Between March and July 2024, an outbreak of CRAB was observed among hospitalized patients in a single ward at IRCCS Fondazione Ospedale Maggiore Policlinico, Milan, Italy. Thirty-three patients (on 280 admissions) had at least one CRAB isolation during hospitalization (Supplementary Fig. 1). They were mainly male (25/33, 75.7%) with a median age of 79 years (interquartile range: 71–86). Of the 33 CRAB isolates, 29 were characterized for whole genome sequencing (WGS) according to Supplementary text [10]. The CRAB isolates represented an infection in 9/29 (31.0%) patients and a colonization in the remaining 20/29 (68.9%).

Genomic analysis of the 29 non-replicated CRAB collected during the outbreak revealed that all isolates belonged to the Pasteur sequence type 2 (ST2). According to the Oxford scheme, 24/29 (82.7%) isolates belonged to the ST208/1806, characterized by a capsular polysaccharide (K) “2” and a lipooligosaccharide outer core locus (OCL) “1”. Four isolates belonged to the ST369, characterized by the KL9 and OCL1. The remaining isolate belonged to ST195/1816, characterized by the KL3 and OCL1.

FDC susceptibility was affected only in the four ST369 isolates by the Kirby-Bauer disk diffusion (zone diameters: 13 mm) [11]. The commercial UMIC® Cefiderocol (Bruker Daltonics, Bremen, Germany) method confirmed these results, assigning to the ST369 isolates a MIC equal to 4 mg/L. Of note, core genome analysis revealed the presence of a stop codon at amino acid position 376 preceded by a stretch of 8 amino acid mutations in the siderophore-drug receptor PirA in all four ST369 strains (E376stop), producing a premature signal to stop translation. Moreover, this stop codon is followed by a 52 amino acid deletion located between the amino acids 377 and 428 of PirA (Fig. 1). This deletion in the siderophore-drug receptor gene causes the loss of a domain involved in the binding of the siderophores, located at the amino acids 402–406 of the Acinetobacter baumannii PirA protein, and the complete and partial loss of the strand β9 and β8, respectively [12]. At the end of this deletion, at position 429 another Methionine was found, suggesting the production of a second truncated protein, containing the loops L7 and L11. As known, the binding site of siderophore receptors is formed by loops NL1, NL2, and NL3 of the plug domain and loops L3, L7, and L11 and strands β7, β8, and β9 of the β-barrel [12,13,14,15]. Thus, these alterations in PirA could result in truncated or highly altered proteins with poor functional siderophore receptor binding sites, and thus likely linked to weaker transportation of FDC and a decrease in FDC susceptibility. This is consistent with the FDC susceptibility results, defined by a zone diameter of 13 mm by disk diffusion and the MIC of 4 mg/L. SNP analyses using Snippy (https://github.com/tseemann/snippy) did not identify other relevant mutations in the iron-uptake genes with respect to ST369 reference sequences. Other amino acid variations in the efflux pump (EmrA/EmrK, n = 1), porin genes (OprD, n = 6), and penicillin-binding protein 1B (n = 1) were detected (Supplementary File 1).

Amino acid alignment of the wild-type PirA and the mutated PirA, detected in the transmission chain of the four ST369 CRABs. An asterisk shows the premature stop codon. The deletion at amino acids 376–428 is shown by dashes

The evaluation of clonal relatedness by a core-genome Maximum Likelihood (ML) tree including 22 ST369 reference strains (Supplementary Table 1) revealed that the four ST369 isolates were overall highly related (Supplementary Fig. 2, panel A) and formed a clear transmission cluster, characterized by a distance of 3 (IQR: 1–6) SNPs (Supplementary Fig. 2, panel B and C). All strains shared also a set of antimicrobial resistance and virulence genes, including the acquired carbapenemase gene blaOXA-23, the blaOXA-66 (OXA-51-like variant) together with genes conferring resistance to aminoglycosides (aph(3′′)-Ib and aph(6)-Id), and tetracyclines (tetA), the cephalosporinase ADC-30, the multidrug efflux pumps of the resistance-nodulation-division (RND) family (adeABC, adeFGHI, and adeJKL), the Non-RND efflux systems abeM and abeS, the sul2 sulphonamide resistance gene, as well as core virulence factors, usually displayed by Acinetobacter baumannii strains [16, 17]. The four ST369 isolates did not carry the blaTEM gene, the beta-lactamase ADC-74, the aminoglycoside-modifying enzyme armA gene, the macrolide resistance genes mphE and msrE, as well as the pse-pathway among the virulence genes. The antimicrobial resistance gene pathway described here was concordant with the MIC values obtained (Supplementary Table 2).

Regarding mobile genetic elements, all strains carried the plasmids tig00000534_pilon and pORAB01-2 (accession numbers: CP026708 and CP015485), and the prophage PHAGE_Acinet_Bphi_B1251_NC_019541.

Surprisingly, all the ST369 isolates were collected from FDC naïve individuals admitted to the medicine ward from May to July 2024 (Fig. 2). All individuals but one received previous antimicrobial therapy. Unfortunately, no information is available regarding the antimicrobial drugs given in the months preceding CRAB detection, which could have contributed to lower susceptibility to FDC.

Estimated maximum likelihood phylogenetic analysis of the four Cefiderocol (FDC) non-susceptible carbapenem-resistant Acinetobacter baumannii isolates. The maximum likelihood was inferred from a core genome alignment of 3,344,425 bp. The phylogeny was estimated with IqTree using the best-fit model of nucleotide substitution HKY + F with 1,000 replicates and fast bootstrapping. The numbers on leaves represent the sample IDs, and bootstrap values are shown on branches. Information regarding the samples was reported: date of isolation, the sequence type (ST), capsular locus (K locus) and lipooligosaccharide outer core (OC locus), the presence (solid squares) or absence of antimicrobial resistance genes, the presence of deletion in pirA gene, the Cefiderocol-non-susceptibility, tested by disk diffusion (according to the last EUCAST guidelines, https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/Breakpoint_tables/v_14.0_Breakpoint_Tables.pdf) and the minimal inhibitory concentration (MIC) value by the UMIC (Bruker) commercial test

Two of the four patients experienced a CRAB-mediated infection. One was treated with an initial FDC-containing regimen; clinical conditions did not improve since the administration of alternative antimicrobial treatment (clinical case manuscript is in progress).

The first ST369 case was isolated in May in a patient admitted to the ward a few days before, without CRAB surveillance being performed at their admittance. Thus, a CRAB acquisition outside the medicine ward cannot be excluded. In the remaining three cases, CRAB at patients’ admission was negative and its isolation occurred more than 48 h from admission, suggesting their acquisition inside the medicine ward.

A modest rate of evolution was observed when the isolates were analyzed alone without reference genomes (Fig. 2). By this analysis, modest genomic divergence can be observed in ID2291 with respect to the other isolates, due to the acquisition of 19 new SNPs and three deletions (Supplementary File 2), and the loss of the chromosomal-encoded aminoglycoside phosphotransferase aph(3’-Ia) gene. This is consistent with the time of acquisition of CRAB in this patient, occurring two months after the first putative case.

As a limitation of our findings, to define the MIC for FDC we used a commercially available test on iron-depleted cation-adjusted Mueller Hinton broth microdilution not yet recommended by EUCAST [18]. A recent paper that evaluated the performances of commercial tests in testing FDC activity [19], showed that the test we used produced few very major errors with respect to the other MIC-based methods and exhibited the highest essential agreement with the standard broth microdilution method. Moreover, this MIC-based test was used in combination with the disk diffusion method, recommended by the EUCAST guidelines v.14.0 [11]. Both tests were concordant in the definition of FDC-non-susceptibility. As a further limitation, we cannot speculate about the event that caused the siderophore-drug receptor PirA alteration, like previous non-FDC antimicrobial treatment exposures, as described in Pseudomonas aeruginosa [5, 6], or other selection mechanisms. Indeed, even if three out of the four received previous antimicrobial therapy, information about the class and type of antimicrobial treatment is unavailable. Information about potential transmission pathways is limited and the environment was not sampled for CRAB presence.

Finally, even though the alterations detected in PirA can suggest a role in affecting the susceptibility to FDC, the perfect correlation between changes in siderophores systems and FDC MICs remains to be defined [20], especially when there may be confounding factors such as levels of expression of beta-lactamases (i.e. AmpC, OXA-23, or OXA-66) or efflux pumps.