Assessment of In Vitro Cefiderocol Susceptibility and Comparators against an Epidemiologically Diverse Collection of Acinetobacter baumannii Clinical Isolates

Cefiderocol is a catechol-substituted siderophore cephalosporin combining rapid penetration into the periplasmic space with increased stability against β-lactamases. This study provides additional data on the in vitro antimicrobial activity of cefiderocol and commercially available comparators against an epidemiologically diverse collection of Acinetobacter baumannii clinical isolates. Antimicrobial susceptibility was tested using pre-prepared frozen 96-well microtiter plates containing twofold serial dilutions of: cefepime, ceftazidime/avibactam, imipenem/relebactam, ampicillin/sulbactam, meropenem, meropenem/vaborbactam, ciprofloxacin, minocycline, tigecycline, trimethoprim/sulfamethoxazole and colistin using the standard broth microdilution procedure in cation-adjusted Mueller–Hinton broth (CAMHB). For cefiderocol, iron-depleted CAMHB was used. A collection of 113 clinical strains of A. baumannii isolated from Argentina, Azerbaijan, Croatia, Greece, Italy, Morocco, Mozambique, Peru and Spain were included. The most active antimicrobial agents against our collection were colistin and cefiderocol, with 12.38% and 21.23% of non-susceptibility, respectively. A high proportion of multidrug-resistant (76.77%) and carbapenem-resistant (75.28%) A. baumannii isolates remained susceptible to cefiderocol, which was clearly superior to novel β-lactam/β-lactamase inhibitor combinations. Cefiderocol-resistance was higher among carbapenem-resistant isolates and isolates belonging to ST2, but could not be associated with any particular resistance mechanism or clonal lineage. Our data suggest that cefiderocol is a good alternative to treat infections caused by MDR A. baumanni, including carbapenem-resistant strains.


Introduction
The emergence of bacteria resistant to currently available antibiotics is steadily increasing but the development of new therapies to combat infections caused by these microorganisms are not following at an appropriate pace. In the clinical setting infections caused by multidrug-resistant (MDR), Gram-negative bacteria are steadily increasing and of major global concern, particularly those caused by clinical isolates of Enterobacterales and non-fermentative Gram-negative bacilli, mainly Pseudomonas aeruginosa and Acinetobacter baumannii. Among the antibiotics towards which these bacteria show a resistant phenotype, carbapenem resistance deserves special attention as carbapenems are used as a therapy of last resort [1].
The present study comprised a collection of 113 bacterial isolates of A. baumannii recovered from several countries all over the world including both carbapenem-resistant and carbapenem-susceptible isolates. Upon testing the antimicrobial susceptibility to a selected panel of antibacterial agents, the lowest MIC 50 values reported among all 113 isolates tested corresponded to cefiderocol and colistin (MIC 50 = 0.5 µg/mL), closely followed by tigecycline and minocycline (MIC 50 = 2 and 4 µg/mL, respectively). The MIC 90 values, however, showed a slightly different pattern. Tigecycline showed the lowest MIC 90 value (MIC 90 = 4 µg/mL) followed by colistin (MIC 90 = 8 µg/mL). Cefiderocol, on the other hand, presented a MIC 90 of >64 µg/mL, only comparable to that of ampicillin/sulbactam (Table 1). While these values were above the range detected in previously published studies, Isler et al. have shown important differences in the reported MIC 50-MIC 90 values depending on the collection of strains being tested [5]. Nevertheless, the most active antimicrobial agents in our collection of isolates were both colistin and cefiderocol, showing only 12.38% of resistant isolates and 21.23% of non-susceptible isolates, respectively. Minocycline also showed good activity overall with 23.89% of resistant isolates. It is worth mentioning that more than 80% of the isolates were resistant to all of the remaining antimicrobial agents and only a handful of isolates showed an overall susceptible profile ( Table 2). As expected, a very high proportion of isolates were non-susceptible to cephalosporins, regardless of the carriage of any carbapenemase or meropenem susceptibility, with 86.76% (98/113) of the isolates being non-susceptible to cefepime and 94.69% (107/113) showing MIC values of ceftazidime/avibactam ≥16 µg/mL.
As much as 87.61% (99/113) of the A. baumannii isolates included in this study were non-susceptible to at least one antimicrobial agent from three classes tested (13 of them being already resistant to colistin) and, hence, were considered as MDR [11]. Among MDR isolates, 76.77% (76/99) were susceptible to cefiderocol with MIC values of ≤8 µg/mL, but all isolates that were resistant to cefiderocol were also MDR. In addition, 72.72% (72/99) and 92.92% (92/99) of MDR isolates were also susceptible to minocycline and tigecycline, respectively, both of them being considered as valuable therapeutic alternatives for the treatment of infections caused by MDR A. baumannii [12]. Among minocyclineresistant A. baumannii, three isolates were resistant to cefiderocol but remained susceptible to tigecycline and colistin. The few tigecycline-resistant isolates were all susceptible to cefiderocol and only two of them showed resistance to colistin ( Table 2).

Activity of Cefiderocol against Carbapenem-Resistant A. baumannii
Of the 113 isolates tested, 89 (78.76%) were resistant to meropenem (MIC≥8 µg/mL) and such resistance was associated with the carriage of either an OXA-type carbapenemase (79/89) or an OXA-type plus and NDM-type carbapenemase (9/89) in all isolates but one, which was negative for all carbapenemase-encoding genes tested and presented a carbapenem MIC of 8 µg/mL, right at the resistance breakpoint ( Table 2). All the remaining meropenem non-resistant isolates were also negative for the presence of carbapenemaseencoding genes except for a single isolate carrying bla OXA-24 that showed a carbapenem MIC of 0.5 µg/mL, thus suggesting very low expression of this carbapenemase. In addition, all carbapenem-resistant A. baumannii isolates but one showed MIC values of meropenem/vaborbactam and imipenem/relebactam of ≥16 µg/mL, suggesting little to no activity of these β-lactam/β-lactamase inhibitors. Vaborbactam and relebactam are novelβ-lactamase inhibitors active against class A and class Cβ-lactamases but are not able to inhibit class B or D carbapenemases, which are common in carbapenem-resistant A. baumannii isolates [13]. The only carbapenem-resistant isolate showing reduced MICs ofmeropenem/vaborbactam and imipenem/relebactam (MIC of 2 µg/mL and 1 µg/mL, respectively) was negative for bla genes encoding class B or class D carbapenemases, thus suggesting the presence of additional resistance mechanisms such as expression of a class A carbapenemase that would be inhibited by the β-lactamase inhibitor in the combination (bla PER or bla VEB were not detected in this isolate either).
Interestingly, Kohira et al. recently suggested that NDM might have a role in cefiderocol non-susceptibility among Enterobacterales and PER might also be involved in a slight MIC increase of cefiderocol in A. baumannii, although the expression of PER enzymes alone did not seem to result in cefiderocol MIC values of >8 µg/mL and PER-positive A. baumannii isolates in such a study were closely clonally related [17]. Notably, only one out of nine isolates harbouring OXA-23/NDM in our study developed resistance to cefiderocol and only 5 out of 24 cefiderocol non-susceptible isolates were positive for bla PER , all but one associated to sequence type 2 (ST2). We also identified a sixth PER-positive meropenem-resistant A. baumannii isolate in our collection that, nevertheless, showed a cefiderocol MIC of 2 µg/mL ( Table 2). The distribution of OXA, NDM and PER enzymes between resistant and susceptible isolates to cefiderocol in our collection, therefore, seemed to suggest that there is no link between cefiderocol-resistance and the carriage of a particular β-lactam-resistance gene.

Geographical and Epidemiological Distribution of Cefiderocol-Resistant Isolates
A. baumannii isolates selected in the present study originated from nine different countries from four different continents in an attempt to provide some degree of epidemiological diversity. In addition, isolates could be grouped in at least 29 different STs from 13 clonal complexes, including isolates from the widespread international clonal complexes CC1, CC2, and CC3 (Table 2) [18]. Interestingly, however, while cefiderocol-susceptible isolates were recovered from all nine participating countries, cefiderocol non-susceptible isolates were only recovered from six out of the nine countries included, and there was a clear predominance towards isolates recovered from Azerbaijan (12/24, 50%) as well as towards isolates belonging to CC2 (15/24, 62,5%), as shown in Figure 1 and Table 2. According to PFGE data, however, there was little clonal homogeneity among the cefiderocol-resistant isolates from Azerbaijan (data not shown) and, in general, there was a strong bias in our collection towards isolates belonging to CC2 as well (62/113, 54.8%), which is just a mere reflection of the worldwide spread of such international clone [18]. For instance, we did not detect cefiderocol-resistance among isolates recovered from Peru, Argentina or Mozambique, despite some of them also belonging to CC2. In view of the above, it is difficult to associate cefiderocol-resistance with a particular clonal lineage, although there seemed to be a certain preference in our study towards isolates from CC2. In this regard, Malik et al. also suggested an association between A. baumannii ST2 isolates and cefiderocol-resistance, linked to the reduced expression of the bacterial siderophore receptor PirA [19]. While the expression of siderophore receptors in different clonal groups has not yet been investigated in our group, it is likely that the emergence of cefiderocol resistance is strain specific and that the putative association with isolates from CC2 arises from the fact that mutations affecting such expression are most likely to be identified first among isolates from predominant clonal lineages, that are far more abundant in clinical settings. Nevertheless, resistance mechanisms in A. baumannii have typically been shown to be multifactorial, often involving modifications on membrane permeability and moiety but also biofilm formation, so additional epidemiological and molecular studies are clearly needed to elucidate the mechanisms behind cefiderocol resistance [20,21].

Minimum Inhibitory Concentration (MIC) Determinations
Antimicrobial susceptibility was performed according to the standard CLSI methodology [22]. After inoculation plates were placed in aerobic conditions and readings were taken after 16-20 h of incubation. The MIC was determined as the lowest concentration of an individual drug that resulted in no visible growth. MIC 50 and MIC 90 values were also calculated and indicated the MIC value of each drug able to inhibit the growth of 50% and 90% of isolates tested, respectively.

Multi-Locus Sequence Typing (MLST)
MLST was performed using the Pasteur scheme for A. baumannii [25]. The allele sequences and STs of selected strains were identified and retrieved from the PubMLST A. baumannii MLST database (http://pubmlst.org/abaumannii/ (last accessed on 17 December 2021). The population structure of STs was assigned to their corresponding clonal complexes (CCs) using the goeBURST software (http://www.phyloviz.net/goeburst/ (last accessed on 17 December 2021)).

Conclusions
Overall, almost 80% of A. baumannii isolates in our study showed susceptibility to cefiderocol, and, more interestingly, as many as 75% of meropenem-resistant isolates were also susceptible to the new drug. Despite the fact that most cefiderocol-resistant isolates belong to CC2, our data seem to suggest that cefiderocol-resistance is strain specific and cannot be related to any particular mechanism of carbapenem-resistance either.
There are already several publications that have investigated the in vitro activity of cefiderocol against clinical isolates of A. baumannii [7,26,27]. Such publications, however, usually fail to consider the clonal relatedness of isolates selected for the study and, given the strong clonal structure of A. baumannii, is it likely that they lack in clonal diversity. In our study we have tried not to focus on including a sheer number of isolates but to provide as much clonal diversity as possible. The susceptibility profiles from our collection of epidemiologically diverse isolates harboring different mechanisms of carbapenem-resistance, nevertheless, reinforce previous results that prompted cefiderocol as a valid alternative to treat infections caused by MDR A. baumannii.