Detection of Salmonella enterica Isolates Producing CTX-M Cephalosporinase in U.S. Livestock Populations

Abstract

We report the detection of Salmonella carrying bla_CTX-M in U.S. livestock populations. We identified 12 of 2,034 (0.6%) Salmonella isolates originating from turkeys, horses, and pigs from at least 6 U.S. states, all carrying bla_CTX-M-1, many on a pandemic sequence type 1 IncN plasmid.

TEXT

We have previously reported commensal Escherichia coli isolated from U.S. livestock harboring bla_CTX-M on conjugative plasmids (15, 29) and subsequently hypothesized that Salmonella spp. bearing bla_CTX-M are present but unrecognized in U.S. livestock populations, posing a potential public health risk. Thus, our objective was to identify and characterize Salmonella carrying bla_CTX-M among veterinary diagnostic submissions to the USDA APHIS National Veterinary Services Laboratories (NVSL).

To accomplish this, we screened 2,034 clinical Salmonella isolates submitted for serotyping to the NVSL between October 2010 and June 2011 (Table 1). Salmonella isolates that had been previously serotyped were tested in June 2011 by streaking to Mueller-Hinton agar containing 8 μg/ml cefepime. We identified a total of 12 (0.6%) Salmonella isolates carrying bla_CTX-M on transferable plasmids (Table 2). We did not detect other classes of β-lactamase resistance genes, including CMY, TEM, SHV, and OXA (11, 14, 22), by PCR. MICs for these Salmonella isolates generally displayed the expected phenotype (Fig. 1). Southern blot hybridization (10, 13, 20) using a bla_CTX-M probe indicated the localization of the gene on the plasmids (Fig. 2).

Table 1.

Salmonella clinical isolates screened^a

S. enterica serovar	No. of represented states	No. of isolates screened by species						Total no. of isolates
		Cattle	Chickens	Horses	Swine	Turkeys	Otherb
Typhimurium var. 5−	31	26	5	7	215	1	17	271
Agona	16	25	1	4	111	6	2	149
Dublin	21	116	0	0	2	0	3	121
Typhimurium	24	30	4	9	54	0	15	112
Cerro	16	100	1	1	5	0	0	107
Derby	16	1	0	0	98	1	3	103
Montevideo	20	61	0	3	6	4	6	80
Infantis	15	4	1	13	55	0	2	75
Heidelberg	16	8	1	1	53	10	1	74
Newport	21	31	0	17	4	0	18	70
Anatum	13	14	0	11	33	0	5	63
Senftenberg	13	1	2	2	34	15	8	62
Enteritidis	13	3	42	2	1	0	8	56
Kentucky	21	30	9	2	3	0	2	46
Mbandaka	13	17	2	1	12	1	4	37
Worthington	16	0	1	1	30	0	0	32
Ouakam	7	0	1	0	8	16	1	26
Bredeney	5	4	0	2	7	3	0	16
Rough O:d:e,n,z15	1	0	0	0	1	0	0	1
All other serotypesc	38	110	13	68	208	31	103	533
Total	41	581	83	144	940	88	198	2,034

^aSummary of results for 2,034 Salmonella clinical isolates originally submitted to the NVSL for serotyping between October 2010 and June 2011 that were screened for the bla_CTX-M phenotype using selective media.

^bIncludes isolates submitted from wild animals, zoo animals, cats, dogs, goats, and sheep and isolates of unknown origin.

^cIncludes 134 additional serotypes that were present in this isolate set.

Table 2.

Salmonella isolates carrying bla_CTX-M^a

Isolate no.	Isolate ID	Serotype	Serogroup	Source	State	Date received (day/mo/yr)	PRTb
1	11-13049	Anatum	E	Equine	TX	11/26/2010	N
2	11-13094	Rough O:d:e,n,z15	E	Swine	MN	11/26/2010	N, I1
3	11-13665	Bredeney	B	Turkey	UNKc	12/16/2010	N
4	11-13933	Anatum	E	Equine	TX	12/21/2010	I1
5	11-104	Ouakam	D	Turkey	AR	1/6/2011	N
6	11-2362	Ouakam	D	Turkey	MO	3/7/2011	N
7	11-2945	Anatum	E	Equine	TX	3/25/2011	nad
8	11-2946	Anatum	E	Equine	TX	3/25/2011	I1, HI1
9	11-3665	Ouakam	D	Turkey	IN	4/12/2011	N
10	11-4809	Anatum	E	Equine	TX	4/29/2011	I1
11	11-4872	Ouakam	D	Turkey	MO	5/3/2011	N
12	11-6604	Ouakam	D	Turkey	AR	5/31/2011	N
13	11-9696e	Ouakam	D	Turkey	NC	8/15/2011	N

^aSummary of results for Salmonella isolates carrying bla_CTX-M identified from clinical isolates originally submitted to the NVSL for serotyping between October 2010 and June 2011.

^bPRT, plasmid replicon type. All IncN plasmids were identified as ST1 on pMLST. None of the IncI1 plasmids matched a reported ST.

^cIsolate 11-13665 state of origin unknown because the submission form was incomplete.

^dIsolate 11-2945 plasmid replicon type could not be identified using our typing procedure.

^eIsolate 11-9696 was not part of the original isolate set used for this study but was identified upon targeted screening of S. Ouakam isolates following completion of the original screening of 2,034 isolates.

Fig 1.

MICs of 26 antimicrobial drugs for 13 Salmonella clinical isolates containing the bla_CTX-M genetic element (numbers of isolates are shown in the body of the figure). Broken lines represent susceptible breakpoints, and solid lines represent resistant breakpoints when available. Corresponding to the concentration listed at the top of each column (μg/ml), the included range of each antimicrobial is shown in gray. a, amikacin MICs were not determined for isolate 11-9696. b, clavulanic acid and tazobactam were included at fixed concentrations of 4 μg/ml. c, isolate 11-13665 (S. Bredeney) was resistant to cefoxitin and the β-lactamase inhibitors. d, isolate 11-13094 (Salmonella with rough serotype O:d:e,n,z15) was susceptible to gentamicin. e, isolates 11-13094 (Salmonella with rough serotype O:d:e,n,z15) and 11-104 (S. Ouakam) were susceptible to kanamycin and streptomycin. f, the 5 S. Anatum isolates were resistant to chloramphenicol, tetracycline, and trimethoprim-sulfamethoxazole. g, the 5 isolates susceptible to sulfisoxazole were the S. Ouakam isolates except 11-9696.

Fig 2.

(A) Plasmid content of 13 Salmonella clinical isolates containing the bla_CTX-M genetic element originally submitted to the NVSL for serotyping. Lanes: 1, BAC-Tracker supercoiled DNA ladder; 2, 11-13049; 3, 11-13094; 4, 11-13665; 5, 11-13933; 6, 11-104; 7, 11-2362; 8, 11-2945; 9, 11-2946; 10, 11-3665; 11, 11-4809; 12, 11-4872; 13, 11-6604; 14, 11-9696; 15, control plasmid DNA containing bla_CMY-2; 16, BAC-Tracker supercoiled DNA ladder. (B) Southern blot hybridization using CTX-M probe of plasmid content of 13 Salmonella clinical isolates containing the bla_CTX-M genetic element originally submitted to the NVSL for serotyping. Lanes: 1, BAC-Tracker supercoiled DNA ladder; 2, 11-13049; 3, 11-13094; 4, 11-13665; 5, 11-13933; 6, 11-104; 7, 11-2362; 8, 11-2945; 9, 11-2946; 10, 11-3665; 11, 11-4809; 12, 11-4872; 13, 11-6604; 14, 11-9696; 15, control plasmid DNA containing bla_CMY-2; 16, BAC-Tracker supercoiled DNA ladder.

We found that 6 of 88 (6.8%) turkey isolates carried bla_CTX-M. Among these, one Salmonella enterica serovar Bredeney isolate was received December 2010 and carried bla_CTX-M-1 on an IncN plasmid. The remaining five turkey isolates were all Salmonella enterica serovar Ouakam (Fig. 3) and were received between March and May 2011, originating from three U.S. states (Arkansas, Missouri, Indiana) that also carried bla_CTX-M-1 on IncN plasmids. We screened an additional 48 S. Ouakam isolates submitted between January 2009 and October 2010 in an attempt to determine if earlier S. Ouakam isolates contained this gene, although none were identified. However, one additional S. Ouakam isolate, received in August 2011 from a turkey clinical diagnostic submission originating from North Carolina, carried bla_CTX-M-1.

Fig 3.

(A) Dendrographic analysis of XbaI PFGE data for five Salmonella serovar Anatum clinical isolates containing the bla_CTX-M genetic element. (B) Dendrographic analysis of XbaI PFGE data for six Salmonella serovar Ouakam clinical isolates containing the bla_CTX-M genetic element.

Turkey production in the United States has consolidated such that there are few large centralized hatcheries that supply day-old turkey poults to grower-finisher operations throughout the United States. Ceftiofur is approved for the control of mortality in day-old turkey poults, and turkey grower-finisher operations can order poults from hatcheries that have been treated with ceftiofur prior to shipment. Turkey poults are maintained in high-population-density environments conducive to the exchange of enteric flora throughout the production system. The mass application of ceftiofur to large populations of day-old poults at the hatchery prior to shipment could provide the selection pressure required to support the emergence, dissemination, and maintenance of a ceftiofur-resistant strain of Salmonella at a central hatchery. The shipment of poults from the hatchery to grower-finisher operations in multiple states could then disseminate the resistant strain over a wide geographic area.

Pulsed-field gel electrophoresis (PFGE) analysis of the S. Ouakam isolates (Fig. 3) with XbaI (18) finds that they are similar but do not represent a single epidemic clone. Resistant Salmonella that disseminated clonally at the serotype level, but not the pulsotype level, have occurred previously, including S. Newport carrying bla_CMY-2 on an IncA/C plasmid (6, 17) and Salmonella serovar Typhimurium DT104 expressing the characteristic ACSSuT resistance phenotype (resistance to ampicillin, chloramphenicol, streptomycin, sulfonamides, and tetracycline) (1, 5).

Of the 940 swine submissions, a single isolate (0.1%) was identified to be a Salmonella isolate, with a rough O:d:enz15 serotype, carrying bla_CTX-M-1. As with turkey poults, ceftiofur is also commonly applied to large groups of piglets at weaning when they are processed into nurseries where they are maintained in a population-dense environment conducive to the spread of enteric pathogens. This mass application, typically with the extended-activity formulation of ceftiofur, may provide the selection pressure required for the emergence of a ceftiofur-resistant Salmonella strain in a population of pigs.

We found 5 of 143 (3.5%) equine isolates, all originating from Texas, were Salmonella enterica serovar Anatum (Fig. 3) carrying bla_CTX-M-1 on multiple plasmid replicon types (Table 2). The PFGE similarity (>97%) of these isolates received over a relatively short time period suggests a common source exposure, as might occur at an equine event or facility. Ceftiofur is approved for the treatment of lower respiratory tract infections in horses, but the extended-activity formulation is applied at some breeding facilities to populations of mares as a postbreeding intrauterine infusion to improve conception rates (12). This application of ceftiofur therapy to mares while transiently housed in a population-dense environment may provide the population-level selection pressure required to support the emergence and spread of a ceftiofur-resistant strain of Salmonella.

None of the 581 cattle, 83 chicken, or 198 other source isolates were found to carry bla_CTX-M. This result suggests that if Salmonella isolates carrying bla_CTX-M are present in cattle or chicken populations in the United States, either they do not produce clinical disease sufficient to initiate a diagnostic investigation, including microbiological culture and serotyping at the NVSL, or their frequency in these populations is below the detection limits of our study.

Using plasmid multilocus sequence typing (MLST) (3, 4), we identified three unique I1 plasmid sequence types (ST) (Table 3). We also identified ST 1 IncN plasmids bearing bla_CTX-M-1 in multiple Salmonella strains representing 4 serotypes from 3 animal species from diverse geographic origins. Sequence type 1 IncN plasmids carrying bla_CTX-M-1 have been previously reported to be epidemic in humans, livestock, and food in Europe (16). The pandemic dissemination of multiresistant organisms, including S. Typhimurium DT104 (5, 28) and E. coli B2-O25:H4-ST131 (9, 26), has been previously reported. However, pandemic plasmid dissemination independent of clonal spread of organisms has not been reported.

Table 3.

pMLST allele variants of IncI1 plasmids from Salmonella spp. harboring bla_CTX-M-1

Isolate ID	Allele variant
	rep1	ardA	trbA-pndC	sogS	pilL
11-13094	1	4	3	6	3
11-13933	1	4	UNa	9	3
11-2946	1	4	UN	9	3
11-4809	1	2	UN	9	3

^aUN, the allele variant has not been characterized in the pMLST database.

Salmonella bearing bla_CTX-M have not been previously reported in livestock in the United States, although they have been recovered from multiple livestock species in Europe and Asia (2, 19, 27). In addition, Salmonella bearing bla_CTX-M have been recovered sporadically from human cases of salmonellosis in the United States (23–25). The presence of these resistant isolates in U.S. livestock populations is important because most human salmonellosis cases in the United States result from food-borne zoonotic transmission in animal products (21). Empirical therapy of resistant infections by physicians without the benefit of microbiological culture and susceptibility results may result in treatment failure and thus increased health care costs (8) and higher risk of death (7).