Introduction

Acinetobacter baumannii (A. baumannii) is considered one of the most challenging pathogens for researchers and clinicians in medical settings all over the world. The threat posed by A. baumannii infections stems from the rapid and unchecked spread of this pathogen (Gerischer 2008) and its naturally low susceptibility to many antimicrobials (Lee et al. 2011). Moreover, A. baumannii can cause different types of infections, including ventilator associated pneumonia, skin and soft tissue infections, urinary tract infections, wound and bloodstream infections, and meningitis (Dijkshoorn et al. 2007; Fernández et al. 2012; McConnell et al. 2013; Rajamohan et al. 2009; Wisplinghoff et al. 2004). These infections are mainly hospital related, especially among intensive care unit patients (Eveillard et al. 2013), and in particular immunocompromised ones (Krahn et al. 2016). Moreover, the microorganism is also capable of causing community-acquired infections, albeit to a lesser extent (Wang et al. 2003). The propensity of A. baumannii to acquire resistance genes (Corbella et al. 2000), in addition to the excessive use of antibiotics in many health care settings caused the emergence of multidrug resistant (MDR) strains (Peleg et al. 2008) leading to the ineffectiveness of many antibiotics including the life-saving carbapenems (Gao et al. 2017).

Carbapenem-resistant A. baumannii (CR-AB) strains have been reported globally (Perez et al. 2007). The mechanisms involved in carbapenem resistance are diverse, including change in permeability of porins in the microorganism outer membrane, efflux pumps, and alteration in the affinity of penicillin binding proteins (Abbott et al. 2013). However, the most relevant mechanism is mediated by the acquisition of carbapenem hydrolyzing β-lactamases, mainly metallo-β-lactamases (MBL): VIM, IMP, SPM, and NDM, and the carbapenem-hydrolyzing class d β-lactamases (CHDLs): OXA-23, OXA-24/40, OXA-58, and OXA-143 and less importantly class A (Evans and Amyes 2014; Palzkill 2013). Most of these genes are carried on plasmids of A. baumannii (Naas et al. 2008). NDM is one of the most recently discovered β-lactamases (Nordmann et al. 2011), being first reported in 2008 in New Delhi, India from Klebsiella Pneumoniae (Yong et al. 2009). It was then detected among Escherichia coli isolates (Kumarasamy et al. 2010), and later in A. baumannii and Pseudomonas aeruginosa (Johnson and Woodford 2013). NDM dissemination was originally confined to the Indian subcontinent, then it spread worldwide in diverse Gram-negative isolates not necessarily epidemiologically linked to the Indian subcontinent (Johnson and Woodford 2013). Despite being first discovered among members of the Enterobacteriaceae, it is thought that blaNDM evolved in Acinetobacter from the fusion of another metallo-β-lactamase and aphA6, a gene encoding aminoglycoside resistance, then was transferred to other Gram-negative bacteria (Toleman et al. 2012).

In Acinetobacter spp., the blaNDM gene is mainly carried on plasmids belonging to the pNDM-BJ01-like family (Hu et al. 2012). These plasmids are usually conjugative which helps in the complex transmission of the gene between strains belonging to different genera (Espinal et al. 2011; Johnson and Woodford 2013). This makes A. baumannii harboring blaNDM a threatening and serious pathogen worldwide (Chen et al. 2011). The natural competency feature of A. baumannii further aggravates the issue (Traglia et al. 2014), rendering the study of plasmid transfer a focal issue to hinder the outbreaks caused by A. baumannii, especially in hospitals (Saranathan et al. 2014).

This study aimed to establish the role played by plasmid harbored blaNDM in mediating carbapenem resistance relative to resistance to other commonly used antibiotics among A. baumannii isolates obtained from patients presenting to Alexandria Main University Hospital (AMUH), the largest tertiary hospital in Alexandria, Egypt, in 2010 and 2015.

Materials and methods

Bacterial isolates

In the present study, 74 CR-AB clinical isolates were collected from Alexandria Main University Hospital (AMUH) from different clinical specimens in 2010 and 2015. The isolates were previously identified by conventional methods such as colony shape and aerobic growth at 44 °C on MacConkey’s agar as well as the Vitek system (Biomerieux, UK). The identity was further confirmed by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI–TOF MS) (Bruker Daltonik, USA) and PCR amplification of the chromosomally intrinsic blaOXA-51 gene. Nine of the A. baumannii isolates were shown by PCR and sequence analysis to carry blaNDM (GenBank accession numbers: MN395910, MN395911, MN395912, MN395913, MN395914, MN395915, MN395916, MN395917, and MN395918). K. pneumoniae ATCC 10031 was used as a reference susceptible strain (Abouelfetouh et al. 2019).

Antimicrobial susceptibility testing of the isolates

The susceptibility of all 74 isolates towards 17 different antibiotics was determined using the standard disc diffusion technique and the results were interpreted according to the guidelines of the Clinical and Laboratory Standards Institute 2018 (CLSI 2018). The antibiotics used were imipenem, meropenem, aztreonam, piperacillin, piperacillin/tazobactam, ampicillin/sulbactam, ceftazidime, cefepime, cefotaxime, ceftriaxone, tetracycline, doxycycline, amikacin, gentamicin, ciprofloxacin, levofloxacin, and sulphamethoxazole/trimethoprim (Oxoid Ltd, England).

Determination of the antibiotics minimum inhibitory concentration

The minimum inhibitory concentration (MIC) of imipenem, meropenem, ertapenem, amikacin, and levofloxacin against the tested isolates was determined using agar dilution technique, whereas MIC of colistin was determined using broth microdilution technique and the results were interpreted according to CLSI, 2018. The antibiotic powders/solutions of pharmaceutical grade were purchased from the Egyptian market as Tienam® (Merck Sharp & Dohme B.V.), Meronem® (Astrazenca, UK), Invanz® (Merck Sharp & Dohme Corp.), Amikacin® (Amoun Pharmaceutical Co.) and Tavanic® (Sanofi-Aventis Ireland Ltd. T/A Sanofi), respectively. Colistin was obtained as colistin sulphate (Sigma Aldrich). Twofold serial dilutions (1–512 µg/mL for all antibiotics, except colistin: 0.25–512 µg/mL) were freshly prepared on the day the experiment was done.

Plasmid extraction and characterization

Plasmids were isolated from the nine isolates harboring blaNDM using the plasmid isolation kit “GeneJET™ Plasmid Miniprep Kit #K0502” (Thermo Scientific, USA) according to the manufacturer’s instructions. The plasmid profiles were analyzed by 1% agarose gel electrophoresis in presence of 1 Kbp DNA ladder (Thermo Scientific, USA). All nine plasmid preps were used as DNA template for PCR amplification of blaNDM using primers NDM-F (5′-CACCTCATGTTTGAATTCGCC-3′) and NDM-R (5′-CTCTGTCACATCGAAATCGC-3′) and amplification conditions of: initial denaturation at 94 °C for 5 min, followed by 30 cycles of 94 °C/40 s, 54 °C/1 min and 72 °C/2 min, and final extension at 72 °C for 5 min (Poirel et al. 2010). The PCR products were resolved on 1% agarose gel in TAE buffer (40 mM Tris, 20 mM acetic acid and 1 mM EDTA, pH 8.3) at 100–120 V, in the presence of 100 bp DNA ladder (New England Biolabs, UK).

Transformation of carbapenem-susceptible A. baumannii cells with plasmids harboring blaNDM

A. baumannii cells are naturally competent (Biswas 2015), which eliminated the need to prepare electrocompetent cells. The cells of the carbapenem-susceptible A. baumannii (CS-AB) isolate A20, selected as the recipient, were transformed with the nine plasmids harboring blaNDM. Briefly, 10 mL of LB broth were inoculated with a single colony of the isolate A20 (MIC = 0.25 μg/mL and blaNDM-negative). After overnight incubation at 37 °C, the culture was used to aseptically inoculate 100 mL LB broth and was allowed to grow with vigorous shaking until optical density at 600 nm (OD600) reached 2.7. Then, 50 μL aliquots of the culture at stationary phase were diluted with 50 μL fresh LB broth and 5 μL of each plasmid preparation was, in turn, electroporated at 1850 v (BTX Harvard Apparatus, USA), followed by incubation at 37 °C for 1 h. Eighty microliters were then plated onto LB agar plates containing 2 μg/mL of imipenem (Huang et al. 2015). The plates were incubated at 37 °C overnight then checked for transformants. Transformation efficiency was calculated by dividing the number of transformants by the initial count of recipient cells (Harding et al. 2013). Carbapenem resistance was confirmed in the transformants by determination of imipenem MIC by agar dilution technique. In addition, PCR amplification of blaNDM in the transformants was performed as previously explained.

Results

Antibiotic susceptibility testing

All blaNDM harboring isolates were shown by susceptibility testing and MIC values to be resistant to imipenem, meropenem and ertapenem, while 100% of blaNDM-negative isolates were found to be resistant to meropenem and ertapenem only. Fifty nine (92.2%) blaNDM-negative isolates were imipenem resistant as evidenced by either susceptibility testing and/or MIC values (Tables 1, 2). Moreover, all blaNDM-positive isolates were also resistant to aztreonam, piperacillin, all tested cephalosporins, tetracycline, amikacin, and ciprofloxacin. On the other hand, only aztreonam was totally ineffective against the blaNDM-negative isolates (Table 1). These findings were mostly confirmed by MIC results (Table 2). MIC50 of imipenem, ertapenem, and levofloxacin was fourfold higher against the blaNDM-positive isolates, relative to the blaNDM-negative ones. In addition, the majority of the blaNDM-positive isolates were inhibited by higher concentrations of imipenem, as evidenced in 77.7% of the isolates being inhibited by concentrations ≥ 32 µg/mL versus 20% of the blaNDM-negative isolates. The same can be said about meropenem and ertapenem where 77.7% and 100%, respectively of the blaNDM-positive isolates were inhibited by ≥ 64 µg/mL of the antibiotics, relative to 56.9% and 90.8% of the blaNDM-negative isolates, respectively. Furthermore, 33.3% of blaNDM-positive isolates were only inhibited by 256 µg/mL of colistin, compared to 12.3% of the blaNDM-negative isolates. Likewise, 128 µg/mL of levofloxacin were needed to inhibit 66.7% of blaNDM-positive isolates, whereas only 26.2% of blaNDM-negative isolates needed as much levofloxacin. However, 16.9% of the blaNDM-negative isolates were only inhibited at 256 µg/mL. On the other hand, 77.8% of the blaNDM-positive group versus 64.6% of the blaNDM-negative group were inhibited by amikacin concentrations ≥ 512 µg/mL (Table 2).

Table 1 Antimicrobial susceptibility of the blaNDM-positive and blaNDM-negative isolates
Full size table
Table 2 Distribution and ranges of the minimum inhibitory concentrations of tested antibiotics among the blaNDM-positive and blaNDM-negative isolates
Full size table

Plasmid profiling/characterization

The extracted plasmids exhibited different profiles, ranging from 1.5 to about > 10 kbp (Fig. 1). PCR amplification of blaNDM and the subsequent resolution of the products on agarose gel showed bands at the expected size of 984 bps, confirming that all nine plasmids carried the gene.

Fig. 1
figure 1

Profiles of the nine plasmids extracted from the blaNDM-positive isolates. Lanes 1–9 represent the nine plasmid preparations and lane M the DNA ladder. Different profiles were obtained, with a size range between 1.5 and > 10 kbp

Full size image

Transformation of blaNDM-carrying plasmids into carbapenem-susceptible A. baumannii

Transformation of a CS-AB isolate, A20, harboring no blaNDM (recipient), with the nine plasmid preparations harboring blaNDM was performed by electroporation. Successful transformants were selected on imipenem plates, with transformation efficiencies that ranged from 1.3 × 10–8 to 2.6 × 10–7. MIC values of imipenem were determined against the transformants and were found to be > 64 µg/mL, which is 256-fold higher than the original MIC of the recipient A. baumannii (MIC 0.25 μg/mL). Moreover, plasmids were isolated from all nine transformants and used as templates to amplify blaNDM gene that was detected in all transformants except one, from plasmid preparation number 6, a representative is shown in Supplementary Fig. 1.

Discussion

Acinetobacter baumannii is a Gram-negative pathogen that is common in the hospital environment (Cerqueira and Peleg 2011). In addition, it has a broad diversity of resistance determinants and is capable of acquiring more resistance phenotypes via horizontal gene transfer (Fournier et al. 2006; Imperi et al. 2011; Perez et al. 2007). These factors, together with the high survival rate of the microorganism on dry surfaces made A. baumannii infections a major healthcare concern, especially among intensive care and immunocompromised patients (Fournier et al. 2006; García-Garmendia et al. 2001; Pogue et al. 2013). As a result of the increasing antibiotic resistance among A. baumannii isolates in the last decades, carbapenems became last option drugs to treat such infections (Meletis 2016). However, CR-AB isolates have emerged (Meletis 2016) and are extensive drug resistant (XDR) in most instances (Viehman et al. 2014), which leaves the second-line treatment options, that are usually more toxic or of controversial efficacy, e.g., colistin and tigecycline (Peleg et al. 2006; Viehman et al. 2014). Moreover, infection with carbapenem-resistant strains was associated with mortality in 16 to 76% of the cases relative to 5 to 53% for infections due to carbapenem-susceptible ones. This is largely attributed to the more severe nature of infection with resistant strains and the initial delay in proper antimicrobial therapy administration (Lemos et al. 2014).

Carbapenem-resistant Acinetobacter isolates have been reported at different rates from around the world. The rates ranged from 84% in a national surveillance study in Switzerland between 2005 and 2016 (Ramette and Kronenberg 2018) and 95% in Turkey between 2011 and 2012 (Cicek et al. 2014). In the Middle East, the rate in the last two decades was 45% in Tunisia (Ben Othman et al. 2007), 65% in Saudi Arabia (Al-Agamy et al. 2014), 19.14% in Kuwait (Al-Sweih et al. 2012), and 47.9% in Algeria (Bakour et al. 2013). In Cairo, Egypt, one study conducted between 2011 and 2012 (Fouad et al. 2013) showed imipenem and meropenem resistance rates of 74% and 100%, respectively among A. baumannii clinical isolates, while a second study between 2012 and 2013 (Abdel Hamid et al. 2016) found that 95.1% of the tested isolates were resistant to carbapenems. Moreover, a more recent study carried out in Mansoura, Egypt reported extensive drug resistance among 100% of the A. baumannii isolates obtained from patients suffering from nosocomial infections. These isolates were simultaneously resistant to penicillins, cephalosporins, fluoroquinolones, aminoglycosides, carbapenems and tigecycline (Elsayed et al. 2019).

One of the main mechanisms driving carbapenem resistance among A. baumannii is the production of carbapenemases that could be acquired or intrinsic (Viehman et al. 2014). NDM is one of the most important acquired metallo-β-lactamases because of its wide substrate specificity and its current dissemination in various regions of the world since its first discovery in India (Chen et al. 2011; Decousser et al. 2013; Palzkill 2013).

The current study included 74 CR-AB clinical isolates, including nine blaNDM-positive ones, collected from AMUH. Susceptibility testing showed that overall antibiotic resistance was higher among the blaNDM-positive isolates, than the blaNDM-negative ones. Moreover, MIC ranges for imipenem, meropenem, ertapenem, colistin, levofloxacin, and amikacin were generally at least twofold higher among the blaNDM-positive group. Among the tested antibiotics, colistin displayed highest activity, being active against 52.3% of blaNDM-negative isolates versus 55.6% of blaNDM-positive ones. These results corroborate the previously reported finding that blaNDM-positive strains are also resistant to all β-lactams, except for aztreonam (Yong et al. 2009). Nevertheless, both blaNDM-positive and -negative isolates reported here were also aztreonam resistant which could be attributed to other resistance determinants as previously reported (Rodríguez-Martínez et al. 2010). In addition, a blaNDM-positive A. baumannii recovered in Brazil in 2013 was also resistant to meropenem, imipenem, all cephalosporins, aztreonam, aminoglycosides, tetracyclines, and sulphamethoxazole/trimethoprim (Pillonetto et al. 2014). A study from Egypt commented on the concomitant resistance to the carbapenems and quinolones, trimethoprim/sulfamethoxazole, and aminoglycosides in a collection of A. baumannii isolates that were 30% blaNDM positive (Benmahmod et al. 2019).

In Acinetobacter, blaNDM is usually carried in a Tn125 composite transposon on pNDM-BJ01-like plasmids which are highly conserved (Chen et al. 2015; Hu et al. 2012). The genetic environment of blaNDM is also conserved in other plasmid families found among non-Acinetobacter (Partridge and Iredell 2012). Besides blaNDM, the plasmid also harbors aphA6 upstream of Tn125, a gene that encodes aminoglycoside resistance (Jones et al. 2015), which could explain the higher range of amikacin MIC against blaNDM-positive isolates in the current study. Moreover, it is believed that blaNDM is a chimeric gene that originated from a recent fusion event between aphA6 and an older metallo-β-lactamase (Toleman et al. 2012). Previous studies have described the presence of aminoglycoside-modifying enzymes as a “main” reason for aminoglycoside resistance (Peleg et al. 2008). A study investigating the transfer of blaNDM-carrying plasmids from Acinetobacter isolates revealed the acquisition of both carbapenem and aminoglycoside resistance in the resultant E. coli transconjugants (Huang et al. 2015), which highlights the link between carbapenem and aminoglycoside resistance determinants on blaNDM-carrying plasmids from Acinetobacter.

In A. baumannii, blaNDM is mostly plasmid mediated, an exception lies in the European isolates where the gene is chromosomal (Hu et al. 2012; Pfeifer et al. 2011). Since these plasmids also carry other resistance determinants (Kumarasamy et al. 2010), it was important to study plasmid transfer among our cohort of A. baumannii isolates. Plasmids were isolated from the nine clinical isolates harboring blaNDM. Profiles of the isolated plasmids were analyzed after agarose gel electrophoresis and ranged from 1.5 to > 10 kbp which agrees with the results reported by an earlier study (Saranathan et al. 2014) in which 2 to > 25 kb plasmids were isolated from CR-AB. blaNDM presence on the isolated plasmids was confimed by PCR using the different plasmid preparations as templates. These findings were in accordance with a previous study which reported carriage of blaNDM on different plasmids (Chen et al. 2011). The nine plasmid preparations were electroporated into CS-AB cells. Transformation efficiency ranged from 1.3 × 10–8 to 2.6 × 10–7. A previous study (Huang et al. 2015) reported an average conjugation frequency in A. baumannii of 7.69 × 10–6 and 7.09 × 10–7 and an even higher frequency among non-pathogenic Acinetobacter spp. which points these strains as potential reservoirs for the transfer of resistance determinants. An important difference between the two studies is the use of conjugation in the previous study (Huang et al. 2015), whereas the current work relied on electroporation, a method recommended for transfer of foreign DNA into A. baumannii (Thompson and Yildirim 2019). Imipenem MIC values against the obtained transformants in the current study were 256-fold higher than the recipient A. baumannii strain. All the obtained transformants, except one (number 6), were shown to carry blaNDM by PCR. This indicated the successful transfer of the plasmids harboring blaNDM. Transfer of other plasmids conferring carbapenem resistance other than the one carrying blaNDM may have contributed to carbapenem resistance in the absence of blaNDM in the odd transformant.

Conclusions

blaNDM was plasmid mediated in the tested CR-AB isolates from Alexandria, Egypt. The increased resistance of these isolates to other antibiotic classes coupled with the natural competence of A. baumannii which facilitates plasmid transfer to CS-AB isolates point to potential loss of the effectiveness of invaluable antimicrobial agents. This warrants further investigation of the genetic context of blaNDM on the plasmids using genomic techniques, towards the design of effective antibiotic stewardship and infection control policies in Egyptian hospitals and the community.