Dilemma of virulence of Streptococcus suis: Canadian isolate 89-1591 characterized as a virulent strain using a standardized experimental model in pigs

Abstract

Virulence of Streptococccus suis capsular type 2 strain 89-1591 has been controversial in literature. A standardized experimental model with specific-pathogen free piglets was used for a new evaluation of this strain. Twenty-nine piglets were allotted in 4 separated groups. Group 1 consisted of negative control animals which received broth medium. Groups 2, 3, and 4 were intravenously challenged with 2 mL of S. suis, strains 1330, 89-1591, and 166′, respectively. The strain 1330 is a recognized avirulent Canadian strain. The strain 166′ is a reference French virulent isolate. Pigs inoculated with strain 1330 did not present clinical signs of a S. suis infection. Contamination in organs and bacterial blood circulation were rare and lesions were almost non-existent. Infection of pigs with S. suis strain 89-1591 (group 3) and 166′ (group 4) caused severe clinical problems, animals infected with S. suis 166′ were the most affected. Pigs presented with clinical signs such as high body temperature, lameness, nervous symptoms, and even mortality. Lesions associated with S. suis were numerous for both strains, but more evident in animals of group 4. It can be concluded that S. suis strain 89-1591 is virulent, although its virulence seems to be lower than that of the French strain. Results of an experimental infection with strain 89-1591 may depend on different factors such as the route of inoculation and the immunological status of the animals used. Using conventional animals, with an unknown status regarding previous S. suis infections, equivocal results may be obtained, and this may explain differences reported by some authors with the same strain.

Streptococcus suis is an important cause of septicemia, meningitis, polyarthritis, polyserositis, endocarditis, and sudden death in swine (1,2). It has also been implicated in human disease (3). Among the 35 serotypes described to date, serotype 2 is the most common serotype recovered from diseased animals. However, the prevalence of this serotype seems to be considerably lower in North America than in Europe (4–6). In addition, phenotypic and genotypic characteristics are different between serotype 2 strains isolated from both continents (7). For example, the hemolysin produced by S. suis strains (suilysin), with toxic properties for epithelial, endothelial, and phagocytic cells (8) is seldom produced by serotype 2 strains of S. suis recovered from diseased animals in North America. In fact, it has been hypothesized that S. suis serotype 2 strains may present different pathogenic capacities and the pathogenesis of the infection caused by European and North American strains may differ (9).

Most studies on S. suis virulence factors have been carried out with serotype 2 strains. Although there is confusion in the description of virulence, researchers agree on the existence of virulent and avirulent strains of this serotype (9,10). Different opinions about the definition of “virulence” for S. suis have hampered the studies on virulence factors of this bacterial species, because the concept of virulence may differ depending on the experimental infection model that is used. In fact, there is no universally accepted model, and different research groups use different animal species, pigs with different health status and different ages, different routes of infection, or variable bacterial doses (11). In addition, the classification of virulent strains on the basis of the production of virulence-related proteins, such as the muramidase-released protein (MRP) and extracellular factor (EF), has been challenged by the frequent isolation of MRP and EF negative strains from serious cases of infection (5,12). A standardized swine infection model for S. suis was reported (10). All parameters, such as the status of pigs, specific pathogen free (SPF) piglet; the housing of animals (experimental units into which air is filtrated); the age of animals (7 wk); and the route of inoculation (intravenous), were standardized. Under these conditions, animals inoculated presented clear manifestations of disease (clinical signs, poor production performances, lesions, and bacteriological isolation) similar to those observed in naturally contaminated pigs in herds. Using this model, some differences in virulence were observed when 2 “virulent” strains were compared (10).

As a consequence of a lack of a common experimental infection model, important discrepancies exist in the literature regarding the virulence of even the same strain. For example, the reference strain (S735) was considered a virulent strain by Charland et al (13), but questioned by Vecht et al (14). Another example is the Canadian strain 89-1591. This strain was recovered from a case of septicemia and was isolated from lungs, liver, heart, and spleen of a diseased pig in a Canadian herd (15). It was reported as avirulent by Rasmussen et al (16) and through a personal communication from Dr. Smith, ID-LDO, Lelystad, The Netherlands. In contrast, strain 89-1591 was shown to be virulent in experimental models applied by Quessy et al (17) and Charland et al (18). This strain also caused clinical disease after intravenous infection of Canadian piglets from high health status herds (M. Gottschalk, unpublished observations). These discrepancies could be due to the use of different experimental animal models. Since the genome sequence of S. suis serotype 2 strain 89-1591 is presently under elucidation (19), the aim of this study was to re-evaluate the virulence of strain 89-1591 in our previously described standardized animal model (10, 20), and to compare it to both a Canadian avirulent and a French virulent strains.

Three strains of S. suis capsular type 2 were used in the present experiment. Two strains, 1330 and 89-1591, were isolated from 2 different Canadian herds from pigs presenting with pneumonia and septicemia, respectively (11,17,21). The 3rd strain, 166′, was isolated in a French herd from a pig with meningitis (10,20). The phenotype MRP, EF, and suilysin was determined using previously described methods (5,12,22). Strain 1330 possesses the EF−/MRP+/suilysin− phenotype; strain 89-1591 was characterized as being EF−/MRP−/suilysin−, and strain 166′ presented the EF+/MRP+/suilysin+ phenotype (result not shown).

Preparation of S. suis isolates for inoculation and the animal model have been described previously (10,20). Bacterial suspensions in Todd-Hewitt broth (THB) with 10% of inactivated bovine serum (Difco Laboratories, Detroit, Michigan, USA) were prepared and adjusted to a concentration of 10⁸ colony forming units (CFU)/mL of S. suis, as confirmed by count plating (10,20). Twenty-nine SPF piglets (3 large-white litters) from a closed experimental herd were housed in independent units with absolute filters to prevent any contact with infectious agents. These 7-week-old piglets were randomly grouped into 4 independent units (as described above), 1 wk before infection. The groups 2, 3, and 4 (7 pigs in each group) were intravenously challenged with 2 mL of S. suis strains 1330, 89-1591 and 166′, respectively. The negative control group (group 1: 8 pigs) received THB medium in the same conditions.

Pigs were monitored daily post inoculation (pi) for clinical signs, notably fever; lameness; and dysfunction of the central nervous system such as tremors, opisthotonos, and nystagmus. Rectal temperature was examined every day and body weight was recorded on days 0, 3, 6, and 8 pi. Necropsies were planned in a haphazard way, on days 2, 6, 8, and 14 after infection (1 or 2 pigs from each group) and lesions were examined. However, 2 piglets (groups 3 and 4) were sacrificed for welfare reasons (severe clinical signs, such as dysfunction of the central nervous system were observed). Finally, pigs were sacrificed as described in Table I. Presence of S. suis in blood samples (on days 1, 2, 3, 6, 8, and 14 pi, for infected groups and until 8 d for group 1) and in organs (post-mortem examinations) was monitored, as previously described: samples were diluted in THB and plated onto Colombia blood agar base (10,20).

Table I.

Schedule of post-mortem examinations

	Groups of pigsa
Days post infection	2	3	4
2	2b	2 (1c)	1
3	—	—	3 (1c)
4	—	—	1
6	2	1	2
8	2	2	—
14	1	2	—
Total of animals	7	7	7

^aAnimals of groups 2, 3, and 4 were intravenously infected with 3 different strains of Streptococcus suis (group 2: 1330; group 3: 89-1591; group 4:166′). Animals of group 1 (negative control) received broth medium and were sacrificed 8 d post-infection

^bNumber of sacrificed animals

^cPigs died naturally 2 or 3 d after experimental infection

Data from the 4 groups were compared simultaneously by using an analysis of variance (ANOVA) (Kruskall-Wallis test) with computer software (Systat, version 7.0; Systat Inc., Evanston, Illinois, USA). When values were significantly different between 2 pig groups, comparisons were performed using the Mann-Whitney U test.

Negative control animals (group 1) did not exhibit clinical signs of an S. suis infection. Post mortem examinations did not reveal any pathological changes in these animals and no S. suis could be isolated during or after the trial.

Infection of pigs with the strain 1330 (group 2) was mild and no clinical signs were observed. Rectal temperature (39.4°C ± 0.0°C to 39.8°C ± 0.2°C) of pigs and daily growth were normal and similar to those observed for the control group (P > 0.05). Streptococcus suis was isolated in only 8/32 blood samples and bacterial levels were low (< 2.8 ± 3.1 CFU [log₁₀]/mL) (Table II). In fact, no bacteria could be recovered from blood on day 6 post-infection. Post-mortem examinations revealed S. suis in only 1 animal presenting abundant synovial liquid in 1 joint. In fact, isolation of S. suis 1330 was rare except from palatine tonsils (Table III). Results confirm previous studies showing a low virulence potential of this strain. In fact, its use as an avirulent live vaccine has also been proposed (23).

Table II.

Results of Streptococcus suis concentration in blood during the experiment

	Groups of pigsa
Days after Infection	1 (control)	2 (isolate 1330)	3 (isolate 89-1591)	4 (isolate 166′)	Statistical analysis
1	0 ± 0 (0/8)	1.0 ± 1.4 (1/7)b	4.2 ± 4.3 (7/7)	4.7 ± 4.6 (7/7)	P = 0.001
2	0 ± 0 (0/8)	1.9 ± 2.1 (2/7)	4.6 ± 4.6 (6/6)	4.7 ± 4.8 (7/7)	P = 0.001
3	0 ± 0 (0/8)	2.8 ± 3.1 (3/5)	4.3 ± 4.4 (5/5)	3.4 ± 3.3 (4/4)	P = 0.001
6	0 ± 0 (0/8)	0 ± 0 (0/5)	4.9 ± 5.1 (5/5)	3.6 ± 3.6 (2/2)	P = 0.001

^aGroup 1 was intravenously inoculated with broth medium. Groups 2, 3, and 4 were intravenously infected with S. suis strains 1330, 89-1591, and 166′, respectively

^bStreptococcus suis concentration (CFU log₁₀/mL) ± standard error (number of pigs)

Table III.

Histological and bacteriological results after Streptococcus suis infections of pigs (post-mortem results)

	Macroscopic lesionsa							Bacteriological resultsb
Groupsc	Lymph nodes	Liver	Heart	Spleen	Lung	Joints	Brain	Lymph nodes	Liver	Heart	Spleen	Lung	Joints	Brain	Tonsils
Isolate 1330 (Group 2; n = 7)	0	0	0	0	0	1	0	2	0	0	1	0	1	0	5
Isolate 89-1591 (Group 3; n = 7)	2	1	1	0	0	6	4	5	2	5	4	5	4	4	4
Isolate 166′ (Group 4; n = 7)	5	4	4	5	6	7	6	7	7	7	7	7	7	7	7

^aNumber of animals with lesions in different organs at necropsy

^bNumber of animals with S. suis isolated from different organs

^cAnimals of groups 2, 3, and 4 (7 pigs in each group) were intravenously infected with S. suis. Animals of group 1 (8 pigs) received broth medium in the same conditions and were sacrificed 8 d after experimental inoculation; no lesions were recorded in these animals and S. suis was not isolated from the different organs described above

Infection of pigs with S. suis strain 89-1591 (group 3) and 166′ (group 4) was severe, animals infected with S. suis 166′ being the most affected. Body temperatures of these groups were higher than 40.0°C from day 1 to day 6 pi, with a peak at 41.7°C ± 0.2°C in group 4. During this time, rectal temperatures in groups 3 and 4 differed significantly (P < 0.05) from those of groups 1 and 2. Significant differences (P = 0.001) were also observed between pigs infected with S. suis 89-1591 and 166′ on days 1 and 2 pi, but not on the following days.

On day 1, all infected pigs in groups 3 and 4 presented lameness and most of them had swollen joints. Nervous signs, such as opisthotonos and incoordination of movements, were also recorded in animals of both groups. Clinical signs relating to the nervous system were more severe in pigs of group 4. In each group, 1 animal died naturally, between days 2 and 3 pi. Bacterial numbers in blood were significantly higher in animals from groups 3 and 4 than those recovered from group 2 (Table II). These differences (P < 0.05) were recovered between groups 3 and 4 only on day 1 pi. Relatively high numbers of bacteria could be recovered from blood of infected animals up to 6 d after infection. Then, the isolation of S. suis was very rare.

At necropsy, gross lesions varied depending on the strain of S. suis (Table III), but, in general, were similar to those already described (24). Animals from group 4 had the highest frequency of lesions in most of the organs studied, especially in lymph nodes, liver, heart, spleen, and lung. Similar levels of lesions in brain and joints were observed for both strains. Postmortem examination revealed arthritis, meningitis, pericarditis, endocarditis, perisplenitis, and perihepatitis. The iliac and mesenteric nodes were enlarged and hemorrhagic. Pulmonary lesions included interstitial fibrinous or fibrinopurulent pneumonia, pleurisy, and pleuritis. Streptococcus suis was also isolated from different tissues, especially from animals infected with strain 166′. Indeed, all organs examined were infected in group 4 (Table III).

Before the experimental infection, the 29 pigs showed a similar body weight (from 20.3 ± 3.1 kg to 20.8 ± 3.5 kg). After inoculation, the body weight gain was lower in pigs of groups 3 and 4 from days 3 to 8 pi, than in groups 1 and 2. Pigs of group 4 showed the lowest mean body weight, 4.9 to 9.6 kg less than that of groups 1 and 2 and 2.8 to 6.6 kg less than that of group 3.

The results confirm previous results showing that this specific animal model is able to differentiate virulence levels between 2 S. suis strains (10). It can be concluded that S. suis strain 89-1591 is virulent, especially when compared to strain 1330, which also has a Canadian origin. Indeed, animals inoculated with S. suis strain 89-1591 presented symptoms, lesions in organs, and S. suis was isolated from blood and organs. However, its virulence seems to be lower than that of the French strain. It is possible that avirulent strains (such as strain 1330) and highly virulent strains (such as strain 166′) would behave similarly in different animal models. In the case of strain 89-1591, the virulence may depend on the route of inoculation, number of CFU, and, mainly, the immunological status of the animals used. In this particular study, highly susceptible animals were used. With conventional animals, with an unknown status regarding previous S. suis infections, equivocal results may be obtained, and this may explain differences reported by some authors with the same strain (16–18). The hypothesis that S. suis serotype 2 European EF+, MRP+, and hemolysin+ strains are more virulent than Canadian EF−, MRP−, and hemolysin-strains remains to be confirmed with a higher number of isolates compared simultaneously in the same animal model. As discussed by Gottschalk and Segura (9), a certain association of these proteins (especially the EF and hemolysin proteins) with virulence seems to exist, and most isolates harbouring these factors are probably virulent. However, the absence of 1 or more of these proteins cannot necessarily be associated with a lack of virulence. It is also possible that the virulence of typical Canadian virulent strains is amplified in herds affected by the highly virulent North American porcine respiratory and reproductive syndrome virus (25,26).