Background
Acinetobacter baumannii is an emerging opportunistic pathogen that can become resistant to multiple antimicrobial agents [
1‐
3]. Over the past 20 years
A. baumannii’s clinical importance has been driven by its ability to obtain antimicrobial resistance factors [
4,
5]. Resistance is acquired through the transfer of integrons, plasmids or transposons that carry groups of genes encoding resistance to several antibiotic families [
6]. This bacterium can survive under a wide range of environmental conditions and on surfaces for extended periods of time and this can lead to both endemicity and outbreaks in healthcare facilities [
4,
7].
Major problems are caused by this pathogen, especially among critically ill hospitalised patients in intensive care units (ICU’s) [
5,
6,
8]. Clinical manifestations that can be associated with
A. baumannii include: bacteraemia, meningitis, pneumonia, urinary tract infection, ventilator-associated pneumonia (VAP) and wound infections [
8‐
10]. Patients with underlying chronic disease, decreased immunity, indwelling catheters and prolonged hospitalisation are more at risk to become colonised and develop subsequent infection with
A. baumannii [
7‐
9,
11].
This pathogen possesses different antimicrobial resistance mechanisms, which include the enzymatic inactivation of β-lactam antibiotics (cephalosporins, carbapenems and monobactams) through the production of extended spectrum β-lactamases, carbapenemases and AmpC-type (ampicillin class C) enzymes [
12‐
15]. The β-lactamases of Gram-negative bacteria belong to Ambler classes A to D [
12,
13]. The inherent class D oxacillinases (OXA) of
A. baumannii belong to the chromosomally encoded OXA-51-like enzyme group [
16‐
18]. However, the OXA-51 enzyme has been under considerable pressure due to antibiotic use and has evolved and now plays an important role in the carbapenem resistance of
A. baumannii isolates [
19]. If IS
AbaI is located upstream of the
bla
OXA genes the
A. baumannii isolate will have an increased OXA-51 expression which leads to carbapenem resistance (imipenem and meropenem) [
19‐
21].
Infections caused by
A. baumannii are difficult to treat because of the resistance to antimicrobial agents, especially β-lactams [
3,
22]. There have been reports of MDR
A. baumannii from hospitals all around the world (Argentina, Brazil, China, Europe and the United Kingdom) [
4,
23,
24]. Only a few antimicrobial agents namely colistin (in combination with carbapenems) and tigecycline are available for the treatment of MDR
A. baumannii [
1,
2,
15]. However, there are now reports of colistin resistance in
A. baumannii isolates, which is of great concern as colistin plays a pivotal role in MDR
A. baumannii treatment [
25‐
28]. Fortunately colistin resistance in
A. baumannii is still rare [
21,
25].
The aim of this study was to detect the prevalence of β-lactamase genes and to determine the genetic relationship among the A. baumannii isolates circulating in our clinical setting. The prevalence of β-lactamase genes is important for the surveillance of circulating A. baumannii strains as it provides more information for the infection control policies of the hospitals.
Methods
Study design and setting
The quantitative study was aimed to determine the prevalence of β-lactamase genes in consecutively collected A. baumannii isolates obtained from clinical specimens and to determine the genetic relatedness of circulating strains. The study was conducted at the Department of Medical Microbiology, Prinshof Campus, University of Pretoria/National Health Laboratory Service (NHLS) and received ethical approval from the Research Ethics Committee, Faculty of Health Sciences, University of Pretoria (121/2013); a waiver for informed consent was provided.
Collection, culturing and storage of A. baumannii clinical isolates
One hundred consecutive clinical A. baumannii isolates were collected without prior knowledge of patient information, hospital ward, specimen type and antibiotic susceptibility status. All 100 A. baumannii isolates were collected from the diagnostic laboratory during May to July 2013. The laboratory processes specimens from a tertiary academic hospital as well as district hospitals and various clinics as part of standard care. Routine analysis was performed as requested on the form and according to sample type. The isolates were identified and tested for susceptibility using the automated VITEK® 2 system (BioMérieux, France) with the VITEK® 2 GN card and the VITEK® 2 AST-N255 card. Gender, age, type of specimen, hospital, ward and the antibiotic susceptibility profiles were collected.
One colony of an isolate (obtained from a blood agar plate) was inoculated into 5 ml tryptone soya broth (TSB) (Oxoid, UK) and incubated on a shaking incubator (Vacutec, South Africa) at 35°C for a period of 18 to 24 h. Growth was indicated by turbidity in the broth. Isolates were stored by adding 900 μl of 50% glycerol (Merck, Germany) to 900 μl of the overnight TSB culture (Oxoid, UK) in 2 ml cryotubes (Greiner Bio-One, Germany) in a one to one ratio. This was done in triplicate and the isolates were stored in a -80°C freezer (New Brunswick, USA). The quality control strains used were the multidrug resistant A. baumannii ATCC® BAA-1605 strain; Klebsiella pneumoniae ATCC 8303 and K. pneumoniae ATCC BAA-2146.
The genomic DNA of each of the 100
A. baumannii isolates was extracted from a single colony grown in TSB (Oxoid, UK) culture media by using the ZR Fungal/Bacterial kit [
29]. The manufacturer’s instructions were followed with some modifications; instead of 50 to 100 mg of cells, 1 000 μl of overnight broth was used, centrifuged (Eppendorf 5417C, Germany) and the pellet was resuspended in phosphate buffered saline (PBS) solution (pH 7.2, Gibco® PBS, Life Technologies Corporation, USA) and 600 μl of the lysis solution were used instead of the prescribed 750 μl. The extracted genomic DNA was stored at -20 °C (Defy Ltd, Multimode, SA) until further analysis.
The detection of β-lactamase genes in A. baumannii by using M-PCR assays
Multiplex PCR assays were performed by using the QIAGEN® M-PCR kit (QIAGEN®, Germany) for the detection of several antibiotic resistance genes in
A. baumannii (Table
1). An initial denaturation step of 95 °C for 15 min, followed by 30 cycles of denaturation at 94 °C for 30 s, an annealing temperature dependent on the melting temperature of the primer pair (multiplex I: 52 °C; multiplex II and III: 57 °C and multiplex IV: 60 °C) and extension at 72 °C for 90 s, followed by the final extension step at 70 °C for 10 min. A negative control (sterile ultrapure water) was included for all M-PCR assays. The following positive controls were used for multiplex I to IV:
A. baumannii ATCC® BAA-1605™ strain;
K. pneumoniae ATCC 8303 [
Klebsiella pneumoniae carbapenemase (KPC) positive control] and
K. pneumoniae ATCC BAA-2146 [New Delhi metallo-β-lactamase (NDM) positive control]. All isolates testing negative for the presence of the inherent OXA-51 gene were sent for identification using matrix assisted laser desorption time of flight mass spectrometry (MALDI-TOF).
Table 1
Oligonucleotide primer sequences used for the detection of β-lactamases in four Multiplex PCR assays of A. baumannii clinical isolates
Multiplex I: |
OXA-23 |
bla
OXA-23 (F) | GATCGGATTGGAGAACCAGA | 501 | |
bla
OXA-23 (R) | ATTTCTGACCGCATTTCCAT |
OXA-24 |
bla
OXA-24 (F) | GGTTAGTTGGCCCCCTTAAA | 246 |
bla
OXA-24 (R) | AGTTGAGCGAAAAGGGGATT |
OXA-51 |
bla
OXA-51 (F) | TAATGCTTTGATCGGCCTTG | 353 |
bla
OXA-51 (R) | TGGATTGCACTTCATCTTGG |
OXA-58 |
bla
OXA-58 (F) | AAGTATTGGGGCTTGTGCTG | 599 |
bla
OXA-58 (R) | CCCCTCTGCGCTCTACATAC |
Multiplex II: |
VIM |
bla
VIM (F) | GATGGTGTTTGGTCGCATA | 390 | |
bla
VIM (R) | CGAATGCGCAGCACCAG |
KPC |
bla
KPC (F) | CATTCAAGGGCTTTCTTGCTGC | 538 |
bla
KPC (R) | ACGACGGCATAGTCATTTGC |
IMP |
bla
IMP (F) | TTGACACTCCATTTACDGb
| 139 |
bla
IMP (R) | GATYGAGAATTAAGCCACYCT |
Multiplex III: |
GES |
bla
GES (F) | AGTCGGCTAGACCGGAAAG | 399 | |
bla
GES (R) | TTTGTCCGTGCTCAGGAT |
PER |
bla
PER (F) | GCTCCGATAATGAAAGCGT | 520 |
bla
PER (R) | TTCGGCTTGACTCGGCTGA |
VEB |
bla
VEB (F) | CATTTCCCGATGCAAAGCGT | 648 |
bla
VEB (R) | CGAAGTTTCTTTGGACTCTG |
Multiplex IV: |
GIM |
bla
GIM (F) | CGTTGCCAGCTTTAGCTCAGG | 279 | |
bla
GIM (R) | GCAACTTGATACCAGCAGTGCG |
SPM |
bla
SPM (F) | GGGTGGCTAAGACTATGAAGCC | 447 |
bla
SPM (R) | GCCGCCGAGCTGAATCGG |
SIM-1 |
bla
SIM (F) | TTGCGGAAGAAGCCCAGCCAG | 613 |
bla
SIM (R) | GCGTCTCCGATTTCACTGTGGC |
NDM – 1 |
bla
NDM-1(F) | CCCGGCCACACCAGTGACA | 129 |
bla
NDM-1(R) | GTAGTGCTCAGTGTCGGCAT |
Analysis of the Multiplex PCR products
A 1.8 % (m/v) MetaPhor agarose gel (Lonza, USA) with 5 μl of ethidium bromide (EtBr) (10 μg.ml−1) (Promega, USA) was used to visualise the PCR products after gel electrophoresis (Bio-Rad, USA). The 1.8 % (m/v) MetaPhor® agarose gel was prepared according to the manufacturer’s instructions (Lonza, USA).
A ready-to-use 100 bp DNA ladder (Fermentas, ThermoScientific, USA) was used to determine the approximate sizes of the amplicons. Gel electrophoresis (Bio-Rad, Germany) was performed at 90 V for 100 min with 1x Tris-Borate-EDTA (TBE) buffer [40 mM Tris-HCl, 20 mM Boric acid (Sigma-Aldrich, USA) and 1 mM EDTA (Sigma-Aldrich, USA), pH 8.0]. The bands were visualised and captured using a UVP Doc It transilluminator (Ultra-violet Products Incorporated, USA).
Determination of the genetic relationship of clinical A. baumannii isolates using the PFGE assay
The Pulsenet PFGE protocol [
30] was followed with some modifications. The optical density (OD) was measured at 630 nm. The absorbance was adjusted to an OD range of 1.4 to 1.6 by using the ELx800 absorbance microplate reader (Bio-TEK, USA). The plugs were incubated overnight at 51°C and after incubation the plug was transferred to a new BD falcon tube (BD Biosciences, USA). The
ApaI enzyme (New England Biolabs, USA) was used and the
Salmonella choleraesuis serovar Braenderup (ATCC BAA-664) was used as a marker. Gel electrophoresis was performed using the following set conditions: interval 25 linear 5; angle 120° constant; voltage 220 V linear to 200 V and ran for 24 h. After the PFGE run was completed, the 1.2 % (m/v) SeaKem® LE agarose (Lonza, USA) gel was stained in 1 L distilled water, containing 250 μl EtBr (10 μg.ml-1) for 15 min and was destained in distilled water for 30 min. The gel was viewed and documented with a UV Doc It transilluminator (Ultra-violet Products Incorporated, USA). The banding patterns were analysed using the GelCompar II software programme (Applied Maths, Belgium). A distance matrix was constructed using the Dice coefficient and a dendogramme was constructed using the unweighted pair group method with arithmetic mean (UPGMA) [
31]. Pulsotype designation was based on isolates showing ≥80 % relatedness, which corresponds to the Tenover criteria of possibly related 4 to 6 band differences between isolates [
32,
33]. Representatives from each major PFGE pulsotype (≥5 isolates) and selective minor pulsotypes (< 5 isolates) with ≥80 % similarity were chosen for MLST analyses.
Molecular epidemiology
Seven housekeeping genes were used for MLST which included the following genes: (i) citrate synthase (
gltA); (ii) DNA gyrase subunit B (
gyrB); (iii) glucose dehydrogenase B (
gdhB); (iv) homologous recombination factor (
recA); (v) 60-kDa chaperonin (
cpn60); (vi) glucose-6-phosphate isomerase (
gpi) and (vii) RNA polymerase sigma factor (
rpoD) (
http://pubmlst.org/abaumannii/info/primers_Oxford.shtml). All the primers were synthesised by Inqaba Biotechnical Industries, South Africa.
All seven housekeeping genes were individually amplified in a 50 μl reaction as described in the Bioline kit’s guideline (Bioline, UK). The following programme was used on the thermal cycler (G-storm Model GS4822, Labtech, UK): an initial denaturation step of 94 °C for 2 min, followed by 35 cycles of denaturation at 94 °C for 30 s, an annealing temperature at 50 °C for 30 s, extension at 72 °C for 30 s, followed by the final extension step at 72 °C for 5 min [
34]. The amplicons were visualised on a 1.8 % (m/v) MetaPhor agarose gel (Lonza, USA) as described previously. All the amplicons were sequenced in both forward and reverse directions by Inqaba Biotechnical Industries, South Africa. An ABI file obtained for each sequence were analysed with the CLC Main Workbench Version 6.0 (CLCbio, USA). The sequences were assigned to corresponding allelic profiles and sequence types by using the MLST Oxford database (
http://pubmlst.org/abaumannii/).
Data analysis
One hundred consecutive A. baumannii isolates were collected, which were identified by the automated VITEK® 2 system (BioMérieux, France). Six isolates were excluded from this study because the isolates were incorrectly identified as A. baumannii by the VITEK® 2 system. These isolates were identified by MALDI-TOF analysis as: A. pittii; K. pneumoniae; Pseudomonas putida (2x); Staphylococcus aureus and S. epidermidis.
The results are reported on 94 clinical
A. baumannii isolates. The proportions of
A. baumannii isolates with Ambler classes A, B and D resistance genes were determined for the 94 isolates. The results were reported as proportions or percentages along with 95 % confidence intervals for the given sample size with an accuracy of within 10 %. The antimicrobial susceptibility and genotypic results obtained in this study were compared to a study done in 2008 in the same clinical setting [
5]. The antimicrobial susceptibility profiles were also compared to the surveillance data from 12 sentinel public hospitals in South Africa (SA) [
35].
Discussion
The increase in resistance profiles of
A. baumannii can be due to the misuse of antibiotics in clinical settings, or because of patients who do not complete the prescribed antibiotic course. The
bla
OXA genes have an impact on treatment because it can already be resistant to the new inhibitors, such as β-lactamase inhibitors, or it can rapidly gain resistance [
19].
The ability of the OXA β-lactamases to confer carbapenem resistance has already had a huge impact on the ability to treat Gram-negative pathogens and the current situation suggests that the problem is only set to increase [
19]. The advantage of using molecular tests, such as the M-PCR assays used in this study, is that resistance genes circulating in clinical settings can be determined as this is not possible with phenotypic tests. Through molecular methods this study has indicated that 99 % of the
A. baumannii isolates contained the OXA-51 gene, compared to Kock et al. [
5] where only 81% out of the 97 isolates contained this gene. In the Kock et al. [
5] study, the species identification obtained by the VITEK® 2 system was not confirmed by an alternative method, such as MALDI-TOF MS and thus possibly includes the other species of the
A. baumannii-calcoaceticus complex. The results obtained in the current study shows the establishment of the species,
A. baumannii in our clinical setting. The OXA-51 cluster is chromosomally encoded and is inherent to all
A. baumannii isolates [
4,
16,
19,
23,
35]
. Seven isolates (of which six isolates were excluded) did not have the inherent OXA-51 gene. These isolates were further analysed using MALDI-TOF, which identified the seven isolates as:
A. baumannii;
A. pittii; K. pneumoniae; Pseudomonas putida (2x)
; Staphylococcus aureus and
S. epidermidis. Studies have shown that the detection of the
bla
OXA-51-like gene can be used as a simple and reliable way of identifying
A. baumannii [
36,
37]. Although it is clear that
bla
OXA-51-like genes are present in at least the vast majority of isolates of
A. baumannii, there has been some debate as to whether they are present in all isolates of this species [
36]. Zander et al. [
38] reported that three
A. baumannii isolates did not harbour the predicted
bla
OXA-51-like gene (353 bp) and instead amplified PCR products of 1.2 kb and 1.6 kb. The authors then used external primers and the amplicons revealed that the one isolate (Ab-511) possessed the
bla
OXA-51 variant
bla
OXA-66, which was interrupted by the insertion sequence IS
Aba15 at nucleotide number 435, while the two other isolates (Ab-508 and Ab-653) possessed the
bla
OXA-51 variant
bla
OXA-78, which was disrupted at nucleotide number 379 by a novel insertion sequence IS
Aba19 [
38]. Further investigation of the genomic environment of the isolate identified as
A. baumannii by MALDI-TOF will reveal the changes in the OXA-51 genetic structure. Additionally the 16S-23S ribosomal spacer region can be sequenced for further species confirmation [
37].
The occurrence of the OXA-23 gene has been reported since 1985 in
A. baumannii isolates and has now disseminated worldwide [
5,
19,
39,
40]. The prevalence of OXA-23 in this study was 77 %, compared to Kock et al. [
5], who reported a 59 % prevalence in the same clinical setting in 2008. It is clear that the prevalence of OXA-23 in
A. baumannii isolates in this clinical setting is increasing. This increased occurrence can be due to poor infection control practices, spread through contact with healthcare workers or from patient to patient and prolonged antibiotic use [
11]. Stricter adherence to infection prevention and control policies by healthcare workers are needed to prevent further dissemination [
4,
41,
42]. Correct use of disinfectants and solutions containing ethanol is particularly important because if applied at very low concentrations
A. baumannii bacterium can become more pathogenic [
41].
The high prevalence of OXA-23 can be due to the acquisition of genetic elements, such as plasmids and transposons [
1,
42]. Liakopoulos et al. [
43] conducted a study in 2010 to 2011 and reported a 95 % prevalence of OXA-23 in Greece. The OXA-23 cluster can be located on a plasmid or chromosome [
5]. Mugnier et al. [
44] reported in 2010 that eleven isolates had the
bla
OXA-23 gene on the chromosome, nine isolates had the
bla
OXA-23 gene on a plasmid and one isolate had two copies of the
bla
OXA-23 gene, one on the plasmid and one on the chromosome. The OXA-23 enzyme has a higher affinity to hydrolyse imipenem than meropenem, ertapenem or doripenem [
19]. The MDR
A. baumannii isolates in this study that contained the OXA-23 gene showed a 69 % resistance to imipenem (1 % had intermediate susceptibility and 6 % were susceptible to imipenem). Imipenem susceptibility may be explained by the absence of the IS
AbaI element upstream of the
bla
OXA-23 gene [
45]. Twenty-two percent of the isolates were positive for the OXA-51 gene only. However, 17 % of these isolates were resistant to imipenem while 5 % were susceptible.
The prevalence rates of isolates that contained both the OXA-23 and OXA-51 genes were 77 %. It is evident that the carbapenem-hydrolysing class D β-lactamases (CHDL’s), such as OXA-23 and OXA-51 were prevalent in clinical A. baumannii isolates in this study. The high prevalence of the OXA-51 and OXA-23 group (77 %) is due to the spread and establishment of several successful clones in our clinical setting.
None of the clinical
A. baumannii isolates contained the genes that were screened for in M-PCR II to IV as well as the OXA-24 and OXA-58 genes that were part of M-PCR I. The GIM, SIM and SPM metallo-β-lactamase (MBL) genes are rarely found in
A. baumannii isolates [
46]. Kock et al. [
5] reported one
A. baumannii isolate that was positive for the VIM-like gene. Safari et al. [
47] reported that 99 % of the isolates in ICU wards of three educational hospitals of Hamadan City in Iran produced metallo-β-lactamases. In 2013, the first outbreak of NDM-1 producing
A. baumannii was reported in France, which occurred in a surgical ICU [
48]. Fortunately NDM-1 in
A. baumannii has not yet been detected in our clinical setting. The NDM-1 gene encodes an MBL carbapenemase that has a high hydrolytic activity for carbapenem antibiotics, and much more [
49].
PFGE is suited for studying isolates from a defined temporal and spatial epidemiologic setting, whereas MLST has an advantage in defining clonal lineages of isolates from larger geographic areas over time [
50]. Three clonal lineages of
A. baumannii, commonly referred to as the pan-European clonal lineages (EUI, EUII, and EUIII), have predominated in many European countries since the 1990s [
50]. Isolates belonging to the pan-European clonal lineages I and II (EUI and EUII) were identified in our clinical setting. EUI (ST258 and ST758) was predominant together with the OXA-23 gene. Other diverse clones, such as ST106 and ST339 are also circulating in our clinical setting. The origin of the
A. baumannii isolates in our setting cannot be attributed to the dissemination of one or two successful clones but rather to the dissemination of a diverse group of successful clones.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
ML, MME, AWD and MMK drafted and finalised the manuscript. ML collected all the clinical A. baumannii isolates and carried out the laboratory assays. ML, MME, AWD and MMK participated in data analysis and interpretation. ML, MME and MMK were involved in the concept design. All authors read and approved the final version of the manuscript.