Staphylococcus aureus Capsular Polysaccharide Types 5 and 8 Reduce Killing by Bovine Neutrophils In Vitro

Abstract

Isogenic variants of Staphylococcus aureus strain Reynolds expressing either no capsule or capsular polysaccharide (CP) type 5 (CP5) or type 8 (CP8) were used to assess the effect of CP on bacterial killing and the respiratory burst of bovine neutrophils. The effects of antisera specific for CP5 and CP8 were also evaluated. The killing of live bacteria by isolated neutrophils was quantified in a bactericidal assay, while the respiratory burst after stimulation with live bacteria in whole blood was measured by flow cytometry. The expression of a CP5 or CP8 capsule protected the bacteria from being killed by bovine neutrophils in vitro (P < 0.001), and the capsule-expressing variants did not stimulate respiratory burst activity in calf whole blood. The addition of serotype-specific antisera increased the killing of the capsule-expressing bacteria and enhanced their stimulating effect in the respiratory burst assay (P < 0.01). When the S. aureus variants were grown under conditions known not to promote capsule expression, there were no significant differences between them. The present study demonstrates that the expression of S. aureus CP5 or CP8 confers resistance to opsonophagocytic killing and prevents the bacteria from inducing respiratory burst of bovine neutrophils in vitro and that these effects can be reversed by the addition of serotype-specific antisera.

Many bacterial species have well-characterized capsules which are important virulence factors that potentiate infections (20). In addition to being a major human pathogen, Staphylococcus aureus is the cause of a variety of infections in animals. Most important is mastitis of dairy cows, which compromises animal welfare and causes large economic losses for the dairy industry. S. aureus produces various surface polysaccharides (18, 19, 22, 40), and most strains express capsular polysaccharides (CPs) in vivo or under defined culture conditions (24). Although at least 11 serotypes based on CPs have been proposed, only four CPs (CP1, CP2, CP5, and CP8) have been chemically characterized (15).

Phagocytosis and killing by neutrophil granulocytes play a key role in defense against S. aureus infections. CP1 and CP2 have been shown to have antiphagocytic properties (9, 17, 25, 41), but such heavily capsulated strains are rarely isolated clinically (2, 31, 43). Most clinical isolates from both human and bovine infections produce either CP5 or CP8, although considerable geographic variations exist in the prevalence of these types in bovine isolates (26, 32, 37).

The relatively small amount of CP produced by serotype 5 and 8 strains has made it difficult to define the role of these capsules in virulence. However, the antiphagocytic properties of the S. aureus CP5 capsule, as measured in in vitro opsonophagocytic killing assays, have been shown to play an important role in the pathogenesis of S. aureus infections, most likely by allowing the organism to resist uptake and killing by phagocytes (21, 35).

Opsonization is of major importance for effective phagocytosis of most bacterial pathogens. Complement and antibodies are the principal serum opsonins, and anti-CP antibodies enhance uptake and killing of capsule-expressing S. aureus by human neutrophils. Vaccines targeted against S. aureus can stimulate the production of antibodies against CP and enhance bacterial killing (16). A combined CP5 and CP8 conjugate vaccine has been evaluated in a phase III clinical study (6, 30), and antibodies specific for CPs have been shown to protect against S. aureus infections in murine models (7). A number of vaccines, some also including various polysaccharide surface antigens, have been designed and used against S. aureus mastitis of dairy cows (14, 23, 39). Both CP5 conjugate and whole-cell vaccines stimulate antibody responses in cattle (8, 38). However, the effects of CP5- and CP8-specific antibodies on the respiratory burst of and phagocytosis by bovine neutrophils have not been evaluated.

The genes involved in capsule biosynthesis by serotype 5 and 8 strains are chromosomal and allelic (29). The predicted amino acid sequences of the cap5 and cap8 gene clusters are almost identical; however, four open reading frames located in the central region bear little homology to each other and are type specific. The isogenic variants used in the present study are the progeny of serotype 5 strain Reynolds and were constructed by deletion of the type-specific region, creating a capsule-negative variant, or by substitution with the corresponding region of the cap8 gene cluster, creating a CP8-producing variant (J. C. Lee, submitted for publication).

In the present study, bacterial killing by isolated bovine neutrophils was quantified in a bactericidal assay, and respiratory burst was measured by a newly described flow cytometric method. In addition to the advantage of using bacteria with intact surface structures, the assays are also suitable for the evaluation of the opsonizing capacity of serum against bacteria with different surface features. The aim of the study was to investigate how the expression of CP5 and CP8 influenced the respiratory burst and the opsonophagocytic killing of S. aureus by bovine neutrophils and whether serotype-specific antibodies had opsonizing properties.

MATERIALS AND METHODS

Bacterial strains.

S. aureus strain Reynolds, a clinical human blood culture isolate, is the prototype CP5-producing strain (13). The isogenic S. aureus variants used in this study are progeny of strain Reynolds and were kindly donated by Jean C. Lee (Channing Laboratory, Harvard Medical School, Boston, Mass.). JL278 is a CP5-producing variant, and the capsule-negative variant JL801 was constructed by deletion of the four type-specific open reading frames located in the central region of the cap5 gene cluster (cap5H through cap5K). CP8-producing variant JL812 was created by substitution of cap5H through cap5K with cap8H through cap8K of the cap8 gene cluster (Lee, submitted). CP production was quantified by an enzyme-linked immunosorbent assay inhibition method as previously described (35, 37). CP5- and CP8-expressing strains JL278 and JL812 produced 259 μg of CP5 per 10¹⁰ CFU and 306 μg of CP8 per 10¹⁰ CFU, respectively. Strain JL801 produced less than 0.7 μg of CP5 per 10¹⁰ CFU (Lee, submitted).

Preparation of bacteria for use in assays.

S. aureus isogenic variants JL278, JL801, and JL812 were grown overnight at 37°C on Mueller-Hinton agar (Difco Laboratories, Detroit, Mich.) containing 2% NaCl to promote capsule production. Single colonies were transferred to new plates, spread, incubated at 37°C for 24 h, removed from the agar, washed in phosphate-buffered 0.9% saline (pH 7.4), centrifuged at 2,500 × g, and plated to determine bacterial concentrations. The bacteria were suspended in heart infusion broth (Difco) containing 15% glycerol at a concentration of 2.5 × 10¹⁰ CFU/ml and stored at −70°C. Before use, the bacteria were diluted in phosphate-buffered 0.9% saline (pH 7.4) or RPMI 1640 without phenol red (Sigma Chemical Co., St. Louis, Mo.) to the required concentrations. The same S. aureus variants were also grown in tryptic soy broth (TSB) (Difco) at 37°C for 12 h, conditions known not to promote capsule expression, and thereafter treated similarly.

Bovine anti-CP sera.

Anti-CP5 serum was obtained from a cow immunized with a CP5-human serum albumin conjugate vaccine (38). Ten months after primary immunization, a booster injection was given. Blood was collected 4 weeks after the booster injection, and serum was stored at −20°C until use. Anti-CP8 serum was prepared from a cow immunized twice with inactivated whole bacteria of CP8-producing S. aureus strain Wright. Anti-CP titers were determined by an enzyme-linked immunosorbent assay as previously described (38). The serum end-point titer was defined as the reciprocal of the serum dilution at which the optical density (OD) at 405 nm was lower than the cutoff value, which was defined as the mean OD of negative control sera plus 2 standard deviations at the lowest serum dilution tested (1:64). The anti-CP5 serum had a titer of 1:16,384 compared to a preimmunization sample with a tite of ≤1:64, while the anti-CP8 titer was 1:8,192. Normal bovine sera were obtained from eight nonimmunized cows, pooled, and stored at −20°C until use. For the bactericidal assay, all sera were heated to 56°C for 30 min, whereas for the flow cytometric whole-blood respiratory burst assay, freshly thawed sera were used.

Animals and blood samples.

Six 2- to 4-year-old, clinically healthy dairy cows, three newborn calves, and eight 1- to 2-month-old calves of the Norwegian dairy cattle breed were used in the present study. Blood was collected with acid-citrate-dextrose or heparin as an anticoagulant. All calf sera had anti-CP titers of ≤1:64, while the titers of the adult cow sera varied between ≤1:64 and 1:256.

Bactericidal assay.

The bactericidal assay was performed essentially as described by Stevens et al. (33). Neutrophils were isolated as described previously (4). One hundred microliters of a suspension of neutrophils (1.0 × 10⁷ cells/ml in RPMI 1640), 50 μl of a suspension of live S. aureus (5.0 × 10⁸ CFU/ml), and 7.5 μl of heat-inactivated serum were incubated at 37°C for 1 h in a 96-well microtiter plate. After incubation, the neutrophils were lysed by the addition of 50 μl of a 0.2% saponin solution (Sigma). Fifty microliters of a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) solution (2 mg/ml) was added to each well, and the plate was incubated for 10 min at room temperature to allow the formation of formazan. The plate was centrifuged, the supernatant was discarded, the precipitate was dissolved in dimethyl sulfoxide, and the absorbance was measured at 550 nm with a microtiter plate reader. The proportion of bacteria killed in each well was determined by comparing the OD with that of a standard curve; the results are reported as the percentage of bacteria killed. All samples were tested in triplicate, and the median was used.

Flow cytometric respiratory burst assay.

The flow cytometric respiratory burst assay was performed essentially as described by Kampen et al. (11). One hundred microliters of heparinized whole blood was incubated for 30 min at 38.5°C with 20 μl of a bacterial suspension (1.0 × 10⁹ CFU/ml) to promote respiratory burst activity. Subsequently, the samples were placed on ice, 20 μl of 100 μM dihydrorhodamine 123 (Molecular Probes, Eugene, Oreg.) was added, and the samples were incubated for an additional 30 min at 38.5°C. Following incubation, the erythrocytes were lysed and the leukocytes were fixed with fluorescence-activated cell sorting lysing solution (Becton Dickinson Biosciences, San José, Calif.). Negative control samples with no stimulant added were prepared from each animal. To investigate the effect of serum addition, 7 μl of either anti-CP5, anti-CP8, or preimmunization serum was added to each tube before incubation. Serial twofold dilutions of anti-CP5 and anti-CP8 sera were also tested. Data were collected from 10,000 cells per sample by use of a FACSCalibur flow cytometer (Becton Dickinson Biosciences) with a 488-nm-wavelength argon laser. A gate was set on the granulocytes, analysis was done with CellQuestPro software (Becton Dickinson Biosciences), and the results are reported as the increase in the geometric mean green fluorescence of all of the gated cells. The mean for duplicate samples was used.

Statistical methods.

The results are presented as the mean and standard error of the mean (SEM) for each group. Groups were compared by using a two-tailed Student t test with a significance level of 5%.

RESULTS

A total of 73.4% (SEM, 2.2) of the capsule-negative bacteria were killed in the MTT bactericidal assay, whereas 2.3% (1.5) of the CP5-producing variants and 10.1% (1.4) of the CP8-producing variants were killed when incubated with neutrophils and heat-inactivated pooled bovine serum (Fig. 1). With the addition of anti-CP5 serum, 61.5% (4.5) of the CP5-producing bacteria were killed, while the proportion of CP8-producing bacteria killed remained unaffected. When anti-CP8 serum was added, 31.3% (3.2) of the CP8-producing bacteria were killed, and the proportion of CP5-producing bacteria killed remained low (Fig. 1). The effect of bacterial capsule production on bacterial resistance to opsonophagocytic killing by isolated neutrophils was significant for both the CP5- and CP8-producing strains (P < 0.001), as was the enhancing effect of serotype-specific antiserum on the killing of each of the capsule-producing strains (P < 0.001).

FIG. 1.

Percentages of noncapsulated, CP5-producing, and CP8-producing S. aureus organisms killed by cow and calf neutrophils in the MTT bactericidal assay with heat-inactivated pooled bovine serum, anti-CP5 serum, or anti-CP8 serum. The results are means (SEMs) for five to eight animals.

The capsule-negative bacteria stimulated respiratory burst activity in whole blood from normal calves in the flow cytometric assay. The increase in the geometric mean fluorescence intensity (GMFI) of 274.5 (SEM, 102.0) in response to the noncapsulated bacteria was significantly higher than that for each of the capsule-producing variants (P < 0.05). The CP5- and CP8-producing variants stimulated little or no activity, producing increases in the GMFI of 13.5 (8.0) and 13.8 (4.8), respectively (Fig. 2 and 3). The addition of anti-CP5 serum to calf whole blood resulted in a stimulating effect of the CP5-producing strain, giving a GMFI increase of 358.8 (120.3), while the response to the CP8-producing strain was unaffected. Likewise, the addition of anti-CP8 serum resulted in a stimulating effect of the CP8-producing strain, giving a GMFI increase of 216.5 (77.9), while there was only a minor response to the CP5-producing strain (Fig. 2 and 3). Preimmunization sera elicited no response.

FIG. 2.

Increase in the fluorescence (GMFI) of calf whole-blood samples stimulated with noncapsulated (Non-capsul.), CP5-producing, or CP8-producing S. aureus without in vitro serum addition or with anti-CP5 serum or anti-CP8 serum supplementation in the respiratory burst assay. The results are means (SEMs) for six or seven animals.

FIG. 3.

(A) Histogram showing green fluorescence of a calf whole-blood sample stimulated with noncapsulated, CP5-producing, or CP8-producing S. aureus in the flow cytometric respiratory burst assay (grey) plotted against the fluorescence of a nonstimulated sample (transparent). (B and C) Histograms showing the green fluorescence of a calf whole-blood sample stimulated with noncapsulated, CP5-producing, or CP8-producing S. aureus with the addition of anti-CP5 serum (grey) or preimmunization serum (transparent) (B) and with the addition of anti-CP8 serum (grey) or preimmunization serum (transparent) (C).

Testing of serial dilutions of anti-CP5 and anti-CP8 sera in the respiratory burst assay showed that stimulation decreased with increasing dilutions of sera (Fig. 4).

FIG. 4.

Increase in the fluorescence (GMFI) with twofold dilutions of anti-CP5 serum (A) or anti-CP8 serum (B) added to calf whole blood stimulated with CP5- or CP8-producing S. aureus in the respiratory burst assay, respectively. The results are means (SEMs) for four animals.

The results of the respiratory burst assay with whole blood from newborn calves collected before intake of colostrum were comparable to those obtained with blood from older calves (Table 1). The noncapsulated strain stimulated respiratory burst activity, whereas the two capsule-producing strains did not. For each of the capsule-producing strains, the addition of capsule-specific antiserum resulted in values higher than those for the noncapsulated strain (Table 1).

TABLE 1.

Respiratory burst assay results for newborn calves^a

Serum	Increase in GMFI in neutrophil granulocytes after stimulation with S. aureus
	Noncapsulated	CP5 capsule	CP8 capsule
None	361.7 (59.7)	6.7 (2.9)	12.2 (3.9)
Anti-CP5	438.8 (51.4)	636.8 (13.3)	12.7 (2.4)
Anti-CP8	473.2 (17.4)	18.7 (7.2)	679.0 (115.0)

^aWhole-blood samples obtained from newborn calves before intake of colostrum were examined in the respiratory burst assay following stimulation with noncapsulated, CP5-producing, and CP8-producing S. aureus and without in vitro serum addition or with anti-CP5 serum or anti-CP8 serum supplementation. The results are the means (SEMs) for three animals.

The noncapsulated, CP5-producing, and CP8-producing strains all stimulated respiratory burst activity in whole blood from six adult cows, with GMFI increases of 416.9 (SEM, 54.1), 362.9 (58.2), and 462.3 (81.4), respectively, and there were no significant differences among the strains. Substitution of the plasma in the calf blood samples with pooled bovine serum resulted in values similar to those obtained with blood from adult cows (data not shown). Three of the calves were repeatedly sampled until the age of 1 year, and the responses to the capsulated strains and the anti-CP titers increased to adult levels during this period (data not shown).

When the bacteria were grown in TSB for 12 h, there were no significant differences between the bacterial variants. The strains genetically capable of producing CP5 and CP8 were killed by cow and calf neutrophils in the bactericidal assay and stimulated respiratory burst activity in calf whole blood to the same extent as the noncapsulated strain (Table 2).

TABLE 2.

Effect of growth in TSB on assay results^a

Assay (measured unit)	Assay result obtained with S. aureus
	Noncapsulated	CP5	CP8
MTT bactericidal (% bacteria killed)	80.5 (2.3)	75.4 (1.3)	77.8 (0.7)
Respiratory burst (increase in GMFI)	356.0 (46.6)	345.9 (49.6)	336.0 (28.3)

^aThe percentage of bacteria killed by isolated neutrophils in the MTT bactericidal assay and the increase in the fluorescence in the whole-blood respiratory burst assay were evaluated with noncapsulated S. aureus and variants genetically capable of CP5 or CP8 expression grown in TSB under conditions known not to promote capsule production. The results are the means (SEMs) for four animals.

DISCUSSION

The present study shows that the expression of S. aureus CP5 and CP8 confers resistance to opsonophagocytic killing and prevents the bacteria from inducing respiratory burst activity in bovine neutrophils in vitro and that these effects can be reversed by the addition of serotype-specific antisera.

A large proportion of the noncapsulated S. aureus variant was killed in the bactericidal assay, whereas for the capsulated variants, few bacteria were killed by bovine neutrophils. The capsule-negative bacteria stimulated respiratory burst activity in calf neutrophils, while the capsule-producing variants stimulated little or no activity. In both assays, the effect of capsule production was significant. CPs have been shown to protect S. aureus from phagocytosis and killing by human neutrophils in vitro (12, 35). The CP layer is believed to make surface antigens and opsonins attached to the bacterial surface inaccessible to phagocyte receptors and thereby to prevent phagocytosis (42).

The addition of CP antisera significantly increased the killing of the capsule-expressing bacteria and enhanced the stimulating effects of these bacteria in the respiratory burst assay, showing that antibodies against CP5 and CP8 are opsonic in cattle. Antisera against CP5 and CP8 have been shown to protect against S. aureus infections in murine models (7). There was no apparent cross-reactivity between the CP5 and CP8 antisera used in the present study, indicating that CP vaccines would confer type-specific protection and that knowledge of the prevalences of the different serotypes in the target population is important.

A small percentage of the CP8-producing strain was killed in the bactericidal assay without the addition of specific antiserum, whereas killing of the CP5-producing strain was negligible. This result might reflect a background level of CP8 antibodies in the serum used, as more than 90% of Norwegian bovine S. aureus isolates belong to serotype 8 (36) and most cattle are likely to have been exposed to CP8-producing strains. However, low titers of antibodies against both CP5 and CP8 were found in sera from the cows used in the present study.

The anti-CP8 serum did not enhance bacterial killing by isolated neutrophils to the same extent as the anti-CP5 serum. This result might be due to the fact that these antisera were produced with different immunization protocols.

In accordance with the findings of Barrio et al. (3), heat-inactivated antisera were sufficiently effective for the opsonization of bacteria in the MTT bactericidal assay. In contrast, Thakker et al. (35) reported that both antibodies and active complement were required for effective phagocytosis and killing of capsule-expressing S. aureus by human neutrophils.

In cows, the placenta does not allow the passage of antibodies, and calves are born with little or no immunoglobulin. The newborn calves in the present study had a marked response to the noncapsulated strain in the whole-blood respiratory burst assay, indicating that antibodies are not required for the opsonization of noncapsulated S. aureus.

In contrast to the findings obtained with calf whole blood, both the noncapsulated and the capsule-producing strains stimulated respiratory burst activity in adult cow whole blood. Adult cattle are likely to have a higher background level of antibodies than calves, but the addition of a small volume of preimmunization cow serum to calf blood samples did not enhance the calf neutrophil response. However, total substitution of the plasma in calf whole-blood samples with pooled serum from adult cows yielded results comparable to those obtained with blood from adult cows, indicating that the observed effect was due to inherent opsonins in the cow sera. The responses to the capsulated strains and the anti-CP titers in calves increased to adult levels during the first year of life. There was no difference between isolated neutrophils from cows and calves in the bactericidal assay, in which serum addition was standardized. When whole-blood methods or assays based on serum addition are used, one should be aware that the combined effects of the intrinsic function of neutrophils and the opsonizing capacity of serum are being measured.

When grown under conditions known not to promote capsule production, S. aureus variants genetically capable of producing CP5 and CP8 were killed by cow and calf neutrophils in the bactericidal assay and stimulated respiratory burst activity in calf whole blood to the same extent as the noncapsulated strain. The expression of S. aureus CP5 and CP8 is greatly influenced by environmental and bacterial growth conditions, such as the culture medium and the growth phase of the organism (5, 27, 34). The growth of CP5-producing strains on solid agar media has been shown to increase CP5 production about 100-fold over that in liquid-grown cultures (35). The differences in the production of CPs by S. aureus strains grown under different conditions are probably the main reason for the inconsistencies in the results of in vitro studies on antiphagocytic capabilities and virulence (1, 35).

There are also differences in the amounts of CPs produced by strains of the same serotype (27, 28, 37). The strain used in the present study is the prototype CP5-producing strain. However, other capsulated strains might not display the same properties and might influence host immune responses differently, even though the CPs of capsule-producing strains of the same serotype are chemically identical. Clinical isolates of S. aureus from cows with mastitis vary in their capsule expression (37), and individual bacteria of the same isolate might display different amounts of CPs (28). S. aureus capsule production might be different in vivo and in vitro and is possibly enhanced by growth in milk (18). CP expression has been demonstrated in situ in experimental bovine mastitis (10), and further studies are needed in order to examine the effects of variations in capsule production on host defense mechanisms.

The present study demonstrates that the expression of CP5 and CP8 protects S. aureus from killing by bovine neutrophils and that anticapsular antibodies are type specific and opsonic for bovine neutrophils. This knowledge is important for the understanding of the pathogenesis of S. aureus infections and the development of efficient vaccines.