Multiplex PCR to detect pAmpC β-lactamases among enterobacteriaceae at a tertiary care laboratory in Mumbai, India

Abstract

Drug-resistance due to AmpC β-lactamases represents a growing problem worldwide. In this study, a previously collected sample of 108 cefoxitin-resistant clinical isolates was assessed for AmpC β-lactamase production through routine phenotypic testing and double-disc cefoxitin/cloxcallin (DD-CC), cefoxitin/phenylboronic acid (CDT-PBA) and AmpC disc tests. The same isolates were characterized by a novel multiplex polymerase chain reaction molecular assay to detect the presence of bla_ACT, bla_DHA, bla_CIT, bla_FOX, bla_MIR and bla_MOX. By phenotypic analysis, 56%, 55% and 48 % were detected as being AmpC β-lactamase producers by the CDT-PBA, DD-CC and AmpC disc tests, respectively. By molecular analysis, 57  % were determined to be AmpC β-lactamase producers, including 34 % bla_FOX, 8 % bla_CIT and 1.6 % bla_DHAas mono-AmpC producers. The production of multiple AmpC molecular types was common, including 30 % with both bla_CIT+FOX and 1.6 % each of bla_CIT+DHA, bla_ACT+MIR, bla_ACT+FOX, bla_ACT+DHA and bla_MIR+FOX. Molecular characterization of AmpC would help detect the prevalence of AmpC β-lactamase producers, facilitate proper patient management and implement infection control practices.

Drug resistance is a major concern in community and hospital settings worldwide. Resistance can develop through multiple different resistance mechanisms, including bacterial production of compounds such as the AmpC β-lactamases, which inactivate commonly used antibiotics. The presence of an AmpC β-lactamase confers resistance to penicillins, cephalosporins, oxyimino-cephalosporins (ceftriaxone, cefotaxime and ceftazidime), cephamycins (e.g. cefoxitin and cefotetan) and monobactams [1, 2]. AmpC β-lactamases are increasingly implicated as major causes of morbidity and mortality and can be characterized by specific genotypes that are either chromosomally or plasmid-mediated [3]. Several clinically important organisms lack chromosomally encoded AmpC (including Klebsiella species, Proteus mirabilis and Salmonella species), while others ( Escherichia coli ) have only negligible chromosomal expression [4]. As a result, the list of common plasmid-mediated AmpC genotypes affecting these species is relatively short, including ACT, DHA, CIT, FOX, MIR, MOX and ACC [5]. When present, these plasmid-mediated enzymes are expressed at high levels, which allows for rapid transfer both within and between species, but could also allow for rapid detection [6].

Unlike for extended-spectrum β-lactamases (ESBLs) and carbapenemases, standard screening and confirmation methods for AmpC β-lactamases are not provided by the Clinical and Laboratory Standards Institute (CLSI). In practice this leads to the use of numerous detection methods, for isolates resistant to cephamycins and susceptible to cefepime, including proxy assays such as the modified Hodge test, the Tris-EDTA test, boronic acid inhibitor-based testing and rapid chromogenic assays to detect AmpC β-lactamases [4, 7]. These methods are associated with varied sensitivity and specificity for AmpC β-lactamase identification, and are costly in terms of both human resources and time [2, 4, 8].

As a result, many laboratories still do not attempt to detect AmpC β-lactamases [9]. One relatively simple alternative – multiplex polymerase chain reaction (PCR) – has been successfully employed by many laboratories to identify AmpC β-lactamases [5]. In order to evaluate the test characteristics of a multiplex PCR assay to detect AmpC β-lactamases, we performed the test on a previously collected samples of cefoxitin-resistant clinical isolates of Enterobacteriaceae .

The study was carried out in the Department of Microbiology of the P. D. Hinduja National Hospital and Medical Research Centre, a tertiary care centre in Mumbai, India. A total of 108 consecutive, non-duplicate clinical isolates with cefoxitin resistance were collected between January and June 2013 from a variety of clinical specimens. Bacterial species identification was performed according to standard procedures [10]. Susceptibility testing was performed by the Kirby–Bauer disc diffusion method for each isolate using the following drug concentrations: amoxicillin/clavulanic acid (30 µg), ampicillin/sulbactam (10/10 µg), piperacillin/tazobactam (100/10 µg), cefazolin (30 µg), cefuroxime (30 µg), cefotaxime (30 µg), ceftriaxone (30 µg), cefoxitin (30 µg), ceftazidime (30 µg), cefaperazone/sulbactam (50/50 µg/disc), ciprofloxacin (5 µg), cefepime (30 µg), trimethoprim/sulfamethoxazole (25 µg), gentamicin (50 µg), netilmicin (30 µg), amikacin (10 µg), nalidixic acid (30 µg), imipenem (10 µg), meropenem (10 µg) and ertapenem (10 µg) . The zone of inhibition was interpreted as ‘susceptible,’ ‘intermediate’ or ‘resistant’ according to CLSI guidelines. [10]

Two steps were followed to confirm the presence of AmpC β-lactamases: the identification of bacterial isolates that were resistant to cefoxitin with a zone size of ≤18 mm and confirmatory testing.

Detecting AmpC by phenotypic methods

Phenotypic detection of AmpC β- lactamases was performed by the MHT, as described elsewhere [11]. The formation of a cloverleaf pattern around the cefoxitin zone for the test strain led to a positive report for the production of AmpC β- lactamases, while isolates with no distortion around cefoxitin zone were reported as negative.

Confirmatory testing for cefoxitin-resistant strains was performed using the following three methods.

1. Detection of class C β-lactamases [double-disc cefoxitin/cloxcallin (DD-CC)]. To confirm the presence of AmpC-β- lactamases, inhibitor-based testing was performed. According to this method, differences in the inhibition zones between cefoxitin (30 µg) alone and in combination with cloxcallin (200 µg) were determined. An increase in the zone size of ≥5 mm was reported as positive for AmpC β-lactamase production [12].

2. Detection of class C β-lactamases (CDT-PBA). Additionally, inhibitor-based testing was performed to detect differences in the inhibition zones between cefoxitin (30 µg) alone and in combination with phenyl boronic acid (300 µg). An increase in the zone size of ≥5 mm was reported as positive for AmpC β-lactamase production [12].

3. Detection of class C β-lactamases (AmpC disc test/Tris EDTA test). The surface of a Mueller–Hilton agar (MH) plate was inoculated with a laboratory standard strain of E. coli (ATCC 25922) in a bacterial suspension of 0.5 MacFarland standard. An AmpC disc was rehydrated with 20 µl saline and several colonies of test organism were applied to the disc for inoculation on the plate. Next, a 30 µg cefotixin disc was placed next to the AmpC disc and the plates were incubated at 35° C. The presence of a zone of inhibition around the AmpC disc indicated that the test strain was an AmpC producer, while the absence of a zone of inhibition indicated that the test strain was not an AmpC producer [4, 12].

Genomic DNA was extracted by the heat-boil method. A multiplex PCR assay was developed to identify the following targets from bacterial isolates: bla_ACT, bla_DHA, bla_CIT, bla_FOX, bla_MIR and bla_MOX. The primers employed are listed in Table 1. PCR was completed with the initial denaturation performed at 94 °C for 3 min, followed by 17 cycles each of 94 °C for 30 s, 45.3 °C for 30 s and 72 °C for 30 s, followed by 17 cycles of 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s, with a final extension of 72 °C for 6 min. Post-amplification products were visualized on 3 % agarose gel electrophoresis.

Table 1. Polymerase chain reaction primers used to identify AmpC β-lactamase production.

Name	Sequence (5′−3′)	Length	Target gene	Variants coverage
ACT-F	AAC CRT CCG RCA TRA GCR AR	120 bp	blaACT	1, 2, 3, 4, 6, 10, 13, 38
ACT-R	RGT ARC CCC AG CRT AAR GR
DHA-F	CTR CAA CAC TRA TRT CCG CT	217 bp	blaDHA	DHA 1–7, 14,15,17,22.
DHA-R	ARTR GTA RTT RCA RTG ACC RGC R
CIT-F	RAT RCC ARC CAC RTT CARGA	262 bp	blaCIT	LAT 1, 3, 4 and CMY 2, 4, 5, 6, 7
CIT-R	RGT ARC TRC CAA ARC CAC CA
FOX-F	RAC CCT RTT CRA GAR TRG CT	325 bp	blaFOX	1–10 (except 2 & 9)
FOX-R	ACR ART TRC RGC CAR GTRG AC
MIR-F	AGC CAC ACT ACT RTA CRCR TCR	401 bp	blaMIR	1–17 (except 14)
MIR-R	ARAGCTCARRCCGRAARGRT
MOX-F	CAA CRA CAA RCC ARC CRG TR	173 bp	blaMOX	MOX 1 to 8, CMY 1, 8 to 11
MOX-R	CAR CCT TRA RCA CGR CCA C

R, G or T; S, A or T.

A total of 108 cefoxitin-resistant Enterobacteriaceae were identified, including 52 samples from urine, 24 from blood, 9 from pus, 4 from sputum, 4 from fluid, 3 from tissue, 3 from bronchoalveolar lavage (BAL) fluid, 2 each from endotracheal secretions, wound swabs, stool and ascitic fluid, and 1 from cerebrospinal fluid (CSF). The majority of samples (69/108, 63.8 %) were referred from other institutions, hospitals, or clinics, while 39 (36.1 %) were obtained from inpatients at our institution. Of these, the majority were E. coli (75), followed by Klebsiella species (28), Enterobacter species (3) and Proteus species (2). Each isolate was tested against at least six different classes of antibiotics. A high rate of resistance was observed against ciprofloxacin/ofloxacin, gentamicin and netilmicin. All of the isolates were resistant to all of the cephalosporins and β-lactam/β-lactamase inhibitor combination drugs tested. All isolates were susceptible to carbapenems . Lower rates of susceptibility were identified for amikacin.

Phenotypic testing identified 51 isolates with AmpC β-lactamase production by the MHT (47 %), 61 by the inhibitor-based test (cefoxitin-PBA, 56 %), 60 by the double-disc diffusion test (cefoxitin/cloxacillin, 55 %) and 52 by the AmpC disc test (48 %).

In-house multiplex PCR identified AmpC production among 62 isolates (57 %), with 46 isolates testing negative for all AmpC mechanisms evaluated (42 %). Among the AmpC producers, 35 (56 %) employed a single mechanism and 26 (42 %) produced AmpC through more than one of the evaluated mechanisms. The most common isolates with single AmpC production mechanisms included 29 bla_FOX producers (34 %), 5 bla_CIT producers (8 %) and 2 bla_DHA producer (1.6 %). Among those with multiple AmpC production methods, 21 (30 %) produced AmpC by bla_CIT+FOX, 1 produced AmpC by bla_CIT+DHA (1.6 %), 1 produced AmpC by bla_ACT+MIR (1.6 %), 1 produced AmpC by bla_ACT+FOX (1.6 %) and 1 produced AmpC by bla_MIR+FOX (1.6 %). The genotypic and phenotypic AmpC results are further characterized in Table 2. AmpC production was more commonly identified from urine, blood, BAL fluid and ascitic fluid than from other sample types. The majority of AmpC production was identified among E. coli isolates (50), followed by Klebsiella species (10), and Enterobacter and Proteus species (1 each). Compared to phenotypic assays, the sensitivity and specificity of genotypic assay for detection of AmpC β-lactamases was 90 and 87 %, 86 and 85 %, 82 and 73 %, and 78 and 71 %, compared to the CDT-PBA test, the DD-CC test, the Tris-EDTA test and the MHT, respectively (Table 3).

Table 2. Summary of AmpC β-lactamase production identified by molecular and phenotypic testing, n=108.

	No. of isolates	Percentage (%)
Molecular characterization
AmpC producers	62	57 %
Non-AmpC producers	46	42 %
Mono-AmpC producers	36	56 %
FOX	29	34.2 %
CIT	5	8%
DHA	2	1.6%
Multiple AmpC producers	26	41.9 %
CIT+DHA	1	1.6 %
ACT+MIR	1	1.6 %
ACT+DHA	1	1.6 %
ACT+FOX	1	1.6 %
MIR+FOX	1	1.6 %
CIT+FOX	21	30 %
Phenotypic AmpC producers
Screening test
MHT (cefoxitin)-positive	51
MHT (cefoxitin)-negative	57
AmpC producers (MHT)	47.2 %
Confirmatory test
Inhibitor-based test – boranic acid CDT-PBA
CDT-PBA-positive	61
CDT-PBA-negative	47
AmpC producers (CDT-PBA)	56.5 %
Double-disc diffusion test – cefotaxime/cloxacillin (DD-CC)
DD-CC-positive	60
DD-CC-negative	48
AmpC producers (DD-CC)	55.5 %
AmpC disc test/Tris-EDTA
Tris-EDTA-positive	52
Tris-EDTA-negative	56
AmpC producers (AmpC disc test)	48.1 %

POS, positive; NEG, negative.

Table 3. Genotypic test vs Phenotypic test to detect AmpC.

n=108	Screening		Confirmatory
	MHT-POS	MHT-NEG	CDT-PBA-POS	CDT-PBA-NEG	DD-CC-POS	DD-CC-NEG	Tris-EDTA-POS	Tris-EDTA-NEG
PCR-POS	40	16	55	6	52	7	43	15
PCR-NEG	11	41	6	41	8	41	9	41
Total	51	57	61	47	60	48	52	56
Sensitivity	78.4 %		90.1 %		86.6 %		82.6 %
Specificity	71.9 %		87.2 %		85.4 %		73.2 %

POS, positive; NEG, negative.

In this study, we evaluated 108 cefoxitin-resistant Enterobacteriaceae for AmpC β-lactamase production through a combination of phenotypic methods and an in-house multiplex PCR assay of common plasmid-mediated AmpC β-lactamase production genotypes. Our multiplex PCR test identified AmpC β-lactamase production among 57 % of isolates. In comparison to a composite measure of positive by any test, the CDT-PBA test was found to be highest sensitivity phenotypic assay, while the multiplex PCR assay was 90 % sensitive and 87 % specific compared to that test.

AmpC β-lactamase production is a major contributor to drug resistance worldwide and has been associated with substantial increases in length of inpatient stay and treatment cost [3]. Nearly 250 distinct AmpC β-lactamases have been reported, with different geographical distributions. In this study, phenotypic testing identified AmpC β-lactamase production among 47, 56, 55 and 48 % of isolates by the MHT, the inhibitor-based test, the double-disc diffusion test and the AmpC disc test, respectively. Plasmid-mediated AmpC represents between 8.4 and 77 % of AmpC β-lactamase production worldwide [13] and between 3.3 and 37 % in India [2]. In this study, 57 % of tested isolates were plasmid-mediated AmpC β-lactamase producers, which is consistent with other studies reporting plasmid-mediated production rates of between 53 and 88 % in isolates [4, 8, 14]. The existing testing based on cefoxitin resistance is nonspecific, with 19–28 % of cefoxitin-resistant isolates found not to be AmpC producers, indicating a need for more reliable testing methods. [15, 16]. In this study, 42 % of cefoxitin-resistant isolates were found not to produce AmpC, or at least not to produce it using the mechanisms evaluated by our multiplex PCR assay. Given that between 9.9 and 36 % of isolates that produce AmpC β-lactamases also produce ESBLs, improved detection and specific identification of β-lactamases could have important implications for treatment [8, 14, 16, 17].

Phenotypic detection of AmpC β-lactamases is time- and human resource-intensive, with poor sensitivity and specificity. As a result, while there are several screening assays available [4, 7], no standard guidelines exist to confirm the presence of AmpC β-lactamases. In this study, 43 % of MHT-positive isolates were found to be AmpC β-lactamase producers, suggesting that there are other mechanisms for other positive isolates, such as decreased porin channels or increased efflux pump expression. Furthermore, to confirm the presence of AmpC β-lactamases, the three-dimensional test is the most reliable assay, with good sensitivity and specificity, but it is labour-intensive and cannot be performed on a routine basis. With certain modifications to the three-dimensional test, the AmpC disc test is widely employed and has good sensitivity and specificity across different studies. The AmpC disc test using Tris-EDTA is one of the most commonly employed tests for AmpC production, with sensitivity ranging from 60 to 97% and specificity ranging from 91 to 98 % [4, 6]. In this study, 48 % of isolates were positive by the AmpC disc test, which represented a sensitivity of 83 % and a specificity of 73 % as compared to a composite positive result from all methods. The double-disc test, which is also widely used, performed better against our composite result, with a sensitivity of 86 % and a specificity of 85 %. Both tests, however, demonstrated inferior performance to the CDT-PBA inhibitor test, which had a sensitivity of 90 % and a specificity of 87 %.

Given the limitations of these phenotypic assays, molecular methods are considered to be the gold standard for the identification of AmpC β-lactamases. In this study, the multiplex PCR assay identified 62 isolates as AmpC β-lactamase producers (57 %), with the most common mechanism of AmpC production being bla_FOX (34 %), followed by bla_CIT (8 %) and bla_DHA (1.6 %), which is consistent with the literature [9]. The presence of multiple AmpC β-lactamases in a single isolate has been observed in several previous molecular studies. In this study, 42 % of cefoxitin-resistant Enterobacteriaceae produced multiple AmpC β-lactamases, among which bla_CIT+FOX (30 %) was the most frequent AmpC producer, although bla_CIT+DHA, bla_MIR+FOX, bla_ACT+MIR, bla_ACT+FOX and bla_{ACT + DHA} were also observed. Even though molecular characterization of AmpC is not widely practised by clinical laboratories, AmpC mechanism typing is crucial as it is correlated with mortality. The presence of the DHA-1 enzyme, for example, is associated with 40 % mortality, while the FOX-7 enzyme is associated with 28 % mortality, and the CMY-1 enzyme is associated with only 14 % mortality [9, 18].

Plasmid-mediated AmpC β-lactamases are common among Enterobacteriaceae in India, with 37 % of E. coli and 24 % of Klebsiella isolates being positive in one study [2]. This study identified AmpC production among 66 % of E. coli , 35 % of Klebsiella species and 33 % Enterobacter species isolates that were resistant to cefoxitin, suggesting a higher rate of AmpC β-lactamase production in our setting. Given such high rates, and the association of different AmpC mechanisms with mortality, it may be worth evaluating bacteria other than Enterobacteriaceae for plasmid-mediated AmpC β-lactamases, since plasmids from these bacteria can be transferred to other species, including Pseudomonas , Acinetobacter .

The presence of AmpC β-lactamases poses an important epidemiological and clinical challenge that impacts directly on patient outcomes. This study evaluated a new multiplex PCR assay to evaluate AmpC β-lactamase production with high sensitivity and specificity compared to existing tests. Such a test has the potential to provide timely information to improve both specific treatment and infection control measures for resistant Gram-negative bacterial infections in India.