|Home | About | Journals | Submit | Contact Us | Français|
This study aimed to evaluate the prevalence of carbapenemase-producing Enterobacteriaceae in Luanda, Angola. A total of 157 rectal samples were collected from children visiting a pediatric hospital in Luanda in March 2015. Fifty-seven imipenem-nonsusceptible enterobacterial isolates were recovered, most of which were non-clonally related. The blaOXA-181 (50/57) and blaNDM-1 (7/57) carbapenemase genes were identified. Notably, OXA-181-producing Escherichia coli isolates rarely coproduced extended-spectrum β-lactamases and consequently remained susceptible to broad-spectrum cephalosporins. The blaOXA-181 gene was always located on an IncX3 plasmid, while the blaNDM-1 gene was located on either IncFIA or IncA/C plasmids. The study identified a high prevalence of OXA-181 among hospitalized children in Angola.
The increasing occurrence of carbapenem resistance in Gram-negative bacteria is one of the major public health problems involving Enterobacteriaceae. Carbapenem resistance arises from two main mechanisms: decreased uptake of carbapenems by porins associated with overexpression of a β-lactamase with weak carbapenemase activity, and the acquisition of carbapenemases (1). Carbapenemases belong to three different Ambler classes (2), with class A including the serine carbapenemases KPC, NMC/IMI, and SME; class B including the metallo-β-lactamases VIM, IMP, and NDM; and class D including OXA-48-like β-lactamases (OXA-48, OXA-181, etc.). Since the initial report of OXA-48 from a Klebsiella pneumoniae isolate in 2003 (3), OXA-48-like enzymes have been recovered in a variety of Enterobacteriaceae worldwide, particularly within the Mediterranean area, including southern Europe and northern Africa, but also in the Middle East (4). Variants of OXA-48 have been described (5), including OXA-181, which differs from OXA-48 by 4 amino acids while sharing a very similar hydrolytic profile (6). Bacteria producing OXA-48-like β-lactamases do not always exhibit high levels of resistance to carbapenems, making their detection challenging (4). The epidemiology of carbapenemases in Africa remains mostly unknown, with the exception of OXA-48, which was identified in the northern part of the continent (Algeria, Egypt, Morocco, and Tunisia) (4, 7), and some reports of KPC- and NDM-like enzymes in South Africa (7). However, there are limited data about the occurrence of carbapenemase producers in most of the other African countries. Therefore, we investigated the occurrence and molecular characteristics of carbapenemase-producing enterobacterial isolates recovered from rectal samples from children visiting a pediatric hospital in Luanda, Angola.
A total of 157 rectal swabs were collected on 30 and 31 March 2015 from hospitalized patients (n = 100) and from ambulatory patients (n = 57). The samples were incubated in Luria-Bertani (LB) broth supplemented with ertapenem (0.25 μg/ml) for 10 h. From each broth, a calibrated inoculated loop (10 μl) was plated onto ChromID CarbaSmart selective medium (bioMérieux, La Balme-les-Grottes, France) to select for carbapenem-resistant isolates. The isolates were identified at the species level using the API20E system (bioMérieux). Antimicrobial susceptibility testing was performed according to the disc diffusion method following the CLSI recommendations (8) and using cation-adjusted Mueller-Hinton (MH) plates (Bio-Rad, Cressier, Switzerland).
MICs were determined by Etest (bioMérieux, La Balme-les-Grottes, France) for imipenem, meropenem, and ceftazidime. Carbapenemase activity was assessed using the Rapidec Carba NP test (9) (bioMérieux) for each isolate growing on the ChromID CarbaSmart plate.
Carbapenemase- and extended-spectrum β-lactamase (ESBL)-encoding genes were identified by PCR amplification using specific primers as described previously (10,–12), followed by sequencing (Microsynth, Balgach, Switzerland).
Considering the high-level resistance to all aminoglycosides observed for some isolates, a search of 16S rRNA methylase-encoding genes was performed by multiplex PCR as described previously (13). Similarly, a search of plasmid-mediated Qnr-like encoding genes involved in reduced susceptibility to quinolones was performed by multiplex PCR, as described previously (14). Finally, a search of the plasmid-mediated colistin resistance mcr-1 gene was performed by real-time PCR (15).
The clonal relationship of the isolates was evaluated by pulsed-field gel electrophoresis (PFGE). Total DNA from K. pneumoniae isolates and Escherichia coli isolates was digested by using the XbaI enzyme (New England BioLabs, Ipswich, MA, USA). The generated fragments were separated by PFGE using a CHEF-DR III System (Bio-Rad), followed by multilocus sequence typing (MLST) (16) for one strain of each different PFGE profile. Sequence types (STs) were assigned using the databases (http://mlst.warwick.ac.uk/mlst/dbs/Ecoli/ and http://bigsdb.web.pasteur.fr/klebsiella/klebsiella.html) for E. coli isolates and for K. pneumoniae isolates, respectively. DNAs of plasmids harboring the blaOXA-181 gene were extracted using the Kieser extraction method (17), electroporated into E. coli TOP10, and selected on LB agar plates supplemented with temocillin (50 μg/ml), which is selective for OXA-48-like β-lactamase producers. Plasmid sizes were evaluated by agarose gel electrophoresis using E. coli NCTC50192 harboring four plasmids of 154, 66, 48, and 7 kb as plasmid size markers. Plasmids carrying blaOXA-181 were characterized by PCR-based replicon typing (PBRT) as described previously (18) by including primers specific for IncX3-type plasmid backbones (IncX3 Fw, 5′-GAG GCT TAT CGT GAA GAC AG-3′; IncX3 Rv, 5′-GAA CGA CTT TGT CAA ACT CC-3′) and by restriction fragment length polymorphism (RFLP) using restriction enzymes PvuII and HindIII, respectively. The genetic environment of the blaOXA-181 gene was investigated by PCR mapping with primers specific for insertion sequence ISEcp1, ISKpn19, or IS3000, as previously described (6, 19), since previous studies showed they might be located upstream of blaOXA-181.
Mating-out assays were performed using the azide-resistant E. coli J53 as the recipient. E. coli J53 and blaOXA-181-carrying donors were separately inoculated overnight into LB broth and incubated. The samples were then mixed at a ratio of 10:1 (donor/recipient) for 5 h and plated onto LB agar plates supplemented with temocillin (50 μg/ml) and sodium azide (100 μg/ml). Susceptibility testing was performed for all E. coli transconjugants, and positivity for blaOXA-181 was checked by PCR.
The protocol was approved by the institutional review board of the hospital. Informed consent was obtained from the guardians of the children after a verbal presentation of the purpose, method, and design of the study.
All the patients were children from 3 months to 13 years old. They were visiting the hospital for several reasons, either for treating infections (pneumonia, peritonitis, malaria, and hepatitis) or for surgery, tumors, leukemia, malnutrition, or renal failure. Limited data in relation to antimicrobial therapy were available; the only important issue was that penicillins and broad-spectrum cephalosporins are quite often used in the hospital, but none of the children had previously received carbapenems.
From the 157 patients, a total of 57 imipenem-nonsusceptible and carbapenemase-producing enterobacterial isolates were isolated (Table 1). They were recovered from a total of 48 patients; 9 patients were colonized with two different carbapenemase producers. Out of the 57 community patients screened, only a single carbapenemase producer was recovered, while 42 out of the 100 hospitalized patients were colonized by at least one carbapenemase-producing isolate.
Among the 57 carbapenemase-producing isolates, 50 were found positive for the blaOXA-181 gene and 7 were positive for the blaNDM-1 gene. None of the isolates coproduced two carbapenemases. Among the blaOXA-181-positive enterobacterial isolates, 25 E. coli isolates, 24 K. pneumoniae isolates, and a single Enterobacter cloacae isolate were identified. The seven blaNDM-1 producers were E. coli (n = 4), K. pneumoniae (n = 1), Providencia stuartii (n = 1), and Providencia rettgeri (n = 1). Moreover, six out of the seven blaNDM-1-positive isolates coharbored the rmtB or rmtH 16S rRNA methyltransferase gene (Table 1). No isolate was positive for the mcr-1 gene.
Notably, the majority of the isolates belonging to either the E. coli or K. pneumoniae species exhibited low carbapenem MIC values, in particular for imipenem, with most isolates showing MIC values of 0.5 or 1 μg/ml.
The majority of the OXA-181-producing E. coli isolates (64%) did not coproduce any ESBL and remained susceptible to broad-spectrum cephalosporins, whereas almost all the K. pneumoniae isolates (96%) were ESBL producers (Table 1).
Mating-out assays followed by plasmid analysis revealed four different-size plasmids (ca. 30 kb, 64 kb, 70 kb, and >150 kb) carrying the blaOXA-181 gene (Table 1). E. coli transconjugants harboring the 30-kb plasmid showed slightly higher MICs of carbapenems than the other transconjugants (4- and 2-fold-increased MICs for imipenem and meropenem, respectively), suggesting that the corresponding plasmid type could likely be present at higher copy numbers and therefore could be enhancing the expression of the carbapenemase gene. Notably, although no additional β-lactamase gene was identified on the other two plasmid scaffolds, PCR amplification showed that the 150-kb plasmid bearing the blaOXA-181 gene coharbored the blaTEM-1 and blaCTX-M-15 β-lactamase genes.
PCR mapping performed on all the positive isolates revealed that a remnant of the ISEcp1 element was located upstream of the blaOXA-181 gene, as previously reported on the IncX3-type plasmid pOXA-181 from a Chinese E. coli isolate (20), with ISEcp1 being truncated by the insertion of IS3000. Downstream of blaOXA-181, and similar to what was observed on pOXA-181, ISKpn19 was identified. Further downstream, the qnrS1 gene, encoding resistance to quinolones and found on pOXA-181, was identified. Overall, the same structure as that identified on pOXA-181 was detected.
The mating-out assays were successful for all the carbapenemase-producing isolates as donors, except for a single blaOXA-181-positive K. pneumoniae isolate (Table 1). PBRT analysis showed that all the different-size plasmids carrying the blaOXA-181 gene belonged to the same IncX3 group, and RFLP confirmed that they shared a common scaffold structure, with similar-size bands (data not shown). The blaNDM-1 gene was identified on an IncFIA-type plasmid for three E. coli isolates and on an IncA/C-type plasmid for a single P. rettgeri isolate. PFGE and MLST analyses identified nine different E. coli clones and 10 different K. pneumoniae clones (Table 1).
Here, we report a high rate of recovery of carbapenemase-producing enterobacterial isolates from children in Angola and an important dissemination of the blaOXA-181 carbapenemase gene, which might be considered endemic in that geographical area, through the diffusion of conjugative plasmids. Indeed, 27.4% of the screened individuals were found positive for carbapenemases, and 88% of those carbapenemase-producing isolates harbored the blaOXA-181 gene. This is particularly noteworthy considering that none of the patients had received any current or past carbapenem-based antimicrobial therapy.
The occurrence of NDM-1-producing isolates is also noteworthy, since the identification of NDM-1 producers in Tunisia, Morocco, Algeria, and South Africa shows that this carbapenemase, known to be widespread in the Indian subcontinent, may also be widespread in Africa.
The rare association of the blaOXA-181 gene with an ESBL-encoding gene in the E. coli isolates contrasts with the frequent associations observed among blaOXA-48-positive K. pneumoniae isolates, in particular with the blaCTX-M-15 ESBL gene. This further underlines the fact that the blaOXA-48 and blaOXA-181 genes, although structurally related, have distinct origins and epidemiologies.
According to our data, the high rate of OXA-181 producers in Luanda results from the spread of some predominant E. coli and K. pneumoniae clones, but also from the dissemination of a self-conjugative IncX3-type plasmid among different enterobacterial isolates.
The plasmids harboring the blaOXA-181 gene were different in size; nevertheless, they all shared the same backbone. It might be hypothesized that they derive from a common IncX3 conjugative plasmid. Notably, although all blaOXA-181-positive plasmids coharbored the qnrS1 gene conferring reduced susceptibility to quinolones, no additional resistance markers were detected, with the exception of a single 150-kb plasmid identified in only one isolate that coharbored the blaTEM-1 and blaCTX-M-15 β-lactamase genes. In addition, this 150-kb plasmid was not self-conjugative, in contrast to the other blaOXA-181-bearing plasmids identified. This might be the consequence of the insertion of the two additional β-lactamase genes in a region of the plasmid that is crucial for its conjugative property.
Interestingly, the genetic environment of the blaOXA-181 gene was identical to that identified recently by Liu et al. (20) in China, also on an IncX3-type plasmid, with the blaOXA-181 gene flanked upstream by ISEcp1 truncated by IS3000 and downstream by ISKpn19 followed by the qnrS1 gene. In other reports from different parts of the world, the blaOXA-181 gene was identified on different plasmid backbones, including IncN and IncT plasmids (19, 21). IncX3-type plasmids harboring the blaOXA-181 gene seem to be predominant in Asia. Indeed, a recent study showed that OXA-181-producing isolates had been imported into Switzerland on fresh vegetables originating from Asia (22).
Our data support the importance of active and continuous surveillance of carbapenemase-producing Gram-negative bacteria in health care facilities. They also show that not only Asia, but also Africa, may act as an important reservoir of OXA-181.
Altogether, most isolates coharbored a high number of resistance determinants that represent a major source of concern in terms of public health.
This work has been funded by the University of Fribourg, Fribourg, Switzerland, and by project PTDC/DTP-EPI/0842/2014 from Fundação para a Ciência e a Tecnologia, Portugal.