A New Twist: The Combination of Sulbactam/Avibactam Enhances Sulbactam Activity against Carbapenem-Resistant Acinetobacter baumannii (CRAB) Isolates

An increasing number of untreatable infections are recorded every year. Many studies have focused their efforts on developing new β-lactamase inhibitors to treat multi-drug resistant (MDR) isolates. In the present study, sulbactam/avibactam and sulbactam/relebactam combination were tested against 187 multi-drug resistant (MDR) Acinetobacter clinical isolates; both sulbactam/avibactam and sulbactam/relebactam restored sulbactam activity. A decrease ≥2 dilutions in sulbactam MICs was observed in 89% of the isolates when tested in combination with avibactam. Sulbactam/relebactam was able to restore sulbactam susceptibility in 40% of the isolates. In addition, the susceptibility testing using twenty-three A. baumannii AB5075 knockout strains revealed potential sulbactam and/or sulbactam/avibactam target genes. We observed that diazabicyclooctanes (DBOs) β-lactamase inhibitors combined with sulbactam restore sulbactam susceptibility against carbapenem-resistant Acinetobacter clinical isolates. However, relebactam was not as effective as avibactam when combined with sulbactam. Exploring novel combinations may offer new options to treat Acinetobacter spp. infections, especially for widespread oxacillinases and metallo-β-lactamases (MBLs) producers.


Sulbactam/Avibactam Susceptibility Testing of Acinetobacter spp. Clinical Strains
We observed that 129/187 (69%) Acinetobacter spp. isolates were inhibited by 4 mg/L of sulbactam when combined with 4 mg/L avibactam ( Figure 1B). Remarkably, 105/116 (91%) of OXA producers showed MIC equal to or less than 4 mg/L for the sulbactam/avibactam combination. NDM producers showed a reduction in the MICs of sulbactam due to the addition of avibactam, although MIC values were above 4 mg/L for all but one isolate ( Figure 1B). The MICs of sulbactam/avibactam also showed a bi-modal distribution but with its modes displaced toward lower MICs. The first mode of 2 mg/L (3-fold lower when compared to sulbactam alone) corresponded to OXA producers, and the second mode of 32 mg/L to NDM producers. The MIC 50 and MIC 90 of sulbactam/avibactam were 2 and 4 mg/L for OXA producers compared to 16 and 64 mg/L for sulbactam alone, respectively. Among NDM producers, the MIC 50 and MIC 90 values of >64 mg/L for sulbactam alone decreased to 32 and 64 mg/L for sulbactam/avibactam, respectively. An experiment using an expanded range of concentrations using as highest value 1024 mg/L on a subset of NDM producers (n = 40/48) showed an MIC 50 of 256 mg/L for sulbactam alone and 32 mg/L for sulbactam/avibactam. This result indicated that there is an enhancement of sulbactam activity by avibactam in this group of strains.

Sulbactam/Relebactam Susceptibility Testing of Acinetobacter spp. Clinical Strains
With the sulbactam/relebactam combination, 73/187 (39%) Acinetobacter spp. isolates were inhibited by 4 mg/L ( Figure 1D). The difference in activity with avibactam was due to a smaller proportion of OXA producers (52/116, 45%) inhibited with the combination of sulbactam/relebactam ( Figure 1E). All 9 Acinetobacter isolates producing class D β-lactamases that were not inhibited by avibactam/sulbactam were also resistant to avibactam/relebactam. The MICs of sulbactam/relebactam also showed a bi-modal distribution with MIC 50 and MIC 90 of sulbactam/relebactam of 8 and 16 mg/L for OXA producers compared to 16 and 64 mg/L for sulbactam alone. Among NDM producers, the MIC 50 values were 64 mg/L for sulbactam/relebactam alone compared to 256 mg/L for sulbactam.
For strains harboring MBLs, colistin and tigecycline were the only tested comparators with uniform in vitro activity (100% and 94% susceptible, respectively). Only 18% of NDM producers were susceptible to amikacin. Conversely, comparators among OXA producers were less active than the sulbactam/avibactam combination: 86% and 85% of susceptibility to colistin and tigecycline, respectively. Ceftazidime was only active against non-ESBL/non-CRAB isolates, while ceftazidime/avibactam was not active against CRAB and showed a modest 14% susceptibility against ESBL producers. Imipenem was uniformly active against non-CRAB isolates, but only 2% of CRABs (all OXA-58 producers) were susceptible to this carbapenem.

Sulbactam/Avibactam Susceptibility Testing of A. baumannii Knockout Strains
The potentiation of sulbactam activity by avibactam was observed in carbapenemases refractory to DBOs inhibition, as OXAs commonly found in Acinetobacter, suggesting a mode of action in these species unrelated to the reported inhibitory capacity. To reveal the potential sulbactam/avibactam targets, twenty-three A. baumannii AB5075 knockout strains were tested to identify potential sulbactam/avibactam targets. Knockout mutant in genes related with efflux pumps, two components system, cell wall synthesis genes, among others, were used. A. baumannii AB5075 MIC to sulbactam was 24 mg/L and showed a 16 times reduction in sulbactam MIC by addition of avibactam (Table 1). The results observed with A. baumannii AB5075 mutants can be classified into three large groups. i.
Mutations affecting the activity of sulbactam alone: when comparing sulbactam MIC in the wild-type strains and isogenic mutants, a decrease in MIC values from 24 to 3 mg/L was observed in the AB5075∆mrcB and AB5075∆mreB, suggesting a potential role in sulbactam susceptibility (interaction). The mrcB and mreB genes, which code for a transglycosylase-transpeptidase and a cytoskeletal protein, respectively, were previously reported as genes involved in the formation of the peptidoglycan synthesis. Potentiation of sulbactam activity by avibactam is also lost in these mutant strains. Additionally, the A. baumannii AB5075∆PBPG MIC of sulbactam is 1.5-fold lower than that measured for A. baumannii AB5075. PBPs are known to be sulbactam targets [36]. Previous reports demonstrated that the antibacterial activity of sulbactam was mediated through inhibition of PBP1 and PBP3.
ii. Mutations that result in loss of avibactam enhancement of sulbactam activity: a 1 or 0-fold decrease in the MIC values comparing sulbactam with sulbactam/avibactam MICs was observed in seven A. baumannii knockout strains (AB5075∆bfmR, AB5075∆P BPG, AB5075∆advA, AB5075∆adeB, AB5075∆mrcB, AB5075∆xerD, and AB5075∆mreB). These results identified potential sulbactam and/or sulbactam/avibactam targets (Table 1). iii. Mutations that do not affect susceptibility to sulbactam or potentiation of sulbactam by avibactam: ≤2-fold decrease in the MIC was observed in twelve A. baumannii AB5075 knockout strains comparing sulbactam with sulbactam/avibactam MIC. We can rule out that these mutated genes are involved in the mechanisms of action (or interaction) of sulbactam and avibactam.

Discussion
Infections with the hospital-acquired bacterium Acinetobacter are extremely difficult to treat. While the development of new antibiotics remains one option in the fight against strains resistant to those currently available, an attractive (and faster) alternative is the development of adjuvant therapies to restore the efficacy of existing antibiotics. A preliminary in vitro study has shown promising results with avibactam adjuvant approaches combined with sulbactam [31]. In this work we contributed new knowledge to support sulbactam/avibactam interaction.
We observed that the addition of DBOs restores sulbactam susceptibility. This interaction between DBOs and sulbactam presented the following characteristics: (i) was more potent in serine carbapenemase-producing strains, such as oxacillinases, but it was also manifested to a lesser extent in NDM-producing strains; (ii) was not related to the DBO concentration as a dose-response relationship was not observed in MICs when both DBOs were tested using twice the recommended concentration (8 mg/L) (not shown); (iv) was not related to the β-lactamase inhibitory capacity described for DBOs; and (v) avibactam was more potent in restoring sulbactam susceptibility than relebactam. The differences observed between avibactam and relebactam would support these hypotheses, the larger molecular size of the second DBO (molar mass 265.24 g/mol g·mol −1 vs. 348.37 g·mol −1 , respectively) could be responsible for a reduced capability to penetrate the outer membrane and reach its targets.
Recently, ETX2514, named durlobactam, a new DBO inhibitor derived from avibactam (resulting from the addition of a double bond between C3 and C4 and a methyl group in C3 position compared to avibactam), significantly increased the susceptibility of Acinetobacter clinical isolates when it was combined with sulbactam [22]. Sulbactam/durlobactam has an MIC 50 /MIC 90 of 1 and 2 mg/L against a global collection of strains. These values are comparable to those obtained in this work for the sulbactam/avibactam combination for OXA-23 producers (2 and 4 mg/L, respectively) [22]. However, the combination with durlobactam was not effective for NDM strains with MIC 50/90 comparable to those of sulbactam alone, unlike avibactam, which showed a slight improvement in sulbactam MICs in these strains. In this context, the potent synergy observed between sulbactam and avibactam indicates that there could be different targets for each DBO.
We were unable to detect a known resistant mechanism compromising sulbactam activity. Only a small proportion of the strains presented narrow-spectrum or extended spectrum β-lactamases that affect sulbactam. No mutations in the genes coding for PBPs were observed in a previously analyzed subset of NDM producers [15]. Taken together, these results suggest that sulbactam/avibactam synergy could have an effect on other targets in MDR strains.
We explored the possible contribution in resistance to sulbactam and its interaction with avibactam of genes related with efflux pumps, two components system, cell wall synthesis genes. We observed that mrcB and mreB were two key components, in addition to the coding genes for PBPs, for phenotypic resistance to sulbactam. mrcB, which codes for a transglycosylase-transpeptidase, and mreB, which codes for cytoskeletal proteins, could play a key role in maintaining vital (or alternate) processes in wall synthesis in the presence of sulbactam, which can explain in part the observed sulbactam MIC reduction when they are inactivated. Potentiation of sulbactam activity by avibactam is also lost in these mutant strains, suggesting that conserved mrcB and mreB are a necessary (but not sufficient) prerequisite for avibactam interactions. More studies to discern the possible interplay of sulbactam and sulbactam/avibactam on mreB and mrcB are needed. We also identified key genes that seem necessary to guarantee the interaction between sulbactam and avibactam. BfmRS, a global regulatory system in Acinetobacter, is a key challenge for antibiotic therapy. BfmRS per se has been shown to increase the ability of Acinetobacter to grow in the presence of diverse antibiotics and tolerate transient, high-level antibiotic exposures [37]. As we observed, avibactam (with sulbactam) may interfere or bypass the normal protection from β-lactam toxicity mediated by BfmRS. advA (antibiotic susceptibility and division protein of Acinetobacter) is another essential gene for A. baumannii growth. It has been shown that reduced AdvA levels modulated antibiotic susceptibility and generate a selective pattern of hypersensitivity to fluoroquinolones and β-lactams [38]. Avibactam could be a hypothetical mediator for sulbactam hypersensitivity through advA interaction. Identification of other genes potentially involved in the interaction between sulbactam and avibactam, such as XerCD, tyrosine recombinases that catalyze the resolution of dimeric chromosomes [39], opens up new possibilities in target identification for designing new drug combinations.
In summary, sulbactam/DBOs, especially in combination with avibactam, demonstrated potent antibacterial activity against recent, geographically diverse clinical isolates of Acinetobacter, including XDR isolates such as widespread oxacillinase producers. The addition of a DBO to sulbactam appears to be a very promising strategy for managing difficult-to-treat Acinetobacter infections. We identified at least 6 genes that could participate in the interaction of sulbactam with avibactam. Understanding this interaction will clarify how A. baumannii optimizes its ability to grow under antibiotics exposition. Further extensive in vitro and in vivo studies have to be performed to confirm the potential use of these adjuvants as therapeutic alternatives.

Antimicrobial Susceptibility Assay
MICs against sulbactam (range 0.12-512 mg/L), sulbactam plus 4 mg/L avibactam (range 0.12-64 mg/L) and sulbactam plus 4 mg/L relebactam (range 0.12-64 mg/L) were performed using the reference methods agar dilution according to CLSI Standards [43]. Because breakpoints are available for sulbactam alone, 4 mg/L was applied for this analysis based on the CLSI susceptible breakpoint of 8/4 mg/L for ampicillin/sulbactam for Acinetobacter spp. [44,45]. The in vitro activity of other comparators, such as imipenem, colistin, tigecycline, amikacin, ceftazidime (alone and in combination with 4 mg/L of avibactam) were also evaluated by dilution methods at one facility.