Molecular epidemiology of macrolide resistant Group A streptococci from Puducherry, India

Introduction: In penicillin allergic patients, macrolides are the most commonly used antibiotics for treating streptococcal infections, irrespective of the higher resistance rates. The objective of this study was to evaluate the comparative prevalence, phenotypes, and genetic determinants of macrolide resistance and associated emm types among different clinical isolates of Streptococcus pyogenes. Methodology: A total of 173 Streptococcus pyogenes isolates were examined for macrolide resistance phenotype by double-disc test, resistance determinants by multiplex PCR and emm genotyping. Results: Erythromycin resistance was found in 51.4% of isolates, with MIC90 ≥ 256 μg/mL Inducible phenotype was commonly found (iMLS, 67.4%) followed by the M phenotype (32.5%). Among these isolates, 65.1% harboured ermB and 32.5% mefA as sole macrolide resistance gene, whereas presence of both, ermB plus mefA was observed in 2.2% cases. The most common types among resistant strains were emm63 (11.2%), emm44 (6.7%), emm42 (5.6%), and emm75.3, emm82, emm85, emm92, emm111.1 (4.4% each). Statistically significant association was observed between emm63, emm44 and erythromycin resistance (p ≤ 0.05). Association of these emm types and macrolide resistance have not been reported earlier. Conclusion: Higher macrolide resistance in this study can be attributed to overuse and misuse of this antibiotic. These findings indicate that macrolides should not be empirically used for treating severe streptococcal infections.


Introduction
Streptococcus pyogenes (Group A streptococcus, GAS) is responsible for a wide array of human illnesses ranging from self-limiting pharyngitis and impetigo to flesh eating necrotizing fasciitis and other life threatening invasive infections [1]. If untreated, mild infections can lead to immune mediated post-infectious sequelae such as glomerulonephritis and acute rheumatic fever, which accounts for majority of streptococcal disease burden in the developing countries [2,3].
GAS remains universally susceptible to penicillin and this is the drug of choice for streptococcal infections. Among patients with penicillin allergy, macrolides, lincosamides and fluoroquinolones are recommended as the treatment options. Macrolides are ideal for treating streptococcal pharyngitis and other respiratory infections whereas clindamycin is preferred for treating patients with serious soft tissue infections. Although penicillin resistance has not been reported till date, resistance to other antibiotic groups has been reported worldwide [4,5].
Globally the rate of erythromycin resistance in streptococci is 20% -40%, that is also in accordance with reports from India [4,6]. Macrolide resistance in streptococci is mediated by two major mechanisms, which include target site modification, and macrolide efflux pumps. Target site modification is affected by rRNA methylases encoded by the ermB gene or ermTR gene of the ermA class and is related to the MLS phenotype (resistance to macrolide, lincosamide and streptogramin B). The expression of these genes may be constitutive (cMLS) or inducible (iMLS). The second mechanism includes efflux pump encoded by the mefA gene and is related to the M phenotype (resistance to 14 and 15 membered ring macrolides). Other minor resistance mechanisms include mutations in the 23S rRNA gene and alterations in riboproteins L4 and L22 [7,8]. It was reported that macrolide resistance in GAS could be associated with certain emm types like emm4, emm12 and emm28 [9].
The present study was undertaken to evaluate the comparative prevalence, molecular epidemiology, phenotypes, and genetic determinants of macrolide resistant Streptococcus pyogenes.

Methodology
This observational study was conducted in the Department of Microbiology, Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER) in Puducherry, South India between November 2013 and December 2015. The study was approved by the JIPMER Scientific Advisory Committee (JSAC) and Institute Ethics Committee (IECNo. JIP/IEC/SC/2014/3/555).

Bacterial strains
All the consecutive, non-repetitive isolates of GAS from various clinical cases (tonsillitis, necrotizing fasciitis, pyoderma, folliculitis, wound infection, abscess and sepsis) were included in the study. Identification was done by routine laboratory techniques involving bacitracin sensitivity, PYR test, latex agglutination (STREP Test kit, Plasmatec, Dorset, UK) and confirmed by spy1258 PCR [10].

Antibiotic susceptibility and MIC determination
Isolates were tested for susceptibility to erythromycin and clindamycin by disc diffusion, minimum inhibitory concentration (MIC) was determined by use of the standard protocol [11], Erythromycin resistant isolates were checked by E-test (Biomerieux, La Balme Les Grottes, France).

Detection of macrolide resistance phenotypes
Macrolide resistance phenotype was detected based on their susceptibility by double disc tests involving erythromycin (15μg) and clindamycin (2μg) discs, as previously described [12].

DNA extraction
Genomic DNA was extracted using mericon DNA Bacteria plus kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.

Detection of macrolide resistance genes
All erythromycin resistant isolates were screened by a multiplex PCR for the presence of genes ermB, ermA and mefA as described previously [13], using 25µl reagent that contains 2×Taq PCR Smart Mix (Origin, Karunagappalli, Kerala, India), 10 pmol of each primers and 10 ng of sample DNA. The PCR cycling conditions were as follows: initial heating at 93°C for 3 minutes, followed by 30 cycles of denaturation at 93°C for 1 minute, annealing at 61.8°C for 1 minute, extension at 65°C for 1 minute and final extension for 3 minutes after the last cycle. The 16SrRNA gene specific for genus Streptococci was included as an internal control (Table 1).

Emm typing
Emm sequencing was performed according to the protocol of the Center of Disease Control (CDC) International Streptococcal Reference Laboratory with slight modifications (http://www.cdc.gov/ncidod/biotech/ strep/protocols.htlm).

Statistical analysis
Analysis were carried out using OpenEpi software (Version3.03a). The prevalence of macrolide resistance was expressed in percentages. The Chi-Square test was used to calculate the association of emm types with macrolide resistance. p value of ≤ 0.05 was considered as statistically significant.

Results
A total of 173 Streptococcus pyogenes were recovered from various clinical samples like pus, wound swab, throat swab, blood, tissue during the study period. Out of 173, 130 (75.14%) were isolated from patients with non-invasive superficial and respiratory  Table 2). M phenotype GAS isolates with (n = 29) macrolide resistance carried only mefA as sole macrolide resistance determinant. Fifty-eight out of sixty with iMLS phenotype carried only the ermB gene as the only resistant determinant whereas one throat and one skin isolate harbored mefA gene along with ermB gene. In the present work, the ermB gene was the most prevalent macrolide resistance determinant followed by mefA gene (Figure 1).

Discussion
GAS erythromycin resistance was first reported in 1955 and since then the resistance has been progressively increasing worldwide. The rates vary between 5% and 40%, with the highest prevalence in Asia and lowest prevalence in Europe and USA. This is a major concern, as macrolides are the main treatment option for streptococcal infections in penicillin allergic patients for treating streptococcal infections. Higher resistance could be attributed to the over prescription and misuse of these antibiotics.
Recent reports show that there is a fluctuating trend in the macrolide resistance pattern across the globe with unexpected reductions in the resistance rates especially from Spain, Portugal and China, where the resistance used to be high [14,15]. Lower macrolide resistance (10-15%) continued to be reported from countries like Finland, Greece, Italy, Germany, Mexico, USA and Canada [8].
Previous studies from India reported that there is a steady increase of GAS erythromycin resistance ranging from 2% to 38.13%. In the present study we report an overall erythromycin resistance of 51.4% which is very high comparing to reports from other parts of India [16,17]. In this study we observed that the erythromycin resistance was 48.8% among invasive  strains (n = 21/43) and 52.3% among non-invasive strains (n = 68/130). Earlier studies from India have reported the prevalence of M type [18], but given our data the iMLS phenotype was predominant (67.4%), followed by M phenotype (32.5%) which is in agreement with a recent study from Chennai [16]. Isolates with the iMLS phenotype showed high-level resistance to erythromycin (MIC90 ≥ 256 µg/mL) and inducible resistance to clindamycin, whereas those with M phenotypes showed lower erythromycin resistance values and good susceptibility to clindamycin. Another notable feature is that M phenotype was predominant in countries with lower resistance. Clindamycin is recommended in treating severe skin and soft tissue infections due to the higher tissue permeability, inhibition of toxin production, and promotion of phagocytic activity. The higher rate of inducible clindamycin resistance (67.4%) in our study is a serious issue since it can result in unexpected treatment failures. The cMLS phenotype is commonly reported from European countries [19]. The complete absence of cMLS phenotype among our study population is in contrast with previous reports, which document the presence of cMLS as one of the minor phenotype in India [18]. The genetic determinants of macrolide resistance in Indian GAS population have not been extensively studied. The only Indian study discussing the genetic basis of macrolide resistance showed the predominance of ermB gene [16]. Our findings reveal a predominance of ermB gene (65.1%) followed by mefA gene (32.5%). A similar result was also previously reported from Belgium, France and Italy [20,21]. Results of the present study also show that the mefA gene is always associated with the M phenotype and is consistent with other reports [22] from the literature. In addition, our data demonstrated that the iMLS phenotype was associated with ermB gene which is in contrast to other studies [7] where iMLS was reported to be associated with ermA gene.
Among erythromycin resistant isolates in the present study, the most common emm types were emm63 (11.2%), emm44 (6.7%), emm42 (5.6%), and emm75.3, emm82, emm85, emm92, emm111.1(4.4% each), which was entirely different from those reported from other geographical areas. In the only other published report from India where such association was studied, detected that emm49 and emm56 were associated with macrolide resistance [23]. The emm28 and emm4 types were reported to be associated with macrolide resistance in Europe, Spain, Finland and Quebec whereas emm12 was found to be the main resistant emm type in Germany, Greece, Italy, Portugal and Israel [13]. In the United States, emm75 and emm12 accounted for majority of erythromycin resistant S. pyogenes [24]. The lack of association between any particular emm type and macrolide resistance may attest to the fact that these two genetic properties are totally unrelated. This could also be due to the relatively large number of emm types (n = 34) encountered in the present study while in the Western studies published in the literature, such associations have been more regularly found. The only importance for such a perceived association would be the possible success of a particular clone in dissemination and/or causation of a particularly severe type of infection, as, an added feature of antibiotic resistance would further complicate the situation. The observation that macrolide resistance in S. pyogenes is associated with particular emm types could be just a suspicion without any clinical or epidemiological conformational relevance.

Conclusions
In the present study, the rate of macrolide resistance is higher compared to reports from other regions of the world, with iMLS as the predominant phenotype and ermB as the main genetic determinant. This may be due to the over the counter availability and misuse of these antibiotics in this area without regular surveillance of resistant GAS genotypes and phenotypes. Findings also indicate that macrolides should not be empirically used for treatment of severe streptococcal infections. In order to control the spread of resistant clones, it is advisable to use erythromycin only after conformational laboratory tests that indicate susceptibility. The prevalent emm types associated with macrolide resistance in this geographical area is entirely different from those of Western countries; the reason requires further investigations.