Genetic characterization and diversity of Streptococcus agalactiae isolates with macrolide resistance

INTRODUCTION

Streptococcus agalactiae [group B streptococcus (GBS)], despite the fact that antibiotic prophylaxis is implemented during the perinatal period in the USA and some European countries, is still one of the major aetiological causes of sepsis and newborn meningitis. Moreover, in recent years, GBS has been isolated more often from immunosuppressed patients (Gherardi et al., 2007; Schrag et al., 2002).

Penicillin is the recommended treatment drug, to which no case of GBS resistance has been reported. However, for people allergic to penicillin, the alternative drugs are macrolides (e.g. erythromycin) or lincosamides (e.g. clindamycin) (Heczko et al., 2008; Schrag et al., 2002).

Two main resistance mechanisms have been described for S. agalactiae. erm (erythromycin ribosome methylase) genes such as ermA and ermB (subclass ermTR), encode methylase 23S rRNA, which is responsible for methylation of erythromycin and clindamycin receptor sites in ribosomes. Expression of these genes is described as the MLS_B phenotype and points to cross-resistance to macrolides, lincosamides and streptogramin B. This resistance can be constitutive macrolide–lincosamide–streptogramin B (cMLS_B) resistance, as well as inductive macrolide–lincosamide–streptogramin B (iMLS_B) resistance. Genes mefA and mefE [macrolide resistance (M) phenotype] are responsible for the second mechanism, in which highly conservative sequences encode a pump that expels antibiotics from inside of the bacterial cell. Expression of these genes is described as the M phenotype (Gherardi et al., 2007).

Epidemiological analyses of GBS isolates are mainly based on capsule serotyping. Currently, ten different GBS serotypes have been described: Ia, Ib and II–IX. Serotype distribution varies with geographical region, ethnic origin and the virulence of clinical isolates (Persson et al., 2004; Poyart et al., 2007; Slotved et al., 2007).

GBS can also be classified on the basis of surface protein antigens. The major surface-localized protein antigens of GBS belong to a family of surface proteins and are the alpha-C protein, rib, alpha-like protein 2, alpha-like protein 3, alpha-like protein 4 and epsilon protein (alpha-protein-like proteins) that are encoded by bca, rib, alp2, alp3, alp4 and epsilon genes, respectively. The possibility of looking at the protein gene profile increases the potential of GBS subtyping (Creti et al., 2004; Persson et al., 2008).

Genotyping methods, including the random amplification of polymorphic DNA (RAPD) method, as well as PFGE, are used to characterize and distinguish specific clones among GBS isolates (Martinez et al., 2000; Skjaervold et al., 2004).

Because in recent years the resistance of S. agalactiae to macrolides, lincosamides and streptogramin B is being reported more often (Acikgoz et al., 2004; Chohan et al., 2006; De Mouy et al., 2001; Hsueh et al., 2001; Schoening et al., 2005), from an epidemiological point of view it is important to conduct detailed characteristics of these isolates. These studies were performed in few specialized scientific centres around the world (Diekema et al., 2003; Domelier et al., 2008; Gherardi et al., 2007; Uh et al., 2005), yet there are no Polish data whatsoever. Therefore, the aims of our studies were: (i) to determine the frequency of the occurrence of resistance phenotypes to macrolides in S. agalactiae isolates originating from pregnant women; (ii) to show correlation between the occurrence of cMLS_B, iMLS_B or M phenotype and the presence of genes encoding resistance to erythromycin, as well as genes coding capsular polysaccharides (CPS) and surface proteins; (iii) to conduct genetic similarity tests of GBS isolates using PFGE and compare the results to RAPD.

METHODS

Research on the macrolide resistance of S. agalactiae isolates was carried out from August 2007 until September 2008. A total of 601 women in the third trimester of pregnancy from Southern Poland underwent examination of GBS colonization according to Centers for Disease Control and Prevention recommendations (Brzychczy-Włoch et al., 2008b; Heczko et al., 2008; Schrag et al., 2002). The study was approved by the Jagiellonian University Bioethical Committee, decision no. KBET/143/B/2007.

Macrolide resistance in 169 GBS isolates originating from pregnant carriers was investigated. The disc diffusion method, with clindamycin (2 μg) and erythromycin (15 μg) (Oxoid), as well as Etests (AB Biodisk), were used. Phenotypes cMLS_B, iMLS_B and M were detected. The results were interpreted according to Clinical and Laboratory Standards Institute guidelines (CLSI, 2009).

Analysis of the genes for resistance to erythromycin ermA, ermB, ermC and mefA/mefE was carried out using PCR with four pairs of primers (Genomed) according to Sutcliffe et al. (1996), as previously described (Brzychczy-Włoch et al., 2008a). The characteristics of CPS were tested serologically using a group B streptococcus GBS serotyping kit (ESSUM). To verify the method, the detection of genes encoding particular CPS was performed using a multiplex PCR method, according to Poyart et al. (2007).

A multiplex PCR method was used to detect the surface protein genes bca, rib, epsilon, alp 2/3 and alp4 with five pairs of primers (Genomed) according to the procedure of Creti et al. (2004). To distinguish between alp2/3 and confirm presence of Alp3 protein, the reverse primer Alp3 (5′-TTT TGG TTC GTT GCT ATC CTT AAG C-3′) was used according to Gherardi et al. (2007), and the universal forward primer was used according to Creti et al. (2004). The PCR products were analysed in 1.5 % agarose gel in the presence of ethidium bromide (0.25 μg ml⁻¹) (Bio-Rad).

The analysis of the genetic similarity of GBS isolates resistant to macrolides was performed using RAPD and PFGE methods. RAPD was performed with primer OPB17 (Genomed) according to Martinez et al. (2000), as previously described (Brzychczy-Włoch et al., 2008a). PFGE was performed according to Tynkkynen et al. (1999), with the following modifications. Bacteria were cultured in tryptic soy broth (Difco) at 37 °C for 24 h and 25 U restriction enzyme SmaI (Fermentas) was used. PFGE was performed using the CHEF-DR III device (Bio-Rad), with the following parameters: pulses 5–40 s; 6 V cm⁻¹; 14 h; 14 °C. The comparison of the genetic profiles obtained from RAPD and PFGE was carried out with Molecular Analyst (Applied Maths) software.

The final pictures from all electrophoresis were analysed using QuantityOne software (Bio-Rad), as well as a GelDoc2000 device (Bio-Rad).

For statistical analysis the χ² test was used. In the case of small sample sizes Fisher's exact test was used. When the frequency was zero, the G² (likelihood ratio) test was used. P values of <0.05 were considered significant.

RESULTS AND DISCUSSION

The increasing resistance to macrolides among group B streptococci observed in recent years in many countries is a therapeutic problem among patients allergic to β-lactams (Acikgoz et al., 2004; Chohan et al., 2006; De Mouy et al., 2001; Hsueh et al., 2001; Schoening et al., 2005). Because of this fact there is a need to monitor the phenomenon and implement rational antibiotic therapy. From the epidemiological point of view, it is essential to perform specific characterization of GBS isolates, to find, among others, an answer to whether the observed phenomenon occurred due to the spread of a specific GBS clone in the population, or whether it is the result of acquired resistance among S. agalactiae strains.

In 2008 in Poland, recommendations were developed to screen pregnant women for GBS colonization and, if needed, implement perinatal antibiotic prophylaxis, according to the Centers for Disease Control and Prevention recommendations (Schrag et al., 2002). Antibiotic therapy used on a wide scale may cause an increase in the resistance of GBS isolates, as has happened in other countries. A good example of this is the USA, where the percentage of GBS isolates resistant to erythromycin in the 1980s century was 1.2 %, in the 1990s was 18 % and at the beginning of 21st century it reached 40 %; similarly in Asia, where resistance to macrolides has now reached up to 50 % (De Mouy et al., 2001; Hsueh et al., 2001).

In Europe the percentage of GBS resistant to erythromycin, according to available literature, is: 11 % in Germany (Schoening et al., 2005), 14 % in Spain (Acikgoz et al., 2004), 16 % in Italy (Gherardi et al., 2007), 21 % in France (De Mouy et al., 2001) and 22 % in Turkey (Acikgoz et al., 2004).

In our study, the susceptibility to erythromycin and clindamycin was determined for isolates of S. agalactiae originating from pregnant women carriers. The frequency of occurrence of GBS strains resistant to erythromycin was determined to be 4.5 % and to clindamycin to be 3 % of the examined female population (n=601), which turned out to be 16 % (n=27) of strains resistant to erythromycin and 10 % (n=17) of strains resistant to clindamycin among GBS isolates (n=169), respectively. In general the phenotypic mode of resistance to macrolides was shown for 27 isolates, among which the cMLS_B phenotype (n=17, 63 %) was most common, then the iMLS_B phenotype (n=7, 26 %) and the M phenotype (n=3, 11 %) (Fig. 1⇓). Using Etest the values for MIC₅₀ and MIC₉₀ were determined for erythromycin and clindamycin (Table 1⇓).

The data obtained correspond with the results of a French study, where 70 % of isolates had the cMLS_B phenotype, 22 % the iMLS_B phenotype and about 8 % the M phenotype (De Mouy et al., 2001), and partially with outcomes of a German research group, where cMLS_B dominated accounting for 41 % of isolates, iMLS_B for 38 % and M for 21 % (Schoening et al., 2005). Slightly different data come from Asia, where in almost 90 % of the analysed phenotypes, the constitutive type of resistance was predominant (Hsueh et al., 2001). Quite different results were described in Turkey, where the iMLS_B phenotype dominated at 80 %, while cMLS_B was present at only 20 % (Acikgoz et al., 2004).

Among investigated GBS isolates with a phenotypically confirmed mechanism of resistance to macrolides, the detection of ermA, ermB, ermC, mefA/mefE genes encoding resistance to erythromycin was performed. The ermB gene dominated and its presence was shown for all strains with the cMLS_B phenotype, and belonging to V serotype (n=17), while mefA/mefE genes were shown in isolates with the M phenotype, representing Ia serotype (n=3). The presence of ermA and ermC genes was not detected in any of the tested strains (Table 1⇑). For our study of the occurrence of genes encoding resistance to erythromycin, similar results were obtained to these found in the literature (De Mouy et al., 2001; Gherardi et al., 2007; Hsueh et al., 2001; Uh et al., 2005).

The detection of genes encoding particular CPS was performed with multiplex PCR. Among strains resistant to macrolides, the V serotype (n=18, 67 %) dominated, then II (n=4, 15 %), Ia (n=3, 11 %) and III (n=2, 7 %). In this pool of isolates none of the other serotypes were found. Based on data collected for all the studied women, showing the distribution of serotypes among S. agalactiae isolates (n=169), the expected numbers of resistant strains that represented particular serotypes were calculated, then checked if they were far different from obtained values. We showed a statistically significant relationship between resistance to macrolides and the V serotype of S. agalactiae; however, there was not a similar statistically significant relationship with the other serotypes (Fig. 2⇓).

At the turn of the last decade there were more frequent reports of the isolation of S. agalactiae strains belonging to serotype V, which in most cases were isolated from different types of infections, especially from older or immunosuppressed people (Gherardi et al., 2007). Nowadays, we think that the spread of the V serotype in the population is one of the more serious threats related to GBS epidemiology. Uh et al. (2005) showed 60 % domination of V serotype in GBS strains resistant to macrolides and similar results were also reported by European centres (Diekema et al., 2003; Gherardi et al., 2007). Our results are consistent with the literature and point to a dangerous phenomenon in recent years, which is an increasing number of GBS isolates being resistant to erythromycin, representing serotype V.

The surface proteins of GBS are likely to play an important role in the pathogenesis of S. agalactiae infection (Creti et al., 2004; Persson et al., 2008). Creti et al. (2004) noted a relationship between serotypes and surface protein genes, but to a lesser degree than previously reported. They found the association of serotypes Ia, Ib and II with the alpha-C protein, of serotype III with Rib, and serotypes V and VIII with Alp3, but it was not absolute (Creti et al., 2004). Among macrolide-resistant isolates investigated in the present study, the alp3 gene significantly dominated (n=19, 70 %). In this pool of isolates we found rib (n=3, 11 %), epsilon (n=2, 7.5 %), bca (n=2, 7.5 %) and alp2 (n=1, 4 %) genes too, but in lesser degree, and no alp4 were present.

Methods used to perform genetic similarity analysis between GBS isolates are, among others, RAPD as well as PFGE (Gherardi et al., 2007; Martinez et al., 2000). The great advantage of RAPD is a relatively short procedure time and simple methodology. This method was successfully used in analysing genetic isolates of S. agalactiae both in the study by Martinez et al. (2000) and in our previous study (Strus et al., 2009).

In this study a better genetic discrimination of GBS isolates was shown thanks to PFGE, where 20 pulsotypes were obtained for 27 isolates, in comparison to RAPD where 14 genotypes were obtained for 27 isolates (Figs 3⇓ and 4⇓). Special attention should be given to the cluster obtained with RAPD, comprising 13 GBS isolates (strain numbers: 53, 82, 97, 139, 212, 83, 7, 22, 52, 114, 126, 190, 248), which were genetically identical (Fig. 3⇓). The use of PFGE allowed differentiation of these 13 strains into 6 different pulsotypes (Fig. 4⇓). Comparing outcomes of PFGE with RAPD results allows us to conclude that PFGE is a more precise tool for analysing the molecular epidemiology of GBS.

The Gherardi et al. (2007) study conducted with PFGE shows a high genetic diversity among GBS strains that are not macrolide resistant. However, they showed an exact correlation between strains belonging to specific pulsotypes and serotypes, the presence of capsular proteins and genes encoding resistance to macrolides. They also found that in the group of GBS strains resistant to erythromycin, isolated from infections, 60 % of the isolates represented one genetic profile (Gherardi et al., 2007). Similar results were shown by Diekema et al. (2003), where GBS isolates resistant to macrolides and belonging to serotype V, collected from blood infections in newborns, in 35 % of cases were of one subtype. According to researchers, the observed phenomenon points to the fact that within the GBS group resistant to macrolides, a clonal spread has occurred in the population prone to the described types of events (Diekema et al., 2003; Gherardi et al., 2007; Schoening et al., 2005).

Analysing the outcomes of our own studies with PFGE, a high genetic diversity was shown between GBS strains resistant to macrolides. Among the analysed features, the highest genetic homogeneity was in the group of isolates representing serotype V (Table 2⇓), which is in agreement with outcomes in the literature (Diekema et al., 2003; Gherardi et al., 2007). It is important to notice that the examined isolates originated from cases of asymptomatic carriage in pregnant women, which could explain the relatively high genetic diversity between strains.

Based on the obtained outcomes, a hypothesis can be made that the observed phenomenon of increasing resistance among GBS isolates originating from colonization not only is the result of clonal spread in the population but also points to horizontal resistance gene transmission between streptococci.