Prevalence and distribution of blaCTX-M, blaSHV, blaTEM genes in extended- spectrum β- lactamase- producing E. coli isolates from broiler farms in the Philippines

Background Antimicrobial resistance is a worldwide problem causing serious health threats. Escherichia coli is one of the most important bacteria that causes resistance problem. These bacteria produce an enzyme called extended-spectrum β-lactamase (ESBL) that allows it to become resistant to a wide variety of penicillins and cephalosporins. Currently, no information or published studies on ESBL-producing E.coli in broilers are available in the Philippines. This cross-sectional study was conducted to determine the prevalence and distribution of extended-spectrum β-lactamase (ESBL)-encoding genes, blaCTX-M, blaSHV, and blaTEM, among E. coli isolates from broiler farms in Luzon, Philippines. Results Results showed a farm prevalence of 66. 67%. A total of 69 (44.23%) ESBL-producing E. coli were isolated from boot swabs and cloacal swab samples from broiler farms. All major blaCTX-M groups except blaCTX-M-25 group were identified in the isolates. The most prevalent group was blaCTX-M-1, 72.46% (CI: 60.38–82.54%), followed by blaCTX-M-2, blaCTX-M-9 group and blaCTX-M-8. The blaTEM and blaSHV genes were identified in 57.97 and 27.54% of isolates, respectively. The blaCTX-M and blaTEM were the most common gene combinations (33.33%). Coexistence of blaCTX-M types was observed in 50 (73.53%) isolates. Conclusion This study shows the high prevalence, diversity of patterns and coexistence of ESBL genes in the E. coli isolates from cloacal and boot swabs from broiler farms which pose risks of possible transmission to the environment, other animals and human.


Background
Antimicrobial resistance (AMR) has become a rapidly growing public health concern worldwide. Infections from resistant bacteria are now too common, and some pathogens have even become resistant to multiple types of antibiotics. The Food and Agriculture Organization of the United Nations (FAO) estimates that around 500,000 human deaths related to antimicrobial resistance occur each year and AMR threat is believed to become more intense by 2050 leading to an estimated 10 million deaths annually [1].
One specific AMR problem with global spread affecting both animals and humans is the extended-spectrum betalactamase (ESBL)-producing E. coli [2]. These bacteria are resistant to penicillins, cephalosporins, and aztreonam mainly due to the production of CTX-M, TEM and SHV β-lactamases which are encoded by bla CTX-M , bla SHV , and bla TEM genes, respectively. These genes can be plasmidmediated or expressed chromosomally. Among these three, CTX-M-enzymes have become the most widespread type of ESBL in animals and humans. The name CTX reflects the potent hydrolytic activity of these βlactamases against cefotaxime and they are not very closely related to TEM or SHV β-lactamases [3,4].
The presence of ESBL-producing E. coli (ESBL-EC) in food animal production systems poses public health concern since it can be transmitted to humans via the food chain [5,6]. Transmission of ESBL-EC in broiler farming was described previously wherein farm workers shared the same plasmid family and E. coli sequence type with broiler isolates [7]. Human infection due to ESBLproducing bacteria is associated with increased mortality, morbidity, high cost of hospitalization, and delay in appropriate therapy [2].
Currently, there is a lack of information on the occurrence of ESBL Enterobacteriaceae in broiler farms in the Philippines unlike the regular antimicrobial resistance surveillance program among humans in various hospitals in the country in the past decades [8][9][10]. The identification of the presence of ESBL genes in isolates from broiler farms will be useful in formulating evidencebased policies on mitigating antimicrobial resistance.
Hence, this study determined the prevalence and distribution of extended-spectrum β-lactamase-encoding genes, bla CTX-M , bla SHV , bla TEM among ESBL-EC isolates from commercial broiler farms in Luzon, Philippines.

Prevalence of ESBL E. coli in farms and samples
The prevalence of ESBL E.coli in the selected farms was 66. 67% (52/78). There is no significant difference in the farm prevalence in four provinces. A total of 69 (44.23%) ESBL-EC were isolated and these came from 47 pooled cloacal swab (60.26%) and 22 boot swab (28.21%) samples from broilers farms in Luzon, Philippines (Table 1). There is a significant difference in the prevalence between cloacal swab and boot swab samples (p < 0.05), with lower ESBL-EC isolates recovered from the latter.

Prevalence of ESBL genes
The most prevalent bla CTX-M group among broiler isolates is bla CTX-M-1 group (72.46%), followed by bla CTX-M-2 group (65.22%) and bla CTX-M-9 group (52.17%). In addition to bla CTX-M genes, bla TEM and bla SHV genes were also identified in 57.97 and 27.54% of poultry isolates, respectively. The prevalence of ESBL-EC resistance genes among cloacal and boot swab samples were summarized in Table 2.

Distribution of ESBL genotypes
The distribution of main ESBL genotypes among isolates was presented in Table 3 while the distribution patterns of bla CTX-M groups in the isolates were presented in Table 4. Coexistence of bla CTX-M types was observed in 50 (73.53%) isolates while 12 (17.65%) and 6 (8.82%) isolates had only bla CTX-M-1 and bla CTX-M-2 , respectively. A total of 9 isolates (13.04%) have genotypic resistance pattern combinations of bla CTX-M-1 , bla CTX-M-2 , bla CTX-M-9 , bla CTX-M-15 , and bla TEM while 7 isolates (10.14%) have the same genotypic pattern, with the addition of bla SHV .

Discussion
ESBL-producing E. coli (ESBL-EC) isolated from livestock and poultry animals is of public health concern since infections with these bacteria can result to treatment failure using commonly used penicillins and cephalosporins which increases the risk of mortality and delay in appropriate treatment [2]. Though ESBL-EC can be susceptible to certain cephalosporins and penicillins/β-lactamase inhibitors combinations, these drugs are rarely used as first line of treatment in E. coli infections. This is the first report of ESBL-EC in broiler farms in the Philippines with the very high prevalence as well as phenotypic and genotypic resistance patterns. A farm prevalence of 66.67% (52/78) is alarming and requires risk assessments and appropriate risk management to minimize the occurrence and spread of this resistant pathogen. With E. coli as a major opportunistic pathogen in broiler chickens and with a potential for zoonotic transfer to human, ESBL-EC represents a major risk both to poultry production and to human health [11]. Seventeen farms have positive isolates both from cloacal and boot swab samples. The rest of the farms were either positive for cloacal swabs or boot swabs.
The most prevalent ESBL encoding gene in this study is bla CTX-M which is similar to published studies in poultry [12][13][14]. In humans, however, the recent report revealed that TEM-type is more prevalent in clinical isolates from Filipinos [8] which is contrary to earlier reports wherein bla CTX-M as the most prevalent type in hospitalized patients [15]. Previous studies suggest that ESBL genotypes can vary between regions and geographical location. Therefore, it is warranted to conduct wider scope and regular surveillance study to determine the prevalence and distribution of these enzymes among broiler farms in the Philippines.
We identified bla CTX-M-1 and bla CTX-M-15 genes as the most prevalent bla CTX-M variants in this study which is similar to other reports on poultry [14] and humans [16]. Some studies also established the relationship of poultry isolates from human isolates suggesting a potential zoonotic transmission [7]. This could be the result of faecal contamination of poultry meat during slaughter, processing, selling and cooking of poultry products [5,17]. Moreover, the high prevalence of bla CTX-M-15 gene in this study has public health concern since it is the most widespread gene type of ESBL-EC in humans [18].
The detection of ESBL-EC in boot swabs in this study suggests the possible spread of the pathogen in the environment which could be a factor for a transmission in farm workers and in the community as previously reported [7,13]. In this study, a significantly lower prevalence of ESBL-EC were isolated from boot swabs compared to cloacal swabs (p < 0.05). This can be expected especially when the farms have good management practices and the floorings are kept dry [11]. Despite the lower number of ESBL-EC isolates in boot swabs, our result shows bla CTX-M-15 as the most prevalent (21.74%) genotype in boot swab samples similar to cloacal swab samples suggesting horizontal transmission to the environment. However, we were not able to establish which came first as the previous study showed that ESBL-EC -free day-old chicks can acquire the pathogen upon entry at the farm [19]. To reduce the risk of transmission, ESBL-EC should be either eliminated from poultry production or reduced the occurrence to levels with lower risk of spread to humans [11].
The bla CTX-M-2 was the third most common bla CTX-M type but the second most common group (since both bla CTX-M-1 and bla CTX-M-15 belong to the bla CTX-M-1 group) in this study. It was previously isolated in chicken meat and in healthy chickens [17,19]. The bla CTX-M-9 gene was observed in 52.17% of isolates in this study. The bla CTX-M-9 gene is widely reported in earlier studies in human infections in Europe, particularly in Spain and UK. A  [20].
Most of the isolates from poultry carry two or more bla CTX-M groups. A total of 23 (33.82%) poultry isolates have three types of bla CTX-M . In this study, co-existence of two or more CTX-M-type β-lactamases in the same strain is common. This coexistence of different types of CTX-M can be a normal scenario since they have many homologous regions which may result in the emergence of recombinant enzymes [18,21]. We speculate that multiple CTX-M types in single isolate could imply that infections caused by these isolates may be more difficult to treat since ESBL expression is more likely to occur phenotypically.
The coexistence of different β-lactamase genes within the same isolates has been reported by several investigators [14,21]. The most common ESBL genotype among our isolates was bla CTX-M and bla TEM (33.33%) which agrees with other studies [22]. The bla CTX-M gene with the bla TEM gene is the most common combination with or without bla SHV in this study which corroborates with the previous report detecting these three genotypes in poultry cloacal swab samples [23]. To our knowledge, this is the first report of high co-resistance pattern among poultry isolates in the Philippines. The presence of multiple ESBL resistance genes could result in retained resistance to β-lactamases despite the reduced expression of one or two genes.
Antimicrobial susceptibility testing showed 100% resistance to ampicillin. Studies have shown that bla TEM gene is highly prevalent in samples of chickens and human with ampicillin resistant-E. coli. [24] Colistin resistance was observed in six isolates. Colistin is considered as a last resort antibiotic for treating multi-drug resistant Enterobacteriaceae. Detection of mcr, the gene responsible for colistin resistance, in ESBL-EC from poultry samples would augment the public health importance of monitoring the antimicrobial usage in poultry farms. Likewise, very high resistance to ciprofloxacin (88.41%) was observed and this points to the possibility of ST131 circulating at high prevalence in the flocks which should be further studied. We also detected carbapenem resistance (2.90%) in our isolates. These findings warrant further investigation of the presence of carbapenem resistance genes since such resistant pathogens are among the list of World Health Organization (WHO) top priority pathogens for the development of antimicrobials. We suggest detecting the presence of a plasmidmediated bla NDM-1 gene encoding the metallo-β-lactamase    NDM-1 which hydrolyze beta-lactam antibiotics including carbapenems. Moreover, plasmids encoding for ESBL can be transferred from the E. coli poultry strains to human while carrying other antibiotic and resistance genes [25]. Some controversies are arising whether antimicrobial usage is the main contributing factor in the positivity of some broiler farms since study have shown the occurrence of ESBL-EC in farms with no or limited use of antibiotics [19,26]. Although we have not yet subjected all the PCR products for DNA sequencing, we believe that the PCR amplification of bla CTX-M -specific products alone and without sequencing usually provides sufficient evidence that a bla CTX-M gene is responsible for the expressed phenotype. However, further analysis should be conducted in bla TEM and bla SHV since sequencing is essential to discriminate between the non-ESBL parent enzymes (TEM1, TEM2, or SHV1) and different variants of TEM or SHV ESBLs (TEM3, SHV2) [27]. In addition, multilocus sequence typing and whole genome sequencing should be performed to further elucidate the chromosomal backgrounds of strains harboring these genes.
We believe that ESBL-EC at low bacterial population in the samples may have not been isolated and identified thus alternatively, we suggest that direct PCR-based detection can be employed. The universal CTX-M primer was not able to detect all positive samples (89.86%) despite showing positive results to other CTX-M group primers. In addition, there were also nine bla CTX-M -15 samples but were negative to the bla CTX-M-1 primer. We suggest the use and development of multiplex PCR to minimize such problems. Further molecular analyses could be performed to establish the relatedness of the ESBL-EC from the broiler samples to human isolates since the antimicrobial resistance genes evaluated in this study can be easily transferred to animal and human strains. In addition, further study on the isolates should be conducted to describe the connection between the presence and degree of expression of the selected genes.

Conclusions
In conclusion, results reveal the occurrence of the three major ESBL genotypes, bla CTX-M , bla TEM , and bla SHV , and the major groupings of CTX-M enzymes in E. coli isolates from cloacal and boot swab samples from broiler farms. The high prevalence, diversity of patterns and coexistence of these genotypes in the bacterial isolates is alarming. Further surveillance study in the Philippines is necessary to document the rapid emergence and spread of multi-resistant ESBL-EC in broiler production system and the food chain.

Farm selection
The four provinces in Luzon (Fig. 2) with the highest broiler production in the central region were selected.
From these provinces, a sampling frame of all broiler farms were constructed using the information on the number of existing farms obtained from the Provincial Veterinary Offices of each province. A total of 391 broiler farms were identified from four study provinces and the sample size was calculated using the following assumptions: 50% prevalence, 10% accepted error and 95% level of confidence. Using probability proportional to size sampling, a total of 78 sample farms were randomly selected from Province 1 (9 out of

Sampling and bacterial isolation
For each selected broiler farm, cloacal swabs were collected using sterile cotton swabs directly from cloaca of 10 randomly selected birds. The cloacal swab samples in each farm were pooled in Falcon tubes containing 25 ml Luria-Bertani (LB) broth (Merck, Darmstadt, Germany). A paired boot swab samples were obtained by walking along the whole length of the broiler house. Boot swab samples were placed in a 500 ml beaker containing 250 ml of LB broth for enrichment. A total of 156 samples (78 pooled cloacal swabs and 78 boot swabs) from 78 broiler farms were processed and subjected to microbiological analysis. Samples were incubated aerobically at 37°C for 18-24 h. Thereafter, a loopful (10 μl) of each enriched sample was streaked onto MacConkey agar plate (Oxoid, United Kingdom) supplemented with 1 mg/L cefotaxime and incubated aerobically at 37°C for 24 h. A replicate MacConkey agar plate without cefotaxime was also prepared for each sample. Subsequently, one bright pink colony, suggestive of lactose-fermenting bacteria and morphologically indicative of E. coli, was picked and streaked in a selective and differential medium, Eosin Methylene Blue agar plate (HiMedia, Mumbai, India) and incubated at 37°C for 24 h. The bacteria isolated from all pooled fecal and swab samples were identified.

Bacterial identification and antimicrobial susceptibility testing
Bacterial identification and antimicrobial susceptibility tests were performed through Vitek® 2 Compact (bio-Mérieux, Craponne, France), an automated microbiology system utilizing growth-based technology, using GN and AST-N261 cards, respectively. Combined disc method was also done on all presumptive ESBL-EC isolates to confirm ESBL production. Both ceftazidime (30 μg) and cefotaxime (30 μg) alone and in combination with 10 μg clavulanic acid were tested. A ≥ 5 mm increase in the zone diameter for either antimicrobial agent tested in combination with clavulanic acid versus its zone when tested alone confirmed the presence of an ESBL [28]. For quality control, E. coli ATCC 25922 (Microbiologics, Minnesota, USA) was used in both Vitek® 2 Compact and Combined disc method for the screening and confirmatory testing of ESBL-producing E. coli as well as antimicrobial susceptibility testing.

DNA extraction
For DNA extraction of bacterial isolates, the column isolation method using NucleoSpin Microbial DNA (Macherey-Nagel, Germany) was performed following manufacturer's protocol.

PCR amplification
PCR amplifications were carried out using the optimized conditions from published studies ( Table 5). All isolates were screened for target genes. The PCR assay was performed in BioRad T100 thermal cycler (BioRad, Herts, United Kingdom) individually for each primer set according to the following amplification conditions: initial denaturation at 95°C for 3 min, 35 cycles of denaturation at 94°C for 1 min, and optimized annealing temperature for each primer set (Table 5). Elongation was set at 72°C for 1 min with final elongation at 72°C for 7 min. One microliter of E.coli DNA lysate was used as template for the PCR reaction mixture containing 0.5 U DNA taq polymerase,  bla TEM TEM-F TEM-R TTGGGTGCACGAGTGGGTTA TAATTGTTGCCGGGAAGCTA 55 506 [35] bla SHV SHV-F SHV-R TCGGGCCGCGTAGGCATGAT AGCAGGGCGACAATCCCGCG 52 628 [35] 1x PCR buffer, 2 Mm MgCl 2 , 1 mM dNTP, 1 uM each of primer pair. A mixture of 3 μl of PCR products and 2 μl of loading buffer was loaded in 1.5% agarose gel and separated through electrophoresis using 0.5x TBE buffer to determine the molecular size of the amplified products per target gene. E. coli strains of ATCC 25922 and ATCC 35218 (β-lactamase-producing strain) (Microbiologics, Minnesota, USA) were used as negative and positive controls in the PCR, respectively. Purified PCR products from few representative isolates were sent to 1st Base Laboratories (Axil Scientific Pte Ltd., Singapore) for DNA sequencing analysis to confirm the target genes. Matches were analysed using Basic Local Alignment Search Tool (BLAST).

Statistical analysis
The data were analyzed descriptively. Farm prevalence was calculated as the number of farms with at least one positive isolate, either from cloacal swabs or boot swabs, over the total number of farms studied. The 95% confidence intervals were determined using exact binomial confidence limits for the proportion with a significance level (alpha) of 0.05, to test for the difference in proportions.