Activity of imipenem/relebactam and comparators against KPC-producing Klebsiella pneumoniae and imipenem-resistant Pseudomonas aeruginosa

Purpose Relebactam is a novel β-lactamase inhibitor, which, when combined with imipenem/cilastatin, is active against both class A and class C β-lactamases. To evaluate in vitro antimicrobial activity of imipenem/relebactam against a collection of recent clinical isolates of carbapenem-non-susceptible P. aeruginosa and K. pneumoniae ST258 and ST512 KPC producers belonging to different lineages from hospitals in Southern Spain. Methods Six hundred and seventy-eight isolates were tested: 265 K. pneumoniae (230 ST512/KPC-3 and 35 ST258/KPC-3) and 413 carbapenem-non-susceptible P. aeruginosa. Imipenem, piperacillin/tazobactam, ceftazidime, cefepime, aztreonam, ceftolozane/tazobactam, meropenem, amikacin, ciprofloxacin, colistin, and ceftazidime/avibactam were used as comparators against P. aeruginosa. Against K. pneumoniae ceftazidime, cefepime, aztreonam, and ceftolozane/tazobactam were not tested, and tigecycline was studied instead. MICs were determined in duplicate by broth microdilution according to EUCAST guidelines. Results Imipenem/relebactam displayed potent in vitro activity against both sequence types of KPC-3-producing K. pneumoniae. MIC50 and MIC90 values were 0.25 mg/L and 1 mg/L, respectively, with percent of susceptible isolates >97%. Only three K. pneumoniae ST512/KPC-3 isolates and one ST258/KPC-3 were resistant to imipenem/relebactam. Relebactam sensitized 98.5% of K. pneumoniae isolates resistant to imipenem. The activity of imipenem/relebactam against P. aeruginosa was moderate (susceptibility rate: 62.7%). Analysis of the acquired and mutational resistome of isolates with high levels of resistance to imipenem/relebactam has not shown a clear association between them. Conclusion Imipenem/relebactam showed excellent activity against K. pneumoniae KPC-3. The activity of imipenem/relebactam against imipenem-resistant P. aeruginosa was moderate. Supplementary Information The online version contains supplementary material available at 10.1007/s10096-023-04735-1.


Introduction
Carbapenem-resistant Gram-negative bacilli (CRGNB) are a major public health problem, recently described by WHO as a global crisis [1].Since nosocomial and healthcare-associated infections caused by CRGNB organisms significantly increase morbidity and mortality, length of hospital stay, and medical costs [2], the development of new antimicrobials or new combinations of β-lactam-β-lactamase inhibitors active against these pathogens has become a priority [3,4].Pseudomonas aeruginosa and Klebsiella pneumoniae are common pathogens in hospitals, have a high propensity to develop antibiotic resistance, and also a high capacity for dissemination in the nosocomial environment [5][6][7].P. aeruginosa carbapenem resistance is driven by different resistance mechanisms, which often act synergistically.The most common mechanisms of imipenem resistance in this microorganism include repression or inactivation of the carbapenem porin OprD coupled with hyperexpression of the chromosomal cephalosporinase AmpC and/or overexpression of efflux pumps, as well as carbapenemase production [5,8,9].In K. pneumoniae, the most frequent resistance mechanism to carbapenem is carbapenemase production, mainly of classes A (KPC), B (MBL), and D (OXA-48-like) β-lactamases [10].In 2017, the WHO published a priority list of pathogens for which the development of new antibiotics was urgently required.Carbapenem-resistant Enterobacterales are included on this list, as well as carbapenem-resistant P. aeruginosa and carbapenem-resistant Acinetobacter baumannii [11].
Relebactam is a novel diazabicyclooctane β-lactamase inhibitor, which in combination with imipenem/cilastatin is active against class A and class C β-lactamase-producing microorganisms [12,13].Imipenem/relebactam was approved by the EMA and FDA in 2020 for the treatment of complicated urinary tract infections, complicated intraabdominal infections, and hospital-acquired and ventilatorassociated bacterial pneumonia in adult patients with limited or no alternative therapeutic options [14,15].
The purpose of this study was to provide data on the comparative in vitro antimicrobial activity of imipenem/ relebactam against a collection of recent clinical isolates of carbapenem-non-susceptible P. aeruginosa and K. pneumoniae ST258 and ST512 KPC producers belonging to different lineages from hospitals in Southern Spain.

Bacterial identification, molecular epidemiology, and genomic characterization
Identification of the isolates was confirmed in the reference laboratory by MALDI-TOF MS (MALDI-TOF Biotyper 3.1; Microflex Bruker, Madrid, Spain).
PFGE analysis of XbaI (Enterobacterales)-and SpeI (P.aeruginosa)-digested DNA was used to determine the degree of genetic relatedness between isolates.Isolates differing by one or more bands in PFGE assays were assigned to different pulsotypes.A dendrogram was created with Bionumerics 8.0 software (BioMérieux), using the Dice coefficient with optimization set at 1% and position tolerance at 1.2% (data not shown).
In-house Miseq sequencing (Illumina, San Diego, CA, USA) was performed on one isolate of each K. pneumoniae pulsetype and in resistant imipenem/relebactam P. aeruginosa with MIC >4 mg/L.Libraries were prepared with the Nextera XT DNA library preparation kit (Illumina) and sequencing with a reagent cartridge, V3 600 cycles (Illumina).CLC Genomic Workbench software (Qiagen, Netherlands) was used for de novo assembly of Illumina reads.Genomes were analyzed in the Center for Genomic Epidemiology resistance and MLST databases from https:// www.genom icepi demio logy.org/.All pulsotypes assigned to the same MLST were considered to belong to the same clone.Whenever possible, isolates from the same clone with different pulsotypes were selected.
The total antimicrobial resistance gene content of the K. pneumoniae sequenced was analyzed in silico using ResFinder v4.1 (https:// cge.food.dtu.dk/ servi ces/ ResFi nder/) and the Comprehensive Antibiotic Resistance Database (CARD) (https:// card.mcmas ter.ca/).K. pneumoniae ATCC10031 was used as a reference to compare porin, PBPs, and efflux bomb aminoacid sequences.In addition, one susceptible isolate was selected for each MIC value for the analysis of mutations related to ß-lactam resistance.

Antimicrobial activity against KPC-producing K. pneumoniae
Figure 1 shows the MIC distributions of imipenem alone or in combination with relebactam for all K. pneumoniae included in the study.The susceptibility rate to imipenem/ relebactam was 98.5% (three K. pneumoniae ST512/KPC-3 producers and one K. pneumoniae ST258/KPC-3 were resistant).With respect to imipenem, only 1.5% of isolates were susceptible.Overall, relebactam sensitized 98.5% of imipenem-resistant isolates (256/260), improving the activity of imipenem by 4 to 8 two-fold dilutions.The MIC range, MIC 50 , MIC 90 values and percentages of susceptibility and resistance for all isolates are shown in Table 1.
In general, all antimicrobials tested showed similar percentages of susceptibility regardless of MLST, except for colistin, with ST258 isolates being more susceptible than ST512 isolates (88.6% vs 39.7%; p < 0.01).All isolates were also resistant to ciprofloxacin and tigecycline.
Analysis of chromosomal mutations in the porin genes and PBPs showed no differences between isolates susceptible and resistant to imipenem/relebactam (Table 2).All isolates had wild-type OmpK36 and AcrAB-TolC regulator (RamR).Mutations in OmpK35 and OmpK37 were detected in two isolates resistant to imipenem/relebactam, although the same mutations were detected in susceptible isolates.
Among the colistin-resistant isolates, 51.8% were found to be susceptible to ceftolozane/tazobactam and ceftazidime/ avibactam, and up to 58.9% were susceptible to imipenem/ relebactam.On the other hand, one-third (32.4%) of the amikacin-resistant isolates were susceptible to relebactam activity.
Regarding WGS results, of the 51 isolates sequenced, only 46 could be analyzed due to technical problems.MLST analysis showed a high variability in our collection.Twentysix different sequences were detected, with two major ST: ST175 (n = 9) and ST274 (n = 4).The remaining STs were represented by 3 or less isolates, and one isolate could not be characterized by this method.The 88.8% of isolates from ST175 had an imipenem/relebactam MIC = 8 mg/L, but among isolates with this MIC value, up to 20 different STs were detected, so there does not show a clear association between clone and MIC in isolates with this MIC value.
The analysis of genes involved in ß-lactam resistance is shown in the supplementary table.With respect to chromosomal ß-lactamases analysis: 19 OXA-50 variants were detected among the isolates, with the wild-type variant being the predominant one (n = 12), followed by the OXA-488 (n = 6) and OXA-486 (n = 4) variants, and 18 AmpC variants (PDC-type).The most frequent PDC variants were PDC-1 (n = 12) and PDC-3 (n = 5).The most frequent polymorphism found in the PDC variants was T105A (71.7%), detected in all PDC alleles except in PDC-1.In general, an association between OXA-50 and PDC variants and ST was detected.The presence of acquired ß-lactamases was detected in only 11 isolates, mostly OXA-10 or OXA-10-like (n = 8).

Discussion
One of the results of the increase in MDR-GNB infections worldwide is that the approved antimicrobials provide few treatment options for systemic infections.There is an urgent need for new antimicrobials active against MDR-GNB, as well as sufficient information to facilitate their use in severe infections.
The collection of bacterial strains selected for this study includes KPC-3-producing K. pneumoniae isolates from the two most prevalent clones worldwide, as well as the most representative imipenem-resistant non-MBL producer P. aeruginosa isolates, both of which cause  The results obtained in the current study showed that the in vitro activity of imipenem/relebactam was superior to that of comparators against recent high-risk clone isolates of K. pneumoniae KPC-3 producers.Imipenem/relebactam showed potent antimicrobial activity, with MIC 90 values of ≤1 mg/L against K. pneumoniae.MIC 90 values showed no differences according to the ST tested, as in previous studies [26].
To date, a small number of imipenem/relebactam-resistant K. pneumoniae KPC-3 producers have been reported.
In our study, 98.5% of KPC-3-producing K. pneumoniae were susceptible to imipenem/relebactam.Our results are consistent with those of previous studies.Hernández-García et al. evaluated the in vitro activity of imipenem/ relebactam against 14 K. pneumoniae KPC-3 producers, all of which were susceptible to imipenem/relebactam [27].Galani et al. analyzed imipenem/relebactam activity against 314 non-MBL carbapenemase-producing K. pneumoniae.Among KPC-producing isolates, 98% were inhibited by this combination, and relebactam effectively restored the in vitro activity of imipenem, with MIC 50 and MIC 90 values decreasing from 32/4 to 0.25/4 mg/L, and from >64/4 to 1/4 mg/L, respectively [28].In a recent study in Spain, 91 KPC-producing isolates were analyzed.The percentage of susceptibility to imipenem was 15.5%, and 100% of the isolates were susceptible to imipenem/relebactam [29].Ceftazidime/avibactam has been positioned as an alternative for the treatment of infections caused by high-risk clones of KPC-producing K. pneumoniae, although the emergence of KPC enzyme variants resistant to this combination has been described, mainly selected after exposure during treatment [30,31].Our results agree with those obtained in previous studies, which show that imipenem/relebactam has excellent in vitro activity and clinical efficacy against KPC-producing isolates, even against variants resistant to ceftazidime/avibactam [27,29,32].In our collection, 2.3% of isolates were resistant to ceftazidime/avibactam and all of them were susceptible to imipenem/relebactam.Moreover, MIC 50 /MIC 90 values were significantly lower than those of ceftazidime/avibactam.Vázquez-Ucha et al. reported similar results, with MIC 50 /MIC 90 values for imipenem/relebactam and ceftazidime/avibactam of ≤0.25/1 mg/L and 1/8 mg/L, respectively [29].
Several studies have previously shown that reduced porin expression decreases the in vitro activity of imipenem/relebactam.Imipenem/relebactam resistance has been associated with mutations resulting in non-functional OmpK35 and OmpK36 porins in KPC-producing K. pneumoniae strains [28,33,34].In our case, we detected four resistant isolates but did not find the mutations associated with resistance to imipenem/relebactam.Since the mutations detected in the porin genes were also present in isolates susceptible to imipenem/relebactam, they cannot explain the resistance to imipenem/relebactam in these four isolates.To our knowledge, there is still very limited data on the clinical efficacy of imipenem/relebactam in patients with severe infections caused by carbapenemase-producing Enterobacterales.The results of the RESTORE-IMI 1 and RESTORE-IMI 2 clinical trials, evaluating the clinical efficacy of imipenem/relebactam for the treatment of infections caused by imipenem-non-susceptible isolates, as well as for treatment of hospital-acquired/ventilator-associated bacterial pneumonia, concluded that imipenem/relebactam was an appropriate treatment option.It should be noted however that the number of carbapenemase-producing isolates was very low [14,15].
With respect to P. aeruginosa, we analyzed a large number of imipenem-resistant non-MBL producer isolates.In our P. aeruginosa collection, a susceptibility rate of 62.7% was detected for imipenem/relebactam.Previous studies have reported similar results.The SUPERIOR and STEP studies found a susceptibility rate of 75.7% [35].Zhang et al. analyzed a collection of 835 non-imipenem-susceptible P. aeruginosa isolates from the global SMART surveillance program, and the susceptibility rates to imipenem/relebactam were 64.4%, and the MIC 50 and MIC 90 values were 2/4 mg/L and >32/4 mg/L, respectively.Compared with our data, the susceptibility percentages were very similar, but the MIC 90 value was 2-fold higher [36].In our study, imipenem/   B).Overall, the MIC value of imipenem/relebactam against 32% of isolates was >4/4 mg/L.This rate was similar to that observed in our study (37.3%), although none of our isolates was a carbapenemase producer [37].
According to the results of our study, approximately one-third of isolates resistant to ceftazidime/avibactam and ceftolozane/tazobactam remain susceptible to imipenem/relebactam.These results are consistent with those previously described in other series [38].
WGS-based analyses of imipenem/relebactam-resistant P. aeruginosa isolates show the presence of several acquired OXA-type ß-lactamases, but they generally occur at low prevalence and do not appear to be responsible for the moderate imipenem/relebactam resistance observed among these isolates.Regarding the variants of the chromosomal ß-lactamases found (PDC-type and OXA-50-type), an association is generally observed between the variant detected and the ST rather than with the imipenem/relebactam MIC values obtained, evidencing that other genetic elements should be implicated in resistance to this combined anitibiotic.These data are in agreement with some published studies that showed no significant relationship between acquired OXAtype ß-lactamases or AmpC variants and resistance to imipenem/relebactam [37,39].This was also noted by Young et al., who describe that in a collection of 2691 isolates, they found no relationship between imipenem/relebactam MIC and PDC alleles detected in their collection, as well as no association between specific alleles and MIC values [37].However, the most prevalent AmpC polymorphism found among our imipenem/relebactam-resistant isolates was T105A, previously associated with imipenem increased resistance [40], which was found in all PDC variant detected with the exception of PDC-1.Furthermore, some of the genes widely reported as AmpC regulators were among those genes with high polymorphism prevalence, suggesting that the over-expression of PDC variants with T105A, alone or combined with other polymorphisms, could be relevant for imipenem/relebactam resistance.Our results also showed that the resistance mechanisms with the highest prevalence of polymorphisms among these isolates were detected in the genes related to the MexXY-OprM pumping system and the OprD porin, which is concordant with those described in other studies.Fraile-Ribot et al. reported that resistance to imipenem/relebactam appears to be very low in non-MBL-producing P. aeruginosa clinical isolates and isogenic laboratory strains with β-lactam resistance mechanisms that include combinations of OprD inactivation and overexpression of AmpC β-lactamase and/or efflux pumps [38].
In addition, some of the polymorphisms found in our collection had a prevalence of more than 90%.Among ß-lactams resistance-related genes with high prevalence polymorphisms, the MexXY efflux pump system stands out especially, as several of these polymorphisms were observed in both structural components (MexX: K329Q and W358R; MexY: T543A), which could be increasing the affinity of this efflux pump for imipenem or relebactam [39,41], and in ArmZ regulator (L88P and V243A), which could lead to over-expression of MexXY [42][43][44].Moreover, the majority of the isolates in our collection presented the polymorphism (I106V) in chromosomal imipenemase Prevalence >90% is highlighted in bold V variability (number of polymorphisms found per gene/number of total of polymorphisms), Pv prevalence (number of isolates that contains polymorphisms per gene/total isolates), A aminoglycosides, B non-carbapenem beta-lactams, C carbapenems, Q quinolones, P polymyxins, IMR imipenem/relebactam † PAO1 was used as reference genome PIB-1, which could be implicated in the increased activity of this enzyme or with a loss of inhibition by relebactam [45].To confirm this implication, further studies should be necessary.Highly prevalent polymorphisms have also been found in regulators of the MexEF-OprN efflux pump system (MexS: D249N; MexT: Q80fs and F172I), whose relationship with imipenem/relebactam resistance could be more associated with decreased expression of OprD than with over-expression of the MexEF-OprN system itself, as there is no clear evidence of ß-lactam efflux through this RND system [46,47].Finally, other polymorphisms were also found in NalC (negative regulator of MexAB-OprM) [48], in ParS (involved in lipopolysaccharide modification and overexpression of some RND efflux pump systems) [49], and in PonA (encoding for PBP1A) [50], all of them with potential involvement in ß-lactam resistance [51], and thus imipenem/relebactam resistance.
Our study has some strengths and limitations.The main strength of this study is that the collection reflects the local epidemiology of a large and specific geographical area.One of the limitations of this study is the absence of WGS data in imipenem/relebactam-susceptible isolates of P. aeruginosa, so the prevalence of the polymorphisms among these isolates is unknown.However, taking into account the genomic heterogeneity of the isolate collection analyzed, which includes a high heterogeneity of clones, it is probable that these polymorphisms are directly or indirectly related to imipenem/relebactam resistance in these isolates.
In conclusion, imipenem/relebactam showed excellent activity against K. pneumoniae KPC-3 isolates, including those resistant to ceftazidime/avibactam, regardless of sequence type.On the other hand, a moderate number of P. aeruginosa isolates were susceptible to imipenem/ relebactam and retained activity against some isolates resistant to ceftazidime/avibactam and ceftolozane/tazobactam.Therefore, this combination could be an option to consider in the treatment of infections caused by these microorganisms.

Table 1
50C range, MIC50, and MIC 90 of the tested antibiotics against K. pneumoniae KPC-3 by clone and resistance rates *EUCAST breakpoints for E. coli.1Forsusceptibility testing purposes, the concentrations of relebactam and avibactam were fixed at 4 mg/L

Table 3
50C range, MIC50, and MIC 90 of the tested antibiotics against non-carbapenemsusceptible P. aeruginosa and resistance rates1For susceptibility testing purposes, the concentrations of relebactam, tazobactam, and avibactam were fixed at 4 mg/L 2 I-susceptible, increased exposure Fig. 2 MIC distribution of imipenem and imipenem/relebactam against P. aeruginosa.Vertical dashed line represents the EUCAST resistance breakpoint for imipenem/relebactam (>2/4 mg/L for P. aeruginosa)

Table 4
Prevalence, function, and variability of mutated genes and polymorphisms detected in more than 50% of imipenem/relebactam-resistant P. aeruginosa isolates