in elevated levels of ClfB-specific serum IgG antibodies, as well as elevated levels of ClfB-specific IgA antibodies in saliva. Control animals had antibody levels similar to prevaccination values. Ten days after the last immunization, the animals were inoculated with 109 CFU S. aureus Smr Newman. Mucosal immunization with rClfB significantly (P = 0.001) reduced the levels of S. aureus nasal colonization compared with the values for control animals immunized with BSA (Table (Table3),3), although the carriage rates were very similar for the two mouse groups.
The cytokeratin 10 binding site of rClfB is located in the N2N3 subdomains of rClfB (36). MAb 3D6 (IgG1κ) recognized the N2N3 protein but not the individual subdomains N2 and N3. Control MAb 6C5 (IgG1κ) bound to N2N3 and subdomain N2 but not to N3. MAb 3D6 blocked the binding of S. aureus Newman to mouse cytokeratin 10 in a dose-dependent fashion, whereas control MAb 6C5 did not (Fig. (Fig.5A).5A). Likewise, biotinylated N2N3 of ClfB preincubated with MAb 3D6 showed a 60% reduction in adherence to cytokeratin 10, whereas the control MAb showed no activity (Fig. (Fig.5B5B).
Because S. aureus nasal carriage was reduced by both systemic and mucosal immunization with rClfB, we performed passive immunization experiments with MAbs to ClfB. Passive immunization of mice with 300 μg MAb 3D6 10 h prior to inoculation with S. aureus Newman significantly (P = 0.0104) decreased the number of CFU/nose compared to the value obtained after immunization with MAb 6C5 (Fig. (Fig.6).6). Similar to the results described above, although the bacterial loads were reduced in the immunized mice, the carriage rates were similar for the two groups. These results indicate that serum ClfB-specific IgG is sufficient to reduce colonization and that a mucosal IgA response is not essential for protection against nasal colonization.
The nose is the primary reservoir of S. aureus in humans, and nasal carriage is an important risk factor for S. aureus infections (19). A comprehensive understanding of the factors involved in asymptomatic carriage of S. aureus in humans is critical to our ability to control the incidence of infection, as well as the transmission of this opportunistic pathogen. Previous studies have shown that both capsular polysaccharide and teichoic acid promote nasal colonization in rodent models of nasal carriage (18, 50). Here we report that the major MSCRAMMs, such as fibronectin binding protein, clumping factor A, and collagen binding protein, which are implicated in staphylococcal adherence to host tissues, had little effect on murine nasal colonization. Likewise, neither the polysaccharide intercellular adhesin nor the global regulator agr had an impact on colonization. However, the observation that two different sortase mutants poorly colonized the mouse nares suggested that at least one cell wall-anchored protein not regulated by agr is crucial for staphylococcal nasal colonization.
Consistent with the fact that rClfB binds to human desquamated nasal epithelial cells and to human and mouse cytokeratin 10 (36), we observed that ClfB was an important determinant of S. aureus nasal carriage in rodents. The S. aureus mutant lacking ClfB showed poor nasal colonization of both mice and rats compared to the parental strain. Previous studies have also confirmed that ClfB is expressed by staphylococci recovered from the nasal cavity of a healthy carrier and that S. aureus mutants defective in ClfB showed decreased in vitro adherence to human desquamated nasal epithelial cells (36). In contrast, mutants defective in SdrC or SdrD proteins showed decreased adherence to human squamous epithelial cells but colonized the mice as well as wild-type bacteria colonized them.
This is the first report of an immunization strategy that reduces S. aureus nasal colonization. Our initial experiments demonstrated that mice immunized i.n. with killed S. aureus showed reduced nasal colonization. Such a strategy has also been shown to be effective in reducing nasal colonization of mice by Streptococcus pneumoniae (55). However, systemic immunization with killed S. aureus did not protect rats or humans against systemic staphylococcal disease (32, 42). Of note, however, was our observation that only immunization with S. aureus harvested in the logarithmic phase of growth provided protection against colonization. These findings suggest that a surface-associated adhesin is important in eliciting protection against colonization, since the binding domain of S. aureus adhesins might be masked by capsular polysaccharide production in postexponential cultures (41). S. aureus ClfB was an obvious target because it is expressed only during the logarithmic phase of growth (33) and because both in vitro and in vivo experiments indicated that it is a key determinant of nasal carriage of S. aureus. Because the clfB gene is well conserved among S. aureus strains (4, 11), it is an attractive candidate for inclusion in a vaccine to prevent nasal colonization by this microbe.
A reduction in S. aureus nasal colonization was observed when mice were actively immunized with rClfB by either the s.c. or i.n. route. We observed elevated levels of ClfB antibodies in the serum of mice immunized s.c. with rClfB (Fig. (Fig.1),1), but ClfB antibodies could not be detected in induced saliva from these animals (unpublished observations). Nonetheless, ClfB antibodies might be present in the nasal secretions at levels below the detection level in pharmacologically induced saliva. Mice immunized i.n. with rClfB had elevated levels of ClfB antibodies in both the serum (IgG) and saliva (IgA) (Table (Table3).3). Thus, it appears that some protection against intranasal colonization by S. aureus can be achieved with serum ClfB antibodies and that IgA antibodies to ClfB (as measured in induced saliva) are not essential to obtain reduced levels of intranasal colonization. Cole et al. reported that S. aureus nasal colonization induces a local inflammatory response in humans (5). These findings suggest that S. aureus exhibits enough tissue invasion or induces sufficient tissue inflammation that serum antibody can play a significant role in reducing nasal carriage. Alternatively, antibodies may be transported onto epithelial surfaces by simple transudation (43). Our findings are concordant with the results of clinical and experimental studies that showed that there was decreased nasopharyngeal carriage of Haemophilus influenzae and S. pneumoniae vaccine serotypes following systemic immunization with conjugate vaccines (2, 10, 16, 38). Moreover, Dryla et al. reported that healthy individuals who were negative for S. aureus nasal carriage had higher levels of serum IgG and IgA against rClfB than individuals who were intermittent or persistent carriers of S. aureus had (9).
The results of our passive immunization experiments further confirmed the protective efficacy of serum antibodies directed against ClfB for reducing S. aureus nasal colonization. Systemic administration of MAbs specific for the N2N3 subdomain of rClfB effectively diminished nasal colonization in naïve mice compared with an isotype-matched MAb. In vitro studies with ClfB antibodies suggested that the keratin binding region of ClfB is located in the N2N3 region (36), and MAbs directed against this region reduced S. aureus binding to cytokeratin 10 (Fig. (Fig.2A).2A). This indicates that ClfB is the major S. aureus MSCRAMM that binds to cytokeratin 10. The failure of the ClfB MAb to completely inhibit S. aureus binding to immobilized keratin is consistent with our observation that the strain Newman clfB mutant exhibited low-level adherence to cytokeratin 10 (unpublished data), suggesting that this strain expresses a second, albeit less effective, cytokeratin adhesin.
Similar to studies of nasal carriage of H. influenzae and S. pneumoniae, rodent models have proven to be useful for elucidating the mechanism of S. aureus nasal colonization and assessing strategies for eradicating carriage. We demonstrated that S. aureus ClfB, a surface-associated adhesin, plays an important role in nasal colonization and that immunization by systemic or mucosal routes, as well as passive immunotherapy with MAbs to ClfB, reduced S. aureus nasal colonization in mice. ClfB is an attractive component for inclusion in a vaccine to reduce S. aureus nasal colonization. Because nasal colonization correlates with susceptibility to staphylococcal infection, such a vaccine might also reduce infections by this medically important pathogen. Rodent models of S. aureus nasal colonization are useful for assessment of targets for vaccine development or antimicrobial intervention.
This study was supported by National Institutes of Health grants AI44136 and AI29040 to J.C.L., by Wellcome Trust grant 061617 to T.F., by Fondazione CARIPLO grant 2003.1640/10.8485 to P.S., and by support from Inhibitex Inc. to T.F. and P.S.
We thank Richard Novick for providing strains RN6734 and RN6911, Joseph Patti for providing strains Phillips and PH100, Olaf Schneewind for providing mutant SKM3 and pSrtA-KO, and Sara Cramton for providing pSC57. We thank Michael Russell for advice on the performance of IgA ELISAs and One Kim for technical assistance.
Editor: J. N. Weiser