In Vitro Activity of Auranofin in Combination With Aztreonam-Avibactam Against Metallo-β-lactamase (MBL)-Producing Enterobacterales

Objectives To assess the efficacy of aztreonam-avibactam-auranofin (ATM-AVI-AUR) against a collection of 88 carbapenemase-producing Enterobacterales (CPE) clinical isolates and 6 in vitro selected ATM-AVI-resistant CPE with CMY-16 Tyr150Ser and Asn346His mutants or transformants. Methods MICs of imipenem, ceftazidime-avibact8am (CAZ-AVI), ATM-AVI, CAZ-AVI-AUR and ATM-AVI-AUR were determined via the broth microdilution method. Genetic background and carbapenemase genes were determined by PCR and Sanger sequencing. Results AUR alone showed little antibacterial activity with AUR MICs were greater than 64 μg/mL for all the 88 clinical CPE isolates. The addition of AUR (16 μg/mL) resulted in an 3-folding dilutions MIC reduction of ATM-AVI MIC50 (0.5 to 0.0625 μg/mL) and a 2-folding dilutions MIC reduction of MIC90 (1 to 0.25 μg/mL) against all 88 clinical CPE isolates, respectively. Notably, the reduced ATM-AVI MIC values were mainly found in MBL-producers, and the MIC50 and MIC90 reduced by 2-folding dilutions (0.25 to 0.0625 μg/mL) and 3-folding dilutions (2 to 0.25 μg/mL) respectively by AUR among the 51 MBL-producers. By contrast, the addition of AUR did not showed significant effects on ATM-AVI MIC50 (0.0625 μg/mL) and MIC90 (0.125 μg/mL) among single KPC-producers. Interestingly, the addition of AUR restored the ATM-AVI susceptibility against the 6 in vitro selected ATM-AVI-resistant CMY-16 Tyr150Ser and Asn346His mutants or transfromants, with the MICs reduced from ≥32 μg/mL (32->256 μg/mL) to ≤8 μg/mL (0.0625-8 μg/mL). Conclusions Our results demonstrated that AUR potentiated the activities of CAZ-AVI and ATM-AVI against MBL-producing isolates in vitro. Importantly, AUR restored the ATM-AVI activity against ATM-AVI resistant mutant strains. As a clinically approved drug, AUR might be repurposed in combination with ATM-AVI to treat infections caused by highly resistant MBL-producing Enterobacterales.


INTRODUCTION
Carbapenem resistance in carbapenem-resistant Enterobacterales (CRE) is driven primarily by the acquisition of various carbapenemases that are able to degrade carbapenem antibiotics. Clinically relevant carbapenemases include both serine b-lactamases (e.g., KPC and OXA-48) and metallo-blactamases (MBLs, e.g., NDM, VIM, and IMP) (Rodrıǵuez-Baño et al., 2018). The rapid worldwide emergence of MBLs in Enterobacterales is especially worrisome due to their broad, constant and efficient carbapenemase activity against almost all the b-lactam antibiotics (except for monobactam, e.g., aztreonam, ATM), potential for horizontal gene transfer, and the absence of clinically available inhibitors (Bush and Bradford, 2020). Although novel agents such as the newer b-lactam-blactamase inhibitor combinations i.e., ceftazidime-avibactam (CAZ-AVI), meropenem-vaborbactam, and imipenemrelebactam, offer safe and effective therapies for serious infections caused by some of the extended-spectrum blactamases, KPC, AmpC, and/or OXA-48 producers, none of them present in vitro activity against MBL-producing Enterobacterales (Yahav et al., 2021). Even though the combination of ATM-AVI showed potency against MBLproducing pathogens, resistance to ATM-AVI has also been documented (Alm et al., 2015;Niu et al., 2020;Ma et al., 2020;Sadek et al., 2020).
Recent study showed that certain metallodrugs can irreversibly abrogate MBL's carbapenemase activity via the displacement of Zn(II) cofactors from their active sites, suggesting a metallodrug-antibiotic combination might serve as an effective strategy to combat Zn-dependent resistant determinants, such as MBLs. The study by Wang et al. showed that an anti-peptic ulcer bismuth drug, colloidal bismuth subcitrate, could re-sensitize MBL-producers to b-lactam antibiotics through abrogating the Zn(II) cofactors by Bi(III) from the MBL active sites, thus disrupting their abilities to hydrolyze b-lactam ring in carbapenems (Wang et al., 2018). Similarly, a recent study by Sun et al. identified an antirheumatic drug, auranofin (AUR), as a potent MBL inhibitor by irreversibly abolishing MBL's carbapenemase activity via the displacement of Zn(II) cofactors from their active sites. AUR demonstrated synergistic effect with imipenem (IPM) on killing a broad spectrum of CRE strains, potently restored the susceptibility of MCR-1-and NDM-5-co-producing pathogens to colistin in a murine peritonitis model, and significantly slowed down the development of carbapenem resistance (Roder and Thomson, 2015;Sun et al., 2020). In addition, a more recent study showed that AUR in combination with colistin was effective against colistin-resistent Gram-negtive bacteria both in vitro and in vivo (Feng et al., 2021).
Here we reported the activities of ATM-AVI-AUR and other antibiotics against a selection of 88 carbapenemase-producing Enterobacterales (CPE) clinical isolates (37 single KPCproducers, 33 single MBL-carriers, 18 dual/triple carbapenemase-producers). In addition, we investigated the impact of AUR addition on ATM-AVI activity among the in vitro-selected ATM-AVI highly resistant CMY-16 Tyr150Ser and Asn346His K. pneumoniae mutants or transformants.

Antimicrobial Susceptibility Testing
MICs for all the CPE strains were determined using standard broth microdilution method. For CAZ-AVI and ATM-AVI MICs evaluation, AVI was tested at a fixed concentration of 4 mg/mL, while CAZ and ATM were added at different concentrations ranged from 0.0625 to 256 mg/mL, respectively. For CAZ-AVI-AUR and ATM-AVI-AUR evaluation, AVI and AUR were added at the fixed concentrations of 4 mg/mL and 16 mg/mL, respectively (Wang et al., 2018), while CAZ and ATM were added at different concentrations from 0.0625 to 256 mg/mL (CLSI, 2020). Susceptibility testing was performed in triplicate on three different days. Escherichia coli ATCC25922 strain was used as quality control for all testing.

In Vitro Selection
Multi-step selection was performed by inoculation of~10 8 cfu in 2-mL LB broth containing ATM-AVI at the 0.5× MICs and incubated for 24 h (Niu et al., 2020). This procedure was repeated daily, each time doubling the ATM concentration up to a maximum of 128 mg/mL, with AVI concentrations fixed at 4 mg/mL. Resistant variants were selected by plating a bacterial suspension on agar plates of corresponding ATM-AVI levels. The same multi-step selections were used to select ATM-AVI-AUR resistance, with AUR and AVI concentrations fixed at 16 and 4 mg/mL, respectively.

In Vitro Activities of CAZ-AVI-AUR and ATM-AVI-AUR Against Clinical CPE Isolates
AUR alone showed little antibacterial activity with AUR MICs were greater than 64 mg/mL for all the 88 clinical CPE isolates. Susceptibility testing data showed that AUR concentrations of ≥16 mg/mL were required to reduce the ATM-AVI MICs from 8 mg/mL to <=0.5 mg/mL for all the 51 clinical MBL-positive isolates tested. And further increasing AUR concentration provided little benefit and, so, AUR was fixed at 16 mg/mL in the following in vitro susceptibility studies. Other agents showed various activities against these CPE isolates ( Table 1). The majority of the isolates (84/88, 95.5%) were resistant to IPM (MIC 50 = 32 mg/mL), and 54.5% (n = 48) were resistant to CAZ-AVI (MIC 50 >64 mg/mL), none of the 88 isolates were resistant to ATM-AVI (MIC 50 = 0.5 mg/mL). The addition of AUR (16 mg/ mL) significantly reduced the CAZ-AVI MIC 50 from >64 to 1 mg/ mL but had no significant effect on the MIC 90 (from >128 to 128 mg/mL) in the 88 clinical CPE isolates. Compared with the result of ATM-AVI alone, the addition of AUR (16 mg/mL) resulted in an 3-folding dilutions MIC reduction of ATM-AVI MIC 50 (from 0.5 to 0.0625 mg/mL) and a 2-folding dilutions MIC reduction of MIC 90 (from 1 to 0.25 mg/mL), respectively.
Activities of ATM-AVI-AUR Against ATM-AVI Resistant K. pneumoniae
In addition, our previous study showed that the CMY-16positive, dual-carbapenemase-producing (NDM-1 and OXA-48), ST101 K. pneumoniae Kp202 can be in vitro selected to be ATM-AVI resistant (Niu et al., 2020). In this study, we examined the in vitro selection ATM-AVI resistance capability of Kp202 under the presence of AUR (16 mg/mL). The results showed 16 mg/mL AUR could completely inhibit the in vitro selected ATM-AVI resistance and no ATM-AVI-AUR resistant isolates were obtained despite multiple attempts.

DISCUSSION
The wide spread of multi-drug resistant Gram negative pathogens raised significant clinical and public concerns. Along with carbapenem-resistant Acinetobacter baumannii and carbapenem-resistant Pseudomonas aeruginosa, CRE top the WHO's priority list of resistant pathogens (WHO, 2019a). CREs are also regarded as the 'Urgent Resistance Threats' by CDC and are prioritized in the UK's 5-Year antimicrobial resistance national action plan (Anon, 2013;. In particular, the global dissemination of MBL-producing Enterobacterales should be considered as the most urgent threat due to potent horizontal transfer and the lack of therapeutic inhibitors, and hence the 2019 WHO review of both the preclinical and clinical antibacterial pipeline stressed MBL-producing Enterobacterales as a neglected target in antibacterial drug development (WHO, 2019a;WHO, 2019b).
Therapy of invasive infections due to MBL-producers is extremely challenging, as most bla MBL -carrying plasmids coharbor multiple additional resistance determinants, resulting in limited and sub-optimal treatment options. Although ATM is the only b-lactam antibiotic that remains stable for hydrolysis by MBLs, it can be hydrolyzed by most serine b-lactamases (SBLs), which are frequently co-produced by MBL-producers (Wu et al., 2019). AVI is a non-b-lactam-b-lactamase inhibitor which can inhibit most SBLs, but is ineffective to MBLs (Abboud et al., 2016). ATM-AVI combination has demonstrated potent in vitro activity against not only SBLs producers, but also MBLproducing Enterobacterales. One study from Sader et al. reported ATM-AVI to be highly active against 60 MBLpositive isolates (Sader et al., 2017), and another large study demonstrated potent activity of ATM-AVI against meropenemnonsusceptible MBL-positive isolates (Karlowsky et al., 2017). A more recent study reported that ATM-AVI had potent activity against a large collection of contemporary Enterobacterales isolates (n = 8787; MIC 50 / 90 , 0.03/0.12 mg/mL), including MBL-producers (n = 110; MIC 50 / 90 , 0.12/0.5 mg/mL), supporting the clinical development of ATM-AVI for the treatment of infections caused by MBL-producing strains (Sader et al., 2021).
Our previous study showed that although ATM-AVI had good in vitro activity against MBL-producing K. pneumoniae isolates, clinical CMY-16-positive, dual-carbapenemaseproducing (NDM-1 and OXA-48), ST101 K. pneumoniae strain can be successfully selected to be ATM-AVI-resistant in vitro (Niu et al., 2020). Worryingly, decreased susceptibilities and resistance to ATM-AVI have also been recently witnessed in clinical E. coli strains from India and China (Alm et al., 2015;Ma et al., 2020), due to amino-acid insertion in PBP3. Furthermore, a recent study of 118 clinical MBL-producing E. coli isolates from various geographical origins (Europe, Africa, Asia, and Australia) alarmingly showed 16.1% of the isolates were ATM-AVI resistant, due to a combination of PBP3 modification and the CMY-42 production (Sadek et al., 2020). These reports suggested that the clinical efficacy of ATM-AVI against MBLproducer infections could be compromised once the resistance emerges. Therefore, new drugs that specifically target MBLs and restore efficacy of last-line drugs against MBL-producing pathogens are urgently needed.
AUR directly inhibited the thioredoxin reductase (TrxR) in Gram-positive bacteria, leading to disruption of thiol-redox homeostasis and cell death. However, AUR alone does not have intrinsic antimicrobial activity against Gram-negative bacteria (Harbut et al., 2015). A recent study showed that imipenem in combination with AUR is effective against a broad spectrum of CRE strains (Sun et al., 2020). In addition, a more recent study showed that AUR in combination with colistin was effective against colistin-resistent Gram-negtive bacteria both in vitro and in vivo (Feng et al., 2021).
Here we demonstrated that AUR can potentiate the activity of both ATM-AVI and CAZ-AVI against a broad range of MBLproducing strains, and restored the activity of ATM-AVI against all the CPE clinical isolates. ATM-AVI-AUR combination was highly efficient against isolates carrying either NDM-, IMP-or VIM-type MBLs, resulting in MICs as low as 0.004-0.5 mg/mL. Our results showed that although the ATM-AVI combination is highly effective against the majority (92.16%, 47/51) of the clinical MBL-producing isolates of diverse MBLs (Table 1), 7.84% (4/51) of the clinical MBL-producing isolates analyzed in this study had the MICs of ATM-AVI at 8 mg/mL (intermediate for ATM based on the CLSI guideline) (CLSI, 2020). However, the addition of AUR successfully reduced the MICs of ATM-AVI by ≥4-folding dilutions, from 8 mg/mL to <=0.5 mg/mL (0.25-0.5 mg/mL). In addition, the combination of ATM-AVI-AUR showed superior effect against than that of CAZ-AVI-AUR against MBL-producers, presumably because AUR together with ATM-AVI may work synergistically to abrogate MBL enzymatic activities.
More importantly, the MICs of ATM-AVI were reduced from ≥32 mg/mL (32->256 mg/mL) to ≤8 mg/mL (0.0625-8 mg/mL) in all the 6 in vitro selected ATM-AVI-highly-resistant mutants and constructs with the AUR combination. Interestingly, our results also indicated that the usage of the triple combination of AUR-ATM-AVI could lower the possibility of in vitro selected ATM-AVI resistance. One limitation of our study is that we didn't have any PBP3 modification caused ATM-AVI resistant strains available for testing, and it is important to examine whether ATM-AVI-AUR remain potent against these ATM-AVI resistant strain in the future.
Taken together, our study showed that AUR in combination with ATM-AVI showed potent activity against clinical CPE isolates, especially MBL-producers. ATM-AVI-AUR might provide as a therapeutic option to treat life-threatening infections caused by MBL-producing Enterobacterales, including strains co-producing more than one carbapenemase. Future studies are need to evaluate its in vivo efficacy and feasibility of clinical applications.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.