Antimicrobial Resistance Profiles in Escherichia coli O157 Isolates from Northern Colorado Dairies

Abstract

Escherichia coli O157 (EcO157) infections can lead to serious disease and death in humans. Although the ecology of EcO157 is complex, ruminant animals serve as an important reservoir for human infection. Dairy cattle are unique because they may be a source of contamination for milk, meat, and manure-fertilized crops. Foodborne dairy pathogens such as EcO157 are of primary importance to public health. Antimicrobial resistance (AMR) is a complex phenomenon that complicates the treatment of serious bacterial infections and is of increasing concern. In the face of recommended use restrictions for antimicrobial agents in livestock operations, current AMR patterns in known foodborne pathogens should be documented. The objective of this study was to document AMR patterns in EcO157 isolates from dairies in northern Colorado using antimicrobial agents commonly found on dairies and representative of medically important antimicrobial drug classes. Seventy-five EcO157 isolates were recovered from three dairies. Six isolates were resistant to at least 1 of the 10 tested antimicrobial agents: four were resistant to streptomycin, sulfisoxazole, and tetracycline; one was resistant to streptomycin and tetracycline; and one was resistant to only tetracycline. All resistant isolates were from a single dairy. Overall, a low prevalence (8%) of AMR was observed among the 75 EcO157 isolates. No significant effects on AMR profiles due to virulence genes, parity, or previous antimicrobial treatments within the current lactation period were detected. The results of this study provide background information for future comparative studies investigating AMR trends. Future studies should include more participating farms and more samples and should control for potential confounding factors of AMR that may underlie individual farm variation.

Escherichia coli O157 (EcO157) infections can lead to serious disease and death in humans (22). EcO157 pathogenicity is related to the presence of virulence genes (12), including stx₁ and stx₂, which encode Shiga toxins, and eaeA, which encodes the protein intimin. Shiga toxin–producing EcO157 that are classified as enterohemorrhagic E. coli (EHEC) are associated with hemorrhagic colitis and hemolytic uremic syndrome (5, 11). EcO157 without the capacity to produce Shiga toxins still may contain the virulence gene eaeA and are classified as atypical enteropathogenic E. coli (aEPEC) (7). aEPEC strains can adhere to epithelial cells leading to attaching and effacing lesions (3, 4).

Although the ecology of EcO157 is complex, ruminants serve as an important reservoir for human infection (18). EcO157 is a commensal in the gastrointestinal tract of cattle (20) and is transmitted to humans by ingestion of contaminated foodstuffs or water or through direct contact with infected cattle or other hosts (8, 25). Some dairy products provide EcO157 with favorable conditions for growth, and dairy cattle are unique in that they may be a source of contamination for milk, meat, and manure-fertilized crops. Thus, foodborne dairy pathogens such as EcO157 are of primary importance to public health (7).

Antimicrobial resistance (AMR) is a complex phenomenon that complicates the treatment of serious bacterial infections and is of increasing concern. Although antimicrobial agents generally are not recommended for treating EcO157 infections (32), administration of antimicrobial agents early in the infection may prevent the progression of disease to hemolytic uremic syndrome (13). Although the role of early antimicrobial therapy in EcO157 infections still is unclear, the emergence and dissemination of AMR among EcO157 and other zoonotic foodborne pathogens has negative clinical implications (19).

Currently, limited data are available related to EcO157 antimicrobial susceptibility on dairies (1, 9, 17). Many factors can affect the frequency of resistance determinants among bacterial populations, including exposure to antimicrobial drugs and environmental conditions that affect the fitness of certain strains of bacteria harboring resistance genes (1, 28). Bacteria use a number of mechanisms to resist the effects of antimicrobial agents. These mechanisms include modifying the antimicrobial agent, altering the agent’s target, decreasing cell wall access to the target, and implementing an alternative metabolic pathway not affected by the antimicrobial agent (16). According to the U.S. Food and Drug Administration (FDA) judicious use guidance (29), the use of antimicrobial drugs that are medically important in human medicine should be limited in animals used for food. The current list of medically important antimicrobial drug classes includes aminoglycosides, lincosamides, macrolides, penicillins, streptogramins, sulfonamides, and tetracyclines.

To better judge the effect of the recommended antimicrobial use restrictions on livestock operations, current AMR patterns in known foodborne pathogens should be documented. Surveillance of AMR among zoonotic foodborne pathogens such as EcO157 is warranted to protect public health (19), and surveillance programs have been implemented worldwide to better understand AMR trends (2, 6, 24). The objective of this study was to document AMR patterns in EcO157 isolates from dairies in northern Colorado with a panel of antimicrobial agents commonly used on dairies and representative of medically important antimicrobial drug classes.

MATERIALS AND METHODS

E. coli O157 strains.

Three free-stall dairies located within a 20-mile (32.3-km) radius of Fort Collins, CO and representing a combined population of 2,750 lactating cattle were sampled monthly for 1 year (July 2013 through June 2014). During sampling, >10 g of feces was collected via rectal palpation from a convenience sample of cows within the first 21 days of lactation (939 fecal samples representing 899 cows) and cows culled on the day of sampling (104 fecal samples representing 104 cows). Samples were collected between 2 and 4 weeks apart, with 40 cows sampled twice during their first 21 days of lactation. Five cows were sampled both during early lactation and as culled cows. Life history features including parity, days in milk, and antimicrobial treatments were obtained for the current lactation from on-farm computer record systems (Dairy Comp 305, Valley Agricultural Software, Tulare, CA; DHI-Plus, DHI Computing Service, Provo, UT).

Each 10-g fecal sample was mixed 1:10 with buffered peptone water for both enrichment culture and direct plating. For direct plating, 100 μl was spread on selective sorbitol MacConkey agar with 5-bromo-4-chloro-3-indolyl-β-d-glucuronide (Oxoid Diagnostic Reagents, Basingstoke, England) containing 1.25 mg of potassium tellurite and 0.025 mg of cefixime (CT-SMAC-BCIG; HiMedia Laboratories, Mumbai, India). Plates were incubated at 37°C for 18 to 24 h (15). On plates containing >100 straw-colored colonies, suspect colonies were confirmed as positive for EcO157 by agglutination using an E. coli O157 latex kit (Oxoid) following the manufacturer’s instructions. Latex-positive isolated colonies were stored at −80°C in 10% sterile glycerol. Before the PCR assay, glycerol was removed from each isolate by centrifugation and then removal of the supernatant. The resulting pellet was resuspended in sterile molecular grade water.

The multiplex PCR assay was performed using previously published primers targeting the stx₁ and stx₂ genes (21) and the rfb_O157 gene (31). All rfb_O157 (and thus O157)-positive isolates, regardless of the presence of stx₁ and/or stx₂, were subsequently used in a PCR assay for the eaeA gene (21). The initial hot start PCR step of 5 min at 95°C is sufficient for lysing the bacteria; therefore, the cells were not prelysed before the PCR (10). Each 25-μl PCR consisted of 12.5 μl of master mix (multiplex PCR plus kit, Qiagen, Limburg, The Netherlands), 2.5 μl of primer mix containing 0.2 μM concentrations of each primer, 5 μl of molecular grade water, and 5 μl of bacterial culture as the direct template. The thermal cycling conditions consisted of an initial incubation step at 95°C for 5 min to activate the polymerase, 40 cycles of amplification with denaturation at 95°C for 30 s, annealing at 57°C for 90 s, and extension at 72°C for 30 s, and a final extension step at 68°C for 10 min. The PCR product was analyzed by agarose gel electrophoresis on a 2% agarose gel. A 100-bp molecular marker (Lonza Group Ltd., Basel, Switzerland) was added to the same gel to aid in the calculation of the size of the amplified DNA fragments.

The diluted fecal sample remaining after direct plating was incubated for 6 h at 37°C and stored overnight at 4°C. The enriched samples not confirmed as EcO157 through direct plating (those not super shedding, i.e., with 0 to <10³ to 10⁴ CFU/g of feces) were subjected to immunomagnetic separation (IMS) using the Dyna-beads anti–E. coli O157 and BeadRetriever System (Life Technologies, Oslo, Norway). IMS samples were subsequently plated onto CT-SMAC-BCIG and incubated for 18 to 24 h at 37°C. Suspect colonies were confirmed by latex agglutination and PCR targeting rfb_O157, stx₁, and stx₂. All rfb_O157 (and thus O157)–positive isolates were subsequently subjected to the PCR assay for eaeA (21).

Antimicrobial susceptibility testing.

Bacterial isolates from frozen stocks were cultured on blood agar plates 24 h before susceptibility testing. The disk diffusion method was used with Mueller-Hinton agar (Difco, BD, Sparks, MD) and interpreted according to Clinical and Laboratory Standards Institute (Wayne, PA) 2006 recommendations using a Biomic V3 system (Giles Scientific, Santa Barbara, CA). E. coli (ATCC 25922 and ATCC 35218), Staphylococcus aureus (ATCC 29213), Enterococcus faecalis (ATCC 29212), and Pseudomonas aeruginosa (ATCC 27853) were used for quality control. EcO157 isolates were tested for AMR to 10 antimicrobial agents active against gram-negative bacteria: amoxicillin–clavulanic acid (AMO), ampicillin (AMP), ceftiofur (CEF), enrofloxacin (ENR), florfenicol (FLO), streptomycin (STR), sulfisoxazole (SUL), tetracycline (TET), trimethoprim-sulfamethoxazole (TMS), and tulathromycin (TUL) (Sensi-Discs, Cockeysville, MD).

Statistical analysis.

The proportions of drug resistant isolates by dairy, virulence genes, parity, and antimicrobial treatments were compared using a chi-square test for equal proportions (PROC FREQ, SAS v. 9.4, SAS Institute, Cary, NC).

RESULTS

Seventy-five EcO157 isolates were recovered from the three participating dairies. Of these isolates, 36 had one or both of the stx₁ and stx₂ genes (designated as EHEC) and 39 had the eaeA gene but neither the stx₁ or stx₂ gene (designated as aEPEC).

Seventy EcO157 isolates, from 70 (7.5%) of the 939 samples, were derived from early lactation cows (i.e., <21 days in milk). Two nonresistant EcO157 isolates were recovered from samples collected 2 weeks apart from the same early lactation cow. Five nonresistant EcO157 isolates were obtained from 5 (4.8%) of the 104 cull cows. One nonresistant EcO157 isolate, from 1 (0.1%) of the 1,043 total samples, was recovered from a super-shedding cow in early lactation.

Table 1 shows the AMR profiles of the isolates by dairy and virulence genes. Six isolates from six different cows were resistant to at least 1 of the 10 tested antimicrobials agents: four isolates were resistant to STR, SUL, and TET; one was resistant to STR and TET; and one was resistant to only TET. All of the resistant isolates were from a single dairy. The four isolates resistant to STR, SUL, and TET were EHECs. The other two isolates were aEPECs. One of the resistant EHEC isolates was recovered from a sample taken on 15 July 2013. The other resistant EHECs and the aEPECs were recovered from samples taken on 7 October 2013.

TABLE 1.

Antimicrobial resistance profiles for 75 EcO157 isolates from three Colorado dairies

		No. of EcO157 isolates resistant tob:
Source, profilea	Total no. of isolates	AMO	AMP	CEF	ENR	FLO	STR	SUL	TET	TMS	TUL
Dairy
A	33	0	0	0	0	0	5	4	6	0	0
B	33	0	0	0	0	0	0	0	0	0	0
C	9	0	0	0	0	0	0	0	0	0	0
Virulence profile
EHEC	36	0	0	0	0	0	4	4	4	0	0
aEPEC	39	0	0	0	0	0	1	0	2	0	0

^aEHEC. enterohemonhagic Ec0157 containing stx₁ and for stx₂ genes,aEPEC, enterohemonhagic Ec0157 containing the eaeA gene but no stx genes.

^bAMO. amoxicillin-clavulanate; AMP, ampicillin; CEF, ceftiofur, ENR, enrofloxacin; FLO, florfenicol; STR, streptomyein; SUL, sulfonamides; TET, tetracycline; TMS, trimethoprim-sulfamethoxazole; TUL, tulathromyein. Four isolates were resistant to three antimierobial agents (STR, SUL, and TET), one isolate was resistant to two antimicrobial agents (STR and TET), and one isolate was resistant to one antimicrobial agent (TET).

Comparisons of the proportion of resistant isolates by dairy revealed a significant difference between dairies (P = 0.01). For the dairy from which the resistant EcO157 isolates were recovered, comparisons of the proportion of resistant strains by virulence genes (P = 0.32), parity (P = 0.12), and previous antimicrobial treatment during the lactation (P = 0.15) revealed no significant differences.

DISCUSSION

Overall, a low prevalence of AMR was observed (8%) in the 75 EcO157 strains. In a previous study, AMR was found in 24% of EcO157:H7 and EcO157:NM strains obtained from animals, foods, and humans (17). In another study, 14% of E. coli O157:H7 strains from cattle carcasses, feces, hides, and ground beef were resistant to at least one antimicrobial agent (1).

Antimicrobial susceptibility testing in this study focused on drugs used on the participating dairies or agents representative of medically important antimicrobial drug classes. Penicillins (AMO and AMP), cephalosporins (CEF), fluoroquinolones (ENR), amphenicols (FLO), sulfonamides (SUL), tetracyclines (TET), trimethoprim-sulfamethoxazole (TMS), and macrolides (TUL) were used on all of the participating dairies. Aminoglycosides were not used on any of the dairies, but STR was tested as a representative aminoglycoside with a labeled use for cattle. Similar to our study, both Meng et al. (17) and Beier et al. (1) found that the most common AMR pattern among EcO157 isolates was to STR, SUL, and TET. Similar AMR profiles involving STR, SUL, and TET have been documented in commensal E. coli and Salmonella across U.S. dairy operations in 21 states (14, 27).

Numerous differences were noted among the dairies in terms of biosecurity, expansion, facilities, environment, herd management, labor indices, and nutritional practices. All sampled animals in the study had been exposed to at least one class of antimicrobial agent at some point during their lifetime for production or therapeutic purposes. In our study, antimicrobial resistant EcO157 isolates were detected on only one dairy. No significant effects on AMR profiles due to virulence genes, parity, or previous antimicrobial treatments within the current lactation period were detected.

AMR is conferred through specific biochemical mechanisms conveyed by resistance genes or factors (26). Multiple mechanisms of AMR can occur in a single isolate, leading to variable resistance patterns (16). Even low concentrations of drugs can select for resistant bacteria (23). The isolation of three EHECs with the same AMR profile on a single day on the same dairy highlights the potential for environmental contamination and subsequent pathogen spread at a given point in time. Many factors may have contributed to the EcO157 AMR profiles at this dairy. The issue is how to prevent and control those pathogens that harbor AMR.

Appropriate health management practices can decrease the need for use of antimicrobial agents on the farm, thereby limiting animal exposure. Such practices may include the use of vaccines, probiotics, immune enhancers, good husbandry practices, and biosecurity. Education of animal producers and veterinarians regarding such strategies is essential for preventing and controlling AMR (26). However, even with appropriate health management, cross-resistance can lead to resistance to a particular drug in bacteria that have not been exposed to that drug. For example, a heterogeneous increase in E. coli resistance to aminoglycosides (specifically STR) has been found in calves fed waste milk without aminoglycoside residues but with residues of other drugs (23). The apparent randomness of the bacterial AMR selection process argues for ongoing surveillance of AMR profiles in food-producing animals and foodborne pathogens.

As per the FDA veterinary feed directive final rule (30) announced in June 2015, an overarching goal of the FDA is to promote the judicious use of medically important antimicrobial agents in food-producing animals. Progress must be made toward eliminating production uses of medically important antimicrobial drugs and bringing the therapeutic uses of such drugs under the oversight of veterinarians (29). Limiting antimicrobial availability, implementing on-farm interventions, and enhancing surveillance are key strategies for reducing AMR in food animal agriculture (26). The results of this study provide background information for future comparative studies of AMR trends. Future studies should include more farms and more samples and should control for potential confounding factors of AMR that may underlie individual farm variation.