In vitro activity of cefiderocol against carbapenem-resistant Acinetobacter baumannii carrying various β-lactamase encoding genes

Objectives This study aimed to determine the in vitro efficacy of cefiderocol in carbapenem-resistant Acinetobacter baumannii (CRAB) isolates and evaluate the disk-diffusion (DD) method as an alternative method to broth-microdilution (BMD). Methods Totally 89 CRAB isolates were included. Cluster analysis was determined by Pulsed-Field Gel Electrophoresis (PFGE). Resistance genes; blaOXA−51, blaOXA−23, blaOXA−24, blaOXA−58,blaPER−1, blaNDM, blaIMP and mcr-1 were screened. Cefiderocol susceptibility testing was performed by both DD and BMD. Interpretation was made according to EUCAST and CLSI. Categorical agreement (CA), minor errors (mEs), major errors (MEs), and very major errors (VMEs) were determined. Results PFGE revealed 5 distinct pulsotypes; 86 of the isolates were extensively drug-resistant (XDR). All the isolates were negative for blaNDM, blaIMP, mcr-1, while positive for blaOXA−58 and blaOXA51. blaPER−1 was positive for 33.7%; blaOXA−23 for 74.2%; blaOXA−24 for 12.3%. According to CLSI, the MEs rate was 1.85%, mEs was 7.86% and there were no VMEs. According to EUCAST, MEs rate was 3.70%, there were no mEs and VMEs. CA was 91% for CLSI and 97.8% for EUCAST. MICs of cefiderocol against A. baumannii isolates ranged from 0.06 to > 128 mg/L, with MIC50 and MIC90 values of 0.5 and > 128 mg/L, respectively. Conclusions Cefiderocol susceptibility was 60.7% in CRAB isolates. MIC50, MIC90 of blaPER−1 positive and blaPER−1 negative groups were > 128/>128 and 0.25/>128 mg/L. A correlation between the presence of blaPER−1 and cefiderocol resistance was observed (p < 0.0001). Among colistin-resistant isolates, the presence of blaPER−1 was 47.1% and 75% of them were resistant to cefiderocol respectively. Supplementary Information The online version contains supplementary material available at 10.1007/s10096-024-04831-w.


Introduction
Antimicrobial resistance (AMR) is one of the most serious worldwide public health threats that poses significant challenges to healthcare systems including mortality, morbidity, and healthcare costs [1].It was estimated that about 5 million deaths were associated with bacterial AMR in 2019 [2].World Health Organization (WHO) published a list of antibiotic-resistant priority bacterial pathogens that pose the greatest threat to human health, in which multi-drug-resistant (MDR) strains of A. baumannii.This situation has forced clinicians to use colistin as one of the last therapeutic options in the fight against such infections [9,10].According to WHO's 2020 Global Antimicrobial Resistance Surveillance System data, the rate of colistin resistance worldwide reaches 37% in some countries such as Lithuania.In a study conducted in Turkiye, A. baumannii was reported as the most common isolate (25.1%) of bacterial lower respiratory tract (LRT) infection and the most effective antibiotic was colistin with a resistance rate of 12.9% [11].The mechanisms of colistin resistance may be due to chromosomal mutations (such as lpxACD, pmrB and pmrA) or transferable plasmid genes called mcr, which can be spread between bacteria by horizontal transfer [12].With the emergence of resistant isolates, last resort carbapenem and lipopeptide class antibiotics are no longer effective, highlighting the urgent need for new and effective treatment options.
Cefiderocol is a newly developed siderophore cephalosporin [13].Like other β-lactam antibiotics, it binds to penicillin-binding proteins (especially PBP-3) and has a bactericidal effect on bacterial cell wall synthesis [14,15].What makes it unique is its increased transition to the bacterial periplasmic area by iron transport systems with its siderophore-like property [14,15].It owes this feature to the catechol part, which is located in the 3rd position of the side chain in its molecular structure and binds to ferric iron [13,16].It shows increased stability against β-lactamases due to the catechol group [14][15][16].Although various mechanisms including amino acid changes in PBP-3, changes in the transcription levels of iron transport protein-encoding genes such as fiu, feoA, and feoB, and downregulation in the TonB-ExbB-ExbD energy transfer system cause cefiderocol resistance in A. baumannii [17][18][19], it is emphasized that some β-lactamases contribute to cefiderocol resistance [17,19].PiuA, a TonB-dependent siderophore receptor, and PirA, a ferric-enterobactin receptor, are very important in cefiderocol uptake [19][20][21].In various studies, disruptions in the piuA gene (such as a frameshift mutation resulting in stop codon formation (K384fs) and F174Y substitution) or in the fepA gene, an ortholog of the pirA gene, (such as a transposon insertion, P635-ISAba125) caused increases in cefiderocol MICs and resistance [20][21][22][23].In another study, piuA and pirA gene deletion mutants showed increased MICs against siderophore-drug conjugates compared to the standard strain without the deletion [22].PER-1 as an Ambler class A ESBL, NDM as an Ambler class B metallo-β-lactamase, and OXA-23, OXA-24, and OXA-58 as Ambler class D carbapenemases have been reported among important β-lactamases to cefiderocol resistance [17,18,24].
In this study, it was aimed to evaluate the in vitro efficacy of cefiderocol in CRAB isolates with a various resistance gene profile.Broth-microdilution (BMD) is recommended as the reference standard method for cefiderocol susceptibility testing [25,26], and iron depleted-cation adjusted Mueller Hinton broth (ID-CAMHB) should be used which is not available as ready-to-use.It was also aimed to evaluate disk-diffusion (DD) as an alternative to the BMD.

Material and method
This study was approved by the Baskent University Institutional Review Board (Project no: KA22/392) and supported by Baskent University Research Fund.

Bacterial isolates
A total of 89 CRAB isolates from the culture collection of Medical Microbiology Laboratory of Baskent University Medical Faculty collected between January to December 2021 were included in the study.The 32.6% (n = 29) of the collected isolates were obtained from lower respiratory tract samples, 23.6% (n = 21) from blood, 20.2% (n = 18) from urine,13.5% (n = 12) from wound, 5.6% (n = 5) from catheter, and 4.5% (n = 4) from biopsy culture.Only one isolate for each patient was included.Identification and antimicrobial susceptibility testing (AST) were performed by the BD Phoenix 50 (BD Diagnostics, USA) automated bacterial identification and antibiotic susceptibility system.The identification was confirmed by the presence of the chromosomal bla OXA−51 gene by PCR.All isolates were resistant to meropenem and imipenem, resistance was also screened by DD.MICs of colistin were determined by the Thermo Fisher Scientific Sensitivity FRCOL kit (Thermo Fisher Scientific, USA).All the AST results were assessed using EUCAST standards [27].

Cluster analysis
Genetic relations among clinical CRAB isolates were determined by PFGE as previously described [28].Genomic DNA was treated with lysis solutions and digested with ApaI restriction enzyme (Takara, Japan).Electrophoresis was performed in a PFGE system (CHEF-DR III System; Bio-RAD, USD).The gel images were processed and cluster analysis was performed using BioNumerics software.The images were normalized by using standard molecular markers, and banding patterns were compared.Pulsotype typing was performed by the interpretative criteria of Tenover et al. [29].

Detection of antibiotic resistance genes by PCR
Bacterial DNA was extracted by boiling the bacterial suspension at 95 o C for 10 min [30].After centrifugation, the supernatant was transferred to new tubes.Amplification of OXA genes, bla OXA−51 , bla OXA−23 , bla OXA−24 , and bla OXA−58 were performed as previously described [31] The bla PER−1 , bla NDM , bla IMP, and mcr-1 genes were screened as described previously [32][33][34].The primer sequences and amplicon sizes were given in supplementary material.Confirmed positive strains of A. baumannii carrying bla OXA51 gene [35], K. pneumoniae CDC529 carrying bla NDM−1 , K. pneumoniae CDC309 bla IMP , and A. baumannii ATCC19606 were used as positive control strains.The PCR products were analyzed by 2% agarose gel containing EtBr.

Cefiderocol susceptibility detection by Kirby Bauer DD method
Cefiderocol susceptibility testing was performed by using MASTDISCS (30 µg) (Mast Group, UK) and cationadjusted Mueller Hinton agar (CA-MHA) (Oxoid, UK).Plates were incubated for 18-24 h at 35 ± 2 °C in aerobic conditions.Zone diameters were evaluated in the reflected light and from the back of the plate.Interpretations of zone diameters were made according to both EUCAST and CLSI [27,36].

Cefiderocol MIC detection by BMD
MIC detection was performed by using ready-to-use broth microdilution plates, ComASP®, available in the range of cefiderocol 0.008-128 mg/L (Liofilchem, Italy) according to the recommendation of the manufacturer.Briefly, a suspension of 0.5 McFarland standard was prepared from fresh bacterial colonies grown in overnight incubation.This prepared suspension was diluted 1:20 in 0.9% NaCl.0.4 ml of the diluted bacterial suspension was added to the tube containing 3.6 ml of ID-CAMHB supplied in the kit.From this suspension, 100 µl were distributed into each well, containing cefiderocol at different concentrations between 0.008 mg/L to 128 mg/L.The plates were incubated for 16-20 h at 35°±2 °C in aerobic conditions.The first well in which the turbidity first disappeared or the turbidity showed a significant decrease compared to the growth control well due to the "trailing effect" was considered as the MIC value.When the "Trailing effect" was observed, 80% inhibition was accepted as the MIC value [25].MIC value was evaluated in the reflected light from the front of the sensititre plate.Interpretation was made according to both EUCAST and CLSI criteria [27,36].E. coli ATCC 25922 and P. aeruginosa ATCC 27853 were used as quality control strains for both methods.

Statistical analysis
Pearson and Likelihood Ratio Chi-square test were used for the comparison of categorical variables between groups, and the significance test of the difference between two proportions was used for the comparison of only one result within two groups.Categorical variables were expressed as f (%).Statistical analysis was performed by using Statistical Package for Social Sciences (SPSS) for Windows Ver25.0.In statistical analysis, the significance level was taken as p < 0.05.
(33.7%) isolates were positive for bla PER−1 gene and 25 (83.3%)bla PER−1 positive isolates were resistant to cefiderocol.While 10 (11.2%) of 59 bla PER−1 negative isolates were resistant to cefiderocol.MIC50 and MIC90 values of bla PER−1 positive and bla PER−1 negative isolates were > 128/>128 and 0,25/>128 mg/L respectively.Totally 66 (74.2%) isolates were bla OXA−23 positive and 28 (42.4%) of them were resistant to cefiderocol.While 7 (30.4%) of 23 bla OXA−23 negative isolates were resistant to cefiderocol.The 11 isolates were bla OXA−24 positive and 1 (9.1%) of them was resistant to cefiderocol while 34 (43.6%)bla OXA−24 negative isolates were resistant to cefiderocol.Table 3 shows the MIC distributions of isolates in the presence of resistance genes and Table 4 shows the susceptibility rates in the presence of resistance genes individually and in combinations.Cefiderocol activity of colistin-resistant isolates in the presence and absence of the bla PER−1 gene were also evaluated and shown in Table 5.

Discussion
Cefiderocol is a siderophore cephalosporin that possesses potent activity against carbapenem-resistant and MDR Enterobacterales, and non-fermentative Gram-negative bacilli [14,16].Cefiderocol received Food and Drug Administration (FDA) approval for the treatment of urinary tract infections in 2019 and treatment of hospital-acquired pneumonia and ventilator-associated pneumoniae in 2020.Although there are methods such as DD and gradient diffusion to test the susceptibility of cefiderocol, the gold standard method is the BMD [25,26].Standard cation-adjusted Mueller Hinton broth is not controlled for iron concentration and iron-depleted cation-adjusted MHB (ID-CAMHB) is required for susceptibility results to be accurate [26,38].The CLSI has approved ID-CAMHB and provides instruction to prepare for determining cefiderocol MICs but preparation of this medium is difficult and impractical for the daily

Susceptibility of cefiderocol by DD and BMD
According to CLSI [36], cefiderocol MIC value of ≤ 4 mg/L in A. baumannii is considered as S, 8 mg/L as I, and ≥ 16 mg/L as R; cefiderocol zone diameter of ≥ 15 mm in A. baumannii is considered as susceptible (S), 11-14 mm as intermediate (I), and ≤ 10 mm as R.There is no I category in EUCAST, and resistance of A. baumannii based on PK/ PD values while recommending breakpoints as ≤ 2 mg/L S and > 2 mg/L as R; cefiderocol zone diameter ≥ 17 mm as S [27].MICs of cefiderocol against A. baumannii isolates ranged from 0.06 to > 128 mg/L, with MIC50 and MIC90 values of 0.5 and > 128 mg/L, respectively.The results of cefiderocol susceptibility test according to both EUCAST and CLSI were given in Table 1.
According to EUCAST, 2 R isolates (3.70%) (zone diameters 15 and 6 mm) by DD were found as S (MICs:2 and 0,5 mg/L) by BMD indicating MEs.There are no VMEs according to EUCAST.Also, there are no mEs due to lack of an I category in EUCAST.Totally 87 (97.8%) isolates were in CA.Table 2 shows the VME, ME, mE, and CA rates of the DD according to CLSI and EUCAST.

In vitro cefiderocol activity according to gene profile
The contribution of bla PER−1 , bla OXA−23 , and bla OXA−24 genes to cefiderocol resistance was evaluated.Totally 30 1 3 laboratory workflow.Also, some gradient diffusion strips were developed but they were reported successful only in P. aureginosa [26].For all these, DD was also tested for all the isolates included in the study as an alternative method in which standard CA-MHA can be used in the cefiderocol susceptibility testing.
If CLSI is considered in terms of compatibility of DD and BMD; the MEs was 1.85%; mEs was 7.86%, there was no VME while the CA was 91%.Acceptance criteria are based on ≥ 90% CA, ≤ 10% mEs, ≤ 3% MEs, and ≤ 1.5% VME [37]; therefore, the DD seems as a candidate to be a practical and easily applicable alternative to the BMD.But based on CLSI, a total of 7 isolates were in category I according to the DD, while all of these isolates were determined as R according to the BMD; so this result indicates that the results interpreted as I by DD might be R and it would be better to confirm them with the BMD.
According to EUCAST, the MEs was 3.70%, there was no mEs and VME while the CA was 97.8%.With an interpretation according to EUCAST, the DD does not seem as a candidate alternative to the BMD although the CA value was found to be higher than CLSI, the MEs rate (3.7%) was observed above the limit value (3.0%).This might be resulted from the lack of category I in EUCAST which ensures no mEs but causes more MEs.
In a study including CRAB isolates, DD was compared with BMD and HardyDisks with MASTDISCS, and CA was slightly higher in MASTDISCS than in HardyDisks [38].They also found that for DD, EUCAST breakpoints demonstrated lower susceptibility at 66% compared to CLSI breakpoints with 89% for MASTDISCS.For MASTDISCS, EUCAST breakpoints resulted in the highest CA, however due to the lack of an I category, higher rates of MEs and VMEs were observed.They have indicated that both DD performed poor activity for A. baumannii complex.DD offers a convenient alternative approach to BMD at gramnegative bacteria for cefiderocol AST, with the exception of A. baumannii complex [38].The results of this study also showed CLSI to be more appropriate than EUCAST as there is no I category [38].Although DD was not reported as an alternative to BMD in A. baumannii complex in this study; according to our results DD can be an alternative to the BMD if CLSI is used for A. baumannii.
The cefiderocol resistance rate was higher in bla OXA−23 positive 28/66 isolates (42.4%) than bla OXA−23 negatives 7/23 (30.4%) but the difference was not statistically significant (p > 0.05).The contribution of bla OXA−23 has been emphasized in the co-existence of bla OXA−23 with other β-lactamase genes [18,43].The co-existence of bla PER−1 and bla OXA−23 rate was 30.4% in the current study and 81.5% of them were resistant to cefiderocol.The PER-1 plays a pivotal role in cefiderocol resistance.All bla PER−1 positive bla OXA−23 negative isolates (n = 3) were resistant to cefiderocol.Among bla PER−1 negative bla OXA−23 positive criteria.Higher resistance rates might have resulted from the fact that 96.6% of the isolates were XDR and the high prevalence of bla PER−1 gene (33.7%;p < 0.0001) in isolates may have contributed to cefiderocol resistance.The limitation of the study is that cefiderocol resistance mechanisms were not demonstrated by performing WGS on cefiderocolsusceptible and resistant isolates.
Cefiderocol is known for its high stability to β-lactamases, including ESBLs, AmpC, and carbapenemases [14][15][16].But it was seen that the presence of bla PER−1 might result higher resistance rate than expected.Even in vitro activity may not reflect the clinical activity.In a study that compared the patient groups treated with the regimens containing cefiderocol to colistin, more microbiological failures were observed -especially in monotherapy-in cefiderocol treated group.All 46 A. baumannii isolates were found as sensitive in vitro according to CLSI and EUCAST.During treatment, microbiological failure was observed in 17.4% (8/46) of patients, and cefiderocol resistance was observed in 8.7% (4/46) [41].
Although many resistance mechanisms have been defined [18,19,21,42], it is stated that the presence of certain β-lactamase genes might be a contributing factor in cefiderocol resistance.Among them, bla PER−1 , bla NDM , isolates (n = 39) the cefiderocol resistance rate was 15.4%.In isolates lacking both genes cefiderocol resistance rate was 20%.So, the presence of the bla OXA−23 gene seems to be only significant in association with the presence of bla PER−1 in cefiderocol resistance in A. baumannii (p < 0.0001).There was no statistically significant difference in cefiderocol resistance between the presence and absence of bla OXA−24 (p > 0.05).The co-existence of bla PER−1 and bla OXA−24 was not detected and among the isolates lacking both genes, cefiderocol resistance rate was significantly lower by 18.8% (p < 0.0001).There seemed to be no bla OXA−24 impact on cefiderocol resistance in A. baumannii isolates.
Colistin is the last resort for the treatment of MDR and XDR A. baumannii infections [9,10] and the rate of colistin-resistant A. baumannii was observed as 12.9% in a study [11].The mechanism of colistin resistance is due to chromosomal mutations (such as lpxACD, pmrB, and pmrA) or transferable plasmid genes called mcr-1, which can be spread between bacteria by horizontal transfer [12].The 19.1% (17/89) isolates were resistant to colistin and none of the A. baumannii isolates was positive for plasmidmediated mcr-1 resistance.The whole genome sequence of one of our colistin-resistant isolates showed that the colistin resistance resulted from the mutation in pmrA gene as G75A, A120G, G192A, C405T, T450C, C463T and C633T (unpublished data).In another study, 8.2% of A. baumannii isolates were found to be resistant to colistin, and similar to our result, in mcr-1-5 genes were not detected [45].Cefiderocol was also declared as an alternate agent against colistin-resistant A. baumannii isolates and 64.7% (11/17) of the colistin resistant isolates and 59.7% (43/72) colistinsusceptible isolates were susceptible to cefiderocol which is similar (p > 0.05).The unexpected cefiderocol resistance rate (40.3%) in colistin-susceptible isolates again resulted from the presence of bla PER−1 gene was 30.6% (22/72), and 86.3% (19/22) of these isolates were resistant to cefiderocol.Among colistin-resistant isolates the presence of bla PER−1 was 47.1% (8/17), and 75% (6/8) of them were resistant to cefiderocol while all 9 bla PER−1 negative colistin R isolates were susceptible to cefiderocol (Table 5).The efficacy of cefiderocol against colistin-susceptible or resistant isolates is inversely proportional to the presence of bla PER−1 .These results indicate that cefiderocol may be effective in colistinresistant isolates; suggesting that the efficacy is related to the presence of the bla PER−1 gene rather than the colistin resistance status in A. baumannii.

Table 1
Cefiderocol susceptibility results according to EUCAST and CLSI.

Table 3
MIC distributions of isolates in the presence of resistance genes

Table 4
Cefiderocol susceptibility rates in the presence of resistance genes individually and in combinations

Table 5
Colistin and cefiderocol susceptibilities of the isolates in the presence and absence of the PER-1