Streptococcus agalactiae in Brazil: serotype distribution, virulence determinants and antimicrobial susceptibility

Streptococcus agalactiae (Group B Streptococcus, GBS) is a leading cause of neonatal morbidity and mortality [1], and is also an important opportunistic agent of infections in pregnant women, as well as non-pregnant adults, especially the elderly or those with underlying medical conditions [2]. This microorganism is commonly found as a colonizer of the genital and the gastrointestinal tracts of both men and women, and vertical transmission from a colonized mother to her newborn during labor can result in life threatening infections.

GBS virulence is complex and multifactorial. Several virulence determinants are involved in the adhesion to and invasion of host cells, as well as in the immune system evasion. Surface components, including a polysaccharide capsule and proteins, such as Cα, Cβ, Rib and the laminin binding protein (LMB), and a number of enzymes (like the C5a peptidase) and toxins/cytolysins, are produced and have been associated with GBS virulence [3].

Capsular polysaccharides are recognized as playing a key role as virulence factors and are important targets for the development of vaccine strategies. GBS capsular polysaccharides have chemical and antigenic differences that enable the subdivision of this species into ten serotypes, denominated Ia, Ib, II-IX [4]. Vaccines against GBS infections must include the more common serotypes associated with disease in different populations and capsular polysaccharide-protein conjugate vaccines are in clinical trials [5]. The epidemiological distribution of these serotypes can vary according to several aspects, including the geographical region, the profile of the population being studied, and the source of the bacterial isolate [6]. Four serotypes (Ia, II, III and V) are usually the most frequently isolated in the United States and in some European countries [6–8], while serotypes VI to IX are rarely described [4, 9]. In Brazil, the occurrence of serotypes Ia, Ib, II, III, IV and V has been described in a few studies conducted with isolates originated from the South and the Southeast regions [10–15].

S. agalactiae is still uniformly sensitive to penicillin, although isolates with reduced susceptibility to penicillin have been reported since 2008 [16–18]. Clindamycin or erythromycin is recommended for GBS intrapartum prophylaxis for penicillin-allergic women with high risk of anaphylaxis or when therapeutic failure is suspected [19]. However, increasing rates of resistance to these antibiotics have been detected in several regions of the world, including Europe [8, 20], Asia [21], North America [22–24], and South America [11, 13, 15, 25].

Despite the clinical and epidemiological impact of GBS infections and the trends to increasing occurrence of antimicrobial resistance, information about isolates from Brazil is still fragmentary and usually restricted to a few areas of this large country. There is no Brazilian study that provides a national overview of the phenotypic and genotypic characteristics of circulating isolates of GBS. In the present report we describe the serotype distribution, antimicrobial susceptibility, presence of virulence-related and antimicrobial resistance genes among GBS isolates collected from the five geographical regions in Brazil. Also, the genetic diversity of macrolide-resistant isolates was assessed and compared to a representative fraction of the macrolide-susceptible isolates, providing a national overview of the characteristics of strains circulating in the country.

Overall, serotypes Ia (120 isolates; 27.6% ± 4.2, 95% CI), II (83 isolates; 19.1% ± 3.7), Ib (81 isolates; 18.7% ± 3.7) and V (59 isolates; 13.6% ± 3.2) were the most predominant among the 434 GBS isolates included in the present study, followed by serotypes IV (35 isolates; 8.1% ± 2.6) and III (29 isolates; 6.7% ± 2.3). The sum of the four most common serotypes (Ia, Ib, II and V) constituted 79% (±3.8) of the isolates while percentages above 90% were obtained with the addition of serotypes III and IV. Serotypes VI, VII, and VIII were not found and 6.2% ± 2.3 (27 isolates) of the isolates were nontypeable.

Serotypes Ia, II and Ib were the most frequent among isolates recovered from infections, which were represented by the South, Southeast and Mid-West regions (Table 1). Although absolute numbers of colonization isolates were small in some regions, particularly the Mid-West, serotypes Ia and II were the most frequently observed in the South, Southeast and Mid-West regions, whereas serotypes Ib and V were the most common in the North/Northeast region (Figure 1).

Two tetracycline resistance genes, tetM and tetO, were detected in 99.3% ± 0.8 and 1.8% ± 1.2 of the total GBS isolates, respectively, and five of them harbored both genes. The tetK and tetL genes were not detected.

One Ery^R strain could not be typed by PFGE, despite several attempts, due to resistance to SmaI digestion. Among the remaining 17 isolates, 5 CCs and 16 different PFGE profiles were identified, and the Simpson’s Index of Diversity (SID) for this subgroup of isolates was 0.897 (95% CI, 0.807-0.987). Among the 25 erythromycin-susceptible strains, 9 CCs and 25 different profiles were observed, with a SID of 0.960 (95% CI, 0.940-0.980). Overall, 9 CCs and 40 different PFGE profiles were identified among the 42 isolates (Figure 2), with a SID of 0.935 (95% CI, 0.891-0.979). Five of the CCs comprised both Ery^R and susceptible isolates. Among the resistant isolates, those harboring different resistance genes shared the same CC. Also, isolates belonging to different serotypes or recovered from different clinical sources were clustered in the same CC.

S. agalactiae is recognized as a frequent colonizing agent in pregnant women and is an important cause of neonatal sepsis and meningitis. Nevertheless, in the past decade GBS has been increasingly associated with invasive disease in non-pregnant adults [1, 2].

The aim of this study was to phenotypically and genotypically characterize 434 GBS isolates obtained from different regions of Brazil, a continental country, with a large ethnic, climatic, social and economic diversity. Most of the previous studies performed in Brazil were limited to isolates from the South and Southeast regions [10–15]. This is the first study in Brazil encompassing the characterization of GBS isolates from 15 different locations representative of the five Brazilian geographical regions. Therefore, although sampling each geographic region and checking for variations among them was not a primary objective of the present study, it seemed useful and informative to display our results according to the geographical region. Nevertheless, our data might not fully and accurately represent each one of the five geographical regions.

Serotype determination has been traditionally applied in epidemiological studies on GBS and is an important aid in the development of broadly protective vaccines containing capsular polysaccharides or polysaccharides conjugated to protein [5]. The serotype distribution varies geographically and studies performed in European countries, United States and in Latin America have shown that serotypes Ia, Ib, III or V are usually the most frequently found [13, 34]. In this study, serotypes Ia, Ib, II and V were the most commonly detected, in frequencies varying according to the region. Serotypes VI to VIII were not found, and these serotypes are still rarely reported from other locations [9]. Our results are in agreement with previous data obtained in earlier Brazilian studies with isolates obtained from the South and Southeast regions [10, 15, 35], as well as with data published in several other countries [24, 25].

Serotypes were found in similar frequencies among isolates obtained from colonization as well as infections (p = 0.1392), suggesting that the capsule, despite being an essential factor for the virulence in GBS is not the only bacterial determinant responsible for the pathogenicity. Other studies have described the predominance of serotypes III and V among cases of infection in neonates and in non-pregnant adults, respectively [25, 34] while serotype Ia was found to predominate among GBS causing invasive infections in non-pregnant adults in Portugal [36]. In Brazil, previous studies showed that serotypes Ia, II, III and V were found in 68.2% of isolates predominantly from vaginal specimens of asymptomatic pregnant woman [14]. On the other hand, serotype IV was detected in 13.1% of the isolates from Curitiba city, Paraná state, Brazil, being most of them from infectious cases [13], and serotypes Ib (34.9%) and Ia (25.6%) were predominant among a cohort of HIV-infected pregnant women in Rio de Janeiro city [35].

The occurrence of a nontypeable isolates rate of 6.2% in this study is consistent with the previously reported data [13]. Nevertheless, it should be pointed out that serotype IX was not investigated in the present study due to the lack of specific antiserum and that the detection of a serotype IX GBS isolate has been recently reported in the South region of Brazil [37]. Therefore, it is possible that the group of NT isolates may include representatives of this serotype.

The uniform susceptibility of GBS to beta-lactam antibiotics detected in this study is consistent with previously published data on S. agalactiae isolates in Brazil [10–15, 30], as well as in several regions in the world [8, 25]. However, an extensive network national surveillance should be implemented since isolates with reduced susceptibility to penicillin have been described in various regions of the world [16–18]. Although susceptibility of GBS to beta-lactam antibiotics was observed in the present study, reduced susceptibility has been reported elsewhere, underscoring the importance of an ongoing surveillance to monitor emergence of increased penicillin MICs among GBS isolates.

The low rates of resistance to erythromycin (4.1%) and clindamycin (3.0%) found in this study, are similar to that obtained in other studies conducted in Brazil, and some other countries in Latin America [11, 13, 15, 25]. On the other hand, higher resistance rates have been observed in studies conducted in Asia, Europe, United States and Canada [6, 22, 23]. No correlation was found among erythromycin resistance and serotype, as previously observed [13]. Nonetheless, the predominance of serotype V among isolates with MLS_B resistance has been reported [8]. Erythromycin-resistant isolates in the present study were obtained from colonization or non-invasive infections, as previously described [13, 15]. These data suggest that the use of erythromycin and clindamycin as second-choice drugs in individuals with penicillin allergy should be accompanied by routine susceptibility testing to ensure proper therapy. If this is not possible, then treatment with vancomycin is recommended by CDC [19].

Among the Ery^R isolates, the ermA gene was the most frequent. Isolates presenting the cMLS_B phenotype were mainly associated with the ermB gene, while those with iMLS_B or M phenotypes harbored the ermA gene. Other authors have shown similar results [38, 39]. The mefA gene was not found in this study, as previously described in Brazil [11, 13], in contrast with the results obtained in other studies, where the mefA gene was detected in the isolates expressing the M phenotype [6, 38]. The causes for the absence of resistance genes in six isolates with MLS_B phenotype might reside in the fact that a great variety of genetic mechanisms can be associated with a single resistance phenotype. Here, we investigated the presence of the genes most commonly reported among GBS strains, however those 6 isolates could harbor less frequent determinants, such as ermC, ermF, ermQ, ermT, msrA or msrD; all of them already described among streptococci [40]. Also, mutations in 23S rRNA or in the genes encoding ribossomal proteins can also be speculated as possible causes of erythromycin-resistance in such strains.

Erythromycin-resistant isolates showed a slightly higher homogeneity by PFGE when compared to the susceptible strains analyzed. Nevertheless, a considerable genetic diversity was observed among the 42 isolates typed by PFGE, which clustered Ery^R and susceptible isolates in the same CC, suggesting that resistance to erythromycin is not necessarily a clonal characteristic among GBS strains included in the present study. On the other hand, the majority (64.7%) of the 17 Eri^R isolates were distributed among five of the nine CCs detected, suggesting that multiple clusters may have arisen simultaneously, mostly by independent acquisition of resistance genes by different strains. Similar results were previously shown in other studies [13, 15, 41]. Likewise, serotype does not seem to have a correlation with the classification of a GBS isolate in a clonal group. Relationships between the serotype and the source of isolation or between the serotypes and PFGE profiles were also not detected in another Brazilian study [14]. This fact could be explained by the occurrence of capsule switching between GBS isolates by horizontal gene transfer, as a result of pressures by immune response of the host [7]. One Ery^R isolate could not be typed by PFGE due to resistance to SmaI restriction. This phenomenon is not rare, and has been reported by other authors from different parts of the world [8, 42, 43]. Nevertheless, PFGE still represents an important molecular typing tool for GBS strains in Brazil, where, to our knowledge, other molecular methods such as MLST and MLVA have not or have rarely [37] been employed.

Resistance to tetracycline was detected in most of the isolates of this study, and the tetM gene was widely disseminated as previously reported in Brazil as well as in other countries [6, 8, 11, 15, 25]. All the isolates resistant to macrolides independent from the phenotypic and genotypic mechanism of resistance also had the tetM gene.

The present study provides an overview of the distribution of serotypes and on antimicrobial susceptibility of GBS isolates from Brazil, the largest country in Latin America. The findings suggest that GBS isolates circulating in Brazil have a wide phenotypic and genotypic diversity and that macrolide-resistant isolates may arise by both clonal spread and independent acquisition of resistance genes. These data are important to help in designing prevention and treatment strategies for GBS infections in the region.