Phenotypic, Biochemical, and Molecular Techniques for Detection of Metallo-β-Lactamase NDM in Acinetobacter baumannii

Abstract

Seven NDM-positive Acinetobacter baumannii isolates of worldwide origin were studied to evaluate the best technique for their identification. Detection of carbapenemase producers based on the measurement of carbapenemase activity by UV spectrophotometry (as for A. baumannii strains producing other types of carbapenemase), or by the modified Hodge test, failed. Inhibition activity using EDTA was a sensitive technique but lacked specificity compared to molecular techniques, which remain the gold standard.

TEXT

The emergence of the metallo-ß-lactamase (MBL) named NDM-1 (New Dehli metallo-ß-lactamase) is one of the latest and most important resistance mechanisms identified in Gram-negative rods (12, 17). The bla_NDM-1 gene, identified initially in Klebsiella pneumoniae and Escherichia coli strains mostly from India and Pakistan, was recently identified extensively in other enterobacterial species worldwide (18, 21). Besides spread of the NDM gene in Enterobacteriaceae, NDM producers in A. baumannii are increasingly being identified worldwide (9–11).

Whereas detection of NDM producers in Enterobacteriaceae is easy (16), our preliminary results indicated that detection of NDM-producing A. baumannii strains may be much more complicated. Therefore, we evaluated several techniques for the detection of NDM-producing A. baumannii clinical isolates.

A collection of seven NDM-positive Acinetobacter baumannii isolates of worldwide origin was studied, to identify the best approach to detect the production of NDM-type carbapenemases among clinical isolates. Another collection corresponding to carbapenemase-producing but NDM-negative A. baumannii isolates of worldwide origin was studied for comparison (Table 1). Those strains produced either Ambler class B carbapenemases (IMP-4, VIM-4, or SIM-1), class D carbapenemases (OXA-23, OXA-40, or OXA-58), or a class A carbapenemase (GES-14). Susceptibility testing was performed by Etest (AB bioMérieux, Solna, Sweden) on Mueller-Hinton agar plates at 37°C, and results were classified according to the updated CLSI guidelines (5). The CLSI breakpoints for A. baumannii of imipenem and meropenem were as follows: susceptible (S), ≤4 μg/ml; resistant (R), ≥16 μg/ml (5). The EUCAST breakpoints for imipenem and meropenem were as follows: S, ≤2 μg/ml; R, ≥8 μg/ml. Those for doripenem were as follows: S, ≤1 μg/ml; R, ≥4 μg/ml (13). All NDM-producing A. baumannii clinical isolates were resistant to all β-lactams, including all carbapenems, according to CLSI and EUCAST guidelines (Table 1).

Table 1.

Comparison of techniques for detection of carbapenemase-producing A. baumannii isolates

Isolated	Acquired carbapenemase(s)	MIC (μg/ml)						Result of:			Specific activity (mU/mg)a			PCR resultc	Reference(s)
		CTX	CAZ	IPM	IPM+EDTA	MEM	DOR	Etest MBL	IPM/IPM+EDTA test (diameter −/+ EDTA) (mm)	Hodge test	IPM	MEM	ERT
SLO	NDM-1	>256	>256	>32	1	>32	>32	+	+ (6/25)	−	5.1	0.6	1.4	+	This study
All	NDM-1	>256	>256	>32	<1	>32	>32	+	+ (6/24)	−	5.9	0.9	1.4	+	9
Ora-1	NDM-1	>256	>256	>32	<1	>32	>32	+	+ (6/22)	−	6.4	0.9	1.1	+	This study
StN	NDM-1	>256	>256	>32	<1	>32	>32	+	+ (6/23)	−	6.3	0.8	1.2	+	This study
Gen	NDM-1, OXA-23	>256	>256	>32	3	>32	>32	+	+ (6/18)	−	7.2	1.1	1.5	+	20
ML	NDM-2	>256	>256	>32	<1	>32	>32	+	+ (6/23)	−	4.7	0.9	1.3	+	10
124	NDM-2	>256	>256	>32	<1	>32	>32	+	+ (6/23)	−	4.5	0.7	1.3	+	This study
IMP	IMP-4	>256	>256	4	<1	8	4	+	−/+ (14/20)	−	718.1	95.6	65.2	−	4
A154	VIM-4	>256	>256	32	<1	32	32	+	+ (8/25)	+	279.8	45.2	33.3	−	6
SIM	SIM-1	>256	>256	3	<1	4	3	−/+e	−/+ (21/29)	+b	58.3	12.6	23.2	−	15
35	OXA-23	>256	>256	32	4	32	24	−/+	−/+ (9/17)	−	2.8	1.1	0.8	−	This study
Acb2	OXA-23	>256	>256	32	4	32	24	−/+	−/+ (11/20)	−	2.1	1.1	0.8	−	2
Acb13	OXA-23	>256	>256	24	4	24	16	−/+	−/+ (11/18)	−	1.9	1.2	0.9	−	2
LAH	OXA-40	>256	128	32	12	>32	32	−/+	−/+ (7/18)	+b	2.1	0.5	0.5	−	This study
MUZ	OXA-40	>256	256	32	12	>32	32	−/+	−/+ (7/16)	+b	1.6	0.6	0.4	−	This study
Acb19	OXA-58	>256	>256	4	2	4	4	−	− (14/18)	−	3.1	1.2	1.5	−	2
MAY	OXA-58	8	2	3	2	2	1	−	− (15/17)	−	2.9	1.1	0.6	−	This study
MAD	OXA-58	256	256	24	16	24	24	−	− (11/15)	+b	3.0	1.2	0.6	−	19
AP	GES-14	>256	>256	24	8	16	8	−	− (11/17)	+b	7.5	1.4	1.1	−	1

^aStandard deviations were within 10% of the means.

^bWeakly positive.

^cPCR was performed under standard conditions with NDM-For and NDM-Rev.

^dAll isolates are A. baumannii except isolate A154, which belongs to genomospecies 16.

^e−/+, indeterminate result.

The Etest MBL strip is one of the recommended tests for detection of MBLs based on inhibition of MBL activity by EDTA (16). This Etest method showed good sensitivity for the detection of MBL production. Susceptibility to imipenem was restored in the presence of EDTA, confirming the significance of MBL as a source of carbapenem resistance among MBL producers. However, several MBL-negative strains producing OXA-23 or OXA-40 gave false-positive results (Table 1). This showed that the intrinsic effect of EDTA on A. baumannii might interfere in the specificity of this test.

Another technique was therefore evaluated to detect MBL production using the same principle based on inhibition by EDTA. This technique consisted of two imipenem disks with or without 10 μl of 0.5 M EDTA (8). An increase of 10 mm in the inhibition zone diameter in the presence of EDTA was considered a positive result. An increase between 5 and 10 mm was considered doubtful and therefore required further investigation. This technique gave results similar to those obtained with Etest MBL strips, with MBL producers being easily detected, except for an IMP-4-producing isolate possessing low MICs of carbapenems (MIC of imipenem, 4 mg/liter) (Fig. 1). However, OXA-23 and OXA-40 producers gave some false-positive results, as observed with Etest MBL strips (Table 1; Fig. 1).

Fig. 1.

Results of the IPM-EDTA synergy test for carbapenemase-producing A. baumannii clinical isolates. (A) Imipenem disk (10 μg) alone. (B) Imipenem disk supplemented with 10 μl of EDTA at 0.5 M. A significant increase of the inhibition zone in the presence of EDTA was considered positive. From left to right, the panels contain an NDM-1-producing A. baumannii strain (isolate SLO), an NDM-2-producing A. baumannii strain (isolate ML), an IMP-4-producing A. baumannii strain (isolate IMP), a VIM-4-producing Acinetobacter genomospecies 16 strain (isolate A154), an OXA-23-producing A. baumannii strain (isolate Acb2), and an OXA-58-producing A. baumannii (isolate MAY).

The modified Hodge test (MHT) has been widely used for screening of carbapenemase activity because of the direct analysis of the carbapenemase activity. We therefore evaluated the MHT as previously described (14) at a turbidity of 0.5 McFarland, using E. coli ATCC 25922 as the indicator organism, K. pneumoniae BIC (producing the carbapenemase OXA-48) as a positive control, and K. pneumoniae COO (a non-carbapenemase-producing but carbapenem-resistant isolate) as a negative control. This test gave negative results for all tested NDM-producing A. baumannii strains (Table 1). VIM-, IMP- and some OXA-type producers gave weak synergistic images. This test was thus poorly sensitive and specific for detecting the carbapenemase activity of any carbapenemase producers.

Biochemical detection of NDM activity in A. baumannii was then analyzed. Ten ml of overnight broth cultures of A. baumannii isolates was centrifuged at 5,000 × g for 15 min and then sonicated in ice. Specific activities for carbapenems were measured using UV spectrophotometry at a wavelength value of 297 nm for imipenem as described previously (3). Detection of carbapenemase activity by UV spectrophotometry performed with crude culture extracts of NDM producers did not give any significant positive result. The standard mean of specific activities obtained for the NDM producers was evaluated at 5.7 mU/mg of proteins. This value was not significant compared to the value obtained with the reference strain A. baumannii AYE strain (7), which does not possess any acquired carbapenemase gene and is susceptible to imipenem (2.6 mU/mg of proteins, which corresponds to the baseline). Specific activities measured using other carbapenem molecules gave similar results but lower levels of carbapenemase activities (Table 1). Similar low-level activities were observed for carbapenem-hydrolyzing class D β-lactamases (CHDL), whereas a high level of activity was observed for class B ß-lactamases of the types VIM, IMP, and SIM (Table 1). The latter results may be related to stronger promoter or/and plasmid locations of those MBL genes.

Molecular techniques performed as previously described (16), using the primers NDM-For (5′-GGTGCATGCCCGGTGAAATC-3′) and NDM-Rev (5′-ATGCTGGCCTTGGGGAACG-3′) for internal gene amplification (660-bp amplicon) and the primers pre-NDM-for (5′-CACCTCATGTTTGAATTCGCC-3′) and pre-NDM-rev (5′-CTCTGTCACATCGAAATCGC-3′) for amplification of the entire gene sequence (984-bp amplicon), are useful tools for identification of NDM producers and the precise identification of the resistance determinant, respectively. In a single case, as reported previously, two carbapenemase genes were identified that were bla_OXA-23 and bla_NDM-1 (20). Using these PCR-based techniques, all NDM producers were detected. Multiplex PCR for detecting several carbapenemase genes should be adapted to A. baumannii, since recently developed multiplex PCR schemes focused on detection of carbapenemase genes only in Enterobacteriaceae. One limit of molecular techniques, however, is the failure to detect carbapenemase producers due to unknown carbapenemase genes.

This work indicates that identification of NDM producers may be suspected by comparing results of MIC of carbapenems with and without EDTA. Detection of carbapenemase activity using either the modified Hodge test or UV spectrophotometry is likely to fail to detect NDM producers in A. baumannii, whereas this detection is efficient with members of the Enterobacteriaceae. The discrepancy between A. baumannii and Enterobacteriaceae could be explained by a weak enzyme production in A. baumannii, likely related to the chromosomal location (a single copy) of the bla_NDM gene. Considering that many microbiology laboratories do not possess the facilities for molecular detection, carbapenem-resistant A. baumannii should be screened first by using EDTA inhibition-based techniques followed by further PCR-based techniques in a reference laboratory.

These observations may have clinical implications for detection and therefore for controlling the spread of NDM producers in A. baumannii, which is in our opinion one of the most difficult tasks, considering the persistence of multidrug-resistant A. baumannii in many health care facilities.