INFECTION AND IMMUNITY, Oct. 1997, p. 4146–4151 0019-9567/97/$04.0010 Copyright © 1997, American Society for Microbiology
Vol. 65, No. 10
Protective Efficacy of Antibodies to the Staphylococcus aureus Type 5 Capsular Polysaccharide in a Modified Model of Endocarditis in Rats JEAN C. LEE,1* JIN-SIR PARK,1 SARA E. SHEPHERD,2 VINCENT CAREY,1
AND
ALI FATTOM2
Channing Laboratory, Department of Medicine, Brigham & Women’s Hospital and Harvard Medical School, Boston, Massachusetts 02115,1 and NABI-Rockville, Rockville, Maryland 208522 Received 4 March 1997/Returned for modification 22 April 1997/Accepted 10 July 1997
The protective efficacy of antibodies to the Staphylococcus aureus type 5 capsular polysaccharide (CP5) was examined in a modified model of catheter-induced endocarditis. Rats were catheterized by surgically passing a polyethylene catheter through the right carotid artery and aortic valve into the left ventricle. The following day, the rats were injected by the intraperitoneal (i.p.) route with immunoglobulin G (IgG) purified from nonimmunized rabbits or from rabbits immunized with a conjugate vaccine composed of CP5 and CP8 linked covalently to recombinant Pseudomonas aeruginosa exotoxoid A. One day after passive immunization, the animals were challenged i.p. with one of three serotype 5 S. aureus isolates (strain Reynolds, Lowenstein, or VP) or nontypeable strain 521. Protection was evaluated by comparing quantitative cultures of blood, endocardial vegetations, and kidneys from control and immune animals. For experiments performed with S. aureus Reynolds and Lowenstein, rats given capsular antibodies (645 mg of CP5-specific IgG) showed a significantly (P < 0.05) lower prevalence of endocarditis than rats injected with nonimmune IgG. Similarly, quantitative cultures of the blood, kidneys, and aortic valve vegetations revealed that fewer S. aureus cells were recovered from rats given capsule-specific IgG than from rats administered nonimmune IgG. Rats challenged with strain VP were protected with 1.145 mg of CP5-specific IgG. Capsular antibodies did not protect against infection elicited by a nontypeable strain. These results demonstrate that capsular antibodies elicited by immunization with a polysaccharide-protein conjugate vaccine protect experimental animals against serotype 5 S. aureus infection in a modified model of endocarditis. munized rabbits and from rabbits immunized with a conjugate vaccine composed of CP5 and CP8 linked covalently to recombinant Pseudomonas exotoxoid A (4).
The goal of this study was to determine whether antibodies to the Staphylococcus aureus capsular polysaccharides (CPs) could protect rats against S. aureus infection in an experimental model of catheter-induced endocarditis. This infection model is attractive because the pathogenesis of the disease parallels the invasive nature of human infection, and the use of rats is economical compared with the costs of purchasing and maintaining larger animals, such as rabbits. Because they are surface associated, limited in antigenic specificity, and highly conserved among clinical isolates, S. aureus CPs may be important in immunity to staphylococcal infections. A previous study (15) demonstrated that antibodies to the S. aureus type 5 CP (CP5) did not protect rats against S. aureus catheter-associated endocarditis. We hypothesized that this lack of protection was a result of two factors. First, the vaccine used for active immunization was composed of killed, whole bacterial cells, and antibodies raised to a whole-cell vaccine may differ in affinity or immunoglobulin class distribution from antibodies raised to a polysaccharide-conjugate vaccine. Second, the bacterial challenge was delivered as a bolus injection of 5 3 104 CFU of S. aureus into the tail veins of the animals. We changed both of these parameters in the current study. We modified the existing animal model of endocarditis by administering the bacterial inoculum by the intraperitoneal (i.p.) route to achieve a slow infusion of S. aureus into the blood. This modified model of staphylococcal endocarditis was used to compare protection afforded animals by passive immunization with immunoglobulin G (IgG) purified from nonim-
MATERIALS AND METHODS Bacteria. The S. aureus isolates used for this study include strains Reynolds (8) and Lowenstein (4). In addition, a fresh bacteremia isolate (strain VP) from the Brigham & Women’s Hospital microbiology laboratory was kindly provided by Andrew Onderdonk. These three strains produced serotype 5 capsules and were sensitive to methicillin. S. aureus 521, isolated from a patient on end-stage renal dialysis, was nonreactive with antibodies to the serotype 5 or serotype 8 capsules. Staphylococci were grown overnight at 37°C on Columbia agar (Difco Laboratories, Detroit, Mich.) supplemented with 2% NaCl. Colonies were scraped from the agar surface and suspended in phosphate-buffered saline (PBS) (10 mM sodium phosphate–0.15 M NaCl, pH 7.3) to an optical density at 650 nm of 0.43. Bacteria were diluted to the appropriate concentration, and CFU were verified by plate counts performed in duplicate on tryptic soy agar (Becton Dickinson Microbiology Systems, Cockeysville, Md.). Conjugate vaccine. S. aureus CP5 and CP8 are serologically distinct, and antibodies reactive with the purified polysaccharides have been shown to be type specific (4, 8). CP5 and CP8 were purified as described previously (4) under Good Manufacturing Practices. Evaluation of the purified CP product showed that it was free of organic phosphate and glycerol, and its purity was confirmed by nuclear magnetic resonance analysis. CP5 and CP8 conjugate vaccines were prepared as described previously (4) with cystamine-derivatized polysaccharides and N-succinimidyl 3-(2-pyridyldithio)propionate-derivatized recombinant exotoxoid A from Pseudomonas aeruginosa. The serotype 5 and 8 conjugates were combined into a bivalent vaccine at a final concentration each of 50 mg/ml. Rabbits were immunized with multiple 50-mg doses of each conjugate. The IgG fraction of normal or immune rabbit serum was purified on a protein G column. After concentration, the total IgG content was estimated by absorption at 280 nm, and the CP-specific IgG content was determined by enzyme-linked immunosorbent assay (ELISA). ELISA. Rabbit IgG specific for S. aureus CP5 and CP8 in rabbit immunoglobulin preparations and in recipient rat sera was measured by ELISA. Ninety-sixwell flat-bottom Immulon 4 microtiter plates (Dynatech Laboratories, Inc., Chantilly, Va.) were coated at ambient temperature with 1 mg of polysaccharide per ml in PBS for 6 h or overnight. The contents of the microtiter plates were aspirated, and the wells were blocked with 1% bovine serum albumin (BSA) in
* Corresponding author. Mailing address: Channing Laboratory, 181 Longwood Ave., Boston, MA 02115. Phone: (617) 525-2652. Fax: (617) 731-1541. E-mail:
[email protected]. 4146
VOL. 65, 1997
CAPSULAR ANTIBODIES IN S. AUREUS ENDOCARDITIS
4147
FIG. 1. Nonimmunized, catheterized rats were challenged i.p. with inocula ranging from 6 3 105 to 6 3 107 CFU of the serotype 5 S. aureus strain Reynolds. Quantitative cultures of blood (A) were performed at 24 or 48 h after challenge with S. aureus. Quantitative cultures of the endocardial vegetations (B) and kidneys (C) were done for the animals 4 days after inoculation. In panels B and C, each point represents a value for one rat. Most of the catheterized rats developed endocarditis at challenge doses of $1 3 107 CFU/rat. All rats with endocarditis were bacteremic and positive for renal abscess formation.
PBS for 1 h at ambient temperature. Twofold dilutions of reference and test serum samples were prepared in PBS containing 0.01% Brij 35 and 1% BSA and added to the washed (0.9% NaCl with 1% Brij 35) microtiter plate. After incubation for 1 h at 37°C, the microtiter wells were washed and a dilution of peroxidase-conjugated affinity-purified goat anti-rabbit IgG (Fc fragment specific; Jackson ImmunoResearch Laboratories, Inc., West Grove, Pa.) was added to the plates. After a 1-h incubation at 37°C, the plates were again washed and TMB (3,39,5,59-tetramethylbenzidine) peroxidase substrate (Kirkegaard and Perry, Gaithersburg, Md.) was added to each well. The microtiter plates were incubated for 10 min at ambient temperature before the reaction was stopped with 1 M phosphoric acid. The absorbance at 450 nm of each well was recorded with a microtiter plate reader (MR4000; Dynatech). The antibody concentration in each test sample was calculated by using the parallel line analysis described by Manclark et al. (13). Opsonophagocytic assay. The opsonic activity of rabbit IgG was determined as described by Xu et al. (20), except that the bacteria were cultivated on Columbia salt agar plates. IgG samples were tested at a 1:10,000 dilution in minimal essential medium (GIBCO Laboratories, Grand Island, N.Y.) containing 1% BSA. Catheter-induced endocarditis. Male Sprague-Dawley rats were obtained from Charles River Laboratories, Kingston, N.Y. They weighed 195 to 206 g upon arrival, and they were given food and water ad libitum. A modification of the rat model of catheter-induced endocarditis described by Baddour et al. (1) was used for this study. Rats were anesthetized with a mixture of ketamine (35 mg/kg of body weight; Aveco Co., Fort Dodge, Iowa), xylazine (10 mg/kg; Miles Inc., Shawnee Mission, Kans.), and atropine (0.09 mg/kg;
Anpro Pharmaceutical, Arcadia, Calif.). A polyethylene catheter (Intramedic PE10; Clay-Adams, Parsippany, N.J.) was passed through the right carotid artery and the aortic valve into the left ventricle. Vigorous pulsation of blood within the catheter indicated correct positioning of the device. The catheter was sealed and tied in place with sterile suturing material, and the incision was closed. Placement of the catheter results in deposition of platelet-fibrin thrombi on the valves, which allows for colonization by bloodborne S. aureus. Rats with indwelling catheters were passively immunized 24 h after surgery by i.p. injection of 1 ml of IgG at a concentration of 15.1 mg/ml (645 mg of CP5-specific IgG and 1,070 mg of CP8-specific IgG per ml; lot 1) or 51.9 mg/ml (1.145 mg of CP5-specific IgG and 2.527 mg of CP8-specific IgG per ml; lot 2). Control animals were injected with an equal concentration of nonimmune IgG (,1 mg of CP5- and CP8-specific IgG per ml). At 48 h after surgery, the rats were bled to determine their CP5-specific antibody levels before i.p. challenge with a 0.2-ml bacterial inoculum. Heparinized blood was collected from each animal by tail vein puncture at 2, 24, 48, 72, and 96 h after inoculation. Quantitative blood cultures were done by plating dilutions of each sample on tryptic soy agar plates containing 5% sheep erythrocytes (Becton Dickinson). At 96 h after challenge, the surviving rats were euthanized, and their hearts and kidneys were removed. The position of the catheter within the heart and the presence or absence of vegetations were noted. The kidneys and aortic valve vegetations were weighed and homogenized in tryptic soy broth. Quantitative plate counts were performed with serial dilutions of the homogenates, and the CFU per gram of tissue was calculated. The minimum detectable numbers of bacteria were calculated as 103 CFU/g of vegetation and 101 CFU/g of kidney tissue. The minimum detectable number of bacteria in the blood depended on the volume of blood collected and plated for culture; the range was 5 3 101 to 5 3 100 CFU/ml of blood. Rats that died before 48 h were excluded from the study if they showed no gross evidence of vegetation formation. Statistical analysis. Bacterial concentrations in the blood, kidneys, and endocardial vegetations were analyzed by the one-sided Mann-Whitney U test, and the numbers of infected animals in each group were compared by the Fisher exact test. Optimally weighted repeated-measure Wilcoxon tests (19) were performed to supplement comparisons of quantitative blood cultures for the same animals at different time points.
RESULTS Modified model of S. aureus endocarditis. The models of endocarditis up to now have used a bolus intravenous (i.v.) injection of microorganisms to inoculate the catheterized animals (1, 14, 16, 17). Reproducible infection of catheterized rats
4148
LEE ET AL.
INFECT. IMMUN.
TABLE 1. Opsonophagocytic killing of S. aureus Reynolds by human polymorphonuclear leukocytes Antibodya
CP-specific IgG Normal IgG None
Complementb
1 2 1 2 1
% Killing (mean 6 SEM) after: 60 min
120 min
50 6 10 19 6 12 363 363 666
93 6 1 24 6 12 10 6 10 060 767
a The IgG preparations (15.1 mg of IgG per ml) were tested at a final dilution of 1:10,000 in the assay. Each combination of antibody and complement (as well as the control with no antibody) was tested three times. b Guinea pig serum (final concentration, 1%) served as a source of complement.
follows i.v. challenge with as few as 5 3 104 CFU of S. aureus. We modified the endocarditis model by challenging the animals by the i.p. route to achieve a slow infusion of S. aureus organisms into the blood. We evaluated the i.p. challenge model of endocarditis in nonimmunized rats at inocula ranging from 4.2 3 105 to 7.4 3 107 CFU of the serotype 5 strain Reynolds. As shown in Fig. 1B, most of the catheterized rats developed endocarditis at challenge doses of $1 3 107 CFU/ rat. All rats with endocarditis were bacteremic at 24 and 48 h (Fig. 1A) and positive for metastatic dissemination to the kidneys (Fig. 1C). In these experiments, blood samples for culture were not taken 2 h after bacterial challenge. Quantitative blood cultures were performed at 72 and 96 h after challenge, but the results are not depicted since many of the rats with endocarditis had died or were moribund at these times. None of the catheterized animals challenged i.p. with S. aureus developed a wound infection at the surgical site, nor did we note any experimental animals with abdominal abscesses or bowel perforation. To evaluate the level of bacteremia achieved by i.p. inoculation of S. aureus, 15 nonimmunized rats were challenged with 4 3 107 CFU of strain Reynolds. Separate groups of three or four animals were bled at 5, 20, 65, and 120 min, and the blood samples were cultured quantitatively. The mean CFU per milliliter of blood did not differ significantly over that period, ranging from 2.16 3 102 CFU of S. aureus per ml at 5 min postinoculation to 1.08 3 102 CFU/ml 120 min after challenge. Rabbit IgG was purified from the sera of normal rabbits or from rabbits immunized with the S. aureus capsule conjugate vaccine. Unless otherwise indicated, the immune rabbit IgG contained 645 mg of CP5-specific IgG and 1,070 mg of CP8specific IgG per ml (lot 1). Both normal and immune IgG samples contained 15.1 mg of total IgG per ml. The CP-specific immunoglobulin preparation (containing antibodies to CP5 and CP8) reacted strongly as determined by immunodiffusion analysis against capsular extracts prepared from S. aureus Reynolds (not shown). Furthermore, the immunoglobulin preparation was opsonic for strain Reynolds cells in an in vitro opsonophagocytic killing assay (Table 1). Human polymorphonuclear leukocytes killed 93% of strain Reynolds cells after a 2-h incubation in the presence of CP-specific IgG and guinea pig serum (provided as a complement source). The phagocytes showed poor killing (10% reduction in CFU per milliliter) of strain Reynolds in the presence of guinea pig serum and normal rabbit IgG. A 7% reduction in CFU per milliliter was seen when strain Reynolds was incubated with complement and no added antibodies. This experiment underscores the importance of capsular antibodies in clearance of encapsulated S. aureus by neutrophils. To evaluate the pharmacokinetics of the CP-specific IgG
FIG. 2. Pharmacokinetics of 1 ml of rabbit CP-specific IgG (15.1 mg of total IgG; 645 mg of CP5-specific IgG and 1,070 mg of CP8-specific IgG per ml) delivered to rats by the i.p. or i.v. route. CP5-specific-antibody levels in serum in recipient animals were analyzed by ELISA (4) and are shown as mean IgG concentrations 6 standard errors of the mean.
delivered to rats, groups of three noncatheterized, noninfected animals were administered 1 ml of the sample by the i.p. or i.v. route (tail vein injection). Recipient animals were bled daily for 7 days, and their CP5-specific-antibody levels in serum were analyzed by an ELISA. As shown in Fig. 2, similar CP5-specific-antibody levels in serum were achieved by the two routes of administration. Peak antibody concentrations of 25 to 30 mg of IgG/ml of serum were observed on day 1 after injection. After 1 week, approximately 50% of the maximal antibody concentration persisted in the sera of the uninfected animals. In animals given nonimmune IgG, CP5-specific-antibody levels in serum were uniformly ,1 mg/ml. Two groups of rats (three animals per group) were passively immunized i.p. with 1 ml of normal or CP-specific serum. Twenty-four hours later, the animals were injected i.p. with 0.5 ml of saline and euthanized. Aliquots of the peritoneal fluid were collected and analyzed for CP5-specific antibodies by ELISA. Rats given immune and nonimmune IgG had 12 and ,1 mg of CP5-specific IgG per ml, respectively, in their peritoneal fluids. In a separate experiment, we collected small samples of peritoneal fluid from immune and nonimmune animals 30 min, 1 h, and 3 h after i.p. challenge with ;3.3 3 107 CFU of S. aureus Reynolds. In stained smears of these samples, we observed similar numbers of peritoneal macrophages (and some neutrophils) with associated staphylococci in the two groups of rats. Thus, local phagocytosis and bacterial killing within the peritoneal cavity contribute to bacterial clearance in this model of infection. Challenge of passively immunized rats with strain Reynolds. Passive immunization with 1 ml of rabbit CP-specific IgG delivered i.p. resulted in levels of 24.5 6 0.99 mg of CP5-specific IgG per ml of rat serum 24 h after injection (just before bacterial challenge). Rats given nonimmune IgG had ,1 mg of CP5-specific IgG per ml in their sera. In the initial passiveimmunization experiments, catheterized rats injected with normal IgG were more resistant to endocarditis than were rats given no IgG. Nonimmunized rats challenged with 107 CFU of strain Reynolds were bacteremic, and all four animals developed endocarditis and renal abscesses (Fig. 1). In contrast, none of six animals given nonimmune IgG and inoculated with
VOL. 65, 1997
CAPSULAR ANTIBODIES IN S. AUREUS ENDOCARDITIS
4149
Log CFU/ml of blood at: 2h 24 h 48 h No. of rats with endocarditis/total Vegetation wt (g) Log CFU/g of vegetation Log CFU/g of kidney
Result with the indicated immunoglobulinb Parameter Normal IgG
CP-specific IgG
9.6 3 106 CFU/rat
Normal IgG
1.70 (1.7–2.0) 1.52 (1.52–2.65) 1.52 (1.52–2.94) 1/7 0.011 (0–0.011) 3.40 (3.4–11.38) 1.07 (1.03–8.66)
P valueb
Result for the indicated variable by inoculum sizea
CP-specific IgG
3.9 3 107 CFU/rat
Normal IgG
0.0073 0.0135
0.0073 0.0073
0.7 (0.7–0.7) 0.7 (0.7–0.7) 0.7 (0.7–1.0) 0/5 None 3.4 (3.4–3.4) 1.05 (0.74–2.09) 0.0079
1.92 (1.48–2.04) 4.31 (4.10–5.63) NDc 4/4 0.012 (0.007–0.021) 11.01 (10.4–11.4) 7.33 (6.76–7.77)
Normal IgG
2.2 (0.7–2.6) 4.61 (0.7–6.25) 5.4 (0.7–5.79) 7/8 0.009 (0–0.022) 11.12 (3.4–11.98) 7.93 (1.04–9.34)
CP-specific IgG
0.0023 0.0147 0.0411 0.0406 0.0117 0.0117 0.0209
P valueb
1.3 3 108 CFU/rat
0.9 (0.7–1.1) 1.1 (0.7–4.06) 1.50 (1.7–5.08) 2/8 0 (0–0.013) 3.4 (3.4–11.14) 2.77 (1.02–7.86)
TABLE 2. Quantitative culture results for rats passively immunized by the i.p. route and challenged with various doses of serotype 5 S. aureus strain Reynolds
Parameter
1.70 (1.7–1.7) 1.52 (1.52–1.52) 1.52 (1.52–1.52) 0/6 None ,3.40 2.41 (1.12–6.5)
a Values are medians, with ranges shown in parentheses. The data were analyzed by the Mann-Whitney U test. ND, not done.
b
c
TABLE 3. Protection elicited by CP-specific antibodies against S. aureus Lowensteina
Log CFU/ml of blood at: 2h 0.7 (0.7–1.7) 24 h 1.3 (0.7–4.8) 48 h 3.0 (0.7–5.6) No. of rats with 8/11 endocarditis/total Vegetation wt (g) 0.0107 Log CFU/g of vegetation 11.0 (3.4–11.4) Log CFU/g of kidney 7.84 (3.0–8.4)
CP-specific IgG
P valuec
0.7 (0.7–2.4) 0.7 (0.7–5.7) 0.7 (0.7–5.3) 4/13
0.605 0.134 0.037 0.0498
0.0029 3.4 (3.4–11.4) 4.85 (3.0–8.8)
0.016 0.095 0.183
a Rats were challenged with an inoculum of 6.2 3 107 CFU/rat after passive immunization with IgG. b Values are medians, with ranges shown in parentheses. c The data were analyzed by the Mann-Whitney U test.
9.6 3 106 CFU of S. aureus developed endocarditis or renal infection (Table 2). When the inoculum was increased fourfold (3.9 3 107 CFU), four of four rats passively immunized with normal IgG sustained bacteremia, and they developed endocarditis and renal abscesses. None of five rats given CP-specific antibodies and challenged with 3.9 3 107 CFU of strain Reynolds developed endocarditis. Similarly, the results of quantitative cultures of blood, kidneys, and aortic valve vegetations from rats given capsular antibodies were significantly lower than those for rats injected with normal IgG (Table 2). When the optimally weighted repeated-measure Wilcoxon test was applied to the quantitative blood culture values determined at 2, 24, and 48 h for the two animal groups, the resulting P value was 0.074. At the highest inoculum tested (1.3 3 108 CFU), rats given CP-specific IgG still showed significant protection compared with rats given normal IgG. Seven of eight control rats were infected with S. aureus, compared with two of eight animals given CP-specific IgG. Quantitative culture results were consistently lower for the rats given capsular antibodies than for control animals (Table 2). The optimally weighted repeatedmeasure Wilcoxon test gave a P value of 0.048 for the sequential blood culture data obtained from the nonimmune and immune animals. Challenge of passively immunized rats with strain Lowenstein. To ensure that the protection observed was not restricted to infection with the prototype strain Reynolds, we inoculated additional groups of catheterized and passively immunized animals with S. aureus Lowenstein (6.2 3 107 CFU/ rat). As shown in Table 3, CP-specific antibodies were also protective against infection induced by this strain. Eight of 11 rats given normal IgG developed endocarditis, compared with 4 of 13 animals that received CP-specific IgG (P 5 0.0498). In addition, the aortic valve vegetations were smaller in the rats given CP-specific antibodies (median 5 0.0107 g) than in the normal rats (median 5 0.0029 g; P 5 0.016). Although the median bacterial densities in the blood, vegetations, and kidneys were lower for the rats given capsular antibodies than for the controls, differences in the quantitative cultures did not reach significance, except for blood cultures performed at 48 h (P 5 0.037). When the optimally weighted repeated-measure Wilcoxon test was applied to the quantitative blood culture values determined at 2, 24, and 48 h for the two animal groups, the resulting P value was 0.15. Challenge of passively immunized rats with strain VP. Additional groups of animals were catheterized and passively immunized to determine whether capsular antibodies could
4150
LEE ET AL.
INFECT. IMMUN.
TABLE 4. Results of passive-immunization experiments in which rats were challenged with S. aureus VPa Expt no.
1 2 3 4 5
Dose (CFU/rat)
8.4 3 107 6.0 3 107 4.0 3 107 1.4 3 107 6.5 3 106
No. of rats with endocarditis/total Nonimmune
Immune
5/5 5/7 5/7 5/7 2/7
3/4 4/5 6/8 1/7 2/6
P valueb
Antibody level (mg/ml) in rats given CPspecific IgG
NS NS NS 0.0513 NS
NDc 23–29 17–24 17–24 22–33
Catheterized rats were given 645 mg of CP5-specific IgG or normal IgG. P values were determined by the Fisher exact test. NS, not significant. c ND, not done. a b
protect against infection with a fresh bacteremia isolate of S. aureus. As shown in Table 4, the results of five experiments with various inocula indicated that passive immunization with CP-specific antibodies did not protect the rats against endocarditis induced by strain VP. The recipient animals in the immune group had CP5-specific-antibody levels in serum ranging from 17 to 33 mg/ml. Only in a single experiment did the antibody treatment exert a substantial effect. At an inoculum of 1.4 3 107 CFU per rat, five of seven control animals developed endocarditis, compared with one of seven rats given CPspecific IgG (P 5 0.0513). S. aureus VP was apparently more virulent than strain Reynolds in this infection model. Five of seven rats given normal IgG and challenged with 1.4 3 107 CFU of strain VP developed endocarditis (Table 4), whereas at a similar inoculum (9.6 3 106 CFU/rat), none of six rats challenged with strain Reynolds were infected (Table 2). In addition, most of the rats inoculated with strain VP died or were moribund within 48 h of bacterial challenge. We considered that the CP5 antibody levels in serum achieved in rats challenged with S. aureus VP were insufficient to protect against infection with this apparently more virulent strain. Therefore, additional groups of rats were passively immunized with a more concentrated IgG preparation (lot 2). Nine rats were injected with 1 ml of CP-specific IgG containing 1.145 mg of CP5-specific IgG and 2.527 mg of CP8-specific IgG (total IgG concentration of 51.9 mg/ml). Control rats were administered 51.9 mg of IgG obtained from nonimmunized rabbits. The results of the protection study, shown in Table 5, TABLE 5. Protection elicited by CP-specific antibodies against S. aureus VPa Parameter
Result with the indicated immunoglobulinb Normal IgG
CP5-specific antibodies ,1 in recipient animals (mg/ml) Log CFU/ml of blood at: 2 hd 0.70 (0.70–1.30) 24 h 3.93 (0.7–4.85) No. of rats with 7/9 endocarditis/total Vegetation wt (g) 0.009 (0.0–0.0166) Log CFU/g of vegetation 10.22 (3.4–11.04) Log CFU/g of kidney 7.20 (2.14–8.31)
P valuec
CP-specific IgG
55–121
0.70 (0.70–0.70) 0.70 (0.7–2.98) 1/9
NS 0.0106 0.0076
0.000 (0.0–0.0096) 0.0672 3.40 (3.4–9.78) 0.0078 2.06 (1.19–7.03) 0.0056
a Rats were challenged with an inoculum of ;6.7 3 107 CFU/rat after passive immunization with IgG. b Values are medians, with ranges shown in parentheses. c The data were analyzed by the Mann-Whitney U test. NS, not significant. d Two-hour blood cultures were performed for six of nine animals in each group.
indicate that higher levels of CP-specific antibodies did protect against endocarditis induced by strain VP at a dose of 6.7 3 107 CFU/rat. Passive immunization with 1 ml of rabbit CP-specific IgG delivered i.p. resulted in levels in serum ranging from 55 to 121 mg of CP5-specific IgG per ml of rat serum 24 h after injection (just before bacterial challenge). Seven of nine rats given normal IgG developed endocarditis, compared with one of nine rats injected with CP-specific IgG. Similarly, the results of quantitative cultures of blood, kidneys, and aortic valve vegetations from rats given capsular antibodies were significantly lower than those for rats injected with normal IgG (Table 5). Too few animals from the control group were alive for quantitative blood culture comparisons at 48 h or thereafter. The optimally weighted repeated-measure Wilcoxon test gave a P value of 0.058 for the sequential blood culture data obtained from the nonimmune and immune animals. Challenge of passively immunized rats with nontypeable strain 521. To ensure that the antibody-mediated protection that we observed in this infection model was specific for strains producing a microcapsule, we challenged a group of nine catheterized and passively immunized animals with nontypeable S. aureus strain 521 (inoculum of 9.3 3 107 CFU/rat). Two of four rats given CP-specific IgG (lot 1) developed endocarditis, whereas three of five rats injected with normal IgG developed endocarditis. The results of quantitative cultures of blood, vegetations, and kidneys did not differ between the two groups of animals (not shown). DISCUSSION S. aureus CPs were one of the earliest targets in vaccine studies aimed at preventing staphylococcal infections. Immunization with polysaccharide antigens extracted from highly encapsulated S. aureus strains (serotype 1 or 2) protected mice against infection with the homologous, but not heterologous, capsule types (3, 6). Furthermore, protection could be passively transferred to immunologically naive animals by injecting them with immune serum. Capsular antibodies could also protect mice against staphylococcal abscesses in a sublethal infection model (12). However, highly encapsulated strains are rarely encountered among clinical isolates of S. aureus. Most strains are microencapsulated, i.e., they produce capsules smaller than those expressed by the highly encapsulated mucoid strains (8). Of the 11 known capsular types, serotypes 5 and 8 comprise the majority (80%) of isolates recovered from humans (18). The protective efficacy of antibodies to the type 5 capsule was evaluated previously in a catheter-induced rat model of endocarditis (15). In that study, rats were actively immunized with killed, serotype 5 S. aureus strain Reynolds or passively immunized with high-titer rabbit antiserum raised against killed bacteria (and adsorbed to remove noncapsular antibodies). Control animals were injected with saline or passively immunized with normal rabbit serum. Both groups were challenged with an i.v. bolus of S. aureus Reynolds (;5 3 104 CFU). Despite having elevated levels of capsular antibodies, the immunized animals were susceptible to staphylococcal endocarditis, and immunized and control animals had similar numbers of bacteria in their blood. A similar lack of protection was observed by Bayer et al. (2), who passively immunized rabbits i.v. with antibodies raised to the CP5-CP8 conjugate vaccine and challenged them with an i.v. bolus of 5 3 106 CFU of the serotype 5 S. aureus strain ST021. Experimental animals injected with a bolus i.v. inoculum of S. aureus cells readily clear the bacteria from the bloodstream. When we injected rats i.v. with 5 3 104 CFU of strain Reynolds, only 10 to 20 CFU/ml could be recovered by quantitative blood cultures performed 60 min after inoculation (15). None-
VOL. 65, 1997
CAPSULAR ANTIBODIES IN S. AUREUS ENDOCARDITIS
theless, animals challenged by this method reproducibly developed endocarditis. This observation suggests that only a few S. aureus cells are necessary to colonize the damaged aortic valve and induce endocarditis. We postulate that immune clearance of a bolus of bacteria is not sufficient to prevent a few organisms from escaping the phagocytic process, adhering to the damaged valve, and initiating the endocardial infection. S. aureus cells possess a number of adhesins (fibronectin-binding proteins A and B, clumping factor, and collagen-binding protein) that have been shown to promote attachment to the damaged aortic valve (7, 9, 14). In this study, we modified the endocarditis model by inoculation of S. aureus by the i.p. route to achieve a gradual bacterial seeding of the bloodstream. This method requires that a greater number of bacteria be used for challenge ($107 CFU/ rat), but bacterial seeding of the damaged valve results from the establishment of a transient bacteremia. Patients who develop S. aureus endocarditis typically have a primary focus of infection (contaminated foreign body, wound, or abscessed organ) that serves to seed the bloodstream with small numbers of bacteria. Our results indicate that passive immunization with antibodies elicited in rabbits by a vaccine composed of S. aureus CP5 and CP8 conjugated to P. aeruginosa exotoxoid A is protective against experimental endocarditis. In addition, quantitative cultures of the blood, kidneys, and aortic valve vegetations from the rats showed that fewer staphylococci were recovered from rats passively immunized with CP-specific IgG than from animals given nonimmune IgG. It is likely that the capsular antibodies were protective because they enhanced opsonophagocytic killing of the circulating staphylococci, thus augmenting blood clearance. This hypothesis is supported by experiments showing that capsular antibodies opsonize S. aureus cells in vitro for phagocytosis and killing (Table 1). The results of antibody quantitation and microscopic examination of peritoneal fluids collected from experimental rats suggest that antibodies within the peritoneal cavity may also enhance local clearance of the bacterial cells, preventing them from gaining access to the bloodstream. To our knowledge, this study is the first demonstration of immune protection against an S. aureus-induced foreign-body infection. A recent study reported by Fattom et al. (5) tested the efficacy of antibodies to CP5 in preventing experimental S. aureus infection in mice. Experimental animals were actively immunized with CP5 conjugated to recombinant exoprotein A from P. aeruginosa or passively immunized with human immunoglobulin obtained from plasma donors vaccinated with the CP5-CP8 conjugate vaccine. Antibodies to the serotype 5 capsule protected mice against lethal challenge with a type 5 S. aureus strain. In addition, the CP-specific immunoglobulin protected the mice against bacteremia from a sublethal staphylococcal inoculum. Animals given capsular antibodies also had reduced numbers of S. aureus organisms in their livers, kidneys, and peritoneal lavage fluids. Ongoing studies in our laboratory suggest that antibodies to the CP5-CP8 conjugate vaccine also protect rats against endocarditis induced by serotype 8 S. aureus (11). S. aureus is a complex bacterial pathogen with an impressive array of secreted and cell-associated virulence determinants (10). Because of the prevalence of antibiotic resistance among clinical isolates, a renewed interest in developing a staphylococcal vaccine has emerged. The challenge for researchers is to define which clinical situations merit use of an immunization strategy and to identify the optimal target antigens for incorEditor: V. A. Fischetti
4151
poration in a vaccine. Although the impact of an S. aureus vaccine on mucosal colonization by S. aureus is difficult to predict, the primary intent of vaccination would be to prevent invasion of the bloodstream by staphylococci at local sites of infection. The results of the present study and that of Fattom et al. (5) suggest that CP5 and CP8 may be important components of a vaccine targeted at preventing serious staphylococcal infection. ACKNOWLEDGMENTS We thank Derek Frederickson and Shu-Min Chang for expert technical assistance. We acknowledge Andrew Onderdonk for providing strain VP from the Brigham & Women’s Hospital and for his assistance in performing the peritoneal fluid cell counts. REFERENCES 1. Baddour, L. M., G. D. Christensen, M. G. Hester, and A. L. Bisno. 1984. Production of experimental endocarditis by coagulase-negative staphylococci: variability in species virulence. J. Infect. Dis. 150:721–727. 2. Bayer, A. S., M. Ing, E. Kim, M. R. Yeaman, S. Shepherd, R. Naso, and A. Fattom. 1996. Role of anticapsular IgG in modifying the course of experimental Staphylococcus aureus endocarditis, abstr. G-096. In Abstracts of the 36th International Conference on Antimicrobial Agents and Chemotherapy 1996. American Society for Microbiology, Washington, D.C. 3. Ekstedt, R. 1963. Studies on immunity to staphylococcal infection in mice. II. Effect of immunization with fractions of Staphylococcus aureus prepared by physical and chemical methods. J. Infect. Dis. 112:152–157. 4. Fattom, A., R. Schneerson, S. C. Szu, W. F. Vann, J. Shiloach, W. W. Karakawa, and J. B. Robbins. 1990. Synthesis and immunologic properties in mice of vaccines composed of Staphylococcus aureus type 5 and type 8 capsular polysaccharides conjugated to Pseudomonas aeruginosa exotoxin A. Infect. Immun. 58:2367–2374. 5. Fattom, A. I., J. Sarwar, A. Ortiz, and R. Naso. 1996. A Staphylococcus aureus capsular polysaccharide (CP) vaccine and CP-specific antibodies protect mice against bacterial challenge. Infect. Immun. 64:1659–1665. 6. Fisher, S. 1960. A heat stable protective staphylococcal antigen. Aust. J. Exp. Biol. 38:479–486. 7. Hienz, S. A., T. Schennings, A. Heimdahl, and J. I. Flock. 1996. Collagen binding of Staphylococcus aureus is a virulence factor in experimental endocarditis. J. Infect. Dis. 174:83–88. 8. Karakawa, W. W., and W. F. Vann. 1982. Capsular polysaccharides of Staphylococcus aureus. Semin. Infect. Dis. 4:285–293. 9. Kuypers, J. M., and R. A. Proctor. 1989. Reduced adherence to traumatized rat heart valves by a low-fibronectin-binding mutant of Staphylococcus aureus. Infect. Immun. 57:2306–2312. 10. Lee, J. C. 1996. The prospects for developing a vaccine against Staphylococcus aureus. Trends Microbiol. 4:162–166. 11. Lee, J. C. Unpublished data. 12. Lee, J. C., N. E. Perez, C. A. Hopkins, and G. B. Pier. 1988. Purified capsular polysaccharide-induced immunity to Staphylococcus aureus infection. J. Infect. Dis. 157:723–730. 13. Manclark, C. R., B. Meade, and D. Burstyn. 1986. Serological response to Bordetella pertussis, p. 388–394. In N. R. Rose, H. Freidman, and J. L. Fahey (ed.), Manual of clinical laboratory immunology, 3rd ed. American Society for Microbiology, Washington, D.C. 14. Moreillon, P., J. M. Entenza, P. Francioli, D. McDevitt, T. J. Foster, P. Francois, and P. Vaudaux. 1995. Role of Staphylococcus aureus coagulase and clumping factor in pathogenesis of experimental endocarditis. Infect. Immun. 63:4738–4743. 15. Nemeth, J., and J. C. Lee. 1995. Antibodies to capsular polysaccharides are not protective against experimental Staphylococcus aureus endocarditis. Infect. Immun. 63:375–380. 16. Santoro, J., and M. E. Levison. 1978. Rat model of experimental endocarditis. Infect. Immun. 19:915–918. 17. Schennings, T., A. Heimdahl, K. Coster, and J. I. Flock. 1993. Immunization with fibronectin binding protein from Staphylococcus aureus protects against experimental endocarditis in rats. Microb. Pathog. 15:227–236. 18. Sompolinsky, D., Z. Samra, W. W. Karakawa, W. F. Vann, R. Schneerson, and Z. Malik. 1985. Encapsulation and capsular types in isolates of Staphylococcus aureus from different sources and relationship to phage types. J. Clin. Microbiol. 22:828–834. 19. Wei, L. J., and W. E. Johnson. 1985. Combining dependent tests with incomplete repeated measurements. Biometrika 17:359–364. 20. Xu, S., R. D. Arbeit, and J. C. Lee. 1992. Phagocytic killing of encapsulated and microencapsulated Staphylococcus aureus by human polymorphonuclear leukocytes. Infect. Immun. 60:1358–1362.
