Introduction

Relebactam is a novel non-β-lactam, bicyclic diazabicyclooctane β-lactamase inhibitor that is structurally related to avibactam, differing by the addition of a piperidine ring to the 2-position carbonyl group.

Relebactam has been combined with imipenem-cilastatin, to restore imipenem’s clinical activity against KPC-producing K. pneumoniae, other KPC-producing Enterobacteriaceae, and Pseudomonas aeruginosa demonstrating carbapenem resistance due to impermeability combined with AmpC expression [1, 2]. Both inhibitors display activity against class A β-lactamases, including KPC-type carbapenemases and class C β-lactamases (AmpC), but unlike avibactam, relebactam does not consistently inhibit isolates producing OXA-48 carbapenemases [2].

Unfortunately, ceftazidime-avibactam resistance has emerged [3, 4]. Little is known about the potential for relebactam to select for resistance; however, inactivation of the porin OmpK36 in K. pneumoniae has been reported to confer resistance to both imipenem-relebactam and meropenem-vaborbactam [5].

In this study, we determined the activity of imipenem combined with relebactam against K. pneumoniae–producing non-MBL carbapenemases isolated from hospitalized patients in 18 hospitals in six Greek cities [6, 7]. Isolates with elevated imipenem-relebactam MIC were further evaluated.

Material and methods

Clinical isolates

Carbapenem non-susceptible Κ. pneumoniae isolates, producing a non-MBL carbapenemase, were selected from two previous studies. The isolates were collected from unique patients hospitalized between November 2014 and December 2016 in 18 hospitals from six cities in Greece [6, 7].

Species identification and primary MIC determination was performed by Vitek®2 (bioMérieux, Marcy-l’Etoile, France).

Antimicrobial susceptibility testing

Susceptibility testing of imipenem with and without relebactam (4 μg/ml) (Merck & Co., Inc., New Jersey, USA), meropenem (Sigma-Aldrich, St. Louis, MO, USA), doripenem (Sigma-Aldrich), gentamicin (AppliChem, GmbH, Darmstadt, Germany), and colistin (Sigma-Aldrich) was performed by standard broth microdilution, according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [8].

Susceptibility testing of ceftazidime-avibactam, fosfomycin, and tigecycline was performed by Liofilchem® MIC Test Strips (MTS) (Liofilchem srl, Roseto degli Abruzzi, Italy), according to manufacturer’s instructions. Antimicrobial susceptibility results were interpreted according to the European Committee on Antimicrobial Susceptibility Testing recommendations (EUCAST 2018, version 8.0) [9]. As susceptibility breakpoints for the combination of imipenem-relebactam are not established by EUCAST or CLSI, imipenem breakpoints were applied. Further, in KPC-producing isolates with elevated MICs to imipenem-relebactam (> 2 mg/L), we determined the MIC to meropenem-vaborbactam by MTS. Escherichia coli ATCC 25922, K. pneumoniae ATCC 700603, and P. aeruginosa ATCC27853 were used as quality control strains. All isolates were stored at − 80 °C and were sub-cultured twice before testing.

Screening for β-lactamase genes and molecular typing

All K. pneumoniae isolates were screened for genes encoding metallo- and serine β-lactamases by PCR as described previously [6]. All isolates with imipenem-relebactam MIC > 2 mg/L were tested further for the presence of genes encoding ESBLs and acquired AmpC β-lactamases [10].

Genetic relatedness of K. pneumoniae isolates was evaluated by pulsed field gel electrophoresis (PFGE) analysis as previously described [6].

Whole genome sequencing analysis of selective isolates

Total DNA was extracted and purified using the PureLink® Genomic DNA mini kit (Life Technologies, Invitrogen™, Carlsbad, CA, USA) following manufacturer’s Gram Negative Bacterial Cell DNA extraction protocol and whole genome sequencing (WGS) data were generated on Ion Torrent™ platform (Life Technologies). Analyses of WGS were performed by tools provided by the Center for Genomic Epidemiology (CGE) (http://cge.cbs.dtu.dk/services/all.php). Capsule and lipopolysaccharide serotype were predicted by Kaptive Web tool (http://kaptive.holtlab.net/). Selected genes were also aligned to reference genes with the NCBI BLAST algorithm (www.ncbi.nlm.nih.gov/BLAST).

Transferability of blaKPC genes

Conjugation and transformation experiments were performed, as previously described [11]. Transconjugant/transformant selection was performed in media containing rifampicin and ceftazidime or ceftazidime alone at 16 mg/L respectively.

Results

A total of 314 Κ. pneumoniae isolates with imipenem and/or meropenem MIC > 2 mg/L were tested. Isolates were obtained from urine (n = 128), blood (n = 62), specimens of lower respiratory tract (n = 53), pus (n = 40), CSF (n = 2), and other sites (n = 29).

All isolates were non-susceptible to imipenem (MICs > 2 mg/L) and doripenem (MICs > 1 mg/L), and 313 (99.7%) were non-susceptible to meropenem (MICs > 2 mg/L). Two hundred ninety-five were KPC-producing, and 19 OXA-48-producing K. pneumoniae isolates. MIC50 and MIC90 against all isolates were 32 and > 64 mg/L, respectively, for imipenem and 0.5 and 2 mg/L for imipenem-relebactam. The addition of relebactam improved imipenem susceptibility rates from 0 to 92.7% (Table 1).

Table 1 MIC and cumulative percent inhibited distributions for imipenem-relebactam and comparators, in relation to the carbapenemase type produced by the 314 K. pneumoniae isolates

KPC-producing isolates

For 295 isolates that harbored blaKPC, the imipenem MIC50 and MIC90 values were 32 and > 64 mg/L, respectively. The addition of relebactam reduced imipenem MICs by a median of 128-fold (range 8- to 512-fold), decreased the MIC50 and MIC90 values to 0.25 and 1 mg/L, respectively, and improved imipenem susceptibility rates from 0 to 98.0% (Table 1).

PFGE genotyping revealed a multiclonal population of KPC-producing K. pneumoniae, with a prevalent PFGE profile (with ten variants) comprising 147 (49.8%) of the isolates, which was detected in 18 centers. Additionally, two PFGE profiles, consisting of three and two variants each, comprised 20 (6.8%) and 18 (6.1%) isolates, respectively. Twenty-four more PFGE profiles were identified that included 2–10 isolates each.

KPC-producing K. pneumoniae non-susceptible to imipenem-relebactam

In six isolates (2.0%), all with imipenem MICs ≥ 128 mg/L, the addition of 4 mg/L relebactam reduced imipenem MIC to 4 mg/L, which corresponds to intermediate imipenem resistance, according to the EUCAST breakpoints (Table 2).

Table 2 Antimicrobial susceptibility and genotype of the imipenem-relebactam non-susceptible clinical KPC-producing K. pneumoniae isolates and their transformant/transconjugant laboratory strains

WGS analysis determined that the isolates belonged to 3 different sequence types (STs): ST258 (n = 3), ST147 (n = 2), and ST15 (n = 1) (Table 2). The three ST258 strains were isolated in different hospitals, in North Greece, Attica, and South Greece. Only the two ST147 isolates were epidemiologically related as they shared the same PFGE pattern and were isolated in the same hospital. Five isolates harbored blaKPC-2 and one blaKPC-23 (Table 2).

The carbapenemase gene was detected on identical Tn4401a transposons, located between ISKpn7 and ISKpn6, while no mutations known to contribute to the expression of blaKPC were evident within promoter sequences P1 and P2 in any isolate. All isolates harbored additional β-lactamase genes (Table 2).

Five isolates harbored a non-functional OmpK35. Isolates belonging to ST258 had a truncated protein at codon 89, while the isolates of ST147 had a truncated protein at codon 173. In the ST15 isolate, a wild-type OmpK35 (GenBank accession number WP_004141771.1) was detected.

OmpK36 of the ST258 isolates matched to the OmpK36_v1 (GenBank accession number WP_002913005.1) were full length with only that of KP-128, to harbor a duplication of amino acids G134–D135, located at the conserved loop L3. This insertion contributes to high-level resistance to carbapenems [12]. ST147 isolates harbored OmpK36_v2 (GenBank accession number WP_099130593.1), while the ST15 isolate harbored the OmpK36_v1 variant with one amino acid substitution (A323P) at the B14 β-strand of the protein. No insertion sequences were identified upstream of ompK36 in any isolate.

Susceptibility testing of E. coli TOP10 transformants and RC85-harboring blaKPC-2- or blaKPC-23-carrying plasmids is shown in Table 2. In all isogenic TOP10 transformants and in RC85/pl105, the addition of 4 mg/L relebactam reduced imipenem MIC from 8 to 16 mg/L to 0.25–0.5 mg/L suggesting that chromosomal genes have contributed to the high imipenem and elevated imipenem-relebactam MICs in parent isolates.

OXA-48-like–producing isolates

No significant difference was observed in imipenem-relebactam vs. imipenem MICs against OXA-48-like carbapenemase producers (median 4 and 8 mg/L, respectively) (Table 1). The addition of relebactam restored the activity of imipenem in only two of the 19 OXA-48-like–producing isolates. All 19 isolates harbored additional β-lactamase genes (Table 3).

Table 3 Antimicrobial susceptibility and genotype of the OXA-48–producing K. pneumoniae isolates

The OXA-48-like–producing K. pneumoniae isolates were multiclonal (eight PFGE clones). Different clones between hospitals and within hospitals were observed.

WGS analysis was performed in isolate KP-205 co-producing KPC-2 and OXA-48, as it was the isolate exhibiting the highest imipenem-relebactam MIC (32 mg/L). The isolate was assigned to ST716, an infrequently reported clone of capsular type KL110 (wzc 941; wzi 89) [13]. blaKPC-2 was detected on a Tn4401a, while blaOXA-48 on a Tn1999 transposon. The isolate harbored additional acquired bla genes coding for CTX-M-15, VEB-1, OXA-10, TEM1B, and OXA-1 and had intact OmpK35 and OmpK36 proteins, with OmpK36 matching to the OmpK36_v6 (GenBank accession number WP_004149145.1).

Comparator antimicrobial agents

Of the comparator agents, only ceftazidime-avibactam showed in vitro activity comparable to that of imipenem-relebactam (S, 99.7%). Ceftazidime-avibactam was active against all but one KPC-producing isolate and against all 19 isolates with OXA-48-like carbapenemases. These isolates were ceftazidime-resistant due to CTX-M-type ESBL co-production and the addition of avibactam reduced MICs to 0.5–2 mg/L. The isolate that was resistant to ceftazidime-avibactam (KP-90; MIC, 16 mg/L) produced KPC-23 was highly resistant to ceftazidime (MIC, > 1024 mg/L) and imipenem (MIC, 512 mg/L) and had an MIC of 4 mg/L to imipenem-relebactam (intermediate resistance to imipenem).

Of the other agents tested, gentamicin was the most active in vitro, with 61.5% of the isolates to be inhibited at ≤ 2 mg/L, followed by fosfomycin (S, 59.8%) and colistin (S, 59.6%). Finally, tigecycline demonstrated 49.4% susceptibility rate.

Discussion

Imipenem-relebactam exhibited potent activity, against KPC-producing isolates. Relebactam restored imipenem susceptibility in 98.0% of the KPC isolates. This is in accordance with other studies reporting that relebactam restored imipenem susceptibility in up to 98% of Enterobacteriaceae-producing KPC [2, 14].

In our collection of KPC-producing K. pneumoniae isolates, six (2.0%) exhibited high MIC to imipenem (≥ 128 mg/L) and elevated MIC to imipenem-relebactam (4 mg/L). These isolates exhibited various defects in OmpK35 or/and OmpK36. As carbapenems cross the outer membrane using both the OmpK35 and OmpK36 porins, the inactivation of OmpK35 or/and OmpK36 shown in those isolates likely contributed to the high carbapenem MICs observed. The resulting outer membrane porin loss in combination with a carbapenemase has been reported to confer high-level resistance to carbapenems [15]. Downregulation of OmpK36 or major mutations in this porin has been reported to increase imipenem-relebactam MIC up to 8 mg/L [16, 17]. Recently, Balabanian et al. reported that disruption of either ompK35 or ompK36 alone did not affect the MIC of imipenem-relebactam. However, when both were disrupted, the MIC of imipenem-relebactam increased to 2–512 mg/L depending on the expression of blaKPC [18]. In our study, disrupted OmpK35 was present in five isolates with elevated imipenem-relebactam MIC, and in four of them was the only porin deficiency identified, suggesting that the role of OmpK35 requires further study. Yet, the level of expression of additional β-lactamase genes, which were not studied further, may also have contributed to increased imipenem-relebactam MICs.

Additionally, our results revealed the association of the GD duplication within the L3 internal loop of OmpK36, with reduced susceptibility to meropenem/vaborbactam, [19], but not to imipenem-relebactam. OmpK36, which has a smaller channel than OmpK35, appears to play a more significant role in the passage of vaborbactam across the outer membrane than OmpK35 [19].

Susceptibility studies of isogenic isolates harboring plasmids from the six isolates revealed that chromosomal factors impacted imipenem-relebactam susceptibilities and contributed to elevated MICs.

A limitation of our study is that the in vitro activity of imipenem-relebactam was not evaluated based on approved susceptibility breakpoints as those have not yet been determined. Rather, we evaluated the in vitro activity of the combination based on the reduction of imipenem MICs to the susceptible range when relebactam was added. Additionally, we have not studied OmpK35 and OmpK36 of all KPC isolates to elucidate the impact of mutations in one or both proteins to the level of imipenem-relebactam MICs. Further, functional studies are needed to define the impact of porin gene disruptions on imipenem-relebactam MICs.

In summary, relebactam restored the in vitro activity of imipenem against a large collection of clinical blaKPC-harboring K. pneumoniae isolates from a country where carbapenemase producers are endemic, suggesting that it is a promising option for difficult-to-treat infections by these bacteria. Production of KPC carbapenemase represents the main cause of carbapenem resistance among K. pneumoniae in Greek hospitals (66.5%) [6] and this carbapenemase appears to be very well inhibited by relebactam. As the global spread of KPC carbapenemases poses a therapeutic challenge for clinicians, the novel combination of imipenem with relebactam is a welcome advancement.