Antimicrobial Resistance and Extended Spectrum β-Lactamases of Escherichia coli isolates from Raw Meat in Greater Accra Region, Ghana

Background Typically, raw meat can be contaminated with antimicrobial resistant pathogens at unhygienic slaughter and sale points. Consumption of meat contaminated with antimicrobial resistant E. coli is associated with grave health care consequences. The aim of this study was to determine the microbial quality of raw meat, the antimicrobial susceptibility and Extended Spectrum Beta Lactamase (ESBL) production in E. coli isolates from raw meat. Methods We collected 205 meat samples from beef (n=81), chevon (n=108) and mutton (n=16) at three slaughter sites in the Greater Accra Region of Ghana. We performed Total Plate Counts (TPC) and Total Coliform Counts (TCC) and isolated E. coli using standard bacteriological methods. Presumptive isolates were conrmed by matrix-assisted laser desorption ionization-time of ight mass spectrometry (MALDI-TOF-MS). We assessed the susceptibility to 11 antibiotics using the Kirby Bauer disk diffusion method and assessed ESBL production using the combination disk test. PCR amplication of associated resistant genes bla TEM, bla SHV, and bla CTX-M was performed for presumptive ESBL isolates. Results Total Plate Counts exceeded the acceptable limit of 5.0 log CFU/ cm 2 in 60.5% (124/205) of raw meat samples. Total Coliform Counts in 70.7% (145/205) of samples were in excess of the acceptable limit of 2.5 log CFU/cm 2 . E. coli was detected in about half of raw meat samples (48%), ranging from 9.5% -79.0% among the slaughter sites. Isolates were susceptible to meropenem (100%), ceftriaxone (99%), cefotaxime (98%), chloramphenicol (97%), gentamycin (97%), ciprooxacin (92%) and amikacin (92%), but resistant to ampicillin (57%), tetracycline (45%), sulfamethoxazole-trimethoprim (21%) and cefuroxime (17%).


Background
The emergence of antibiotic resistance in bacteria commonly found in food animals has garnered attention globally for its potential contribution to human colonization with antibiotic resistant bacteria. Antibiotics are used in food animals for therapeutic and non-therapeutic purposes. The continuous use of antibiotics in food animal production is cited as a major determinant for carriage of antibiotic resistant bacteria in food animals (1) .A proportion of antimicrobial classes are used in both animals and humans, creating the need to monitor the spread of resistant bacteria from animals to humans at all stages of the transmission pathway (2).
Food products of animal origin contaminated with drug resistant bacteria may provide a direct route for human colonization (3). Meat is an important part of the Ghanaian diet and is often sourced from open markets and cold stores (4). Although selective pressure due to antibiotic use in primary production is considered as a major source of antibiotic resistant bacteria in livestock products, sanitary conditions at slaughter, sale and processing points may affect the pattern and intensity of spread along the food chain (5).
Escherichia coli is a common colonizer of the intestinal tract of humans and animals. It is frequently isolated as the causative organism in urinary tract infections, neonatal meningitis and bacteremia in humans (6). The presence of pathogenic strains of E. coli in meat and dairy products has been associated with foodborne disease outbreaks in humans (7).
Extended spectrum beta lactamase (ESBL), an enzyme produced by some gram negative bacteria, pose a peculiar challenge for treatment as they are resistant to penicillins, cephalosporins and monobactams (8). ESBL-producing bacteria such as E. coli have been reported in patients with urinary tract infections worldwide (9). The contribution of the food chain to the occurrence of ESBL bacteria in humans is widely debated with fecal-oral and nosocomial transmissions largely in uencing its emergence and spread in humans (10).
Monitoring the trend of spread of ESBL positive E. coli through the food chain is necessary in settings where humans have routine contact with livestock and their products, because of the increased risk of spread of resistance genes in these interactions.
In Ghana, it has been found that, resistance in enterobacteriaceae to commonly used antimicrobial agents is widespread (11). Research evidence on microbial contamination of meat and their antibiotic resistance patterns is limited and thus, it is di cult to assess the risk posed to human health through the food chain. Yet, such evidence is critical for guiding future antimicrobial therapy. The aim of this study was to determine the level of microbial contamination of raw meat, presence of E. coli and their resistance to commonly prescribed antibiotics for human infections. This study also detected the occurrence of ESBL production among E. coli isolates recovered from raw meat.

Study sites and sample collection
In this study, we collected raw meat samples from three slaughter sites in the Greater Accra region of Ghana. The sites consisted of a public slaughterhouse (PS), a slaughter slab (SS) and a privately-owned slaughterhouse (PO). Livestock such as cattle, sheep and goats were slaughtered daily at all three sites. Butchers operating at these slaughter sites obtained their livestock mainly from the three northern regions of Ghana. All three sites were points of meat sale and had the basic modules of production, which included a slaughter oor and lairage/holding pen. All parts of carcasses are sold as meat in large cuts and small cuts at these sites. The daily throughputs of livestock slaughtered at these sites ranged from 30-150 for large stock (cattle) and 20-300 for small stock (sheep and goats). Livestock at each site were selected randomly on predetermined sampling days. A total of 205 surface swab samples were collected from beef (n=81), chevon (n=108) and mutton (n=16) at the three slaughter sites (PS=63, SS=79, PO=63). Sample collection was done immediately after carcass dressing with a sterile cotton swab and a sterile template of size 100 cm 2 for cattle and 25 cm 2 for goat and sheep carcasses. For each beef carcass, surface swabs were collected from the thigh, brisket and ank and pooled into a single tube containing 10ml of Buffered Peptone Water (BPW). Swabs from goat and sheep carcasses was obtained from thigh, brisket and mid-loin (12). All samples were transported to the laboratory within 2 hours of sample collection for processing.

Enumeration of bacteria
Total plate counts and total coliform counts were performed for all samples using the pour plate method. Serial dilutions of ten-fold units of each sample were plated on Plate Count Agar (Oxoid) and incubated for 48 hours at 30°C. Following incubation, plates with colonies ranging from 30-300 were counted and expressed in CFU/cm 2 . The total coliform count was performed in the same way using Brilliance E. coli/ Coliform selective agar (Oxoid). The results of both counts were interpreted using guidelines set by the Ghana Standards Authority and the European Commission Regulation on the microbial criteria for food stuffs (13,14). Limits to total coliform counts were not speci ed in these documents hence limits set for Enterobacteriaceae counts in both documents were adopted for total coliform counts.

Identi cation of E. coli in meat
To isolate E. coli, samples were pre-enriched in Brain Heart Infusion (BHI) at 37°C for a period of 24 hours. Ten (10µl) microliters of each sample were then plated on MacConkey agar and incubated overnight at 37°C. We identi ed E. coli by colonial morphology and con rmed the isolates using Matrix Assisted Laser Desorption/Ionization -Time of Flight Mass Spectrometry (MALDI-TOF MS). Colonies from fresh overnight cultures were spotted on the MALDI-TOF MS target plate. One (1) µl of formic acid was then added and allowed to dry for 15 min. One (1) µl of matrix preparation was placed on each sample and left to dry for a further 15 minutes. MALDI-TOF MS was then conducted and ionization peaks (spectra) generated were matched against the integrated reference library of the MALDI system for con rmation of species of bacteria.

Phenotypic and genotypic detection of ESBL production
The combination disk method was used to identify ESBL producing E. coli. Brie y, ceftazidime disks and ceftazidime combined with clavulanic acid disks were positioned 30 mm apart on the inoculated plates, and incubated overnight at a temperature of 37°C. Isolates were classi ed as ESBL positive if the difference between the inhibition zone diameter of ceftazidime combined with clavulanic acid and diameter of the inhibition zone for the ceftazidime-only disk was ≥5mm (17). Cefotaxime disks and cefotaxime combined with clavulanic acid disks were used concurrently with ceftazidime for con rmation of ESBL production. DNA was extracted from presumptive colonies for the genotypic detection of ESBL production.
Bacterial DNA extraction consisted of placing pure colonies of overnight growth in 1ml of distilled water. The mixture was boiled for 10 minutes and centrifuged for 5 minutes, at 1000 rotations per minute. Using the supernatant as a DNA template, PCR was done to detect ESBL resistance genes (bla TEM, bla SHV, and bla CTX-M ).
The nal volume of 25µl contained 2µl DNA template, 10mM of each primer (Table 1) (18), PCR grade water (10ul) and multiplex PCR master mix (13ul) (QIAGEN). Multiplex PCR was carried out in a thermal cycler with the following cycling conditions: 95°C for 5 minutes, 35 cycles of 95°C for 30 seconds, 60°C for 30 seconds, 72°C for 2 minutes, and a nal extension lap at 72°C for 10 minutes.

Statistical analysis
Descriptive statistics were performed on carriage rates of E. coli and their resistance patterns. The Chi-square test was used to assess differences in the proportion of E .coli detected in raw meat between the three slaughter sites. A Kruskal-Wallis test was performed to determine if there were signi cant differences in microbial counts in raw meat among the slaughter sites. Statistical analyses were performed at 95% con dence level using STATA version 15.0.

Microbial contamination of raw meat
Overall, 124 (60.5%) samples from raw meat in this study exceeded the maximum limit of 5.0 log CFU/cm 2 , for total aerobic counts for cattle, sheep and goat carcasses set by the Ghana Standards Authority and the European Commission Regulation (EC) on the microbial criteria for food stuffs. Total Plate Counts for all 205 samples ranged from 2.86 log CFU/cm 2 to 7.26 log CFU/cm 2 with a median of 5.28 log CFU/cm 2 ( Table 2).
More than half of all samples (70.7%) had coliform counts exceeding the acceptable limit of 2.5 log CFU/cm 2 .
Total coliform counts ranged from 0.11 log CFU/cm 2 to 5.76 log CFU/cm 2 with a median of 3.12 log CFU/cm 2 ( Table 2). The proportion of samples exceeding the maximum limit for coliform counts was 98.4 % at PS, 97.5% at SS and 11.1 % at PO. The TPC and TCC distribution among the three slaughter sites is shown in Figures 1 and 2.
The Kruskal-Wallis test showed that there was a statistically signi cant difference in total plate counts between the three slaughter sites, χ 2 (2) =73.81, p=0.0001. The distribution of coliforms in samples from the three sites were also found to differ signi cantly, χ 2 (2) =110.94, p=0.0001.

Discussion
The study ndings show that raw meat, ready for sale at the selected study sites in most cases (> 60%) had microbial loads exceeding the acceptable limit and thus, were of poor quality. The majority of slaughtering in Ghana is conducted at slaughterhouses and slabs with sub-standard conditions of slaughter and minimal oversight from food safety authorities (19). The absence of hygiene standards at points of slaughter has been linked to increased microbial contamination of carcasses (20). High coliform and plate counts observed at the Public Slaughterhouse (PS) and Slaughter Slab (SS) may stem from the introduction of contaminants from oor surfaces, hides and faeces of livestock in the absence of standard operating procedures that ensure sanitary standards are met in the slaughter environment. Contamination of meat in such situations may also be exacerbated by inadequate slaughter infrastructure that ensures that work areas are properly segregated (21). The consequence for human health lies in the potential exposure of consumers to disease causing organisms and carriage of drug resistant bacteria (22). E. coli is commonly used as an indicator organism for assessing food and water hygiene (23). Although E. coli was detected in close to half of raw meat sampled, wide-ranging proportions of E .coli contamination was observed at the slaughter sites (9.5-79%). Previous studies on E .coli in beef from retail joints in Northern Ghana similarly reported wide-ranging values of 0-100% (24). While high E. coli contamination rates often indicate poor slaughter hygiene, the presence of other environmental contaminants in heavily contaminated settings may reduce recovery rates of E. coli. Slaughter slabs have less adequate service modules in comparison to slaughterhouses in Ghana. Privately owned slaughterhouses often have better slaughter infrastructure and compliance to hygienic practices in an effort to maintain customer satisfaction. This phenomenon may account for the low levels of E.coli contamination (9.5%) observed for the Privately Owned Slaughterhouse (PO).
The highest rate of resistance in E .coli was observed for ampicillin (57%). This follows a commonly reported pattern of β lactam antibiotic resistance rates of more than 50% in foods of animal origin (25,26). While these observations were attributed to the common use of ampicillin and penicillin derivatives in food animals, the scale of resistance is largely dependent on the food animal type and route of administration (27). Subtherapeutic doses of β lactams administered through feed and water in poultry and the pig industry may yield higher rates of resistance than in cattle where parenteral administration is common. The levels of resistance to tetracycline (44%) and SXT (17%) observed in this study have similarly been reported in E. coli from beef in Ghana (44% and 18%) (28). Tetracyclines are commonly used for therapy in humans and livestock and for growth promotion in intensive farming systems through feed (29). High transferability of tetracycline resistance determinants in gut bacteria of livestock in such settings may contribute to the observed rates of tetracycline resistance in E. coli isolated from meat, meat products and the environment (30).
We documented lower rates of resistance of E. coli to cefuroxime in comparison to isolates recovered from clinical specimens in Ghana (31). In comparison to its use in humans, second generation cephalosporins are less commonly used in food animals and are broadly approved for the treatment of mastitis in dairy cattle (32). Among SXT resistant E. coli isolates of animal origin, ampicillin and tetracycline have been identi ed in previous studies as common co-transferred resistant phenotypes (33). The MDR rate of 22% in E. coli isolates was similar to studies carried out by Saud et al on raw buffalo meat while signi cantly higher rates were observed for chicken in the same study (34). These ndings can be attributed to the routine use of antibiotics in poultry feed and less common use of oral antibiotics in large animals. The detection of the TEM gene in four E. coli isolates is comparable to studies by Sheikh et al that found a lower incidence of the TEM gene in ground beef as compared to other meat types (35). Several factors may account for the low proportion of ESBL producing E. coli recovered in this study. Studies conducted by Aldeyab et al posited that community incidence of ESBL's was signi cantly linked to amoxicillin-clavulanic acid use while hospital incidence was linked to uoroquinolone use in health care settings (36). ESBL production in enterobacteriaceae recovered from foods of animal origin may re ect the burden of resistance to certain classes of antimicrobial drugs in gut bacteria in different food animals. The presence of ESBL-producing E. coli in raw meat is a major concern for food safety as these enzymes are transferred easily amongst Enterobacteriaceae capable of causing infections in humans. The consequences thereof for human health include prolonged hospital stays, increased mortality and morbidity and increased costs (37) .
Our isolates showed high susceptibility rates to meropenem, ceftriaxone, chloramphenicol, and gentamycin.
Third generation cephalosporins and phenicols are not commonly used in livestock in Ghana, thus their e cacy has been maintained (38). Preserving the e cacy of these drugs is a necessary measure as 3rd generation cephalosporins fall in the category of critically needed antimicrobials used to treat life-threatening infections caused by E .coli in humans.
While this study provides evidence of the presence of drug resistant bacteria in raw meat, its interpretation is limited to known determinants of resistance in the primary production chain, as environmental determinants were not assessed.

Conclusions
Our study found that more than half of meat samples had unacceptable levels of microbial contamination.
Contamination with E. coli was found in 48% of meat samples with multidrug resistance observed in 22% of isolates. Meat handlers in such environments and consumers are at risk of foodborne infections from E. coli including ESBL producing E. coli that are resistant at various levels to most antibiotics in use. This calls for further investigation of the sources of meat contamination and an assessment of hygiene standards in slaughterhouses in the Greater Accra Region to improve meat quality. Enhanced surveillance in food products of animal origin is required to monitor trends in resistance patterns and promptly detect emerging resistant bacteria.  PS-Public slaughterhouse, SS-Slaughter slab, PO-Privately-owned slaughter facility Table 4. Antimicrobial resistance profile of E. coli isolates from raw meat at the selected slaughter sites in the Greater Accra region of Ghana