Characterisation of Streptococcus suis Isolates in the Czech Republic Collected from Diseased Pigs in the Years 2018–2022

As in other countries, in the Czech Republic, Streptococcus suis infection in pigs is considered an economically significant disease for the pig industry, though little is known about its population structure. We collected S. suis isolates from 144 farms in the years 2018–2022. All samples were taken from animals suffering from symptoms indicating possible S. suis infection. Serotyping revealed the presence of 23 different serotypes, and 18.94% were non-typable strains. The most common was S7 (14.96%), while other serotypes had frequencies of less than 10%. Sequence typing identified 56 different sequence types, including 31 newly assigned sequence types together with 41 new alleles in genes in the MLST schema. A large portion of isolates (25.70%) were of unknown sequence type. The most common sequence types were ST29 (14.77%) and ST28 (10.04%); the other sequence types had frequencies of less than 10%. In total, 100 different combinations of serotypes and sequence types were identified. Among them, S7ST29 was found in 72 isolates, representing 13.63% of all isolates, and was significantly associated with the central nervous system. Many other isolates of particular serotype and sequence type combinations were found in a few cases, and a number of isolates were non-typable.


Introduction
Streptococcus suis is known as a major swine pathogen that is responsible for significant economic losses in the swine industry. Pathogenic strains are associated with meningitis, arthritis, endocarditis, polyserositis, and septicaemia in piglets and growing pigs [1,2]. Some pathogenic S. suis strains have zoonotic potential, causing meningitis in humans [3]. Pigs can be carriers of commensal strains that colonise the upper respiratory tract. Some strains can appear as opportunistic pathogens associated with coinfection with other bacterial or viral pathogens, but if S. suis is isolated from the tissue, indicating systemic infection (brain, meninges, heart, and joints), it is considered the primary pathogen [1,4,5].
The S. suis population is highly heterogeneous, as different serotypes, phenotypes, and genotypes have been found [6]. Serotyping is the most basic method used to distinguish strains. Currently, there are 29 S. suis serotypes [7], which have been described based on serological reactions against capsular polysaccharide (CPS). Methods traditionally used for serotyping are the agglutination and coagglutination tests using serotype-specific antisera [8,9]. However, these serological methods are expensive, time-consuming, and not always clearly readable [10]. Recently, PCR methods targeting the cps locus, enabling the rapid serotyping of S. suis strains, have been developed [11,12]. However, with standard serological testing or with later-developed specific PCRs, serotype 2 cannot be clearly distinguished from serotype 1/2 and serotype 1 cannot be distinguished from serotype 14 due to the similar genetic contents of their cps loci and their serological cross-reactivity [10,13,14]. Differentiation of these serotypes is very important, because serotypes 2 and 14 are associated with human diseases [3]. Additional methods, such as high-resolution melting analysis [15] or a simple and low-cost method for detecting cpsK gene polymorphisms based on PCR restriction fragment length polymorphisms (PCR-RFLPs) [16], must be applied to distinguish the mentioned serotypes. Moreover, a whole-genome sequencing (WGS) technique can be used to accurately identify strains [17]. The predominant S.suis serotypes isolated from clinical cases in pigs are serotypes 1/2, 1, 2, 6, 7, 9, and 14 [15,18,19], notably in Europe. Besides these serotypes, serotypes 3 and 8 are frequently isolated in North America [20,21]. However, a relatively high percentage of S. suis isolates remain non-typable [15,19]. It was shown that more serotypes could be isolated simultaneously in one animal and that different genotypes of the same serotype could be isolated at one timepoint from the same animal [22].
Multilocus sequence typing (MLST) is a method commonly used for the genotyping of pathogens, and numerous sequence types (STs) exist within the S. suis species [6]. The MLST method is widely implemented in research to determine the population diversity and global distribution and to understand the epidemiology of S. suis [23,24]. The MLST schema, which is widely used today, was originally developed by King et al. in 2002 [23]. The advantage of this method is the database available on the internet (https://pubmlst.org/organisms/streptococcus-suis, accessed on 2 November 2022) which enables the comparison of isolates worldwide, and newly discovered STs from different areas are constantly being added to the database. Global MLST studies of S. suis identified ST1, ST25, and ST28 as the most prevalent STs in swine [6,25].
Little is known about the S. suis serotypes and sequence types present in the Czech Republic. Previously, information on isolates causing human disorders has been made available [26].
The aim of this study was thus to assess the genetic diversity of S. suis strains isolated from diseased pigs in the Czech Republic. The aim was also to determine the associations of serotypes and sequence types with the virulence potentials of the strains specified by the sites of isolation. Due to the possible occurrence of different genotypes of S. suis on the same farm or even in the same animal, it is necessary to distinguish virulent strains from less virulent or commensal strains. Identification of the major disease-causing strains could enable the development of specific treatments and effective control plans, including the selection of a vaccine strain.

Isolates
Pigs with suspected presence of S. suis were sampled, and S. suis isolates were retrieved from organs and body cavity swabs of dead pigs or from nasal swabs of diseased animals. Symptoms indicating possible S. suis infection were evaluated by field veterinarians on the farms and were as follows: sudden death, fever, depression, loss of balance, lameness, paralysis, paddling, shaking, and convulsion. Samples from healthy animals were not collected in this study. Some isolates were taken from different organs of the same animal. In total, 528 isolates (Supplementary Table S1) were obtained from 144 farms located in the Czech Republic during the years 2018-2022. The data regarding farm identifications were anonymised. In the year 2018, we obtained 104 isolates; we obtained 96 isolates in 2019, 100 isolates in 2020, 146 isolates in 2021, and 82 isolates in the first half of the year 2022.
Blood Agar Base possessing enhanced nutritional properties suitable for the cultivation of Streptococci (Blood Agar Base No. 2 CM0271, OXOID, Basingstoke, UK) with 5% defibrinated ram blood (LabMediaServis s.r.o., Jaroměř, the Czech Republic) was used for the primoculture (22-24-48 hours, at 37 • C ± 10 • C) of sectional material or swabs. Suspected S. suis colonies were further cultivated in microaerophilic conditions (CampyGen, OXOID, Basingstoke, UK) on blood agar plates for 22-24 h at 37 • C. The pure bacterial culture was tested with a standard biochemical examination using a commercial STREP-TOTEST 24 (ERBA LACHEMA, Brno, the Czech Republic), rapid slide latex agglutination (DiaMondiaL Strep Kit; DiaMondiaL, Vienna, Austria), and, finally, the MALDI-TOF method (Matrix-Assisted Laser Desorption/Ionisation-Time of Flight), using a Bruker Microflex mass spectrometer equipped with Maldi Biotyper Compass 4.1.100.10 software, Database RUO 12.0 (updated version) (Bruker, Billerica, MA, USA). The threshold for species identification with a high degree of certainty was log 2.00-3.00, but log 1.70-1.99 was used for species identification with a low degree of certainty. The achieved scores for our investigated isolates were in the range of 1.70-2.6.
Antisera prepared via the immunisation of rabbits with reference strains were used for coagglutination tests [29] for all isolates non-typable by PCR. No positive coagglutination was detected within these isolates.

MLST
Allelic determination of seven housekeeping gene loci (aroA, cpn, dpr, recA, thrA, gki, and mutS) was performed using multilocus sequence typing (MLST), according to a previously published method [23]. Automatic DNA extraction was performed with the MagNA Pure 96 DNA and Viral NA Small Volume Kit on the MagNA Pure 96 Instrument (Roche Diagnostics International AG, Rotkreuz, Switzerland). PCR products were sequenced by the Sanger sequencing method (Eurofins Genomics, Cologne, Germany) using primers with modifications at the 5 end to obtain better sequencing results (SEQplus sequencing primers, Generi Biotech, Hradec Králové, Czech Republic).

Association Testing
Associations between pathotypes and serotypes or sequence types, or combinations of serotypes and sequence types, identified among isolates were tested using Pearson's chi-squared test. Bonferroni corrections for the α level were calculated for each table of combinations (α = 0.05/number of combinations). Due to the very low α calculated, combinations with α < 0.05 were also considered significant.

Pathotypes
The term pathotype was used for a group of isolates that were isolated from a particular organ or tract. The site of isolation may correspond with pathogenicity. The pathotypes were divided into eight categories: Isolates from the lower respiratory tract were far more prevalent, representing nearly half of all isolates, followed by isolates from the circulatory and lymphatic systems (16%) and the central nervous system and upper respiratory tract (both 13%) (Supplementary Table S1).
Out of 528 isolates, 27 tested RecN-negative (Supplementary Table S1). Since all of these isolates were isolated from clinically ill pigs, we decided to include them in the study. Among the 27 RecN-negative isolates, 13 were identified as known serotypes, according to Kerdsin et al. 2014. The identified serotypes were 5, 7, 8, 11, and 12. The remaining 14 isolates were of unknown serotypes.

Relationships between Serotypes and Sequence Types
Within 528 isolates, 100 different combinations of serotypes and sequence types were found (Supplementary Table S1). The most prevalent combination of serotype and sequence type was S7 and ST29: out of 79 S7 isolates, 72 were ST29. Similarly, among 44 S1/2 isolates, 30 of them were ST28. Nearly all S1 isolates were ST1 (19 out of 22). S2 isolates were ST28 (22 isolates) and ST1 (17 isolates), with six newly identified as ST2074 and three as ST1805. Among the 29 isolates of S8, 16 were ST87; the other S8 isolates were of different STs. Not surprisingly, the least well-known STs were noted among isolates without serotype determinations.

New Alleles within MLST Genes
Among previously unidentified alleles in MLST genes, 41 were identified in at least two different isolates. These alleles were submitted to the S. suis MLST database (https://pubmlst.org/ssuis, accessed on 11 November 2022) and received assignments, as shown in Supplementary Table S2.

New Sequence Types
New sequence types were identified by new combinations of previously known MLST gene alleles or they included newly identified alleles. A new ST was assigned when it was found in at least two different isolates. In total, 31 new STs were identified and deposited in the PubMLST database (Supplementary Table S3).

Associations between Pathotypes, Serotypes, and Sequence Types
The possible relationships between serotypes, sequence types, and pathotypes (represented by the site of isolation) were analysed using multiple correspondence analysis (MCA) (Figure 2), which showed the first and second dimensions that represented less than 5% of the data. Pathotypes that could be considered less pathogenic (upper respiratory tract, urogenital system, gastrointestinal tract, and skin) were slightly shifted from other, more pathogenic pathotypes, whose confidence ellipses overlapped and formed a cluster of very similar features (central nervous system, circulatory and lymphatic systems, lower respiratory tract, and joints).

Figure 2.
Multiple correspondence analysis. Multiple correspondence analysis of three variables-pathotype (site of isolation), serotype and sequence type. Confidence ellipses represent 95% of isolates in each pathotype. Grey triangles represent the three variables (pathotype, serotype, sequence type); dots represent each isolate and are coloured according to pathotype. The names of isolates and variables are not shown due to a high number of entities.

Associations between Serotypes and Pathotypes
Associations between serotypes and pathotypes are depicted in Figure 3. For 24 different serotypes and 8 different categories of pathotypes, the α level calculated was α = 0.05/192 = 0.00026. Significant associations with α < 0.00026 were found for S7 with the central nervous system and for S21 and S29 with the upper respiratory tract. Associations at α < 0.00026 were also found among the lower respiratory tract and urogenital system with isolates of non-detectable sequence types. Associations at α < 0.05 were found for S8 with the lower respiratory tract, S11 with the upper respiratory tract, S1/2 and S9 with the circulatory and lymphatic systems, and S7 with joints.

Associations between Sequence Types and Pathotypes
As with serotypes and pathotypes, sequence types were also tested for associations with pathotypes ( Figure 4). For 57 different STs and 8 different pathotypes identified, the α level calculated was α = 0.05/456 = 0.000109. No associations of STs with pathotypes were significant at this level; however, for the association of ST29 with the central nervous system, the α calculated was 0.00015. Other associations at α < 0.05 were as follows: ST1 was associated with the circulatory and lymphatic systems and with the lower respiratory tract; ST15, ST16, ST28, and ST87 were also associated with the lower respiratory tract; ST29 was associated with joints; and isolates with no detectable ST were associated with the upper respiratory tract.

Associations between Particular Serotype and Sequence Type Combinations
When testing the associations of 100 different combinations of serotypes and sequence types identified in our study with 8 pathotypes, the α level calculated was α = 0.05/800 = 0.0000625. The only association at this α level was found for S7ST29 and the central nervous system. The associations significant at α < 0.05 were for the combinations of S2 and ST28 and S8 and ST87, both with the lower respiratory tract (Supplementary Table S1).

Discussion
The data presented in our study showed high diversity among S. suis isolates collected in the Czech Republic from diseased pigs during the years 2018-2022. We identified 23 different serotypes, 56 sequence types, and 100 combinations of serotypes and sequence types. Moreover, a significant proportion of isolates were non-typable with respect to serotypes or sequence types or both.
Isolates from the lower respiratory tract were the most frequent among isolates from organs or systems. These isolates, together with isolates from the upper respiratory tract, gastrointestinal tract, urogenital system, and skin, may not have been the direct cause of death [6]. On the other hand, in multiple correspondence analysis, lower respiratory tract isolates clustered together with isolates from the central nervous system and the circulatory and lymphatic systems, as discussed below. The clinical relevance of this main group of isolates is thus uncertain. Isolates from the central nervous system and the circulatory and lymphatic systems we considered to be highly pathogenic.
Isolates of serotype 7 represented more than 10% of all isolates. Isolates of other serotypes (S1/2, S1, S2, S3, S4, S8, and S9) known to be present in pathogenic strains [6,32] were less frequent but represented at least 3% of all isolates. Ten different serotypes represented up to nine isolates (below 1.8%), but more than 18% of isolates were nontypable for known serotypes using multiplex PCR [27]. This finding demonstrates the high diversity of S. suis in the Czech Republic, which is higher than the diversities currently reported in Switzerland [33], North America [6], and China [34], but similar to that reported in Germany [32]. Interestingly, serotype 6, recognised together with serotype 9 as a main isolate serotype (14.8% and 15.9%, respectively) in Switzerland [33], was not identified in any of our isolates.
Similar to previous works [6,33,35], serotype 7 was associated with pathogenicity. In our case, S7 was associated significantly (after Bonferroni correction) with presence in the central nervous system. Serotype 7 thus could be, at least in Europe and North America, considered to be pathogenic, such that attention should be paid to it. Strong associations were found, also, for non-typable isolates with the urogenital system and the lower respiratory tract. This finding is surprising, because the group of non-typable strains should be considered genetically diverse, as 18 different sequence types are documented within this group. Associations significant at α lower than 0.05 but higher than α after Bonferroni correction were also found for isolates of S8 with the lower respiratory tract, S1/2 and S9 with the circulatory and lymphatic systems, and S7 with joints.
Isolates of serotypes 21 and 29 were both associated with the upper respiratory tract and thus should be considered as having low pathogenicity or as being non-pathogenic, despite the fact they were isolated from diseased animals. Serotype 29 was considered a commensal and serotype 21 a commensal or opportunistic pathogen also by Estrada [6]. Moreover, isolates of serotype 11, not detected in Switzerland [33] or North America [6], but detected in China [34] and Germany [32], were associated at α < 0.05 with the upper respiratory system and thus should be considered as having low pathogenicity or as commensal.
In our study, we included 27 isolates that tested RecN-negative and which thus could be considered not S. suis, according to typing developed by Ishida [28]. The reason for including them in the study was that they were isolated from clinically ill pigs and could thus be considered pathogenic to pigs. Moreover, 13 RecN-negative isolates were typed for known S. suis serotypes. This may open discussion about the specificity of RecN for determining S. suis species identity.
Associations of pathotypes with serotypes were more frequent than associations with sequence types or serotype-sequence type combinations. Moreover, the significance of associations of pathotypes with serotypes was stronger, as five different serotype-pathotype associations were lower than the α level after Bonferroni correction. However, this fact was influenced by the very low α level calculated by Bonferroni correction with the increasing number of tests in the case of sequence types and serotype-sequence type combinations.
Similar to serotypes, only sequence types 29 and 28 were found in more than 10% of isolates. These two sequence types, belonging to the same clonal complex [33], represented nearly one-quarter of all isolates. Moreover, ST29 was significantly associated with the central nervous system, as the only ST within our study, and ST29 isolates currently present in the Czech Republic should thus be considered highly pathogenic, similar to ST29 present in Germany [32]. This finding is in agreement with previous work, where ST29 was associated with pathogenicity also in North America [6]. In contrast to our own and German [32] collections, ST29 was found to be a rare sequence type in North America [6].
The ST1 isolates were the only ones associated with the circulatory and lymphatic systems, and ST1 was thus confirmed to possess high pathogenic potential. These findings are in concordance with previous observations [6,33].
Multiple correspondence analysis based on pathotypes, serotypes, and sequence types showed clustering into four main clusters. One cluster was composed of isolates from the upper respiratory tract, the second was composed of isolates from the urogenital system, and the third was composed of isolates from skin and the gastrointestinal system, while the rest of the isolates were from the central nervous system, the circulatory and lymphatic systems, and the lower respiratory tract. Although the authors of a previous work [6] recognised only three categories of pathotypes (commensal, possibly opportunistic, and pathogenic), their clustering was very similar to our results.
Relatively high numbers of serotypes or sequence types were present in only a few isolates. This was particularly apparent for newly described sequence types: we identified 41 new alleles in genes within the MLST schema and 31 new sequence types. A possible explanation of this fact is that a substantial proportion of these isolates were isolated only by chance and were not primary pathogens causing death or major signs of disease. On the other hand, all samples were collected from animals suffering from disease with symptoms common to the S. suis course of infection, as reported by field veterinarians. S7ST29 was significantly associated with the central nervous system, similar to S7ST29 isolates in Germany [35]. Due to the similar pathogenicity of our S7ST29 and German S7ST29 isolates, we can speculate that the S7ST29 isolates present in the Czech Republic and Germany may represent the same strain. This possibility is supported, also, by extensive pig transfer between the Czech Republic and Germany [36].
The other sources of diversity are the group of non-determined serotypes and the group of non-determined sequence types. With respect to the non-determined serotype group, it was significantly associated with the lower respiratory tract and also with the urogenital system; we will pay attention to this group in further studies. In addition, while non-determined serotypes have been found by other authors in their works, the size of this group varies among studies. In a recent study in North America from the years 2014-2017, the percentage of isolates of unknown serotypes was 11% [6]; in Switzerland, in the last decade, it was 9% [33]; but in Germany, in the years 2015-2016, it was only 4.2% [32]. On the other hand, in China, in the years 2015-2017, it was 14% [34]. In our work, it was nearly 19%, although 2.6% of non-typable isolates may be other Streptococcus species. This high portion of non-typable isolates may be due to the predominance of autochthonous serotypes in Eastern Europe before borders were opened in the late nineties of the last century. There are no studies on serotypes in the Czech Republic in those years, and, to the best of our knowledge, our work is the first ever study on S. suis diversity in pigs in the Czech Republic.
Although the information presented in our work is based on data collected in the Czech Republic, we believe that the findings may be of interest to Western European countries due to the high mobility of pigs in the European Union and their being the source of a significant amount of breeding material and also of weaned piglets on Czech farms [36].