Introduction

Among the wide range of Pneumococcal diseases, Pneumococcal pneumonia, which can manifest as invasive or non-invasive, accounts for 36% of overall childhood pneumonia [1]. Farooqi et al. (2015) estimated that in 2010, India witnessed 0.56 million severe episodes of pneumococcal pneumonia and 105 thousand pneumococcal deaths in children younger than five years of age [2]. Erythromycin was first introduced in 1952 [3], but it was not frequently used for the treatment of pneumonia, until the 1980s. The 1990s marked the rise in the incidence of pneumococcal penicillin resistance [4]. Macrolides, including erythromycin, possess anti-inflammatory and immunomodulatory activity in the eukaryotes [5], making them an optimal drug for upper respiratory infections, especially pneumonia. The increasing incidence of penicillin resistance led to the widespread use of macrolides, which exerted an intense selective pressure that resulted in macrolide-resistant pneumococci [6]. Recent reports show that macrolide resistance among S. pneumoniae is geographically variable, ranging from a minimum of 30 to 50% globally [7]. Erythromycin resistance is higher in Asia [8], with a trend of steady progression over the last decade compared to other continents [9, 10]. Until 2004, there were no reports of macrolide resistance from Asia, although several studies from Europe, America, and Africa reported an increasingly high percentage of macrolide resistance [11, 12]. In 2004, Song et al. (ANSORP study) reported a high incidence of macrolide resistance in Asia [11]. In India, macrolide resistance in invasive pneumococcal disease (IPD) among children less than five years increased from 13% in 1999 [13] to 50% in 2019 [14].

Macrolides bind reversibly to the 23S rRNA at a site near the peptidyl transferase, in the center of the 50S ribosomal subunit, where they interfere with the development of peptide bonds during protein elongation, thereby inhibiting protein biosynthesis [15]. The erythromycin-ribosome methylase (erm) gene encodes adenine-specific N-methyltransferases that confers macrolide resistance by the target site modification. The erm gene methylates 23S rRNA and thereby inhibits the binding of macrolide antibiotics [16]. In S. pneumoniae, erm(B) is the principal ribosomal methylase, whereas the other gene subclasses such as erm(A) [17] and erm(TR) [18] are rare. The erm(B) gene mediates high-level resistance, while the macrolide efflux pump mef(A/E) gene mediates low-level resistance. The point mutations in 23S rRNA, L4, or L22 ribosomal proteins can also mediate erythromycin resistance [16]. Although there have been reports of increasing erythromycin resistance in India [19,20,21], there are very little data on the resistance determinants. Hence, we aimed to perform the molecular characterization of erythromycin-resistant invasive pneumococcal isolates by screening for erm(B), mef(A/E) genes and by mutational analysis of both the 23S rRNA and the ribosomal proteins. We also aimed to describe the pneumococcal serotype distribution within the erythromycin-resistant isolates.

Materials and Methods

Bacterial Isolate Characterization and Antimicrobial Susceptibility Testing

This study included 250 archived non-duplicate erythromycin-resistant (blood, cerebrospinal fluid, and other sterile body fluids) invasive pneumococcal isolates in all age groups, collected from July 2014 to December 2019. This laboratory-based study was conducted at Christian Medical College (CMC), Vellore, India, the WHO Pneumococcal reference laboratory in Southeast Asia. The 250 isolates included 200 isolates from CMC, and the remaining 50 isolates from other Indian hospitals located in New Delhi (Chacha Nehru Bal Chikitsalaya, Maulana Azad Medical College), and Chennai (Kanchi Kamakoti CHILDS Trust Hospital). Most of the isolates (76%) were from blood, followed by CSF (14%) and pleural fluid (10%). The isolates from children and adults were 65% and 35%, respectively. The isolates were identified based on the typical colony morphology, gram staining, optochin sensitivity test (Oxoid Company, Britain), and the bile solubility test. Serotyping was performed by the Co-agglutination method (neufeld antisera obtained from Statens Serum Institut, Copenhagen, Denmark) and customized conventional sequential multiplex PCR according to the Centers for Disease Control and Prevention (CDC) protocol [22]. The minimum inhibitory concentration (MIC) was determined using the Vitek systems II method for erythromycin, and the results were interpreted based on the CLSI (Clinical and Laboratory Standards Institute, M100, 30th edition, 2020) guidelines. ATCC 49619 S. pneumoniae strain was the reference for antimicrobial susceptibility testing. For erythromycin, the breakpoints used were ≤ 0.25, 0.5, and 1 µg/mL for susceptible, intermediate, and resistant, respectively. The MIC value for all the study isolates was more than 1 µg/mL.

DNA Extraction, MLST, and PCR

According to the manufacturer's instructions, DNA extracted from an overnight culture grown at 37 °C on blood agar, using QIAamp DNA Mini Kit and the QIAsymphony SP instrument (Qiagen, Hilden, Germany). To detect erythromycin resistance genes mef(A/E) and erm(B), PCR was performed on all the isolates using primers described elsewhere [23]. Briefly, the PCR reaction volume consisted of 12.5 μL of master mix, 8.5 μL of nuclease-free H2O, 2 μL of primer mix, and 2 μL of DNA template, with a total volume of 25 μL. The PCR cycling conditions used were: initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 1 min, annealing at 48 °C for 1 min, and extension at 72 °C for 1 min and then final extension at 72 °C for 10 min. Among the 250 isolates, 40 were subjected to Multilocus Sequence Typing (MLST) described elsewhere [24]. The sequence type was assigned based on the allelic profile of the seven housekeeping genes using the pubmlst database (https://pubmlst.org/bigsdb?db=pubmlst_spneumoniae_seqdef). Nine isolates, negative for the presence of mef(A/E)/erm(B), underwent mutational analysis of 23S rRNA, ribosomal proteins L4 and L22 [23]. The amplified gene products were Sanger sequenced, edited, and merged using the Bioedit free software. The mutational analysis was performed using S. pneumoniae R6 as the reference sequence. The amino acid mutations were determined by translation of the DNA sequences using the ExPASy translate tool (https://web.expasy.org/translate/) and then compared with the corresponding reference S. pneumoniae R6 ribosomal proteins. The alignments were performed using ClustalW sequence alignment software (http://www.ebi.ac.uk/Tools/msa/clustalo/).

Results

Detection of Macrolide-Resistant Genes, erm(B) and mef (A/E)

Among the 250 erythromycin-resistant studied isolates, 46% (n = 114) and 35% (n = 87) isolates individually carried mef(A/E) gene and erm(B) genes, respectively. Both genes, mef(A/E) and erm(B), coexisted in 8% (n = 20) of the isolates, and neither of them were detected in 12% (n = 29) of the studied isolates.

Serotype Distribution

The predominant serotypes were 19F (n = 31), 14 (n = 32), 6A (n = 19), 6B (n = 26) , 19A (n = 16), 23F (n = 22), 9 V (n = 19), and 38 (n = 8). The serotype distribution in children and adults is shown in Fig. 1. The expected serotype coverage in children for vaccines PCV13, PCV10SII (Serum Institute of India), PCV10(GSK), PCV15, and PCV20 are 53%, 51%, 43%, 54%, and 57%, respectively. In adults, PPSV23 provides a serotype coverage of 61%. Non-vaccine serotypes 38 (8%) and 7B (4.5%) were predominant among adults. Although there were fewer samples from CSF than from blood, the predominant serotype in CSF was 14, whereas in blood and pleural fluid the predominant serotype was serotype 19F.

Fig. 1
figure 1

Serotype distribution among children and adults. Serotypes 2, 34, 36, 23A, 24F, 6C, 39, 10C, 15A, 22A, 35A, 6D, 24F, and 25F were of one each

Full size image

MLST

Among the 40 isolates, 29 sequence types were observed within 13 serotypes. Global comparison of studied STs revealed the grouping of the 29 STs into 13 clonal complexes (CC) and six singletons. The predominant CC was CC320 (n = 13), followed by CC230 (n = 4) and CC63 (n = 4). The PMEN (Pneumococcal Molecular Epidemiology Network) clones associated with these CCs are Taiwan19F-ST236, Spain 6B-2 ST90, Portugal 19FST177, Sweden15A -ST63, Denmark 14-ST230, and Sweden1-ST21 (Table 1).

Table 1 Serotype, sequence type, resistant genes, and erythromycin MIC values for study isolates
Full size table

Detection of Mutation in 23S rRNA, L4, and L22

Nine isolates, negative for erm(B) and mef(A/E) genes, were screened for mutations in both, the 23S rRNA and ribosomal proteins (L4 and L22). All the isolates showed mutations in the first part of the sequence, including domain II, whereas no mutations in the region of domain III–V was found. The mutations observed were A138G, T389C, A682G, C1180T, C1618A, T1745A, A2000C, A2288G, G2519T, G2590T, and A2751T. The earlier reported major mutations associated with erythromycin and clindamycin resistance in 23S rRNA, such as A2060C, A2061G, and C2613G, were absent. None of the isolates showed mutations in L22. The L4 ribosomal protein showed only two mutations, E161G (n = 2) and S20N (n = 3), in five isolates, while the conserved site (between the 63 and 74 amino acid positions) had no mutations (Table 2).

Table 2 Mutations observed in the 23S rRNA, ribosomal proteins L4 and L22
Full size table

Discussion

This study detected the macrolide resistance determinants prevalent among invasive pneumococcal isolates in India. In the present study, mef(A/E) was the most prevalent gene, followed by erm(B). These findings are similar to the reports of the ANSORP (Asian Network for Surveillance of Resistant Pathogens) (12) study, and contradictory to the findings by Peela et al., where erm(B) was predominant (20).

India, compared to other countries, such as South Africa, Australia, Turkey, and the USA [25,26,27,28], has reported a low percentage of isolates bearing both mef(A/E) and erm(B) genes. Isolates harboring dual macrolide resistance determinants are associated with the genetic elements of multidrug-resistant clonal complexes; hence, they are known to display resistance to multiple classes of antimicrobial agents [29]. The STs of the studied isolates having both erm(B) and mef(A/E) genes were predominantly from CC320 and have evolved from the Taiwan 19F-14 (ST236) clone, which initially had mef(A/E) and then later acquired erm(B) [30] genes. In 2004, ST236 and ST81 were the most significant erythromycin-resistant lineages reported in Asia, but the study included only one isolate from India [11]. In 2011, ST320 replaced ST236 and ST81 as the predominant lineage of erythromycin resistance in Asia [31]. However, this study involved many isolates collected from Korea and very few from India. The current study is the first to include STs of erythromycin-resistant invasive pneumococcal isolates from India.

Observation of similar penicillin MIC value of the S. pneumoniae isolates of a clonal complex to the associated individual PMEN clones explains the gradual evolution of these clones by recombination [32]. Therefore, there is clonality observed within the sequences of penicillin binding proteins(PBPs). While erythromycin resistance in S. pneumoniae is mainly due to mobile genetic elements, the clonality seen in penicillin resistance is less seen in erythromycin-resistant clones and ensures the rapid spread of resistance. In this study, the major mutations in the 23S rRNA domain V responsible for high erythromycin resistance observed in other studies, such as A2058G, C2611G(MLSB), A2059G, and C2610G, were not detected. In the current study, mutations in the 23S rRNA (A138G, G260A, T389C, A682G, C1180T, C1618A, T 1745A) were within the first 2000 bp, which encodes for the domain I and II, and no mutations in domain V of the 23S rRNA were observed. Of these, A138G and T389C have been reported previously. This result is in concordance with the findings of Canu et al. on telithromycin resistant S. pneumoniae [33]. Mutations in the domain II and IV of 23S rRNA mainly affects the macrolide resistance in S. pneumoniae [33, 34]. Among the L4 ribosomal protein mutations, only S20N has been reported earlier from Germany [35]. The absence of major ribosomal mutations indicates decreased spread of resistance mediated by ribosomal mutations. The mobile genetic elements are the predominant mode of erythromycin resistance in Indian isolates of S. pneumoniae.

The most prevalent resistant serotypes were of PCV 13 (77%). The recently introduced PCV13 could protect against invasive disease and reduce antimicrobial resistance among the vaccine serotypes in India. Serotype 14 (13%, Fig. 1) was the major serotype, followed by 19F, the same as previous reports from India [14, 21, 28]. These two major serotypes correlate with the predominant CCs associated serotypes: CC320, which is associated with Taiwan19F-ST236 and Denmark14-ST230 clones. Recent reports of change in the serotypes, antimicrobial susceptibility, and clonality associated with the introduction of pneumococcal childhood vaccination [36,37,38] mean that it will be imperative to monitor the emergence of non-vaccine serotypes with resistance due to vaccine pressure.

Conclusion

This study contributes to a better understanding of the baseline resistance determinants associated with the serotypes and sequence types of Indian invasive S. pneumoniae resistant to erythromycin. This study reports the predominance of mef(A/E) gene-mediated resistance rather than the ribosomal mutations, with the majority being PCV13 serotypes. The rapid spread of erythromycin resistance mediated by mobile genetic elements highlighted the need to discontinue the misuse of macrolides to treat upper respiratory tract infections. The presence of erythromycin-resistant non-vaccine serotypes demands monitoring the prevalent serotypes and their antimicrobial resistance profiles continuously.