UvA-DARE ( Digital Academic Repository ) Slaughterhouse Pigs Are a Major Reservoir of Streptococcus suis Serotype 2 Capable of Causing Human Infection in Southern Vietnam

Streptococcus suis is a pathogen of major economic significance to the swine industry and is increasingly recognized as an emerging zoonotic agent in Asia. In Vietnam, S. suis is the leading cause of bacterial meningitis in adult humans. Zoonotic transmission is most frequently associated with serotype 2 strains and occupational exposure to pigs or consumption of infected pork. To gain insight into the role of pigs for human consumption as a reservoir for zoonotic infection in southern Vietnam, we determined the prevalence and diversity of S. suis carriage in healthy slaughterhouse pigs. Nasopharyngeal tonsils were sampled from pigs at slaughterhouses serving six provinces in southern Vietnam and Ho Chi Minh City area from September 2006 to November 2007. Samples were screened by bacterial culture. Isolates of S. suis were serotyped and characterized by multi locus sequence typing (MLST) and pulse field gel electrophoresis (PFGE). Antibiotic susceptibility profiles and associated genetic resistance determinants, and the presence of putative virulence factors were determined. 41% (222/542) of pigs carried S. suis of one or multiple serotypes. 8% (45/542) carried S. suis serotype 2 which was the most common serotype found (45/317 strains, 14%). 80% of serotype 2 strains belonged to the MLST clonal complex 1,which was previously associated with meningitis cases in Vietnam and outbreaks of severe disease in China in 1998 and 2005. These strains clustered with representative strains isolated from patients with meningitis in PFGE analysis, and showed similar antimicrobial resistance and virulence factor profiles. Slaughterhouse pigs are a major reservoir of S. suis serotype 2 capable of causing human infection in southern Vietnam. Strict hygiene at processing facilities, and health education programs addressing food safety and proper handling of pork should be encouraged. Citation: Hoa NT, Chieu TTB, Nga TTT, Dung NV, Campbell J, et al. (2011) Slaughterhouse Pigs Are a Major Reservoir of Streptococcus suis Serotype 2 Capable of Causing Human Infection in Southern Vietnam. PLoS ONE 6(3): e17943. doi:10.1371/journal.pone.0017943 Editor: J. Fitzgerald, University of Edinburgh, United Kingdom Received October 5, 2010; Accepted February 16, 2011; Published March 28, 2011 Copyright: 2011 Hoa et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This research was supported by the Wellcome Trust Major Overseas Programme. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: hoant@oucru.org


Introduction
Infections of Streptococcus suis constitute a major health problem in the swine industry worldwide. In neonatal and infant pigs, S. suis is the cause of outbreaks of septicaemia, meningitis, arthritis and pneumonia whilst in adult pigs asymptomatic nasopharyngeal carriage is well documented. S. suis is also an emerging zoonotic pathogen in Asia, and is increasingly recognized as a leading cause of bacterial meningitis cases and septicaemia in adult humans [1,2,3,4]. Among the 33 described serotypes, serotype 2 is most frequently associated with invasive disease in both pigs and humans [5,6]. Additional serotypes reported to cause invasive disease in pigs include serotypes 1 to 9, 1/2 and 14 [1,3]. However, these strains are rarely found in human infection. Reported risk factors for S. suis infection in humans suggest the importance of both occupational and food-borne transmission [5,7]. We identified consumption of pork with a high risk of contamination with S. suis as an independent risk factor for S. suis meningitis in southern Vietnam (Nghia et al, submitted for publication). Consumption of undercooked pork was identified as the most common risk factor in a retrospective cohort study in 66 patients infected with S. suis in Thailand [7].
Genetic characterization of S. suis serotype 2 strains isolated from human patients with meningitis in southern and central Vietnam, showed a highly conserved population structure [2]. In addition, these strains all carried the genes encoding the putative virulence factors Extracellular Protein Factor (EF) or one of its larger size variants EF*, and the hemolysin (suilysin), whilst 69.6% of these strains produced Muramidase Released Protein (MRP) [2,8,9,10,11]. These findings indicate the existence of a virulent clone of S. suis serotype 2, capable of persisting within the pig population and with a propensity to infect humans. Alternatively, the findings suggest a predominance of S. suis serotype 2, ST1 amongst the S. suis serotypes and sequence types prevalent in the southern Vietnamese pig population. To gain insight into the population structure of S. suis carried in healthy pigs, in particular pigs for human consumption which may constitute a reservoir for zoonotic infection, we determined the prevalence of S. suis serotype 2 and other serotypes in nasopharyngeal tonsils of slaughterhouse pigs in southern Vietnam. Strains were then subjected to molecular genetic characterization to further delineate population structure and to facilitate comparisons between the porcine isolates and representative human clinical isolates from southern and northern Vietnam.

Ethical Statement
The study protocol was reviewed and approved by the Sub Department of Animal Health of Ho Chi Minh City (1242/SNN-NN). All sampling was performed under the supervision of government staff following standard rendering procedures.

Tonsil collection
Tonsils were collected prospectively from pigs submitted to three slaughterhouses receiving animals from 6 provinces in southern Vietnam (Dong Nai, Ben Tre, Binh Duong, Binh Thuan, Long An, Tien Giang) and from the greater Ho Chi Minh City area (Hoc Mon and Cu Chi) ( Figure 1). The number of pigs slaughtered on a daily basis at each slaughterhouse varied between 40-50 and 800-1000.
Pigs were inspected for gross health status and were kept in groups according to delivery by the seller, for a maximum of 16 hours after arrival at the slaughterhouse. The age and province of origin of each pig was recorded. Visually healthy pigs were killed by electrocution in group based batches. Directly after decapitation, the pharyngeal tonsil was removed aseptically and transferred to the laboratory in separate sterile container on ice within 8 hours and stored at 220uC until culture.
Each slaughterhouse was sampled once during repetitive sampling rounds of four-weeks. To ensure sampling from multiple herds originating from multiple provinces, a maximum of 50 tonsils from consecutively slaughtered pigs was collected during each sampling occasion at each slaughterhouse, until a minimum total of 500 tonsils was collected. Sample size calculations indicated that a sample size of 500 tonsils is sufficient to reliably estimate S. suis carriage rates. An observed carriage rate of 50% would result in an exact 95% confidence interval (CI) of 45.5 to 54.4%, whilst a rate of 5% would result in a 95% CI of 3.3 to 7.3%.

Tonsil culture
The inner section of each tonsil was minced on a sterile petri dish. A loop of minced tonsil was streaked onto blood agar plates containing Streptococcal Selective Reagent (Oxoid, UK) and crystal violet. Plates were incubated at 37uC in 5% CO 2 for 16- 18 hours. A maximum of six colonies suspected for S. suis was selected from each plate. S. suis was identified on the basis of colony morphology and alpha-hemolysis, Gram-stain, catalase reactivity, and biochemical identification using API20Strep (BioMerieux, Vietnam). Serotyping was performed by slide agglutination using antisera against all 33 polysaccharide capsule antigens (Statens Serum Institute, Denmark). Multiple colonies of different morphology were tested from each tonsil sample. Isolates were screened using polyvalent antisera, followed by agglutination with appropriate monovalent antisera. Strains S. suis 31533 and Lactococcus lactis spp lactis strain FX256-6 were used for quality control purpose.

Molecular typing
All non-duplicate S. suis serotype 2 strains were characterized by multi locus sequence typing (MLST) and pulse field gel electrophoresis (PFGE) as described previously [2,12]. The multi locus sequence type assignment was based on the sequence of the alleles at each locus of 7 housekeeping genes included in the MLST scheme, using the MLST database [13]. Primer sequences used were as described and published on the MLST website (www. mlst.net), except for the recA and mutS alleles for which adjusted primers were designed with the following sequences; ssrecA-f: ATAGCCACGTTGGTTTGCAG, ssrecA-r: CATAATGAA-CACGTATCTTGCGG; ssmutS-f GAGCAGATGGAAGATCC-TAAGC, ssmutS-r ACAAACTACCATGCTTCTTGCC. MLST results were analyzed using eBURST (www.mlst.net) [13].
PFGE with SmaI digestion was performed as described elsewhere [2] and the images of gels were analysed using Bionumerics (Applied Maths, Belgium). Six strains, each representing one of the six predominant PFGE types which were identified in S. suis serotype 2 strains isolated from human patients with meningitis in southern and northern Vietnam [2,14], were included in the PFGE analysis to facilitate comparison of strains isolated from humans and pigs.

Virulence factor profiling
The presence of genes encoding the putative virulence associated factors EF (epf/epf*), and suilysin (sly) was determined for all strains belonging to the predominant serotypes found, using multiplex PCR as described [15]. MRP expression was determined by Western blot for all serotype 2 strains using polyclonal antibodies against MRP (provided by H. Smith) [16]. S. suis serotype 2 strains 31533 and 89-1591 were used as positive and negative controls, respectively.

Antimicrobial resistance profiling
Antimicrobial susceptibility testing was done on all S. suis serotype 2 isolates using E-test (AB Biodisk, Sweden), according to instruction of the manufacturer. The antimicrobials tested were penicillin, chloramphenicol, erythromycin, tetracycline, ciprofloxacin and vancomycin. Interpretative criteria were used according to CLSI guidelines for viridans group streptococci [17]. Streptococcus pneumoniae strain ATCC 49619 was used for quality control purpose. Erythromycin resistance phenotypes were identified using the triple disk diffusion test as described [18].

Tonsil culture
Nasopharyngeal tonsils from 542 pigs were collected between September 2006 and November 2007 from six southern provinces and the Ho Chi Minh City metropolitan area (Table 1). Three hundred and four (56%) pigs originated from large-scale production units with herd sizes more than 500 head; 142 (26%) pigs originated from smaller semi-intensive farms with average herd sizes of 200 to 500 head, and 96 (18%) pigs were from backyard farms with less than 200 head. The majority of pigs (513, 94.6%) were between 3.5 and 6 months old, and the remaining (29, 5.4%) were sows.
S. suis carrier pigs were identified from all provinces studied ( Figure 1). Prevalence rates of S. suis carriage of all serotypes and of S. suis serotype 2 specifically varied from 14.7% to 48.5% and 5% to 20% respectively, across provinces. Nineteen (6.3%) S. suis serotype 2 strains were isolated from 304 pigs reared in intensive farms;19 (13.4%) from 142 pigs reared in semi-intensive farms, and 7 (7.3%) strains were isolated from 96 pigs reared in backyard farms. The proportion of S. suis serotype 2 strains originating from semi-intensive farms was significantly higher than those from either intensive or small farms (p,0.021, Pearson's chi-square test).

Characterization of S. suis serotype 2 strains
MLST analysis indicated that 36 of 45 (80%) S. suis serotype 2 strains were sequence type 1 (ST1). On eBURST analysis ST1 was the most commonly found sequence type, representing clonal complex 1 (CC1). One single locus variant of ST1 was identified based on detection of a new allele for recA, and this strain was assigned to ST158. Nine strains (20%) were assigned to ST28, belonging to CC27.
On PFGE analysis, strains could be assigned to two major clusters, corresponding with the ST distribution. To facilitate comparison with strains isolated from human patients, 6 S. suis serotype 2 strains of ST1 isolated from human patients in southern Vietnam were included in the PFGE analysis. These strains (indicated by A-F in Figure 2) represent the 6 major PFGE clusters observed during previous PFGE analyses of 128 strains of S. suis serotype 2 isolated from human patients in southern and northern Vietnam [2,14]. All human strains clustered together with isolates from pig tonsils. Four additional clusters (clusters 5, 7, 8 and 13 in Figure 2) contained only strains isolated from pig tonsils. Identical band patterns were observed for strains isolated from pigs reared in different provinces (cluster 4), for strains isolated from pigs slaughtered in different slaughterhouses (cluster 9), and for strains isolated from pigs sampled in different sampling rounds (cluster 10), indicating the clonal spread of epidemiologically unrelated S. suis serotype 2 strains of ST1. However, differences in resistance determinant profiles were observed for strains with identical band patterns within PFGE clusters 8, 10, 11 and 13 ( Figure 2). Among the 9 strains of ST28, 3 clusters were observed in PFGE analysis ( Figure 2). Similar to the strains of ST1, pigs which carried these 9 strains originated from different provinces and were sampled in different sampling rounds.

Discussion
We studied the prevalence of S. suis in slaughterhouse pigs in southern Vietnam by systematic sampling of pig tonsils in three slaughterhouses which received pigs for slaughter from different farms in 6 provinces and HCMC area in southern Vietnam. We were particularly interested in the prevalence of S. suis serotype 2 strains compared to other serotypes as strains carrying the serotype 2 polysaccharide capsule are responsible for the vast majority of human infections in Vietnam as well as worldwide [1,6].
We observed an overall S. suis carriage rate of 41% among pigs entering slaughter facilities in southern Vietnam. The prevalence rate of carriage of the pathogenic serotype 2 was 8%. Colonization of pigs with S. suis occurs at an early stage of life, often through vertical transmission from carrying sows. S. suis carriage rates may vary between herds and can range from 0% to up to 80-100% [22,23,24] The finishing pigs sampled in our study were aged between 3.5 to 6 months and originated from multiple farms in multiple provinces in southern Vietnam. Therefore, the carriage rates of 41% for S. suis and 8% for S. suis serotype 2, are likely to represent a reliable estimate of S. suis carriage prevalence in pigs raised in the region during the sampling period. Our prevalence estimates were based on bacterial culture rather than molecular screening methods, and are thus likely to be a conservative estimate of true prevalence [25] S. suis serotype 2 was the most common serotype isolated from the sampled pigs, indicating that S. suis serotype 2 is highly prevalent in slaughterhouse pigs in southern Vietnam. In contrast, in a study of slaughterhouse pigs in Korea, S. suis serotype 2 strains were absent, whilst serotype 9 was the most common serotype [26]. A large survey amongst 1043 healthy sows in 10 regions in China showed a similar S. suis carriage rate (40.4%) but a lower carriage rate of serotype 2 strains (3.0%) [27]. S. suis serotype 2 strains of MLST CC1 are considered to represent a population of particularly virulent strains causing invasive infection in both pigs and humans [1,2,12]. All S. suis serotype 2 strains isolated from humans in southern Vietnam for which MLST was reported, belonged to this clonal complex, as well as the strains causing outbreaks in China in 1997 and 2005 [2,4,28]. All S. suis serotype 2 strains of ST1 in this study were positive in a PCR detecting the presence of the virulence-associated genes epf/epf* and sly and almost 70% of these strains expressed MRP. These results are consistent with our observations in strains isolated from adult human patients [2]. In addition, strains representing the prevalent genotypes of S. suis serotype 2 isolated from humans, all clustered with the pig carriage isolates on PFGE analysis.
Despite the clonal structure of the S. suis serotype 2 population under study, differences were observed in the antimicrobial resistance determinants content of strains, in particular for genes encoding tetracycline and erythromycin resistance. We observed multiple combinations of tetracycline resistance encoding genes, including the presence of tet(L), in the presence or absence of genes encoding erythromycin resistance (Figure 2). These results indicate circulation of multiple descendants of MLST ST1 lineage in multiple provinces in southern Vietnam.
Human infections with S. suis serotype 2 strains of MLST CC27, in particular strains with ST28, have been reported from Japan and Thailand but to date have not been reported from Vietnam or China [29,30]. The ST28 strains present in our study did not possess the virulence-associated marker genes epf and sly, suggesting that reduced virulence may explain their absence amongst strains causing human infections in Vietnam. However, the possible emergence of ST28 strains in human clinical cases in Vietnam cannot be ruled out considering their presence within the slaughterhouse pig population.
We observed a high prevalence of S. suis strains which reacted with antibodies against capsular antigens of serotypes 9/31/32. Strains were identified using APIStrep, a well validated phenotypic identification method. However, as it is known that S. suis carriage strains are diverse and may be untypeable, additional genotypic confirmation methods for identification are recommended [1]. We therefore studied 7 randomly selected strains in more detail. MLST and16S rDNA sequencing results further confirmed that these strains are indeed S. suis (data not shown). MLST results also suggested that the population of strains reacting with all three of the 9, 31, and 32 antibodies is heterogeneous and it is likely that these strains represent a diverse population carried in healthy pigs (data not shown). None of the strains possessed any of the genes typically used as virulence markers of S. suis. Despite their relatively high prevalence in slaughterhouse pigs in southern Vietnam, strains of serotype 9, 31, 32 have never been reported from human patients in the region nor in human patients in other geographical areas, supporting their presumed limited virulence.
Other serotypes found amongst the S. suis carriage isolates included serotype 3, serotype 21, serotype 7, serotype 16 and serotype 4. Of these, only serotype 16 has been reported in a human patient [31]. Serotype 3 strains were found amongst S. suis strains causing invasive infection in pigs in China [32] but have never been reported to cause disease in humans. Despite its occurrence in human infection in southern Vietnam, as well as in Thailand and North America [33,34,35], strains of serotype 14 were not found in our study population. To the best of our knowledge, this is the first study reporting the prevalence of S. suis in healthy pigs in Vietnam and data on the most prevalent serotypes causing invasive disease in Vietnam are lacking.
In conclusion, our study shows that S. suis carriage is common in healthy slaughterhouse pigs in southern Vietnam and that serotype 2 is the most prevalent serotype. S. suis serotype 2 strains found in slaughterhouse pigs are highly similar to those isolated from human patients indicating that pigs for consumption could be a source of human infection. Human exposure to virulent S. suis strains through handling of pork for consumption may be particularly hazardous in an environment where food items are kept at high ambient temperatures, resulting in high bacterial counts, as is the case in southern Vietnam. Therefore, health education programs addressing the proper handling of pork and other pig products should be encouraged.