Invasion of Porcine Brain Microvascular Endothelial Cells by Streptococcus suis Serotype 2

Abstract

Streptococcus suis is an important swine pathogen that mainly causes meningitis and occasionally causes other infections, such as endocarditis, arthritis, and pneumonia. The pathogenesis of S. suis infection has not been completely defined. However, in order to cause meningitis, S. suis has to cross the blood-brain barrier (BBB) made up of brain microvascular endothelial cells. The objective of this work was to study the interactions of S. suis serotype 2 with porcine brain microvascular endothelial cells (PBMEC). The ability of North American and European S. suis serotype 2 strains to adhere to PBMEC and, most importantly, to invade PBMEC was demonstrated by using an antibiotic protection assay and was confirmed by electron microscopy. The polysaccharide capsule of S. suis seemed to partially interfere with the adhesion and invasion abilities of the bacterium. Our results showed that intracellular viable S. suis could be found in PBMEC up to 7 h after antibiotic treatment. Inhibition studies demonstrated that invasion of PBMEC by S. suis required actin microfilaments but not microtubular cytoskeletal elements or active bacterial RNA or protein synthesis. At high bacterial doses, suilysin-positive strains were toxic for PBMEC. The role of suilysin in cytotoxicity was confirmed by using purified suilysin, electron microscopy, and the lack of toxicity of a suilysin-negative mutant. In swine, the invasion of endothelial cells of the BBB could play an important role in the pathogenesis of the meningitis caused by S. suis.

Streptococcus suis serotype 2 is an important swine pathogen that mainly causes meningitis and occasionally causes other infections, such as endocarditis, arthritis, and pneumonia (21). Of the 35 serotypes described, serotype 2 is the serotype that is most frequently associated with disease. This organism is also recognized as a zoonotic agent (2). Knowledge concerning the virulence factors of S. suis is still limited. Several molecules have been suggested to be virulence factors; these molecules include the capsule polysaccharide (CPS), a hemolysin (suilysin) (19, 27), a 136-kDa muramidase-released protein (MRP), and a 110-kDa extracellular factor (EF) protein (52). Recently, it has been proposed that a fibronectin- and fibrinogen-binding protein plays a role in the colonization of affected organs after experimental infection (13). So far, the CPS is the only proven critical virulence factor since unencapsulated isogenic mutants were shown to be completely avirulent and rapidly cleared from the circulation in both pig and mouse models of infection (8, 45). The CPS is composed of five different sugars, rhamnose, galactose, glucose, N-acetylglucosamine, and N-acetylneuraminic acid (sialic acid), and it defines the serotype (15). However, natural nonvirulent S. suis serotype 2 strains are also encapsulated and have an amount of sialic acid in the CPS similar to the amount found in virulent strains (9). On the other hand, suilysin, MRP, and EF protein have been associated with the virulent phenotype of European strains, but they are absent in most virulent North American strains (11, 17, 41). While the role of MRP and EF protein is unknown (46, 47), suilysin is involved in cytotoxic effects for several types of cells, including epithelial, endothelial, and phagocytic cells (10, 29, 35, 42, 43). This hemolysin belongs to the family of cholesterol-binding toxins with a multihit mechanism of action (19). It has also been suggested that the pathogenesis of the infection caused by suilysin-positive strains and the pathogenesis of the infection caused by suilysin-negative strains may be different (18).

The pathogenesis of S. suis infection is not fully understood, and many steps are probably involved in this process. In order to reach the central nervous system (CNS) and cause meningitis, circulating S. suis has to cross the blood-brain barrier (BBB). This barrier is responsible for maintaining the homeostasis within the CNS and is characterized by tight intercellular junctions that regulate the movement of cells, solutes, and macromolecules across the BBB (26, 50). The BBB is composed of the brain microvascular endothelial cells (BMEC), which are closely associated with pericytes and outgrowths of astrocytes (the so-called astrocytic end feet), as well as the epithelial cells of the choroid plexus (26). The primary site of breakdown of the BBB in most bacterial meningitis appears to be the BMEC (49). However, it is not clear how circulating S. suis cells cross the BBB. Moreover, S. suis serotype 2 has been shown to adhere to and activate human BMEC (10, 51), but unlike other meningeal pathogens, such as Escherichia coli K1, Streptococcus pneumoniae, and group B Streptococcus (GBS), invasion does not happen in this model (25, 50). Interestingly, it has been shown that the capacities of adhesion and invasion by E. coli K1 (36) and GBS (29) varied for different types of endothelial and epithelial cells, respectively, suggesting that there are cell-specific interactions. In this regard, the adhesion of S. suis seems to be specific for certain cell types since the organism does not significantly adhere to human umbilical vein endothelial cells (HUVEC) (10).

In an attempt to further understand the pathogenesis of meningitis caused S. suis in swine, the objective of this study was to evaluate the ability of S. suis serotype 2 to adhere to, invade, and damage porcine brain microvascular endothelial cells (PBMEC).

MATERIALS AND METHODS

Bacterial strains and growth conditions.

S. suis serotype 2 suilysin-positive MRP⁺ EF⁺ virulent strain 31533 (28) was used as the reference strain in this study. In selected experiments, S. suis serotype 2 suilysin-negative MRP⁻ EF⁻ virulent strain 89-1591 was used. Both of these strains have previously been used in studies with human brain microvascular endothelial cells (HBMEC) (10). Two isogenic mutants derived from strain 31533, suilysin-negative mutant SX911 obtained by allelic replacement (kindly provided by P. Willson, Veterinary Infectious Disease Organization, Saskatoon, Canada) (30) and unencapsulated mutant B218 obtained in our laboratory by allelic exchange and corresponding to a transposon-derived mutant described previously (8), were also used in this study. S. suis isolates used in comparative toxicity studies are listed in Table 1. Bacteria were grown overnight on sheep blood agar plates at 37°C, and isolated colonies were used as inocula for Todd-Hewitt broth (THB) (Difco Laboratories, Detroit, Mich.), which was incubated for 8 h at 37°C with agitation. Working cultures were obtained by inoculating 10 μl of 10⁻³ dilutions of the cultures into 30 ml of THB and incubating the cultures for 16 h at 37°C with agitation. Stationary-phase bacteria were washed twice in phosphate-buffered saline (PBS) (pH 7.3) and were appropriately diluted in cell culture medium before infection (see Results). The number of CFU per milliliter in the final suspension was determined by plating samples onto THB agar. Encapsulated GBS type III strain COH1 (kindly provided by C. Rubens, Children's Hospital and Regional Medical Center, Seattle, Wash.) was used in selected experiments for comparison purposes. In addition, Streptococcus gordonii strain Challis was used as a negative control (10, 34). S. suis mutant SX911 was grown in the presence of 200 mg of spectinomycin (Sigma, Oakville, Ontario, Canada) per ml.

TABLE 1.

Cytotoxicities of different S. suis serotype 2 strains for PBMEC, as determined by measurement of LDH release^a

Strain	Geographic origin or type	Suilysin production	Cytotoxicity (%)
89-1591	Canada	−	6 ± 5
90-1330	Canada	−	4 ± 5
98-B575	Canada	−	5 ± 7
98-B099	Canada	−	1 ± 3
98-8993	Canada	−	0 ± 0
94-623	France	−	3 ± 4
SX332	United States	+	96 ± 4
95-8242	Canada	+	74 ± 14
166′	France	+	91 ± 8
31533b	France	+	53 ± 17
B218	Mutantc	+	72 ± 12
SX911	Mutantc	−	0 ± 0

^aThe cytotoxicity assay was performed by measuring LDH release after 4 h of incubation in the presence of S. suis (10⁸ CFU/ml).

^bReference strain used in this study.

^cThe isogenic unencapsulated mutant B218 and the isogenic suilysin-negative mutant SX911 were derived from S. suis strain 31533 by allelic exchange.

Cell culture.

The PBMEC line PBMEC/C1-2 was previously immortalized by lipofection with pRNS-1 encoding the small and large T-antigens of simian virus 40 and was shown to maintain its morphological and functional characteristics (48). Cells were grown in complete IF medium, which is a 1:1 mixture of Iscove's modified Dulbecco's medium and Ham's F-12 medium (Gibco, Burlington, Vt.) supplemented with 7.5% (vol/vol) heat-inactivated fetal bovine serum, penicillin-streptomycin (Gibco), sodium bicarbonate, l-glutamine, human transferrin (ICN Biomedical Inc., Aurora, Ohio), N-acetyl-cysteine, hypoxanthine, porcine heparin, human recombinant fibroblast growth factor-basic (Sigma), and β-mercaptoethanol (Bio-Rad Laboratories, Mississauga, Ontario, Canada). Flasks (Falcon; Becton Dickinson, Bedford, Mass.) and Primaria 96- and 24-well tissue culture plates (Falcon) were precoated with type A gelatin (1%) from porcine skin (Sigma) to support the cells. Cells were incubated at 37°C with 5% CO₂ in a humid atmosphere and were used before passage 20 for all experiments. For assays, cells were trypsinized by adding a 0.05% trypsin-0.03% EDTA solution (Gibco) and were diluted in culture medium to obtain a concentration of 8 × 10⁴ cells/ml, and the suspension was distributed into tissue culture plates and incubated until confluence was reached (48 h). Just before the experiments, the medium was removed from the plates and was replaced by medium without antibiotics.

PBMEC invasion and adhesion assays.

The invasion assay was performed as previously described (10, 34), with some modifications. Stationary-phase bacteria were pelleted, washed twice with PBS, and resuspended in fresh cell culture medium without antibiotics at different concentrations (see Results). Confluent monolayers of PBMEC grown in 24-well plates were infected with 1-ml aliquots of a bacterial suspension. The plates were centrifuged at 800 × g for 10 min to bring the bacteria to the surface of the monolayer and incubated for different times (see Results) at 37°C with 5% CO₂ to allow cell invasion by the bacteria. The monolayers were then washed twice with PBS, and 1 ml of cell culture medium containing 100 μg of gentamicin per ml and 5 μg of penicillin G (Sigma) per ml was added to each well. The plates were then incubated for 1 h at 37°C with 5% CO₂ to kill extracellular and surface-adherent bacteria. The monolayers were washed three times with PBS and incubated for 10 min at 37°C in the presence of 200 μl of 0.05% trypsin-0.03% EDTA. After this incubation period, 800 μl of ice-cold deionized water was added, and the cells were disrupted by scrapping the bottom of the well and by repeated pipetting to liberate intracellular bacteria. Serial dilutions of the cell lysate were plated onto THB agar and incubated overnight at 37°C. To confirm that 100% of the extracellular bacteria were killed after the antibiotic treatment, a 100-μl sample of the last PBS wash solution was plated on THB agar (data not shown). Levels of invasion were expressed as the total number of CFU recovered per well.

An adhesion assay to quantify the total cell-associated bacteria (intracellular bacteria plus surface-adherent bacteria) was performed like the invasion assay, but the cells were vigorously washed five times to eliminate nonspecific bacterial attachment and no antibiotic treatment to kill the extracellular bacteria was used. The levels of adhesion were expressed as the total number of CFU recovered per well.

Intracellular survival assay.

An invasion assay was performed as described above, except that after a 2-h invasion period, the initial gentamicin-penicillin G treatment was lengthened for different times up to 7 h. To ensure that the number of intracellular bacteria was not affected by the entry of antibiotics during the incubation period, some experiments were performed by reducing the antibiotic concentrations to 25 μg of gentamicin per ml and 1.25 μg of penicillin G per ml after an initial incubation for 1 h in the presence of 100 μg of gentamicin per ml and 5 μg of penicillin G per ml. The data were expressed as the total number of CFU recovered per well after antibiotic treatment.

Invasion inhibition studies.

For experiments to test the effects of bacterial RNA and protein synthesis inhibitors, bacteria were pretreated with rifampin and tetracycline (Sigma), respectively. S. suis strain 31533 (10⁶ CFU/ml) was treated with the appropriate concentration of inhibitor at 37°C for 30 min prior to infection, as well as during the 2-h invasion period. The MICs (determined by using NCCLS recommendations) for strain 31533 were 0.03 μg/ml for rifampin and 1.0 μg/ml for tetracycline. In addition to MICs, concentrations that were one-half and two times the MICs were also used for each antibiotic.

For experiments in which the effects of microfilament and microtubule formation inhibitors were tested, PBMEC monolayers were preincubated with cytochalasin D and colchicine (Sigma), respectively. Different concentrations of the inhibitors were added at 37°C for 30 min prior to infection with strain 31533 (10⁶ CFU/ml), as well as during the 2-h invasion period. Cytochalasin D concentrations of 0.25, 0.5, 1.0, and 2.0 μg/ml and colchicine concentrations of 0.5, 1.0, 5.0, 10, and 20 μg/ml were used. The results were expressed as the percentage of invasion relative to the level of invasion without inhibitor (considered 100% invasion).

PBMEC cytotoxicity assay.

The cytotoxic effects of bacteria were evaluated by measuring the release of lactate dehydrogenase (LDH) enzyme as previously described (34), with some modifications. Briefly, bacteria were grown and diluted as described above, and cells grown in 96-well plates were infected with 100-μl aliquots of a bacterial suspension at different concentrations. The plates were centrifuged at 800 × g for 10 min to bring the bacteria to the surface of each monolayer and were then incubated for different times at 37°C with 5% CO₂ (see Results). Noninfected cells and bacteria in IF medium without a PBMEC monolayer were used as negative controls, whereas cells lysed in 0.025% Triton X-100 were used as positive controls (100% toxicity). At the end of the incubation period, a 50-μl aliquot of each supernatant was transferred to a 96-well plate, which was centrifuged at 2,300 × g for 20 min to pellet the bacteria. LDH was measured by using 20-μl aliquots of each centrifuged supernatant and a miniaturized version of the Sigma colorimetric assay as described by Nizet et al. (33). The percentage of cytotoxicity was calculated as follows: [(sample OD₄₁₄ − OD_0%)/(OD_100% − OD_0%)] × 100, where sample OD₄₁₄ is the optical density at 414 nm of the sample, OD_0% is the optical density at 414 nm of noninfected cells, and OD_100% is the optical density at 414 nm of Triton X-100-lysed cells. Purified suilysin (kindly provided by T. Jacobs, Intervet International, Boxmeer, The Netherlands) was also evaluated at different concentrations (1 to 5 μg/ml) in cell medium without antibiotics. The suilysin was reactivated by addition of 2-mercaptoethanol (1%) to the culture medium during the assay (27). In parallel, 2-mercaptoethanol (1%) was added to the culture medium to ensure that it was not toxic to cells (data not shown).

Electron microscopy studies.

For transmission electron microscopy (TEM), PBMEC monolayers were grown on 13-mm Thermanox coverslips in a 24-well culture plate. Invasion, adhesion, and cytotoxicity assays were performed as described above. After two washes with PBS, the monolayers were fixed for 1 h at room temperature with 2% (vol/vol) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3), and then samples were postfixed in 2% (vol/vol) osmium tetroxide in deionized water. Specimens were dehydrated in a graded series of ethanol solutions and embedded in ferm Spurr resin. Thin sections were cut with a diamond knife by using a Leica Ultracut ultramicrotome and were poststained with uranyl acetate and lead citrate. Samples were observed with a model 420 electron microscope (Philips Electronics). For scanning electron microscopy (SEM), samples were processed like the samples for TEM were processed, except that they were not postfixed with osmium tetroxide. Samples for SEM were dehydrated in a graded series of ethanol solutions and covered with gold after critical point drying and were examined with a Hitachi S-3000N microscope.

Statistical analysis.

All data are expressed as means ± standard deviations. Data were analyzed by a two-tailed, unpaired t test. A P value of <0.05 was considered significant. All assays were repeated at least three times.

RESULTS

Some S. suis strains can damage PBMEC. In order to determine if S. suis could be cytotoxic to PBMEC, LDH release measurements were obtained. As shown in Table 1, all suilysin-producing strains were toxic for PBMEC, while suilysin-negative strains were not toxic. In addition, suilysin-negative mutant SX911 was not toxic, in contrast to its parental strain 31533. The role of suilysin was confirmed by adding the purified hemolysin at a concentration of 1 to 5 μg/ml to cells. Suilysin injured PBMEC in a concentration-dependent manner, and the highest toxicity observed occurred at a concentration of 5 μg/ml (97% ± 4%). Kinetic studies showed that suilysin-positive strain 31533 injured PBMEC in a time-dependent manner (Fig. 1A). In contrast, suilysin-negative strain 89-1591 was not cytotoxic during the same period. As shown in Fig. 1B, cytotoxicity caused by suilysin-positive strain 31533 after 4 h of incubation was concentration dependent, while suilysin-negative strain 89-1591 remained noncytotoxic even at a concentration of 10⁸ CFU/ml. However, it should be noted that the levels of cytotoxicity of both strains (Fig. 1B) were very low and similar at concentrations ranging from 10⁴ to 10⁷ CFU/ml (P > 0.05). At a concentration of 10⁸ CFU/ml, the level of cytotoxicity due to suilysin-positive strain 31533 was significantly higher than the levels of cytotoxicity due to mutant SX911 and strain 89-1591 (P < 0.001). Interestingly, similar levels of toxicity were obtained with the unencapsulated mutant B218 and its parental strain 31533 (P > 0.05), indicating that there was no additive toxic effect due to surface-exposed cell wall components. TEM was used to confirm the toxicity of suilysin-positive strain 31533 to PBMEC (Fig. 2C). Injury was manifested by a loss of cell membrane integrity and by disappearance of the nucleus. The PBMEC integrity after 2 h of incubation with strain 89-1591 (Fig. 2B) was comparable to the integrity of noninfected control cells (Fig. 2A). In the context of these results, the bacterial concentration used for adhesion and invasion assays (10⁶ CFU/ml) was not toxic to cells.

FIG. 1.

(A) Kinetics of cytotoxicity of S. suis (10⁸ CFU/ml) for PBMEC. (B) Effect of S. suis concentration on PBMEC injury. The cytotoxic effect of bacteria was evaluated by measuring LDH release in the presence of different concentrations of an S. suis strain after 4 h of incubation. The data are the percentages of cytotoxicity in infected wells compared with the cytotoxicity in control wells with PBMEC alone. The error bars indicate standard deviations. See Table 1 for a description of the strains.

FIG. 2.

TEM micrographs showing the PBMEC injury caused by S. suis at a concentration of 10⁸ CFU/ml. (A) Noninfected control cells. Bar = 2 μm. (B and C) Cells incubated for 2 h with suilysin-negative strain 89-1591 (B) or suilysin-positive strain 31533 (C). Bars = 1 μm. The PBMEC integrity after 2 h of incubation with strain 89-1591 was comparable to that of noninfected control cells. Injury was manifested by a loss of cell membrane integrity and disappearance of the nucleus. The arrow indicates an S. suis 89-1591 coccus. The arrowheads indicate disruption of the cell membrane. M, cell membrane; N, nucleus.

S. suis adheres to PBMEC.

The kinetics of adhesion to PBMEC were compared for different strains of S. suis. As shown in Fig. 3A to D, adhesion to PBMEC was time dependent. The kinetics of adhesion were similar for strains 31533 and 89-1591 (P > 0.05), and they reached a plateau after between 4 and 6 h of incubation. Experiments were performed to optimize and standardize the adhesion assay for S. suis by using the lowest bacterial concentration that allowed recovery of significant numbers of colonies on THB agar. These conditions were optimal for an inoculum of 10⁶ CFU of PBMEC per well (a multiplicity of infection of approximately 4:1). Unencapsulated mutant B218 exhibited earlier adhesion and a higher level of adhesion to PBMEC (P < 0.01) than its parental strain 31533 (Fig. 3A and B). Suilysin-negative mutant SX911 adhered to PBMEC in a way similar (P > 0.05) to the way that parental strain 31533 adhered during this period (Fig. 3A and C), reaching a plateau after between 4 and 6 h of incubation. Adhesion of S. suis to PBMEC was confirmed by TEM (Fig. 2B) and SEM (Fig. 4A).

FIG. 3.

(A to D) Kinetics of S. suis adhesion to PBMEC. Results were determined after exposure of PBMEC to S. suis (10⁶ CFU/ml), followed by extensive washing of nonadherent bacteria and cell lysis to obtain S. suis viable plate counts on THB agar. (E to H) Kinetics of PBMEC invasion by S. suis (10⁶ CFU/ml). Results were determined as described above except that after washing, the bacteria and cells were exposed to antibiotics to kill extracellular bacteria. See Table 1 for a description of the strains.

FIG. 4.

SEM micrographs showing S. suis (10⁶ CFU/ml) interactions with PBMEC after 5 min (A) and 2 h (B and C) of incubation. First, cocci were shown to adhere to PBMEC (A). Next, they were observed in close contact with cells (B). Finally, streptococci were found within invagination structures, as well as underneath the cell surface, behind the PBMEC membrane (C). (A and C) Bars = 1 μm. (B) Bar = 1.5 μm. The arrows indicate S. suis cocci. The arrowheads indicate S. suis cocci behind the cell membrane.

S. suis invades PBMEC.

The kinetics of invasion of PBMEC were examined for different strains of S. suis at a concentration of 10⁶ CFU/ml (Fig. 3E to H). The levels of invasion of PBMEC by S. suis strains did not increase significantly during the period of time studied, except for strain 31533 between 4 and 6 h (P < 0.01). The increase in the level of invasion was not due to an increase in extracellular bacterial growth (data not shown). The levels of invasion for strain 31533 were higher (P < 0.05) than those observed with strain 89-1591. Similar to the results for adhesion, unencapsulated mutant B218 exhibited earlier invasion and a higher level of invasion of PBMEC (P < 0.05) than parental strain 31533, and saturated levels of invasion were observed during the experiment (Fig. 3E and F). Invasion of PBMEC by suilysin-negative mutant SX911 was similar to that by parental strain 31533 (P > 0.05) (Fig. 3E and G). No cytotoxicity was detected when cells were incubated for more than 4 h with the bacterial concentration used for the adhesion and invasion studies.

Invasion of PBMEC by S. suis was confirmed by SEM (Fig. 4) and TEM (Fig. 5). Figure 4 shows sequential interactions between S. suis and PBMEC. First, after 5 min of incubation, cocci were found to adhere to PBMEC (Fig. 4A). Next, after 2 h of incubation, they were observed in close contact with cells (Fig. 4B) and underneath the cell surface, behind the PBMEC membrane (Fig. 4C). Figure 5 confirms that there was intracellular invasion of PBMEC by S. suis. First, S. suis cells were observed in close contact with the PBMEC and within an invagination (Fig. 5A). Finally, Fig. 5B shows an intracellular S. suis coccus in a membrane-bound vacuole within a PBMEC.

FIG. 5.

TEM micrographs showing invasion of PBMEC by S. suis at a concentration of 10⁶CFU/ml after 4 h of incubation. S. suis cells were observed in close contact with PBMEC and within invaginations (A). An intracellular S. suis coccus is present in a membrane-bound vacuole in a PBMEC in panel B. (A) Bar = 1 μm. (B) Bar = 1.25 μm. The arrows indicate S. suis cocci. M, cell membrane; N, nucleus.

Time courses of PBMEC invasion were also determined for GBS strain COH1 at a concentration of 10⁶ CFU/ml (data not shown). The invasion resulted in a peak of internalization after 4 h of incubation (1.7 × 10⁵ ± 0.7 × 10⁵ CFU recovered/well). The level of invasion observed for GBS was higher than the levels of invasion observed with S. suis strains (P < 0.01). S. gordonnii strain Challis at concentrations up to 10⁸ CFU/ml was used as a negative control for invasion assays, and no bacteria were recovered during the 6-h invasion period (data not shown).

S. suis intracellular survival in PBMEC.

As shown in Fig. 6, after an initial 2-h invasion period and a minimal 1-h antibiotic treatment, viable S. suis strains 31533 and 89-1591 could be found inside PBMEC up to 7 h of additional incubation, which was performed in the presence of antibiotics to avoid any extracellular source of bacteria. The results showed that there were similar (P > 0.05) gradual decreases in the number of viable intracellular bacteria for the two strains. In experiments in which reduced concentrations of gentamicin and penicillin G were used during the survival assay, no significant differences in the numbers of viable intracellular bacteria compared to the numbers obtained with the usual antibiotic doses were observed (P > 0.05) (data not shown). In contrast to the decreasing levels of S. suis, the intracellular GBS levels remained constant throughout this period (2.7 × 10⁴ ± 0.9 × 10⁴ CFU recovered/well), confirming that the decreasing levels of viable intracellular S. suis cells were not related to antibiotic uptake.

FIG. 6.

Intracellular survival of S. suis in PBMEC. After 2 h of invasion by S. suis (10⁶ CFU/ml), the initial gentamicin-penicillin G treatment was lengthened for different times up to 7 h, and cells were lysed to quantify intracellular bacteria by viable plate counting on THB agar. An asterisk indicates the incubation time from which the number of intracellular bacteria recovered per well is significantly different (P < 0.05) from number of intracellular bacteria obtained after an initial 1-h antibiotic treatment.

Effect of bacterial RNA and protein synthesis inhibitors on PBMEC invasion by S. suis.

To evaluate the possible role of bacterial RNA and protein synthesis in S. suis invasion of PBMEC, bacteria were pretreated with rifampin or tetracycline. When used at its MIC, neither rifampin nor tetracycline inhibited PBMEC invasion by S. suis strain 31533 (90% ± 37% and 100% ± 40% invasion, respectively [P > 0.05]). Similar levels of invasion were observed with concentrations corresponding to one-half and two times the MICs of the two antibiotics (data not shown).

Effects of microfilament and microtubule formation inhibitors on PBMEC invasion by S. suis.

To evaluate the possible roles of microfilaments and microtubules in S. suis invasion of PBMEC, cells were pretreated with cytochalasin D and colchicine, respectively. As shown in Fig. 7, cytochalasin D inhibited the invasion of PBMEC by S. suis strain 31533 in a concentration-dependent manner. At a concentration of 0.5 μg/ml, more than 60% inhibition of invasion was observed (P < 0.05). The highest level of inhibition (more than 80%) was observed at a concentration of 1.0 μg/ml or higher. On the other hand, colchicine did not inhibit the invasion of PBMEC by S. suis strain 31533 (P > 0.05) despite the use of inhibitor concentrations up to 20 μg/ml (Fig. 7). None of the products used as inhibitors of PBMEC invasion was found to be toxic for cells at the concentrations used (data not shown).

FIG. 7.

Inhibition of invasion of PBMEC by S. suis strain 31533 with cytochalasin D and colchicine. PBMEC monolayers were preincubated with the appropriate concentrations of inhibitors at 37°C for 30 min prior to infection, as well as during the 2-h invasion period. An asterisk indicates that the P value is <0.05 for a comparison with the level of invasion without inhibitor (considered 100% invasion).

DISCUSSION

The pathogenesis of meningitis caused by S. suis is poorly understood. The presence of high levels of bacteria in the bloodstream of diseased animals is correlated with the presence of clinical signs and symptoms in such animals (6). In order to gain access to the CNS and cause the disease, circulating S. suis has to cross the BBB. In fact, S. suis is frequently isolated from brains of diseased pigs with clinical meningitis (22, 31). Interactions between BMEC and S. suis have not been completely characterized, and no data are available on such interactions with cells of porcine origin. Indeed, this study shows for the first time that S. suis is able to interact with PBMEC, and this interaction may play an important role in the development of meningitis.

Adhesion of S. suis to PBMEC is time dependent, as previously reported by Charland et al. for HBMEC (10). Despite the fact that the presence of the capsule does not seem to prevent bacterium-cell contact, a significantly higher rate of adherence was observed in the absence of CPS, as observed with unencapsulated S. suis mutant B218. Similar observations were reported previously for S. suis with epithelial cells of swine origin (29). An adhesin(s) responsible for cell adhesion might be present in the cell wall and, although exposed at the bacterial cell surface, might be partially covered by the CPS. Thus, preliminary studies showed that purified S. suis cell wall partially inhibits attachment of whole bacteria to PBMEC (unpublished observations). Cell wall components of other well-encapsulated bacterial pathogens, such as S. pneumoniae and group A and B streptococci, have been described as being responsible for adhesion to host cells (1, 5, 32). Interestingly, a previous study showed that an S. suis strain and its unencapsulated mutant adhered similarly to human endothelial cells (10). As discussed below, adhesins and cell receptors involved in the interaction between S. suis and endothelial cells of human and swine origin may be different.

This study demonstrated that S. suis does invade PBMEC. To the best of our knowledge, this is the first report of S. suis invasion of endothelial cells of porcine brain origin. Invasion of BMEC by free bacteria has also been reported for other meningeal pathogens, like GBS (34), E. coli K1 (3), S. pneumoniae (38), Citrobacter freundii (4), and Listeria monocytogenes (54). In contrast, previous studies showed that although S. suis was able to adhere to and activate HBMEC, it was not able to invade this type of cells (10, 51). Interestingly, these studies were carried out with strains 31533 and 89-1591, which were also used in the present work. The fact that S. suis invades PBMEC and not HBMEC could be related to the cell origin, since the natural host for S. suis is the pig. On the other hand, the lack of invasion of the HBMEC cell line could be related to the lack of an S. suis-specific receptor(s) for invasion at the cell surface. While adhesion without invasion was reported by Lalonde et al. (29) for epithelial cells of porcine origin and human origin, Norton et al. demonstrated by using differential fluorescence that some S. suis strains were able to adhere to and invade (as a rare event) human Hep-2 epithelial cells (35). Thus, the specific and clear invasion of PBMEC by S. suis reported here may have a significant impact on our understanding of the pathogenesis of S. suis meningitis. Adhesion to and invasion of PBMEC by S. suis were confirmed by SEM and TEM. First, cocci were shown to adhere to PBMEC by SEM and TEM. Later, S. suis bacteria were found within invagination structures, as well as behind the PBMEC membrane (by SEM) and inside a membrane-bound vacuole within a PBMEC (by TEM).

The presence of a capsule seems to partially interfere with invasion since unencapsulated S. suis mutant B218 invaded PBMEC to a greater extent than the parental encapsulated strain. Similar interference of the capsule with invasion has been reported for HUVEC invasion by Haemophilus influenzae (53) and for BMEC invasion by S. pneumoniae (38). It should be noted that the higher level of invasion by the unencapsulated S. suis mutant may be due to the higher level of adherence of this mutant to PBMEC. As mentioned above, it is possible that the polysaccharide capsule causes steric interference with certain bacterial cell wall receptor-ligand interactions which may be important in the invasion process. Interestingly, as observed in adhesion studies, preliminary data showed that purified S. suis cell wall partially inhibits PBMEC invasion by S. suis (unpublished observations). The fact that the S. suis CPS is essential for resisting phagocytosis but not for the BBB invasion process is in agreement with data obtained with E. coli K1 (24). On the other hand, it has been reported recently that the hemolysin produced by GBS contributes significantly to cell invasion (14). This is not the case for S. suis, as the results obtained in this study showed that the suilysin-negative mutant has in vitro invasion properties similar to those of its parental strain.

Our results showed that S. suis is able to survive, to a certain extent, inside PBMEC. However, it should be noted that the number of intracellular viable S. suis cells gradually decreased. In contrast to the levels of S. suis, the levels of viable GBS remained constant during the same period in PBMEC. These results are in agreement with the constant levels of viable intracellular GBS in BMEC reported by Nizet et al. (34). The ability to cross an in vitro transwell model of BBB has been reported for S. pneumoniae in BMEC, although a clear in vitro decrease in the number of internalized viable cocci was reported for this pathogen (38). In fact, it has been suggested that many invading pathogens enter endothelial cells in a transient process (23).

In order to preliminarily characterize S. suis invasion of PBMEC, bacteria were pretreated with bacterial RNA and protein synthesis inhibitors. Results obtained in this study suggest that active bacterial RNA and protein synthesis may not be essential for penetration of the PBMEC. This is in agreement with results reported by Virji et al. for invasion of HUVEC by H. influenzae (53). However, Nizet et al. (34) reported that invasion of BMEC by GBS was dependent on such synthesis. Thus, the requirement for novel RNA and protein synthesis for invasion of host cells is not a generalized mechanism in bacterial species. In this case, the interaction of a putative surface adhesin that is already present with a cell receptor would be sufficient to trigger bacterial internalization (25, 49). This was further demonstrated by Sinha et al. (44), who showed that Staphylococcus aureus cell invasion occurred even with formaldehyde-treated bacteria. To evaluate the possible role of microfilaments and microtubules in S. suis invasion of PBMEC, eukaryotic cells were pretreated with the appropriate inhibitors. Cytochalasin D, but not colchicine, inhibited invasion, suggesting that actin microfilaments of the host cytoskeleton, but not microtubular cytoskeletal elements, are required for internalization of S. suis. This is in agreement with the actin-dependent and microtubule-independent mechanism of HUVEC invasion by S. pyogenes and GBS reported by Greco et al. (20).

In addition to adhesion and invasion, under certain conditions and with certain strains, bacterial cytotoxicity was observed. According to results obtained in this study, suilysin seems to be the S. suis factor responsible for the in vitro PBMEC cytotoxicity. Indeed, only suilysin-positive strains were cytotoxic for PBMEC. Moreover, the suilysin-negative mutant SX911 was shown to be noncytotoxic even at a high concentration. In previous studies workers have reported toxicity of suilysin for human endothelial cells (10), swine epithelial cells (29, 35), and murine and human phagocytic cells (42, 43). Use of purified suilysin confirmed the toxic potential of this cholesterol-binding toxin. Interestingly, an unencapsulated mutant did not have a higher cytotoxic capacity than the encapsulated parent strain, indicating that the well-exposed cell wall does not seem to be toxic to PBMEC. Similarly, although the pneumococcal cell wall was shown to be toxic for HUVEC (16), it has been recently demonstrated that the pneumolysin (another member of the cholesterol-binding toxin family) is the only factor of S. pneumoniae that is responsible for the toxicity for bovine BMEC (55). Like the effects of other toxins, the combined toxic effect with inflammatory potential (7, 12, 14, 37, 39) recently reported for the suilysin (30, 51) may play an important role not only in BBB permeability but also by increasing leukocyte influx. Indeed, histopathological findings indicating that there is necrosis of vessel walls in association with inflammatory cellular aggregates have been reported for S. suis-affected pigs (40).

In conclusion, in this paper we describe for the first time the interactions between S. suis and PBMEC. In contrast to what was reported previously for HBMEC, S. suis is able to invade BMEC of swine origin. Further studies are needed to characterize the molecule(s) responsible for adherence to and/or invasion of PBMEC. In swine, invasion of endothelial cells forming the BBB could play an important role in the pathogenesis of meningitis caused by S. suis. The capacity to invade combined with the suilysin-related cytotoxicity (for suilysin-positive strains) may have a direct impact on the pathological potential of an S. suis strain.