Prevalence of Derepressed AmpC Mutants and Extended-Spectrum β-Lactamase Producers among Clinical Isolates of Citrobacter freundii, Enterobacter spp., and Serratia marcescens in Korea: Dissemination of CTX-M-3, TEM-52, and SHV-12

Abstract

The resistance mechanism of extended-spectrum cephalosporins in clinical isolates of Citrobacter freundii, Enterobacter spp., and Serratia marcescens was studied. Of 152 isolates, 45 isolates (29.6%) were derepressed AmpC mutants and 39 isolates (25.7%) produced extended-spectrum β-lactamase (ESBLs). The most prevalent ESBLs were CTX-M enzymes, followed by TEM-52 and SHV-12.

In recent years, extended-spectrum β-lactamases (ESBLs) have become more and more prevalent in species characterized by inducible class C cephalosporinase (AmpC), such as Enterobacter spp., Citrobacter freundii, or Serratia marcescens, which frequently segregate mutants with high-level constitutive production of AmpC enzymes (5, 8-9, 12, 16). Although less common than AmpC hyperproduction, ESBLs among these species are a problem of great concern due to the potential transmission of resistance to other bacterial species and because ESBLs are usually encoded by plasmids that also harbor genes for resistance to non-β-lactam antibiotics such as aminoglycosides (13, 17).

In Korea, 47% and 51% of Enterobacter cloacae and S. marcescens isolates, respectively, were resistant to cefotaxime and resistance rates to ceftazidime were 48% and 25%, respectively, in the 1998 survey (4). Although the resistance rates of E. cloacae and S. marcescens to third-generation cephalosporins are considerably high, studies on the resistance mechanism of extended-spectrum cephalosporins among these species have been rarely performed in Korea. Therefore, we aimed to study the resistance mechanism of extended-spectrum cephalosporins in clinical isolates of C. freundii, Enterobacter spp., and S. marcescens and determine the prevalence rate of derepressed AmpC mutants or ESBL-producing organisms among these species.

Between June and November 2003 from three university hospitals located in three different cities in Korea, 152 consecutive nonduplicate nosocomial isolates, including 21 isolates of C. freundii, 15 of E. aerogenes, 44 of E. cloacae, and 72 of S. marcescens, were collected. All isolates were subjected to the double-disk synergy test (6), isoelectric focusing (10), antimicrobial susceptibility test (11), and PCR and sequencing for β-lactamase genes with the primers in Table 1.

TABLE 1.

Oligonucleotide primers used for detection of β-lactamase-encoding genes

Primer	Tma (°C)	Nucleotide sequence	GenBank accession no.; positions	Expected amplicon size (bp)
CTX-M-9-S	50	5′-TAT TGG GAG TTT GAG ATG GT-3′	AF4546633.2; 742-761	932
CTX-M-9-AS		5′-TCC TTC AAC TCA GCA AAA GT-3′	AF4546633.2; 1655-1674
CTX-M-3-S	55	5′-CGT CAC GCT GTT GTT AGG AA-3′	AJ632119.1; 180-209	780
CTX-M-3-AS		5′-ACG GCT TTC TGC CTT AGG TT-3′	AJ632119.1; 941-960
TEM-S	50	5′-ATA AAA TTC TTG AAG ACG AAA-3′	AB103506; 166-186	1,080
TEM-AS		5′-GAC AGT TAC CAA TGC TTA ATC-3′	AB103506; 1225-1245
SHV-S	55	5′-TGG TTA TGC GTT ATA TTC GCC-3′	AY223863; 166-186	865
SHV-AS		5′-GGT TAG CGT TGC CAG TGC T-3′	AY223863; 1015-1031
OXA-1-S	55	5′-AGC CGT TAA AAT TAA GCC C-3′	AV162283.2; 1052-1070	908
OXA-1-AS		5′-CTT GAT TGA AGG GTT GGG CG-3′	AV162283.2; 1941-1960
ACT-1-S	50	5′-AAACCTGTCACTCCACAAAC-3′	U58495; 244-264	887
ACT-1-AS		5′-GGGTTCGGATAGCTTTTATT-3′	U58495; 1111-1130
DHA-1-S	50	5′-GTT ACT CAC ACA CGG AAG GT-3′	AY205600; 75-94	869
DHA-1-AS		5′-TTT TAT AGT AGC GGG TCT GG-3′	AY205600; 925-944

^aAnnealing temperature used for PCR.

According to the characteristics of β-lactamase production, we defined the isolates as follows: derepressed AmpC mutants are those with a cefoxitin MIC of ≥32 mg/liter and a cefotaxime MIC of ≥32 mg/liter and without ESBL production; ESBL producers are isolates which produced ESBLs regardless of the cefoxitin or cefotaxime MIC; and inducible AmpC-producing strains are those with a cefotaxime MIC of ≤16 mg/liter and without ESBL production. When we categorized the 152 clinical isolates of C. freundii, Enterobacter spp., and S. marcescens according to the definition described above, 45 (29.6%) isolates were derepressed AmpC mutants, 39 (25.7%) isolates were ESBL producers, and 68 (44.7%) isolates were inducible AmpC-producing strains (Table 2).

TABLE 2.

Inferring β-lactam antibiotic resistance mechanism in clinical isolates of C. freundii, Enterobacter spp., and S. marcescens collected from three university hospitals

Group (no. of isolates)a	DDSTb	No. (%) of isolates
		C. freundii	E. aerogenes	E. cloacae	S. marcescens	Total
Derepressed mutants (45)	−, 45	4 (19.0)	4 (26.7)	21 (47.7)	16 (22.2)	45 (29.6)
Strains producing ESBLs (39)	+, 37; −, 2	4 (19.0)	6 (40.0)	7 (15.9)	22 (30.6)	39 (25.7)
Inducible strains (68)	−, 68	13 (62.0)	5 (33.3)	16 (36.4)	34 (47.2)	68 (44.7)
Total no. of isolates	152	21	15	44	72	152

^aEach group is defined in the text.

^bDDST, double-disk synergy test. The numbers of isolates with positive (+) and negative (−) results are shown.

Distribution of ESBLs among the clinical isolates of C. freundii, Enterobacter spp., and S. marcescens are shown in Table 3. The most prevalent ESBL types were CTX-M type ESBLs, which have been rarely found in Korea, followed by TEM-52 and SHV-12. Two C. freundii isolates produced a novel SHV-type β-lactamase with a pI of 7.6, and the bla gene encoding this enzyme differs from that encoding bla_SHV-1 by only one amino acid change, at position 54 (Asp → Glu). Because these two isolates were positive by the double-disk synergy test, we regarded this enzyme as an ESBL.

TABLE 3.

Distribution of ESBLs among the clinical isolates of C. freundii, Enterobacter spp. and S. marcescens collected from three university hospitals in Korea

Species	No. of strains	No. DDST+a	ESBL (no. of isolates)
			TEM type	SHV type	CTX-M type	AmpC typeb	Mixed ESBLs
C. freundii	4	4		SHV-2a	CTX-M3/OXA-1 like
				TEM-1/SHV-Nc (2)
E. aerogenes	6	6			TEM-1/SHV-1/CTX-M-15/OXA-1 like (3)	SHV-11/DHA-1 (3)
E. cloacae	7	5		SHV-12 (3)	CTX-M-9		TEM-52/DHA-1
							TEM-1/SHV-12/ACT-1
							SHV-12/CTX-M-9
S. marcescens	22	22	TEM-52 (9)	SHV-12 (3)	CTX-M3(4)
			TEM-52/OXA-1 like		CTX-M3/OXA-1-like (5)
Total no. of strains	39	37	10	9	14	3	3

^aNumber of strains positive by the double-disk synergy test (DDST).

^bPlasmid-mediated AmpC-type β-lactamases.

^cSHV-N is a novel SHV-derived ESBL with a pI of 7.6.

ESBLs among the isolates of Enterobacter spp., C. freundii, and S. marcescens have been described from several countries worldwide and become more and more prevalent (1, 2, 5, 9, 12-14, 16, 17). Initially, the ESBLs among these species were typical TEM or SHV enzymes (9, 13, 16), but enzymes of the CTX-M class have been described more recently (1-3, 5, 8, 12). The CTX-M-3 enzyme was the most common ESBL type among the S. marcescens isolates in Poland (12), and CTX-M-9 and CTX-M-14 were detected in E. cloacae isolates from China (3). In Korea, TEM-52, SHV-2a, and SHV-12 are the most prevalent ESBL types among the family Enterobacteriaceae and CTX-M type ESBLs have been rarely found (7, 14-15). But in this study, CTX-M enzymes such as CTX-M-3, CTX-M-9, and CTX-M-15 were commonly identified among the isolates of Enterobacter spp. and S. marcescens, suggesting dissemination of the enzymes among these species. To our knowledge, this is the first report of CTX-M-3, CTX-M-9, and CTX-M-15 among the clinical isolates of Enterobacter spp., C. freundii, and S. marcescens in Korea and of CTX-M-15 in E. aerogenes isolates in the world.

The MIC ranges and MICs at which 50% of the isolates tested are inhibited (MIC₅₀s) of several β-lactam antibiotics are given in Table 4. For cefepime, 33% (13 of 39) of isolates producing ESBLs were nonsusceptible to cefepime whereas only 6% (3 of 45) of derepressed AmpC mutants were nonsusceptible to cefepime. The data suggest that cefepime still has activity against derepressed mutants but not against ESBL producers. Thus, in order to use cefepime safely for the treatment of infections by Enterobacter spp., Citrobacter, or S. marcescens, differentiation of ESBL producers from derepressed mutants is necessary for those clinical isolates.

TABLE 4.

Percent antimicrobial susceptibilities of C. freundii, Enterobacter spp., and S. marcescens according to the resistance mechanism

Antibiotic	Derepressed mutants (n = 45)				Strains producing ESBLs (n = 39)				Inducible strains (n = 68)
	Disk diffusion test (%)a			MIC50, MIC range	Disk diffusion test (%)			MIC50, MIC range	Disk diffusion test (%)			MIC50, MIC range
	R	I	S		R	I	S		R	I	S
Cefoxitin	100			≥512, 64-≥512	67	18	15	512, 4-≥512	57	5	38	32, 4-128
Cefotetan	98	2		NDb	28	3	69	ND		1	99	ND
Cefotaxime	82	18		256, 16-≥512	72	8	20	512, 4-≥512		10	90	4, ≤1-16
Ceftazidime	84		16	128, 8-512	46	8	46	64, ≤1-256			100	2, ≤1-16
Aztreonam	33	31	36	32, 4-64	59	5	36	128, ≤1-512		1	99	2, ≤1-16
Cefepime	4	2	94	ND	28	5	67	ND	1	1	98	ND
Imipenem	2	33	65	≤1, ≤1-16	8	5	87	≤1, ≤1-16			100	≤1

^aR, resistant; I, intermediate; S, susceptible.

^bND, not done.

In 45 strains of derepressed mutants, the resistance rate to cefotetan was 98%, but none of 68 isolates of inducible strains were resistant to cefotetan. Pai et al. (14) also showed similar data that the MIC of cefotetan was ≥32 mg/liter for 94% of derepressed or partially derepressed AmpC mutants but the MIC of cefotetan was ≤16 mg/liter for 96.5% of inducible strains regardless of ESBL production. Therefore, susceptibility to cefotetan would help to differentiate derepressed mutants from inducible strains.

Antimicrobial susceptibilities to trimethoprim, ciprofloxacin, chloramphenicol, tetracycline, and aminoglycosides such as tobramycin, gentamicin, and amikacin were also examined. ESBL production was associated with high levels of resistance to trimethoprim and ciprofloxacin compared to both derepressed chromosomal AmpC mutants and inducible AmpC mutants. The resistance rates of ESBL producers to trimethoprim, ciprofloxacin, tobramycin, and gentamicin were very high at 54%, 72%, 95%, and 74%, respectively. Since ESBL producers express their β-lactamase genes from plasmids, these findings suggest that genes coding for ESBLs and genes coding for resistance to these antibiotics may reside within the same plasmids and therefore be spread together. This means that resistance to two different kinds of drugs may be coselected by the use of either one and all of these antimicrobials could be a selective pressure for spreading of such isolates.