Mutation-Driven β-Lactam Resistance Mechanisms among Contemporary Ceftazidime-Nonsusceptible Pseudomonas aeruginosa Isolates from U.S. Hospitals

Abstract

OprD loss and hyperexpression of AmpC, MexAB-OprM, MexCD-OprJ, MexEF-OprN, and MexXY-OprM were evaluated among 120 Pseudomonas aeruginosa isolates collected during 2012 in U.S. hospitals and selected based on ceftazidime MIC values (1 to >32 μg/ml). AmpC derepression (10-fold greater than that with the control) and OprD loss (decreased/no band) were the most prevalent resistance mechanisms: 47.5 and 45.8% of the isolates were considered positive, respectively. Elevated expression of the efflux pumps MexAB-OprM, MexCD-OprJ, MexEF-OprN, and MexXY-OprM was observed in 32.5, 8.3, 0.0, and 28.4% of the isolates, respectively. A total of 21 different combinations of resistance mechanisms were noted, and the most prevalent included AmpC derepression with OprD loss with and without efflux hyperexpression (38 and 10 isolates, respectively). A total of 26 isolates had no changes in the resistance mechanisms tested and had lower MIC values for all β-lactams or β-lactam/β-lactamase inhibitor combinations analyzed. OprD loss had a strong correlation with elevated MIC results for imipenem and meropenem (median MIC values of 8 and 4 μg/ml, respectively), with all combinations displaying OprD loss also displaying elevated median MIC values for these carbapenems (4 to >8 μg/ml). AmpC expression levels were greater in isolates displaying elevated cefepime, ceftazidime, or piperacillin-tazobactam MIC values (≥4, ≥4, and ≥16 μg/ml, respectively). Isolates displaying derepressed AmpC had ceftolozane-tazobactam MIC values ranging from 1 to 16 μg/ml. No strong correlation was noticed with MIC values for this β-lactam/β-lactamase inhibitor combination and OprD loss or hyperexpression of efflux systems. Two KPC-producing isolates were detected among 16 isolates displaying ceftolozane-tazobactam MIC values of ≥8 μg/ml.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic nosocomial pathogen that is associated with significant mortality rates and represents a challenge for antimicrobial chemotherapy due to its intrinsic low permeability to various antimicrobial agents and its potential to acquire multiple resistance mechanisms (1, 2). Among valuable therapeutic options for P. aeruginosa infections, cephalosporins with antipseudomonal activity and carbapenems are important due to their activity and low toxicity profiles. Resistance against these β-lactam agents in P. aeruginosa is usually multifactorial (3) and might include the derepression of the chromosomal cephalosporinase AmpC, upregulation of resistance-nodulation-division (RND) efflux systems, and loss of the outer membrane channel OprD, which allows the entry of carbapenems into the cell (1–3). Furthermore, the presence of acquired β-lactamases, including carbapenemases, has been highlighted as a prevalent resistance mechanism against cephalosporins and carbapenems among P. aeruginosa isolates from certain geographic areas (4–6).

Until recently, ceftazidime was considered the most active cephalosporin against P. aeruginosa strains; however, this antimicrobial agent is affected by upregulation of efflux systems, mainly MexAB-OprM, by derepression of AmpC, and by the acquisition of β-lactamases (7). Ceftolozane-tazobactam is an antibacterial consisting of ceftolozane, an antipseudomonal cephalosporin that has improved stability against the chromosomal AmpC produced by P. aeruginosa, and tazobactam, a penicillanic acid-sulfone β-lactamase inhibitor that has activity against constitutive Ambler class C and various class A enzymes (7, 8). Ceftolozane-tazobactam was at least 4-fold more potent than ceftazidime against a large collection of global P. aeruginosa isolates, including multidrug-resistant and extremely drug-resistant isolates (7–9).

In this study, we evaluated the expression of mutation-driven mechanisms of resistance to β-lactams among 120 P. aeruginosa clinical isolates collected in U.S. hospitals during 2012 displaying ceftazidime-nonsusceptible MIC values (≥16 μg/ml) and subsets of isolates displaying modestly elevated ceftazidime MIC values (4 and 8 μg/ml) or susceptible MIC values (1 and 2 μg/ml). Additionally, isolates displaying elevated ceftolozane-tazobactam MIC values (≥8 μg/ml) were screened for the presence of the most common acquired β-lactamases.

MATERIALS AND METHODS

Bacterial isolates.

A total of 998 P. aeruginosa clinical isolates were collected during 2012 in 27 U.S. hospitals and tested against ceftolozane-tazobactam and comparator antimicrobial agents as part of a surveillance initiative (8). The surveillance study included one isolate per patient per episode that was deemed by the participant investigator to be the cause of infection. Isolates were susceptibility tested using the broth microdilution method as described by the Clinical and Laboratory Standards Institute (CLSI), and categorical interpretations applied were those found in CLSI document M100-S24 (10) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) website (11). Quality control was performed using Escherichia coli ATCC 25922 and P. aeruginosa ATCC 27853. All quality control results were within specified ranges as published in CLSI documents.

A total of 120 (12.0% overall) P. aeruginosa isolates displaying ceftazidime MIC values at ≥1 μg/ml were further evaluated (Table 1). These isolates were randomly selected from multiple hospitals to increase geographic and genetic diversity. Species identification was confirmed using the MALDI Biotyper system (Bruker Daltonics, Billerica, MA, USA) according to the manufacturer's instructions.

TABLE 1.

Distribution of P. aeruginosa isolates tested in this study and overall sampling collected in U.S. hospitals during 2012

P. aeruginosa group	No. (%) of isolates	No. (% overall) of isolates inhibited at ceftazidime MIC (μg/ml) of:
		≤1	2	4	8	16	32	>32
Overall isolates	998	156 (15.6)	448 (44.9)	146 (14.6)	76 (7.6)	33 (3.3)	44 (4.4)	75 (7.5)
Isolates selected	120 (12.0)	10 (6.4)	10 (2.2)	15 (10.3)	15 (19.7)	15 (45.5)	20 (45.5)	35 (46.7)

Expression analysis of the chromosomally encoded AmpC and efflux pumps.

The expression of ampC, mexA (MexAB-OprM), mexC (MexCD-OprJ), mexE (MexEF-OprN), and mexX (MexXY-OprM) was determined by quantitative real-time PCR (qRT-PCR) using DNA-free RNA preparations. Total RNA was extracted from mid-log-phase bacterial cultures (cell density at an optical density at 600 nm [OD₆₀₀] of 0.3 to 0.5) using RNA Protect reagent and an RNeasy minikit (Qiagen, Hilden, Germany) in the Qiacube workstation (Qiagen), and residual DNA was eliminated using RNase-free DNase (Promega, Madison, WI, USA). Quantification of mRNA and sample quality was assessed using the RNA 6000 Pico kit on the Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's instructions. Only preparations with an RNA integrity number (RIN) of >8 that showed no visual degradation were used for experiments (12). Relative quantification of target genes was performed in triplicate by normalization to an endogenous reference gene (rpsL) on the StepOne Plus instrument (Life Technologies, Carlsbad, CA, USA) using the Power SYBR green RNA-to-CT kit (Life Technologies) and custom-designed primers showing efficiency of >96.0%, previously validated (5, 13). Transcription levels were considered significantly different if an at least 5- or 10-fold difference was noted compared with P. aeruginosa PAO1 for efflux pumps and AmpC expression, respectively.

Porin detection.

Outer membrane proteins were purified in the FastPrep-24 instrument (MP Biomedicals, Solon, OH, USA), according to the manufacturer's instructions. Normalized concentrations of purified outer membrane proteins were electrophoretically separated and transferred onto polyvinylidene difluoride (PVDF) membranes. Western blots were probed with an affinity-purified polyclonal antibody raised in rabbits using the synthetic OprD peptide N′-SDKTGTGNLPVMNDGKPPD-C' (ThermoFisher Scientific, Rockford, IL, USA) and revealed by use of the WesternBreeze chromogenic kit (Life Technologies) (14). P. aeruginosa PAO1 and two OprD downregulated laboratory constructs were used as positive and negative controls for comparative analysis.

Detection of acquired β-lactamases.

Sixteen isolates displaying ceftolozane-tazobactam MIC values of ≥8 μg/ml were screened by a combination of conventional reference PCR and/or microarray-based (Check-points, Wageningen, The Netherlands) assays for the presence of bla_TEM, bla_SHV, bla_CTX-M, bla_GES, bla_VEB, bla_PER, bla_PSE, and oxacillinases with extended-spectrum β-lactamase (ESBL) spectrum (bla_OXA-2, bla_OXA-10, and bla_OXA-30 groups, bla_OXA-18, and bla_OXA-45). Additionally, isolates were evaluated for the presence of bla_IMP, bla_VIM, bla_KPC, bla_NDM, and bla_SME by PCR (15). Amplicons were sequenced on both strands, and the nucleotide sequences and deduced amino acid sequences were analyzed using the Lasergene software package (DNASTAR, Madison, WI, USA). Sequences were compared with others available via Internet sources (http://www.ncbi.nlm.nih.gov/blast/).

RESULTS AND DISCUSSION

A total of 120 P. aeruginosa isolates collected during 2012 in U.S. hospitals were selected according to the ceftazidime MIC values as displayed in Table 1. These isolates collected in 27 hospitals from 21 U.S. states were recovered from the following sources: nosocomial respiratory infections (n = 65), skin/soft tissue infections (n = 21), bloodstream infections (n = 17), intra-abdominal infections (n = 8), urinary tract infections (n = 6), or other or unknown sources (n = 3).

Overall, 57 (47.5%) isolates were considered to have derepressed chromosomal AmpC, and these isolates displayed expression levels at least 10-fold greater than the baseline levels of isolate P. aeruginosa PAO1 (Table 2). AmpC-derepressed isolates displayed ceftazidime, cefepime, and piperacillin-tazobactam MIC values at ≥4, ≥4, and ≥16, respectively, and isolates that did not overexpress AmpC displayed a median MIC 4- to 8-fold lower for these β-lactam agents (Table 3). MIC values for ceftolozane-tazobactam ranged from 1 to 16 μg/ml (4 μg/ml median) for isolates overexpressing AmpC, and isolates with basal AmpC expression levels had values ranging from 0.5 to >32 μg/ml and a slightly lower median MIC (1 μg/ml) (Table 3).

TABLE 2.

Expression of chromosomal cephalosporinase (AmpC), efflux pumps, and phenotypic presence/absence of outer membrane protein for P. aeruginosa isolates collected from U.S. hospitals

Resistance determinant and difference from control	No. (%) of isolates
AmpC
<5-fold	55 (45.8)
5.0- to 9.9-fold	8 (6.7)
10.0- to 99.9-fold	24 (20.0)
≥100-fold	33 (27.5)
MexAB-OprM
<5-fold	81 (67.5)
5.0- to 9.9-fold	23 (19.2)
10.0- to 99.9-fold	16 (13.3)
≥100-fold	0 (0.0)
MexCD-OprJ
<5-fold	110 (91.7)
5.0- to 9.9-fold	4 (3.3)
10.0- to 99.9-fold	6 (5.0)
≥100-fold	0 (0.0)
MexEF-OprNa
<5-fold	100 (100.0; 74 [61.6%])
5.0- to 9.9-fold	0 (0.0; 26 [21.6%])
10.0- to 99.9-fold	0 (0.0; 20 [16.8%])
≥100-fold	0 (0.0)
MexXY-OprM
<5-fold	86 (71.6)
5.0- to 9.9-fold	5 (4.2)
10.0- to 99.9-fold	29 (24.2)
≥100-fold	0 (0.0)
OprD (Western blot)
No band	55 (45.8)
Similar to control	65 (54.2)

^aP. aeruginosa PAO1 was used as a baseline for the analysis of the results. Due to high expression levels of MexEF-OprN by PAO1, the expression of this efflux system was also analyzed using another P. aeruginosa wild-type laboratory strain (13).

TABLE 3.

MIC ranges for six β-lactam agents according to results for expression of chromosomal cephalosporinase (AmpC), efflux pumps, and phenotypic presence/absence of outer membrane protein among P. aeruginosa isolates collected from U.S. hospitals

Resistance mechanism	No. of isolates	MIC or range (median if range is noted) (μg/ml)
		Ceftazidime	Cefepime	Ceftolozane-tazobactam	Piperacillin-tazobactam	Imipenem	Meropenem
AmpC overexpression
Positive (≥10-fold compared with baseline strain)	57	4–>32 (16)	4–>16 (16)	1–16 (4)	16–>64 (64)	0.5–>8 (1)	0.12–>8 (2)
Negative (<10-fold compared with baseline strain)	63	1–>32 (4)	0.5–>16 (4)	0.5–>32 (1)	0.5–>64 (8)	0.5–>8 (1)	0.06–>8 (0.5)
OprD loss
Positive (no band)	55	2–>32 (16)	2–>16 (8)	0.5–>32 (2)	4–>64 (32)	0.5–>8 (8)	0.12–>8 (8)
Negative (similar to control)	65	1–>32 (8)	0.5–>16 (4)	0.5–>32 (1)	0.5–>64 (8)	0.5–>8 (1)	≤0.06–>8 (0.5)
MexAB-OprM overexpression
Positive (≥5-fold compared with baseline strain)	39	4–>32 (16)	4–>16 (8)	0.5–16 (2)	8–>64 (32)	0.5–>8 (1)	0.25–>8 (2)
Negative (<5-fold compared with baseline strain)	81	1–>32 (8)	0.5–>16 (8)	0.5–>32 (1)	0.5–>64 (8)	0.25–>8 (1)	≤0.06–>8 (0.5)
MexCD-OprJ overexpression
Positive (≥5-fold compared with baseline strain)	10	4–>32 (32)	2–>16 (16)	0.5–4 (2)	4–>64 (32)	1–8 (1)	0.12–>8 (0.5)
Negative (<5-fold compared with baseline strain)	110	1–>32 (8)	0.5–>16 (8)	0.5–>32 (1)	0.5–>64 (16)	0.25–>8 (1)	≤0.06–>8 (1)
MexXY-OprM overexpression
Positive (≥5-fold compared with baseline strain)	34	2–>32 (16)	2–>16 (16)	0.5–>32 (2)	4–>64 (32)	0.5–>8 (1)	≤0.06–>8 (2)
Negative (<5-fold compared with baseline strain)	86	1–>32 (8)	0.5–>16 (8)	0.5–8 (1)	0.5–>64 (8)	0.25–>8 (1)	≤0.06–>8 (0.5)

A total of 55 (45.8%) (Table 2) P. aeruginosa strains showed an OprD decrease/loss, and no band was observed after specific antibody hybridization for all of these isolates. OprD-deficient strains displayed a median MIC value for imipenem and meropenem of 8 μg/ml, whereas isolates displaying the presence of OprD had median MIC values of 1 and 0.5 μg/ml, respectively, for these carbapenems (Table 3).

Among isolates tested, 39 (32.5%) had elevated expression of the MexAB-OprM system (≥5×), and this resistance mechanism was expressed in greater levels (≥10×) in 16 (13.3%) isolates (Table 2). Isolates overexpressing this efflux system displayed a broad range of MIC values for all agents analyzed (Table 3). Piperacillin-tazobactam and meropenem were the only β-lactam agents displaying a difference greater than 2-fold in the median MIC for the group of isolates expressing this resistance mechanism (32 and 2 μg/ml, respectively) compared with those with a basal expression level (8 and 0.5 μg/ml, respectively).

Elevated expression of MexCD-OprJ (≥5×) was noted among 10 (8.3%) of the isolates (Table 2), while the hyperexpression of MexXY-OprM (≥5×) was observed among 32 (28.4%) of the isolates tested (Table 2). The median ceftazidime MIC was 4-fold greater for the isolates that displayed MexCD-OprJ hyperexpression (median MIC, 32 μg/ml) than for those not hyperexpressing this efflux system (median MIC, 8 μg/ml) (Table 3). Similar differences were noted for piperacillin-tazobactam and meropenem when comparing isolates hyperexpressing MexXY-OprM displaying MIC median values of 32 and 2 μg/ml, respectively, and with those with a basal level of expression (median MIC, 8 and 0.5 μg/ml, respectively) (Table 3).

Among all agents analyzed and the groups producing different mutation-driven resistance mechanisms, cefepime and ceftolozane-tazobactam median MIC values were significantly different (>2-fold) only for those isolates displaying derepressed chromosomal cephalosporinase, and differences were also observed for imipenem and OprD presence/absence. On the other hand, ceftazidime, piperacillin-tazobactam, and meropenem seemed influenced by the presence/absence of two or more mutation-driven resistance mechanisms, although meropenem MIC values were still low for mechanisms other than OprD loss (Table 3).

Compared with PAO1 as a baseline, none of the isolates tested displayed elevated expression of MexEF-OprN; however, when the results in this subset were analyzed using a collection of 13 susceptible strains as baseline (5), 46 isolates had modestly and highly elevated expression of this efflux system (Table 2). We have previously observed that P. aeruginosa PAO1 may not be a good control for MexEF-OprN, since it has a baseline expression level that is higher than those of other laboratory and clinical isolates tested (5); however, since MexEF-OprN expression levels have been demonstrated to have limited impact on β-lactams MIC values (16) and for consistency throughout the study, P. aeruginosa PAO1 was used as a baseline for further analysis.

The resistance mechanisms described above were observed alone (21 isolates) or in combinations of two to five mechanisms in the isolates tested (Table 4). Overall, 21 different combinations of resistance mechanisms were noted, and AmpC derepression alone was the most prevalent resistance mechanism (11 strains), followed by a combination of AmpC derepression and OprD loss with or without elevated expression of efflux systems (38 and 10 isolates, respectively) (Table 4).

TABLE 4.

Mutation-driven resistance mechanisms detected among 120 P. aeruginosa isolates and MICs/ranges for six β-lactam agents tested

Resistance mechanism(s)	No. of isolates	MIC or range (median if range is noted) (μg/ml)
		Ceftazidime	Cefepime	Ceftolozane-tazobactam	Piperacillin-tazobactam	Imipenem	Meropenem
AmpC overexpression	11	8->32 (32)	4->16 (16)	1–4 (2)	16->64 (>64)	0.5–4 (1)	0.12–2 (0.5)
OprD loss, AmpC + MexAB-OprM overexpression	11	4->32 (32)	8->16 (16)	1–8 (2)	16->64 (>64)	1->8 (>8)	4->8 (>8)
OprD loss, AmpC overexpression	10	16->32 (>32)	8->16 (16)	1–8 (4)	>64 (>64)	1->8 (8)	0.5->8 (8)
OprD loss, AmpC + MexXY-OprM overexpression	8	4->32 (>32)	8->16 (16)	1–4 (4)	16->64 (>64)	8->8 (>8)	4->8 (8)
AmpC + MexXY-OprM overexpression	6	16->32 (32)	8->16 (16)	1–16 (2)	32->64 (>64)	0.5->8 (1)	0.25->8 (1)
AmpC + MexAB-OprM overexpression	6	16->32 (>32)	8->16 (>16)	1–8 (8)	64->64 (64)	0.5–1 (1)	0.25->8 (1)
OprD loss	6	2–16 (4)	2–8 (4)	0.5–2 (1)	4–16 (8)	0.5->8 (8)	0.12->8 (4)
OprD loss, AmpC + MexAB-OprM + MexXY-OprM overexpression	6	16->32 (>32)	16->16 (>16)	1–16 (4)	32->64 (>64)	8->8 (>8)	4->8 (8)
OprD loss, MexAB-OprM overexpression	6	4–8 (4)	4–8 (8)	0.5–1 (0.5)	8->64 (16)	2->8 (8)	2–8 (8)
OprD loss, MexXY-OprM overexpression	5	8->32 (>32)	8->16 (16)	1->32 (4)	8->64 (>64)	1->8 (8)	8->8 (>8)
MexAB-OprM overexpression	4	4->32 (32)	4->16 (>16)	1–4 (2)	16->64 (64)	0.5–1 (1)	0.25–2 (1)
AmpC+ MexCD-OprJ overexpression	3	>32	16->16 (>16)	4	>64	1	0.25–1 (0.25)
MexAB-OprM + MexXY-OprM overexpression	3	8–16 (8)	8–16 (8)	0.5–2 (1)	16–64 (16)	0.5->8 (1)	0.25->8 (0.25)
MexXY-OprM overexpression	3	2->32 (2)	4->16 (8)	1->32 (1)	4->64 (8)	0.5–1 (1)	≤0.06->8 (0.25)
AmpC+ MexAB-OprM + MexCD-OprJ + MexXY-OprM overexpression	1	>32	>16	4	>64	1	0.25
MexCD-OprJ overexpression	1	8	4	1	16	1	0.5
MexCD-OprJ + MexXY-OprM overexpression	1	4	2	0.5	4	1	0.12
OprD loss, AmpC+ MexAB-OprM + MexCD-OprJ overexpression	1	32	16	1	>64	8	2
OprD loss, AmpC + MexAB-OprM + MexCD-OprJ + MexXY-OprM overexpression	1	32	>16	1	64	8	4
OprD loss, AmpC+ MexCD-OprJ overexpression	1	32	16	2	64	8	8
OprD loss, MexAB-OprM + MexCD-OprJ overexpression	1	8	8	1	32	4	>8
None	26	1–32 (2)	≤0.5–16 (2)	0.5–4 (0.5)	≤0.5->64 (4)	0.25->8 (1)	≤0.06–1 (0.25)

Resistance mechanisms were analyzed according to MIC values for six β-lactams and β-lactam/β-lactamase inhibitor combinations, and median MIC values were compared (Table 4). Isolates displaying derepression of chromosomal AmpC alone had elevated median MIC values for ceftazidime, cefepime, and piperacillin-tazobactam (32, 16, and >64 μg/ml, respectively) (Table 4). Additionally, for 11 isolates showing only loss of OprD, carbapenem MIC values were elevated (median MIC, 8 and 4 μg/ml for imipenem and meropenem, respectively), and among four isolates hyperexpressing MexAB-OprM, ceftazidime, cefepime, and piperacillin-tazobactam had elevated median MIC values (32, >16, and 64 μg/ml, respectively) (Table 4). These results confirm the results of the analysis of the resistance mechanism separately (Table 3).

Among prevalent combinations of resistance mechanisms detected, the combination of OprD loss and elevated expression of AmpC without (10 isolates) or with elevated expression of MexAB-OprM (11 isolates), MexXY-OprM (8 isolates), or both (6 isolates) displayed elevated median MIC values for all agents analyzed except ceftolozane-tazobactam (median MIC, 2 to 4 μg/ml) (Table 4). However, when analyzing isolates carrying AmpC derepression with hyperexpression of one or more tripartite efflux systems, the median MIC values were elevated for ceftazidime, cefepime, and piperacillin-tazobactam (median MIC ranges, 32 to >32, 16 to >16, and 64 to >64 μg/ml, respectively) but not for carbapenems (Table 4). In a recent study, Cabot et al. (17) evaluated meropenem-induced mutants and observed that these strains displayed OprD deficiencies but also had hyperexpression of AmpC with or without efflux systems elevating MIC values for other β-lactam agents tested. Interestingly, ceftolozane-tazobactam MIC values for these mutant strains were still low (17).

Six isolates displaying OprD loss and derepression of MexAB-OprM had modestly elevated median MIC values for cephalosporins and β-lactam/β-lactamase inhibitor combinations (from 0.5 to 16 μg/ml) and elevated median MIC values for imipenem and meropenem (8 μg/ml for both carbapenems) (Table 4). Isolates displaying OprD loss with MexXY-OprM hyperexpression (5 isolates) had elevated median MIC values for all agents except ceftolozane-tazobactam (4 μg/ml); however, isolates expressing only MexXY-OprM (n = 3) had much lower median MIC values for all agents (0.25 to 8 μg/ml), suggesting that OprD loss is needed to encode resistance to these antimicrobial agents when MexXY-OprM is overexpressed.

Twenty-six (21.6%) isolates carried none of the resistance mechanisms analyzed, and 22 had ceftazidime MIC values of ≤8 μg/ml (susceptible according to CLSI criteria); 10 of those had ceftazidime MIC of 1 μg/ml. These isolates displayed ceftolozane-tazobactam MIC results at ≤4 μg/ml (median MIC, 0.5 μg/ml), and meropenem MIC results were ≤1 μg/ml (Table 4).

Due to the apparent limited correlation of ceftolozane-tazobactam MIC values to specific mutation-driven resistance mechanisms (the exception being derepressed AmpC), we evaluated the presence of acquired β-lactamases among isolates displaying ceftolozane-tazobactam MIC values of ≥8 μg/ml.

Ceftolozane-tazobactam resistance seems to be more frequent among P. aeruginosa and Enterobacteriaceae clinical isolates when an acquired resistance mechanism, such as production of β-lactamases, is detected than with the mutation-driven mechanisms (7, 18). Additionally, development of mutation experiments performed with P. aeruginosa laboratory strains exposed to elevated concentrations of ceftolozane-tazobactam and other efflux pump substrates showed that mutant strains had generally low MIC values for this cephalosporin/β-lactamase inhibitor combination. These mutants carried a combination of alterations that included OprD loss and AmpC derepression with or without MexAB-OprM hyperexpression (17).

The 16 isolates displaying ceftolozane-tazobactam MIC values of ≥8 μg/ml presented six different combinations of mutation-driven mechanisms, and OprD loss with AmpC derepression with or without MexAB-OprM was observed in 8 of these strains and AmpC derepression and MexAB-OprM overexpression among 3 other strains (Table 5). Among these 16 isolates, only 5 carried the acquired β-lactamase-encoding genes tested (Table 5). Two isolates carried genes encoding KPC-2 or -3, and three other isolates carried genes encoding the ESBL genes SHV-12, OXA-17, and OXA-226. As expected, KPC-carrying P. aeruginosa displayed elevated MIC values for all β-lactam agents reported, and both strains showed hyperexpression of MexXY-OprM with AmpC derepression or OprD loss (Table 5). Isolates carrying genes encoding OXA-17 and OXA-226 also displayed OprD loss and AmpC derepression alone or with MexAB-OprM and MexXY-OprM hyperexpression (OXA-226 producer). The isolate carrying bla_SHV-12 showed MexXY-OprM overexpression in addition to this ESBL gene. OXA-226 has not been described previously, and this enzyme displayed one amino acid alteration (W158R) compared with OXA-2.

TABLE 5.

Acquired β-lactamases detected among 16 P. aeruginosa isolates displaying ceftolozane-tazobactam MICs of ≥8 μg/ml

City, state	MIC (μg/ml)						Acquired β-lactamase	Mutation-driven resistance mechanisms
	Ceftazidime	Cefepime	Ceftolozane-tazobactam	Piperacillin-tazobactam	Imipenem	Meropenem
Charlottesville, VA	>32	>16	8	>64	1	1	None	AmpC + MexAB-OprM overexpression
Long Beach, CA	>32	>16	8	>64	>8	8	None	OprD loss, AmpC overexpression
Long Beach, CA	>32	>16	8	>64	1	1	None	OprD loss, AmpC overexpression
Galveston, TX	>32	>16	16	>64	8	4	None	OprD loss, MexXY-OprM overexpression
Akron, OH	>32	16	>32	64	8	>8	OXA-226	OprD loss, AmpC + MexAB-OprM + MexXY-OprM overexpression
Houston, TX	>32	>16	16	>64	0.5	1	None	AmpC + MexXY-OprM overexpression
Houston, TX	>32	>16	16	>64	>8	>8	KPC-2	AmpC + MexXY-OprM overexpression
Detroit, MI	>32	>16	8	>64	>8	>8	None	OprD loss, AmpC overexpression
Ewa Beach, HI	>32	>16	8	>64	1	0.25	None	AmpC + MexAB-OprM overexpression
Aurora, CO	>32	>16	8	>64	8	4	OXA-17	OprD loss, AmpC overexpression
West Roxbury, MA	>32	>16	>32	>64	0.5	1	SHV-12	MexXY-OprM overexpression
Tampa, FL	>32	>16	8	>64	>8	>8	none	OprD loss, AmpC + MexAB-OprM overexpression
New York, NY	>32	>16	8	>64	1	1	None	AmpC + MexAB-OprM overexpression
Charlottesville, VA	>32	>16	8	>64	>8	>8	None	OprD loss, AmpC + MexAB-OprM + MexXY-OprM overexpression
New York, NY	>32	>16	>32	>64	>8	>8	KPC-3	OprD loss, MexXY-OprM overexpression
Lexington, KY	>32	>16	8	>64	>8	8	None	OprD loss, AmpC + MexAB-OprM overexpression

In summary, AmpC derepression and OprD loss were the most common mutation-driven resistance mechanisms detected in selected P. aeruginosa strains collected in U.S. hospitals. In all cases, OprD loss was essential for carbapenem resistance, being present in all isolates with elevated MIC values for these agents. Slightly higher MIC values for imipenem than for meropenem in strains lacking OprD were observed previously (1, 2) and were noted in this clinical sampling. Furthermore, isolates harboring this resistance mechanism alone or with hyperexpression of AmpC and/or efflux systems had a tendency to display higher MIC values for all agents, which aligns with data previously presented on laboratory strains (2). Similarly, AmpC derepression seemed to increase ceftazidime and cefepime MIC values, but isolates hyperexpressing MexAB-OprM alone or with OprD loss also had elevated MIC values for these two cephalosporins, and this finding corroborates the data generated using laboratory mutants selected with different cephalosporins (19).

Among mutation-driven resistance mechanisms against β-lactam agents, OprD loss and AmpC derepression seem to be more important for the development of resistance, and mainly the latter can be a complex process and a result of the inactivation of AmpD, inactivation of the transcriptional regulator AmpR, or PBP4 inactivation (18, 20, 21).

In this study, hyperexpression of efflux pumps in most cases (71/83 isolates) was observed among isolates that also harbored AmpC derepression and/or OprD loss. In general, the expression of efflux pumps did not display strong correlations with MIC values of the antimicrobial agents selected. Additionally, the presence of acquired β-lactamases did not seem to play an important role among P. aeruginosa isolates displaying elevated ceftolozane-tazobactam MIC values; however, two isolates carried KPC-encoding genes that are not very common among this species and were observed to confer resistance to all β-lactam agents tested. Evaluations of mutation-driven resistance mechanisms in P. aeruginosa have often been performed with laboratory mutants (1, 2, 19). Although these studies are extremely important for understanding resistance mechanisms in a homogeneous background with controlled variants, the study of these mechanisms in a contemporary clinical isolate population not only serves to validate the findings with laboratory mutants but also gives insight into how these mechanisms accumulate and interact in regard to the susceptibility of β-lactams and β-lactam/β-lactamase inhibitor combinations.