First Emergence of NDM-5 and OqxAB Efflux Pumps Among Multidrug-Resistant Klebsiella pneumoniae Isolated from Pediatric Patients in Assiut, Egypt

Introduction

Currently, there is increased awareness of the impact of Gram-negative bacteria that are resistant to carbapenem as it is the last choice treatment for bacterial infections.^1,2 Enterobacteriaceae exhibit resistance to carbapenems by three possible mechanisms: production of carbapenemase enzymes, efflux pump activity, and porin loss.^3,4 Carbapenemase production constitutes the basic mechanism of carbapenem resistance and study of this mechanism is essential for the best choice of recently effective antibiotics.^5,6

New Delhi metallo-β-lactamase (NDM) is considered the latest threat for public health and numerous variants of NDM-type carbapenemases have been identified among Gram-negative bacteria worldwide.⁷ Among them, NDM-5 was a variant with increased carbapenemase activity in comparison with NDM-1.⁸ The bla_NDM-5 gene has been identified in E. coli and Klebsiella pneumoniae from many countries.⁹ In addition, bla_NDM-5 was reported to be carried on different incompatibility typing plasmids that are responsible for transferring as IncF, IncN and IncX3.¹⁰ These plasmids can facilitate the dissemination of bla_NDM-5 among the members of Enterobacteriaceae through horizontal gene transfer¹¹ that cause the bacterial cells to become a severely hazardous strain.¹²

Efflux pumps are found in nearly all bacteria and are responsible for mediating resistance to antimicrobials by reducing intracellular concentration of antibiotics and promoting site mutation accumulation. Presence of OqxAB efflux pump confers resistance to many drugs such as quinolones, quinoxalines, tigecycline, chloramphenicol and nitrofurantoin.^13,14 Efflux pumps are considered as one of the key mechanisms of antibiotic resistance in K. pneumoniae isolates. OqxAB, encoded by the oqxA and oqxB genes, is a plasmid-encoded multi-drug efflux pump first identified in E. coli (located on a conjugative plasmid). Their presence was found to develop multi-drug resistance against different groups of antimicrobials, detergents, and disinfectants and on the chromosome of K. pneumoniae.¹⁵ This study aimed to determine NDM variants among K. pneumoniae strains isolated from pediatric patients in Assiut governorate, Egypt, analyze the presence of OqxAB genes and carry out the molecular characterization of bla_NDM-5-producing K. pneumoniae.

To the best of our knowledge, this study is the first to isolate K. pneumoniae carrying bla_NDM-5 and OqxAB from pediatric patients in Egypt.

Materials and Methods

K. pneumoniae Isolates and Identification

Fifty-six K. pneumoniae isolates were isolated from pediatric patients admitted to different departments at Assiut University Children’s Hospital over a period of 6 months, from September 2022 to February 2023. This study was approved by scientific research ethics committee at the faculty of pharmacy, Al-Azhar University, Egypt (reference number of ZA-AS/PH/12/C/2022). K. pneumoniae were isolated from blood samples, sputum, endotracheal aspirate samples, urine samples, stool, wound or throat swabs. Isolates were identified by API20E kit (BioMerieux, Marcy L Etoile. France). Demographic and clinical data of patients were obtained from the hospital information system.

Antimicrobial Susceptibility Testing

The antimicrobial susceptibility of the isolated K. pneumoniae was tested by the Kirby-Bauer disc diffusion method according to recommendations of the clinical laboratory standards institute.¹⁶ The following 14 commercial antimicrobial discs were used: ampicillin (10 μg), amoxicillin-clavulanic acid (20/10 μg), piperacillin (100 μg), ceftriaxone (30 μg), ceftazidime (30 μg), cefoperazone (75 μg), meropenem (10 μg), trimethoprim/sulfamethoxazole (1.25/23.75 μg), aztreonam (30 μg), amikacin (30 μg), nitrofurantoin (300 μg), levofloxacin (5 μg), tetracycline (30 μg), and cefepime (30 μg). The E. coli ATCC 25922 isolate was used for quality control. The test was repeated twice for each isolate.

Detection of Carbapenemase-Producing Isolates

 Isolates were considered carbapenemase producing by modified carbapenem inactivation methods (mCIM) test.¹⁷ In addition, the minimum inhibitory concentrations (MIC) of meropenem against the tested K. pneumoniae were tested using Meropenem E-test strips (bioMérieux, Solna, Sweden).

Phenotypic Evaluation of Colistin Resistant K. pneumoniae

Minimum inhibitory concentrations (MICs) were determined by broth microdilution method according to European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints for colistin against Enterobacteriaceae.¹⁸

PCR for Detection of Resistant Genes

DNA was extracted from overnight grown culture using Wizard Genomic DNA Purification Kit (Promega, WI, USA). The extracted DNA was stored at −20°C. K. pneumoniae isolates were tested by PCR for the detection of resistant genes. K. pneumoniae isolates were analyzed for the presence of the ESBL gene (bla_CTX-M), carbapenemase genes (bla_NDM, bla_KPC)¹⁹ and for colistin resistant genes (mcr1, mcr2).^20,21 Amplified DNA fragments of bla_NDM were purified using QIAquick PCR Purification Kit (QIAGEN, Crawley, UK) and sequenced in both directions. Nucleotide and deduced amino acid sequences were analyzed and compared by BLAST, as implemented by the National Center for Biotechnology Information website (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Detection of oqxA and oqxB Genes Efflux Pump

The oqxA and oqxB genes were screened using a PCR-based technique. The presence of oqxA and oqxB was detected using primers oqxAF (5′-CTCGGCGCGATGATGCT-3′) and oqxAR (5′-CCACTCTTCACGGGAGACGA-3′), with products of 392 bp, and oqxBs (5′-TTCTCCCCCGGCGGGAAGTAC-3′) and oqxBa2 (5′-CTCGGCCATTTTGGCGCGTA-3′), with products of 512 bp, as described previously.¹⁵

Statistical Analysis

R-packages (version 4.2.3) were used in data analysis and visualization. Wilcoxon rank sum test and Mann–Whitney U-test were used to compare meropenem MIC between NDM-5 and NDM-1-positive isolates and meropenem MIC before and after addition of CCCP, respectively. P-value of less than 0.05 was regarded as significant. Clustering of samples was elucidated based on the Euclidean distance matrix.

Phenotypic Evaluation of the Efflux Pump in NDM-5-Producing K. pneumoniae

Phenotypic evaluation of the efflux pump in NDM-5-producing K. pneumoniae was assessed by measuring the minimum inhibitory concentrations (MICs) for ciprofloxacin and meropenem before and after exposure to the efflux pump inhibitor (EPI), carbonyl cyanide 3-chlorophenylhydrazone (CCCP) (Sigma-Aldrich, Dorset, UK) at a concentration of 20 mg/L. If the MIC values decreased 4-fold or greater in the presence of EPI, this was defined as a significant inhibition effect.²²

Conjugation Experiment and Plasmid Replicon Typing

The horizontal gene transfer ability of bla_NDM-5 was determined using liquid mating assay for nine K. pneumoniae isolates that harbored bla_NDM-5. E. coli J53 was used as recipient isolate, and transconjugants selection was performed on MacConkey agar plates containing meropenem (2 µg/mL) and sodium azide (100 µg/mL). Transconjugants were tested for bla_NDM-5 by PCR.²³ PCR-based replicon typing (PBRT) was done to determine the major plasmid incompatibility group by using 14 pairs of primers in which multiplex PCR for FII, FIA and FIB was performed while 11 simplex PCRs were performed for IncL, IncM, IncT, FIC, FIIK, IncN, IncX3, IncH12, IncW, IncY and IncA/C.²⁴

Genotyping by ERIC

PCR amplification was performed using the primers ERIC1 (5′-ATGTAAGCTCCTGGGGATTCAC-3′) and ERIC2 (5′-AAGTAAGTGACTGGGGTGAGCG-3′) for K. pneumoniae isolates.²⁵ The gel bands were presented by PyElph version 1.4.²⁶ The dendrogram was computed by the unweighted pair group method with arithmetic averages (UPGMA).

Results

Bacterial Isolates and Patient Details

Fifty-six K. pneumoniae isolates recovered from 56 (male 31, female 25) pediatric patients were analyzed; 21 patients were from chest unit and eight from gastroenterology, six from the intensive care unit, five from neurology, and five from hematology wards. Most of the isolates were recovered from endotracheal aspirate (25%, 14/56), followed by sputum (23.21%, 13/56) and blood (21.42%, 12/56). The clinical details of all patients are given in Table 1.

Table 1 Clinical Details and MIC of Colistin and Meropenem for All Isolates

Antimicrobial Susceptibility Testing

The antibacterial susceptibility profiles to all isolates showed that all tested isolates (n = 56) were resistant to ampicillin, amoxicillin-clavulanate, piperacillin, ceftriaxone, ceftazidime, cefoperazone and aztreonam (Figure 1).

Figure 1 Antibacterial resistance pattern of K. pneumoniae isolates.

Isolates Producing Carbapenemase

mCIM test showed that 46 isolates (82.1%) of the tested K. pneumoniae were carbapenemase enzyme producers.

The MIC Values for Colistin and Meropenem

It was found that 67.85% (38/56) of K. pneumoniae isolates were meropenem resistant [MIC ≥4 μg/mL], 17.85% (10/56) isolates were meropenem sensitive [MIC ≤1 μg/mL] and 14.28% (8/56) isolates were meropenem intermediate [MIC >1<4 μg/mL] as illustrated in Table 1.

MIC values for all 56 isolates for colistin by broth microdilution ranged from 0.5–32 µg/mL and meropenem MIC by E-test ranged from 0.19–32 µg/mL (Table 1). There is a significant increase of meropenem MIC of NDM-5-positive isolates (P-value = 1.704e-14) than MIC of NDM-1-positive isolates.

Detection of Resistant Genes and OqxAB by PCR

Out of 56 isolates, 29 isolates were positive for bla_NDM-1, 9 isolates were positive for bla_NDM-5, and 15 isolates were positive for bla_KPC carbapenemase genes. In addition, 38 isolates were positive for bla_CTX-M ESBL gene, and 15 isolates were positive for mcr1. Both OqxA and OqxB were detected in 26 isolates. Thirteen isolates were positive for OqxA only and 11 isolates were positive for OqxB only (Figure 2).

Figure 2 Dendrogram based on the Euclidean distance matrix between resistance factors of K. pneumoniae isolates in all samples. The presence and absence of the resistance factors are represented by black or white blocks, respectively.

Effect of the Efflux Pump in NDM-5-Producing K. pneumoniae

MIC of ciprofloxacin in NDM-5-producing K. pneumoniae ranged from 4–10 µg/mL due to the inhibition of efflux pump by CCCP. Eight of the nine (8/9) NDM-5-producing K. pneumoniae had 4-fold or more decrease in MIC after CCCP exposure. Only isolates St-26 and St-42 had an MIC fold change of 2 after CCCP addition (P-value = 0.00042). Conversely, no change in MIC of meropenem after CCCP addition was observed (Table 2).

Table 2 Characterization of NDM-5-Producing K. pneumoniae

Plasmid Replicon Typing and Conjugation Experiment

All NDM-5-producing isolates except one isolate (St-29) could transfer their plasmids by conjugation to their corresponding transconjugants (E. coli J53). The bla_NDM-5 gene was located on six different types of plasmids (FII, FIIK, FIB, FIC, L and Inc M) and the FII type was predominant in NDM-5-producing K. pneumoniae (Table 2).

Genotyping by ERIC

Dendrogram of NDM-5-producing K. pneumoniae presented by PyElph showed that similar strains were found between these isolates (St-37, St-29, St-40) from hematology, neurology and chest and similarity was also detected between two isolates (St-42, St26) from the chest and gastroenterology departments (Figure 3).

Figure 3 Dendrogram of NDM-5-producing K. pneumoniae presented by PyElph.

Discussion

Carbapenem antibiotics are used in the treatment of infections caused by multi-drug resistant K. pneumoniae. However, the emergence and spread of NDM-producing K. pneumoniae has been a serious challenge to manage in the clinic because of the rapid worldwide spread of multi-drug resistance.²⁷ As one main type of carbapenemases, New Delhi metallo-β-lactamase (NDM) can resist almost all β-lactams, including carbapenems. NDM-5, compared with NDM-1, has two amino-acid substitutions (Val88→Leu) and (Met154→Leu), which confer enhanced hydrolytic activity against carbapenems.¹¹ Most studies focus on the dissemination of carbapenemase-producing Gram negative strains among adult patients more than from children. There was a spread of carbapenem-resistant K. pneumoniae and endemicity of NDM was reported in many hospitals of our region.^28,29

In this study 56 K. pneumoniae isolates were recovered from hospitalized pediatric patients at Assiut University Children’s Hospital. Mostly patients were from the chest unit and most of the specimens were recovered from endotracheal aspirate. We observed previously that many MDR K. pneumoniae isolates can maintain and spread easily in the ward of PICU of this hospital.³⁰ So, it is essential to estimate molecular characterization of this pathogen from other wards in the same hospital and to show the responsibility of this pathogen to cause a great challenge for infection monitoring among pediatric patients. The antibacterial susceptibility profiles of all isolates showed that all tested isolates were multi-drug resistant as each isolate were resistant to at least three different classes of antimicrobial agents.³¹

Among the 56 K. pneumoniae isolates, 46 isolates were confirmed phenotypically to produce carbapenemase by mCIM. However, only 40 isolates were found to produce carbapenemase genes by PCR and six other isolates exhibited other unknown mechanisms of carbapenem-resistance. The high level of production of carbapenemase genes among the isolates is that mostly of pediatric patients treated with carbapenem empirically. Sequencing of NDM-producing isolates revealed that the majority of bla_NDM genes were bla_NDM-1 type as with many studies^32–34 and we found nine bla_NDM-5 type. It was the first to isolate NDM-5 type from Klebsiella pneumoniae isolates in Egypt but recovered previously from Escherichia coli isolates with associated OXA-181.³⁵ There is a significantly higher meropenem MIC of NDM-5-positive isolates than of NDM-1-positive isolates. NDM-5-producing strains (containing the V88L substitution) have been reported repeatedly to exhibit higher MIC against carbapenem than those against NDM-1-producing strains which suggests that this substitution may have a significant impact on carbapenemase activity even though not being located at the active site, and the mechanism of enhanced activity remains uncertain.⁸ NDM-5-positive isolates are continuing to be a critical challenge for treatment and the continuous monitoring of infection control policies is critically important.

Out of 56 isolates, 15 colistin-resistant isolates were positive for mcr-1 gene (26.7%) but mcr-2 gene was not detected at all among the isolated K. pneumoniae as colistin is not used in this hospital to treat infections. Moreover, extended spectrum b-lactamase was shown by the tested isolates as 38 isolates were positive for bla_CTX-M gene. The present study revealed massive co-existence of different resistance genes among the tested isolates which may greatly contribute to the observed raised variability in resistance genotypes among K. pneumoniae in Egypt.^36,37

It was known that efflux pump is one of the mechanisms that is involved in antibiotic resistance in K. pneumoniae,³⁸ and the presence of efflux pump genes (OqxAB) in many Gram-negative pathogens causes an increase in the rates of resistance to many different antibiotic families including beta-lactams, carbapenems, macrolides, fluoroquinolone, tetracyclines that results from their inherent ability to develop resistance.^39,40 So, we need to know if efflux pump plays a role in carbapenem and fluoroquinolone resistance or not by testing the effect of CCCP on meropenem and ciprofloxacin resistance. There is a significant decrease in MIC after CCCP is added to ciprofloxacin which indicates that the drug efflux system is involved in resistance to ciprofloxacin. Our study showed no effect of CCCP on meropenem resistance among isolates.

PCR was used to screen for oqxA and oqxB genes. Both OqxA and OqxB were detected in 26 isolates, 13 OqxA only and 11 OqxB only were also detected. Presence of these efflux pump genes seems to contribute to the high resistance of clinical isolates of K. pneumoniae to other antibiotics. Previous research found that OqxAB was detected in all of K. pneumoniae chromosome isolates and suggested the genome of K. pneumoniae is a possible reservoir of oqxAB.⁴¹

In the current study we could not observe any effect of CCCP on susceptibility to meropenem as no change in MIC of meropenem after CCCP addition was detected (Table 2). However, the addition of the CCCP decreased resistance to ciprofloxacin (4-fold or more in MIC) in NDM-5- producing K. pneumoniae isolates. Meanwhile only St-26 and St-42 isolates had an MIC fold change of 2 upon CCCP addition. We observed that eight of the nine NDM-5-producing K. pneumoniae had 4-fold decrease in MIC after CCCP exposure. Similarly, other studies reported greater reductions in MIC of ciprofloxacin with the use of efflux pumps inhibitors compared with their initial MIC.^42–44 In the present study, the resistance to meropenem could be greatly affected by the presence of beta lactamases, which are not targeted by efflux pump inhibitors⁴⁵ and the use of CCCP may not improve in vitro susceptibility of K. pneumoniae to meropenem.⁴⁶

All NDM-5-producing isolates except one isolate could transfer their plasmids carrying bla_NDM-5 by conjugation to their corresponding transconjugants (E. coli J53). Our results supported what we had previously reported about the vital role of the plasmid in the transfer of resistant genes.⁴⁷ This plasmid may promote the rapid dissemination of bla_NDM-5 among Gram-negative bacterial pathogens. This may indicate the possible transition of this type (bla_NDM-5) between different species by conjugative plasmid and further highlights the importance of carbapenemase genes which could facilitate their dissemination on mobile genetic elements.⁴⁸

Plasmid replicon typing showed that the FII type was predominant in NDM-5-producing K. pneumonia. It has been reported that bla_NDM-5 closely associated with different types of plasmids; among these, the IncF plasmid was found mainly in Enterobacteriaceae and has the ability to transfer its antimicrobial resistance determinants.⁴⁹ On the other hand, genotyping similarity between strains was found between these isolates (St-37, St-29, St-40) and between two isolates (St-42, St26). This indicated the capability of these NDM-5-producing strains to disseminate between different wards of this pediatric hospital.

Conclusions

The highly resistant K. pneumoniae producing bla_NDM-5 type was firstly isolated from pediatric patients. The association of efflux pump genes such as OqxAB is involved in resistance to ciprofloxacin but not to meropenem. This highlights the severity risk of bla_NDM-5-positive K. pneumonia as it could disseminate bla_NDM-5 to other bacterial cells and has more resistance against carbapenems and other antibiotics. This emphasizes the importance of continuous monitoring of infection control policies, and the urgent need for a national antimicrobial stewardship plan in Egypt.