Evaluation of a Commercial Microarray System for Detection of SHV-, TEM-, CTX-M-, and KPC-Type β-Lactamase Genes in Gram-Negative Isolates

Abstract

We evaluated the ability of a commercial microarray system (Check KPC/ESBL; Check-Points Health BV) to detect clinically important class A β-lactamase genes. A total of 106 Gram-negative strains were tested. The following sensitivity and specificity results were recorded, respectively: for bla_SHV, 98.8% and 100%; for bla_TEM, 100% and 96.4%; and for bla_CTX-M and bla_KPC, 100% and 100%.

The spread of class A or group 2be extended-spectrum β-lactamases (ESBLs) represents an emerging public-health concern (1, 17). Among the organisms of the Enterobacteriaceae family (e.g., Klebsiella pneumoniae and Escherichia coli), the most frequently detected and clinically important ESBLs belong to the TEM, SHV, and CTX-M families (17). While TEM- and SHV-type ESBLs arise via substitutions in strategically positioned amino acids (e.g., Gly238 and Arg164) from the natural narrow-spectrum TEM-1, TEM-2, or SHV-1 β-lactamase genes, all currently identified CTX-M enzymes demonstrate an ESBL phenotype (7, 14).

The ability to rapidly identify narrow-spectrum β-lactamases (e.g., SHV-11 and TEM-1) and ESBLs (e.g., SHV-5 and SHV-12 or TEM-10) has important clinical implications. Usually, Enterobacteriaceae species producing narrow-spectrum enzymes are resistant to penicillins and narrow-spectrum cephalosporins, whereas those producing ESBLs manifest resistance to extended-spectrum oxyimino-cephalosporins and aztreonam (14). Since resistance to quinolones and aminoglycosides is frequently observed among ESBL producers, carbapenems represent one of the therapeutic options of last resort for life-threatening infections due to these organisms (6, 14).

In some geographic areas, the spread of carbapenemases belonging to class A (e.g., KPCs), class B (e.g., VIMs and IMPs), and class D (e.g., OXA-48) has significantly compromised the clinical use of carbapenems, consigning clinicians to the use of “last-line” antimicrobials such as colistin (2, 19). In particular, the KPC β-lactamases (primarily KPC-2 and KPC-3) are the serine carbapenemases that are most widespread in the United States, and strains producing these enzymes are responsible for numerous outbreaks with high mortality rates (3, 9, 13). Although nine KPC-type β-lactamases have been described, their susceptibility profiles are similar, rendering the differentiation of these variants less clinically relevant (13, 22).

Prompt and appropriate antibiotic treatment for infections due to ESBLs- and/or KPC-producing Enterobacteriaceae may positively affect the final outcome for infected patients (6, 13). Unfortunately, standard and confirmatory phenotypic tests may fail to identify ESBL- and, more frequently, KPC-producing organisms. For the latter group, the use of the modified Hodge test delays the final report by an additional 24 h (11, 12, 21). Therefore, a rapid and reliable method is needed to perform a quick and accurate analysis of the most important bla genes possessed by clinical isolates.

Microarray technologies are promising genotyping systems that possess a high multiplexing capacity and can be used for detecting different β-lactamase genes that are present in a single strain (8, 10, 23). This ability can assist clinicians in directing antimicrobial therapy. In the present work, we evaluated the ability of Check KPC/ESBL (Check-Points Health BV, Wageningen, Netherlands), the first rapid, commercially available, microarray-based diagnostic test system for detection and identification of bla genes belonging to the TEM, SHV, CTX-M, and KPC types. This system can detect single nucleotide polymorphisms found in the most important TEM- and SHV-type ESBLs (www.lahey.org/studies), including single mutations corresponding to amino acid positions Val84Ile, Glu104Lys, Arg164Ser/His/Cys, and Gly238Ser in TEMs and Gly238Ser/Ala and/or Glu240Lys in SHVs (7).

A total of 102 Enterobacteriaceae and four Acinetobacter baumannii isolates possessing different bla genes were tested (Table 1; see also Table S1 in the supplemental material). The majority of strains (n = 61) had previously been characterized (3-5, 15), whereas the bla genes of the remaining isolates were characterized by PCR amplification, standard DNA sequencing, and analytical isoelectric focusing (aIEF) as previously described (4). In this collection, isolates possessed an average of three different bla genes (range, one to five; see Table S1 in the supplemental material). The collection also included K. pneumoniae ATCC 700603, which produces the SHV-18 ESBL (20), and six E. coli DH10B control strains in which single bla genes are carried in different plasmid vectors (see Table S1 in the supplemental material).

TABLE 1.

Performance of the Check KPC/ESBL microarray assay in identification of β-lactamase genes^a

Species	blaKPC		blaSHV						blaTEM						blaCTX-M		Isolates with all bla genes correctly reported
	No. of isolates positive by PCR/Seq	Agreementi (%)	ESBLd		Non-ESBL		Total		ESBLd		Non-ESBL		Total		No. of isolates positive by PCR/Seq	Agreementi (%)	Agreementi (%)
			No. of isolates positive by PCR/Seq	Agreementi (%)	No. of isolates positive by PCR/Seqc	Agreementi (%)	No. of isolates positive by PCR/Seq	Agreementi (%)	No. of isolates positive by PCR/Seq	Agreementi (%)	No. of isolates positive by PCR/Seqe	Agreementi (%)	No. of isolates positive by PCR/Seq	Agreementi (%)
K. pneumoniae (n = 79)b	52	79/79 (100)	41	41/41 (100)	38	31/38 (81.6)	79	72/79 (91.1)	3	3/3 (100)	64	64/64 (100)	67	79/79 (100)	8	79/79 (100)	72/79 (91.1)
E. coli (n = 22)b	4	22/22 (100)	3	2/3 (66.7)	2	2/2 (100)	5	21/22 (95.5)	2	2/2 (100)	7	7/7 (100)	9	21/22 (95.5)	9	22/22 (100)	20/22 (90.9)
A. baumannii (n = 4)	0	4/4 (100)	0		0		0	4/4 (100)	2	2/2 (100)			2	4/4 (100)	2	4/4 (100)	4/4 (100)
Enterobacter cloacae (n = 1)	1	1/1 (100)	0		0		0	1/1 (100)	0		1	1/1 (100)	1	1/1 (100)	0	1/1 (100)	1/1 (100)
Total (n = 106)	57	106/106 (100)f	44	43/44 (97.7)	40	33/40 (82.5)	84	98/106 (92.5)g	7	7/7 (100)	72	72/72 (100)	79	105/106 (99.1)h	19	106/106 (100)f	97/106 (91.5)

^aResults represent comparisons of the Check KPC/ESBL microarray assay to PCR amplification followed by DNA sequencing (PCR/Seq).

^bIncludes control isolates.

^cOne E. coli isolate containing bla_SHV-38 was reported as a non-ESBL SHV type; six K. pneumoniae isolates containing a bla_SHV-non-ESBL gene (i.e., a bla_SHV-1 or bla_SHV-11 gene) were reported as representing SHV-ESBL strains; one K. pneumoniae isolate containing bla_SHV-11 was reported as SHV negative (see Table S1 in the supplemental material).

^dIncludes isolates containing bla_ESBL (e.g., bla_SHV-12) or bla_ESBL plus bla_non-ESBL genes (e.g., bla_SHV-12 and bla_SHV-11).

^eOne bla_TEM-negative E. coli isolate was reported as a non-ESBL TEM strain.

^fSensitivity, 100%; specificity, 100%; positive predictive value, 100%; negative predictive value, 100%.

^gSensitivity, 98.8%; specificity, 100%; positive predictive value, 100%; negative predictive value, 95.7%.

^hSensitivity, 100%; specificity, 96.4%; positive predictive value, 98.8%; negative predictive value, 100%.

ⁱThe value preceding the slash is the number of isolates whose Check KPC/ESBL microarray results agreed with their PCR/Seq results, and the value following the slash is the total number of isolates tested by both assays.

Genomic DNA of strains was extracted from overnight colonies grown on blood agar (BBL, Sparks, MD) by the use of a DNeasy blood and tissue kit (Qiagen Sciences, Germantown, MD). Microarray assays were performed according to the instructions of the manufacturer (Check-Points Health BV). Briefly, templates of the target bla DNA sequences are generated during the ligation step. These templates are then amplified, and the products are hybridized in specific array tubes. Tubes are then inserted in the array tube reader upon completion of the detection reaction, and images are acquired and interpreted with software supplied by the manufacturer (Fig. 1). For 50 isolates, the complete procedure (i.e., from genomic DNA extraction to results) can be performed in approximately 8 h.

FIG. 1.

Schematic flowchart representing the different steps used by the Check-Points KPC/ESBL platform to recognize specific bla genes. (Step A) Target DNA recognition and ligation (thermocycling conditions, 95°C for 3 min, 24 cycles of 95°C for 30 s and 65°C for 5 min, and 98°C for 2 min [total, 2.75 h]). Each target-specific probe consists of two oligonucleotide probe arms that are used to detect single nucleotide polymorphisms (SNPs). These two probe arms are connected by the ligase, generating a single probe molecule only when they perfectly match the target sequence. Only connected probe arms produce the labeled amplification products detected in step C. Every target-specific probe is equipped with the same consensus primer pair necessary for step B and a unique “ZIP code” necessary for step C. (Step B) PCR amplification of the target DNA sequences (thermocycling conditions, 95°C for 10 min, 35 cycles of 95°C for 5 s, 55°C for 30 s, and 72°C for 30 s, and 98°C for 2 min [total, 1.5 h]). Using a common primer pair, target ligated sequence templates labeled with specific “ZIP codes” are multiplied. (Step C) Detection (requiring approximately 1 to 2 h of processing, depending on the number of samples). Amplification products are targeted to specific addresses on the microarray. This targeting is dependent on the specific “ZIP code.” (Step D) Immediate acquisition of images by scanning of the microarray using the array tube reader and immediate interpretation of the acquired pictures by the use of dedicated software.

Overall, the Check KPC/ESBL system correctly identified representatives of the four bla gene families tested, including differentiation between non-ESBL and ESBL genes, in 97 of 106 isolates (91.5%). Specificities of 100% were recorded for the bla_KPC, bla_SHV, and bla_CTX-M genes, whereas one false positive was reported for bla_TEM genes (specificity of 96.4%). The system detected all bla_KPC-, bla_TEM-, and bla_CTX-M-possessing isolates, including differentiation of ESBL from non-ESBL bla_TEM-containing strains (Table 1). Notably, all bla_CTX-M genes detected were classified into the appropriate family group (i.e., group I, CTX-M-1-like; group II, CTX-M-2-like; group III, CTX-M-8-like; group IV, CTX-M-9-like; group V, CTX-M-25/CTX-M-26) according to the classification method of Pitout et al. (16) (see Table S1 in the supplemental material).

Detection and recognition of the bla_SHV genes showed 92.5% agreement, with sensitivity and specificity of 98.8% and 100%, respectively (Table 1). Only 1 in 44 bla_ESBL-positive strains (i.e., bla_SHV-38-positive strains) was not identified (97.7% agreement). SHV-38 is a very rare chromosomal ESBL enzyme (group 2be) that was found in a single clinical isolate. It possesses a unique amino acid substitution (i.e., Ala146Val) and is capable of conferring resistance to ceftazidime and imipenem (18). The amino acid at position 146 is not included in those analyzed by the Check KPC/ESBL system.

Six strains with non-ESBL bla_SHV genes were misclassified as ESBLs (Table 2). Notably, three of these were bla_SHV-11-positive K. pneumoniae isolates (non-ESBL), which showed β-lactamase bands at pIs of 7.6 and 8.2 by aIEF and double spikes at positions 238 and/or 240 in the DNA sequencing traces of the bla_SHV gene. This pattern is consistent with the possible production of an SHV-ESBL (along with the non-ESBL SHV-11) that was not detected with a cloning and DNA sequencing method that we previously employed (4). Therefore, bla_SHV-positive total agreement and the overall agreement (i.e., all bla genes correctly reported) would improve by 2.8% if these three strains were classified as ESBL producers (Table 1).

TABLE 2.

Details of discrepancies between conventional PCR plus DNA sequencing results and Check KPC/ESBL genotyping results

Strain	Source or reference	Gene identified by PCR and DNA sequencing	Gene identified by Check KPC/ESBL genotyping	Comment(s)
E. coli DH10B	This study	blaSHV-38	SHV-non-ESBL	Rare chromosomal genotype found in a single K. pneumoniae isolate (18); mutation conferring ESBL phenotype not assayed with current microarray primers
K. pneumoniae VA-361a	4	blaKPC-2, blaTEM-1, blaSHV-11	KPC, TEM-non-ESBL, SHV-ESBL	Possible production of an SHV-ESBL (e.g., SHV-5/SHV-12)
K. pneumoniae VA-388	4	blaKPC-3, blaTEM-1, blaSHV-11	KPC, TEM-non-ESBL	SHV genes were not detected
K. pneumoniae VA-392a	4	blaKPC-3, blaTEM-1, blaSHV-11	KPC, TEM-non-ESBL, SHV-ESBL	Possible production of an SHV-ESBL (e.g., SHV-5/SHV-12)
K. pneumoniae VA-412a	4	blaKPC-2, blaTEM-1, blaSHV-11	KPC, TEM-non-ESBL, SHV-ESBL	Possible production of an SHV-ESBL (e.g., SHV-5/SHV-12)
K. pneumoniae VA-414	4	blaKPC-3, blaTEM-1, blaSHV-11	KPC, TEM-non-ESBL, SHV-ESBL	SHV-11 is not an ESBL
K. pneumoniae 111b	15	blaTEM-10-like, blaSHV-1-like	TEM-ESBL, SHV-ESBL	SHV-1 is not an ESBL
K. pneumoniae 438b	15	blaTEM-2-like, blaSHV-1-like	TEM-non-ESBL, SHV-ESBL	SHV-1 is not an ESBL
E. coli 25	This study	blaCTX-M-9-like	CTX-M-(IV),c TEM-non-ESBL	blaTEM genes were not detected by PCR analysis (including using internal primers); aIEF showed only one β-lactamase band at a pI of 6.7, possibly related to the CTX-M enzyme expression

^aAnalytical isoelectric focusing (aIEF) revealed β-lactamase bands with pIs of 7.6 and 8.2 and initial standard DNA sequencing of bla_SHV showed double spikes in amino acid position 238 and/or 240 (4).

^bPartial DNA sequencing (i.e., from amino acid 35 to 274 for TEMs and from 8 to 249 for SHVs). The bla_SHV and bla_TEM sequences did not show substitutions conferring an ESBL phenotype.

^cNumber in parentheses following CTX-M designation indicates the family group for CTX-M-type ESBLs (16).

The data presented above also support the previous observation that standard DNA sequencing of PCR amplification products fails to accurately detect more than one bla gene of a given family (4). In particular, many K. pneumoniae isolates possessing both bla_SHV-11 (non-ESBL) and bla_SHV-12 (ESBL) genes were initially identified incorrectly as bla_SHV-11-positive isolates only with standard DNA sequence analysis (4). In contrast, the microarray can accurately identify the bla_ESBL gene (e.g., bla_SHV-12) regardless of the coexistence of additional bla_non-ESBL genes (e.g., bla_SHV-1 and/or bla_SHV-11) (see Table S1 in the supplemental material).

In conclusion, the results of the present work show that the microarray Check KPC/ESBL system is a highly accurate tool for detection of the clinically important β-lactamase genes found among contemporary Gram-negative organisms. Due to its rapid performance, this platform could be used in epidemiological or infection control studies in which large collections of isolates need to be characterized. Furthermore, the use of Check KPC/ESBL in clinical practice may lead to more appropriate use of antimicrobial agents, reduction of costs, and improved patient outcomes. More-extensive evaluations (e.g., using clinical isolates possessing bla genes not tested in this study) are needed to establish the full potential of this methodology for detecting different resistance genes.