Association of blaVIM-2, blaPDC-35, blaOXA-10, blaOXA-488 and blaVEB-9 β-Lactamase Genes with Resistance to Ceftazidime–Avibactam and Ceftolozane–Tazobactam in Multidrug-Resistant Pseudomonas aeruginosa

Ceftazidime–avibactam and ceftolozane–tazobactam are approved for the treatment of complicated Gram-negative bacterial infections including multidrug-resistant (MDR) Pseudomonas aeruginosa. Resistance to both agents has been reported, but the underlying mechanisms have not been fully explored. This study aimed to correlate β-lactamases with phenotypic resistance to ceftazidime–avibactam and/or ceftolozane–tazobactam in MDR-P. aeruginosa from Qatar. A total of 525 MDR-P. aeruginosa isolates were collected from clinical specimens between 2014 and 2017. Identification and antimicrobial susceptibility were performed by the BD PhoenixTM system and gradient MIC test strips. Of the 75 sequenced MDR isolates, 35 (47%) were considered as having difficult-to-treat resistance, and 42 were resistant to ceftazidime–avibactam (37, 49.3%), and/or ceftolozane–tazobactam (40, 53.3%). They belonged to 12 sequence types, with ST235 being predominant (38%). Most isolates (97.6%) carried one or more β-lactamase genes, with blaOXA-488 (19%) and blaVEB-9 (45.2%) being predominant. A strong association was detected between class B β-lactamase genes and both ceftazidime–avibactam and ceftolozane–tazobactam resistance, while class A genes were associated with ceftolozane–tazobactam resistance. Co-resistance to ceftazidime–avibactam and ceftolozane–tazobactam correlated with the presence of blaVEB-9, blaPDC-35, blaVIM-2, blaOXA-10 and blaOXA-488. MDR-P. aeruginosa isolates resistant to both combination drugs were associated with class B β-lactamases (blaVIM-2) and class D β-lactamases (blaOXA-10), while ceftolozane–tazobactam resistance was associated with class A (blaVEB-9), class C (blaVPDC-35), and class D β-lactamases (blaOXA-488).


Introduction
Gram-negative bacteria (GNB) represent a major healthcare burden due to their association with a variety of community and healthcare-associated infections (HAIs) [1,2]. Multidrug-resistant (MDR) Pseudomonas aeruginosa is a challenging cause of HAIs given the limited effective treatment options, increased morbidity and mortality, as well as the cost of medical care [3,4]. Although pathogens have multiple mechanisms of resistance, it has been established that β-lactamase genes are the cornerstone of antimicrobial resistance (AMR), particularly in GNB [5]. To overcome these challenges, parallel critical measures are needed, including the prevention of AMR spread, while simultaneously developing new therapeutic modalities [3,6].
Ceftazidime-avibactam and ceftolozane-tazobactam have been approved by the United States Food and Drug Administration (US-FDA) and the European Medicines Agency (EMA) for the treatment of complicated infections caused by GNB, including ventilation-associated pneumonia, urinary tract and intra-abdominal infections [7]. The broader activity of ceftazidime-avibactam has been attributed to the addition of avibactam, a non-β-lactam β-lactamase inhibitor capable of inhibiting class A, class C, and most class D β-lactamases [8], whereas ceftolozane is a novel cephalosporin demonstrating favorable activity against P. aeruginosa isolates with AmpC hyper-production and overexpressed efflux mechanisms [9]. The addition of tazobactam to ceftolozane extended its activity against many, but not all, extended-spectrum β-lactamase (ESBL)-producing GNB [10]. Despite the initial remarkable success, ceftazidime-avibactam and ceftolozanetazobactam resistance are being increasingly reported in MDR-GNB including P. aeruginosa [11,12]. A previous in vitro study from Qatar evaluated the efficacy of ceftazidimeavibactam and ceftolozane-tazobactam against MDR-P. aeruginosa isolates and reported that AMR for both agents was higher in Qatar compared to other regions worldwide [13]. The underlying mechanisms of resistance to these agents have not been fully explored. The present study aimed to characterize the β-lactamases and identify their potential correlations to phenotypic resistance of ceftazidime-avibactam and/or ceftolozane-tazobactam in MDR-P. aeruginosa from Qatar.

Methods
Ethical approval for the study (protocol number IRGC-01-51-033) was obtained from the Institutional Review Board at the Medical Research Council, HMC, Qatar, which complies with international ethical standards. Bacterial samples were collected prospectively between 2014 and 2017 from various clinical specimens received at the Microbiology Division of the Department of Laboratory Medicine and Pathology (DLMP), Hamad Medical Corporation (HMC), Qatar, as part of the routine care. MDR-P. aeruginosa was identified and subsequently stored at −80 • C for further molecular analysis. MDR-P. aeruginosa was defined as resistant to at least one agent from three or more different antimicrobial classes [14], whereas difficult-to-treat resistance (DTR) was defined as non-susceptibility to all firstline agents, including β-lactams, carbapenem, monobactam and fluoroquinolones [15]. A total of 525 MDR-P. aeruginosa were tested with ceftazidime-avibactam and ceftolozanetazobactam, of which 75 isolates were randomly selected and subsequently processed for whole-genome sequencing (WGS), including 42 that were resistant to ceftazidimeavibactam and/or ceftolozane-tazobactam.

Bacterial Identification and Antimicrobial Susceptibility Testing
Bacterial identification and initial antimicrobial susceptibility tests (AST) of P. aeruginosa species were performed using the BD Phoenix TM automated system, and identification was confirmed using matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF MS, Bruker Daltonics MALDI Biotyper, Billerica, MA, USA), according to the manufacturer's instructions. The minimum inhibitory concentrations (MICs) were determined using Liofilchem ® MIC Test Strips (Liofilchem, Rosetodegli Roseto Degli Abruzzi, Italy), and the results were interpreted using the Clinical and Labo- ratory Standards Institute (CLSI) reference breakpoints [16]. The standard reference strains, Escherichia coli ATCC 25922, E. coli ATCC 35218, and P. aeruginosa ATCC 27853, were used for quality control.

Genomic and Phylogenetic Analyses
Seventy-five MDR-P. aeruginosa isolates were sent to Eurofins Genomics (GATC Biotech GmbH, Konstanz, Germany) for WGS using the Illumina HiSeq 2000 system (Illumina, San Diego, CA, USA). Following quality control assessment, trimmed reads were assembled using SPAdes, Version 3.13.0. [17]. Multilocus sequence typing (MLST) based on the seven housekeeping genes of P. aeruginosa isolates was performed on the MLST server 1.8 provided by The Center for Genomic Epidemiology [18]. The Comprehensive Antibiotic Resistance Database (CARD), Version 1.2.0 (McMaster University, Hamilton, Ontario) was used to annotate the antibiotic resistance genes (ARGs) in the assembled genomes [19].

Statistical Analysis
Descriptive and inferential statistics were used to characterize the study samples and test hypotheses. Susceptibility patterns of MDR-P. aeruginosa to the tested antibiotics were presented as frequency (percentages). The association between β-lactamase classes (e.g., class A, B, C, and D) and susceptibility patterns were analyzed using Pearson Chisquare and Fisher Exact test as appropriate. Cohen's Kappa (k) was used to measure agreement between ceftazidime-avibactam and ceftolozane-tazobactam susceptibility results. MIC values of ceftazidime-avibactam against the sequence types (STs) of 75 MDR-P. aeruginosa were plotted using box-plot, and the median MIC values were compared by using the non-parametric Kruskal-Wallis test. The potential association between genes and antibiotic resistance to ceftazidime-avibactam and ceftolozane-tazobactam were visualized using a correlation matrix based on presence-absence data and was constructed using pairwise Spearman's correlation between β-lactamase genes and resistance phenotype of all the P. aeruginosa isolates included in the study (both susceptible and non-susceptible to ceftazidime-avibactam and ceftolozane-tazobactam). Spearman's correlation coefficient (p) cut-off of ≥0.4 was considered as a statistically reliable association, while a network of associations between β-lactamase genes and resistance to ceftazidime-avibactam and ceftolozane-tazobactam was constructed in Gephi using Fruchterman Reingold layout (http://gephi.org) (accessed on 15 November 2021). The co-occurrence network was further categorized into sub-networks based on the modularity index (http://gephi.org) (accessed on 15 November 2021). A p value < 0.05 (two-tailed) was considered statistically significant. All statistical analyses were conducted using IBM Statistical Package for Social Sciences (SPSS) for Windows, version 25.0 (IBM Corp, Armonk, NY, USA).

Correlation of Specific β-Lactamase Genes to Ceftazidime-Avibactam and Ceftolozane-Tazobactam Resistance
The co-occurrence network of all the β-lactamase genes and the phenotypic resistance to ceftazidime-avibactam and ceftolozane-tazobactam in 75 MDR P. aeruginosa is shown in Figure 3. Spearman's correlation showed that resistance to ceftazidime-avibactam was associated with the presence of bla VIM-2 (0.53) or bla PDC-35 (0.51) in the isolates, while resistance to ceftolozane-tazobactam correlated with the presence of bla VEB-9 (0.53), bla PDC-35 (0.49), and bla OXA-10 (0.45). Co-occurrence of other β-lactamase genes did not correlate to either ceftazidime-avibactam or ceftolozane-tazobactam resistance. Results are expressed as number (percentage), † p-value was calculated using the Fisher Exact test to examine the association between the β-lactamase gene and susceptibility patterns (susceptible, resistant, extremely resistant) to ceftazidime-avibactam and ceftolozane-tazobactam.

Discussion
The present study reported 42 MDR-P. aeruginosa isolates that were phenotypically resistant to ceftazidime-avibactam and/or ceftolozane-tazobactam, and many were also considered as DTR. They belonged to 12 different STs, with a high frequency of ST235, ST357, ST233, and ST308. These notorious epidemic clones exhibiting high MIC values for ceftazidime-avibactam and ceftolozane-tazobactam are responsible for spreading resistance globally, including in the Middle East region [20][21][22]. Nearly all resistant isolates harbored at least one gene of class C and class D β-lactamases, while nearly half of the resistant isolates had class A and class B β-lactamases ( Table 1). The results are comparable to other studies reported in the region [23,24]. Furthermore, our results demonstrate a significant association between the presence of a class B β-lactamase and resistance to both ceftazidime-avibactam and ceftolozane-tazobactam, while class A β-lactamase was significantly associated with resistance only to ceftolozane-tazobactam (Table 2). Similarly, it has been established that the production of class B enzymes is linked with resistance to both combinations, [25] while different class A ESBLs in P. aeruginosa are linked to the observed resistance to ceftolozane-tazobactam [26]. In addition, bla VIM-2 is a metallo-βlactamase (MBL) that utilizes Zn 2+ as a nucleophile in the active site for the hydrolysis of β-lactams [25]. β-lactamase inhibitors such as tazobactam and avibactam can inhibit class A, C, and some of class D (serine β-lactamases), but not MBL, such as bla VIM-2 [25]. This corresponds well with our results, where bla VIM-2 was detected in half of the isolates that were found to be resistant to ceftazidime-avibactam and ceftolozane-tazobactam, and bla VIM-2 and were mainly associated with ST235 and ST233.
Class A β-lactamase bla VEB-9 which was mainly associated with ST235 and ST357, was predominant in 45.2% of the resistant isolates and significantly associated with high resistance (MIC 256 ) to ceftolozane-tazobactam and, to a lesser degree, ceftazidime-avibactam (Table 3). Although bla VEB-9 has been described by different regions worldwide, [27][28][29] its role in ceftolozane-tazobactam resistance warrants further exploration. It is worth highlighting that class C β-lactamase genes were detected in nearly all the resistant isolates with a predominance of ESBL, bla PDC-3 , and bla PDC-35 . These extended-spectrum cephalosporinases have been previously shown to be associated with ceftolozane-tazobactam resistance [30,31].
In our collection, bla PDC-35 , with 99.75% resemblance to bla PDC-2 , has only one amino acid substitution, glycine to alanine at position 391 (G391A) (NCBI Reference Sequence: NG_049907.1). Intriguingly, bla PDC-35 was exclusively detected in ST235 and was significantly associated with high-level resistance to ceftolozane-tazobactam (MIC 256 ) and, to a lesser degree, ceftazidime-avibactam (Table 3). A similar association between bla PDC-35 and resistance to ceftolozane-tazobactam and ceftazidime-avibactam was recently reported [32]. Moreover, a previous study identified different mutations in class C AmpC, glycine to aspartate substitution in position 183 (G183D) in class C AmpC associated with a low-level of resistance to ceftolozane-tazobactam [33].
The highly prevalent Class D β-lactamase genes included bla OXA-488 (GenBank: TKV86805.1) and bla OXA-486 (GenBank: QBY97442.1), which are variants of the intrinsic oxacillinase bla OXA-50 , had only two-point amino acid substitutions at position T16A and Q25R and R49C and D109E, respectively, and this necessitated their reclassification. Furthermore, bla OXA-10 and bla OXA-17 , which is a variant of bla OXA-10 with substitutions of asparagine by serine at position 77, [34] were found to be associated with extreme resistance to ceftolozane-tazobactam in the present study. Conversely, bla OXA-4 was closely linked to ST233 and has only been detected in isolates resistant to both combinations, while bla OXA-10 , which is mainly linked to ST235 and ST357, was almost universally resistant to ceftolozane-tazobactam and highly resistant to ceftazidime-avibactam. A recent study reported the association between bla OXA-10 derivatives and resistance to both combination drugs [35]. Furthermore, bla OXA-488 was linked to ST235 and ST1284 and has a noticeable resistance to both combinations. Despite the emergence of many new variants of the OXA-type β-lactamases, few studies have been conducted to evaluate their possible roles such as the bla OXA-10 , [12], and bla OXA-50 family (i.e., bla OXA-486, bla OXA-488 ) [31,36] detected in the present study. On the other hand, the mutant bla OXA-4 and the selection for an extended-spectrum bla OXA-2 derivative (bla OXA-539 ) results in ceftazidime-avibactam and ceftolozane-tazobactam resistance [12,37].
To the best of our knowledge, this is one of the largest collections from the Middle East, examining 75 MDR-P. aeruginosa by genomic characterization of resistance to ceftazidime-avibactam and ceftolozane-tazobactam. We acknowledge that a combination of multiple resistance mechanisms encompassing β-lactamases, modification of outer membrane proteins, and upregulation of efflux pumps, which was not examined here, may also contribute to the high resistance to ceftazidime-avibactam and ceftolozane-tazobactam in some of the P. aeruginosa isolates [21,38]. A possible study limitation may be the use of a gradient test Liofilchem MIC test strips as a standard susceptibility testing method for ceftazidime-avibactam and ceftolozane-tazobactam, since subsequent studies suggested that broth microdilution methods might be more accurate in defining resistance [36,39].
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/antibiotics11020130/s1. Table S1: Antimicrobial susceptibility of 75 MDR-P. aeruginosa isolates collected from Qatar between October 2014 -September 2017. Table S2: Clinical diagnosis and demographic profile of 42 patients with MDR-P. aeruginosa infections at participating Qatar hospitals.  Informed Consent Statement: Patient consent was waived because no identifying data were used. The waiver was approved by the Institutional Ethics Committee of Hamad Medical Corporation, Doha, Qatar.