Multi-drug resistance in commensal bacteria of food of animal origin in Uganda

This study investigated the levels and patterns of antibiotic resistance in resistanceindicator bacteria; Escherichia coli and Enterococci of food animal origin. Isolates were obtained from fecal samples collected from food animals (chickens, pigs, cattle, goats and sheep) and tested against selected panels of antibiotics. High resistance was observed for both Enterococci species and E. coli. Enterococci isolates showed high resistance against erythromycin (60.5%), gentamycin (58.9%) and tetracycline (46.8%) and lowest resistance against penicillin. E. coli isolates showed highest resistance against erythromycin (96.0%), tetracycline (61%) and ampicillin (55.3%), and showed least resistance to gentamycin (6.9%), ciprofloxacin (6.5%) and aztreonam (8.8%), and no resistance to meropenem. Comparison of resistance between E. coli and Enterococci isolates from the same animal hosts for different antibiotics indicated that resistance was significantly higher in E. coli (p < 0.05) for erythromycin, cotrimoxazole, and ampicillin than it was in Enterococci but it was significantly lower than in Enterococci for gentamycin, chloramphenicol and ciprofloxacin. Overall 35 and 46 percent of Enterococci and E. coli were resistant to five or more antibiotics, respectively. The high level of multidrug resistance to clinically important antibiotics in commensal indicator bacteria is reservoir of a resistant gene pool that may spread resistance to pathogens in animals and has public health implications.


INTRODUCTION
There is an increasing concern for the antimicrobial resistance problem.Developing countries in particular have received limited attention regarding this problem (Okeke et al., 2007).Several studies have demonstrated that animals and humans share several ecological systems including microorganisms in those environments and they spread easily globally facilitated by globalization and the current ease of travel across the world.Administration of antimicrobial agents affects both targeted pathogenic organisms as well as non-target commensals.Thus frequent antimicrobial use creates a pool of resistant commensal bacteria that contribute to the general increase and dissemination of bacterial resistance worldwide, and can be a source of resistance genes for pathogens (Andremont, 2003).There is a comparatively higher use of antimicrobials in developed countries than in developing countries both for prophylaxis and therapy (Mitema et al., 2001) although there is higher therapeutic use in developing countries than prophylactic use.In both cases, antimicrobial use in animal production leads to a high level of colonization of animals with antimicrobial-resistant bacteria that can then contaminate the food and subsequently spread to humans.
Previous studies have shown that animal drugs are grossly misused in Uganda, are readily available without prescription, are sold in markets under direct sunlight, and farmers administer the drugs themselves (Byarugaba, 2004).This creates a pool of resistant organisms both commensals and pathogenic.Aarestrup et al. (2008) have demonstrated that the use of antimicrobials in veterinary practice greatly influences the prevalence of resistance in animal bacteria and poses a great risk for the emergence of antibiotic resistance in human pathogens as well.Most studies on antimicrobial resistance of bacteria of animal origin in developing countries have mostly examined pathogenic bacteria (Byarugaba, 2004).Data based on pathogenic bacteria is however less accurate as these pathogenic strains are frequently isolated from cases following treatment failures and therefore the resistance patterns are influenced by preceding antimicrobial treatment.Despite the potential threat of this problem in most developing countries, few systematic investigations have been conducted in livestock using the commensal indicator bacteria that can be used to estimate the exact resistance problem in animal populations.
Escherichia coli and Enterococci are frequently used as Gram-negative and positive resistance indicator bacteria, respectively, because of their high prevalence in the feces of healthy animals and because of their ability to harbor several resistance determinants (Aarestrup et al., 1998).Commensal bacteria represent an underappreciated means of assessing resistance (Marshall et al., 2009).Monitoring antimicrobial resistance using indicator bacteria avoids overestimating resistance levels by use of pathogenic bacteria, which is less accurate as the resistance patterns of pathogenic strains from treatments failures can be influenced by the preceding antimicrobial treatment.Thus, susceptibility patterns of indicator bacteria derived from healthy animals are suggested as good predictors of the resistance situation in the bacterial population as a whole (van den Bogaard and Stobberingh, 2000).
Knowledge of antibiotic resistance in food animals and their resistance mechanisms provides critical information about antibiotic problems and provides the information required to formulate strategies for containment of the problem of antimicrobial resistance and food safety.Food of animal origin has been demonstrated to be a source of a majority of food-borne bacterial infections caused by Campylobacter, Yersinia, E. coli 0157, non-typhoid Salmonella, and other pathogens (Threlfall et al., 2003;Padungtod et al., 2006;Miles et al., 2006;Meyer et al., 2008).Particular antibiotic resistance genes first described in human specific bacteria have also been found in animal-specific species of microorganisms and vice versa, suggesting that those bacterial populations can share and exchange these genes (O'Brien, 2002).Therefore, it is critical to understand the gravity of the problem and this requires that information about the levels and patterns, as well as the mechanisms of resistance to specific antibiotics of bacteria, be available.This enables a scientific risk assessment to be carried out and the impact of antibiotic use to be made, and baselines for intervention to be established (Caprioli et al., 2000).Moreover, long-term surveillance data are needed to evaluate the impact of any interventions (Caprioli et al., 2000;Aarestrup et al., 1998;Verhoef and Fluit, 2006).This study investigated the current levels and patterns of bacterial resistance in food animals to common antibiotics, using standard resistance indicator bacteria as a first step towards containment of antimicrobial resistance and reduction in the impact of antibiotic use in food animals in Uganda.

Sample collection
Samples were collected for isolation of resistant indicator bacteria E. coli and Enterococcus spp.from each of the four main food animals, namely small ruminants (sheep and goats), chickens, pigs, and cattle.Approximately 10 to 20 g of fresh fecal samples were collected in sterile plastic containers containing about 5 ml of Stuart's transport Media (Oxoid, England), from abbattoirs and slaughter slabs in Kampala.The samples were got from the animals per rectum or cloaca (broilers) before slaughter.Samples were collected over a period of six months during 2008.They were then transported in a cold box on ice to the laboratory and plated within 24 h of collection.

Isolation and identification of E. coli
Each 10 g sample was emulsified in peptone water and, using a sterile cotton swab, inoculated directly onto Mac-Conkey agar (Mast group, Merseyside, UK) by streaking, and incubated aerobically at 37°C over night.A single pink colony was sub-cultured onto Mac-Conkey agar and incubated aerobically at 45°C over night.Pink colonies were then sub cultured on XLD (Xylose Lysine Deoxycholate) agar and E. coli chromogenic agar (Oxoid, England) and the E. coli suspected colonies confirmed with API-20E identification kits (Biomerieux, USA).Pure isolates were stored in 30% glycerol at -20°C until further use.

Isolation and identification of Enterococci spp.
Enterococci were isolated by using a sterile cotton swab to pick a sample from each of the 10 g samples, emulsified in peptone water, inoculated onto bile esculine azide agar (Oxoid, UK) and incubated for 2 days at 42°C.Colonies showing a morphology typical of Enterococci were sub-cultivated onto bile esculine azide agar medium.Esculine-positive, white colonies were identified according to the following criteria: motility, production of arginine dihydrolase and the ability to ferment mannitol, sorbitol, arabinose, raffinose and melibiose.All isolates were stored at -20°C until further use in suceptibility testing.
The inoculum was prepared as follows: a saline suspension was made from a bacterial colony at a turbidity equivalent to a 0.5 McFarland standard.A sterile cotton swab was placed in the bacterial suspension and excess fluid was removed by pressing and rotating the cotton against the inside of the tube.Each swab was surface spread uniformly onto Mueller Hinton agar (Oxoid, England) plate to yield uniform growth.Antimicrobial paper disks were then applied to the surface of the plate.After incubation at 37°C for 18 to 24 h, plates were inspected for growth and inhibition zone diameters measured.The presence of a growth inhibition zone larger than the break point diameter as defined by standard procedures in the NCCLS M31A Manual (NCCLS, 1999) was considered to indicate susceptibility to the agent.The presence of a growth inhibition zone smaller than the break point diameter, the absence of any inhibition zone or the presence of isolated colonies growing inside an inhibition zone of any size were considered indicative of resistance as defined (NCCLS, 1999).In general all susceptibility testing and the interpretation of data was performed according to standard procedures as described in the NCCLS M31A Manual fo MIC determination (NCCLS, 1999).The recommended reference strains E. coli ATCC 25922 for E. coli isolates and Enterococcus faecium ATCC 2912 were used for quality control.

Data handling and analysis
The data were entered into a computer using Microsoft Excel and analyzed using SAS versions 9.13.The descriptive statistics were computed for frequency counts, relative cumulative frequency or presented graphically.The phenotypic resistance was presented as the percentage of the total isolates tested that was resistant.Multidrug resistance was defined as one isolate being resistant to three or more antibiotics tested.The proportions of the resistant isolates to the various antibiotics were compared between species of bacteria isolated and between the different food animals from which the bacteria were isolated, by using chi-square tests at a 5% level of significance.

RESULTS
A total number of 387 and 441 isolates of Enterococci species and E. coli respectively were isolated from the four species of animals as shown in Table 1, for susceptibility testing.Overall, high resistance was observed for Enterococci species to gentamycin (63.4%), erythromycin (60.4%) and tetracycline (46.8%).For E. coli the highest resistance was seen against erythromycin (96%), tetracycline (61%) and ampicillin (55%), with E. coli isolates showing a generally higher trend (Table 2).However, Enterococci species showed more resistance to chloramphenicol, gentamycin and ciprofloxacin than to E. coli isolates.E. coli isolates showed least resistance to gentamycin, ciprofloxacin and aztreonam and none were resistant to meropenem.Enterococci isolates were tested for susceptibility against (erythromycin, gentamycin, tetracycline, cotrimoxazole, chloramphenicol, ampicillin and penicillin).Figure 1 shows the levels and patterns of the resistance among the isolates from the different food animals.Generally there was highest resistance observed for isolates against erythromycin followed by tetracycline and the lowest resistance was observed in isolates against Penicillin.Compared with those from other hosts, isolates from chickens were observed to be generally more resistant, with least resistance observed from isolates obtained from small ruminants, although the highest resistance was observed against gentamycin in small ruminants.
Analysis was also made for multi-drug resistance among the Enterococci isolates.It was observed that Byarugaba et al. 1541 60% of the isolates were multi-drug resistant (Table 3).Moreover 10.6% of the isolates were resistant to seven or more drugs and about 35% were resistant to 5 or more antibiotics.When statistical comparison of the proportions of resistance was made on the isolates from the different food animal species for each antibiotic, a significant difference (p < 0.05) in the resistance was observed for all antibiotics (Table 4) with higher resistance being observed in chickens for erythromycin and tetracycline, while the highest resistance among all host species and antibiotics was resistance to gentamycin in small ruminants.E. coli isolates were tested for susceptibility against erythromycin, tetracycline, cotrimoxazole, chloramphenicol, ampicillin, streptomycin, cephalothin, nalidixic acid, gentamycin, meropenem, aztreonam.Figure 2 shows the levels and patterns of the resistance among the isolates from the different food animals.Generally there was highest resistance observed for  isolates against erythromycin followed by tetracycline, and no resistance was demonstrated against meropenem.Isolates from chickens and pigs were observed to be more resistant in general than those from cattle and small ruminants, with the least resistance observed from isolates obtained from small ruminants.Analysis for multi-drug resistance among the E. coli isolates as defined indicated 77.6% of the isolates were multi-drug resistant (Table 5).Moreover 46% of the isolates were resistant to five or more drugs while 3.6% were resistant to 9 or more antibiotics.A statistical comparison of the proportions of resistance was made of the isolates from the different food animal species for each antibiotic.A significant difference (p < 0.05) in the resistance was observed for all antibiotics except for ciprofloxacin where no significant difference was observed among the isolates (p>0.05)(Table 6).
Resistance was also compared between E. coli and Chicken Swine Cattle Small ruminants  Enterococci isolates from the same species for different antibiotics.It was found that resistance was generally higher in E. coli than Enterococci isolates for erythromycin, cotrimoxazole, tetracycline, and ampicillin while it was higher in Enterococci for gentamycin, chloramphenicol and ciprofloxacin (Figure 3).It was found that the resistance between E. coli and Enterococci was significantly different (p < 0.05) for isolates from all food animal species against all antibiotics except for isolates from small ruminants against chloramphenicol and ciprofloxacin, as well as for isolates from swine against ciprofloxacin and those from cattle against tetracycline where there was no significant difference observed (Table 7).

DISCUSSION
Studies of the prevalence of antimicrobial resistance in commensal microflora are very useful in monitoring and understanding the process of antimicrobial-mediated selection in a population (Caprioli et al., 2000).Most of the previous studies on resistance from developing countries have concentrated on pathogenic bacteria (Byarugaba, 2005;Kassa et al., 2007).The present study investigated the levels and patterns of resistance against a panel of antibiotics for two important resistanceindicator commensal bacteria; E. coli and Enterococci.
The study revealed high levels of resistance to the commonly used antibiotics; tetracycline, penicillins, and erythromycin and less resistance to antibiotics that are less commonly used in Uganda and especially in the animal industry, such as monobactams (aztreonam), and no resistance was demonstrated in any of the isolates against carbanepems like meropenem.However, there were some differences observed between the present study and the previous studies.Previous studies from bacteria isolated from clinical cases in Uganda (Byarugaba, 2005;Nakavuma et al., 1994) have reported higher levels of resistance to penicillin, erythromycin, ampicillin, than in the present study while the levels for some of the antibiotics are relatively similar.The explanation for this difference could be the fact that data based on pathogenic bacteria is less accurate as the resistance patterns of pathogenic strains isolated from clinical cases usually follows treatment failure and is influenced by preceding antimicrobial treatment.Investigations based on commensal bacteria provide a more accurate reflection of the resistance situation in the population than the data based on pathogenic isolates.The high prevalence against most antibiotics tested observed in this study suggests that bacteria of food animal origin can be a significant reservoir of resistant bacteria as has been suggested in other studies (Harada et al., 2007;Young et al., 2009).Commensal bacteria have also been reported to play a role in transmission of resistance (Andremont, 2003).Particularly, the observed resistance to aztreonam, a monobactam and other   antibiotics that are not currently used in the animal industry in Uganda, indicates that there is a possibility of exchange of resistant bacteria between humans and animals.A similar phenomenon has also been reported in studies elsewhere (Srinivasan et al., 2007).Transmission of resistant bacteria has been confirmed between bacteria of human origin and those of animal origin in several studies using robust molecular methods (Oloya et al., 2009) and it is possible that the resistance to aztreonam seen in this study could have originated from humans.Recent studies have shown that enteric bacteria isolates from humans and animals had the same antimicrobial resistance determinants (Johnson et al., 2007;Skov et al., 2007).This also confirms the current understanding that antimicrobial use is the most important factor that is responsible for inducing resistance although transmission of already resistant bacteria also plays a significant role.It is therefore possible that the high resistance seen in this study is indeed due to the selective pressure exerted by the use of antibiotics in the management of bacterial infections in animals and humans in the country.Antimicrobial resistance is a natural biological phenomenon that often results from antibiotic use pressure in humans, agriculture and the widespread use of disinfectants in farm and household chores (van den Bogaard and Stobberingh, 1999).When it gets amplified many times, it results in serious public health concerns and long term shifts in resistance levels (Houndt and Ochman, 2000).Indeed antimicrobial agents are grossly misused in many developing countries leading to high selective pressure on microorganisms (Byarugaba, 2005).Beta-lactams, tetracycline and erythromycin the most widely used antimicrobials in poor countries, not only in food animals but also in humans because they are cheap and readily available (Mitema et al., 2001).This correlates with the observation made in this study of high resistance seen in some of the commonly used antibiotics.
Significant differences were observed between bacteria isolated from different food animal species in the present study.This was both seen for Enterococci isolates and E. coli isolates.This may be attributed to the different formulations, the mode of administration, the different types of antibiotics used and the amounts of antibiotics used in the different food animal species in the country.In the present study, chicken generally showed the highest resistance followed by swine with least resistance observed in small ruminants.It is known that there is a lot of antibiotic use in chickens and the mode of administration in water or food usually exposes the whole flock to the antibiotics considering the fact that there are lots of erythromycin and tetracycline formulations available on the market (Byarugaba, 2004).
The use of antibiotics, whether for prophylaxis or chemotherapy, does not only affect the pathogenic bacteria but also the commensal bacteria.This maintains a pool of resistant bacteria with a pool of resistance genes in the population which further contributes to the general increase and dissemination of bacterial resistance and can be a source of resistance genes for pathogens (Abatih et al., 2009;Blake et al., 2003).The level of intensification in both chickens and swine production dictates frequent use of antibiotics and many bacteria become resistant in order to survive (Smith et al., 2007).The application of some antibiotics in feed or water in these management systems also leads to development of resistance in bacteria from the entire flock, leading to a high overall population of resistant bacteria.Whereas it is known that there is little use of antibiotics in small ruminants in Uganda, thus explaining the low resistance observed, there was still some resistance observed which may be likely due to acquisition of already resistant bacteria from the environment or from other animals with which they are grazed.The largest small ruminant production in Uganda currently is among the pastoralist communities who keep them together with cattle.Uganda animal keepers access without prescription all types of antibiotics in any amount they wish and many keep them on their farms and administer them usually to animals whenever they wish (Byarugaba, 2004).
The comparison of resistance between the Grampositive resistance indicator (Enterococci) and Gramnegative resistance indicator (E.coli) further confirmed that all commensal bacteria are affected.The present study demonstrated that there were no significant Byarugaba et al. 1547 differences for some of the antibiotics for either Enterococci or E. coli isolated from the same hosts.However, differences were observed in the levels of resistance between Enterococci and E. coli in agreement with previous studies which have reported that bacterial species more readily develop (or acquire) resistance than other microorganisms, when exposed to apparently similar selective pressures (Ellner et al., 1987).Whether particular organisms become resistant to a particular antimicrobial agent depends on many factors which may include extent of exposure to antibiotics, (Parry, 1989).Today, the emergence of bacterial strains which display resistance to a variety of drugs is a major cause of failure of treatment of infections worldwide and a serious concern to animal and public health (Vergidis and Falagas, 2008).Multi-drug resistance was demonstrated in both Enterococci and E. coli isolates with 60 and 77.6%, respectively, in the present study.Moreover, some isolates were resistant to a large number of drugs; for example, about 11% of the Enterococci were resistant to seven or more drugs while about 4 were resistant to all the eight drugs tested.For E. coli, an even a worse situation was demonstrated with about 1% being resistant to 10 or more drugs while 35% were resistant to 6 or more drugs.Multi-drug resistance is a serious challenge to disease treatment with possibility of a complete treatment failure occurring and is occurring with much more frequency (Besser et al., 2000).This phenomenon has been reported in other bacteria and is more of a rule rather an exception (Spera and Farber, 1994;Savage 2001; Lee et al., 2002;Dargatz and Traub-Dargatz, 2004;Gebreyes and Thakur, 2005;Davis et al., 2007).It is often mediated by genetic mobile elements such as plasmids, transposons and intergrons that carry resistance gene cassettes with multiple resistance genes (Andremont, 2003;Byarugaba, 2009).The high level of multi-drug resistance observed in the commensal bacteria of food animal origin poses serious threats to both animal and public health has been reported (Wegener et al., 1999) and they can transfer the genetic determinants of resistance with a lot of ease.

Figure 1 .
Figure 1.Levels and patterns of resistance among Enterococci spp.isolated from different food animals.

Figure 2 .
Figure 2. Levels and patterns of resistance among E .coliisolated from different food animals.

Figure 3 .
Figure 3.Comparison of resistance levels between E. coli and Enterococci spp.for different antibiotics.

Table 1 .
Distribution of bacteria isolated from the different food animal species.

Table 2 .
Overall resistance of E. coli and Enterococci from all the food animal species against different antibiotics.

Table 3 .
Multi-drug resistance among the Enterococci isolates.

Table 4 .
Comparison of the resistance proportions to antibiotics for Enterococci spp. in the different food animal species.

Table 6 .
Comparison of resistance proportions to antibiotics for E. coli isolated from the different food animal species.

Table 7 .
A comparison of resistances levels to antibiotics between E. coli and Enterococci spp.isolated from different food animals.