Acinetobacter baumannii (AB) and its close relatives—Acinetobacter calcoaceticus, Acinetobacter pittii, Acinetobacter nosocomialis, Acinetobacter seifertii, and Acinetobacter dijkshoorniae—comprise the Acinetobacter calcoaceticus-baumannii (ACB) complex, which mainly causes hospital-acquired pneumonia and bacteremia (Towner, 2009; Chusri et al., 2014). The ACB complex is regarded as a group of phenotypically indistinguishable opportunistic pathogens. Among the ACB complexes, infections with AB are a major public health threat in health care settings worldwide (Doi et al., 2015). Non-baumannii ACB (NB-ACB) species also have more been reported as pathogens causing invasive infections than before (Chuang et al., 2011; Lee et al., 2011; Said et al., 2021).

In the recent decades, outbreaks caused by carbapenem resistant AB (CRAB) have been reported in many countries. Super-infection by carbapenem resistant AB and other gram negative organsims was observed in up to 20% of the patients hospitalized with coronavirus disease-2019 (COVID-19), and studies have shown that CRAB had independent association with 14-day mortality in COVID-19 respiratory sub-intensive care units (Casale et al., 2023; Iacovelli et al., 2023). The production of carbapenemases is the most common mechanism of carbapenem resistance, and so far, the most common carbapenemases found in AB are OXA-23 (Mugnier et al., 2010). In addition, CRAB isolates carrying NDM-1 have also been rapidly increasing worldwide (Chen et al., 2011; Pillonetto et al., 2014; Alcock et al., 2020). To make matters worse, there have been reports of a global emergence of resistant to colistin, which is essentially the last line of defense in treating extensively-drug resistant (XDR) AB (Cai et al., 2012).

In South Korea, expansion of ST191 and ST451 that belong to clonal complex (CC) 208 in adult patients from multiple centers have been observed. Clones that belong to CC208 have been shown to exhibit higher resistance to antimicrobials, therefore their expansion have been a global concern due to difficulties in treatment (Jun et al., 2021; Jun et al., 2023).

Currently most of the data on invasive AB infections are derived from adults, and data on the molecular epidemiology and antimicrobial resistance mechanism of AB and NB-ACB causing invasive infections in the pediatric population is very limited. Therefore, this study aimed to identify AB and NB-ACB genotypes causing invasive diseases in children and analyze their antibiotic resistome by using whole genome sequencing (WGS).

AB has emerged as a major global threat, especially due to its exceptional ability to acquire resistance genes to all classes of antibiotics currently available. This study aimed to identify circulating AB and NB-ACB genotypes causing invasive diseases in Korean children, analyze their antibiotic resistomes, and investigate the mechanism of carbapenem and colistin resistance. Compared to NB-ACB isolates, AB had a significantly higher mortality rate with an overall low susceptibility to antibiotics, with carbapenem and colistin resistances in 77.6% and 27.1% of the isolates, respectively. In particular, CC92 AB were exclusively CRAB and MDR, with the main resistance mechanism found to be OXA23 β-lactamase. Colistin resistance was not correlated with the presence of pmr gene and/or mutations in lpxA/C genes.

ACB complex is usually reported as AB from the laboratory using commercial bacterial identification system. However, correctly identifying ACB subspecies is important because of the poorer prognosis of AB than NB-ACB infections (Kuo et al., 2013; Xiao et al., 2017; da Silva et al., 2018). In Japan, between 2001 and 2014, among 155 invasive Acinetobacter species strains, 27.1% was A. pittii and 25.8% was AB, and imipenem nonsusceptibility was detected only in 4 strains (Kiyasu et al., 2020). In the previous study in Mexico, a total of 88 strains were identified as ACB complex by VITEK II during 2015–2017. Among them, AB accounted for 89.8%; A. pittii, 6.8%; and A. nosocomialis, 3.4%. In addition, 44.3% were MDR strains and, 11.4% were XDR (Mancilla-Rojano et al., 2020). In the study from Russia and Kazakhstan in 2016–2022, 234 Acinetobacter isolates were identified as NB-ACB isolates, which comprised 6.2% of Acinetobacter spp. Most NB-ACB isolates were susceptible to all antibiotics; however, sporadic isolates were resistant to carbapenems (Sheck et al., 2023). As shown above, the proportion of NB-ACB strain and its subspecies among Acinetobacter species and their AMR vary according to country, period, specimen type, and isolation method. Also, in the current study, among the 68 invasive isolates reported as AB, one-fourth was NB-ACB, which had much lower antimicrobial resistance and mortality than AB. Among the NB-ACB species, the most abundant were A. nosocomialis and A. pittii as in the previous studies (Mancilla-Rojano et al., 2020; Sheck et al., 2023).

AB emerged as a global pathogen due to the successful expansion of a few epidemic lineages producing acquired OXA-type carbapenemases. Particularly ST2 (Pasteur scheme), known as international clone (IC) II, the most common clone globally, are distributed in Europe, Asia, and Latin America (Levy-Blitchtein et al., 2018; Villalon et al., 2019). Genotype replacement with CC92 (Oxford scheme), which is the same strain to ST2 (Pasteur scheme), has been a global phenomenon, especially in China where CC92 CRAB harboring bla _OXA23 has been increasing (Ruan et al., 2013; Chen et al., 2017). The spread of this drug-resistant pathogen should be closely monitored and managed, particularly in critical care. In a previous single center study in Korea, 27 AB were isolated from blood of patients in PICU, and MLST analysis showed that ST138 (Oxford scheme, CC92) was predominant (70%) (Kim et al., 2022). Also in the current study, majority of AB were ST2, which showed XDR characteristics, with compete resistance to 3rd/4th generation cephalosporin, piperacillin-tazobactam, amikacin, ciprofloxacin, and even carbapenem. Furthermore, 36.4% were resistant to colistin.

In our previous study, we identified carbapenem nonsusceptibility as a risk factor for mortality due to AB in children. An early administration of appropriate antibiotics should be enacted, especially in children with neutropenia (Choe et al., 2019). The bla _OXA-23 gene is one of the most prevalent β-lactamase genes on the genome of CRAB and is distributed worldwide (Mugnier et al., 2010; Ruan et al., 2013). Among 86 AB isolates obtained during 2008–2015 in Central Illinois, 70.9% were nonsusceptible to carbapenems. Among the CRAB isolates, bla _OXA-23 was the most frequently detected carbapenemase gene (52%) (Koirala et al., 2020). However, it is suggested that ISAba1 is providing the promoter for bla _OXA-51 and bla _OXA-23 genes (Turton et al., 2006). In the current study, the presence of bla _OXA-23 genes and ISAba1 upstream of bla _OXA-23 were exactly matched to carbapenem resistance in AB. Any other β-lactamase or other AMR mechanism could be differentiated between CRAB and CSAB. We may conclude that the main mechanism of carbapenem resistance in AB isolated in Korean children during 2015–2020 was ISAba1-bla _OXA-23. Because the detection of AB with bla _OXA-23 and MDR phenotye is associated with increased mortality, prompt identification and management is extremely important in the clinical field. Similar results of high mortality associated with MDR AB carrying bla _OXA-23 was observed globally in different STs, and the bla _OXA-23 gene found to be present on a conjugative transferable plasmid indicated the advantage in transferability of the genetic element (Lopes et al., 2015; da Silva et al., 2018).

AmpC β-lactamase coded in bla _ADC gene could induce resistance to 3rd-generation cephalosporin of gram-negative bacteria during the use (Tamma et al., 2019). In the current study, all ACB complex isolates harbored bla _ADC gene, but they did not show resistance to 3rd-genderation cephalosporin unless they also had ISAba1-bla _OXA-23, regardless of the presence of upstream ISAba1. Thus, we may conclude that the harboring of bla _ADC itself is not the requirement of AMR, and upstream ISAba1 is not sufficient to activate bla _ADC within AB.

Because colistin is usually the only remaining treatment available, the emergence of colistin-resistant CRAB is devastating. Colistin resistance due to the acquisition of plasmid-mediated mobile colistin-resistant (mcr) genes has been described in numerous Enterobacterales, especially Klebsiella pneumonia, however also found in AB. In Pseudomonas aeruginosa, mcr-5 has been identified, however, activation of histidine kinase (PmrB) or the response regulaor (PmrA) following an amino-acid substitution has been found to be a key mechanism for colistin resistance (Liu et al., 2016; Jeannot et al., 2021; Novović and Jovčić, 2023). To date only chromosomally encoded colistin resistance mechanisms have been reported in AB. It is suggested that mutations in the pmr and lpx genes confer colistin resistance via modification and complete loss of lipopolysaccharide, respectively (Adams et al., 2009; Moffatt et al., 2010; Nie et al., 2020). In the current study, two isolates were found to harbor the pmr gene. The significant mutations in pmrA gene related to colistin-resistance (D82G and S119T) were not found, but mutations in pmrB gene (G21V and V227A) were found in both isolates. Furthermore, lpxA and lpxC were found in all the isolates, but the change in the amino acid sequence of lpxA found in one isolate, which were unlikely to be associated with colistin resistance mechanisms in our isolates. This shows that there are other mechanisms that have not yet been uncovered, that are associated with colistin resistance.

The most common mechanism of resistance to fluoroquinolone in AB involves alterations in the genes that encode subunits of the quinolone targets DNA gyrA and parC (Park et al., 2011). Phenotypical quinolone resistance was perfectly matched to both gyrA and parC gene mutations in the current study. All quinolone resistant AB were CRAB, which harbored ISAba1-bla _OXA-23. None of CSAB had either gyrA or parC gene mutations.

There are two main mechanisms underlying aminoglycoside resistance in AB: aminoglycoside-modifying enzymes, which were encoded by aac, aad, ant, and aph genes, and antibiotic target alteration by 16S rRNA methylases (encoded by armA) (Bakour et al., 2014). In this study, genes encoding aminoglycoside-modifying enzymes were found in 30.3%~69.7% of CC92 AB, 18.8%~31.3% of non-CC92 AB, and none of NB-ACB. The armA gene was also detected in 100%, 37.5%, and 0%, respectively. In a previous report from 2 hospitals in South Korea, invasive AB isolates from adults during 2004–2005 showed 85.2% of harboring armA (Cho et al., 2009).

The resistance-nodulation-cell division (RND) pumps are the most prevalent efflux pumps in AB and one of the most important determinants of MDR. Typically, efflux pumps (especially those chromosomally encoded) are controlled by negative regulators. Mutations within the negative regulators which releases the ‘block’ on the efflux pump leads to increased expression of the pump causing antibiotics resistance (Fernando and Kumar, 2013; Davin-Regli et al., 2021; Kornelsen and Kumar, 2021). A high prevalence of this pump was observed in up to 97% of MDR AB isolates worldwide (Nowak et al., 2015). In the invasive AB strains from this study, the ade RND pumps were present in 100%. Furthermore, abe, which is a chromosomally encoded drug efflux pump of the small multidrug resistance (SMR) family, was also found in 100% of the isolates in this study. Although just the presence of these efflux pumps did not correlate with MDR in AB, it might show us the potential of AB to be MDR pathogen whenever they get the certain promotor.

Rapid detection of AB and main carbapenemase genes can allow prompt intervention in patients with invasive infections. By identifying whether the identified AB is MDR phenotype, early initiation of adequate antibiotics can improve patient outcome. Commercial molecular rapid testing are available and have already implemented in routine diagnostics in many centers (Dunbar et al., 2022). Improvements have been made in providing rapid phenotypic AST (Briggs et al., 2021), the benefits and limitations of these rapid platforms are currently under study and more studies are needed in critically-ill children.

There were several limitations of this study. First, although the three centers included were representative centers in South Korea with the highest number of beds for treating critically ill children in the pediatric intensive care unit, the resistance profiles of AB causing nosocomial infections may differ at each hospital throughout Korea. Second, the total number of cases included is limited, and because this was a multicenter study, we were unable to obtain all clinical data including information on treatment regimens and duration for all patients. However, this is the first study to investigate, via WGS, the antibiotic resistome profiles of AB dominant in strains isolated from critically ill children, and is deemed to be valuable for the monitoring and treatment of invasive AB infections.

To conclude, we found that CC92 AB have an exceptional ability to acquire a wide variety of antibiotic resistomes which enhance their AMR and survival fitness. Non-CC92 isolates also harbored a variety of resistance genes for which they were phenotypically susceptible to, depicting the necessity for continuous monitoring. Many efforts are needed to prevent their outbreak in critical patients, including methods such as decreasing overall antibiotic use and relieving selective pressure, strict contact precaution and hand hygiene of medical staff to prevent transmission, and environmental disinfection to decrease sources of possible infection. Carbapenem resistance in AB was likely induced by blaOXA-23 carbapenemases promoted by ISAba1. Although only one-third of the isolates were shown to be resistant to colistin, we found that all AB harbored mutations in enzymes that produce lipid A, meaning all AB have the potential to become resistant to colistin. Further studies are necessary to understand resistance mechanisms in order to combat the highly virulent XDR AB, and evaluate susceptibility to novel approved drugs such as cefiderocol and sulbactam/durlobactam.

Original datasets are available in a publicly accessible repository:The original contributions presented in the study are publicly available. This data can be found here: NCBI, PRJNA1127342.

The studies involving humans were approved by institutional review board of Seoul National University Hospital (no. H-1812-080-995). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

HK: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Validation, Writing – original draft, Writing – review & editing. KK: Data curation, Investigation, Methodology, Writing – review & editing. GK: Data curation, Investigation, Methodology, Writing – original draft. D-GL: Data curation, Investigation, Methodology, Writing – review & editing. YK: Data curation, Formal Analysis, Investigation, Supervision, Writing – review & editing. EC: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Project administration, Supervision, Writing – review & editing. JL: Data curation, Formal Analysis, Investigation, Methodology, Supervision, Validation, Writing – review & editing. KW: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Validation, Writing – original draft, Writing – review & editing.