Overexpression of the adeB Gene in Clinical Isolates of Tigecycline-Nonsusceptible Acinetobacter baumannii without Insertion Mutations in adeRS

Abstract

Thirteen clinical isolates of multidrug-resistant Acinetobacter baumannii resistant to carbapenems (MRAB-C) with tigecycline nonsusceptibility were collected from individual patients in this study. None of the 13 isolates shared the same strain characteristics in molecular typing. All of them showed increased adeB transcription, as predicted. However, none of these tigecycline-nonsusceptible MRAB-C isolates were found to possess previously known adeRS mutations. Upregulation of adeB transcription may result from cross stimulation by other mechanisms.

Multidrug-resistant Acinetobacter baumannii resistant to carbapenems (MRAB-C) has been increasingly reported worldwide, raising serious concerns about the limited antimicrobial treatment options (3, 6). Tigecycline (TGC) was therefore recommended for treating MRAB-C infections (4, 7, 13). However, overexpression of the adeABC efflux system regulated by the adeRS two-component system has been reported to result in tigecycline nonsusceptibility of A. baumannii (5, 11, 12, 17, 22). We noted more frequent isolation of TGC-nonsusceptible MRAB-C in clinics in Taiwan after the general recommendation of TGC usage against MRAB-C in March 2008. In this study, we investigated clinical characteristics, therapeutic management, and outcomes for patients infected with TGC-nonsusceptible MRAB-C. The mechanism of TGC nonsusceptibility of these clinical isolates was also explored.

A total of 17 clinical isolates of MRAB-C were collected at the Tri-Service General Hospital (TSGH) in 2008 and 2009 and subjected to further studies. They were categorized into 3 groups: 4 TGC-susceptible MRAB-C isolates (group I), 4 TGC-intermediate MRAB-C isolates (group II), and 9 TGC-resistant MRAB-C isolates (group III). These strains were collected from different individuals who had no overlapping stays in the same intensive care unit (ICU) or the same ward. A. baumannii ATCC 15151 was used as a control. All the isolates were identified using the Vitek 2 system (bioMérieux, Marcy l'Etoile, France), and identifications were confirmed by sequencing of the intergenic spacer (ITS) region of the 16S-23S rRNA genes (2). An A. baumannii isolate was defined as MRAB-C according to a previously described standard (3, 8). The MIC was determined by using the Vitek 2 system and Etest (AB Biodisk, Solna, Sweden). The U.S. Food and Drug Administration-recommended TGC susceptibility breakpoints for Enterobacteriaceae (susceptible, ≦2 mg/ml; intermediate, 4 mg/ml; resistant, ≧8 mg/ml) were used as MIC interpretation criteria. Mueller-Hinton (M-H) agar plates with or without 64 mg/liter 1-(1-naphthylmethyl)-piperazine (NMP; Steinheim, Germany), an efflux pump inhibitor, were used to verify the efflux pump effect on TGC resistance. All isolates were repeatedly investigated for their adeB transcripts by using the ABI Prism 7700 sequence detection system (Applied Biosystems, Warrington, United Kingdom) on three separate occasions. A cutoff threshold (CT) value was used to represent adeB transcripts quantitatively. The ΔCT for adeB transcripts was calculated for each isolate by dividing the CT value for the 16S RNA housekeeping gene by that for the adeB gene. adeB expression relative to that in the ATCC 15151 strain was represented by the ΔCT ratio, which was calculated by dividing the ΔCT for the control stain (ATCC 15151) by that for each individual isolate. To identify the strain variation among all isolates, OXA carbapenemase multiplex PCR (23), integron cassette detection (9), and pulsed-field gel electrophoresis (PFGE) typing (18, 19) were performed. PFGE dendrograms were compared by the unweighted-pair group method using average linkages (UPGMA). Isolates were considered to be of the same cluster if they showed 85% similarity. Total adeR and adeS genes were amplified and sequenced (15).

Table 1 summarizes demographic data, therapeutic management data, and clinical outcomes for 13 patients infected with TGC-intermediate or -resistant MRAB-C. Sputum (10) or urine (3) was the source of isolates. The mean patient age was 75.4 years, and 10 (76.9%) of the patients were male. Five patients (38.5%) had been hospitalized in the ICU. Only seven patients (53.8%) had been treated with TGC for MRAB-C infection, while the remaining six patients had never been treated with TGC. Besides TGC, a median of 3 antibiotics (range, 1 to 9) had been administered to the patients. Piperacillin-tazobactam (given to 69.2% of patients), carbapenems (given to 46.2%), ceftazidime (given to 38.5%), and vancomycin (given to 38.5%) were the most prescribed. Susceptibilities to major antibiotics and characteristics of the isolates are listed in Table 2. All isolates showed resistance to imipenem (MIC range, 1.5 to 32 mg/liter) either in the Vitek 2 system or by the Etest. In the presence of NMP (64 mg/liter), MICs for the TGC-nonsusceptible isolates were lower. In the multiplex PCR analysis for integrase genes, 15 isolates were found to have the class 1 integrase gene (intI1⁺) and only one type of gene cassette in a segment of approximately 2.3 kb (aacA4-cat8-aadA1). The isolates were positive for OXA-51, OXA-23, and OXA-58. PFGE analyses of the 17 isolates and the ATCC 15151 control strain revealed 12 distinct genotypes that were classified into 9 clusters (A to I). Two main clusters, A and B, included 64.7% of the MRAB-C isolates (11 of 17). Although the 7 TGC-resistant isolates belonged to the A or B cluster, there was no temporal or spatial overlap among the infected patients during hospitalization. After confirming the individuality of each isolate, we further investigated adeB gene transcripts by using real-time reverse transcriptase (RT) PCR. Comparing the adeB gene transcripts in the isolates to those in the ATCC 15151 control strain, we found 16.8, 22.9, and 57.6 times more adeB transcripts in group I, group II, and group III isolates, respectively (Fig. 1). The level of transcription of the adeB gene in group III was significantly higher than that in group I or group II (P, <0.05 by one-way analysis of variance [ANOVA] followed by a Student-Newman-Keuls [SNK] post hoc test for group III versus group I or group II). To elucidate the mechanism of adeABC overexpression in the group III isolates, the total sequences of adeR and adeS were determined. There was no insertion of ISABA-1 (17) and no significant point mutation.

TABLE 1.

Summary of data for the 13 patients infected with tigecycline-nonsusceptible carbapenem-resistant A. baumannii

Isolate group	Isolate designation	Patient ward	Patient age (yr)/gendera	Preexisting condition(s) and/or comorbidity(ies)b	Other organism(s) isolated (sample type[s])c	Length of hospital stay (days)	Days of TGC usaged	Day of strain isolation	Other antibiotic(s)e used		Patient outcome
									Before MRAB isolation	After MRAB isolation
II	AB127	ICU	82/M	Pancreatic carcinoma, CHD, CRF		67	None	55	TZP, LEV	TZP, LEV, CAZ, IPM	Death
	AB166	Non-ICU	60/M	DM, CRF, CHD, hypertension, gout	VRE (rectal swab)	56	27-40	32	MXF, VAN	CPO, CPO, TZP, VAN	Cure
	AB098	Non-ICU	68/M	Hypertension, lung carcinoma	MRAB (wound specimen), VRE (wound specimen, urine)	98	46-66	92	CAZ, CIP, GM, TZP, FEP, AMK, TEC, MEM, LEV	LEV	Cure
	AB099	ICU	75/M	Head injury		55	None	27	TZP, LEV, MEM	MEM, LEV	Cure
III	AB106	Non-ICU	89/M	ARF, gastric ulcer, pneumonia	MRAB (blood), VRE (rectal swab, urine), ESBL-KP (sputum)	240	None	27	TZP, MXF, IPM	FEP, CPO, TZP, CL, MEM, VAN, IPM, DAP, CAZ, FLU	Death
	AB113	ICU	97/F	TB, pneumonia, gout	VRE (rectal swab)	59	None	3	CAZ, IPM, DAP	Nil	Cure
	AB118	Non-ICU	87/M	Pleurisy, pneumonia		18	None	2	TZP	Nil	Cure
	AB125	Non-ICU	60/M	Asthma, empyema	MRAB (wound specimen), VRE (rectal swab)	94	17-28	48	CPO, IPM, LEV	FEP, CAZ, NET, TZP, VAN	Cure
	AB134	Non-ICU	77/M	Colon carcinoma, peptic ulcer	MRAB (urine, CVP catheter tip sample), VRE (rectal swab)	97	27-48	47	INN, FLU, TZP, CIP, CAZ, VAN, AMK	CL, CAZ, VAN	Cure
	AB168	ICU	85/F	CHD, hypertension	ESBL-KP (sputum)	34	None	30	CAZ, TZP, VAN, ETP	TZP, FLU, IPM, CIP	Death
	AB173	Non-ICU	66/F	Pneumonia, UTI, ARF	VRE (rectal swab, urine)	86	25-30	57	ETP, CIP, CAZ, IPM, VAN, AMS	AMS, FLU, TZP	Cure
	AB097	Non-ICU	47/M	DM, esophageal carcinoma	MRAB (wound specimen), VRE (wound specimen, urine)	235	43-64	55	FEP, FLU, TZP	FLU, CPO, ETP, TZP	Death
	AB096	ICU	87/M	CVA, UTI	MRAB (urine, blood), VRE (rectal swab, urine)	127	11-35	26	VAN, TZP, FEP, ETP, CPO	CPO, ETP, TZP, VAN, DAP, AMS, AMK	Cure

^aM, male; F, female.

^bCHD, congenital heart disease; CRF, chronic renal failure; DM, diabetes mellitus; ARF, acute respiratory failure; TB, tuberculosis; UTI, urinary tract infection; CVA, cerebrovascular accident.

^cESBL-KP, extended-spectrum β-lactamase-producing K. pneumoniae; CVP, central venous pressure.

^dDays during hospital stay on which TGC was administered (the day of admission to the hospital was designated day 1).

^eIPM, imipenem; VAN, vancomycin; AMK, amikacin; TZP, piperacillin-tazobactam; MEM, meropenem; GM, gentamicin; FEP, cefepime; ETP, ertapenem; CPO, cefpirome; FLU, fluconazole; CIP, ciprofloxacin; CAZ, ceftazidime; TEC, teicoplanin; LEV, levofloxacin; MXF, moxifloxacin; AMS, ampicillin-sulbactam; DAP, daptomycin; INN, flomoxef.

TABLE 2.

Characteristics of 13 carbapenem-resistant A. baumannii isolates

Isolate group	Isolate designation	MICa (mg/liter) of:						PFGE group	PFGE subgroup	Integrase gene PCR resultb	Integron 1 PCR result	OXA carbapenemase(s) detected
		TGC	TGC + NMP	CIP	IPM	LEV	TZP
Control	ATCC 15151	0.125	0.25	0.5	0.25	0.12	16/4	D	D	Neg	Neg	OXA-51
I	AB018	2	2	>32	8	6	>128/4	G	G	1	Positive	OXA-58, OXA-51
	AB039	2	4	>32	>32	8	>128/4	B	B1	1	Positive	OXA-23, OXA-51
	AB095	2	3	>32	32	>32	>128/4	E	E	1	Positive	OXA-23, OXA-51
	AB110	2	4	>32	1.5	12	>128/4	C	C	1	Positive	OXA-51
II	AB127	4	4	>32	32	>32	>128/4	I	I	1	Positive	OXA-23, OXA-51
	AB166	4	4	>32	6	>32	>128/4	A	A3	1	Positive	OXA-51
	AB098	6	4	>32	>32	>32	>128/4	F	F	Neg	Neg	OXA-23, OXA-51
	AB099	6	4	>32	>32	>32	>128/4	H	H	1	Positive	OXA-23, OXA-51
III	AB106	8	6	>32	>32	32	>128/4	B	B1	1	Positive	OXA-23, OXA-51
	AB113	8	4	>32	32	>32	>128/4	B	B1	1	Positive	OXA-23, OXA-51
	AB118	8	6	>32	>32	>32	>128/4	B	B2	1	Positive	OXA-23, OXA-51
	AB125	8	8	>32	>32	>32	>128/4	A	A1	1	Positive	OXA-23, OXA-51
	AB134	8	3	>32	>32	>32	>128/4	A	A2	1	Positive	OXA-23, OXA-51
	AB168	8	4	>32	>32	>32	>128/4	A	A1	1	Positive	OXA-23, OXA-51
	AB173	8	4	>32	32	>32	>128/4	B	B1	1	Positive	OXA-23, OXA-51
	AB097	12	4	>32	>32	>32	>128/4	B	B1	1	Positive	OXA-23, OXA-51
	AB096	16	8	>32	>32	>32	>128/4	B	B1	1	Positive	OXA-23, OXA-51

^aDrug abbreviations are as listed for Table 1. Values for TZP are expressed as follows: MIC of piperacillin/MIC of tazobactam.

^bNeg, negative; 1, a class 1 integrase gene was detected.

FIG. 1.

Expression of adeB mRNA transcripts as assessed by RT-PCR. adeB expression in the tigecycline-susceptible clinical isolates (group I), the tigecycline-intermediate isolates (group II), and the tigecycline-resistant isolates (group III) was quantified relative to that in the reference ATCC 15151 strain. The expression of the 16S rRNA gene was used as the internal control. Each isolate was tested in triplicate. Each bar represents the average of results from three independent experiments, and the error bars represent the standard deviations. Data were analyzed using one-way ANOVA followed by the SNK post hoc test (*, P < 0.05).

As summarized in Table 1, most patients infected with tigecycline-nonsusceptible MRAB-C presented with severe underlying conditions (e.g., diabetes mellitus, a cerebrovascular accident, acute respiratory failure, chronic renal failure, urinary tract infection, and congenital heart disease) and were characterized by comorbidities, long hospital stays, and previous exposure to multiple antibiotics for multidrug-resistant bacterial infections and to invasive devices. All patients except two (those yielding isolates AB127 and AB099) had infection with other multidrug-resistant bacteria (such as vancomycin-resistant enterococci [VRE], methicillin-resistant Staphylococcus aureus [MRSA], and extended-spectrum β-lactamase-producing Klebsiella pneumoniae). Additionally, TGC-nonsusceptible MRAB was isolated not only from TGC-treated patients (7, 16), but also from the patients without any TGC prescription (1, 10, 14). In this study, we obtained 6 MRAB isolates from patients who had never received TGC treatment. The MICs for these isolates were about 4 to 8 mg/liter. Both the PFGE patterns of the isolates and the patients' clinical contact histories indicated that the infections were not nosocomial events. Although the reason for the rapid emergence of TGC nonsusceptibility among MRAB-C isolates is poorly understood, selective pressure caused by other antibiotics may contribute to nonsusceptibility to TGC.

The pharmacodynamic studies indicated that the validated TGC regimen could achieve a maximum steady-state serum drug concentration of 0.63 ± 0.28 mg/liter (mean ± standard deviation). Thus, the results do not support TGC use for bloodstream infections caused by organisms for which TGC MICs are >1 mg/liter (16, 20, 21). In Taiwan, the MIC₅₀ and MIC₉₀ of TGC for MRAB were reported to be 2 and 4 mg/liter, respectively (10). However, another study reported higher MICs (1 to 128 mg/liter) of TGC for clinical MRAB isolates. It concluded that the MIC₅₀ and MIC₉₀ of TGC were as high as 16 and 32 mg/liter, respectively (14). These results may collectively suggest that the recommended TGC regimen may be insufficient for treating MRAB-C infection, and it may even result in the emergence of TGC-nonsusceptible MRAB-C in Taiwan.

Overexpression of the adeABC efflux pump may be caused either by the ISABA-1 insertion upstream of the adeABC operon or by point mutations in adeR (Pro116Leu) and adeS (Thr153Met or Gly30Asp) (5, 11, 12, 17). In this study, we demonstrated that overexpression of the adeABC efflux pump resulted in TGC nonsusceptibility of these isolates by quantizing transcripts of the adeB gene and demonstrating conversion of the TGC resistance pattern in the presence of an efflux pump inhibitor without any previously known mutation.

In conclusion, the present study was the first attempt to describe patient characteristics and the molecular mechanism associated with TGC nonsusceptibility in clinical MRAB-C isolates. Our study may be limited by its retrospective design, the small number of isolates, and geographical confinement, with the increasing prevalence of MRAB-C in Taiwan. TGC nonsusceptibility among MRAB-C strains that had not previously been exposed to the drug is our greatest concern, and further effort will be directed toward elucidating the mechanism.

Nucleotide sequence accession numbers.

The GenBank accession numbers for the nucleotide sequences determined and reported in this study are GU647201 to GU647217 for the adeRS regions of isolates AB018, AB039, AB095, AB096, AB097, AB098, AB099, AB106, AB110, AB113, AB118, AB125, AB127, AB134, AB166, AB168, and AB173.