ABSTRACT
Metallo-β-lactamases (MBLs) are transmissible carbapenemases of increasing prevalence in Gram-negative bacteria among health care facilities worldwide. Control of the further spread of these carbapenem-resistant bacteria relies on clinical microbiological laboratories correctly identifying and classifying the MBLs. In this study, we evaluated a simple and rapid method for detecting IMP, the most prevalent MBL in Japan. We used an immunochromatography (IC) assay for 181 carbapenem-nonsusceptible (CNS) (nonsusceptible to imipenem or meropenem) strains comprising 74 IMP-producing and 33 non-IMP-producing strains of non-glucose-fermenting Gram-negative rods (NFGNR), as well as 64 IMP-producing and 10 non-IMP-producing Enterobacteriaceae strains. The IC assay results were compared to those from the double-disk synergy test (DDST), the MBL Etest, and the modified Hodge test (MHT) (only for Enterobacteriaceae). The IMP type was confirmed by specific PCR and direct sequencing. The IC assay detected all of the IMP-type MBLs, including IMP-1, -2, -6, -7, -10, -11, -19, -20, and -22 and IMP-40, -41, and -42 (new types), with 100% specificity and sensitivity against all strains tested. Although the sensitivity and specificity values for the DDST and MHT were equivalent to those for the IC assay, the MBL Etest was positive for only 87% of NFGNR and 31% of Enterobacteriaceae due to the low MIC of imipenem, causing an indeterminate evaluation. These results indicated that the IC assay might be a useful alternative to PCR for IMP MBL detection screening.
INTRODUCTION
The recent worldwide emergence and dissemination of carbapenemase-producing Gram-negative rods (GNR) that are resistant to carbapenems is a significant concern with respect to patient care and infection control strategies (1). The transmissible carbapenemases are divided into three different classes, class A (serine carbapenemases, such as Klebsiella pneumoniae carbapenemase [KPC]), class B (metallo-β-lactamases [MBLs], such as IMP, VIM, and NDM), and class D (OXA carbapenemases, such as OXA-23 and OXA-48) (1, 2). Rapid and adequate detection of carbapenemases is very important for appropriate antimicrobial chemotherapies and infection control measures. Various phenotypic confirmation tests for detecting carbapenemases have been performed, including inhibition tests of carbapenemase activity, the modified Hodge test (MHT), and detection of carbapenem hydrolysis (1–8). However, there are no complete assays available to confirm and specify carbapenemases correctly because carbapenemase-producing bacteria, notably Enterobacteriaceae, show variable carbapenem MIC distributions (even under the breakpoint) and sometimes have carbapenemase-independent mechanisms, such as reduced permeability by porin alternations, active efflux pumping, and hyperproduction of class C β-lactamases (e.g., AmpC) or extended-spectrum β-lactamases (ESBLs) that operate with or without carbapenemase activity (1–4). Moreover, phenotypic assays cannot specify types within each class of carbapenemases, such as IMP, VIM, NDM, SIM, and GIM in MBLs (1–4). Therefore, molecular confirmation of carbapenemases is recommended for suspected carbapenemase-producing strains (1–4). However, although molecular detection methods such as PCR and sequencing of carbapenemase genes are reliable for confirmation of carbapenemases, it is difficult to perform such tests in routine clinical microbiology laboratories because of the skill level required, the higher cost, and the special equipment required (1–4). A simple and rapid alternative method is thus needed to confirm carbapenemase presence in bacteria.
In Japan, IMP MBL is the most prevalent transmissible carbapenemase, particularly members of the IMP-1 group (9, 10), while KPC is quite rare and OXA-48 has not been reported (11). The first IMP MBL was described in Pseudomonas aeruginosa in Japan (12) and is now found worldwide in non-glucose-fermenting Gram-negative rods (NFGNR) other than P. aeruginosa and Enterobacteriaceae (1–4, 8, 13). Recently, Kitao et al. (14) developed an immunochromatography (IC) assay for the production of IMP MBL in P. aeruginosa and Acinetobacter. This assay is easy to perform and rapid (≤20 min required), requires no special equipment, and detects the 24 established IMP types. In addition, it shows excellent correlation with PCR results. In countries like Japan, wherein IMP MBL is the most prevalent mechanism of carbapenem resistance, this assay provides a useful alternative to PCR for classifying MBLs in clinical microbiology laboratories. Since it is uncertain that this system can detect IMP MBL in Enterobacteriaceae, we evaluated the usefulness of this IC assay in carbapenem-nonsusceptible (CNS) Enterobacteriaceae and NFGNR strains with an MIC of imipenem (IPM) or meropenem (MEM) of >1 μg/ml.
MATERIALS AND METHODS
Strains.A total of 181 CNS strains were used, including Pseudomonas aeruginosa MRY 06-352 (producing IMP-1) and Serratia marcescens MRY 06-353 (producing IMP-19) as IMP-positive controls provided by Y. Arakawa (National Infectious Disease Institute, Japan). The strains used also included 74 NFGNR (Pseudomonas aeruginosa, Pseudomonas spp., Acinetobacter spp., and Achromobacter xylosoxidans) and 64 Enterobacteriaceae (Citrobacter freundii, Enterobacter cloacae, Escherichia coli, Klebsiella spp., Providencia rettgeri, and Serratia marcescens) strains that produce IMP MBL. An additional 43 CNS, but carbapenemase-negative, strains (33 NFGNR and 10 Enterobacteriaceae) were also used as negative controls. The 179 strains, excluding the IMP-positive controls, were collected in the Miroku Medical Laboratory from 2001 to 2012. Each strain was species identified by the MicroScan Breakpoint Combo Panel Type 6.23J (Siemens Healthcare Diagnostics, Tarrytown, NY), and all of the NFGNR strains were reidentified using matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) (Microflex LT, Bruker Daltonik GmbH, Leipzig, Germany) with MALDI Biotyper software (version 3.0, Bruker Daltonik) (15, 16). Identification of Pseudomonas spp. other than P. aeruginosa and Acinetobacter spp. was confirmed by 16S rRNA gene, rpoB, or gyrB sequencing (17, 18). Strains used are listed in Table 1, while the clinical sources and the place and date of isolation are presented in Table S1 in the supplemental material.
IMP-producing and non-IMP-producing strains used in this study
MIC determination.The MICs of IPM and MEM were determined by the broth microdilution method according to the CLSI M07-A9 guideline (19) and the supplement M100-S21 (20). Nonsusceptibility to carbapenem was considered to be present in strains designated intermediate or resistant to IPM or MEM (MIC > 1 μg/ml).
Phenotypic detection of MBL.Double-disk synergy tests (DDSTs) with sodium mercaptoacetate (SMA) (metallo-β-lactamase SMA Eiken; Eiken Chemical Co., Ltd., Tokyo, Japan) were performed according to the manufacturer's instructions based on the method described previously (21). A McFarland 0.5 standard suspension of each test strain was inoculated on Mueller-Hinton agar (Nippon Becton-Dickinson, Fukushima, Japan). Two commercial Kirby-Bauer (KB) disks (Nippon Becton-Dickinson) containing 30 μg of ceftazidime (CAZ) or 10 μg IPM were placed on the plate and an SMA disk was placed at a distance of 10 mm (edge to edge). Each agar plate was incubated at 35°C overnight. The presence of a synergistic inhibition zone of CAZ or IPM (≥5 mm of enlargement with the SMA disk side) was interpreted as positive. The MBL Etest and MHT were performed according to a previous report (22) and the CLSI guideline (20).
Determination of IMP MBL genes.Screening of carbapenemase genes was carried out by PCR as described previously (9, 23). Strains carrying transmissible carbapenemases other than IMP (VIM, SIM, GIM, AIM, DIM, GIM, NDM, KPC, BIC, and OXA-48) were excluded from this study. If blaIMP gene sequencing was positive by screening PCR, blaIMP types were determined using each IMP-specific PCR to amplify the whole length of blaIMP, and direct sequencing was performed. Primers used are listed in Table 2. PCR was performed in 50-μl reaction mixtures that comprised 2.5 U of Ex Taq DNA polymerase (TaKaRa Bio Inc., Shiga, Japan), 0.2 mM deoxynucleoside triphosphate (dNTP), 25 pmol of each primer, and 2 μl of DNA template. PCR conditions were as follows: initial denaturation at 95°C for 10 min, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 62°C for 1 min, and DNA extension at 72°C for 1 min, with final extension at 72°C for 10 min. PCR products were visualized under UV light exposure after 1% agarose gel electrophoresis with ethidium bromide. Amplicons obtained from each PCR were sequenced using M13F and M13R primers, the BigDye Terminator v3.1 Cycle Sequencing Kit (ABI, Carlsbad, CA), and an ABI sequence analyzer 3730XL (ABI). Each type of IMP was determined by a BLAST search and data based on all the blaIMP sequencing described previously (http://www.lahey.org/studies/other.asp#Table 1) (24).
Primers for PCR and sequencing of IMP genes
Detection of IMP MBL by immunochromatography.The IMP MBL IC assay kit (Quick Chaser IMP) was kindly provided by the Mizuho Medy Co., Ltd. (Saga, Japan) and used according to the manufacturer's instruction, based on a previous report (14). Briefly, fresh cultured colonies were dispensed into 700 μl of extraction reagent solution with nonionic detergent at McFarland 4.0 standard. After vigorous vortexing, 3 drops of this suspension were applied onto the sample area of the IC assay test plate. These plates were incubated for 15 min at room temperature and results were interpreted visually (14).
Statistical analysis.Receiver operating characteristic (ROC) curves and the area under the ROC curve with its standard error were used to analyze the IC assay, DDST, and MHT method results, using PCR results for IMP as the gold standard. Statistically significant differences were evaluated by comparing 95% confidence intervals of the corresponding areas. The statistical analysis was performed using STATA 11.0 (Stata Corp LP, College Station, TX).
RESULTS
Species identification of test strains and determination of IMP type.The results of identification of all the strains of A. xylosoxidans and P. aeruginosa were identical for MicroScan and MALDI-TOF MS. Both IMP-producing and non-IMP-producing Pseudomonas spp. strains were identified as P. putida or P. moteilii, which belonged to the Pseudomonas putida group by MALDI-TOF MS, but were distinguished from any species in this group by 16S rRNA, rpoB, or gyrB sequencing, suggesting a new species. By rpoB sequencing, the 17 IMP-producing Acinetobacter strains were identified as 10 A. pittii strains, 3 Acinetobacter genomospecies 13 strains, 3 A. soli strains, and 1 A. johnsonii strain (18), whereas the 10 non-IMP-producing Acinetobacter strains were identified as A. baumannii (Table 1).
The most prevalent IMP was IMP-1, comprising 48 of 74 (64%) NFGNR and 49 of 64 (76%) Enterobacteriaceae strains. In this study, three new IMP-type MBLs were discovered and designated IMP-40, IMP-41, and IMP-42. IMP-40 is closely related to IMP-10, with a nucleotide alteration of T206C in blaIMP-10 causing an F69S amino acid substitution, according to the standard numbering scheme of MBL (25). IMP-41 is similar to IMP-11, with a G145T nucleotide change causing a V49F amino acid substitution. IMP-42 shows a G45R amino acid substitution of IMP-1, with a G133A nucleotide change. IMP-40, IMP-41, and IMP-42 were found in two P. aeruginosa, one P. aeruginosa, and two A. soli strains, respectively (Table 1; see also Table S1 in the supplemental material). One A. pittii and one E. coli strain showed a C294T nucleotide change in blaIMP-1 and a C306T change in blaIMP-6, respectively, but neither base change altered the amino acid sequence (Table 1).
Detection of IMP in CNS-NFGNR by three different methods.Table 3 shows the results of three IMP detection methods in CNS-NFGNR. The IC assay detected all IMP MBL regardless of IMP type, including the new IMPs, IMP-40, IMP-41, and IMP-42, with 100% sensitivity and specificity (area under the curve of the ROC, 1.000). DDST also showed good sensitivity and specificity at 96% and 98%, respectively. The areas under the ROC curve for the IC assay and DDST were not statistically different. However, there were fewer positive detections with the MBL Etest (87%) than with the IC assay and DDST because some strains showed a low MIC of IPM, and four tests (5%) were evaluated as “indeterminate” due to the MIC being out of range. Moreover, the MBL Etest showed negative results for three strains of IMP-1-producing P. aeruginosa, one IMP-1-producing A. xylosoxidans, and one IMP-11-Acinetobacter genomospecies 13. The MBL Etest results were excluded from statistical analysis because of data including indeterminate values.
Detection of IMP in CNS-NFGNR by three different methods
Detection of IMP in CNS Enterobacteriaceae by four different methods.Table 4 details results for the four IMP detection methods in CNS Enterobacteriaceae. In these bacteria, the IC assay also showed 100% sensitivity and specificity (area under the ROC curve, 1.000), and those of DDST and MHT were 95% and 96%, respectively; the areas under the curve were not statistically different among these three detection methods. In comparison, the MBL Etest was inadequate for 68% of IMP-producing CNS Enterobacteriaceae, because 59% of these strains showed low susceptibility to IMP within the Etest range (≤4 μg/ml). In addition, the area under the ROC curve could not be calculated for the MBL Etest results because some data were deemed indeterminate.
Detection of IMP in CNS-Enterobacteriaceae by four different methods
DISCUSSION
Many types of carbapenemase-producing Gram-negative rods have emerged worldwide, with various carbapenemase types disseminated among countries and continents. IMP MBL is one of the most prevalent carbapenemases in Asia, Europe, and some areas of North and South America and Australia (13). Currently, IMP MBL has spread through most Enterobacteriaceae, including E. coli and K. pneumoniae, both of which are prevalent in community-acquired and health care-associated infections (1–5, 13). Since some IMP-producing Enterobacteriaceae strains, especially E. coli and K. pneumoniae, show low-level resistance or even sensitivity to carbapenems, the CLSI breakpoints of carbapenems in Enterobacteriaceae have changed since June 2010, as follows. IPM breakpoints of ≤4 µg/ml (susceptible [S]), 8 µg/ml (intermediate [I]), and ≥16 μg/ml (resistant [R]) and MEM breakpoints of ≤4 μg/ml (S), 8 μg/ml (I), and ≥16 μg/ml (R) have moved to ≤1 μg/ml (S), 2 μg/ml (I), and ≥4 μg/ml (R) and ≤1 μg/ml (S), 2 μg/ml (I), and ≥4 μg/ml (R) (3, 19). Moreover, the European Committee on Antimicrobial Susceptibility Testing (EUCAST) has established epidemiological cutoff values (ECOFFs) that discriminate wild-type isolates lacking any carbapenem-resistance mechanisms from those possessing resistance as IPM (1 to 4 μg/ml) and MEM (0.125 to 0.25 μg/ml), in addition to carbapenem clinical breakpoints (2). If a strain of Enterobacteriaceae for which carbapenem MICs are below ECOFFs is detected and becomes prevalent, clinical microbiology laboratory staff should consider surveillance of carbapenemase-producing strains (2). Ideally, such surveillance by routine clinical microbiological laboratories would require a rapid, reliable, inexpensive, and simple detection method of IMP. The IC assay evaluated in this study showed excellent sensitivity (100%) and specificity (100%) across MIC ranges (area under the ROC curve, 1.000). Current phenotypic detection of MBL in clinical laboratories uses testing by DDST or MHT, and the IC assay results in this study are statistically comparable with those acquired using these two tests by ROC analysis. However, both DDST and MHT require overnight culture (>16 h), compared with only 20 min for the IC assay. The DDST is also not recommended by the Dutch Working Party on the Detection of Highly Resistant Microorganisms because the sensitivity depends on the optimal distance between the disks, and this cannot be predicted (4). In our results, the MBL Etest was inappropriate for detecting IMP, especially in Enterobacteriaceae, mainly because of the low MIC of IPM. These results confirm those of similar studies previously described (26–28). Moreover, the MBL Etest had five false-negative results involving four IMP-1- and one IMP-11-producing strains. In addition, Laraki et al. (29) reported that even 10 mM EDTA did not inhibit the enzyme activity of IMP-1 isolated from one Japanese strain. Poor inactivation of IMP by EDTA might account for such false-negative results of the MBL Etest. The IC assay detected all IMP types, including 3 new types, IMP-40, -41, and -42, and did not react with VIM-2 in two P. aeruginosa strains, NDM-1 in one K. pneumoniae isolate, OXA-23 in one A. baumannii strain, and KPC-2 in K. pneumoniae ATCC BAA-1705 (data not shown). Two monoclonal antibodies designated 4C9-C/F6 and 4E7-C/F6 are used in this IC assay system (14). Since 4C9-C/F6 and 4E7-C/F6 recognize the highly conserved amino acids 124 to 130 (the H2 region) and 134 to 140 (the S6 region) in IMP (14, 24), and since the H2 and S6 regions of the new IMP types, IMP-40, -41, and -42, are well conserved, the IC assay detects such IMP types. Thus, the IC assay might also be used for detecting emerging types of IMP that retain the original H2 and S6 regions. The cost of an IC assay is also comparable to other conventional or molecular detection methods of MBL at ¥1,000 per test.
In conclusion, the IC assay is considered an alternative and suitable method to detect IMP in routine clinical microbiology testing, including IMP surveillance, with advantages in cost, processing time, reliability, and convenience, and no requirement for special equipment. In countries like Japan, where IMP MBL is highly prevalent, and in areas of newly developed MBL endemicity, IC assay testing would be quite useful for the early detection and control of nosocomial and community spreads of bacteria with IMP MBL.
ACKNOWLEDGMENTS
We thank Y. Arakawa for providing the IMP-positive-control strains. We also thank T. Mizutani for assistance with the statistical analysis.
This study was supported in part by a Grant-in-Aide (S0991013) from the Ministry of Education, Culture, Sport, Science, and Technology, Japan (MEXT), for the Foundation of Strategic Research Projects in Private Universities.
FOOTNOTES
- Received 30 January 2013.
- Returned for modification 25 February 2013.
- Accepted 21 March 2013.
- Accepted manuscript posted online 27 March 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/JCM.00234-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
