Emergence of KPC-31, a KPC-3 Variant Associated with Ceftazidime-Avibactam Resistance, in an Extensively Drug-Resistant ST235 Pseudomonas aeruginosa Clinical Isolate

ABSTRACT

A ceftazidime-avibactam-resistant KPC-producing Pseudomonas aeruginosa strain was isolated in Argentina from a tracheal aspirate. The patient was treated with ceftazidime-avibactam in combination with other agents for 130 days. Whole-genome sequencing of P. aeruginosa identified a D179Y substitution in the Ω loop of KPC-3, corresponding to KPC-31, integrated at the chromosome. The strain belonged to the sequence type 235/O11 (ST235/O11) high-risk clone. Evaluation of carbapenemase detection assays most used by clinical laboratories failed to identify the isolate as a KPC producer.

INTRODUCTION

Pseudomonas aeruginosa is an opportunistic pathogen that causes a variety of serious nosocomial infections with high morbidity and mortality rates (1). The World Health Organization has categorized carbapenem-resistant P. aeruginosa (CRPA) as a high-priority pathogen for the research and development of antibiotics. The common mechanisms underlying resistance in CRPA include impermeability due to the inactivation of OprD, the overexpression of efflux pumps, the hyperexpression of chromosomally encoded PDC, and/or the acquisition of carbapenemases, mostly metallo-β-lactamases. However, an increasing number of KPC-producing P. aeruginosa (KPC-PA) has been reported worldwide. In Argentina, bla_KPC-2-producing P. aeruginosa has been detected since 2008, causing regional outbreaks in Patagonian and Buenos Aires districts (2).

P. aeruginosa has a nonclonal epidemic population structure, but some sequence types (ST111, ST175, ST235, ST244, and ST395) have been described to be associated with outbreaks worldwide (1). ST235 is the most prevalent of these high-risk clones and has been associated with poor clinical outcomes in part due to high-level antibiotic resistance (1, 3).

New β-lactamase inhibitor combinations, such as ceftazidime-avibactam (CZA), showed good antipseudomonal activity to treat several infections, including those caused by multidrug-resistant/extensively drug-resistant (MDR/XDR) P. aeruginosa, especially KPC-PA. Nevertheless, some isolates showing different mechanisms of resistance affecting CZA activity have been reported, including alterations in PDC, acquisition of metallo-β-lactamase, impermeability, and efflux (4, 5). Unlike Enterobacterales, KPC allelic variants that confer resistance to CZA have only been exceptionally reported in P. aeruginosa (5). The aim of this work is to describe a CZA-resistant P. aeruginosa clinical isolate harboring bla_KPC-31.

A 20-year-old patient with a history of Friedreich's ataxia was admitted on 18 February 2021 to perform a spine corrective surgery. On 1 March, the patient evolved with multiple infections as follows: aspiration pneumonia due to ampicillin-resistant Klebsiella pneumoniae and an upper respiratory infection (bronchorrhea) due to P. aeruginosa susceptible to meropenem, ceftazidime, CZA, and colistin. Initially, the patient was treated with a combination of meropenem plus colistin. On 2 April, a surgical wound toilette was performed, recovering extended-spectrum β-lactamase (ESBL)-producing K. pneumoniae showing susceptibility to CZA plus a carbapenem-resistant P. aeruginosa isolate susceptible to ceftazidime, CZA, and colistin. The presence of carbapenemases was phenotypically ruled out in both isolates following the recommendations of the National Reference Laboratory (NRL) (see http://antimicrobianos.com.ar/2021/11/algoritmos-de-deteccion-de-carbapenemasas-2021/). Treatment with tigecycline, ceftazidime, and colistin was implemented for 28 days. Patient was admitted to the operating room on several occasions to perform surgical drainage of a persistent spine collection and removal of necrotic tissue. The case was assumed as chronic osteomyelitis, which required prolonged combination treatment with CZA (2.5 g every 8 h [q8h], 2-h infusion, 130 days) plus colistin (150 mg/12 h, 1-h infusion, 60 days). The patient was cultured intratreatment at several opportunities for infectious complications (Enterococcus central catheter-associated bacteremia) and noninfectious complications (pneumothorax). On 8 July 2021, a routine tracheal aspirate revealed colonization by an XDR P. aeruginosa isolate (named M27432). Antimicrobial susceptibility testing was performed by agar dilution or gradient diffusion (6, 7). P. aeruginosa M27432 was resistant to penicillins, cephalosporins, monobactam, carbapenems, ciprofloxacin, gentamicin, and amikacin (Table 1). Additional MIC testing revealed high-level resistance to CZA (≥256 μg/mL) and ceftolozane-tazobactam (≥256 μg/mL). P. aeruginosa M27432 was only susceptible to colistin by broth disk elution (7) and imipenem-relebactam. Addition of 4 μg/mL of avibactam achieved a MIC reduction for imipenem to susceptible values (MIC, 2 μg/mL) but failed to reduce the MICs of aztreonam, meropenem, or ceftolozane (Table 1). The MIC of aztreonam but not CZA and other selected beta-lactam agents was reduced in the presence of the efflux inhibitor phenyl-arginine β-naphthylamide (PAbN). During the first days of August, antimicrobial treatment was suspended and the patient was referred to a rehabilitation center until mid-October where he was discharged alive.

TABLE 1.

Antimicrobial susceptibility profile of P. aeruginosa M27432^a

Antibiotic(s) or test	Test result	MIC (μg/mL)
		MH	MH+PAbN
Antibiotic
Imipenem		8	8
Imipenem + avibactam		2
Imipenem + relebactam		1	1
Meropenem		≥32	≥32
Meropenem + avibactam		32
Ceftazidime		≥256	≥256
Ceftazidime + avibactam		≥256	≥256
Ceftaroline		96
Ceftaroline + avibactam		48
Aztreonam		64	8
Aztreonam + avibactam		32
Ceftolozane + tazobactam		≥256	≥256
Ceftolozane + avibactam		≥256
Piperacillin + tazobactam		≥128
Colistin		≤2
Amikacin		≥64
Gentamicin		≥64
Ciprofloxacin		≥32
Carbapenemase production test
O.O.K. lateral flow	Negative
BCT and CNPd	Negative
mCIM	Negative
THT	Negative

^aMIC values were determined in Mueller Hinton media (MH) with or without supplementation with the efflux pump inhibitor phenyl-arginine β-naphthylamide (PAbN) (at 100 mg/L). Avibactam and relebactam were tested at a fixed 4 μg/mL (final concentration). BCT, Blue-Carba test; CNPd, CARBA NP direct.

M27432 was screened for carbapenemase production with the modified carbapenem inactivation method (mCIM) test yielding a negative result (7). Additionally, no synergism between a carbapenem and EDTA disks was observed, discarding a possible metallo-β-lactamase production (8). Triton-Hodge test (THT), colorimetric assays for carbapenemase production (CARBA NP-Direct and Blue Carba test), and lateral flow immune assays (Resist-3 O.O.K. K-SeT, Coris Bioconcept, and NG-Test CARBA 5; NG Biotech) were negative to detection of a KPC target (9–12). An in-house multiplex PCR for bla_KPC, bla_OXA-48-like, bla_NDM, bla_IMP, and bla_VIM gene detection yielded a positive result for bla_KPC. Short read plus long read whole-genome sequencing of P. aeruginosa M27432 was performed to obtain a hybrid assembly sequence using Unicycler. Only one circular contig (6,868,572 bp) was obtained. blaKPC-31 carbapenemase gene, a variant of bla_KPC-3, was detected (accession number CP101885). Additional bla_OXA-1, bla_OXA-129, bla_OXA-488, and bla_PDC-35 β-lactamase genes were detected (Table 2). KPC-31, a D179Y variant of KPC-3, was previously described in CZA-resistant K. pneumoniae clinical isolates (13–15) but not in P. aeruginosa. To the best of our knowledge, this would be the first identification of KPC-31 in P. aeruginosa having the codifying gene integrated into the chromosome. The D179Y mutation is part of the Ω-loop (163 to 179) that surrounds the active site of KPC, and this modification has been associated with a reduced hydrolytic capacity against carbapenems, causing negative assays for phenotypic detection of carbapenemases (15). Additionally, this D179Y mutation on KPC affects the detection by lateral flow immune assays as reported (15). bla_KPC-31 was located in the Tn4401a transposon, flanked by both ATTGA target site duplications. A 19,318-bp fragment, including the Tn4401a (9,907 bp) showed >99% hit with the pGMI16-005_01 plasmid from K. pneumoniae CFSAN054110 (accession number CP028181) (Fig. 1). Flanking regions of this 19,318-bp fragment match with P. aeruginosa genomes, suggesting acquisition through a no-insertion element (IS)-mediated recombination. Additional intrinsic and acquired resistance mechanisms detected are listed on Table 2. Nine amino acid modifications plus a glutamine-67 deletion were observed on OprD porin compared with P. aeruginosa PAO1, previously described in carbapenem-resistant P. aeruginosa clinical isolates (Table 2) (16). Additional amino acid changes on MexR, NalC, NalD, and PBP3a were also detected, while PBP2 and PBP3 showed identical amino acid sequences compared to P. aeruginosa PAO1 (Table 2). Natural polymorphism on these genes was previously reported; therefore, these changes could not directly associate to resistance phenotypes (17). Several mutations in the quinolone resistance-determining regions (QRDR) were observed as follows: GyrA (T83I, D87N), GyrB (V539L), ParC (S87L), and ParE (D533E) (Table 2).

TABLE 2.

Antimicrobial resistance mechanisms, virulence factors, and multilocus sequence typing of P. aeruginosa M27432 by whole-genome sequencing

Parameter	Description
Acquired resistance
β-lactamase genes	blaKPC-31; blaOXA-1; blaOXA-129; blaOXA-488; blaPDC-35
Aminoglycoside-modifying enzyme-coding genes	aadA6; aph(6)-Id; aph(3′)-Iib; aph(3′')-Ib; aac(6′)-Ib3; ant(2′')-Ia; aac(6′)-Ib-cr
Quinolones	aac(6′)-Ib-cr; qnrB2
Phenicol	catB3; catB7; cmx
Fosfomycin	fosA
Trimethoprim	dfrA5; dfrA25
Sulfonamide	sul1
Macrolide	mph(A)
Rifampicin	arr-3
Others	qacE
Nucleotide differences compared with P. aeruginosa PAO1
AmpD	N111S; G148A
OprD	Q67-Deletion; T103S; K115T; F170L; E185Q; P186G; V189T; R310E; A315G; and G425A
MexR	V126E
NalC	G71E; E153Q; S209R
NalD	T11N
MexX	A30T; K329Q; L331V; W358R
MexY	T543A
DnaK	None
PBP2	None
PBP3 (ftsI)	None
PBP3a	A104P; V466I
PBP4	None
Quinolone resistance-determining region	GyrA (T83I; D87N); GyrB (V539L); ParC (S87L); ParE (D533E)
Virulence factor
Virulence factor genes	toxA; exoS; exoT; exoU; exoY; plcH; plcN; phzM; phzS; lasA; lasB; aprA; pvdA; algD
Serotype	O11
Multilocus sequence type
Sequence type	235

FIG 1.

Chromosomal location of bla_KPC-31 in P. aeruginosa M27432. Comparison with plasmid pGM116-005_01 from K. pneumoniae DFSAN054110 and the chromosome of P. aeruginosa AR_0357 suggests a plasmidic origin of an ~19-kb fragment containing bla_KPC-31. The light blue regions between DNA sequences indicate nucleotide identity >99 by BLASTn. Arrows indicate predicted open reading frames (ORFs). Red arrows represent bla_KPC gene variants.

Multilocus sequence typing (MLST) analysis revealed that the isolate belonged to the high-risk ST235 P. aeruginosa clone (3). T83I at GyrA and S87L at ParC are two QRDR mutations highly conserved on the ST235 P. aeruginosa clone (3).

We described a KPC variant, KPC-31, in a CZA-resistant P. aeruginosa isolate that arises after 130 days of CZA exposure. KPC-31 was previously described in Enterobacterales in several hospitals from Argentina but with no apparent epidemiological link with the patient included in this study (18). CZA resistance caused by KPC or GES variants in P. aeruginosa isolates were recently described, showing that exposure to CZA in P. aeruginosa could lead to specific mutations on acquired resistance genes like bla_KPC or bla_GES genes, leading to resistance to CZA and restoring susceptibility to carbapenems (5, 19).

CZA resistance is of concern for public health. The emergence of this resistance through a KPC variant in a ST235 P. aeruginosa high-risk clone represents a challenge for antimicrobial management. Notably, the most common carbapenemase detection methods were unable to detect KPC-31 in this isolate. However, isolates with high-level resistance to CZA and ceftolozane-tazobactam should be suspicious of producing this kind of carbapenemases and have to be considered for further studies. Thus, we advocate routine CZA susceptibility testing of P. aeruginosa.