Molecular Epidemiology and Mechanisms of High-Level Resistance to Meropenem and Imipenem in Pseudomonas aeruginosa

Introduction

Pseudomonas aeruginosa is a gram-negative opportunistic organism that causes severe global healthcare-associated infections such as respiratory tract infections, sepsis, urinary tract infections and surgical site infections, especially in immunocompromised patients.¹ The development of multidrug-resistant (MDR) strains that display resistance to nearly all antibiotics except for one or two classes is becoming an uppermost public health problem, increasing morbidity, mortality and length of hospital stay.² Unfortunately, P.aeruginosa harbors a wide array of MDR mechanisms, including the existence of outer-membrane barriers (porin OprD), over-expression of multidrug efflux pumps and endogenous antimicrobial inactivation.³ Accordingly, the choice of an appropriate treatment is a formidable problem in hospitals in many regions worldwide.⁴

Imipenem, meropenem, and doripenem are members of the carbapenems (β lactam class) and are routinely used to manage P. aeruginosa infections.³ The mechanism of action of carbapenems is the inhibition of the peptidoglycan-assembling transpeptidases (penicillin-binding proteins [PBP]) placed on the outer part of the plasma membrane.⁵ Although this class of antibiotics has a great efficacy in treatment of MDR P. aeruginosa infections, an expansion of carbapenem-resistant strains has been detected in the recent years.⁴

Carbapenems resistance in P. aeruginosa could be attributed to a reduction in the outer membrane permeability, up-regulated expression of the efflux pumps genes and production of metallo-β-lactamases (MBL), which inactivate these drugs efficiently.⁶

The outer membrane protein OprD permits the transport of amino acids, peptides and carbapenems. Downregulation of oprD generally causes resistance to imipenem and meropenem.⁷ Another major mechanism of Carbapenem resistance in P. aeruginosa is decreasing the antibiotic concentration through efflux systems belonging to the resistance-nodulation- division (RND) family (MexAB-OprM, MexCD-OprJ, MexEF-OprN, MexXY-OprM, MexJK and MexVW).⁸ Meropenem is affected by high regulation of MexAB-OprM, MexCD-OprJ, and MexXY-OprM, whilst imipenem is not affected.⁹ MexAB-OprM and MexXY-OprM pumps are the principal means of natural resistance against antimicrobial and disinfectant compounds. The MexAB-OprM system confers resistance against a wide range of antibiotics including tetracycline, chloramphenicol quinolones, trimethoprim and most β lactams and MexXY-OprM contributes to the resistance against aminoglycosides, tetracycline, and erythromycin.¹⁰

Carbapenems withstand relatively well hydrolysis by most of the beta lactamases; however, metallo-b-lactamases (MBL) can inactivate them.¹¹ This group of enzymes include imipenemase (IMP), Verona imipenemase (VIM), New Delhi MBL (NDM), Seoul imipenemase (SIM), Sao Paulo MBL (SPM), German imipenemase (GIM), Adelaide imipenemase (AIM), and Dutch imipenemase (DIM).¹² The genes that encode the production of MBLs are typically part of class 1 integron structures and are transmitted by mobile genetic elements.¹³ Other non-metallo carbapenemases also exist, including Ambler class A β-lactamases: K. pneumonia carbapenemase (KPC) and Guiana Extended spectrum (GES) as well as class D β-lactamases: Oxa-48.¹⁴

The aims of this study were to evaluate the frequency of carbapenem resistance among P. aeruginosa isolated from different clinical specimens (urine, sputum, pus, etc.) in Minia governorate, Egypt and to investigate the correlation between the MICs obtained and overexpression of certain efflux pump genes, downregulation of outer membrane porin, OprD and the presence of metallo-b-lactamases (MBL) genes.

Materials and Methods

This study was conducted at the Minia University Hospital (the largest tertiary hospital in Minia-Upper Egypt, receiving referrals from 9 districts and serving a population of more than 5 million) in the period between December 2017 to November 2018.

Bacterial Isolates

A total of 634 P. aeruginosa isolates were collected from different clinical specimens (160 urine, 103 sputum, 227 wound, 116 tracheal aspirates, 88 bloodstream infections, and 56 bronco-alveolar lavage samples) as part of the routine hospital laboratory procedure. Isolates were identified using standard laboratory methods including bacteriological and biochemical tests such as Gram stain, oxidase test, growth on cetrimide agar medium, the ability to grow at 42°C, O/F (Oxidation-Fermentation) test and pigment production. The identified P. aeruginosa isolates were stored at −70°C in trypticase soy broth (Merck, Darmstadt, Germany) supplemented with 10% glycerol until further investigation. P. aeruginosa ATCC 27853 was used as the reference strain throughout the study. A set of 32 carbapenem-sensitive clinical isolates were used as negative controls.

Antimicrobial Susceptibility Testing

The minimum inhibitory concentrations (MICs) of meropenem and imipenem were measured on Muller Hinton Agar plates (Oxoid, Basingstoke, UK) by the agar dilution method and interpreted according to CLSI breakpoint.¹⁵ Meropenem and Imipenem were purchased from Himedia, Mumbai (India). Antimicrobial profile against different anti-pseudomonas drugs was done for the carbapenem-resistant isolates using disc-diffusion method according to CLSI.¹⁵

Detection of Carbapenemase Genes

DNA Extraction and PCR

Total DNA from P. aeruginosa isolates was extracted by boiling method according to Rostami et al.¹⁶ The tubes were stored at −20°C prior to being used in PCR amplification as a DNA template.¹⁰

All carbapenem-resistant P. aeruginosa isolates were screened by PCR for the bla_IMP1, bla_IMP2, bla_VIM1, bla_VIM2, bla_SPM and bla_NDM genes using specific primers and conditions listed in Table 1.

Table 1 Primers Used in This Study

Efflux Pump and Porin Expression

RNA Extraction and Reverse Transcription to cDNA

For total RNA extraction, P. aeruginosa isolates were cultured in 1.5 mL LB broth and placed in a shaking incubator (180 rpm) at 37°C for 18–24 hrs. Tubes were incubated overnight at 37°C and then were centrifuged (8000rpm) at 4°C for 7 mins; afterwards the pellet was processed according to the manufacturer’s instructions using the GeneJET RNA Purification Kit (Thermo-Scientific^TM, USA). To prepare cDNA, 5 µL of extracted RNA was added to15 µL deionized water and converted to cDNA in a total volume of 20µL according to the manufacturer’s instructions using Revert Aid First Strand cDNA synthesis Kit (Thermo-Scientific^TM, USA). The cDNAs were stored at −20°C and were used within 1 week.

Real-Time (RT)-PCR to Measure mexA, mexC, mexE, mexY and oprD Expression Levels

The expression of mexA, mexC, mexE, mexY and oprD in the P. aeruginosa isolates that showed resistance to carbapenems (imipenem and meropenem) was determined by real-time PCR (RT-PCR). Primers that were used for the amplification of mexA, mexC, mexE, mexY, oprD and rpsL are specified in Table 1 and the housekeeping gene rpsL was used as the normalizing gene. Expression levels of the above genes for clinical isolates of P. aeruginosa were compared to P. aeruginosa PAO1.

The quantification of transcripts was carried out by SYBRGreen PCR Master Mix (Thermo-Scientific^TM, USA). Applied Biosystems 7500-Fast Real-Time PCR system (Thermo-Scientific^TM-USA). The relative gene expression was calculated using the ΔΔCt method.²⁶ Expression of the house-keeping gene rpsL was determined in parallel to normalize the transcriptional levels of target genes and they were furthermore standardized against target gene expression by the wild-type reference strain PAO1. Overexpression of the efflux systems MexAB-OprM, MexCD-OprJ, MexEF-OprN and MexXY-OprM was considered when mexA, mexC, mexE or mexY transcriptional levels were at least two-, two-, 10-and 4-fold higher than that of PAO1, respectively.^27,28 All the experiments were repeated three times.

Random Amplified Polymorphic DNA (RAPD) PCR

The isolated P. aeruginosa strains were sent for RAPD-genotyping using primer 272-AGCGGGCCAA as previously designated by Mahenthiralingam and co-workers.²⁹ The master mixture was prepared using 2.5 µL 10x PCR buffer; 2.5mM MgCl₂; 300µM of dNTPs; 1.7U Taq DNA polymerase; and 3 µL genomic DNA (40 ng) in a final volume of 25 µL using the following thermocycler conditions: denaturation 5 mins at 95°C, annealing 5 mins at 36°C, and elongation 5 mins at 72°C, for 4 cycles and 31 cycles consisting of 94°C for 1 min, 45°C for 1min, and 72°C for 2 mins, followed by a final extension at 72°C for 10 mins. For reproducibility, each reaction was performed in triplicate. For analysis of RAPD products, PCR amplicons were separated on 1.% (w/v) agarose gels and the DNA band size was determined in comparison to 1Kb DNA ladders (250 bp–10,000 bp band sizes). Similarity between isolates was assessed according to Dice similarity coefficient and Unweighted Average Pair Group Method (UPGMA), using FreeTree and TreeView software. Only major reproducible bands of intensity were put into consideration for similarity matrix calculation. Cut-off values of ≥80% were used for determination of potential clonal relatedness.³⁰

Statistical Analysis

Data were analyzed using SPSS statistical package version 20 (Chicago, USA). The following had been done accordingly: mean (X~) and standard deviation (SD) for quantitative data, Chi-squared test for categorical data and Bivariate Pearson correlation analysis for association analysis. For all tests, probability (p) was considered: non-significant if ≥ 0.05, significant if < 0.05, Highly significant if < 0.01 and very highly significant if <0.001. Grade of correlation or association were calculated as follows: 0.00 to 0.24: weak or no association, 0.25 to 0.49: fair association, 0.50 to 0.74: moderate association and 0.75+: strong association.

Results

Out of 634 P. aeruginosa isolates, 32 (5%) were imipenem resistant of which 29 were meropenem resistant. High-level resistance (≥64mg/mL) to imipenem was found in 27 isolates (84.3%) out of 32 isolates, while high-level resistance (≥64mg/mL) to meropenem was found in 28 (96.5%) out of the 29 meropenem-resistant isolates. The antimicrobial profile to different antimicrobials used for treating P. aeruginosa infection is illustrated in Table S1.

Carbapenemase genes were found as follows: 31 isolates were positive for bla_VIM-1, while bla_NDM was detected in 4 isolates, which co-harbored bla_VIM-1 also. None of the isolates possessed either bla-_VIM-2, bla_IMP-1, bla_IMP-2 or bla_SPM.

Regarding gene expression, most of the isolates showed over-expression of CDJ (75%) followed by XY efflux pump (62%), while ABM and EFN efflux pumps were overexpressed in 21.8% and 18.7% of the isolates, respectively. None of the isolates showed low expression of oprD (Table 2). There was a significant difference between sensitive and resistant isolates regarding XY and CDJ efflux pump expression, which were overexpressed in 14.3, 22.9% vs 62.1, 79.3% among sensitive and resistant groups, respectively (p <0.001) (Table 3). Only three of the isolates showed over-expression of the four tested efflux pumps. Expression levels of different genes are demonstrated in Table S2.

Table 2 MIC and Characterization of Efflux Pump Overexpression and Carbapenemase Genes in the Carbapenem-Resistant P.aeruginosa Isolates

Table 3 Correlation of Efflux Pump Gene Expression with Carbapenem-Resistant and Sensitive P.aeruginosa Isolates

A highly significant fair association was found between XY efflux pump expression with meropenem, imipenem resistance (r=0.363, 0.327, p=0.003, 0.008) respectively. In addition, there was a very highly significant moderate correlation between CDJ efflux pump expression and meropenem, imipenem resistance (r=0.532, 0.654, p<0.001, <0.001) respectively (Table 4).

Table 4 Bivariate Correlation Analysis Between Meropenem, Imipenem Resistance and Efflux Pump Over-Expression

The carbapenem-resistant P. aeruginosa isolates were typed by RAPD-PCR analysis and Figure 1 illustrates their clustering. RAPD fingerprinting produced 6–10 fragments with band sizes between 250 bp and 3 kb. There were three major groups: group 1(10 isolates), group 3 (9 isolates), group 2 (8 isolates) while the fourth group (4) consisted of 4 isolates. Only one isolate (PA-15) was a unique isolate forming a separate cluster.

Figure 1 Dendrogram based on RAPD analysis of 32 P.aeruginosa isolates. The dendrogram was developed using gelJ software. Abbreviation: PA, Pseudomonas aeruginos.

Discussion

In this study, carbapenem resistance was detected in 5% of P. aeruginosa isolates, with the majority showing high-level resistance to both imipenem and meropenem. Regarding the incidence of carbapenem resistance, our findings are lower than previously reported in the same region,³¹ probably due to an improvement in abiding to infection control measures. Nevertheless, this is the first report of high-level resistance to carbapenems in Egypt, which is disturbingly snowballing in clinical P.aeruginosa isolates world-wide.³² We further investigated the mechanisms of this high-level carbapenem resistance among the isolates and interestingly we found that 31 of the isolates (>96%) harbored at least one carbapenemase gene. This was in agreement with previous reports in Egypt,^33,34 which are all in agreement with finding that MBLs can overtake efflux pump up-regulation or loss of oprD as principal elements leading to carbapenem resistance.³⁵ However, our results show a dominance of blaVIM-1, which is not in agreement with earlier studies carried out in Egypt showing a dissemination of bla-VIM-2 instead or the universal bla-VIM genes.^34,36 Interestingly, we report here for the first time in Upper Egypt the emergence of New Delhi metallo-beta-lactamase −1 gene (bla-NDM) harboring isolates, which possessed bla-VIM-1 as well and showed high-level resistance. There are no previous reports of bla-NDM in Egypt, except a case report of two isolates recovered from a tertiary hospital in 2012.³⁷ The patients who harbored the bla-NDM isolates had no reported contact with a foreign population or a history of travelling to or from India or any foreign country. The presence of this gene could be explained as described earlier by Zafer and co-workers as a result of antimicrobial abuse or unreported contact and cross-contamination with a positive case³⁷ (perhaps a contact with the previously described cases in Cairo could be the source, although not confirmed).

This occurrence of MBL-producing isolates can produce grave infections that are problematic to treat and their existence in different sources in Egypt is of great countrywide importance.

Different efflux pumps are described in P. aeruginosa, among which MexAB-OprM, MexCD-OprJ, MexEF-OprN and MexXY-OprM are the best-characterized systems that are correlated with antibiotic resistance in clinical P. aeruginosa isolates.³⁸ In our study, the majority of the isolates demonstrated over-expression of MexCD-OprJ and MexXY-OprM efflux pumps. A strong correlation was found between MexCD-OprJ over-expression and both meropenem and imipenem resistance, concurring with Bubonja-Sonje and co-workers.³⁹

Although the MexXY-OprM system is known as a substantial element in aminoglycoside resistance in only P. aeruginosa,⁴⁰ it is also over-expressed in multi-drug resistant P. aeruginosa⁴¹ and similar reports to our study have shown its over-expression in carbapenem-resistant P. aeruginosa strains.^7,42,43

MexAB-OprM over-expression was observed in 21.8% of the isolates, concurring with Xavier and co-workers.⁴⁴ In contrast, Dantas et al⁷ and Chalhoub et al³² reported a much higher frequency of MexAB-OprM over-expression in P. aeruginosa isolates. Surprisingly, no down-regulation of oprD was detected in this study, which could be claimed to the presence of mutations eg, in loops L2 and L3 as described by Muderris et al.⁴⁵ However, Chalhoub et al found no role for oprD mutation in high-level resistance compared to efflux pump overexpression.³²

Finding bacterial genetic similarity and relatedness is crucial for cross-infection assessment, for which the value of various genotyping techniques has been recognized.^46,47 Of those techniques, RAPD-PCR is advantageous in being fast, simple, and reproducible in addition to having a high discriminatory power.⁴⁷

In the present study, we detected 4 distinct RAPD clusters among 32 isolates (12.5% of polymorphisms), which unfortunately suggests possible hospital cross-contamination.

Conclusion

Our results demonstrate that high-level carbapenems resistance reported in this region is attributable to carbapenemase production as well as to concomitant efflux-pump over-production. A high probability of cross-acquisition is concluded from phylogenetic results suggesting an infection control problem, which necessitates more strict procedures.

Limitations: Studying the mutation of oprD was not done in this study due to lack of funding. There was no positive control for the over-producers.