In vitro activity of aztreonam in combination with newly developed β-lactamase inhibitors against MDR Enterobacterales and Pseudomonas aeruginosa producing metallo-β-lactamases

Abstract

Objectives

To evaluate the in vitro activity of aztreonam in combination with novel β-lactamase inhibitors, namely avibactam, nacubactam, taniborbactam and zidebactam, against MDR MBL-producing Enterobacterales and Pseudomonas aeruginosa clinical isolates.

Methods

MIC values of aztreonam, aztreonam/β-lactam inhibitor but also cefiderocol as comparator were determined for 64 and 39 clinical Enterobacterales or P. aeruginosa isolates, respectively, producing representative MBLs, i.e. derivatives of NDM (n = 64), VIM (n = 32), IMP (n = 8) and SPM (n = 2). MICs were also determined for Escherichia coli TOP10 and P. aeruginosa PAO1 harbouring recombinant plasmids producing the different β-lactamases under isogenic backgrounds (n = 22 and n = 11, respectively). Fifty percent inhibitory concentrations were additionally determined for the abovementioned β-lactamase inhibitors using β-lactamase crude extracts.

Results

The susceptibility rate for aztreonam was 17.1% among MBL-producing Enterobacterales, while it was very high with aztreonam/zidebactam (98.4%), and to a lower extent with aztreonam/nacubactam (84.4%) and aztreonam/taniborbactam (75%), compared with aztreonam/avibactam (70.3%) and cefiderocol (39.1%). Among MBL-producing P. aeruginosa isolates, the susceptibility rates were 53.8% with aztreonam, 66.7% with aztreonam/nacubactam and aztreonam/taniborbactam combinations, and 69.2% with aztreonam/avibactam, aztreonam/zidebactam and cefiderocol.

Conclusions

Altogether, these results showed that combinations including aztreonam and novel β-lactamase inhibitors, such as zidebactam, nacubactam or taniborbactam, have a very significant in vitro activity against MDR MBL-producing Enterobacterales clinical isolates, the aztreonam/zidebactam combination being the best option. On the other hand, aztreonam/zidebactam is equivalent to aztreonam/avibactam and cefiderocol among MBL-producing P. aeruginosa isolates.

Introduction

The worldwide spread of carbapenemases in Enterobacterales and Pseudomonas aeruginosa is one of the most rising concerns.¹ Class B β-lactamases (MBLs) such as NDM, VIM or IMP confer high-level resistance to all β-lactams including carbapenems, with the exception of aztreonam, which may be considered as a sparing β-lactam against MBL-producing P. aeruginosa strains.² However, aztreonam is often inactive against MDR isolates co-producing MBLs and other β-lactamases, such as ESBLs or AmpC β-lactamases (in case of overproduction), leading to very few therapeutic options.^2–5 Although aztreonam/avibactam combination and cefiderocol are being considered as interesting therapeutic options for treating infections caused by MDR Gram-negative bacteria co-producing MBLs and ESBLs,^6–8 recent studies showed that several clavulanic acid-inhibited ESBLs (PER, BEL, SHV) significantly reduce the susceptibility to cefiderocol.⁹ In combination with other mechanisms, including modification of the PBP3 sequence, production of CMY-type enzymes (and particularly of CMY-42), resistance to aztreonam/avibactam may be observed among MBL-producing Escherichia coli.¹⁰ Hence, therapeutic options for infections caused by MBL producers are still limited, and new therapies are needed.

Recently, novel β-lactamase inhibitors have been developed, such as zidebactam and nacubactam, classified as diazabicyclooctane (DBO) molecules,^11,12 similarly to the commercially available avibactam. Zidebactam and nacubactam possess interesting features since they exhibit a triple mechanism of action involving (i) selective and high-affinity Gram-negative PBP2 binding; (ii) an ‘enhancer’ mechanism potentiating the partner β-lactams in binding to PBP3; and (iii) a β-lactamase inhibitory activity. These combined actions allow an effective activity against Gram-negative bacteria producing MBLs, those latter enzymes being inhibited by DBOs.^13–16 Furthermore, taniborbactam belongs to boronic acid derivative agents, also corresponding to a new class of clinically relevant β-lactamase inhibitor, and additionally possesses β-lactamase inhibitory activity against serine-β-lactamases (KPC or OXA-48), but also against MBLs, including VIM and NDM enzymes (but not IMP).^17,18

The objective of our study was therefore to evaluate the in vitro activity of aztreonam in combination with the recently developed β-lactamase inhibitors zidebactam, nacubactam and taniborbactam compared with aztreonam/avibactam and cefiderocol against MDR MBL-producing Enterobacterales and P. aeruginosa clinical isolates.

Material and methods

Bacterial isolates

A collection of 83 MDR MBL-producing Enterobacterales and P. aeruginosa clinical isolates that had been collected by the Swiss National Reference Center for Emerging Antibiotic Resistance across all of Switzerland was included in the study. Clinical isolates included E. coli (n = 34), Klebsiella pneumoniae (n = 15), Enterobacter cloacae (n = 8), Citrobacter freundii (n = 4), Klebsiella aerogenes (n = 2), Serratia marcescens (n = 1) and P. aeruginosa (n = 39). They produced the main types of MBLs, namely NDM-like (n = 64), VIM-like (n = 32), IMP-like (n = 8), SPM-like (n = 2) and PME-like (n = 1). Most of them co-produced other β-lactamases, mainly represented by CTX-M-like (n = 38), OXA-type (n = 47), TEM-like (n = 28), CMY-like (n = 25) or SHV-like (n = 14) enzymes. All the clinical isolates have been characterized and selected for their β-lactamase content (Table S1, available as Supplementary data at JAC Online). In addition, a series of β-lactamase-encoding genes, without the original promoter or signal peptides, was cloned into the shuttle vector pUCp24 and expressed in E. coli TOP 10 or P. aeruginosa PAO1 recipient strains.¹⁹ Those β-lactamase genes were derivatives of bla_TEM, bla_SHV, bla_CTX-M, bla_SPM, bla_PER, bla_GES, bla_BEL, bla_OXY, bla_CMY, bla_KPC, bla_VIM, bla_NDM, bla_IMP and bla_OXA.

Susceptibility testing

The MIC values were determined in duplicate by the broth microdilution method according to EUCAST guidelines.²⁰ MICs of aztreonam were determined alone or in combination with a fixed concentration of zidebactam (4 mg/L) (aztreonam/zidebactam), nacubactam (4 mg/L) (aztreonam/nacubactam), avibactam (4 mg/L) (aztreonam/avibactam) and taniborbactam (4 mg/L) (aztreonam/taniborbactam). Interpretation was based on EUCAST breakpoints for aztreonam and cefiderocol, respectively.²⁰ Resistance to aztreonam/β-lactamase inhibitor combinations were defined by referring to the aztreonam breakpoints, corresponding to resistance at >4 mg/L for Enterobacterales and >16 mg/L for P. aeruginosa. Cefiderocol resistance was considered for MICs >2 mg/L for both Enterobacterales and P. aeruginosa. In order to further evaluate the contribution of zidebactam that possesses significant antibacterial activity on its own, MICs of aztreonam/zidebactam were also determined at a 1:1 ratio (further labelled aztreonam/zidebactam*), along with MICs of zidebactam alone.²¹

β-Lactamase activities

Cultures of E. coli TOP10 harbouring recombinant plasmids and therefore producing the different β-lactamases tested were grown overnight at 37°C in 50 mL of Luria broth with gentamicin (50 mg/L). The bacterial suspension was pelleted, resuspended in 10 mL of 100 mM phosphate buffer (pH 7) for E. coli TOP10 harbouring β-lactamases except for MBLs, or HEPES buffer (pH 7.5) for E. coli TOP10 producing MBLs, and then sonicated using a Vibra-cell™ 75186 sonicator (Thermo Fisher Scientific), followed by centrifugation for 1 h at 11 000 × g and 4°C. The presence of the β-lactamase was monitored using nitrocefin (200 μM). Kinetic measurements were performed at room temperature in 100 mM sodium phosphate (pH 7.0), or HEPES buffer (pH 7.5) supplemented with ZnSO₄ (5 μM) for MBLs using a UV/visible ULTROSPEC 2100 pro spectrophotometer (Amersham Biosciences, Buckinghamshire, UK). The following wavelengths and absorption coefficients were used: benzylpenicillin (232 nm/Λɛ = −1100 M⁻¹cm⁻¹), cefalotin (262 nm/Λɛ = −7960 M⁻¹cm⁻¹) and imipenem (297 nm/Λɛ = −9210 M⁻¹cm⁻¹).^22,23 IC₅₀ values were determined for clavulanic acid, tazobactam, avibactam, zidebactam, taniborbactam and nacubactam. Various concentrations of these inhibitors were pre-incubated with the crude extract of the enzyme for 3 min at room temperature to determine the concentrations that reduced the hydrolysis rate of 100 µM β-lactam by 50%. Results were expressed in micromolar units. The total protein content was measured by using a Bradford assay.²³

Results

Antimicrobial susceptibility to the different aztreonam/β-lactamase inhibitor combinations

Among all MBL-producing Enterobacterales clinical isolates tested here, only 17.2% of strains were susceptible to aztreonam. Among the different combinations, aztreonam/zidebactam was the most effective (98.4% MIC_{aztreonam/zidebactam},  ≤ 4/4 mg/L) followed by aztreonam/zidebactam* (96.9% MIC_{aztreonam/zidebactam*},  ≤ 4/4 mg/L) and aztreonam/nacubactam (84.4% MIC_{aztreonam/nacubactam},  ≤ 4/4 mg/L). The activity of aztreonam/taniborbactam (75% MIC_{aztreonam/taniborbactam}, ≤ 4/4 mg/L) was also higher than for aztreonam/avibactam (70.3% MIC_{aztreonam/avibactam}, ≤ 4/4 mg/L) and cefiderocol (39.1% MIC_cefiderocol, ≤ 2 mg/L) (Table 1). On the other hand, aztreonam/zidebactam had the same susceptibility rate as cefiderocol or aztreonam/avibactam (69.2% MIC,  ≤ 16 mg/L) and these were the most active agents in the MBL-producing P. aeruginosa collection followed by combinations including aztreonam and nacubactam or taniborbactam (66.2% MIC_{aztreonam/nacubactam or taniborbactam}, ≤ 16 mg/L). Notably, zidebactam alone showed excellent activity against E. coli (97% MIC_zidebactam, ≤ 4 mg/L), which was significantly higher than the one observed against Klebsiella spp. (17.6% MIC_zidebactam, ≤ 4 mg/L) and P. aeruginosa (2.6% MIC_zidebactam, ≤ 4 mg/L). MIC results for all Enterobacterales and P. aeruginosa clinical isolates co-producing MBLs and other β-lactamases are shown in Table S1, including their β-lactam resistance determinants.

Susceptibility testing of recombinant E. coli and P. aeruginosa strains

In order to precisely evaluate the impact of variable β-lactamases on susceptibility to aztreonam/β-lactamase inhibitor combinations, MIC values of cefiderocol, aztreonam and aztreonam/β-lactamase inhibitors were determined using a series of isogenic E. coli TOP10 recombinant strains (Table 2). Interestingly, all combinations including zidebactam showed lower MICs compared with other agents, mainly due to the antibacterial action of zidebactam, although it exhibited comparable inhibitory activity compared with other inhibitors (Table 3). Higher MIC values for aztreonam/nacubactam and aztreonam/taniborbactam compared with aztreonam/avibactam were observed for SHV-12, CTX-M-15, CMY-42 and PER-like E. coli TOP10 producers. Those results are in line with the lower capacities of taniborbactam and nacubactam to inhibit those latter enzymes compared with avibactam (with the exception of taniborbactam versus avibactam) (Table 3). A reduced efficacy of aztreonam/avibactam was also observed against E. coli TOP10 producing PER-like or CMY-like enzymes. Finally, high heterogeneity was observed in terms of susceptibility to cefiderocol for recombinant E. coli TOP10 strains, higher MIC values being actually observed for the SHV-12, PER-like, CMY-like, VIM-2 and NDM-like producers.

The same approach was then used to evaluate the impact of a series of clinically relevant β-lactamases in P. aeruginosa, using a series of isogenic recombinant P. aeruginosa PAO1 strains. As expected, none of the MBLs conferred resistance to aztreonam; however, production of PER-like enzymes had a significant impact on aztreonam susceptibility. Of note, most of the tested combinations allowed recovery of the susceptibility to aztreonam except for PER-like-producing strains, thus showing that none of those DBOs was efficiently inhibiting these ESBLs. The only effective combination was the one with zidebactam, but MIC values of zidebactam alone showed that this was due to the intrinsic activity of that DBO against P. aeruginosa PAO1. Nevertheless, our data indicated that zidebactam was not significantly inactivated by any of the β-lactamases tested, thus preserving its intrinsic activity. As already observed in E. coli TOP10, production of MBLs or of PER-like enzymes in P. aeruginosa PAO1 led to significantly increased MICs of cefiderocol.

Resistance to the aztreonam/zidebactam combination

Resistance to the aztreonam/zidebactam combination was observed for two isolates only. One of these corresponded to a K. pneumoniae isolate co-producing VIM-2 and SHV-5, with an MIC value of aztreonam/zidebactam being higher than 128/4 mg/L, and >8 mg/L for zidebactam alone. Such a high MIC value could be explained by several possible mechanisms, including a weak capacity of zidebactam to inhibit SHV-5, a weaker intrinsic antibacterial activity of zidebactam due to modifications in the PBP2 binding target, and modifications in the PBP3 sequence leading to higher resistance to aztreonam.

Another isolate exhibiting poor susceptibility to aztreonam/zidebactam (MIC at 4/4 mg/L) was an S. marcescens isolate, also showing a high MIC of zidebactam alone. This resistance could be explained by the loss of, or lower, antibacterial action of zidebactam, previously reported in S. marcescens.²¹

Inhibitory activities of DBOs against ESBLs

In order to better evaluate the respective activities of the different β-lactamase inhibitors, namely the ‘old ones’ clavulanic acid and tazobactam, the ‘new one’ avibactam, and the ‘future ones’ nacubactam, taniborbactam and zidebactam, against clinically relevant β-lactamase inhibitors, determination of the IC₅₀ values was performed using crude enzyme extracts from a series of E. coli recombinant strains. As expected, taniborbactam was the only effective inhibitor antagonizing the activity of MBLs (such as VIM-2 and NDM-5). KPC-3 was well inhibited by all inhibitors except clavulanic acid and tazobactam showing partial activities. Of note, avibactam and taniborbactam showed the highest inhibitory activities against KPC-3.

Avibactam and tazobactam were the best inhibitors of the ESBL CTX-M-15 activity, while taniborbactam, nacubactam and zidebactam performed similarly to clavulanic acid. Avibactam was also identified as the inhibitor with the highest activity against the ESBL SHV-12.

Results obtained with the PER-7 enzyme, as representative of the PER-type ESBLs, showed that most inhibitors performed at moderate levels, with clavulanic acid showing the highest inhibitory activity. This unexpected observation is in line with the obtained MIC data showing relatively high MICs for the PER-1-, PER-6- and PER-7-producing E. coli strains when testing the aztreonam/avibactam, aztreonam/nacubactam and aztreonam/taniborbactam combinations (Table 2).

Finally, results obtained with the CMY-42 enzyme as a representative of the acquired AmpC-type β-lactamases frequently identified in the resistome of the tested isolates (Table S1) showed that the different DBOs were overall significant inhibitors of its hydrolytic activity, even though this inhibitory capacity was much lower than the ones observed with the class A ESBLs SHV-12 and CTX-M-15 (Table 3).

Discussion

In this study, we showed that combinations including aztreonam and any of those recently developed β-lactamase inhibitors zidebactam, nacubactam and taniborbactam had higher effective in vitro activity compared with the one combining aztreonam and avibactam, namely aztreonam/avibactam that is going to be commercially available in a near future. The efficacy of those combinations was also better than that of cefiderocol against MDR MBL-producing Enterobacterales clinical isolates, therefore providing opportunities for new therapeutic options. On the contrary, our data showed that aztreonam/β-lactamase inhibitor combinations did not exhibit better activity against MBL-producing P. aeruginosa compared with cefiderocol.

Although β-lactamase inhibitors belonging to the DBO class including zidebactam and nacubactam do not inhibit the activity of MBLs, they are promising molecules due to their direct antibacterial activities by binding to PBP2, and acting as enhancers with the β-lactam partner binding PBP3, corresponding to aztreonam in the present study.^21,24 Many studies did already show a high in vitro or in vivo effectiveness of the cefepime/zidebactam combination against MDR clinical isolates,^{13,16,21,25,26} but few studies evaluated the in vitro activity of a combination including a monobactam with a DBO.^27,28 Livermore et al.¹⁵ reported that almost 100% of MBL-producing Enterobacterales were susceptible to a combination of aztreonam and nacubactam at 4/4 mg/L. However, in that study, approximately 30% of the isolates remained susceptible to aztreonam, suggesting that there was no ESBL or AmpC associated, which differs from our strain collection tested. Moreover, Vázquez-Ucha et al.,²⁹ by studying a collection of 32 ESBL- and MBL-co-producing Enterobacterales, showed that 93.8% and 71.9% of those isolates had MIC values of ≤ 2 mg/L for cefepime/zidebactam and cefepime/taniborbactam, respectively. This is consistent with our results, the latter ones suggesting that even higher activity could be achieved if combining those DBOs with aztreonam.

Interestingly, our data showed higher activity of the aztreonam/zidebactam combination compared with aztreonam/nacubactam. In previous studies, it was evidenced that MICs of combinations containing zidebactam were indeed lower than those containing nacubactam in Enterobacterales, likely explained by a higher intrinsic antibacterial activity of zidebactam at a fixed concentration at 4 mg/L, in line with our observations.^21,24

Only two strains tested here showed reduced susceptibility to the aztreonam/zidebactam combination, even if a significant number of the strains tested were actually resistant to zidebactam alone. Hence, this further highlights the significant action as a β-lactamase inhibitor of that molecule, besides its antibacterial activity.

Altogether, our study showed that the aztreonam/DBO combinations seem particularly interesting to treat infections caused by MBL-producing Enterobacterales. To the best of our knowledge, this is the first study reporting in vitro activity of aztreonam/taniborbactam. This combination exhibits excellent activity against MBL-producing Enterobacterales, which is higher than the one observed with cefiderocol and comparable to that observed with aztreonam/avibactam. We must however acknowledge that the rate of aztreonam/avibactam-resistant isolates included in our strain collection was higher than in other previously published studies.^30–33 In fact, our collection included a series of E. coli isolates co-producing NDM-5 and CMY-like β-lactamases and possessing modifications of their PBP3 sequences eventually leading to resistance to aztreonam/avibactam.^10,34,35 For those latter isolates, our data also suggest that aztreonam/zidebactam might be a very effective therapeutic option. Notably, cross-resistance between aztreonam/avibactam, aztreonam/taniborbactam and aztreonam/nacubactam was observed in a few isolates, suggesting similar potential resistance-associated mechanisms.

On the other hand, the aztreonam/β-lactamase inhibitor combinations did not overall display satisfactory activities against MBL-producing P. aeruginosa. This might be related to different features: (i) the efficacy of aztreonam in P. aeruginosa might be affected by non-β-lactamase mechanisms such as overproduction of efflux pumps or permeability defects;^36,37 and (ii) some MBL-producing P. aeruginosa strains may co-produce broad-spectrum β-lactamases hydrolysing aztreonam at a high level and being poorly inhibited by β-lactamase inhibitors (such as the ESBL PER-1).⁹ Moreover, overproduction of the intrinsic AmpC-type PDC β-lactamase but also of acquired OXA-type β-lactamases significantly hydrolysing aztreonam are commonly observed in P. aeruginosa, and the efficacy of DBOs as inhibitors of those latter β-lactamases remains unknown.³⁸

To conclude, this in vitro study highlights that combinations including aztreonam with zidebactam, nacubactam or taniborbactam might be very interesting therapeutic options against MBL-producing Gram-negative bacteria, particularly for Enterobacterales.

Funding

This work was financed by the University of Fribourg, Switzerland, the NARA, and by the Swiss National Science Foundation (grant FNS 310030_1888801).

Transparency declarations

None to declare.

Supplementary data

Table S1 is available as Supplementary data at JAC Online.

References