Summary

Introduction: Increasing ceftazidime-avibactam (CZA) resistance in P. aeruginosa isolates is of serious concern worldwide. Therefore, studies on the underlying resistance mechanisms gain even more importance. The aim of this study was to investigate the presence of beta-lactamase genes reported to be associated with resistance, including metallo-beta-lactamases (MBL), in CZA-resistant P. aeruginosa strains and the clonal relationship between isolates.
Materials and Methods: Ceftazidime-avibactam resistant P. aeruginosa strains isolated from various clinical specimens between December 2021 and March 2023 were included in this study. MALDI-TOF (VITEK-MS, bioMérieux, France) was used for the identification of strains. Ceftazidime-avibactam susceptibility was determined by disc diffusion method and resistant strains were also studied by gradient diffusion method (MIC strip, Liofilchem, Italy). Antibiotic susceptibility results were evaluated according to the European Committee for Antibiotic Susceptibility Testing criteria. blaKPC, blaNDM, blaIMP, blaVIM, blaOXA-48, blaOXA-2, blaOXA-10, blaGES, blaPER, blaVEB genes were determined using polimerase chain reaction, and clonal relationship was investigated by pulsed field gel electrophoresis.
Results: A total of 38 CZA resistant P. aeruginosa strains were analyzed. The MIC50 and MIC90 values determined by the gradient diffusion method for CZA were ≥256 mg/l and ≥256 mg/l, respectively; the MIC range was found to be 32-256 mg/l. Among of 38 isolates, 11 (28.9%) isolates including 10 blaVIM (26.3%) and one blaIMP (2.6%) were positive for the MBL genes. The blaOXA-10 gene was found to be co-positive in seven of the blaVIM positive isolates and in one blaIMP positive isolate. In addition, blaOXA-10 alone was found in nine (23.6%) of the isolates, blaPER in two (5.2%) and blaGES in one (2.6%). No genes were found in 15 (39.4%) isolates. A total of 15 different genotypes consisting of eight different clusters were identified.
Conclusion: It was determined that MBL production, especially blaVIM, was the most common cause of CZA resistance. blaOXA-10, blaPER and blaGES positivity were also found remarkable. Regarding molecular epidemiology, it was observed that the strains isolated predominantly from ICU patients had different genotypes and exhibited a polyclonal diversity.

Introduction

P. aeruginosa, a gram-negative non-fermentative bacterium, is an important pathogen causing healthcare-associated infections such as bacteremia, pneumonia, and urinary tract infections in the intensive care unit (ICU)^[1]. The fact that P. aeruginosa has an extraordinary capacity to acquire antimicrobial resistance and the rapid spread of its high-risk clones worldwide has placed it at the top of the list of important pathogens for which new antimicrobials should be developed, as reported by the World Health Organization^[2-4]. Ceftazidime-avibactam (CZA), which was developed as an alternative and is also used in our country due to the increasing carbapenem resistance seen in P. aeruginosa isolates worldwide, consists of the combination of ceftazidime, a broad-spectrum cephalosporin, and avibactam, a non-beta-lactam beta-lactamase inhibitor. Since ceftolozane-tazobactam, which is reported to be more effective against AmpC-type beta-lactamases frequently seen in P. aeruginosa isolates, is not available in our country, CZA can be used as an alternative option if it is found susceptible by in vitro methods. When examined in terms of carbapenemases, while CZA is effective against class A (KPC) and class D (OXA-48) carbapenemases, it is not effective against class B metallo-beta-lactamases (MBL) (NDM, VIM, IMP) due to the absence of serine residues in the active site^{[5, 6]}. The in vitro activity initially observed for CZA around the world was soon replaced by rapidly progressing resistance reports^[7-9]. This situation, which causes serious concern, has accelerated research on CZA resistance mechanisms^{[1, 3, 5, 10, 11]}. These studies reported that, in addition to MBL production, P. aeruginosa can develop resistance to CZA in different ways. These are stated as enzymatic resistance causing inactivation of antibiotics, chemical modification of the antibiotic target or expression of an alternative target, and changes in cell permeability or expression of efflux pumps^{[5, 6]}.

In addition to determining the underlying resistance mechanisms in CZA resistant isolates, it is also important to closely monitor the clonal relationship between isolates with molecular epidemiological studies in order to determine the spread within the hospital and take control measures^{[1, 3]}.

In this study, several beta-lactamase genes reported to be associated with resistance (blaKPC, blaOXA-48, blaOXA-10, blaOXA-2, blaPER, blaGES, blaVEB), including MBLs (blaVIM, blaIMP, blaNDM), in CZA resistant P. aeruginosa isolates and whether the isolates were clonally related to each other were investigated. This study, which is the first to our knowledge on the molecular epidemiology of CZA-resistant P. aeruginosa in Turkey, aims to contribute to both local and global literature.

Methods

Bacterial Isolates

CZA resistant P. aeruginosa strains isolated from various clinical samples in the microbiology laboratory of our hospital between December 2021 and March 2023 were included in this study.

Bacterial Identification and Antibiotic Susceptibility Tests

Matrix-mediated laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF) (VITEK-MS, bioMérieux, France) was used to identify the strains, and VITEK 2 automated system (bioMérieux, France) was used to determine antibiotic susceptibility. Ceftazidime-avibactam susceptibility was determined by the standard disk diffusion method using 14 microgram disks (Bioanalyse, Turkey). Isolates with an inhibition zone diameter of ≥17 mm were considered susceptible, and those <17 mm were considered resistant. Strains found to be resistant to CZA were also studied with the gradient diffusion method (MIC strip, Liofilchem, Italy). Those with minimal inhibitory concentration (MIC) values ≤8 mg/l were considered susceptible, and those >8 mg/l were considered resistant. Colistin susceptibility was studied with the commercial broth microdilution kit MIC-COL (Diagnostics I.n.c., Galanta, Slovakia). Antibiotic susceptibilities of the isolates were determined according to the  European Committee on Antibiotic Susceptibility Testing (EUCAST) criteria^[12].

Determination of Resistance Genes by Polymerase Chain Reaction (PCR) Method

DNA isolation  was performed using the QIAamp DNA midi kit (Qiagen, Hilden, Germany) on the QIAsymphony automated DNA extraction system (Qiagen, Hilden, Germany). The DNA extracts were kept at -80 °C until use. blaKPC, blaNDM, blaIMP, blaVIM, and blaOXA-48 genes were examined by multiplex PCR method as previously described by Poirel et al.^[13]. The presence of blaOXA-2, blaOXA-10, blaGES, blaPER, and blaVEB genes was determined by multiplex PCR method as described by Kiratisin et al.^[14]. DNA amplification was carried out by a thermocycler, GeneAmp PCR System 9700 (Applied Biosystems, Waltham, MA, USA).

The amplicons were electrophoresed for 1 h at 100V in 1.5% agarose gel and stained by ethidium bromide before their UV images were taken using a Kodak Gel Logic 200 (1708x1280, Kodak Company, Rochester, USA).

Molecular Epidemiological Analysis

Isolates were stored at −80°C refrigerator until genotyping. Pulsed field gel electrophoresis (PFGE) method was performed as previously described by Selim et al.^[15], with minor modifications. Agarose plugs prepared for each of the 38 P. aeruginosa isolates were incubated with 30 units of SpeI restriction enzyme. Band profiles were analyzed using the GelCompar II software system (version 6.5; Applied Maths, Sint-Martens-Latem, Belgium). Dice correlation coefficient was used to make similarity calculations for band analysis and UPGMA method was used for cluster analysis (optimization 1.0, tolerance 1.0). Isolates with band similarities over 85% were considered to be the same clone.

Statistical Analysis

Antibiotic susceptibility results and distribution of resistance genes are given as numbers and percentages.

Results

A total of 38 CZA-resistant P. aeruginosa strains were isolated during the study period. Isolates were obtained mostly from endotracheal aspirate cultures (n=22, 57.9%), followed by blood (n=6, 15.8%), tissue biopsy (n=6, 15.8%) and urine cultures (n=4, 10.5%), respectively. All but one of the strains were detected in clinical samples of patients hospitalized in the ICU.

The inhibition zone diameters obtained by the disk diffusion method in the isolates ranged between 0-13 mm, and no inhibition zone was formed in 84.2% (n=32) (Table 1). MIC50 and MIC90 values determined by the gradient diffusion method for CZA were found to be ≥256 mg/l and ≥256 mg/l, respectively, and the MIC range was 32-256 mg/l. It was determined that the gradient method and disc diffusion methods were 100% compatible with each other.

Table 1: Distribution of resistance genes and antimicrobial activity in ceftazidime-avibactam-resistant P. aeruginosa clinical isolates

Of 38 CZA resistant P. aeruginosa, 92.1% (35/38) were resistant to amikacin, 76.3% (29/38) to ciprofloxacin, 73.6% (28/38) to levofloxacin, 100% (38/38) to imipenem and meropenem. All isolates were susceptible to colistin, and MIC values were found to be between 0.5-2 mg/l.

Resistance Genes

Among a total of 38 isolates, 11 (28.9%) isolates were positive for MBL genes, 10 of which were blaVIM (26.3%) and one was blaIMP (2.6%). The blaOXA-10 gene was found to be positive together in seven of the blaVIM positive isolates and 1 blaIMP positive isolate. Additionally, blaOXA-10 alone was detected in nine of the isolates (23.6%), blaPER in two (5.2%) and blaGES in one (2.6%). None of the other blaNDM, blaKPC, blaOXA-48 and blaVEB genes, which were the subject of the study, were found in 15 (39.4%) isolates (Table 1).

Molecular Epidemiology and Clonality

Fifteen different genotypes were detected among a total of 38 P. aeruginosa isolates, and isolates showing clustering were collected in eight different clusters (tolerance 1.0, optimization 1.0, cutoff 85%). Thirty-one of the thirty-eight P. aeruginosa isolates were located in any cluster, and the clustering rate was found to be 81.5%. The largest cluster was genotype 1 cluster with nine isolates, and the second largest clusters were genotype 4 and 5 clusters with five isolates each. Other clusters were genotypes 2 and 8, with three isolates each, and genotypes 9, 10 and 12, with two isolates each (Figure 1).

Figure 1: PFGE dendogram of ceftazidime-avibactam resistant P. aeruginosa isolates
PFGE: Pulsed field gel electrophoresis

When examined in terms of molecular epidemiological data, it was seen that 37 of 38 isolates were obtained from patients in the ICU and one was obtained from a patient in the general surgery ward. Although the isolate of the patient hospitalized in general surgery ward and the isolates of two patients hospitalized in ICU had similar genotypes (genotype 8), it was determined that the patient hospitalized in general surgery ward did not receive treatment in the ICU during his hospitalization. When the isolation dates of the strains belonging to the remaining 37 patients hospitalized in ICU were compared, no clusters that would suggest a cross-contamination were observed.

Discussion

Shortly after CZA became available for use in carbapenem-resistant P. aeruginosa infections, the reporting of resistant isolates worldwide emerged as a clinical challenge and increased concerns about the global spread of this microorganism^{[1, 8]}. Therefore, research on understanding the mechanisms that cause CZA resistance has become even more important. In these studies on the molecular causes of resistance, it has been reported that the presence of MBL, which can be carried between microorganisms via plasmids, has a significant rate^{[5, 9, 16]}. In the large-scale INFORM surveillance study, P. aeruginosa isolates collected from 42 medical centers in nine countries in the Asia-Pacific region were evaluated and 7.4% (151/2038) were found to be resistant to CZA, and MBL-positive strains accounted for 48.3% (73/151) of resistant isolates^[16]. When previous studies were examined, it was observed that the type and frequency of MBL showed regional variations, and VIM enzyme was generally dominant, but in some countries IMP ranked first^{[3, 9, 7, 17]}. In the study by Mojica et al.^[3], which included data from five Latin American countries, MBL positivity was found at a rate of 38.5% (n=42) in 109 CZA resistant P. aeruginosa isolates. Among the isolates, VIM (28.4%) was the most common, followed by IMP+KPC (6.4%), IMP (2.7%) and SPM (0.9%) positivity. In China, Hu et al.^[9] detected MBL in 42.5% of 80 CZA resistant P. aeruginosa isolates, 16.2% of which were VIM and 26.3% were IMP. Weber et al.^[7] found 36.3% of the isolates to be VIM positive and did not detect IMP or NDM. Schaumburg et al.^[17] found MBL (mostly IMP) positivity in 49.3% and they stated that MBL positivity was the most likely cause of CZA resistance in P. aeruginosa. Similarly, Sid Ahmed et al.^[11] detected the presence of MBL in 64.9% of 37 CZA resistant P. aeruginosa isolates. However, unlike most studies mentioned above, Castenheira et al.^[18] found MBL positivity in CZA resistant P. aeruginosa isolates to be only 4.3%, while Babouee Flury et al.^[10] in Switzerland did not detect MBL in any isolate. Although there are publications reporting that VIM and IMP constitute the majority of MBL detected in carbapenem-resistant P. aeruginosa isolates in our country^{[19, 20]}, there is very little information about CZA-resistant P. aeruginosa isolates. In the study by Hoşbul et al.^[21], VIM enzyme was found in only one of 10 CZA resistant P. aeruginosa isolates, while NDM and IMP were not detected. We detected the MBL positivity rate as 28.9% including IMP, mostly VIM in our isolates, which was consistent with the general literature.

In recent studies, it has been stated that OXA-10 and OXA-2 enzymes may be associated with CZA resistance in P. aeruginosa isolates^{[11, 22, 23]}. Arca-Suárez et al.^[23] determined that CZA resistance, which developed in vivo during treatment in a patient infected with extensively drug-resistant (XDR) P. aeruginosa, was caused by a modification in the OXA-10 enzyme. Sid Ahmed et al.^[11] also detected OXA-10 in 15% of CZA resistant P. aeruginosa isolates and stated that the presence of OXA-10 was associated with the development of resistance. In our study, in addition to OXA-10 positivity alone in 26.3% of the isolates (n=9), we also detected OXA-10 enzyme in seven of the VIM-positive isolates and one IMP-positive isolate. Accordingly, considering the previous data, it is thought that the OXA-10 enzyme may cause CZA resistance, especially in isolates where it is found alone. Fraile-Ribot et al.^[22] found that duplication of residue D149 in OXA-2 (OXA-539) led to resistance to CZA in vivo. Mojica et al.^[3] detected OXA-2 positivity in 17 of 109 resistant isolates, but did not find any duplication in the whole genome sequence analysis. We did not observe OXA-2 positivity in any of our isolates. Apart from all these, similar to some other studies, OXA-48 and KPC enzymes were not found in any of our isolates^{[17, 18]}.

It has been stated that PER and GES enzymes, which are extended-spectrum beta-lactamases that are common worldwide, may also be the source of CZA resistance in P. aeruginosa isolates^{[3, 24, 25]}. Ortiz de la Rosa et al.^[24] demonstrated that blaPER-1 alone can confer resistance to CZA when expressed at high level in the recombinant plasmid in PAO1. Babouee Flury et al.^[10] also detected PER positivity accompanied by other resistance mechanisms in a small number of P. aeruginosa isolates. Mojica et al.^[3] found the PER positivity to be 2.8%. In our study, we found PER positivity in two isolates (5.2%). In a recent study by Li et al.^[25], it was stated that overexpression of the blaGES-1 gene carried in the class 1 integron of the Tn6584 complex transposon contributed to CZA resistance, and furthermore, such overexpression of blaGES-1 combined with an efflux pump caused high CZA resistance. Similarly, Recio et al.^[26] reported that they isolated a large number of CZA resistant P. aeruginosa isolates carrying only GES enzymes in their hospitals in a short time. Weber et al. detected GES positivity in 0.9% of CZA resistant isolates^[7], Hu et al.^[9] in 1.25%, and Mojica et al.^[3] in 4.6%. In our study, we detected GES enzyme in 2.6% of our isolates, consistent with these rates.

Metallo-beta-lactamases, which are one of the causes of CZA resistance, are important because they can be easily transported between bacteria through mobile genetic elements, thus causing cross-contamination within the hospital. Therefore, it is necessary to be careful about large-scale spread and possible epidemics that may occur with multidrug-resistant bacteria in units where broad-spectrum antibiotic pressure is intense, such as ICUs^[4]. The fact that all but one of the CZA resistant P. aeruginosa that we isolated consecutively during the study period were obtained from patients in the ICU was a warning for us in terms of a possible cross-contamination and epidemic. Thereupon, in the PFGE analysis we conducted to determine whether the strains were clonally related to each other, we detected 15 different genotypes among 38 P. aeruginosa isolates. We found that the isolates showing clustering among them were collected in eight different clusters. The largest cluster was genotype 1 cluster with nine isolates, and the second largest clusters were genotype 4 and 5 clusters with five isolates each. Other clusters were genotypes 2 and 8, with three isolates each, and genotypes 9, 10 and 12, with two isolates each. This result we obtained suggested that there was a polyclonal diversity among our isolates.

When the isolates were examined in terms of epidemiological relationship, it was seen that the isolate of the patient in general surgery ward and the isolates of two patients in ICU had similar genotypes (genotype 8) and were isolated on the same date. However, the fact that the general surgery patient did not be treated in the ICU during his hospitalization suggested that contamination might occur through healthcare workers. It was determined that strains belonging to genotype 1, which included the most isolates, were first seen in 2 patients in January 2022, and re-emerged 6 months later and were isolated in 7 more patients. When the isolation dates of other strains were examined, it was observed that there was no chronological clustering in isolates with the same genotype that would suggest a cross contamination. However, it was found remarkable that certain genotypes continued to exist, albeit intermittently, throughout the study period of approximately 1.5 years. Apart from this, we should also point out that the enzymes responsible for resistance differ in some isolates with the same genotype.

Study Limitations

Our study had some limitations. First, other potential resistance mechanisms that might cause CZA resistance, such as efflux pumps, loss of porins, overexpression of blaPDC, decreased permeability, and overproduction of AmpC, were not investigated. Secondly, in molecular epidemiological studies conducted with CZA resistant P. aeruginosa isolates, it was observed that the multilocus sequence typing (MLST) method was used, which allowed comparison with international clones^{[7, 11]}. Due to technical limitations, MLST could not be studied and a literature comparison could not be made because there were differences in methods in the evaluation of clonal relationship.

Conclusion

As a result, as stated in previous studies, CZA resistance mechanisms observed in P. aeruginosa isolates are quite diverse and complex^{[3, 10]}. In our study, MBL production, especially VIM, was found to be the most common cause of CZA resistance. Apart from this, OXA-10, PER and GES positivity were also found to be remarkable. In the molecular epidemiological examination, it was observed that the strains isolated predominantly from ICU patients had different genotypes and exhibited polyclonal diversity. Close monitoring of resistance to CZA, one of the alternative antibiotics, in multidrug-resistant P. aeruginosa infections is important for effective treatment strategies. It is thought that our data will make a significant contribution to the national and global literature. However, it is obvious that more comprehensive studies are needed to determine the mechanisms responsible for CZA resistance in P. aeruginosa.

Acknowledgments

We would like to thank Prof. Dr. Barış OTLU from İnönü University Faculty of Medicine, Department of Medical Microbiology, and Prof. Dr. Sebahat AKSARAY from University of Health Sciences Turkey, Hamidiye Faculty of Medicine, Department of Medical Microbiology, for their support in the design and analysis phase of our study.

Ethics

Ethics Committee Approval: The study was approved by the Haydarpaşa Numune Training and Research Hospital Clinical Research Ethics Committee on 22.05.2023 with the decision number HNEAH-KAEK 2023/KK/96.

Informed Consent: Informed consent was not obtained since the study was conducted with bacterial isolates.

Peer-review: Externally peer-reviewed.

Authorship Contributions

Surgical and Medical Practices: N.A., E.S.T., Concept: N.A., E.S.T., Design: N.A., E.S.T., Data Collection or Processing: N.A., E.S.T., Analysis or Interpretation: N.A., E.S.T., Literature Search: N.A., Writing: N.A., E.S.T.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.