FormalPara Key Summary Points
The majority (94.1%) of carbapenem-resistant A. baumannii (CRAB) isolates from cancer patients harbored more than one carbapenemase-encoding genes.
We observed a high prevalence of metallo-β-lactamase-encoding genes including blaNDM, blaOXA-23-like, and blaKPC.
MLST analysis identified different STs that belong to the highly resistant easily spreadable International clone II.
Measures should be implemented to control the spread of this clone.

Digital Features

This article is published with digital features, including a summary slide, to facilitate understanding of the article. To view digital features for this article go to https://doi.org/10.6084/m9.figshare.13134965.

Introduction

Cancer patients are high-risk, immunocompromised, and may experience long hospital stays. Thus, they are more prone to infections with opportunistic bacteria such as Acinetobacter baumannii [1]. A. baumannii is a Gram-negative, aerobic, non-motile coccobacillus. Moreover, it is an opportunistic pathogen that can cause severe, life-threatening healthcare-associated infections such as bacteremia, pneumonia, meningitis, and endocarditis [2]. The fairly recent emergence and increased prevalence of multidrug-resistant (MDR) A. baumannii is worrisome. A. baumannii is now listed by the World Health Organization as one of the critical pathogens, which highlights the need for the development of new antimicrobials [5]. This can be attributed to the increased resistance to multiple antibiotics, including last resort antibiotics, such as carbapenems, which are reserved for cases when all alternatives have been exhausted (typically used with MDR bacteria in hospitalized patients) [6]. The high resistance patterns of A. baumannii are due to the upregulation of intrinsic antimicrobial resistance genes in addition to their genomic plasticity, allowing the acquisition of new resistance genes through mobile genetic elements such as plasmids and transposons [7]. Many mechanisms can decrease susceptibility of A. baumannii to carbapenems, including the production of carbapenemase enzyme [10], loss of outer membrane proteins [11], overexpression of multidrug efflux pumps [12], and alterations in penicillin-binding proteins [2].

Among the previous mechanisms, the production of carbapenemase enzyme is considered to be the main mechanism of resistance to carbapenems [10]. Serine carbapenemases and metallo-β-lactamases are two carbapenemase groups that have been defined according to their active sites. Serine carbapenemases include class (A) penicillinases and class (D) oxacillinases, whereas class (B) carbapenemases belong to metallo-β-lactamases (MBL) that are inhibited by EDTA. Class (A) carbapenemases include the IMI/NMC, SME, KPC, and GES enzymes, whereby KPC and GES enzymes are plasmid-encoded and thus highly spreadable [13]. Class (B) MBL include IMP, VIM, GIM, SIM, and NDM enzymes, whose genes are mainly found in transferable plasmids. The NDM-encoding gene was first detected in a Klebsiella pneumoniae isolate [14]. NDM-1 has spread worldwide and is one of the most common carbapenemases in Enterobacteriaceae and A. baumannii [15]. Class (D) β-lactamases are also known as oxacillinases, for ‘oxacillin-hydrolyzing’, or OXA β-lactamases. Genes encoding OXA β-lactamases are present in plasmids and chromosomes. Worldwide, the most common OXA-encoding gene groups in A. baumannii are the OXA-23, OXA-24/40, and OXA-58 groups, whereas OXA-143 has only been detected from A. baumanii isolates in Brazil [16]. OXA-48 can typically be detected in K. pneumoniae [13]. The oxacillinase are relatively lower in activity than other types of carbapenemases, but overexpression of these genes has been observed in the presence of insertion sequences (IS) upstream of these genes which can provide additional promoters [17]. There are three “worldwide” clonal lineages (International clones: ICs I, II, and III) for A. baumannii [18]. The international clone II shows worldwide spread in many hospitals which can be attributed to the ability of this lineage to incorporate new genes and their adaptation to hospital environment [2]. The rapid spread of multidrug-resistant A. baumannii clinical isolates among cancer patients in the last two decades is worrisome because infection by this bacteria is associated with a high rate of mortality among this vulnerable group [8].

The aim of the current study was to investigate the dissemination of carbapenemase-encoding genes among carbapenem-resistant A. baumannii (CRAB) isolates from cancer patients followed by the genotypic analysis of these isolates. This will help to tailor the antimicrobial protocols in healthcare settings and to improve infection control policies.

Methods

Bacterial Isolates

A total of 520 isolates were recovered from blood samples in cases of blood infection of cancer patients at the National Cancer Institute (NCI), Giza, Egypt, from July 2017 to January 2018. Ethical approvals were obtained from the Ethics committees of the NCI and the Faculty of Pharmacy, October University for Modern Sciences and Arts. Consents from patients were obtained before the inception of the study. Samples were streaked on CHROMagar Acinetobacter supplemented with CR102 (CHROMagar, France) for isolation of multidrug-resistant Acinetobacter sp., then isolated colonies were identified using the VITEK2 automated system (BioMerieux, Marcy-l’Étoile, France). Identification was confirmed by polymerase chain reaction (PCR) amplification of the intrinsic blaOXA−51−like [19], in addition to using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF/MS) (Microflex LT; Bruker Daltonics) [20]. All isolates were preserved in glycerol broth at – 20 °C.

Antimicrobial Susceptibility Testing

Antimicrobial susceptibility patterns were determined by the VITEK2 system and the results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines [21]. The classes of antibiotics used in this test included β-lactams, aminoglycosides, quinolones, folate pathway inhibitors, glycylcycline, and polymyxin. Minimum inhibitory concentrations (MICs) of meropenem and colistin, for all CRAB isolates, were determined by the agar dilution, and micro-broth dilution methods, respectively, according to the CLSI [21] guidelines.

Detection of the Carbapenemase-Encoding Genes

Pure colonies of carbapenem-resistant isolates were used for DNA extraction using the Thermo Scientific™ GeneJet™ genomic DNA purification kit (Thermo Scientific, MA, USA), according to the manufacturer’s recommendations and kept at − 20 °C. PCR was performed to screen for the presence of carbapenemase-encoding genes belonging to Ambler classes (A), (B), and (D), using the primers and annealing temperatures described in Table 1. Genes were amplified by initial denaturation at 94 °C for 5 min, followed by 35 cycles consisting of 3 phases: DNA denaturation at 94 °C for 0.5 min, annealing according to Table 1 for 0.5 min, and elongation at 72 °C for time (1 Kb/1 min); finally, elongation at 72 °C for 10 min. A combination of forward ISAba1 primer and reverse primers of class (D) encoding genes were used to detect the presence of the ISAba1 insertion element upstream carbapenemase-encoding genes [17]. Negative control was included in all PCR assays.

Table 1 Primers used for amplification, expected product lengths, and annealing conditions

Molecular Typing of CRAB Isolates by Enterobacterial Repetitive Intergenic Consensus–PCR (ERIC–PCR) and Multilocus Sequence Typing (MLST)

Clonal relatedness of collected isolates was examined via enterobacterial repetitive intergenic consensus (ERIC)–PCR, which was carried out as described previously by Versalovic et al. [22] using ERIC2 primer (Table 1). Electrophoretic patterns were analyzed by using the Bionumerics software v.7.6 (Applied Maths, Sint-Martems-Latem, Belgium). The BioNumerics analysis was performed using the Dice coefficient and the unweighted pair group method of averages (UPGMA) with a 1% tolerance limit and 1% optimization. Isolates that clustered with ≥ 80% similarity were grouped into one ERIC type. Representative isolates from ERIC clusters of 100% similarity were subjected to typing by multilocus sequence typing (MLST), which was performed according to the Oxford scheme protocol using primers listed in the A. baumannii MLST database website (https://pubmlst.org/abaumannii/). Amplification of the seven conserved housekeeping genes (gltA, gryB, gdhB, recA, cpn60, rpoD, and gpi) was performed according to the protocol proposed by Bartual et al. [23]. Sanger sequencing was carried out using the ABI 3730xl DNA Analyzer at the Macrogen sequencing facility (Macrogen®, South Korea). The allelic numbers and sequence types (STs) were defined by means of the A. baumannii MLST database. Clusters of related STs (defined as clonal complexes; CCs) were analyzed using the Global Optimal eBURST (goeBURST) by the Phyloviz 2.0 software (https://www.phyloviz.net/goeburst). Analysis of clonal complex was carried out at the level of single locus variant (SLV) and double loci variant (DVL).

Results

Antibiotic Susceptibility of Collected Isolates

Forty-eight A.baumannii non-duplicate isolates were recovered from 520 blood samples representing 15.1% of total isolates. The A.baumannii isolates were recovered from 43 hospitalized patients and 5 outpatients. The identification of the A.baumannii isolates up to the species level was confirmed by the detection of the intrinsic blaOXA-51-like gene and the MALDI-TOF/MS. The antibiotic resistance of the isolates was determined by VITEK2, and six resistance patterns were detected. A total of 34 out of the 48 isolates (70.8%) were found to be resistant to meropenem and ertapenem and were, hence, designated as CRAB. The 34 CRAB isolates were recovered from cancer patients (5 outpatients and 29 inpatients). Resistance to tigecycline and colistin (last-resort antibiotics) was detected in 7/34 (20.5%) and 1/34 (2.9%) of CRAB isolates, respectively (Table 2). The MICs of meropenem against tested isolates ranged from 8 to ≥ 128 µg/ml (Table  3).

Table 2 Antimicrobial sensitivity patterns of carbapenem resistant A. baumannii (CRAB) isolates
Table 3 Minimum inhibitory concentration (MIC), carbapenemase-encoding genes, ERIC type of 34 carbapenem-resistant A. baumannii isolates

Detection of Carbapenemase-Encoding Genes

Co-existence of more than one carbapenemase-encoding gene was detected in 28/34 (82.3%) of CRAB isolates. The carbapenemase-encoding resistance genes blaNDM, blaOXA-23- like, and blaKPC showed the highest prevalence of 67.7%, 55.9%, and 50% of CRAB isolates, respectively. The Ambler class (B) MBL genes were detected in 31/34 (91%) of isolates. Among the MBL genes detected, in addition to blaNDM, were blaGIM, blaSPM, blaSIM, and blaIMP in 38.2%, 29.4%, 8.8%, and 5.8% of CRAB isolates, respectively, while blaVIM was not detected in any of the collected isolates. The class (D) oxacillinase coding genes blaOXA24/40 and blaOXA58 were detected in 26.4% and 2.9% of CRAB isolates, receptively. Class (A) GES gene was detected in 9/34 (26.4%) of isolates. The insertion element was detected upstream of blaOXA 23-like in two isolates (ID.27 and 41), blaOXA24/40-like in two isolates (ID 38 and 41), while it was upstream of blaOXA 51-like in one isolate (ID.7) (Table 3).

Molecular Typing of CRAB Isolates

ERIC-PCR typing revealed that the 34 CRAB isolates were grouped into 9 clusters and classified into 19 ERIC types according to 80% cut off. Eric type “3” was the most prevailing and represented by 7 isolates (Fig. 1). The predominant cluster was G which contains 7 isolates, followed by D(5), H (5), C (4), B (3), E (3), F (3), A (2), and I (2). Isolates with 100% ERIC typing similarities were found to be collected within 0–3 months. MLST was carried out for six isolates as representative for each of the groups in Table 3, and belonging to ERIC clones showing 100% similarity. The most prevalent MLST type was ST 286 (3 isolates), followed by ST 195, ST 1114, and ST 1632, represented by one isolate each (Table 4). Isolates typed as MLST ST 286 were found to be isolated from the same hospital floor but from different wards. ST relatedness to CCs was analyzed by the eBURST algorithm. eBURST analysis showed that all the identified STs belong to a founder ST208. ST 286 and ST 195 were SLV of gpi locus from founder, while ST 1114 and ST 1632 were SLV from ST437 and DLV from ST 208 founder in gyrp and gpi loci (Fig. 2). All identified STs were found to belong to CC92 (International clone II).

Fig. 1
figure 1

Dendrogram generated by Dice coefficient and the UPGMA clustering method with 1% tolerance limit and 1% optimization, showing the genetic similarity among A. baumannii isolates by enterobacterial repetitive intergenic consensus (ERIC) genotyping

Table 4 Multilocus sequence types (MLSTs), allele profiles, and clonal complex of six carbapenem-resistant A. baumannii isolates
Fig. 2
figure 2

eBURST analysis of A. baumannii strains using the Phyloviz 2.0 software. a  Related and unrelated ST groups to the identified isolates were compared with 2087 sequences (in the MLST Oxford database). b STs 286 (dark green), 1114 (red), 1632 (blue), and 195 (light green) are related to the founder ST 208 and belong to CC92. A circle represents an ST, and its size corresponds to the number of isolates

Discussion

Cancer patients are at higher risk of acquiring A. baumannii infections due to several factors, including their immunocompromised state and lengthy hospital stays [10]. Infections by CRAB isolates pose a great threat for cancer patients because they are associated with a high mortality rate. The ability of A. baumanni to resist the reserved antibacterial agents including carbapenems is alarming, hence raising the importance of studying its prevalence and mechanism of resistance to control its spread.

Carbapenem insensitivity was detected in 70.8% (34/48) of collected isolates, which is similar to other studies carried in Egypt by Sultan and Seliem [24]. Others have shown a higher prevalence for CRAB reaching 90% [10, 25, 26], which suggests that Egypt is the topmost country in the region in CRAB prevalence [27]. CRAB treatment typically relies on other last-resort antibiotics such as colistin and tigecycline and sometimes antibiotic combinations [28]. This becomes more challenging when isolates also exhibit resistance to last-resort antibiotics. About 21% of CRAB isolates were resistant to tigecycline; a similar prevalence of tigecycline resistance was detected by Kamel et al. [29].

Co-occurrence of a variety of intrinsic and acquired carbapenemase-encoding genes has been detected with increased prevalence for acquired carbapenemase-encoding genes known to be carried on mobile elements. Isolates co-harboring more than one acquired carbapenemase-encoding genes account for 32/34 (94.1%) of CRAB isolates. Clones carrying multiple carbapenemase-encoding genes have been detected in many studies carried out in the Middle East region [30,31,32], and in China [33].

Ambler class (A) blaKPC and blaGES genes were detected in 50% and 26.5% of CRAB isolates, respectively. The increased spread of the blaKPC and blaGES in A. baumannii clinical isolates in Egypt reached a prevalence of 56% and 48%, respectively, in Benmahmod et al.’s [26] study. The high spreading capacity of the blaKPC and blaGES genes could be attributed to their linkage to mobile elements such as Tn4401 located on conjugative plasmids [34] and integrons [35], respectively, facilitating their horizontal transfer. The blaGES-encoding gene is usually associated with a low level of carbapenem resistance (MIC 4–16 µg/ml) [9], while in the current study carbapenem MIC in isolates harboring blaGES-encoding gene ranged from 16 to ≥ 128 µg/ml due to the co-existence of other carbapenemase-encoding genes. Infection with A. baumannii carrying blaKPC is usually associated with a high level of morbidity and mortality [36].

MBL-encoding genes were detected in 91% of CRAB isolates which is worrisome, because these genes are usually correlated with high MIC [37], and they are characterized by rapid spread and high transferability between bacteria [38]. MBL-producing bacteria are a potentially great threat to modern intensive care treatment protocols [39]; hence, rapid detection and good infection control are required to reduce their impact. MICs in MBL-carrying isolates ranged from 16 to ≥ 128 µg/ml except for one isolate (ID 9) which might indicate mutation in the MBL gene of this isolate. Class (B) carbapenemase-encoding genes including blaNDM, blaGIM, blaSPM, blaSIM, and blaIMP were detected in 67.7%, 38.2%, 29.4%, 8.8%, and 5.8% of CRAB isolates, respectively.

High prevalence of blaNDM gene (67.7% of CRAB isolates) was observed in our study compared to previous studies carried out in Egyptian hospitals which showed blaNDM prevalence of 8% [40] and 39.3% [25] among CRAB isolates. The blaNDM-1 and blaNDM-2-encoding genes were first described in A. baumannii isolated from Egyptian patients [41] and then a noteworthy spread of blaNDM-positive A. baumannii was detected in the Middle East [42]. There are several proved mechanisms for horizontal transfer of blaNDM-encoding gene. Jamal et al. [43] found that the transfer of MBL carbapenemase genes was associated with Tn125-type transposon and can be harbored by a plasmid or rarely integrated into the chromosome. The plasmid-harboring blaNDM-1 gene in A. baumannii was found to exhibit high transformation frequency via outer membrane vesicles [44]. Another mechanism for horizontal spread of blaNDM gene among A. baumannii is by phage transduction [45]. A noteworthy observation in this study was the co-existence of MBL-encoding genes in some isolates which has been detected in many studies carried out in countries which suffer from uncontrolled antibiotic misuse [9, 31, 32, 46]. Lee and his colleagues found that strains harboring multiple plasmids that encode different carbapenemases showed increased fitness and virulence even in the absence of antibiotics which could increase the spread of this strain and emphasize the need for a strategy to combat this strain [47].

The increased prevalence of co-existing MBL could be attributed to the association of these genes by class 1 (sometimes class 3) integrons, which, in turn, are embedded in transposons, resulting in a highly transmissible genetic apparatus [48]. Integrons facilitate movement of resistance genes between integrons in plasmids, and the plasmids allow transfer of genetic material to different bacteria.

Oxacillinase-encoding genes can be intrinsic (blaOXA-51-like) or acquired (blaOXA-23-like, blaOXA-24/40- like, blaOXA-58-like). Two isolates (ID 2, 4) were found to harbor solely the blaOXA-like genes and they were found to be inhibited by relatively lower concentrations of meropenem compared to isolates harboring other types of carbapenemase genes. The oxacillinase enzymes only weakly hydrolyze carbapenems, but it was found that the insertion of a sequence such as ISAba1 upstream of the blaOXA-like genes may enhance the gene expression by conferring strong promoter activity. This insertion sequence could also explain the high capacity of the blaOXA-like genes for horizontal transfer and increased clonal diversity [49]. Numerous studies in Egypt and the Mediterranean regions have classified blaOXA-23 like gene as the most common carbapenemase-encoding gene [9, 40, 50,51,52]. In our study, the blaOXA-23 like gene was detected in 64% of CRAB isolates, which is similar to the previously mentioned results. Our findings showed that clusters carrying ISAba1 are widely distributed in our hospital, reaching 21/34 (61%), which might explain the high spread of acquired resistance genes among isolates. The blaOXA-58 gene prevalence reported here (2.9%) is lower than previously published rates from Tunisia (4%), Egypt (9.1%) [53], and Algeria (14.7%) [54].

The diversity of ERIC patterns obtained in our study suggests dissemination of carbapenem-hydrolyzing β-lactamase-encoding genes among genetically unrelated isolates of A. baumannii. This may be attributed to horizontal gene transfer of plasmids carrying resistance determinants. Isolates with ERIC typing similarity of 100% showed similarity in antibiotic resistance pattern, except for isolates (ID.6, 24), (ID 2, 5), and (ID14,23).

MLST was carried out for six isolates as representative for ERIC clones showing 100% similarity, and which was isolated from the same hospital ward. The Oxford scheme of MLST typing was chosen due to its high discriminatory power, which is comparable to the resolution obtained with pulsed-field gel electrophoresis [55]. MLST analysis identified different sequence types: ST286 (3 isolates), ST195 (1 isolate), ST1114 (1 isolate), and ST1632 (1 isolate). ST195 was previously detected in Egyptian hospitals [10, 56], and it is also prevalent in the Gulf area [57] and many countries, including China [58]. The ST195 strain was previously isolated from environmental as well as clinical samples (https://pubmlst.org/abaumannii/). ST1114 was detected in an outbreak in a tertiary hospital in Egypt [59]. ST 1632 was not detected in previous Egyptian studies, but this type was SLV of ST437 which has been previously isolated from Mediterranean countries such as Italy and Greece (https://pubmlst.org/abaumannii/). All detected ST types were either SLV or DLV to ST 208 which is the common ST type detected in previous studies in Egypt [10], and also in Saudi Arabia [57]. All identified STs belong to international clone II (IC II), which is the most widely distributed highly resistant clone worldwide [40, 53, 56], and responsible for the majority of outbreaks reported around the world [55]. A correlation between the IC II and the prevalence of resistance genes and mobile elements has been detected which necessitates the need for adequate control for the spread of this clone.[60]. Some research have explained the spreadibility of IC II due to prevalence of blaOXA-23-like gene in this clone conferring high resistance for this strain [10], but, in contrast to our result, no blaOXA-23-like gene was detected in the ST195 MLST typed isolate.

The main mechanism of carbapenem resistance is the production of carbapenemase enzyme. However, this does not eliminate the possibility of the presence of other mechanisms of carbapenem resistance. One limitation of the current study is that it does not cover these other possible mechanisms of resistance. Deciphering the role of efflux pumps in carbapenem resistance could be the scope of a future study.

Conclusion

The high prevalence of carbapenemase-encoding genes (including MBL-encoding genes such as blaNDM) and their co-existence in CRAB isolates is worrisome. This is due to the potential for high spreadability and the possibility of further dissemination of the highly antibiotic-resistant genes to other bacteria. This can make treatment of these cases very challenging, especially in the immunocompromised cancer patient population. Overall, the findings are alarming and call for strict control measures to prevent the spread of these genes among bacteria causing infections in immunocompromised patient populations.