Ceftazidime-Induced Overexpression of MexJK-OprM Confers Resistance to Ceftolozane/Tazobactam in an Isolate of Pseudomonas aeruginosa

Introduction

Pseudomonas aeruginosa is a major causative organism of nosocomial infections such as ventilator-associated pneumonia (VAP), and antimicrobial resistance has become a worldwide problem.^1,2 In particular, the effective antimicrobial agents for infections caused by carbapenem-resistant P. aeruginosa were extremely limited.^2,3 Ceftolozane/tazobactam (CTLZ/TAZ) is a novel cephalosporin antibiotic combined with a β-lactamase inhibitor, developed to combat antimicrobial-resistant P. aeruginosa infections.^4,5

Recent surveillance data from Europe, the United States, and Japan showed CTLZ/TAZ resistance rates in Pseudomonas aeruginosa of 6.7%, 3.4% and 5.4%, respectively.^6–8 However, resistance rates are considerably higher in China and South Africa, ranging from 13 to 18.1%,^9,10 likely due to the widespread dissemination of carbapenemase-producing strains.^11,12 The primary mechanism of CTLZ/TAZ resistance involves mutations and overexpression of AmpC β-lactamase.^13,14 Overexpression of AmpC in P. aeruginosa is a well-established resistance mechanism against anti-pseudomonal cephalosporins such as ceftazidime (CAZ) and cefepime (CFPM).^13,14 In contrast, CTLZ/TAZ is largely unaffected by this mechanism, and P. aeruginosa has generally been considered unlikely to develop resistance to CTLZ/TAZ.¹⁵ However, recent studies have reported the emergence of CTLZ/TAZ-resistant P. aeruginosa, with mutations in the ampC β-lactamase gene identified as a potential resistance mechanism.^16,17 Specific mutations in AmpC, such as G183D, E247K, and T96I, alter the structure of the substrate-binding site, resulting in an expanded substrate-binding pocket of AmpC and leading to increased hydrolytic activity against CTLZ.^13,14 These mutations, when coupled with AmpC overexpression, raise the minimum inhibitory concentration (MIC) of CTLZ/TAZ to 64–128 μg/mL.¹⁸ Upregulation of efflux pumps and downregulation of the D₂ porin have also been occasionally reported to contribute to CTLZ/TAZ resistance,¹⁹ but the involvement of MexJK in resistance has been underexplored.

In this study, we investigated the mechanism of CTLZ/TAZ-resistance in P. aeruginosa strain isolated from a patient with a lung abscess who had received multiple antimicrobial agents. Additionally, based on this clinical case, we conducted in vitro exposure experiments with multiple antimicrobial agents to evaluate how specific combinations influence the acquisition of resistance.

Materials and Methods

CTLZ/TAZ-Resistant P. aeruginosa Isolated from Sputum of Patients with Lung Abscess

A CTLZ/TAZ-resistant strain of P. aeruginosa was isolated from an 84-year-old Japanese man with colorectal cancer. His only medical history was type 2 diabetes, with no history of respiratory disease. The patient was administered with cefmetazole for 5 days as prophylaxis against surgical site infection following a colectomy. However, on postoperative day five, he developed a fever, and abdominal ultrasound findings suggested acute cholecystitis. As a result, percutaneous transhepatic gallbladder drainage was performed, and the antibiotic was switched to tazobactam/piperacillin (TAZ/PIPC). At this time, a P. aeruginosa strain susceptible to antibiotics (Pa-S) was detected in the blood culture (Figure 1A; allow A). Despite 8 days of TAZ/PIPC treatment, the patient’s condition did not improve significantly. Therefore, the antibiotic regimen was changed to levofloxacin (LVFX) combined with metronidazole to enhance treatment against anaerobic bacteria. One week later, based on respiratory symptoms and chest CT (cystography) (Figure 2), a lung abscess caused by P. aeruginosa was diagnosed. Consequently, the antimicrobial agent was changed to ceftazidime (CAZ), which was administered for 13 days. However, due to lack of improvement, a combination of CTLZ/TAZ and amikacin (AMK) was initiated. After 2 weeks of treatment, there was still no improvement in chest radiographic opacities, and a CTLZ/TAZ-resistant P. aeruginosa strain (Pa-R) was isolated from the patient’s sputum (Figure 1A; allow B).

Figure 1 Clinical course of patient with lung abscess who isolated ceftolozane/tazobactam-resistant Pseudomonas aeruginosa (A) and the antimicrobial susceptibility of clinical isolates (B). Allow A indicates detection of antimicrobial-susceptible P. aeruginosa (Pa-S); Allow B indicates detection of ceftolozane/tazobactam–resistant P. aeruginosa (Pa-R). Abbreviations: AMK, amikacin; BT, body temperature; CAZ, ceftazidime; CMZ, cefmetazole; CPFX, ciprofloxacin; CRP, C-reactive protein; LVFX, levofloxacin; MNZ, metronidazole; CTLZ/TAZ, ceftolozane/tazobactam; TAZ/PIPC, tazobactam/piperacillin; WBC, white blood cells.

Figure 2 Chest X-ray and CT image when diagnosed with a lung abscess. Due to the patient’s respiratory condition and the presence of Pseudomonas aeruginosa in his blood, he was diagnosed with a lung abscess caused by the bacterium.

Both the Pa-S and Pa-R strains were analyzed using the PCR-based open reading frame typing (POT) method with the Cica Geneus^TM Pseudo POT KIT (KANTO KAGAKU, Tokyo, Japan).²⁰ Additionally, the presence or absence of carbapenemase genes (IMP-1, IMP-6, VIM, KPC, NDM, OXA-48, GES) was determined by multiplex PCR (CFX Connect; Bio-Rad, Hercules, CA) using the Cica Geneus^TM Carbapenemase Genotype Detection KIT 2 (KANTO KAGAKU).²¹

Antimicrobial Susceptibility Testing and Resistance Mechanism Analysis

MIC of CTLZ/TAZ (Merck & Co., Inc., Rahway, NJ), TAZ/PIPC (Taiho Pharmaceutical Co., Ltd., Tokyo, Japan), CAZ (WAKO, Osaka, Japan), avibactam/ceftazidime (AVI/CAZ; Selleck Chemicals, Co., Ltd., Houston, TX), cefepime (CFPM; WAKO), imipenem (IPM; WAKO), MEPM (WAKO), AMK (WAKO), ciprofloxacin (LKT Laboratories Inc., MN) and LVFX (LKT Laboratories Inc.) were determined by microdilution according to the CLSI document.²² Continuously, 1000 μg/mL cloxacillin (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) as AmpC inhibitor and 20 μg/mL carbonyl cyanide 3-chlorophenylhydrazone (CCCP; Tokyo Chemical Industry Co., Ltd.) as efflux pump inhibitor were used to evaluate the contribution of AmpC and efflux pump in CTLZ/TAZ resistance.^14,23 Each DNA sequence of ampC and ftsI which encoded AmpC and penicillin binding protein 3 (PBP-3), respectively, was determined by Sanger’s dye terminator cycle sequencing method.^24,25 In addition, each expression level of ampC, oprD and mex family mRNA was assayed by real-time RT-PCR (CFX Connect; Bio-Rad) using the iTaq Universal SYBR Green One-Step Kit (Bio-Rad).^26–29

RNA Transfection

siRNA transfection was performed using HiPerFect reagents (Qiagen, Stockholm, Sweden), according to the manufacturer’s instructions.³⁰ It was placed in room temperature for 30 minutes after mexJK-specific dsRNA (5’-GGU GUC ACA UGU ACC GCC AdTdT-3’ and 5’-UGG CGG UAC AUG UGA CAC CdTdT-3’) was mixed with this reagent. One-hundred μL of Pa-R bacterial suspension prepared with 10⁶ CFU/mL and 5 μL of dsRNA mixed reagent were inoculated into a 96-well plate, and incubated for 6 hours at 37°C. After the transfection, MIC of CTLZ/TAZ was determined as described above.

In vitro Serial Passage

An in vitro serial passage was performed for confirmation by a combination of repetitive administration of each antimicrobial agent whether Pa-S strain acquired CTLZ/TAZ-resistance.³¹ Ten μL of the bacterial suspension prepared with McFarland no. 0.5 was inoculated onto Mueller-Hinton agar plates including serially diluted antibiotic of 8–0.5 × MIC. Plates were incubated for 24 hours at 37°C, and then colonies grown on sub-MIC plates were re-inoculated into other dilution series. This process was repeated 14 times, the MIC of CTLZ/TAZ was determined by microdilution method every time.

Statistical Analysis

The data were calculated as the mean and standard deviation per experimental group and compared by an unpaired t-test. Statistical significance was set at P <0.05. All experiments were performed in duplicate on another day. All data were curated from at least four samples to assess an assumption of normality, and the results are presented as mean ± standard deviation.

Results

Antimicrobial Susceptibility and Resistant Mechanisms of Clinically Isolated P. aeruginosa

Both Pa-S and Pa-R strains showed the same POT types, confirming that they are identical strains (Figure 3). Both strains had no carbapenemase genes. The MIC of TAZ/CTLZ for Pa-S and Pa-R were 0.5 and 32 μg/mL, respectively (Figure 1B). Pa-R showed resistance to TAZ/PIPC and CAZ but remained susceptible to LVFX and AMK. The MIC of CTLZ/TAZ in Pa-R decreased to 4 μg/mL when combined with the efflux pump inhibitor CCCP (Table 1). mRNA expression analysis showed an approximately 30-fold increase in only the mexJK gene (Figure 4; P <0.05). RNA transfection with mexJK-specific dsRNA suppressed mexJK mRNA translation, reducing the MIC of CTLZ/TAZ to 4 μg/mL. Combined use of the AmpC inhibitor cloxacillin did not changed CTLZ/TAZ susceptibility. No mutations in AmpC were identified, and no changes in AmpC mRNA expression were observed. No mutations were identified in PBP-3.

Table 1 MIC Value for P. aeruginosa Strains with Inhibitor of AmpC β-Lactamases and Efflux Pumps

Figure 3 Agarose gel electrophoresis patterns of PCR-based open-reading-frame typing. Lane M, 100-bp ladder marker, lane PC, positive control. Mixture 1 is constructed PCR PC, Genomic islet-1, −2, −3, −4, −5, VIM, Prophage-1, −2, and −3 for 506, 336, 281, 235, 201, 175, 151, 126, 103, 85 bp, respectively. Mixture 2 is constructed for PCR PC, Genomic islet-6, −7, −8, −9, −10, Prophage-4, −5, and IMP for 506, 324, 271, 238, 204, 176, 150, 124, 105 bp, respectively.

Figure 4 mRNA expression in ceftolozane/tazobactam -resistant Pseudomonas aeruginosa clinical isolate. Pa-S means ceftolozane/tazobactam -susceptible P. aeruginosa isolated before antibiotic treatment; Pa-R means ceftolozane/tazobactam -resistant P. aeruginosa isolated after multiple antibiotic treatment. *P <0.05.

Acquisition of Tazobactam/Ceftolozane Resistance by the in vitro Serial Passage

The Pa-S strain acquired resistance to CTLZ/TAZ (MIC: 32 μg/mL) at 38 days later (Pa-S^d38) when exposed in vitro as re-appearance in a case in CTLZ/TAZ following TAZ/PIPC and CAZ for 2 weeks, respectively (Figure 5A). However, the elevation of the MIC value remained to 2 μg/mL only by having made CTLZ/TAZ exposed for two weeks (Figure 5B). To determine whether pre-exposure to CAZ or TAZ/PIPC affected CTLZ/TAZ resistance development, either CAZ or TAZ/PIPC was exposed for 2 weeks followed by 2 weeks of CTLZ/TAZ exposure (Figure 5C and D). The MIC value showed 32 μg/mL following CAZ 24 days after CTLZ/TAZ exposure, and the Pa-S strain acquired resistance to CTLZ/TAZ (Pa-S^d24). In the Pa-S^d38 and Pa-S^d24, mexJK expression levels increased 38 and 34 times, respectively, compared with its parental strain (Figure 6; P <0.05).

Figure 5 Change of ceftolozane/tazobactam MIC in clinical isolate Pa-S through an in vitro serial passage under the antibiotic pressure. (A) tazobactam/piperacillin, ceftazidime and ceftolozane/tazobactam were exposed for 2 weeks each (total 42 days); (B) ceftolozane/tazobactam only was exposed for 2 weeks; (C) ceftazidime followed by ceftolozane/tazobactam, exposed for 2 weeks (total 28 days); (D) tazobactam/piperacillin followed by ceftolozane/tazobactam, exposed for 2 weeks (total 28 days). Abbreviations: CAZ, ceftazidime; CTLZ/TAZ, ceftolozane/tazobactam; TAZ/PIPC, tazobactam/piperacillin.

Figure 6 Changes of mRNA expression by the in vitro serial passage. Pa-S (wt): Pa-S clinical isolate (wild type), Pa-S (d24) and Pa-S (d38): Pa-S who acquired resistance by the in vitro serial passage, *P <0.05.

Discussion

Many ampC-mutated strains also harbor mutations in DNA mismatch repair genes such as mutS and mutL, impairing the correction of ampC mutations and promoting stable resistance.^32,33 In general, CTLZ/TAZ resistance mediated by ampC mutations can be reversed by co-administration of the AmpC inhibitor cloxacillin.¹⁴ However, in this study, the CTLZ/TAZ-resistant strain (Pa-R) remained resistant even in the presence of cloxacillin.

No known ampC mutations were detected in the Pa-R strain nor were mutations found in the DNA mismatch repair genes mutS or mutL (data not shown). These findings suggest that resistance in this isolate is not attributable to AmpC alterations but instead involves an alternative mechanism. Co-administration of the efflux pump inhibitor CCCP restored CTLZ/TAZ susceptibility in Pa-R, implicating efflux-mediated resistance.

Quantitative analysis of mex family gene expression revealed that most mex genes were expressed at levels comparable to those in the susceptible Pa-S strain. However, mexJK expression was markedly elevated. The MexJK efflux pump, which operates in conjunction with OprM, has previously been associated with resistance to erythromycin and tetracycline.³⁴

To directly assess the involvement of MexJK, we conducted an RNA transfection assay using a synthetic small RNA designed to hybridize with the 5′ untranslated region of mexJK mRNA, thereby inhibiting its translation. Transfection of this RNA into the Pa-R strain significantly restored susceptibility to CTLZ/TAZ, supporting the conclusion that overexpression of the MexJK-OprM efflux system underlies resistance in this isolate. Notably, mexJK expression is known to be upregulated under quorum-sensing conditions.²⁷ Since the Pa-R strain was isolated from a patient with a pulmonary abscess—a microenvironment likely characterized by a high bacterial density—activation of the quorum-sensing system may have facilitated MexJK overexpression.^35,36 In contrast, other major efflux systems in P. aeruginosa, such as MexAB and MexCD, are typically downregulated in response to quorum sensing,³⁷ which may explain their lack of involvement in the resistance phenotype observed in this case.

Recent in vitro studies have shown that exposure to meropenem (MEPM) can induce CTLZ/TAZ resistance via ampC mutations and overexpression.³⁸ However, the patient from whom Pa-R was isolated had no history of MEPM use, and no AmpC-mediated resistance mechanism was detected. Instead, prior administration of CAZ or TAZ/PIPC was considered a potential factor contributing to resistance development. In vitro sequential exposure assays revealed that pre-treatment with CAZ followed by CTLZ/TAZ induced resistance, accompanied by increased mexJK expression. CAZ has previously been reported to upregulate efflux pumps such as MexAB in P. aeruginosa,³⁹ and our findings suggest that it may similarly induce mexJK. Nonetheless, the regulatory pathways governing mexJK expression remain poorly understood and warrant further investigation.⁴⁰

Conversely, prolonged exposure to TAZ/PIPC for 14 days had no impact on CTLZ/TAZ susceptibility. Given that TAZ/PIPC resistance in P. aeruginosa is primarily mediated by AmpC overexpression—which does not impair CTLZ/TAZ efficacy—this observation aligns with previous reports.

Limitation

As this study was limited to a single clinical isolate, further investigations involving a broader range of clinical strains are needed to validate and generalize these findings.

Conclusion

We identified overexpression of the MexJK-OprM efflux system, rather than AmpC mutation or overexpression, as a novel CTLZ/TAZ resistance mechanism in P. aeruginosa. Our results also suggest that prior CAZ administration may contribute to the emergence of resistance, underscoring the importance of careful therapeutic transitions from CAZ to CTLZ/TAZ. In cases where CAZ has been pre-administered, a switch to a carbapenem, such as MEPM, should be considered.