Abstract
In recent years, nosocomial infections due to multidrug resistant Klebsiella pneumoniae strains have been increasing, associated with growing trend of resistance to beta-lactams and fluoroquinolones (FQs) worldwide. In this study, prevalence of beta-lactamase genes and resistance mechanisms to FQ were analyzed in 191 clinical K. pneumoniae isolates derived from respiratory tract infections in a teaching hospital in Yangon, Myanmar. The major extended-spectrum beta-lactamase gene was blaCTX-M, which was detected in 33% of isolates, with CTX-M-15 being dominant. Fourteen isolates (7.3%) harbored carbapenemase genes that were genotyped as blaNDM-1, blaNDM-5, or blaNDM-7. The most common plasmid-mediated quinolone resistance (PMQR) gene was aac6′-Ib-cr (51.8%), followed by qnrB (41.9%), oqxAB (23%), and qnrS (15.2%). In quinolone-resistance determining region of GyrA, eight different types of mutation were identified for FQ-resistant isolates, with double mutations at two positions (S83F, D87A) being most common (54.6%). Isolates with double mutations (three patterns) showed higher minimal inhibitory concentration to levofloxacin (LVX) (≥64 μg/mL) than those with a single mutation. PMQR gene profiles, including aac6′-Ib-cr and any other gene(s), were generally related to higher resistance level to LVX. K. pneumoniae isolates with different profiles of beta-lactamase genes and FQ-resistance determinants were mostly classified into ST15 or its single-locus variant (SLV). The most common NDM gene, blaNDM-5, was detected in ST975 (ST15-SLV) isolates and an ST4000 isolate. The present study revealed the wide spread of FQ-resistant K. pneumoniae clinical isolates acquiring various FQ-resistance determinants and beta-lactamases that were presumably derived from a single clonal lineage in a hospital in Myanmar.
Introduction
K
Beta-lactamases, that is, the most common cause of resistance to beta-lactams, have been classified into four major groups; penicillinases, cephalosporinases, extended-spectrum beta-lactamases (ESBLs), and carbapenemases. ESBLs consist of genetic variants of TEM and SHV and CTX-M, among which CTX-M is dominant globally. 3 In addition to ESBL, plasmid-mediated AmpC (pAmpC) beta-lactamases are also responsible for resistance to broad-spectrum cephalosporins among Enterobacterales. 4 Carbapenem resistance in Enterobacterales is primarily mediated by carbapenemases, which have been classified into Ambler class A (KPC, IMI, GES types, etc.), B (NDM, IMP, VIM types, etc.), and D (OXA types) enzymes. 5 Among them, class B carbapenemases (metallo beta-lactamases) exhibit a broad spectrum of activity to all penicillins, cephalosporins, and carbapenems except for aztreonam. NDM is recognized as an emerging class B carbapenemase, and NDM-producing Gram-negative bacteria have been reported almost worldwide. 6
FQs, that is, new quinolones, are one of the main therapeutic choices for infections caused by bacteria belonging to Enterobacterales. They are the most frequently used antimicrobial agents worldwide due to their broad-spectrum antimicrobial activity. 7 However, as their use becomes widespread, FQ resistance has occurred and increased among several clinically important bacteria, including K. pneumoniae.8,9 FQ resistance is primarily caused by occurrence of mutation(s) in GyrA subunit of DNA gyrase and ParC subunit of topoisomerase IV. 10 In addition, plasmid-mediated quinolone resistance (PMQR) determinants represented by Qnr, AAC(6′)-Ib-cr, and QepAB/OqxAB confer reduced susceptibility to FQs. 11 It was described in some reports that previous FQ use is an independent risk factor for infections caused by ESBL-producing Escherichia coli, 12 as well as carbapenem-resistant K. pneumoniae infection. 13 This may be partly explained by the fact that multidrug resistance phenotype is possible to be conferred by a single mechanism such as multidrug efflux pumps, which is activated by FQ. 14 Moreover, when PMQR determinants coexist with ESBLs or carbapenemase genes on the same plasmid that have cross-species/genera transferability, 15 bacterial strains with resistance to broad spectrum beta-lactams could spread by selective pressure exerted by FQ. Accordingly, it is important to monitor prevalence of FQ resistance along with beta-lactam resistance. However, in Myanmar, such study has scarcely been performed for clinical isolates of K. pneumoniae, while there is only a study on E. coli. 16 In the present study, the prevalence of beta-lactamases/carbapenemase genes, quinolone-resistance determining region (QRDR) mutations, and PMQR determinants was investigated for clinical K. pneumoniae isolates derived from patients with respiratory tract infection.
Materials and Methods
Bacterial isolates
A total of 191 nonduplicate consecutive clinical isolates of K. pneumoniae derived from patients with respiratory infections were analyzed. These isolates were collected from the North Okkalapa General Hospital (NOGH), a tertiary care teaching hospital of University of Medicine (2), Yangon, during 1-year period starting January 2018. Median age of patients was 51.5 years (ranging from 16 to 87 years old), and sex ratio (male/female) of patients was 1.2 (105/86). The clinical specimens (sputum) were inoculated onto MacConkey agar and blood agar plates at 37°C overnight aerobically. Identification of bacterial species was performed by automated system (Vitek 2 compact; Biomerieux). Furthermore, K. pneumoniae was confirmed by PCR targeting infB, a locus included in multilocus sequence typing (MLST) of this bacteria species. 17 Isolates were stored in Microbank (Pro-Lab Diagnostics, Richmond Hill, Canada) at −80°C and recovered when they were analyzed. This study was approved by the Research Ethics Committee of University of Medicine (2), Yangon, Myanmar.
Antimicrobial susceptibility testing
For all K.pneumoniae isolates, antimicrobial susceptibility test was initially performed by automated system (Vitek 2 compact; Biomerieux) for following antimicrobial agents: cefepime, cefuroxime, cefotaxime, ceftriaxone, ceftazidime, amoxicillin–clavulanic acid, ciprofloxacin, levofloxacin (LVX), piperacillin–tazobactam, ampicillin/sulbactam, cefoperazone/sulbactam, gentamicin, amikacin, aztreonam, trimethoprim–sulfamethoxazole, tetracycline, meropenem, imipenem, and colistin. Resistance and susceptibility were distinguished according to the breakpoints defined in the Clinical Laboratory Standard Institute (CLSI) guidelines. Minimal inhibitory concentration (MIC) of LVX was measured by broth microdilution method for all the quinolone resistant isolates according to CLSI guidelines. 18 Broth microdilution test was also used to determine MIC to colistin and meropenem for confirmation.
Detection and characterization of beta-lactamase genes and PMQR genes
Beta-lactamase genes blaCTX-M, blaTEM, and blaSHV were detected by multiplex PCR as described previously. 19 Four blaCTX-M subgroups (group 1, 2, 9, and 8/25/26) were discriminated by multiplex PCR assay. 20 For all the isolates, presence of carbapenemase genes (blaNDM, blaVIM, blaIMP, blaSPM, blaAIM, blaGIM, blaBIC, blaSIM, blaDIM, blaKPC, blaIMI, blaGES, and blaOXA-48) was identified by multiplex/uniplex PCR using primers and conditions as described previously.21–23 In addition, six families of AmpC beta-lactamase genes were detected by multiplex PCR as described by Perez-Perez and Hanson. 24 Full-length nucleotide sequences of blaCTX-M, blaTEM, blaSHV, and carbapenemase genes (blaNDM) were determined directly from PCR products amplified by primers listed in Supplementary Table S1, using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) on an automated DNA sequencer (ABI PRISM 3100). To assign subtypes of beta-lactamase genes, homology search for related sequences was performed using standard nucleotide BLAST (Basic Local Alignment Search Tool) available at the NCBI website.
Identification of PMQR genes (aac 6′-Ib-cr, qnrA, qnrB, qnrC, qnrD, qnrS, oqxAB, and qepA) was also performed by multiplex PCR using primers and conditions as described previously 25 and confirmed by direct sequencing of the PCR products. Presence of mutation in QRDR of DNA gyrase (GyrA) and topoisomerase IV (ParC) was analyzed for all the isolates by PCR and direct sequencing, using primers listed in Supplementary Table S1. For all the carbapenem resistance isolates, detection of mcr genes was attempted by PCR, as described previously. 26
Genetic analysis of ESBL/carbapenemase-producing K. pneumoniae
Sequence type (ST) of selected K. pneumoniae isolates was determined based on the scheme of the Pasteur Institute, by determination of partial sequences of seven housekeeping genes (gapA, infB, mdh, pgi, phoE, rpoB, and tonB). 18 Phylogenetic dendrogram of concatenated sequence of the seven ST loci was constructed by the maximum likelihood method using the MEGAX software.
GenBank accession numbers
The nucleotide sequence of beta-lactamase genes encoding NDM-1, -5, and -7, CTX-M-15, -55, -27, -9, and -3, TEM-1, and SHV-1, -11, -28, -75, -202, -213, and -27 were deposited in the GenBank database under accession numbers MN786376 - MN786391 (Supplementary Table S2).
Results
Among a total of 191 K. pneumoniae isolates, resistance rates to antimicrobials in the decreasing order are as follows: cefuroxime, 75.4%; trimethoprim–sulfamethoxazole, 72.3%; amoxicillin–clavulanic acid, 70.2%; ciprofloxacin, 62.3%; LVX, 62.3%; tetracycline, 61.3%; ceftriaxone, 60.2%; ceftazidime, 60.2%; cefotaxime, 56.5%; ampicillin/sulbactam, 56%; cefepime, 53.9%; piperacillin–tazobactam 52.4%; gentamicin, 45%; aztreonam, 38.2%; cefoperazone/sulbactam, 32.5%; amikacin, 21.5%; meropenem, 14.7%; imipenem, 14.7%; and colistin, 4.7%.
Beta-lactamase genes blaTEM, blaSHV, blaCTX-M, and blaNDM were identified with detection rates 45.5%, 36.1%, 33.0%, and 7.3%, respectively (Table 1). These beta-lactamase genes were detected mostly within LVX-resistant isolates (n = 119), while LVX-susceptible isolates contained only three isolates with blaTEM. Among the FQ-resistant isolates, prevalence of blaCTX-M was 52.9%. According to the genotyping of the beta-lactamase genes, ESBL gene was mostly blaCTX-M, with CTX-M-15 being dominant. In contrast, except for SHV-27, all the TEM and SHV genes were assigned to non-ESBL types, including dominant types TEM-1 and SHV-28. NDM gene was identified in 14 isolates and assigned into three types, NDM-5 (9 isolates), NDM-1 (4 isolates), and NDM-7 (an isolate), while other carbapenemase genes (blaVIM, blaIMP, blaKPC, and blaOXA-48) were not detected. No isolates were positive for pAmpC gene. Although nine isolates, including four isolates with NDM-5 and two isolates with NDM-1, showed colistin MIC 4 μg/mL that represent resistance, mcr genes (mcr-1, mcr-2, mcr-3) were not detected in any isolates.
Prevalence of Beta-Lactamase Genes and Plasmid-Mediated Quinolone Resistance Genes in Clinical Isolates of Klebsiella Pneumoniae Isolates (n = 191)
Following genes were not detected in any isolate: qnrA, qnrC, and qnrD.
LVX, levofloxacin; PMQR, plasmid-mediated quinolone resistance.
Among the FQ-resistant isolates (n = 119), 106 isolates (89.1%) harbored any of the five PMQR genes (aac-(6)′-Ib-cr, qnrB, qnrS, oqxAB, and qepA). aac-6′-Ib-cr was the most commonly detected (71.4%), followed by qnrB (59.7%), oqxAB (37%), and qnrS (21.8%) (Table 1). In the FQ-susceptible isolates (n = 72), aac-(6)′-Ib-cr, qnrB, and qnrS were found at low rate (4–20%). Among CTX-M-positive isolates, prevalence of aac-(6)′-Ib-cr was 71.4% (45/63), which was higher than those of qnrS, qnrB, and oqxAB (25.4–50.8%) (Supplementary Table S3).
Mutations in QRDR of GyrA were detected in all the 119 isolates showing resistance to LVX (MIC: 8– ≥128 μg/mL) and assigned into any of the 8 patterns at position 83 and/or 87 of GyrA (Table 2), while no mutation was found in ParC. The most common GyrA mutation was a double mutation S83F+D87A (54.6%), followed by S83I (20.2%). All the isolates with any of the three double mutations (S83F+D87A, S83L+D87N, S83Y+D87A) showed ≥64 μg/mL of LVX-MIC, which was higher than most of isolates with single mutation (8–32 μg/mL). Only some isolates with single mutation S83I and S83Y showed high LVX-MIC (≥64 μg/mL).
Frequency of GyrA-Quinolone Resistance Determining Region Mutations and Resistance Level Against Levofloxacin of Fluoroquinolone-Resistant K. pneumoniae Isolates
MIC, minimal inhibitory concentration; QRDR, quinolone-resistance determining region.
PMQR genes were detected for 106 FQ-resistant isolates and classified into 15 profiles, among which aac-(6)′-Ib-cr+qnrB (26.9%) and aac-(6)′-Ib-cr+qnrB+oqxAB (19.3%) were the most frequently observed (Table 3). Although relatedness of the presence of individual PMQR genes to LVX-MIC was not generally evident, high MIC to LVX (≥64 μg/mL) was observed for some PMQR gene profiles, that is, aac-(6)′-Ib-cr+qnrB, aac-(6)′-Ib-cr+oqxAB, aac-(6)′-Ib-cr+qnrB+oqxAB, and aac-(6)′-Ib-cr+qnrB+qnrS. Among isolates with all the seven PMQR profiles with aac6′-Ib-cr and any other gene(s), 77.6% (59/76) of isolates were related to higher resistance level to LVX (MIC, ≥64 μg/mL).
Plasmid-Mediated Quinolone Resistance Genes, Plasmid-Mediated Quinolone Resistance Gene Profiles, and Levofloxacin Minimal Inhibitory Concentration of Fluoroquinolone-Resistant K. Pneumoniae Isolates
STs were determined for the selected 30 FQ-resistant K. pneumoniae isolates with different profiles of beta-lactamase genes and PMQR genes (Table 4). These isolates mostly belonged to ST15 and its single-locus variants (SLVs), that is, ST14, ST975, ST4702, and ST4703, among which ST4702 and ST4703 were novel STs identified in this study. Four isolates with NDM-1 and CTX-M-15 genes were assigned to ST15, while other four isolates were harboring NDM-5, SHV-28, and TEM-1 genes to ST975. One isolate each with blaNDM-7 and blaNDM-5 were classified into ST16 and its triple-locus variant (ST4000), respectively. Among the 14 NDM-positive isolates, 9 and 10 isolates possessed blaCTX-M and aac-(6)′-Ib-cr, respectively, and double mutation in QRDR of GyrA was identified in 9 isolates (Supplementary Table S4). An isolate (ST16) with NDM-7 gene harbored four PMQR genes.
Genetic Characteristics of the Fluoroquinolone-Resistant K. pneumoniae Isolates with Different Beta-Lactamase Genes
SLV; TLV.
Novel STs identified in this study.
Underline represents variant locus number of SLV/TLV shown in the left column.
F, female; M, male; SLV, single-locus variant; ST, sequence type; TLV, triple-locus variant.
Concatenated sequences of STs identified in the present study were analyzed phylogenetically, together with those of representative drug-resistant clones reported previously.1,27 As shown in Fig. 1, STs in our study mostly belonged to clonal group (CG) 15 (CG15), CG17, and CG147, among which CG15 and CG17 included STs of carbapenemase-positive isolates.

Phylogenetic dendrogram of concatenated sequence of the seven ST loci constructed by the maximum likelihood method using the MEGAX software. Trees was statistically supported by bootstrapping with 1000 replicates, and genetic distances were calculated by Kimura two-parameter model. Variation scale is described at the bottom. Percent bootstrap support is indicated by the values at each node (the values <80 are omitted). Filled circle and open circle indicate STs identified in the present study with and without carbapenemase genes, respectively. Clonal groups are indicated on the right. ST, sequence type.
Discussion
The present study revealed prevalence of ESBL and carbapenemase genes, PMQR genes, and QR-related GyrA mutations for K. pneumoniae first time in Myanmar. The prevalence of ESBL gene (blaCTX-M) in K. pneumoniae from respiratory infections was 33%, which was higher than overall detection rate of ESBL in K. pneumoniae from intra-abdominal infections (24.3%) in Asia-pacific region (2010–2013). 28 However, the ESBL rate in our study was comparable to those in Thailand (36.5%), Vietnam (30.4%), Malaysia (28.5%), and so forth. 28 Although the TEM and SHV genes showed higher prevalence than CTX-M-type beta-lactamase, they were mostly assigned to non-ESBL type, and blaCTX-M was exclusive ESBL gene identified, with CTX-M-15 being predominant. This finding may be reflected by the global trend showing the spread of CTX-M-15 among K. pneumoniae through dissemination of plasmids and transposons encoding blaCTX-M-type ESBLs. 29
Carbapenemases that are produced by K. pneumoniae have been classified into three major groups, that is, KPCs (class A), NDMs (class B), and OXA-48-like (class D), 1 among which blaKPC has become the most prevalent globally. While being endemic in the Indian subcontinent, K. pneumoniae harboring blaNDM has been spreading worldwide. 30 In Myanmar, carbapenemase gene-positive rate in clinical isolates of E. coli was 6.1% (26/426) in our previous study. 16 Another report in Myanmar described this rate as 3.4% among all the isolates of Enterobacterales. 31 In both studies, most of carbapenemases were identified as NDM-type, including NDM-1, NDM-4, NDM-5, and NDM-7, with NDM-5 being dominant.16,31 Similarly to these findings on E. coli, our study indicated that carbapenemase genes in K. pneumoniae were all identified as blaNDM with detection rate of 7.3%. Accordingly, it is suggested that NDM-type carbapenemase has been disseminating among Gram-negative pathogenic bacteria in Myanmar. Furthermore, detection of NDM-type beta-lactamase genes was reported in isolates from environment (sewage) and foodstuffs in Yangon, Myanmar.31,32 These findings suggested the wide distribution of carbapenemase-producing bacteria that may pose an increased risk of refractory infections to local people.
It was of note in our present study that ST15 and its SLVs were identified in most of all the K. pneumoniae isolates harboring ESBL and/or carbapenemase genes, as well as quinolone-resistance determinants. K. pneumoniae strains belonging to CC258 (ST258, ST11, ST340, ST512) have been known as the major global clonal groups with multidrug resistance that carries carbapenemase genes (mainly blaKPC).1,27 CC15 (ST14, ST15) is described as the second most prevalent clonal group of multidrug resistant K. pneumoniae that caused outbreaks. 1 ST15, as well as ST147, K. pneumoniae were shown as those being linked with multidrug resistance worldwide, with geographically diverse distribution, while being most prevalent in Europe. 33 It was indicated that multidrug-resistant ST15 K. pneumoniae in the UK, Ireland, have genetically divergent nature and are related to global isolates, including those from Asia (Nepal, Singapore, Laos, Indonesia).33,34 Accordingly, it is suggested that Myanmar might have been also involved in the spread of this clone, due to movement of people and any genetic features of this clone allowing its successful transmission.
NDM-producing K. pneumoniae isolates belong to various STs, including ST14, ST15, ST16, ST17, ST20, ST37, ST101, ST147, ST258, ST340, ST512, and ST972, among which ST14, ST147, and ST340 have been found as those associated with blaNDM in many countries. 30 Particularly, ST11 (CG258) and ST147 (CG147) were the major clones harboring blaNDM and found in India. 35 Therefore, ST15 (CG15) and ST16 (CG17), which were identified in our study, are considered minor clones of K. pneumoniae producing NDM-type carbapenemases. However, STs of CG15 were identified most frequently in our present study, irrespective of the presence of carbapenemase genes, suggesting the predominance of CG15 K. pneumoniae in Myanmar. Meanwhile, ST15 K. pneumoniae harboring blaNDM-1 has been reported in Nepal 33 and China, 36 thus its potential spread in Asia may be of concern and needs to be monitored.
The high incidence of ST15 and its SLVs among various isolates with different profiles of beta-lactamases and quinolone resistance determinants indicated a clonal dissemination of multiple-drug resistant K. pneumoniae in the hospital of this study. According to the only available information for K. pneumoniae clinical isolates in Myanmar, ST147 and ST101 were major genotypes among blaNDM-positive isolates, which were mostly derived from blood and urine. 31 In contrast, from foodstuffs, ST15, ST16, ST39, ST40, ST395, and ST1537 were reported for blaNDM-positive isolates. 32 Because only isolates from respiratory infections (sputum) were analyzed in our present study, it is possible that distinct K. pneumoniae clones are prevalent depending on different infection types or specimens. Thus, it remains to be determined whether ST15 K. pneumoniae with blaNDM is prevalent in diseases other than respiratory infections.
Prevalence of diverse PMQR genes in Gram-negative bacteria has been increasingly associated with rise in quinolone resistance worldwide. 37 PMQR, which causes reduced susceptibility to quinolones, is considered to promote occurrence of mutation in QRDR of GyrA/ParC leading to high level FQ resistance. 38 However, higher prevalence of QNR genes than incidence of the gyrA/parC mutations, presumably due to protection of GyrA/ParC by Qnr, was described in China.39,40 In contrast, in our present study, mutation(s) in GyrA was detected in all the 119 isolates showing resistance to FQ, among which ∼90% of isolates harbored any of PMQR genes. Similarly high incidence of both GyrA mutations and PMQR genes was reported in a university hospital in Egypt. 41 These findings indicate that incidence of GyrA mutations is not predictable by prevalence of PMQR genes.
As reported previously, presence of double mutations in GyrA causes higher resistance level to quinolones than single mutation, 8 which was also observed in our present study. Furthermore, carriage of qnr alleles or qepA was also implicated in higher FQ resistance level. 8 However, PMQR genes were not clearly correlated to higher MIC to LVX in our study, while PMQR profiles with multiple genes with aac(6′)-Ib-cr appeared to be associated. Accordingly, resistance level to FQ may not be defined by PMQR gene alone, but primarily by presence of single or double mutations in GyrA, in combination with PMQR genes. Resistance level defined by PMQR may be variable due to difference in plasmid copy number and gene expression, as suggested previously. 8
Among the mutations in QRDR of GyrA, S83L and D87N/A have been the most frequently reported for FQ-resistant K. pneumoniae.40–42 In the present study, the most common mutations were S83F and D87A, among which S83F had been rarely identified, while there are reports of its detection in China 42 and Spain. 43 It is suggested that prevalence of the rare type of GyrA mutation might be caused by clonal spread of a specific ST15 clone. Such rare mutation in this study may be utilized as a marker to distinguish other FQ-resistant strains of different sources, to analyze spread of the ST15 multidrug-resistant K. pneumoniae over other infection types and other areas in Myanmar. Antimicrobial resistance is global concern, and its control needs government sectors and community participation. As described previously, 44 an antimicrobial stewardship program (ASP) should include written guidelines for prescription of appropriate antibiotics at correct dosage, shortest duration of treatment, carrier isolation, and rapid diagnosis of viral infections that reduce unnecessary usage of antibiotics. Further promotion of ASP is anticipated in Myanmar for the control of multidrug-resistant bacteria.
This study revealed high prevalence of ESBL and quinolone-resistance determinants, along with emergence of NDM-type carbapenemase genes among K. pneumoniae from respiratory infections. The results underscore the importance of further surveillance of multidrug resistant K. pneumoniae in Myanmar.
Footnotes
Disclosure Statement
The authors of this article have no commercial associations that might create a conflict of interest in connection with the submitted article.
Funding Information
This study was supported, in part, by JSPS (Japan Society for the Promotion of Science) KAKENHI Grant No. 17H04664.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
