Research ArticleLaboratory and Clinical Evaluation of DNAMicroarray for the Detection of Carbapenemase Genes inGram-Negative Bacteria from Hospitalized Patients

Yi Song 12 Fengna Dou12 Sha He3 Yu Zhou 4 and Qiqi Liu 12

1Beijing Institute of Radiation Medicine Beijing China2Beijing Key Laboratory of New Molecular Diagnosis Technologies for Infectious Diseases Beijing China3e First Hospital of Hebei Medical University Hebei China4Clinical Laboratory of Second Medical Center of PLA General Hospital Beijing China

Correspondence should be addressed to Yu Zhou zhouy427163com and Qiqi Liu liuqiqibmiaccn

Received 14 December 2018 Revised 12 March 2019 Accepted 2 April 2019 Published 13 May 2019

Guest Editor Payam Behzadi

Copyright copy 2019 Yi Song et al This is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Background The prevalence of a variety of carbapenemases in Gram-negative bacteria (GNB) has posed a global threat on clinicalcontrol and management Monitoring and controlling the carbapenemase-producing GNB became imperative tasks for manyhealthcare centers The aim of this study was to develop a high-throughput specific sensitive and rapid DNA microarray-basedmethod for the diagnosis phenotypic confirmation and molecular epidemiological study of carbapenemase genes MethodsWe targeted a panel of eight carbapenemase genes including 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM and119887119897119886DIM for detection Ultrasensitive chemiluminescence (CL) detection method was developed and used to simultaneously detecteight carbapenemase genes and plasmids were established as positive or limit of detection (LOD) reference materials Antibioticsusceptibility was determined by disk diffusion according to Clinical and Laboratory Standards Institute (CLSI) guidelines in orderto screen clinical isolates resistant to carbapenemantibiotics aswell as Sanger sequencingwhichwas used to confirm the reliability ofthe results presented byDNAmicroarrayResults Eight carbapenemase genes could be detectedwith high sensitivity and specificityThe absolute LOD of this strategy to detect serially diluted plasmids of eight carbapenemase genes was 102- 103copies120583LThen 416specimens collected fromhospital were detected and the results showed 966 concordance between the phenotypic andmicroarraytests Compared with Sanger sequencing a specificity and sensitivity of 100were recorded for 119887119897119886NDM-1 119887119897119886IMP 119887119897119886VIM and 119887119897119886DIMgenes The specificity for 119887119897119886KPC 119887119897119886OXA-23 119887119897119886OXA-48 and 119887119897119886OXA-51 genes was 100 and the sensitivity was 985 976 957and 979 respectively The overall consistency rate between the sequencing and microarray is 978 Conclusions The proposedultrasensitive CL imaging DNA hybridization has high specificity sensitivity and reproducibility and could detect and differentiateclinical specimens that carried various carbapenemase genes suggesting that the method can conveniently be customized for high-throughput detection of the carbapenemase-producing GNB and can be easily adapted for various clinical applications

1 Introduction

Carbapenems are a class of 120573-lactam antibiotics with a broadspectrum and served as the last line against ESBLs efficientlyand stably [1] With wide and heavy use of various antibi-otics carbapenem-resistant isolates have become worldwidepublic-health issue with a widespread distribution broadrange of activities against 120573-lactams and increased patientmorbidity mortality and lengths of hospital stay [2] particu-larly among elderly patients infants and patients with severe

underlying disease In the clinic it is imperative to develop arapid simple and accurate test to detect and identify the clin-ical strains that produce carbapenemase and this is criticalfor the management and control of the increasing prevalenceof carbapenemase-producing strains worldwide [3]

The detection of carbapenemase producers in clinicalspecimens is based first on the analysis of susceptibilitytesting according to CLSI updated in 2017 [4] followed byconfirmatory genotypic tests with methods like PCR andmass spectrometry [3 5] However these tests are not the

HindawiBioMed Research InternationalVolume 2019 Article ID 8219748 13 pageshttpsdoiorg10115520198219748

2 BioMed Research International

best fit for detecting the carbapenemase-producing strainswith desirable specificity and sensitivity [6] Even the recom-mended CLSI methods are limited by their inherent disad-vantages the phenotypic tests for carbapenemase-producingstrains suffer problems of false positives and false negatives[5 7]Mass spectrometry cannot providemolecular epidemi-ological information and is cost-ineffective Addressing thesedeficiencies we herein propose a novel DNA microarray-based method for rapid sensitive and specific detectionof clinical carbapenamase-producing samples The proposedassay may have important implications in the diagnosisphenotypic confirmation and the molecular epidemiologystudies of carbapenemase-producing bacteria

The Ambler molecular classification system based onprotein homology categorizes 120573-lactamases into four classes(A to D) The most extensive distribution class A enzymewith carbapenemase activity is KPC The class B enzymesare metalloenzymes of the most IMP or VIM series BesidesNDM and DIM belong to class B enzymes which are widelyconcerned in Asia [8] Many OXA enzymes (OXA-23-likeOXA-48-like class D) are considered to be responsible forthe worldwide resistance epidemics as well as their detailedproperties which have been extensively reported [9 10] Thepresence of 119887119897119886DIM-1 has been considered lower once beforefor a Pseudomonas stutzeri isolate from the Netherlands [11]However in a recent report 119887119897119886DIM-1 was found in hospitalisolates belonging to the families Enterobacteriaceae Pseu-domonadaceae Burkholderiaceae and Comamonadaceaeand forty percent of the isolates were found to contain119887119897119886DIM-1 among the tested isolates [12] The 119887119897119886OXA-51 and119887119897119886OXA-23 genes are mainly responsible for the resistancein Acinetobacter baumannii [13] whereas the 119887119897119886OXA-23-likegenes are the most prevalent carbapenem-resistant genesidentified in China [13 14] The 119887119897119886OXA-51-like genes have beenconsidered exclusively chromosomally encoded intrinsicoxacillinase genes of Acinetobacter baumannii and are usedby many investigators for species identification and straintyping [15] However a number of recent reports indicate thatthe 119887119897119886OXA-51-like genes have beenmobilized and are spreadingto other Acinetobacter spp by conjugative plasmids [16 17]While further work is required to determine carbapenemresistance of OXA-51-like group the enzymes of this groupremain amajor concern as they present the possibility that allA baumannii isolates may be capable of becoming resistantto the carbapenems [18] Therefore in the current studywe included 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51119887119897119886IMP 119887119897119886VIM and 119887119897119886DIM in the detection panel for the DNAmicroarray assays

DNA microarrays have had wide applications includinggene expression analysis [19] disease diagnosis [20] andpathogenic microorganism detection [21] This technique ischaracterized by miniaturization high-throughput manage-ability and easiness of automatization The purpose of thiswork is to develop a rapid reliable and high-throughputDNA microarray method for detection of clinically relevantcarbapenemase-encoding genes In this assay a reliable andportable ultrasensitive chemiluminescence (CL) imagingDNA hybridization was developed to simultaneously detecteight genes Plasmids were established as positive or limit

of detection (LOD) reference materials The specificity andsensitivity of the method was validated in 416 actual samples

2 Materials and Methods21 Ethics Statement All patients provided informed con-sent in accordance with requirements of the Declarationof Helsinki and the research project was approved by theEthical Committee of Chinese Peoplersquos Liberation Army(PLA) General Hospital

22 Specimen Collection and Processing The samples col-lected from PLA General Hospital of China were sputumurine and bacteria isolates Sputum sample was liquefied by4 NaOH for 30 min at room temperature with shakingand then centrifuged at 13000 rpm for 2 min Pellets werecollected and washed The DNA in sputum pellets and urinesamples were isolated by QIAamp DNA Mini Kit accordingto themanufacturers instructions Bacteria isolates were pre-pared by modified boiling method Briefly colonies of eachisolate were picked from Luria-Bertani plates and suspendedin 100 120583L of sterilized H2O followed by boiling at 95

∘C for 15min and centrifugation at 12000 rpm for 5min Supernatantswere harvested and transferred to new tube and served astemplates for PCR and microarray assays [13]

23 Primer and Probe Design The DNA sequences of thecarbapenemase (ie 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM and 119887119897119886DIM) were downloaded fromGenBank (httpwwwncbinlmnihgovgenomes) Primersand probes were designed by DNAMAN and Primer Pre-mier Specific primers were designed for 119887119897119886NDM-1 119887119897119886OXA-48and 119887119897119886DIM-1 For 119887119897119886KPC 119887119897119886IMP 119887119897119886VIM 119887119897119886OXA-23-like and119887119897119886OXA-51-like primers were chosen in the conserved upstreamor downstream regions Microarray probes ranging from 33to 42 nucleotides were synthesized for these genes Thesegenes were amplified by multiplex PCR in separate tubesand an internal standard probe was included for each tubefor process monitoring Finally ten primers and seventeenprobes with favorable specificities were selected (Tables 1and 2) All the primers and probes were verified by BLAST(httpblastncbinlmnihgov)

24 Construction of Reference Plasmids Oligonucleotidesof 119887119897119886DIM and 119887119897119886VIM in this study were spliced by bigprimer amplification method in which 40 bp fragmentswere concatenated into final sequences of about 280 bpCarbapenemase-producing samples of OXA-23 OXA-48and OXA-51 were collected from Chinese PLA GeneralHospital in Beijing and samples of KPC IMP and NDM-1 were collected from Chinese PLA Academy of MilitaryMedical Sciences These eight well-characterized referencestrains each carrying 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM or 119887119897119886DIM were used for the designand validation of the microarray probes and primers DNAfragments of KPC NDM-1 OXA-23 OXA-48 OXA-51 IMPVIM DIM mitochondrial DNA (mtDNA) and 16S rRNAwere amplified by PCR followed by digestion with PGM-T and cloning into DH5120572 The cloned fragments were con-firmed by sequencing the entire regionsMitochondrial DNA

BioMed Research International 3

Table 1 The primer sequences for microarray

Primer a Sequence (51015840-31015840) Positions b Targeted geneGenBank accession

KPC-F CTGGGCAGTCGGAGACAAAA 681-700 KPC KX6196221KPC-R AGACGGCCAACACAATAGGT 765-784NDM1-F GAATGTCTGGCAGCACACT 168-186 NDM-1 KX249707NDM1-R TGGCATAAGTCGCAATCC 407-424OXA23-F GCAGTCCCAGTCTATCAGGA 379-398 OXA-23NG 049726OXA23-R CCCAACCAGTCTTTCCAA 641-658OXA48-F TCGGGCAATGTAGACAGTT 548-556 OXA-48 NG 049762OXA48-R CACCAGCCAATCTTAGGTTC 746-765OXA51-F GCTCGTCGTATTGGACTTGA 406-425 OXA-51 KX609247OXA51-R TGTGCCTCTTGCTGAGG 523-539IMP-F GTAATTGACACTCCATTTAC 291-309 IMP NG 049172IMP-R GCGGACTTTGGCCAAGCTTC 674-693VIM-F TGGTGAGTATCCGACAG 190-206 VIMNG 050336VIM-R ATGAAAGTGCGTGGAG 433-448DIM-F GCTTGTCTTCGCTTGCTAA 38-56 DIMNG 049077DIM-R ATTCCTGCGGTTCTATCCT 293-311mtDNA-F GTCGAAGGTGGATTTAGCAGTAA 1413-1435 mtDNAMG182040mtDNA-R GTAAGGTGGAGTGGGTTTGGG 1684-1704182-F AGAGTTTGATCMTGGCTCAG 1-20 16S rRNA LN612729756-R CGTATTACCGCGGCTGCTG 518-530a F forward primer R reverse primer all reverse primers have biotin conjugated at 51015840-endsb Positions refer to the nucleotide numberings of the corresponding GenBank genes

Table 2 The probe sequences for microarray

Probe Sequences (51015840-31015840) a Targeted geneKPC-P CAAATGACTATGCCGTCGTCTGGCC KPCNDM1-P1 ACCGATGACCAGACCGCCCAGATCCTCAAC NDM-1NDM1-P2 TCAGGACAAGATGGGCGGTATGGAC NDM-1OXA23-P1 TTTTAGAAGAGAGTAATGGCTACAAAA OXA-23OXA23-P2 ATTGGACAGCAGGTTGATAATTTCTGG OXA-23OXA48-P1 CGAATTTCGGCCACGGAGCAAATCAGCTT OXA-48OXA48-P2 CAGCGTATTGTCAAACAAGCCATGC OXA-48OXA51-P1 GAAGTGAAGCGTGTTGGTTATG OXA-51OXA51-P2 ATATCGGTACCCAAGTCGATAATTTTTGGC OXA-51IMP-P1 GGCTAGTTAAAAATAAAATTGAAG IMPIMP-P2 CCCACGTATGCRTCTGAATTAAC IMPVIM-P1 TGGTGTTTGGTCGCATATCGCAACG VIMVIM-P2 CTCATTGTCCGTGATGGTGATGAG VIMDIM-P1 GTCAGTTCAAACGGCCTTGTTGTCATAGATT DIMDIM-P2 CTTGGTCAGACCGAGATACAGAAACGCTCG DIMmtDNA-P ATGTCCTTTGAAGTATACTTGAGGAGTT mitochondria551-P ACTCCTACGGGAGGCAGCAGTT 16S rRNAQuality control b TTTTTTTTTTTTTTTTTTTT Oligo dT

20

a An oligonucleotide of 12 Trsquos with an amino-labeled 31015840-end was conjugated to the 31015840-ends of all probes b An oligonucleotide of 20 Trsquos with an amino-labeled31015840-end biotin-labeled 51015840-end was used as microarray quality control

4 BioMed Research International

① KPC-P② NDM1-P1③ NDM1-P2④ OXA23-P1⑤ OXA23-P2⑥OXA48-P1⑦OXA48-P2⑧Negative control⑨ OXA51-P1⑩ OXA51-P2

⑪IMP-P1⑫IMP-P2⑬VIM-P1⑭VIM-P2⑮DIM-P1⑯DIM-P2⑰ Internal control probe of tube A⑱ Internal control probe of tube B⑲Quality control probe

⑲⑲⑲⑲⑲⑲⑲⑲

①②③④⑤⑥⑦⑧

①②③④⑤⑥⑦⑧

①②③④⑤⑥⑦⑧

⑨⑩⑪⑫⑬⑭⑮⑯

⑨⑩⑪⑫⑬⑭⑮⑯

⑨⑩⑪⑫⑬⑭⑮⑯

⑰⑰⑰⑰⑱⑱⑱⑱

(a)

Probe

C-H

o

N-H

H

C-H

o

N-H

H

C-H

N

C-H

N

Biotin

A B

Streptavidin

Horseradish Peroxidase

C-H

N

C-H

N

C-H

N

C-H

N

Luminol

C-H

N

C-H

N

CL

Biotin

C D E

（22

(b)

Figure 1 (a) The microarray layout of the target carbapenemase genes that are detected in this study (b) Principle of hybridization andCL imaging of microarray (A) Captured probes were fixed to the aldehyde-chip surface (B) denatured PCR products were hybridized withcapture probes (C) horseradish peroxidase modified streptavidin was bond to biotin incorporated in hybridization and (D) adding HRPsubstrates luminal and H

2O2 and (E) CL signal was detected by catalyzed substrates

and 16S rRNA were collected from cultured Hela cell andstandard strains of Escherichia coli ATCC25922 respectively

25 Microarray Fabrication All microarray probes were syn-thesized by the Chinese PLA Academy of Military MedicalSciences An oligo(dT)12-with an amino-labeled 31015840-end wasconjugated to the 31015840-end of all the probes such that it servedas a linker arm to be immobilized on the aldehyde modifiedglass surface (Baio Technology Shanghai Co Ltd ShanghaiChina) An Oligo(dT)20 with an amino-labeled 31015840-end anda biotin-labeled 51015840-end served as a quality control (QC)probe Each probe (50 120583Mof final concentration) was spottedthrice repeatedly by using a noncontact inkjet Nano-plotter21 (GeSim Dresden Germany) onto an aldehyde-chip after

mixing with printing buffer [5 glycerol 01 sodiumdodecyl sulfate (SDS) 6times saline-sodium citrate buffer (SSC)and 2 (wtvol) Ficoll 400] Quality control (QC) probe wasincluded tomanage the standard of operation and used at 125120583M final concentration It was spotted eight times repeatedlyin the horizontal direction to calibrate the CL signal valuesEach aldehyde slide was divided into 10 blocks (11times 11mm) bya waterproof film to detect 10 different samples Microarrayswere placed in a dryer for 24 h at room temperatureUnboundprobes were washed off by 02 SDS and distilled water priorto use The layout is shown in Figure 1(a)

26 Multiplex PCR All reverse primers for target genes andinternal controls (ie mtDNA and 16S rRNA) were labeled

BioMed Research International 5

DIMKPC NDM-1 OXA-23 VIMIMPOXA-48 OXA-51

Figure 2 Representative microarray hybridization images with the reference carbapenemase plasmids (positive controls)

by biotin at the 51015840-ends for acting a CL reaction MultiplexPCRs were performed for 119887119897119886KPC 119887119897119886OXA-23 119887119897119886VIM 119887119897119886DIMand 16S rRNA in one tube (Tube A) and 119887119897119886IMP 119887119897119886NDM-1119887119897119886OXA-48 119887119897119886OXA-51 and mtDNA in another (Tube B) PCRreaction was in 30 120583L and contained 15 120583L of 2times MultiplexPCR Mix (cwBiotech Beijing China) and 3 120583L of mixedDNA templates For Tube A the concentrations of 119887119897119886OXA-23119887119897119886VIM 119887119897119886DIM and 16S rRNA forward and reverse primersall were 01 120583M and 05 120583M and for 119887119897119886KPC the forward andreverse primer were 02 120583M and 1 120583M respectively For TubeB the concentrations of the forward and reverse primers were01 120583M and 05 120583M respectively PCR was performed on aThermalCycler PCR system (AppliedBiosystems FosterCityCA) using the following conditions 10 min at 95∘C 37 cyclesof 30 s at 94∘C 30 s at 55∘C and 45 s at 72∘C and a finalextension of 5 min at 72∘C

27Hybridization and SignalDetection PCRproducts ampli-fied by the two multiplex PCR reactions using the sametemplate were mixed The mixtures were denatured at 95∘Cfor 5min and placed on ice immediately for 5 min Then 5120583L of denatured PCR mixtures were blended with 5 120583L ofhybridization buffer [8times SSC 06 SDS 10 formaldehydeand 10timesDenhardt] Hybridization reaction was proceeded ina hybrid-box by incubation for 1 h at 45∘C After hybridiza-tion slide was washed successively in 1times SSC and 02 SDS02times SSC and 01times SSC for 30 s Slide was air dried atroom temperature To detect the CL signals microarray wasincubated in 37∘C for 30 min with 10 120583L of streptavidinhorseradish peroxidase (Str-HRP Sigma-Aldrich StLouisUSA) followed bywashwith PBST (1timesPBS 005Tween-20)for 30 s at room temperature Dried microarray was coveredin 10 120583L of premixed CL HRP substrate luminal solution andH2O2 (Millipore Corporation Boston USA) (Figure 1(b))Then immediately subject to scanning by Biochip Chemi-luminescence Imager a microlight level imaging systemdeveloped in our laboratory Signal intensitieswere calculatedby Array Vision 70

28 Identification of Carbapenem-Resistant and SusceptibleStrains The strains isolated from samples were tested usingthe Kirby-Bauer (K-B) method of disk diffusion accordingto the recommendations of the CLSI to determine theirsusceptibilities to imipenem (10120583g) and meropenem (10120583g)Escherichia coli ATCC25922 and Pseudomonas aeruginosaATCC27853 were used as control strains for susceptibilitytesting Isolates were considered to have a carbapenemasephenotype if they were resistant to at least one carbapenem(ie meropenem or imipenem) [22]

Four hundred sixteen (416) antibiotic-resistant bacterialsamples were evaluated in this studyThe initial samples wereobtained from patients who had been hospitalized for a longtime (gt1 year)

29 Confirmation of the Resistance Genes by SequencingThe resistance genes including 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM and 119887119897119886DIM that hadbeen detected in antibiotic-resistant samples by microarrayhybridization were validated by Sanger sequencing

3 Results

31 Determination of reshold Signal Intensity To deter-mine the threshold value for differentiating positive andnegative microarray signal intensities we have performedpilot microarray hybridization experiments using the Gram-positive bacterial strain S aureus 04018 as negative controlsand carbapenemase plasmids of 3times103 copies120583L as positivecontrols under the conditions specified in Materials andmethods If the signal intensity value is 10 times of thebackground intensity value the probe was considered to bepositive

32 Specificity and Sensitivity of Microarray Test To evaluatethe specificity of the microarray method we performedmicroarray hybridization assays for the reference carbapen-emase plasmids (ie positive controls) (Figure 2) clinicalcarbapenem-resistant samples (Figure 3) and ten negativecontrols that were fromATCC standard strains and were sen-sitive to carbapenem (Figure 4) As shown by the microarrayimages our method could effectively distinguish between thecarbapenem resistance and carbapenem-sensitive genotypesamong the clinical bacterial specimens with high specificity

To evaluate the sensitivity of the microarray assay wediluted the reference carbapenemase plasmids into vari-ous concentrations (ie from 3times101 copies120583L to 3times105copies120583L) The different copy numbers of DNA werehybridized to the microarrays The detection images wereshown in Figure 5(a) In general the microarrays yieldedsatisfactory sensitivity For most reference plasmids thedetection limit was as low as 30 copies120583L (Figure 5(a))Diagnostic Kit for Bacterial Resistance Gene KPC (PCR-Fluorescence Probing) andDiagnostic Kit for Bacterial Resis-tance Gene NDM-1 (PCR-Fluorescence Probing) (PuruikangBiotech Shenzhen China) were also used to detect KPCand NDM-1 reference plasmids (3times101 to 3times105 copies120583L)respectively Real-time PCR amplified by ABI Prism 7500real-time PCR apparatus (Applied Biosystems Foster City

6 BioMed Research International

lowast1-8

9-16

17-24

25-32

Figure 3 Representative images of microarray assays for clinical carbapenemase-producing specimens lowast the number of clinicalcarbapenemase-producing specimens (Supplementary Table)

33-37

38-42

Figure 4 Images of microarray assays for carbapenemase-sensitive strains (negative control strains) 33 Escherichia coli (ATCC 25922)34 Enterococcus faecium (ATCC 35667) 35 Enterobacter cloacae (ATCC 13047) 36 Enterococcus faecalis (ATCC 29212) 37 Pseudomonasaeruginosa (ATCC27853) 38Acinetobacter baumannii (ATCC 19606) 39 Staphylococcus aureus (ATCC25923) 40 Streptococcus pneumoniae(ATCC 49619) 41 Klebsiella pneumonia (ATCC700603) 42 Streptococcus pneumonia (ATCC 9007(serotype C)) ATCC American TypeCulture Collection

US) and the results of sensitivity comparison betweenmicroarray assay andReal-time PCR showed they had similarsensitivities (Figures 5(b) and 5(c)) indicating that our DNAmicroarrays could be applied to the clinical detection of thecarbapenemase-producing samples

33 Stability of Microarray Assay Diluted carbapenemaseplasmids (3times105 copies120583L) and negative controls S aureus04018 were used to evaluate how the microarray assays per-formed as far as repeatability is concernedThe 3times103copy120583L

plasmid of each target gene was selected as the template todetect for the determination of interchip and intrachip varia-tion and a negative control was set up without template Theexperiment was repeated three times and the repeatability ofinterchip and intrachip variation were evaluated Coefficientof variation (CV) = SDmean of signal intensitiestimes100

Statistical analysis showed that for all target gene probesthe intrachip and interchipCVvalues of 8 probes ranged from358 to 1102 (below 15) suggesting a favorable repeata-bility of the DNA microarray detection method (Table 3)

BioMed Research International 7

Negativecontrol

KPC

NDM-1

OXA-23

OXA-48

OXA-51

IMP

VIM

DIM

copy numbers

13 times 105

3 times 104

3 times 103

3 times 102

3 times 101

(a)

KPC

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

A B C D E F G H

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(b)

NDM-1

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(c)

Figure 5 (a) Sensitivity of the microarray assays The indicated carbapenemase reference plasmids were subject to series dilutions andhybridized to the DNA microarrays The plasmids copy numbers are diluted from 3times101 to 3times105 copies120583L (b) KPC reference plasmids(3times101 to 3times105 copies120583L) were detected by real-time PCR (c) NDM-1 reference plasmids (3times101 to 3times105 copies120583L) were detected byreal-time PCR

34 Phenotypic Resistance About 78 clinical specimens(326416) were resistant to both imipenem and meropenem77 samples were susceptible to both imipenem andmeropenem three samples were resistant to imipenembut sensitive to meropenem and five samples were reverselysensitive to imipenem but resistant to meropenem as

well as five samples were intermediary to imipenem butsensitive to meropenemMost carbapenem-resistant sampleswere nonsusceptible to diverse antibiotics containingcephalosporins fluoroquinolone aminoglycosides etc Therates of nonsusceptibility to different antimicrobial agentswere commonly gt90 Clinical information and phenotypic

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

4 BioMed Research International

① KPC-P② NDM1-P1③ NDM1-P2④ OXA23-P1⑤ OXA23-P2⑥OXA48-P1⑦OXA48-P2⑧Negative control⑨ OXA51-P1⑩ OXA51-P2

⑪IMP-P1⑫IMP-P2⑬VIM-P1⑭VIM-P2⑮DIM-P1⑯DIM-P2⑰ Internal control probe of tube A⑱ Internal control probe of tube B⑲Quality control probe

⑲⑲⑲⑲⑲⑲⑲⑲

①②③④⑤⑥⑦⑧

①②③④⑤⑥⑦⑧

①②③④⑤⑥⑦⑧

⑨⑩⑪⑫⑬⑭⑮⑯

⑨⑩⑪⑫⑬⑭⑮⑯

⑨⑩⑪⑫⑬⑭⑮⑯

⑰⑰⑰⑰⑱⑱⑱⑱

(a)

Probe

C-H

o

N-H

H

C-H

o

N-H

H

C-H

N

C-H

N

Biotin

A B

Streptavidin

Horseradish Peroxidase

C-H

N

C-H

N

C-H

N

C-H

N

Luminol

C-H

N

C-H

N

CL

Biotin

C D E

（22

(b)

Figure 1 (a) The microarray layout of the target carbapenemase genes that are detected in this study (b) Principle of hybridization andCL imaging of microarray (A) Captured probes were fixed to the aldehyde-chip surface (B) denatured PCR products were hybridized withcapture probes (C) horseradish peroxidase modified streptavidin was bond to biotin incorporated in hybridization and (D) adding HRPsubstrates luminal and H

2O2 and (E) CL signal was detected by catalyzed substrates

and 16S rRNA were collected from cultured Hela cell andstandard strains of Escherichia coli ATCC25922 respectively

25 Microarray Fabrication All microarray probes were syn-thesized by the Chinese PLA Academy of Military MedicalSciences An oligo(dT)12-with an amino-labeled 31015840-end wasconjugated to the 31015840-end of all the probes such that it servedas a linker arm to be immobilized on the aldehyde modifiedglass surface (Baio Technology Shanghai Co Ltd ShanghaiChina) An Oligo(dT)20 with an amino-labeled 31015840-end anda biotin-labeled 51015840-end served as a quality control (QC)probe Each probe (50 120583Mof final concentration) was spottedthrice repeatedly by using a noncontact inkjet Nano-plotter21 (GeSim Dresden Germany) onto an aldehyde-chip after

mixing with printing buffer [5 glycerol 01 sodiumdodecyl sulfate (SDS) 6times saline-sodium citrate buffer (SSC)and 2 (wtvol) Ficoll 400] Quality control (QC) probe wasincluded tomanage the standard of operation and used at 125120583M final concentration It was spotted eight times repeatedlyin the horizontal direction to calibrate the CL signal valuesEach aldehyde slide was divided into 10 blocks (11times 11mm) bya waterproof film to detect 10 different samples Microarrayswere placed in a dryer for 24 h at room temperatureUnboundprobes were washed off by 02 SDS and distilled water priorto use The layout is shown in Figure 1(a)

26 Multiplex PCR All reverse primers for target genes andinternal controls (ie mtDNA and 16S rRNA) were labeled

BioMed Research International 5

DIMKPC NDM-1 OXA-23 VIMIMPOXA-48 OXA-51

Figure 2 Representative microarray hybridization images with the reference carbapenemase plasmids (positive controls)

by biotin at the 51015840-ends for acting a CL reaction MultiplexPCRs were performed for 119887119897119886KPC 119887119897119886OXA-23 119887119897119886VIM 119887119897119886DIMand 16S rRNA in one tube (Tube A) and 119887119897119886IMP 119887119897119886NDM-1119887119897119886OXA-48 119887119897119886OXA-51 and mtDNA in another (Tube B) PCRreaction was in 30 120583L and contained 15 120583L of 2times MultiplexPCR Mix (cwBiotech Beijing China) and 3 120583L of mixedDNA templates For Tube A the concentrations of 119887119897119886OXA-23119887119897119886VIM 119887119897119886DIM and 16S rRNA forward and reverse primersall were 01 120583M and 05 120583M and for 119887119897119886KPC the forward andreverse primer were 02 120583M and 1 120583M respectively For TubeB the concentrations of the forward and reverse primers were01 120583M and 05 120583M respectively PCR was performed on aThermalCycler PCR system (AppliedBiosystems FosterCityCA) using the following conditions 10 min at 95∘C 37 cyclesof 30 s at 94∘C 30 s at 55∘C and 45 s at 72∘C and a finalextension of 5 min at 72∘C

27Hybridization and SignalDetection PCRproducts ampli-fied by the two multiplex PCR reactions using the sametemplate were mixed The mixtures were denatured at 95∘Cfor 5min and placed on ice immediately for 5 min Then 5120583L of denatured PCR mixtures were blended with 5 120583L ofhybridization buffer [8times SSC 06 SDS 10 formaldehydeand 10timesDenhardt] Hybridization reaction was proceeded ina hybrid-box by incubation for 1 h at 45∘C After hybridiza-tion slide was washed successively in 1times SSC and 02 SDS02times SSC and 01times SSC for 30 s Slide was air dried atroom temperature To detect the CL signals microarray wasincubated in 37∘C for 30 min with 10 120583L of streptavidinhorseradish peroxidase (Str-HRP Sigma-Aldrich StLouisUSA) followed bywashwith PBST (1timesPBS 005Tween-20)for 30 s at room temperature Dried microarray was coveredin 10 120583L of premixed CL HRP substrate luminal solution andH2O2 (Millipore Corporation Boston USA) (Figure 1(b))Then immediately subject to scanning by Biochip Chemi-luminescence Imager a microlight level imaging systemdeveloped in our laboratory Signal intensitieswere calculatedby Array Vision 70

28 Identification of Carbapenem-Resistant and SusceptibleStrains The strains isolated from samples were tested usingthe Kirby-Bauer (K-B) method of disk diffusion accordingto the recommendations of the CLSI to determine theirsusceptibilities to imipenem (10120583g) and meropenem (10120583g)Escherichia coli ATCC25922 and Pseudomonas aeruginosaATCC27853 were used as control strains for susceptibilitytesting Isolates were considered to have a carbapenemasephenotype if they were resistant to at least one carbapenem(ie meropenem or imipenem) [22]

Four hundred sixteen (416) antibiotic-resistant bacterialsamples were evaluated in this studyThe initial samples wereobtained from patients who had been hospitalized for a longtime (gt1 year)

29 Confirmation of the Resistance Genes by SequencingThe resistance genes including 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM and 119887119897119886DIM that hadbeen detected in antibiotic-resistant samples by microarrayhybridization were validated by Sanger sequencing

3 Results

31 Determination of reshold Signal Intensity To deter-mine the threshold value for differentiating positive andnegative microarray signal intensities we have performedpilot microarray hybridization experiments using the Gram-positive bacterial strain S aureus 04018 as negative controlsand carbapenemase plasmids of 3times103 copies120583L as positivecontrols under the conditions specified in Materials andmethods If the signal intensity value is 10 times of thebackground intensity value the probe was considered to bepositive

32 Specificity and Sensitivity of Microarray Test To evaluatethe specificity of the microarray method we performedmicroarray hybridization assays for the reference carbapen-emase plasmids (ie positive controls) (Figure 2) clinicalcarbapenem-resistant samples (Figure 3) and ten negativecontrols that were fromATCC standard strains and were sen-sitive to carbapenem (Figure 4) As shown by the microarrayimages our method could effectively distinguish between thecarbapenem resistance and carbapenem-sensitive genotypesamong the clinical bacterial specimens with high specificity

To evaluate the sensitivity of the microarray assay wediluted the reference carbapenemase plasmids into vari-ous concentrations (ie from 3times101 copies120583L to 3times105copies120583L) The different copy numbers of DNA werehybridized to the microarrays The detection images wereshown in Figure 5(a) In general the microarrays yieldedsatisfactory sensitivity For most reference plasmids thedetection limit was as low as 30 copies120583L (Figure 5(a))Diagnostic Kit for Bacterial Resistance Gene KPC (PCR-Fluorescence Probing) andDiagnostic Kit for Bacterial Resis-tance Gene NDM-1 (PCR-Fluorescence Probing) (PuruikangBiotech Shenzhen China) were also used to detect KPCand NDM-1 reference plasmids (3times101 to 3times105 copies120583L)respectively Real-time PCR amplified by ABI Prism 7500real-time PCR apparatus (Applied Biosystems Foster City

6 BioMed Research International

lowast1-8

9-16

17-24

25-32

Figure 3 Representative images of microarray assays for clinical carbapenemase-producing specimens lowast the number of clinicalcarbapenemase-producing specimens (Supplementary Table)

33-37

38-42

Figure 4 Images of microarray assays for carbapenemase-sensitive strains (negative control strains) 33 Escherichia coli (ATCC 25922)34 Enterococcus faecium (ATCC 35667) 35 Enterobacter cloacae (ATCC 13047) 36 Enterococcus faecalis (ATCC 29212) 37 Pseudomonasaeruginosa (ATCC27853) 38Acinetobacter baumannii (ATCC 19606) 39 Staphylococcus aureus (ATCC25923) 40 Streptococcus pneumoniae(ATCC 49619) 41 Klebsiella pneumonia (ATCC700603) 42 Streptococcus pneumonia (ATCC 9007(serotype C)) ATCC American TypeCulture Collection

US) and the results of sensitivity comparison betweenmicroarray assay andReal-time PCR showed they had similarsensitivities (Figures 5(b) and 5(c)) indicating that our DNAmicroarrays could be applied to the clinical detection of thecarbapenemase-producing samples

33 Stability of Microarray Assay Diluted carbapenemaseplasmids (3times105 copies120583L) and negative controls S aureus04018 were used to evaluate how the microarray assays per-formed as far as repeatability is concernedThe 3times103copy120583L

plasmid of each target gene was selected as the template todetect for the determination of interchip and intrachip varia-tion and a negative control was set up without template Theexperiment was repeated three times and the repeatability ofinterchip and intrachip variation were evaluated Coefficientof variation (CV) = SDmean of signal intensitiestimes100

Statistical analysis showed that for all target gene probesthe intrachip and interchipCVvalues of 8 probes ranged from358 to 1102 (below 15) suggesting a favorable repeata-bility of the DNA microarray detection method (Table 3)

BioMed Research International 7

Negativecontrol

KPC

NDM-1

OXA-23

OXA-48

OXA-51

IMP

VIM

DIM

copy numbers

13 times 105

3 times 104

3 times 103

3 times 102

3 times 101

(a)

KPC

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

A B C D E F G H

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(b)

NDM-1

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(c)

Figure 5 (a) Sensitivity of the microarray assays The indicated carbapenemase reference plasmids were subject to series dilutions andhybridized to the DNA microarrays The plasmids copy numbers are diluted from 3times101 to 3times105 copies120583L (b) KPC reference plasmids(3times101 to 3times105 copies120583L) were detected by real-time PCR (c) NDM-1 reference plasmids (3times101 to 3times105 copies120583L) were detected byreal-time PCR

34 Phenotypic Resistance About 78 clinical specimens(326416) were resistant to both imipenem and meropenem77 samples were susceptible to both imipenem andmeropenem three samples were resistant to imipenembut sensitive to meropenem and five samples were reverselysensitive to imipenem but resistant to meropenem as

well as five samples were intermediary to imipenem butsensitive to meropenemMost carbapenem-resistant sampleswere nonsusceptible to diverse antibiotics containingcephalosporins fluoroquinolone aminoglycosides etc Therates of nonsusceptibility to different antimicrobial agentswere commonly gt90 Clinical information and phenotypic

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

BioMed Research International 5

DIMKPC NDM-1 OXA-23 VIMIMPOXA-48 OXA-51

Figure 2 Representative microarray hybridization images with the reference carbapenemase plasmids (positive controls)

by biotin at the 51015840-ends for acting a CL reaction MultiplexPCRs were performed for 119887119897119886KPC 119887119897119886OXA-23 119887119897119886VIM 119887119897119886DIMand 16S rRNA in one tube (Tube A) and 119887119897119886IMP 119887119897119886NDM-1119887119897119886OXA-48 119887119897119886OXA-51 and mtDNA in another (Tube B) PCRreaction was in 30 120583L and contained 15 120583L of 2times MultiplexPCR Mix (cwBiotech Beijing China) and 3 120583L of mixedDNA templates For Tube A the concentrations of 119887119897119886OXA-23119887119897119886VIM 119887119897119886DIM and 16S rRNA forward and reverse primersall were 01 120583M and 05 120583M and for 119887119897119886KPC the forward andreverse primer were 02 120583M and 1 120583M respectively For TubeB the concentrations of the forward and reverse primers were01 120583M and 05 120583M respectively PCR was performed on aThermalCycler PCR system (AppliedBiosystems FosterCityCA) using the following conditions 10 min at 95∘C 37 cyclesof 30 s at 94∘C 30 s at 55∘C and 45 s at 72∘C and a finalextension of 5 min at 72∘C

27Hybridization and SignalDetection PCRproducts ampli-fied by the two multiplex PCR reactions using the sametemplate were mixed The mixtures were denatured at 95∘Cfor 5min and placed on ice immediately for 5 min Then 5120583L of denatured PCR mixtures were blended with 5 120583L ofhybridization buffer [8times SSC 06 SDS 10 formaldehydeand 10timesDenhardt] Hybridization reaction was proceeded ina hybrid-box by incubation for 1 h at 45∘C After hybridiza-tion slide was washed successively in 1times SSC and 02 SDS02times SSC and 01times SSC for 30 s Slide was air dried atroom temperature To detect the CL signals microarray wasincubated in 37∘C for 30 min with 10 120583L of streptavidinhorseradish peroxidase (Str-HRP Sigma-Aldrich StLouisUSA) followed bywashwith PBST (1timesPBS 005Tween-20)for 30 s at room temperature Dried microarray was coveredin 10 120583L of premixed CL HRP substrate luminal solution andH2O2 (Millipore Corporation Boston USA) (Figure 1(b))Then immediately subject to scanning by Biochip Chemi-luminescence Imager a microlight level imaging systemdeveloped in our laboratory Signal intensitieswere calculatedby Array Vision 70

28 Identification of Carbapenem-Resistant and SusceptibleStrains The strains isolated from samples were tested usingthe Kirby-Bauer (K-B) method of disk diffusion accordingto the recommendations of the CLSI to determine theirsusceptibilities to imipenem (10120583g) and meropenem (10120583g)Escherichia coli ATCC25922 and Pseudomonas aeruginosaATCC27853 were used as control strains for susceptibilitytesting Isolates were considered to have a carbapenemasephenotype if they were resistant to at least one carbapenem(ie meropenem or imipenem) [22]

Four hundred sixteen (416) antibiotic-resistant bacterialsamples were evaluated in this studyThe initial samples wereobtained from patients who had been hospitalized for a longtime (gt1 year)

29 Confirmation of the Resistance Genes by SequencingThe resistance genes including 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIM and 119887119897119886DIM that hadbeen detected in antibiotic-resistant samples by microarrayhybridization were validated by Sanger sequencing

3 Results

31 Determination of reshold Signal Intensity To deter-mine the threshold value for differentiating positive andnegative microarray signal intensities we have performedpilot microarray hybridization experiments using the Gram-positive bacterial strain S aureus 04018 as negative controlsand carbapenemase plasmids of 3times103 copies120583L as positivecontrols under the conditions specified in Materials andmethods If the signal intensity value is 10 times of thebackground intensity value the probe was considered to bepositive

32 Specificity and Sensitivity of Microarray Test To evaluatethe specificity of the microarray method we performedmicroarray hybridization assays for the reference carbapen-emase plasmids (ie positive controls) (Figure 2) clinicalcarbapenem-resistant samples (Figure 3) and ten negativecontrols that were fromATCC standard strains and were sen-sitive to carbapenem (Figure 4) As shown by the microarrayimages our method could effectively distinguish between thecarbapenem resistance and carbapenem-sensitive genotypesamong the clinical bacterial specimens with high specificity

To evaluate the sensitivity of the microarray assay wediluted the reference carbapenemase plasmids into vari-ous concentrations (ie from 3times101 copies120583L to 3times105copies120583L) The different copy numbers of DNA werehybridized to the microarrays The detection images wereshown in Figure 5(a) In general the microarrays yieldedsatisfactory sensitivity For most reference plasmids thedetection limit was as low as 30 copies120583L (Figure 5(a))Diagnostic Kit for Bacterial Resistance Gene KPC (PCR-Fluorescence Probing) andDiagnostic Kit for Bacterial Resis-tance Gene NDM-1 (PCR-Fluorescence Probing) (PuruikangBiotech Shenzhen China) were also used to detect KPCand NDM-1 reference plasmids (3times101 to 3times105 copies120583L)respectively Real-time PCR amplified by ABI Prism 7500real-time PCR apparatus (Applied Biosystems Foster City

6 BioMed Research International

lowast1-8

9-16

17-24

25-32

Figure 3 Representative images of microarray assays for clinical carbapenemase-producing specimens lowast the number of clinicalcarbapenemase-producing specimens (Supplementary Table)

33-37

38-42

Figure 4 Images of microarray assays for carbapenemase-sensitive strains (negative control strains) 33 Escherichia coli (ATCC 25922)34 Enterococcus faecium (ATCC 35667) 35 Enterobacter cloacae (ATCC 13047) 36 Enterococcus faecalis (ATCC 29212) 37 Pseudomonasaeruginosa (ATCC27853) 38Acinetobacter baumannii (ATCC 19606) 39 Staphylococcus aureus (ATCC25923) 40 Streptococcus pneumoniae(ATCC 49619) 41 Klebsiella pneumonia (ATCC700603) 42 Streptococcus pneumonia (ATCC 9007(serotype C)) ATCC American TypeCulture Collection

US) and the results of sensitivity comparison betweenmicroarray assay andReal-time PCR showed they had similarsensitivities (Figures 5(b) and 5(c)) indicating that our DNAmicroarrays could be applied to the clinical detection of thecarbapenemase-producing samples

33 Stability of Microarray Assay Diluted carbapenemaseplasmids (3times105 copies120583L) and negative controls S aureus04018 were used to evaluate how the microarray assays per-formed as far as repeatability is concernedThe 3times103copy120583L

plasmid of each target gene was selected as the template todetect for the determination of interchip and intrachip varia-tion and a negative control was set up without template Theexperiment was repeated three times and the repeatability ofinterchip and intrachip variation were evaluated Coefficientof variation (CV) = SDmean of signal intensitiestimes100

Statistical analysis showed that for all target gene probesthe intrachip and interchipCVvalues of 8 probes ranged from358 to 1102 (below 15) suggesting a favorable repeata-bility of the DNA microarray detection method (Table 3)

BioMed Research International 7

Negativecontrol

KPC

NDM-1

OXA-23

OXA-48

OXA-51

IMP

VIM

DIM

copy numbers

13 times 105

3 times 104

3 times 103

3 times 102

3 times 101

(a)

KPC

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

A B C D E F G H

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(b)

NDM-1

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(c)

Figure 5 (a) Sensitivity of the microarray assays The indicated carbapenemase reference plasmids were subject to series dilutions andhybridized to the DNA microarrays The plasmids copy numbers are diluted from 3times101 to 3times105 copies120583L (b) KPC reference plasmids(3times101 to 3times105 copies120583L) were detected by real-time PCR (c) NDM-1 reference plasmids (3times101 to 3times105 copies120583L) were detected byreal-time PCR

34 Phenotypic Resistance About 78 clinical specimens(326416) were resistant to both imipenem and meropenem77 samples were susceptible to both imipenem andmeropenem three samples were resistant to imipenembut sensitive to meropenem and five samples were reverselysensitive to imipenem but resistant to meropenem as

well as five samples were intermediary to imipenem butsensitive to meropenemMost carbapenem-resistant sampleswere nonsusceptible to diverse antibiotics containingcephalosporins fluoroquinolone aminoglycosides etc Therates of nonsusceptibility to different antimicrobial agentswere commonly gt90 Clinical information and phenotypic

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

6 BioMed Research International

lowast1-8

9-16

17-24

25-32

Figure 3 Representative images of microarray assays for clinical carbapenemase-producing specimens lowast the number of clinicalcarbapenemase-producing specimens (Supplementary Table)

33-37

38-42

Figure 4 Images of microarray assays for carbapenemase-sensitive strains (negative control strains) 33 Escherichia coli (ATCC 25922)34 Enterococcus faecium (ATCC 35667) 35 Enterobacter cloacae (ATCC 13047) 36 Enterococcus faecalis (ATCC 29212) 37 Pseudomonasaeruginosa (ATCC27853) 38Acinetobacter baumannii (ATCC 19606) 39 Staphylococcus aureus (ATCC25923) 40 Streptococcus pneumoniae(ATCC 49619) 41 Klebsiella pneumonia (ATCC700603) 42 Streptococcus pneumonia (ATCC 9007(serotype C)) ATCC American TypeCulture Collection

US) and the results of sensitivity comparison betweenmicroarray assay andReal-time PCR showed they had similarsensitivities (Figures 5(b) and 5(c)) indicating that our DNAmicroarrays could be applied to the clinical detection of thecarbapenemase-producing samples

33 Stability of Microarray Assay Diluted carbapenemaseplasmids (3times105 copies120583L) and negative controls S aureus04018 were used to evaluate how the microarray assays per-formed as far as repeatability is concernedThe 3times103copy120583L

plasmid of each target gene was selected as the template todetect for the determination of interchip and intrachip varia-tion and a negative control was set up without template Theexperiment was repeated three times and the repeatability ofinterchip and intrachip variation were evaluated Coefficientof variation (CV) = SDmean of signal intensitiestimes100

Statistical analysis showed that for all target gene probesthe intrachip and interchipCVvalues of 8 probes ranged from358 to 1102 (below 15) suggesting a favorable repeata-bility of the DNA microarray detection method (Table 3)

BioMed Research International 7

Negativecontrol

KPC

NDM-1

OXA-23

OXA-48

OXA-51

IMP

VIM

DIM

copy numbers

13 times 105

3 times 104

3 times 103

3 times 102

3 times 101

(a)

KPC

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

A B C D E F G H

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(b)

NDM-1

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(c)

Figure 5 (a) Sensitivity of the microarray assays The indicated carbapenemase reference plasmids were subject to series dilutions andhybridized to the DNA microarrays The plasmids copy numbers are diluted from 3times101 to 3times105 copies120583L (b) KPC reference plasmids(3times101 to 3times105 copies120583L) were detected by real-time PCR (c) NDM-1 reference plasmids (3times101 to 3times105 copies120583L) were detected byreal-time PCR

34 Phenotypic Resistance About 78 clinical specimens(326416) were resistant to both imipenem and meropenem77 samples were susceptible to both imipenem andmeropenem three samples were resistant to imipenembut sensitive to meropenem and five samples were reverselysensitive to imipenem but resistant to meropenem as

well as five samples were intermediary to imipenem butsensitive to meropenemMost carbapenem-resistant sampleswere nonsusceptible to diverse antibiotics containingcephalosporins fluoroquinolone aminoglycosides etc Therates of nonsusceptibility to different antimicrobial agentswere commonly gt90 Clinical information and phenotypic

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

BioMed Research International 7

Negativecontrol

KPC

NDM-1

OXA-23

OXA-48

OXA-51

IMP

VIM

DIM

copy numbers

13 times 105

3 times 104

3 times 103

3 times 102

3 times 101

(a)

KPC

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

A B C D E F G H

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(b)

NDM-1

Cycle2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40

14000001300000120000011000001000000900000800000700000600000500000400000300000200000100000

0minus100000

Amplification Plot

ΔRn

3 times 105

3 times 104

3 times 103

3 times 102

(c)

Figure 5 (a) Sensitivity of the microarray assays The indicated carbapenemase reference plasmids were subject to series dilutions andhybridized to the DNA microarrays The plasmids copy numbers are diluted from 3times101 to 3times105 copies120583L (b) KPC reference plasmids(3times101 to 3times105 copies120583L) were detected by real-time PCR (c) NDM-1 reference plasmids (3times101 to 3times105 copies120583L) were detected byreal-time PCR

34 Phenotypic Resistance About 78 clinical specimens(326416) were resistant to both imipenem and meropenem77 samples were susceptible to both imipenem andmeropenem three samples were resistant to imipenembut sensitive to meropenem and five samples were reverselysensitive to imipenem but resistant to meropenem as

well as five samples were intermediary to imipenem butsensitive to meropenemMost carbapenem-resistant sampleswere nonsusceptible to diverse antibiotics containingcephalosporins fluoroquinolone aminoglycosides etc Therates of nonsusceptibility to different antimicrobial agentswere commonly gt90 Clinical information and phenotypic

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

8 BioMed Research International

Table 3 Statistics of the microarray repeatability

Repeatability (CV)Intra-chip experiments Inter-chip experiment

Repeat times (n) 1 2 3 3KPC 523 475 985 1016NDM-1 367 412 644 671OXA-23 447 503 782 754OXA-48 358 387 683 642OXA-51 445 531 809 788IMP 684 695 1024 1102VIM 587 604 997 1019DIM 402 389 652 667

results of all clinical samples were shown in SupplementaryTable (available here)

35 Detection of Carbapenemase-Producing Strains byMicroarray in Clinical Samples A total of 416 clinicalsamples collected from Chinese PLA General Hospital weretested The majority of the samples had previously beenwell-characterized as carbapenem resistance using K-Bmethod The microarrays for these bacteria revealed that78 (325) of the samples carried one or more carbapenemasegenes and that in some samples more than one bla gene hadbeen identified (ie 496 genes were found in 325 specimens)The genotyping results of the clinical samples are listedas follows 256 (62) carried 119887119897119886KPC 137 (33) carried119887119897119886OXA-51 40 (96) carried 119887119897119886OXA-23 22 (53) carried119887119897119886OXA-48 27 (65) carried 119887119897119886NDM-1 5 (12) carried 119887119897119886IMP3 (07) carried 119887119897119886VIM and 6 (14) carried 119887119897119886DIM (Table 4)Therefore 119887119897119886KPC was the most frequent carbapenemase genefound in the antibiotic-resistant sample Most interestinglywe found that 119887119897119886OXA-51 frequently coexisted with otherbla genes For example 89 samples that carried 119887119897119886KPCalso carried 119887119897119886OXA-51 all 40 samples of 119887119897119886OXA-23 carried119887119897119886OXA-51 9 samples carried all three bla genes 119887119897119886KPC119887119897119886OXA-51 and 119887119897119886OXA-23 and 5 samples carrying 119887119897119886OXA-48also carried both 119887119897119886OXA-51 and 119887119897119886KPC It is worth notingthat all carbapenemase genes in the 325 antibiotic-resistantsamples had been verified by sequencing In parallel we alsoamplified the isolates that were susceptible to carbapenemsperformed in phenotypic tests as well as recruited inour microarray assays The overall concordance betweenthe microarray-based assay and the reference methods(standard DNA sequencing) was 978 suggesting that ourmicroarray is a highly reliable method for detecting thecarbapenemase-producing GNB in the clinic

Fourteen samples which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified determining 966concordance between the phenotypic and microarray testsTaken together the newly developed microarray detectionmethod is comparable to the conventional antibiotic sus-ceptibility test and therefore may be suitable for clinicalapplications

4 Discussion

In this study we developed a novel microarray method todetect carbapenemase genes that could be applied to clinicaldiagnosis and identification of carbapenemase-producingGNB We included eight carbapenemase genes that havebeen shown or potential display to be most clinically rele-vant 119887119897119886KPC 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP119887119897119886VIM and 119887119897119886DIM in the design of the microarray chipsMtDNA and 16S rRNA were chosen as internal controlsbecause our samples were taken from human and thedetected target genes were from bacteria The high copynumber sequence of mtDNA was used to monitor andcontrol all PCR reactions and hybrids operations and 16SrRNA was used to prove that the DNA microarray systemcould detect bacterial genes from all samples The specificitysensitivity and reproducibility of the proposed method werehighly favorable for clinical applications Most importantlythe microarray results of the 416 clinical samples showedhighly consistent agreement with results obtained fromdirectsequencing or antibiotic susceptibility tests

It was intriguing for us to identify a number of resistantstrains coharboring two or more carbapenemase genes forinstance 119887119897119886KPC frequently coexisted with other genes and119887119897119886OXA-51 always coexisted with 119887119897119886OXA-23 inAcinetobacter baumannii indicating a more serious threat than before whenit came to the control and management of the extremelydrug-resistant bacterial infections [23] Of the transferablemolecular class B metallo-120573-lactamases IMP VIM andNDMwere common while DIMwas endemic [10] Howeverin our study 119887119897119886IMP and 119887119897119886VIM were detected in 5 and 3samples respectively yet 119887119897119886DIM was in 6 specimens whichindicated that it might be necessary to improve the attentionfor DIM in the later study (Table 4 and Supplementary Table)

The DNA microarray could detect bacterial carbapene-mase genes from not only clinical sputum urine samplesand colony or bacterial culture in this study but alsospecimens of pleural effusion cerebrospinal fluid oral swaband throat swab even environmental swab (data not shown)directly This remarkable capability of compatibility to detectof several original samples made it faster to obtain resultthan similar microarrays Check-MDR CT102 [24 25] andVITEK2 [26] that should detect cultured bacterial isolates

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

BioMed Research International 9

Table4Microarrayresults

onvario

usclinicalsam

ples

harboringcarbapenem

aseg

enes

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Noaof

samples

263

4236

555

43

11

12

21

416

Nowith

119887119897119886

KPC

gene

PCRb

193

157

380

23

01

10

00

260

Array

c190

157

370

23

01

10

00

256

Con

-cordance

of119887119897119886

KPC

results

984

100

100

974

100

100

100

100

100

100

100

100

100

985d

Nowith

119887119897119886

NDM-1

gene

PCR

192

05

01

00

00

00

027

Array

192

05

01

00

00

00

027

Con

-cordance

of119887119897119886

NDM-1

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100e

Nowith

119887119897119886

OXA-

23gene

PCR

831

10

01

00

00

00

041

Array

830

10

01

00

00

00

040

Con

-cordance

of119887119897119886

OXA-

23results

100

968

100

100

100

100

100

100

100

100

100

100

100

976

f

Nowith

119887119897119886

OXA-

48gene

PCR

201

11

00

00

00

00

023

Array

191

11

00

00

00

00

022

Con

-cordance

of119887119897119886

OXA-

48results

95100

100

100

100

100

100

100

100

100

100

100

100

957g

Nowith

119887119897119886

OXA-

51gene

PCR

8440

57

02

10

10

00

0140

Array

8239

57

02

10

10

00

0137

Con

-cordance

of119887119897119886

OXA-

51results

976

100

100

100

100

100

100

100

100

100

100

100

100

979

h

Nowith

119887119897119886

IMP

gene

PCR

41

00

00

00

00

00

05

Array

41

00

00

00

00

00

05

Con

-cordance

of119887119897119886

IMP

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100i

Nowith

119887119897119886

VIM

gene

PCR

30

00

00

00

00

00

03

Array

30

00

00

00

00

00

03

Con

-cordance

of119887119897119886

VIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100j

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

10 BioMed Research International

Table4Con

tinued

Species

Paerugi-

nosa

A bau-

mann

Ecoli

Kpn

eumoniae

Smarcescens

Smaltophilia

Flavobacteriu

mmeningosep-

ticum

Rmannitolilytica

Achrom

obacter

xylosoxidans

Burkholderiacepacia

Klebsiella

oxytoca

Enterobacter

cloacae

Eaerogenes

Total

Nowith

119887119897119886

DIM

gene

PCR

41

00

00

01

00

00

06

Array

41

00

00

01

00

00

06

Con

-cordance

of119887119897119886

DIM

results

100

100

100

100

100

100

100

100

100

100

100

100

100

100k

Totaln

oof

strains

thatagreetotaln

oof

strains

257263

3840

3636

5455

55

44

33

1111

1122

22

1140

7416

agreem

entfor

all

strains

977

95100

982

100

100

100

100

100

100

100

100

100

978

aAllteste

dsamples

inclu

desputum

urin

eandcultu

rediso

latesthe

specificc

linicalinform

ationandtestresults

ofeach

samplea

reshow

nin

theS

upplem

entary

Table

bRe

sults

obtained

with

classicalPC

Rsequ

encing

cMicroarrayresults

wereo

btainedusingcarbapenem

ase-arraydesig

nedin

thisstu

dy

dAsensitivityof

985specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of975

forK

PCdetection

eAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forN

DM-1detection

fAsensitivityof

976specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

23detection

gAsensitivityof

957specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of997

forO

XA-

48detection

hAsensitivityof

979specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of989

forO

XA-

51detection

IAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forIMPdetection

j Asensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forV

IMdetection

kAsensitivityof

100

specificity

of100

positive

predictiv

evalue

of100

and

negativ

epredictivev

alue

of100

forD

IMdetection

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

BioMed Research International 11

The microarray was rapid and portable when starting fromthe clinical sample less than 7 hours with overall 2 hoursof hands-on time enabling one day analysis Whole detec-tion operation of the DNA microarray consisted of 5 stepsand costs 4-5 h including PCR amplification and wholedetection did not need sophisticated instrument which wasfar simpler than Check-MDR CT102 The DNA microarraycould detect 9 specimens at one chip one time and the costper sample was below five dollars which was far cheaperthan Check-MDR CT102 and VITEK2 In microarray assaya proprietary CL imaging system was developed in ourlaboratory The Biochip Chemiluminescence Imager reliedon charge-coupled device (CCD) camera imaging technologyand equipped with a power supply unit for portable useThe new CL imager had a lower cost ($3000) than othercommercial CCD imaging technology CL imagers (egAmersham Imager 600 GE Healthcare Life Sciences) andmuch faster than other visual microarray system whichwas based on quantum dot-catalyzed silver deposition Thenewly designed DNA microarray system had yielded highspecificity sensitivity and reproducibility in detecting theeight carbapenemase genes among the clinical specimensFor 119887119897119886NDM-1 119887119897119886OXA-23 119887119897119886OXA-48 119887119897119886OXA-51 119887119897119886IMP 119887119897119886VIMand 119887119897119886DIM there were two different probes (22-30 nt)designed to hybridize respectively which could effectivelyminimize the false hybridization signals Our study hasdemonstrated that themicroarrays with greatly simplified theprotocol to determine the cabapenemase-encoding genes inthe clinical samples and can offer an efficient means for themolecular epidemiological studies of cabapenemase genesthe emergence of the resistant genes may potentially betraced back to their origins where community- and hospital-based bacterial infections were frequently happening On theother hand microarray hybridization is highly sensitive Inthis study they showed similar sensitivities as the real-timePCR kit We could detect as low as 30 copies120583L of DNAtargets Furthermore the DNA microarrays can generallydetect target DNAswithmuch larger dynamic ranges Finallythe microarray method is also highly reproducible we haveshown that the averaged coefficient of variations (CV ) forinterchip and intrachip experiments were low and most ofthem were less than 10Therefore we propose that the newmicroarray method has a great potential to be applied toclinical studies

The microarray based on multi-PCR which made itpossible to detect multiple resistant genes at the same time ina single tube But it could not effectively identify a variety offragments by electrophoretic separation Direct sequencingof DNA is not suitable for identification of multiple PCRproducts The microarray is used to confirm the results ofmultiple PCR The concordance of microarray to PCR ofdetection 8 carbapenemase genes ranges from 957-100(Table 4)

The microarray had some limitations because resistanceto carbapenems in particular was produced through severalmechanisms such as synthesis of carbapenem 120573-lactamasesefflux pumps [27] loss of membrane permeability [28] andpenicillin-binding proteins variants The DNA microarraycould not cover genes form all the mechanisms Fourteen

samples in our study which failed to detect any genes bymicroarray but resistant to either imipenem or meropenemin phenotypic test could not be amplified indicating resistantgenes fromothermechanismswere beyondof detection reachof the microarray Whole genome sequencing (WGS) haspotential to detect many different molecular mechanismsleading to resistance Our lab has takenWGS as new powerfultechnology to find novel resistant genes located in bacterialplasmid or genome DNA as well as new resistant mech-anisms However WGS-based antimicrobial susceptibilitytesting in clinical laboratories remain the current high-cost and taking more time [29] In clinical testing usingPCR-based microarray to detect some specific resistancegenes simultaneously was more economical and faster Inthis study multi-PCR amplification was divided into twotubes and reducing PCR amplification systems (amplifying 8carbapenemase genes in one tube) might increase sensitivity[30]

In conclusion we developed a new microarray detectionsystem that could directly detect eight carbapenemase genesfrom several kinds of clinical specimens It was convenientreadily to be customized for high-throughput detection andcould be easily adapted for clinical applications

Data Availability

The date generated and analyzed during this study are avail-able from the corresponding author on reasonable request

Disclosure

The funder had no role in study design data collectionand analysis decision to publish or preparation of themanuscript

Conflicts of Interest

The authors declare that the research was conducted in theabsence of any commercial or financial relationships thatcould be construed as potential conflicts of interest

Authorsrsquo Contributions

Yi Song and Fengna Dou contributed equally as first authorsQiqi Liu and Yu Zhou conceived and designed the study YiSong Fengna Dou Sha He and Qiqi Liu were responsible forthe experimental analysis Yi Song wrote the paper and othercoauthors contributed to the final draft All authors read andapproved the final manuscript

Acknowledgments

Thanks are due to the team of Peoplersquos Liberation Army301 Hospital of China and the Cancer Hospital of Chi-nese Academy of Medical Sciences and Peking UnionMedical College This work was financially supported byMajor Special Projects on Infectious Diseases of China

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

12 BioMed Research International

(no 2013ZX10004802-008) and National Science and Tech-nology Major Project ldquoPrevention and Control of MajorInfectious Diseases such as AIDS and Viral Hepatitisrdquo (no2018ZX10711001-003-003)

Supplementary Materials

Additional file 1 Supplementary Table Clinical informationand detection results of carbapenemase-producing samples(Supplementary Materials)

References

[1] S D Braun O S Dorneanu T Vremera A Reiszligig S Mon-ecke and R Ehricht ldquoCarbapenemase-producing Enterobacte-riaceae a 2-year surveillance in a hospital in Iasi RomaniardquoFuture Microbiology vol 11 no 3 pp 391ndash401 2016

[2] S Kuo Y Lee S Yang et al ldquoEradication of multidrug-resistant Acinetobacter baumannii from the respiratory tractwith inhaled colistin methanesulfonate a matched case-controlstudyrdquoClinicalMicrobiology and Infection vol 18 no 9 pp 870ndash876 2012

[3] A Carricajo P O Verhoeven S Guezzou N Fonsale andG Aubert ldquoDetection of carbapenemase-producing bacte-ria by using an ultra-performance liquid chromatography-tandem mass spectrometry methodrdquo Antimicrobial Agents andChemotherapy vol 58 no 2 pp 1231ndash1234 2014

[4] CLSI Performance Standards for Antimicrobial SusceptibilityTesting Clinical and Laboratory Standards Institute Pa USA27th edition 2017

[5] P Nordmann and L Poirel ldquoStrategies for identificationof carbapenemase-producing enterobacteriaceaerdquo Journal ofAntimicrobial Chemotherapy vol 68 no 3 pp 487ndash489 2013

[6] P Wang S Chen Y Guo et al ldquoOccurrence of false positiveresults for the detection of carbapenemases in carbapenemase-negative Escherichia coli and Klebsiella pneumoniae isolatesrdquoPLoS ONE vol 6 no 10 p e26356 2011

[7] D Rawat and D Nair ldquoExtended-spectrum szlig-lactamases ingram negative bacteriardquo Journal of Global Infectious Diseasesvol 2 no 3 pp 263ndash274 2010

[8] D van Duin and Y Doi ldquoThe global epidemiology ofcarbapenemase-producing Enterobacteriaceaerdquo Virulence vol8 no 4 pp 460ndash469 2017

[9] P Nordmann T Naas and L Poirel ldquoGlobal spread of car-bapenemase producingEnterobacteriaceaerdquoEmerging InfectiousDiseases vol 17 no 10 pp 1791ndash1798 2011

[10] S-S Jean W-S Lee C Lam C-W Hsu R-J Chen and P-RHsueh ldquoCarbapenemase-producingGram-negative bacteriacurrent epidemics antimicrobial susceptibility and treatmentoptionsrdquo Future Microbiology vol 10 no 3 pp 407ndash425 2015

[11] L Poirel J M Rodriguez-Martinez N Al Naiemi Y J Debets-Ossenkopp and P Nordmann ldquoCharacterization of DIM-1 anintegron-encoded metallo-120573-lactamase from a Pseudomonasstutzeri clinical isolate in theNetherlandsrdquoAntimicrobial Agentsand Chemotherapy vol 54 no 6 pp 2420ndash2424 2010

[12] T A Leski U Bangura D H Jimmy et al ldquoIdentification ofblaOXA-51-like blaOXA-58 blaDIM-1 and blaVIM carbapen-emase genes in hospital enterobacteriaceae isolates from sierraleonerdquo Journal of Clinical Microbiology vol 51 no 7 pp 2435ndash2438 2013

[13] P Li W Niu H Li et al ldquoRapid detection of Acinetobac-ter baumannii and molecular epidemiology of carbapenem-resistant A baumannii in two comprehensive hospitals ofBeijing Chinardquo Frontiers in Microbiology vol 6 p 997 2015

[14] S Ji Y Chen Z Ruan et al ldquoPrevalence of carbapenem-hydrolyzing class D 120573-lactamase genes in Acinetobacter sppisolates in Chinardquo European Journal of Clinical Microbiology ampInfectious Diseases vol 33 no 6 pp 989ndash997 2014

[15] G Patel and R A Bonomo ldquoStatus report on carbapenemaseschallenges and prospectsrdquo Expert Review of Anti-infective er-apy vol 9 no 5 pp 555ndash570 2011

[16] T Chen Y Lee S Kuo et al ldquoEmergence and distribution ofplasmids bearing the blaOXA-51-like gene with an upstreamISAba1 in carbapenem-resistant acinetobacter baumannii iso-lates in Taiwanrdquo Antimicrobial Agents and Chemotherapy vol54 no 11 pp 4575ndash4581 2010

[17] Y Lee S Kuo M Chiang et al ldquo Emergence of Carbapenem-Resistant Non-baumannii Species of Acinetobacter Harboringa bla OXA-51 -Like Gene That Is Intrinsic to A baumannii rdquoAntimicrobial Agents and Chemotherapy vol 56 no 2 pp 1124ndash1127 2012

[18] B A Evans and S G B Amyes ldquoOXA 120573-lactamasesrdquo ClinicalMicrobiology Reviews vol 27 no 2 pp 241ndash263 2014

[19] D Mazzatti F-L Lim A OrsquoHara I S Wood and P TrayhurnldquoA microarray analysis of the hypoxia-induced modulation ofgene expression in human adipocytesrdquo Archives of Physiologyand Biochemistry vol 118 no 3 pp 112ndash120 2012

[20] K A Kolquist R A Schultz A Furrow et al ldquoMicroarray-based comparative genomic hybridization of cancer targetsreveals novel recurrent genetic aberrations in the myelodys-plastic syndromesrdquo Cancer Genetics vol 204 no 11 pp 603ndash628 2011

[21] E C Chen S A Miller J L DeRisi and C Y Chiu ldquoUsing apan-viral microarray assay (Virochip) to screen clinical samplesfor viral pathogensrdquo Journal of Visualized Experiments no 502011

[22] D L Paterson ldquoResistance in gram-negative bacteria enter-obacteriaceaerdquo American Journal of Medicine vol 119 no 6 ppS20ndashS28 2006

[23] C Rizek L Fu L C dos Santos et al ldquoCharacterizationof carbapenem-resistant Pseudomonas aeruginosa clinical iso-lates carrying multiple genes coding for this antibiotic resis-tancerdquo Annals of Clinical Microbiology and Antimicrobials vol13 no 43 2014

[24] T Naas G Cuzon P Bogaerts Y Glupczynski and P Nord-mann ldquoEvaluation of a DNA microarray (check-MDR CT102)for rapid detection of TEM SHV and CTX-M extended-spectrum 120573-lactamases and of KPC OXA-48 VIM IMP andNDM-1 carbapenemasesrdquo Journal of Clinical Microbiology vol49 no 4 pp 1608ndash1613 2011

[25] J Cohen Stuart G Voets J Scharringa A C Fluit and M ALeverstein-Van Hall ldquoDetection of carbapenemase-producingEnterobacteriaceae with a commercial DNA microarrayrdquo Jour-nal of Medical Microbiology vol 61 no 6 pp 809ndash812 2012

[26] S Dally K Lemuth M Kaase S Rupp C Knabbe and JWeile ldquoDNA microarray for genotyping antibiotic resistancedeterminants in Acinetobacter baumannii clinical isolatesrdquoAntimicrobial Agents and Chemotherapy vol 57 no 10 pp4761ndash4768 2013

[27] B D Ling L Zhang and X Z Li ldquoAntimicrobial resistanceand drug efflux pumps in Acinetobacterrdquo in Efflux-Mediated

BioMed Research International 13

Antimicrobial Resistance in Bacteria X Z Li C Elkins and HZgurskaya Eds pp 329ndash358 Springer International Publish-ing Cham Switzerland 2016

[28] E Fonseca E Scheidegger F S Freitas R Cipriano and A CVicente ldquoCarbapenem-resistantAcinetobacter baumannii fromBrazil role of carO alleles expression and blaOXA-23 generdquoBMCMicrobiology vol 13 no 1 p 245 2013

[29] M Ellington O Ekelund F Aarestrup et al ldquoThe role of wholegenome sequencing in antimicrobial susceptibility testing ofbacteria report from the EUCAST Subcommitteerdquo ClinicalMicrobiology and Infection vol 23 no 1 pp 2ndash22 2017

[30] Y Zhang Q Liu D Wang S Chen X Wang and S WangldquoGenotyping and detection of common avian and humanorigin-influenza viruses using a portable chemiluminescenceimaging microarrayrdquo SpringerPlus vol 5 no 1 p 1871 2016

Hindawiwwwhindawicom

International Journal of

Volume 2018

Zoology

Hindawiwwwhindawicom Volume 2018

Anatomy Research International

PeptidesInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Journal of Parasitology Research

GenomicsInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Neuroscience Journal

Hindawiwwwhindawicom Volume 2018

BioMed Research International

Cell BiologyInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Biochemistry Research International

ArchaeaHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Genetics Research International

Hindawiwwwhindawicom Volume 2018

Advances in

Virolog y Stem Cells International

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Enzyme Research

Hindawiwwwhindawicom Volume 2018

International Journal of

MicrobiologyHindawiwwwhindawicom

Nucleic AcidsJournal of

Volume 2018

Submit your manuscripts atwwwhindawicom

Hindawiwwwhindawicom

International Journal of

Volume 2018

Zoology

Hindawiwwwhindawicom Volume 2018

Anatomy Research International

PeptidesInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Journal of Parasitology Research

GenomicsInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Hindawiwwwhindawicom Volume 2018

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Neuroscience Journal

Hindawiwwwhindawicom Volume 2018

BioMed Research International

Cell BiologyInternational Journal of

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Biochemistry Research International

ArchaeaHindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Genetics Research International

Hindawiwwwhindawicom Volume 2018

Advances in

Virolog y Stem Cells International

Hindawiwwwhindawicom Volume 2018

Hindawiwwwhindawicom Volume 2018

Enzyme Research

Hindawiwwwhindawicom Volume 2018

International Journal of

MicrobiologyHindawiwwwhindawicom

Nucleic AcidsJournal of

Volume 2018

Submit your manuscripts atwwwhindawicom