ABSTRACT
Here, we evaluated the immunochromatographic assay NG-Test Carba 5v2 (NG-Biotech), with improved IMP variant detection on 31 IMP producers, representing the different branches of the IMP phylogeny, including 32 OXA-48, 19 KPC, 12 VIM, 14 NDM, and 13 multiple carbapenemase producers (CPs), 13 CPs that were not targeted, and 13 carbapenemase-negative isolates. All tested IMP variants were accurately detected without impairing detection of the other carbapenemases. Thus, NG-Test Carba 5v2 is now well adapted to countries with high IMP prevalence and to the epidemiology of CP-Pseudomonas aeruginosa, where IMPs are most frequently detected.
TEXT
Carbapenem resistance among Gram-negative bacilli (GNB) (Enterobacterales, Pseudomonas, and Acinetobacter species) has become a major public health issue (1). Resistance to carbapenems can stem from production of carbapenemases or other mechanisms, such as decreased permeability, overproduction of extended-spectrum β-lactamases (ESBL) or cephalosporinases, efflux pumps, or combinations of these mechanisms (2). Carbapenemase producers (CPs) are by far the most worrisome; thus, their rapid detection and identification are essential to help physicians to quickly implement appropriate infection control measures, to adapt antibiotic treatment rapidly, and to optimize care strategies and outcomes (3). Based on their amino acid sequence, carbapenemases are divided into different molecular classes, A, B, and D, of the Ambler classification. Class A (mainly KPC enzymes) and D (mostly OXA-48-like enzymes) carbapenemases are serine active-site enzymes, while class B carbapenemases, which are also called metallo-β-lactamases (MBLs) (mostly enzymes of NDM, VIM, and IMP types), require zinc ions to be active (2). The IMP family of carbapenemases is a very heterogeneous family of enzymes (sharing only 79% amino acid sequence identity), rendering their detection difficult (4–6) (Fig. 1).
Phylogenetic tree of IMP variants. Amino acid sequences were from the BLDB web site (http://www.bldb.eu/alignment.php?align=B1:IMP). Alignment and phylogenetic reconstructions were performed using the function “build” of ETE3 v3.0.0b32 (23), as implemented on the GenomeNet (https://www.genome.jp/tools/ete/). A maximum-likelihood tree was inferred using PhyML v20160115 run with model JTT and parameters −f m –pinv e −o tlr –alpha e –nclasses 4 –bootstrap −2 (24). Tested isolates are in boldface, and boxed isolates are those that were not detected with version 1 of the NG-Test Carba 5.
Recently, the NG-Test Carba 5 immunochromatographic assay (ICA) (NG Biotech, Guipry, France) was developed to detect the five most widespread carbapenemase families in Enterobacterales (CPEs) (i.e., KPC, NDM, VIM, IMP, and OXA-48-like enzymes). It was demonstrated to accurately identify the claimed enzymes in culture and also from positive blood cultures growing with Enterobacterales (7, 8). Unlike other ICAs developed to detect CPEs (RESIT-4 OKVN; Coris Bioconcept, Gembloux, Belgium), the NG-Test Carba 5 also targets the IMP type enzymes, which are more prevalent in CPEs from the Asian continent and in nonfermenters worldwide (5, 9). In a recent study, Potron et al. have evaluated the performance of the NG-Test Carba 5 test for the detection of carbapenemase-producing Pseudomonas spp. and Acinetobacter spp. encountered in France (6). The NG-Test Carba 5 allows the detection of 14 out of the 21 IMP variants, with the 7 false-negative results corresponding to the IMP-13 clade (IMP-13 and IMP-37), IMP-14, IMP-15, IMP-18 clade (IMP-18 and IMP-71), and IMP-63 (6). Here, we have evaluated the NG-Test Carba 5v2, a novel version of the test with additional antibodies for the detection of all IMP variants.
Monoclonal antibodies (MAb) derived from IMP-immunized mice (7) were further tested against IMP-13 enzyme. To select the best MAb pairs for the development of the two-site lateral flow immunoassay with IMP-13, a combinatorial analysis was carried out using each MAb either as a capture or gold-labeled antibody as previously described (10). Briefly, the strips were prepared by spotting 0.5 μl of MAb (500 μg/ml in 50 mM phosphate buffer, pH 7.4) and then dried. One hundred microliters of a crude IMP-13 extract from P. aeruginosa (11) or extraction buffer alone and 10 μl of colloidal gold-labeled MAb were mixed in microtiter plate wells (Greiner, Paris, France) and allowed to react for 5 min before dipping the strip into the solution. After 30 min of migration, signals were analyzed by eye. The parameters used to select the best MAb pairs were the intensity of the visual signals obtained with IMP-13 crude extract and the absence of signal without IMP-13 (nonspecific signal). The NG-Test Carba 5v2 tests (strip plus cassette) were manufactured by NG Biotech using the additional IMP-13-selected MAbs.
In order to see whether the novel version, NG-Test Carba 5v2, is now able to detect all IMP variants without modifying the biological performances for the other targets, a collection of 147 isolates with whole genome sequence-characterized β-lactamase content were used to evaluate the NG-Test Carba 5v2 and the results compared with those obtained with the NG-Test Carba 5v1. Both assays were used as recommended by the manufacturer. Results were eye read and the time for appearance of a red band specific for a given carbapenemase was recorded. After 15 min of migration, in the absence of a band, the test was considered negative (Table 1).
Results of the NG-Test Carba 5 v1 and v2 on the collection of Enterobacterales, Pseudomonas spp., and Acinetobacter spp.
This collection included 134 carbapenemase producers and 13 non-carbapenemase producers. The produced carbapenemases included 14 NDM, 12 VIM, 31 IMP, 19 KPC, 32 OXA-48, 13 multiple carbapenemases, and 13 other carbapenemases not targeted by the NG-Test Carba 5v2 (Table 1). The 31 IMP variants corresponded to very diverse enzymes, as illustrated in Fig. 1.
The NG-Test Carba 5v2 gave results similar to those of NG-Test Carba 5v1, except for IMP results. The median time for a positive signal was comparable between the two assays, indicating that the addition of novel antibodies did not interfere with the other targets. The NG-Test Carba 5v2 allows the detection of 100% of the IMP variants tested (n = 31/31), unlike the NG-Test Carba 5v1, which detects only 67% of the tested variants (n = 21/31). The NG-Test Carba 5v2 identified all VIM (n = 12) and NDM (n = 14) producers with no false-positive results. All OXA-48 carbapenemase variants were correctly identified, including difficult-to-detect variants (OXA-244 and OXA-519), using biochemical tests such as Carba NP (12, 13). It also detects very recently identified variants that have never been tested by ICA (OXA-484, -505, -517, and -793; Table 1) and the distantly related OXA-535 variant (14), which is the progenitor of OXA-436, a distantly related OXA-48 variant responsible for an outbreak associated with several enterobacterial species in Denmark (14–16). Moreover, OXA-163 and OXA-405, two OXA-48 variants that lack significant imipenemase activity (17) but with strong expanded-spectrum hydrolytic activity, were consistently detected. It is still debated whether these enzymes are or are not carbapenemases (17, 18). From the point of view of diagnostic tests based on imipenem hydrolysis, these variants are not carbapenemases (17). These results are in line with an 820-fold decrease in OXA-163 catalytic efficiency for imipenem (19). However, only a 4-fold decrease in catalytic efficiency for meropenem and doripenem compared to that of OXA-48 was observed, suggesting that OXA-163 can still be considered a meropenemase or doripenemase (19). Similarly, as the specific carbapenemase activities of OXA-405 and OXA-163 are comparable (20), it is likely that OXA-405 behaves in a similar manner with respect to meropenem activity. As it has been shown that enzymes with low carbapenem-hydrolyzing activities, such as OXA-163, may develop clinically significant carbapenem resistance when expressed in bacterial isolates with decreased outer membrane permeability (12, 21), it is important that these two OXA variants are efficiently detected. Among the KPC-positive isolates, 90% (n = 17/19) were correctly identified. The two isolates that yielded negative results after 15 min of incubation corresponded to KPC-31 and KPC-33, two rare D179Y variants of KPC-3 and KPC-2, respectively, known to be responsible for ceftazidime/avibactam resistance and for not being detected using ICA tests such as KPC K-SET (Coris BioConcept) and NG-Test Carba 5v1 (NG Biotech). Of note, this D179Y variant impaired the carbapenem hydrolytic activity of this enzyme, which cannot be considered a carbapenemase anymore, leading to negative results with biochemical tests such as the Carba NP test (22). All 13 non-carbapenemase-producing isolates and all 13 isolates producing a carbapenemase different from those targeted by the NG-Test Carba 5v2 gave negative test results, demonstrating no cross-reactivity between carbapenemases.
In summary, the NG-Test Carba 5v2 is a significant improvement, as all of the tested IMP variants were detected without impairing the detection of the other four carbapenemases. The NG-Test Carba 5v2 is now well adapted to countries where the epidemiology of IMP producers is high, especially to CP-P. aeruginosa isolates, where the proportion of IMP producers is higher, e.g., <1% in CPEs and 8% in P. aeruginosa in 2017 in France (http://www.cnr-resistance-antibiotiques.fr/ressources/pages/Rapport_CNR_2017_VF_2.pdf). Considering the French epidemiology of CP-P. aeruginosa in 2017, the NG-Test Carba 5v2 might have correctly detected 97.4% of them (missing 10 GES and one DIM producer).
ACKNOWLEDGMENTS
We acknowledge the NG Biotech Company for providing the NG-Test Carba 5 v1 and v2 assays.
This work was partially funded by the University Paris-Sud, France. L.D., T.N., S.O., and D.G. are members of the Laboratory of Excellence in Research on Medication and Innovative Therapeutics (LERMIT), supported by a grant from the French National Research Agency (ANR-10-LABX-33).
L.D. is coinventor of the Carba NP test, a patent for which has been licensed to bioMérieux (La Balmes les Grottes, France).
FOOTNOTES
- Received 24 September 2019.
- Returned for modification 22 October 2019.
- Accepted 28 October 2019.
- Accepted manuscript posted online 4 November 2019.
- Copyright © 2019 American Society for Microbiology.