Molecular Epidemiology of Carbapenem-Resistant Acinetobacter baumannii Isolates from Northern Africa and the Middle East

At the Bundeswehr Hospitals of Hamburg and Westerstede, patients repatriated from subtropical war and crisis zones of Northern Africa and the Middle East were medically treated, including microbiological assessment. Within a six-year interval, 16 Acinetobacter spp. strains, including 14 Acinetobacter baumannii (Ab) isolates with resistance against carbapenems and origins in Afghanistan (n = 4), Iraq (n = 2), Libya (n = 2), and Syria (n = 8) were collected. While clonal relationships of Libyan and Syrian strains had been assessed by superficial next generation sequencing (NGS) and “DiversiLab” repetitive elements sequence-based (rep-)PCR so far, this study provides core genome-based sequence typing and thus more detailed epidemiological information. In detail, sequencing allowed a definitive species identification and comparison with international outbreak-associated Ab strains by core genome multi locus sequence typing (cgMLST) and the identification of MLST lineages, as well as the identification of known resistance genes. The sequence analysis allowed for the confirmation of outbreak-associated clonal clusters among the Syrian and Afghan Ab isolates, indicating likely transmission events. The identified acquired carbapenem resistance genes comprised blaOXA-23, blaOXA-58, blaNDM-1, and blaGES-11, next to other intrinsic and acquired, partly mobile resistance-associated genes. Eleven out of 14 Ab isolates clustered with the previously described international clonal lineages IC1 (4 Afghan strains), IC2 (6 Syrian strains), and IC7 (1 Syrian strain). Identified Pasteur sequence types of the 14 Ab strains comprised ST2 (Syrian), ST25 (Libyan), ST32 (Iraqi), ST81 (Afghan), ST85 (Libyan), and ST1112 (Syrian), respectively. In conclusion, the study revealed a broad spectrum of resistance genes in Ab isolated from war-injured patients from Northern Africa and the Middle East, thereby broadening the scarcely available data on locally abundant clonal lineages and resistance mechanisms.


Introduction
During the last two decades, military conflicts have affected many regions of Northern Africa, the Middle East and neighboring countries, including Afghanistan, Iraq, Libya, cgMLST analysis, which showed that the isolates clustered with the previously identified international clonal lineages IC1 (the four Afghan strains), IC2 (six out of seven Syrian strains), IC7 (one out of two Libyan strains) [36], as shown in Figure 1, while for three isolates (one from Iraq, one from Syria, one from Libya), no matching IC could be identified.
"Bundeswehr Hospital Westerstede" and HBG for "Bundeswehr Hospital Hamburg"), the clonal relationship that had already been suggested by rep-PCR and superficial sequencing (for comparison, also see Figure 2 in reference [33]) could be confirmed by cgMLST, making transmission events prior to hospital admission likely.
All isolates were typed based on two 7-loci multi-locus sequence typing (MLST) schemes, i.e., the Oxford scheme and the Pasteur scheme (Table 1) [37,38]. Notably, six of the A. baumannii were Pasteur ST-2 which corresponds to IC2, but were subdivided into four Oxford STs, while four isolates were Pasteur ST-81 (Oxford ST-498) which is a single locus variant of ST-1 and corresponds to IC1. The other isolates, including the non-A. baumannii, were represented by single STs.  Of note, cgMLST ( Figure 1) confirmed the close clonal relationship of the four IC1 strains which had been isolated in the course of a small local outbreak in Afghanistan, suggesting a common source of infection. Regarding the Syrian A. baumannii strains of the IC2 cluster comprising isolates from two different Bundeswehr Hospitals (WEST for "Bundeswehr Hospital Westerstede" and HBG for "Bundeswehr Hospital Hamburg"), the clonal relationship that had already been suggested by rep-PCR and superficial sequencing (for comparison, also see Figure 2 in reference [33]) could be confirmed by cgMLST, making transmission events prior to hospital admission likely.
All isolates were typed based on two 7-loci multi-locus sequence typing (MLST) schemes, i.e., the Oxford scheme and the Pasteur scheme (Table 1) [37,38]. Notably, six of the A. baumannii were Pasteur ST-2 which corresponds to IC2, but were subdivided into four Oxford STs, while four isolates were Pasteur ST-81 (Oxford ST-498) which is a single locus variant of ST-1 and corresponds to IC1. The other isolates, including the non-A. baumannii, were represented by single STs. Table 1. Analysis of antimicrobial resistance determinants, ordered by strain, MLST type, and the international clonal lineage of the assessed Acinetobacter spp. isolates. N.a. = not applicable, ST = Sequence type, IC = International clone. Clusters are indicated in the same colors as in Figure 1.

Identified Molecular Resistance Mechanisms and Comparison with Phenotypic
Resistance Testing Table 1 summarizes the analysis of antimicrobial resistance determinants, ordered by strain, MLST type and international clonal lineage. Identified molecular resistance determinants of the carbapenem-susceptible Iraqi A. dijkshooriniae strain and the Syrian A. radioresistens strain are shown as well.
In most cases, isolates that were related carried the same resistomes. The most frequently detected resistance mechanisms among the assessed isolates comprised betalactamases as well as sulfonamide and aminoglycoside resistance genes. This is reflected by observed phenotypic resistance as indicated in the Appendix A Table A2, indicating high resistance rates against these substance groups, in particular for the isolates WEST-S6-47 and HBG-S2-62 which were resistant to amikacin, gentamicin, and tobramycin. These two isolates also possessed the same array of aminoglycoside modifying enzymes, as well as being the only isolates with dfrA7, aac(6 )Ib-cr-like, cmlA1-like, and bla GES-11 , suggesting a shared plasmid or resistance island ( Table 1). All A. baumannii were carbapenem resistant and this was mediated through possession of bla OXA-23 in the Afghan strains, by bla OXA-23 and/or bla GES-11 in the Syrian isolates, by bla OXA-58 in the Iraqi isolate, and bla NDM-1 in the Libyan isolate. Fluoroquinolone resistance was observed for all A. baumannii isolates (Appendix A Table A2). Analysis of the gyrA and parC genes revealed that the resistant isolates had a Ser83-Leu substitution in GyrA, while all but one (Iraq-2) also harbored a Ser80-Leu substitution in ParC. Resistance determinants to sulfonamides were present in all isolates resistant to co-trimoxazole despite a dfr gene being present in only two isolates, suggesting other mechanisms such as efflux might be playing a role, while tetB-like genes were detected in four isolates. All isolates remained susceptible to colistin.
All A. baumannii were in possession of the intrinsic bla OXA-51-like, the variants of which correlated with their Pasteur STs and IC clonal designation (Table 1). ISAba1 was not associated with this gene. The A. dijkshoorniae possessed bla OXA-819 , which is similar to the instrinic bla OXA-213-like from the closely related A. pittii, while the A. radioresistens had bla OXA-815 which is a bla OXA-23-like that is intrinsic to this species. Furthermore, all A. baumannii isolates carried their intrinsic bla ADC .

Discussion
The assessment was conducted to expand the knowledge about the micro-epidemiology of A. baumannii isolates with resistance against carbapenems from patients from Northern Africa and the Middle East, thus extending the results of a previous study published in 2018 [33]. A total of 16 Acinetobacter spp. strains including 14 A. baumannii from the strain collection of the Department of Microbiology and Hospital Hygiene, Bundeswehr Hospital Hamburg, were included in the study. Therefore, in addition to the previous study, clinical isolates from patients from Afghanistan and Iraq, which had been first introduced in a technical evaluation scheme on diagnostic identification of Acinetobacter spp. by fluorescence in situ hybridization (FISH) [39], could be included in the analysis. In comparison to the previous epidemiological assessment based on rep-PCR and superficial sequencing [33], cgMLST-and MLST-based typing applying the Pasteur and Oxford schemes [37,38] provided considerably more details, allowing origin-specific assignments to international clones and MLST-based sequence types. The comparably small size of the collection of those rare strains made a complete assessment of all available isolates, even of the non-A. baumannii strains, possible without any economic restrictions. Of note, all assessed A. baumannii isolates could be assigned to MLST sequence types of the Oxford and Pasteur schemes [37,38].
The cgMLST approach confirmed two nosocomial transmission clusters. Cluster 1 comprised the four Afghan isolates, which were isolated during a small local outbreak, while five out of seven Syrian A. baumannii strains which were isolated in 2013 at the Bundeswehr Hospitals Westerstede and Hamburg formed cluster 2. The remaining five isolates were singletons. The latter clustering as observed with the cgMLST-typing confirmed likely transmission events between Syrian patients prior to admission at the Bundeswehr Hospitals Westerstede and Hamburg, which had already been suspected based on the results of rep-PCR and superficial sequencing [33]. Those transmission events might either have occurred in medical facilities in the patients home country but also, as previously discussed for Syrian and Ukrainian war-injured patients treated at German military hospitals [33,40,41], during the evacuation flights out of the patients' crisis-struck home countries. Of note, two strains isolated from Syrian patients at the military hospital in Westerstede ("WEST"-strains) showed completely identical sequences and other isolates of this cluster, which were isolated at the Bundeswehr Hospital Hamburg ("HBG"-strains) showed only one, two, and nine allele differences, respectively (Figure 1). Such very close genetic relationship makes recent transmission events highly likely.
When comparing the cluster 2 from Figure 1 of the results chapter with the epidemiology at the Bundeswehr hospitals as recently described [33], the resulting likely nosocomial transmission events were as follows. All patients had been admitted to the German Bundeswehr Hospitals at the same day in 2013. Focusing on the Bundeswehr Hospital Westerstede, the strain WEST-S1-31, which had been isolated at day 21 after admission from the anus of a patient, was most likely nosocomially acquired from another patient, from which strain WEST-S6-50 had been isolated at day 1 after admission from a wound. Focusing on the cluster-related strains HBG-S1-56, HBG-S4-64 and HBG-S7-73 isolated at the Bundeswehr Hospital Hamburg, the first two isolates had already been isolated at day 2 after admission from a deep wound and the perineal skin of the patients, respectively, so nosocomial transmission had most likely occurred prior to hospital admission. The strain HBG-S7-73, however, had been isolated at day 9 from the pharynx of a patient. The sequence homology as indicated in Figure 1 makes nosocomial transmission from the patient carrying the strain HBG-S4-64 at the perineal skin highly likely.
As an interesting side-effect, cgMLST confirmed the clonal relationships as suggested by rep-PCR in the recent assessment [33], thus additionally confirming the basic reliability of this comparably cheap and easy-to-perform typing approach for local outbreak scenarios, in spite of the undeniably higher resolution of genome-based typing [34,40,42]. A comparison with other A. baumannii isolates from Ukrainian conflict zones that we have investigated did not show any clonality to those in the current study (data not shown) [40].
When looking at the detected beta-lactam resistance genes, the quantitatively most relevant one was bla OXA-23 , which could be identified in strains from all assessed geographic regions, and which is also frequent in resistant A. baumannii in Europe and beyond [43]. Other beta-lactamase genes with relevance for the observed carbapenem resistance comprised bla OXA-58 , the class A beta-lactamase gene bla GES-11 , and bla NDM-1 , respectively. It is interesting to note that isolate WEST-S6-47 had the lowest carbapenem-MICs of the A. baumannii in this study and it only had the acquired bla GES-11. In addition, and especially when looking for the resistome beyond just beta-lactamases, considerable diversity could be seen even in spite of an otherwise low number of allelic differences between some isolates within the observed clonal clusters, which is suggestive for the abundance of plasmids and mobile genetic elements. Despite the high-level of resistance seen in these isolates, colistin still retained activity.
Any comparisons with previous reports from the geographic regions of interest are limited by a paucity of data from respective war and crisis zones. Further, the relatively low number of available isolates is, of course, insufficient to provide a comprehensive overview on locally abundant resistant A. baumannii clones as well as their molecular resistance determinants, an admitted limitation of the study. Nevertheless, the bla OXA-58 gene as detected in the Iraqi A. baumannii strain is one of the genes mediating carbapenem resistance which had been reported already in 2006 and thus early after the beginning of the local conflict [4]. In a similar way, the occurrence of bla NDM-1 and bla OXA-23 in A. baumannii isolated in Libya had been described before [27,29,30], making those detections not further surprising. In so far, several of the reported findings are not completely new but just complement previously available epidemiological information.

Patient Isolates
All Acinetobacter strains with reduced susceptibility towards carbapenems and origin from North African and Middle Eastern war and crisis zones available at the laboratory of the Bundeswehr Hospital Hamburg were included in the assessment. There were no exclusion criteria. In detail, the strains Iraq-1 and Iraq-2 were isolated from wounds of injured US-American soldiers in the Iraqi conflict and kindly provided by the US Landstuhl Regional Medical Center (LRMC) as described elsewhere [39]. The strains AFG-1, AFG-2, AFG-3, and AFG-4 isolated from casualties of the Afghanistan conflict were shipped via the German field hospital in Mazar-e Sharif as previously detailed as well [39]. No patient specific data were provided for those strains for security reasons. The origin of the strains from patients injured in the conflicts in Libya and Syria and treated at the Bundeswehr Hospitals Hamburg and Westerstede (HBG-L1 (1), HBG-L2 (9) from Libya; WEST-S1 (31), WEST-S5 (44), WEST-S6 (47), WEST-S6 (50), HBG-S1 (56), HBG-S2 (62), HBG-S4 (64), HBG-S7 (73) from Syria) was described in detail elsewhere [33] next to epidemiological links and a preliminary molecular resistance analysis based on superficial NGS and rep-PCR typing. No environmental isolates were available. A. baumannii were identified using MALDI-TOF-MS and gyrB multiplex PCR [44,45].

DNA Extraction and Whole Genome Sequencing
DNA extraction and whole genome sequencing was performed exactly as described recently [40]. In detail, MagAttract HMW DNA Kits (Qiagen, Hilden, Germany) were used to extract the genomic DNA of the isolates according to the manufacturer's instructions for whole genome sequencing (WGS). Nextera XT library prep kits (Illumina GmbH, Munich, Germany) were applied for the preparation of sequencing libraries for 250 bp paired-end sequencing runs on an Illumina MiSeq sequencer. The Velvet assembler integrated in the Ridom SeqSphere+ v.7.0.4. software allowed the de novo assembling of the obtained reads. Any raw sequencing reads from the project were submitted to the European Nucleotide Archive (https://www.ebi.ac.uk/ena/, last accessed on 10 March 2021) with the Accession numbers PRJEB42650.

Molecular Epidemiology and Determination of Antibiotic Resistance Genes
All isolates were initially screened for the presence of common carbapenemases by PCR [46,47]. Using the pubMLST website (https://pubmlst.org/abaumannii/, last accessed on 10 March 2021)), the genome assemblies of all isolates were used to identify sequence types (STs) according to the Oxford and the Pasteur 7-loci MLST schemes [17,37,38].
Based on a validated core genome multi-locus sequence typing (cgMLST) scheme [34] using the Ridom SeqSphere + v. 7.0.4 software (Ridom GmbH, Münster, Germany), the isolates were further analyzed. The clonal relatedness was visualized using the same software by generating a minimum spanning tree including 2390 target alleles, also including an in-house library with the established international clones.

Phenotypic Resistance Testing and Spectrometry-Based Discrimination
A preliminary discrimination of the strains to the A baumannii complex (ABC) level had been performed applying MALDI TOF MS as described previously [33]. In detail, a Shimadzu/Kratos "AXIMA Assurance" MALDI TOF MS (Shimadzu Deutschland GmbH, Duisburg, Germany) and the "IVD-mode VitekMS-ID" database version 3.2.0.-6 (bioMérieux, Marcy-l Étoile, France) had been used. Regarding phenotypic resistance testing, colistin susceptibility was assessed applying the microbroth dilution assay MICRONAUT-S (MERLIN Diagnostika GmbH, Bornheim, Germany). All other tested substances were assessed applying AST-N248 cards in a VITEK-II automated device (bioMérieux).

Ethical Clearance
Ethical clearance for the molecular typing of the Acinetobacter strains was obtained from the ethics committee of the medical association of Hamburg (WF-042/15) in line with National German laws, without the need for informed consent.

Conclusions
In conclusion, this assessment provided epidemiological information on carbapenemresistant A. baumannii strains as isolated from patients from Northern African and Middle Eastern war and crisis zones and deposited at the Bundeswehr Hospital of Hamburg, Germany. Respective epidemiological assessments, further analyzing the molecular background of the distribution and spread of antimicrobial resistance in such difficult-to-access regions in the world, are neglected so far, as technological resources for molecular surveillance purposes are scarce in politically unstable areas. Certainly, our current small assessment of more or less randomly isolated strains collected over several years cannot replace continuous surveillance programs but may provide a glimpse on information which is otherwise difficult to obtain and usually relies on small cross-sectional assessments .
Considering the ongoing problem of infections due to carbapenem resistant A. baumannii isolates worldwide, the provided data may serve as a piece in the puzzle of global antimicrobial resistance without diminishing the need for larger future studies on the local epidemiology of carbapenem resistance in Northern African and Middle Eastern A. baumannii isolates in order to amend and broaden this preliminary information. Data Availability Statement: All relevant data are provided in the manuscript, its figure, its table, and its Appendix A. Raw sequence data are deposited and available as stated in the methods chapter. Further raw data can be made available on reasonable request.

Conflicts of Interest:
The sponsors did not have any role in the collection, analysis, or interpretation of data, in the writing of the report, or in the decision to submit the article for publication. Table A2. Phenotypic resistance testing of the Acinetobacter spp. isolates based on micro-broth dilution (applied for colistin) and automated resistance testing using a VITEK-II device (BioMéreux, Nürtingen, Germany; applied for all other assessed antimicrobial substances). MICs (minimum inhibitory concentrations) in µg/mL are shown in brackets. Interpretation was performed according to EUCAST NAK (European Committee on Antimicrobial Susceptibility Testing "Nationales Antibiotia-Sensitivitätstest-Kommittee"). S = susceptible. I = Susceptible at increased dosage. R = Resistant. N.d. = not defined.