Antimicrobial resistant Escherichia coli were prevailing in retail chicken, beef and turkey meat samples collected from 19 census divisions across Alberta, Canada. Prevalence of E. coli in retail chicken, beef and turkey meat was 96, 82 and 85 percent, respectively. Current study showed high prevalence of E. coli in retail meats. High prevalence of E. coli in retail meats is also recorded in other provinces of Canada: 99 percent in retail chicken in Saskatchewan; 97 percent in Ontario; 94 percent in Quebec and 75 percent in British Columbia and 77 percent in retail beef in Saskatchewan; 81 percent in Ontario; 53 percent in Quebec and 79 percent in British Columbia (CIPARS, 2006). High frequency of E. coli isolation from poultry and beef products is not surprising as E. coli is normal inhabitant of intestinal tract and may contaminate meat during evisceration (Yolanda et al., 2001). However, there is lot of variation in prevalence of E. coli in different countries. In Washington, DC, USA, prevalence of E. coli, in chicken, beef and turkey meats is 39, 19 and 12 percent, respectively, (Zhao et al., 2001). In Korea, prevalence of E. coli, in retail chicken and beef is 6 and 4 percent, respectively (Lee et al., 2009). High prevalence of E. coli in all meats in Canada indicates that the meat processing plants do not process the meat under suitable hygienic and sanitary conditions or meat handling practices at the retail stores may not be of required hygienic level.
In the present study, 56 percent chicken and 62 percent turkey isolates were found resistant to tetracycline. There is irrational use of tetracycline in poultry feed as growth promoter since the inception of poultry industry so tetracycline resistance is high (82 percent) in avian isolates (Schroeder et al., 2003 and Miles et al., 2006). More than 49 percent turkey isolates of E. coli are resistant to ciprofloxacin (Yolanda et al., 2001).
Resistance of E. coli to category-I antimicrobials is of high importance in human medicine (CIPARS, 2006). Chicken isolates (33 percent) were resistant to amoxicillin/clavulanic acid, 32 percent to Ceftrioxone and 25 percent to ceftiofur. In other provinces of Canada, 29 percent chicken isolates are resistant to amoxicillin/clavulanic acid and 22 percent isolates are resistant to ceftiofur (CIPARS, 2006). Ceftiofur is most commonly used in hatcheries in Canada. Public Health Agency of Canada has proposed that resistance of cephalosporin in human is related to widespread off-label use of this class of antibiotic in poultry hatcheries (Diarrassouba et al., 2007 and Webster, 2009).
More than 16 percent beef isolates of E. coli were resistant to tetracycline. A number of factors can contribute its high prevalence such as its use as growth promoter and in disease control during cattle rearing, the availability of pools of tetracycline resistance genes in the environment and their ability to spread via mobile genetic elements (Bryan et al., 2004 and Roberts, 2003). Seven of the isolates were resistant to sulfisoxazole.
In present study, one percent of the isolates were resistant to chloramphenicol. Low resistant of E. coli to chloramphenicol could be due to discontinuation of its use in Canada (White et al., 2000). Low prevalence (2 percent) of cefoxitin resistance beef isolates of E. coli could be due the reason that the antibiotic is not used in livestock production (Donkersgoed et al., 2003) but cefoxitin resistant strains of E. coli are reported in patients in Canada due to mutation in ampC gene promoter and attenuator that results in the overproduction of ampC enzyme (Mulvey et al., 2005). Resistance of E. coli isolates to cefoxitin in present study could be due to presence of human origin E. coli in the abattoir, processing facility, environment at retail point.
None of the E. coli isolates from chicken, beef and turkey meats was resistant to amikacin and ciprofloxacin but 32 percent chicken isolates were resistant to ceftriaxone. National reports in Canada shows that E. coli isolates from any source are not resistant to amikacin, ciprofloxacin and ceftriaxone (CIPARS, 2006 and 2007). This could be due to the fact that CIPARS used a much higher resistance breakpoint which has recently been revised to >4 ug/ml. E. coli isolates from cow calf herds are susceptible to ceftriaxone, ciprofloxacin and nalidixic acid (Gow et al., 2008). The isolation of susceptible E. coli to ciprofloxacin is of utmost importance for public health because flouroquinolones are most commonly used for egg sanitizing and in humans to treat acute or extra-intestinal tract salmonellosis, for elimination of the excretion of Salmonella in the feces, and for treatment of E. coli induced infections (Poppe et al., 1995). Salmonella infection in children is treated with ceftriaxone, which has low prevalence of resistance (David et al., 2001). Absence of resistance against amikacin and ciprofloxacin favors health care workers in devising strategies to use these therapeutic agents in human medicine.
E. coli resistance to category-I antimicrobials, amoxicillin-clavulanic acid, ceftrioxone and ceftiofur is of utmost importance. Ceftiofur, which is third generation cephalosporin, is a veterinary antimicrobial and is closely related with ceftrioxone which is used to treat certain types of infections in humans (CIPARS, 2007). Resistance to ceftiofur is generally associated with resistance to amoxicillin-clavulanic acid, cefoxitin, ampicillin and ceftrioxone (CIPARS, 2006). Amoxicillin/clavulanic acid is used in veterinary sector to treat lower respiratory tract infection caused by Actinobacillus, Haemophilus and Pasteurella (Pozzi and Ben-David, 2002). In human amoxicillin-clavulanic acid is used to treat respiratory, skin and urinary tract infection. Presence and transmission of category - I resistant E. coli to food chain may transfer resistant genes to normal flora of human and may help in development of resistance and ultimately therapeutic failure (Hammerum and Heuer, 2009).
Multi-drug resistant (MDR) E. coli were prevailing in chicken, beef and turkey meats. MDR E. coli are prevailing in various geographical regions such as China, Hong Kong, Philippines, Japan, Taiwan and Singapore (Bell and Turnidge, 2002). In the present study, 79 percent E. coli isolates from chicken, beef and turkey meats were multi drug-resistant (>2) comprising tetracycline, amoxicillin/clavulanic acid, ampicillin, cefoxitin, ceftiofur, ceftrioxone, chloramphenicol, kanamycin, streptomycin and sulisoxazole. It is observed that the avian isolates resistant to tetracycline, are more likely to be resistant to additional antimicrobial agents making the tetracycline conserved in bacterial populations over time, regardless of selection pressure (Miles et al., 2006). It has been suggested that exposure of E. coli to low levels of tetracycline induces an expression of genetic loci that regulates susceptibility to cephalosporins, penicillin, chloramphenicol, tetracycline, nalidixic acid and flouroquinolones (Oguttu et al., 2008).
A strong association exists among amoxicillin-clauvulanic acid, ampicillin, cefoxitin, ceftiofur and ceftrioxone. It might be due the fact that all these antimicrobials belong to same class of antibiotic (β-lactam antibiotic). The β-lactam antibiotics are extensively used for dairy animals. The antibiotics are transported into the colostrum that is used for feeding claves (Dolejska et al., 2008). The most common β - lactamase enzyme produced by E. coli is TEM-I and SHV-I. These β - lactamases are plasmid mediated and can be transferred from species to species by conjugation and confer resistance to ampicillin, cephalothin, piperacillin, ticarcillin, cefazolin and cefamandole. Through selective pressure, various bacterial strains have been isolated carrying TEM-I and SHV-I. β - lactamases not only able to hydrolyze 1st generation cephalosporins but also cefoxitin and cefotetan (Kenneth and Sanders, 1992).
A strong association of chloramphenicol resistance with sulfisoxazole and tetracycline was observed. A strong association of gentamicin resistance with sulfisoxazole was also recorded. There is correlation of resistance amongst E. coli isolates to tetracycline, chloramphenicol, kanamycin and neomycin resistance (Sherley et al., 2004). This could be due the fact that genes of these antimicrobials might be located on a common mobile element called integrons or plasmid (Carattoli, 2001 and Boerlin et al., 2005). The integrons and their associated cassettes, move as a group as independent unit (Martinez-Freijo et al., 1999). Tetracycline, chloramphenicol, kanamycin and neomycin resistance genes appear primarily to be transferred vertically or via large scale recombination events. Resistance ability of E. coli to trimethoprim, sulfisoxazole and streptomycin also co-occur that trimethoprim, sulfisoxazole and streptomycin resistance genes transport horizontally through transposable genetic elements (Sherley et al., 2004).
Results from AMR genes revealed that a high prevalence of tetA genes was observed in chicken E. coli isolates. From turkey isolates, tetB gene was observed most frequently. The tetA and tetB genes associated with tetracycline efflux pumps were predominant in E. coli isolates from livestock and food animals (Soufi et al., 2009). From beef E. coli isolates, tetC gene was observed more frequently which is in contrast to the findings of previous reports suggesting absence of this gene in E. coli isolates from cattle (Bryan et al., 2004). The tetC gene in E. coli isolates from commercial beef processing plan, human and pigs is common (Aslam et al., 2009). This could be due to contamination of meat with human pathogens at the time of slaughtering, meat processing and retail supply of the meat or contamination of beef meat with pork meat in slaughter house as same equipment and place is used for slaughtering and processing of pork and beef meat.
There is poor relationship between genotype and phenotypic expression of E. coli. Several isolates of the E. coli phenotypically resistant to multiple antibiotics did not contain any gene cassettes conferring resistance to these antimicrobials. It indicates that some other mechanisms might be contributing the resistance to the antimicrobials (Zhao et al., 2001 and Gow et al., 2008). Enhanced persistence of antimicrobial resistance even in absence of antimicrobial exposure may be due to incorporation of resistance determinants into the chromosome (Daly and Fanning, 2000)
The β-lactamase gene (blaCMY-2) was found in 98, 98, 92, 90 and 61 percent of the isolates showing resistance to TIO, CRO, FOX, AMC and AMP, respectively. While Aslam et al. (2009) has observed blaCMY-2 in 12 percent of the isolates. The gene is present antimicrobials resistant E. coli in food animals (Winokur et al., 2001, Zhao et al., 2001 and Yan et al., 2004). Isolation of β-lactamse containing E. coli from retail meats is of great concern for human health as β-lactam antimicrobials are grouped into category-I and some in category-II by Health Canada due to its significant importance in human medicine (Health Canada, 2009).
The resistance of STR was significantly associated with strA/B and aadA genes. The strA/B gene was observed in 95 percent of streptomycin resistant E. coli isolates and very few 22 percent of the isolates were having aadA gene. Previous reports also showed a high prevalence of strA and strB genes in streptomycin resistant E. coli isolates (Sunde and Norstrom, 2005) but Aslam et al. (2009) has reported 76 percent prevalence of strA and strB genes in streptomycin resistant E. coli isolates.
The tetA, tetB and tetC genes were observed in (38, 40 and 11 percent) E coli isolates from chicken, turkey and beef, respectively but resistance of TET was 54, 52 and 3 percent with tetA, tetB and tetC genes, respectively. Although prevalence of individual TET phenotype and tetC genotype found high in this study but relationship of TET and tetC is very low. The mechanism of antimicrobial resistance development is very complex at molecular level and lots of genetic determinant supports antimicrobial resistance development even in the absence of its specific gene. Similarly many factors can suppress the expression of resistant genes (Gow et al., 2008).
On the basis of antimicrobial resistant ability, there were no association between chicken E. coli isolates collected from independent and chain stores however, on the basis of their resistance potential to gentamycin (p = 0.024) and nalidixic acid (p = 0.008), there was strong association amongst the isolates collected from independent stores. Similarly, on the basis of antimicrobial resistant genes, there was no association between beef E. coli isolates collected from independent and chain stores but on the basis of tetA and aadA genes, there was strong association amongst the isolates collected from independent stores. On the basis of antimicrobial resistance and their genes, there was no association between turkey isolates of E. coli collected from independent and chains stores. The chain stores might have supply of meat from the areas where the meat is handled in the clean and hygienic environment. Antimicrobial resistant bacteria in abattoir environment may contaminate the carcass (Gill and McGinnis, 2000). The difference in rates of enteric organism contamination of retail meats, particularly chicken carcasses, might be due to differences in store handling practices, sampling times, and product batches (Zhao et al., 2001).
A significant number of chicken isolates of E. coli showed resistance to antimicrobials and were having antimicrobial resistant genes. A non significant number of beef or turkey isolates of E. coli showed resistance to antimicrobials but significant number of the beef isolates were having tetC and turkey isolates of E. coli have aac3 (IV) genes. This might be due to difference in methods of antimicrobials administration to the different animal species. Antimicrobials are generally administered through water or feed to birds in a poultry farms. On such farms, the sick birds usually take less feed and water than healthy birds. Consequently, all the healthy birds will have more drug than those of sick birds. Resistances may develop in the gut of birds receiving less dosage of antimicrobials (Mayrhofer et al., 2006) and also extensive use of tetracycline in beef production in Canada (Carson et al., 2008).
It is observed that meat samples collected in various seasons have significant association with antimicrobial resistant E. coli isolated from various retail meats. Chicken isolates of E. coli showed resistance to various antimicrobial agents and antimicrobial genes in all seasons of the year. Beef strains of E. coli collected during winter season showed resistance to tetracycline and were having tetB gene but the isolates collected during spring season were having blaCMY-2 gene. Turkey isolates of E. coli collected during spring season showed resistance to NAL. The resistance of non-type-specific E. coli and C. jejuni to certain drugs was more prevalent during the fall season in the newly arrived animals compared to other groups. This difference may have occurred because calves born during fall are at the highest risk for bovine respiratory disease and generally tend to be treated more commonly than spring born calves (Sangeeta et al., 2009). In poultry farming, most of the drugs are administered through drinking water (Mayrhofer et al., 2006), irrespective of the season of the year, that's why antimicrobial resistance in E. coli isolates from chicken meat was observed in all four seasons of the year. E. coli from turkey meat showed resistance to NAL during spring month may be due to less production of turkey as compared to chicken. According to Statistics of Canada, 2006 census revealed production of 640,281 thousands chicken during 2006 comparing 22,849 thousand turkey during same year (Statistics Canada Census, 2006). Another report from USA reveals that seasonality component of microbial contamination of retail meat with E. coli and Campylobacter was observed more in warmer months. But statistically no significant difference was observed when data of warm and cold months was compared. This report also suggested that for seasonality component analysis, larger sample size is required with long analysis period (Zhao et al., 2001).
The pulse field gel electrophoresis (PFGE) analysis showed that E. coli types fall into 465 PFGE types, out of which 179 PFGE types showed >90% similarity with diverse banding patterns. Some of the previous studies also reported high genetic diversity in E. coli populations by using various fingerprinting techniques (Faith et al., 1996; Pacheco et al., 1997; Grif et al., 1998; Akiba et al., 1999; Jarvis et al., 2000; Galland et al., 2001; Radu et al., 2001). Most of the E. coli isolates, showing 100 percent resemblance, were considered cloned. Some of the isolates from same sample showed varying degree of results and were not cloned. This might be due to contamination of meat with more than one E. coli strain or cross contamination of meat. The observation of cross contamination by slicer in processing plant is documented (Sheen and Hwang, 2010). Also dissemination of E. coli from beef cattle through slaughter and processing to ground beef has been documented (Aslam et al., 2003). According to Canadian Food Inspection Agency, the mechanical process can spread E. coli through the meat. The modern biology has made it possible to develop DNA fingerprints of bacteria isolated from various sources and to compare them with others at genomic level (Grif et al., 1998). PFGE is one of the fingerprinting tools to generate fingerprinting patterns by chromosomal restriction. It is considered as expensive way of identifying genetic relatedness and also time consuming (Son et al., 2009) but on the other hand PFGE is considered a good tool with sufficient discriminatory power to distinguish various strains of the species. The significance of PFGE in strain identification during E. coli outbreaks has been documented (Anonymous, 1997).
Present study is unique in its own that various genotypic and phenotypic patterns were observed in various PFGE types. Some of the PFGE types consisted of E. coli isolates from different meat showed >90 percent similarity but varying AMR pheno and genotypes. Some of the colonel PFGE patterns were susceptible to all antimicrobials tested. While some of the colonel PFGE patterns showed similar genotypic and phenotypic patterns. Some of the isolates were resistant to one antimicrobial agent or having one resistance gene with similar DNA fingerprint. Also there were PFGE types with common phenotypic patterns but different genotypic profiles and vice versa. PFGE patterns of the isolates that were susceptible to all antimicrobials showed a close genetic relationship (Aslam et al., 2010). The patterns showing no genetic similarity but identical AMR phenotypic patterns could be due to other mechanisms of resistance (Yolanda et al., 2004). The difference in phenotypic and genotypic profile of the E. coli might be due to failure of expression of genes due to methylation of DNA or transposons (Marinus, 1987) or unfavorable environment.
Some of the E. coli isolates were having identical AMR pheno and genotypes but different PFGE patterns. This might be due to difference in cleavage of DNA by Xba1. The Xba1 targets the sequence 5'-T'CTAGA-3' and cleaves double-stranded DNA asymmetrically (Sales et al., 2006). The restriction enzyme cleaves the DNA at its recognition site and modification in cleavage site due to its methylation leads to generation of one or more additional bands in PFGE gels (Barrett et al., 2006). Point mutation may insert or delete a restriction site for Xba1 that may subsequently result either increase or decrease bands in PFGE patterns (Barrett et al., 2006). Some minor difference in PFGE patterns were observed in E. coli O157 (Bohm and Karch, 1992). A single band difference has been observed previously suggesting use of some other techniques along with PFGE to interpret the results (Grif et al., 1998).
Conclusion
Resistance to antimicrobials used for treating human infections was higher in E. coli recovered from chicken.
Significant statistical associations exist among various AMR genes in E. coli isolated from retail meats.
Phenotypic AMR in E. coli was highly correlated with the presence of AMR conferring genes.
Statistical associations between some AMR pheno- and genotypes suggest that the use of one antimicrobial may result in the selection of E. coli resistant to multiple antimicrobials.
