Antimicrobial Susceptibility and Prevalence of bla and qnr Genes in Salmonella Enterica Isolated from Slaughtered Pork

Background: Salmonella enterica is known as one of the most common foodborne pathogens worldwide. While salmonellosis is usually self-limiting, severe infections may require antimicrobial therapy. However, increasing resistance of Salmonella to antimicrobials, particularly uoroquinolones and cephalosporins, is of utmost concern. The present study aimed to investigate the antimicrobial susceptibility of S. enterica isolated from pork, the major contributor in Philippine livestock production. Results: Our results show that 61.2% of the isolates carried antimicrobial resistance genes qnrS and bla TEM . While qnrA (12.9%) and qnrB (39.3%) were found less frequently, co-carriage of bla TEM and one to three qnr subtypes was observed in 45.5% of the isolates. Co-carriage of bla TEM and bla CTX-M was also observed in 3.9% of the isolates. Antimicrobial susceptibility testing revealed that majority of the isolates were non-susceptible to ampicillin and trimethoprim/sulfamethoxazole, and 13.5% of the isolates were multidrug-resistant. Conclusions: High prevalence rates of S. enterica carrying antimicrobial resistance genes (ARG), specically the presence of isolates co-carrying resistance to both ß-lactams and uoroquinolones, raise a concern on antimicrobial use in the Philippine hog industry and on possible transmission of ARG to other bacteria.


Background
Salmonella infections, or salmonellosis, are commonly acquired through consumption of contaminated food of animal origin. In the Philippines, Salmonella enterica was shown to be the leading cause of foodborne disease outbreaks from 1995-2018 [1,2]. While the disease is usually self-limiting, it may require antimicrobial therapy when the infection becomes invasive. The uoroquinolone cipro oxacin and the extended-spectrum cephalosporin (ESC) ceftriaxone are the current treatments of choice because the emergence of antimicrobial resistance (AMR) has rendered several drugs such as ampicillin, chloramphenicol, and trimethoprim/sulfamethoxazole obsolete in salmonellosis therapy [3,4].
Resistance to ß-lactams, such as ESCs, is most commonly attributed to the bla genes of subtypes TEM, SHV, and CTX-M which encode for ß-lactamases that hydrolyze the ß-lactam ring, thereby rendering the drug inactive [5,6]. In contrast to ß-lactam resistance, uoroquinolone resistance is typically attributed to chromosomal mutations in the quinolone targets DNA gyrase and topoisomerase IV, and overexpression of e ux pumps that reduce drug accumulation [7]. However, plasmid-mediated quinolone resistance (PMQR), such as qnr genes, may also occur. These genes are broadly distributed worldwide and are commonly found in association with genes encoding for ß-lactamases [8][9][10][11]. Consequently, bla and qnr genes have been increasingly found in bacteria isolated from livestock animals [8,9,[12][13][14][15][16][17]. If motile, resistance determinants may accelerate the spread of AMR when these are taken up by non-pathogenic or pathogenic bacteria alike.
There is evidence that substantial use of antimicrobials in food-producing animals may drive the emergence of drug-resistant strains [12,18,19]. While the use of certain antimicrobials such as nitrofurans and chloramphenicol has been banned in livestock production in several parts of the world, AMR in agriculture remains a global challenge [12,19]. Monitoring AMR development in livestock and meat allows early detection of AMR emergence and prevalence [20] which can be used to design interventions to improve antimicrobial therapy and reduce resistance selection pressure [21,22]. This is generally accomplished by antimicrobial susceptibility testing (AST) and detection of antimicrobial resistance genes (ARG).
In the Philippines, pork makes up the majority of livestock production and amounts to a 3.8M USD industry [23]. The country's rapidly growing population is expected to further increase pork consumption and production. If left unchecked, AMR may lead to challenges in food production, food security, food safety, economic losses to the hog industry, and AMR spillover to the surrounding environment [12,18,22]. Therefore, this study aimed to investigate the antimicrobial susceptibility and prevalence of ßlactamase-encoding genes (bla CTX−M , bla SHV , and bla TEM ) and plasmid-mediated quinolone resistance (qnrA, qnrB, and qnrS) in S. enterica from slaughtered pork in Metro Manila, Philippines.

Results
A total of 178 isolates were analyzed in this study. Vitek® 2 AST revealed that the isolates were generally resistant to ß-lactams, but susceptible to quinolones. A large number were non-susceptible to ampicillin (72.5%) and trimethoprim/sulfamethoxazole (70.8%). Non-susceptibility to key drugs, ceftazidime, ceftriaxone, and cipro oxacin were observed in 8.4%, 7.9%, 15.7% of the isolates, respectively (Table 1). Multidrug resistance was observed in 24 (13.5%) isolates; most of which were non-susceptible to four classes of antimicrobial agents ( Table 2). One ESBL-producing isolate was also detected. Table 1 Non-susceptibility levels of 178 S. enterica isolates against different antimicrobial agents. Antimicrobials are classi ed into categories based on the recommendations of Magiorakos et al. [24]. Non-susceptibility to non-ESCs, cephamycins, and aminoglycosides are not shown as these antimicrobial agents are not clinically effective, although they may appear active in vitro.  and bla TEM was observed in seven isolates. For qnr genes, 12.9%, 39.3%, and 61.2% were harboring the qnrA, qnrB, and qnrS genes, respectively. Co-carriage of bla TEM and one to three qnr subtypes were found in 45.5% of the isolates (Fig. 1).

Discussion
Since it has been established that ampicillin and trimethoprim/sulfamethoxazole have become obsolete in salmonellosis therapy, high non-susceptibility rates to these antimicrobials were expected. In many countries, aminopenicillins, which include ampicillin, trimethoprim, sulfamethoxazole, and trimethoprim/sulfonamide combinations are among the most frequently used antimicrobials in livestock production [12,26]. These antimicrobials are generally administered in all phases of hog production [26]. In this study, non-susceptibilities to ampicillin and trimethoprim/sulfamethoxazole were observed in 71.9% and 70.8% of S. enterica, respectively. Phongaran et al. [13] reported that 69.0% of Salmonella isolated from hogs in Thailand were resistant to ampicillin. However, in this study, only 35.7% were resistant to trimethoprim/sulfamethoxazole. One study conducted among hogs in Vietnam reported that 36.7% of Salmonella isolates were resistant to trimethoprim/sulfamethoxazole and only 41.3% to ampicillin [27]. On the other hand, low rates of resistance (< 5%) to ESC and cipro oxacin were reported in both studies [13,27], while this present study reported rates which were slightly higher (< 10%).
In this study, multidrug resistance was observed in 13.5% of S. enterica isolates. Reports of (multidrugresistant) MDR Salmonella isolated from hogs in other Southeast Asian countries are higher (30-40%) (13,27). In other countries, even higher rates (70-80%) of MDR Salmonella isolated from pork and the pork production chain were observed [15,28]. Out of 24 MDR S. enterica isolates in the present study, 15 and 8 were non-susceptible to ESC and uoroquinolones, respectively, the current drug options in treating salmonellosis. Multidrug resistance is a challenge as it narrows down the options for antimicrobial therapy.
Majority of studies on bla genes and livestock animals in Southeast Asian countries are focused on E. coli in which bla TEM and bla CTX−M are the most frequently identi ed bla genes [12]. In Salmonella, bla TEM appears to be the most common. In India, Lalruatdiki et al. [14] observed that 30% of Salmonella isolated from a pig population were carrying bla TEM , and 10% were carrying bla CTX−M . Co-carriage of bla CTX−M and bla TEM has also been observed in extended-spectrum ß-lactamase (ESBL)-producing Salmonella from pigs [14,29]. In the present study, co-carriage of bla TEM and bla CTX−M was found in seven (3.9%) isolates.
However, none of these isolates were ESBL-producing which could suggest that these are only carrier of silent bla genes. The only ESBL-producing Salmonella in this study was carrying only bla TEM . While most bla TEM in the study possibly confer only broad-spectrum ß-lactam resistance considering the high rates of non-susceptibility to ampicillin, its presence in combination with other resistance determinants could render an isolate multidrug-resistant.
We report in this study that 71.3% of S. enterica isolates harbored PMQR. The genes qnrA, qnrB, and qnrS were observed in 12.9%, 39.3%, and 61.2% of the isolates, respectively. While Qnr proteins offer only low resistance against quinolones, the high incidence of PMQR may be a cause for concern since it has been shown to broaden the mutant selection window in bacteria [7]. Lin et al. [16] demonstrated that cipro oxacin resistance conferred by PMQR is even comparable to that of quinolone target mutations. Prevalence rates of qnr genes appear to vary among samples and geographical locations. Cameron-Veas et al. [15] reported that 15% of S. enterica isolated from a pork production chain in Brazil were carrying qnrB, and none were carrying qnrA and qnrS. A separate study in China reported the prevalence of qnrA (0%), qnrB (16%), and qnrS (66%) [16] in foodborne Salmonella. In Thailand and in Laos, Sinwat et al. [17] found only 1-8% of S. enterica isolated from pork to be carrying the same qnr genes. This highlights the importance of a national surveillance of ARG since it appears individual countries seem to have different prevalence rates.
Several studies have also reported the association of qnr genes with bla genes. One MDR Salmonella isolated from a piglet in Spain was carrying both qnrB and bla CTX−M . Moawad et al. [8] found that 33% of Salmonella from poultry and beef in Egypt were carrying qnr genes and either bla CTX−M , bla TEM , or both.
Whether qnr and bla genes reside within the same plasmid was not con rmed in either of the studies. However, Penha Filho et al. [9] recently isolated Salmonella from poultry in Brazil which carried both bla CTX−M−2 and qnrB in the same plasmid. In clinical isolates of S. enterica, E. coli, and K. pneumoniae, qnr genes have also been found within the same plasmid as that of bla TEM or bla CTX−M [10,11]. In the present study, 81 bla TEM -carrying isolates and all 9 bla CTX−M -carrying isolates were harboring one to three qnr subtypes.

Conclusions
The increasing prevalence of MDR Salmonella in livestock animals has been widely reported [13,15,27,28] and is mainly attributed to the inappropriate use of antimicrobial agents in veterinary medicine [18,19]. We report that 89.4% of S. enterica isolated from slaughtered pork were non-susceptible to at least one antimicrobial agent and 13.5% were MDR. Majority of the isolates were also harboring bla TEM which possibly encode broad-spectrum ß-lactamases, and qnrS which could facilitate emergence of mutations that target quinolone resistance. While worldwide AMR surveillance has allowed the determination of the evolution of resistance, national surveillance will allow countries to create policies that would t their needs. Generation of local information on AMR and antimicrobial consumption in the veterinary and agricultural sectors will allow the development of relevant approaches to tackle AMR [21,22]. This is highly important for low-and middle-income countries (LMICs) as strategies proven effective to work in developed countries may not be suitable for LMICs. Attention to AMR in the agricultural sector began in the Philippines only recently, and further surveillance is necessary to identify emerging resistant S. enterica in the pork production chain.

Sample collection
The study population consisted of freshly slaughtered hogs from six abattoirs across the different districts of Metro Manila, Philippines. Informed consent was obtained from the Philippine National Meat Inspection Service, hence, ethics approval was waived for this particular study. Animal slaughter and evisceration were performed according to national regulations. Informed consent was also obtained from veterinarians in charge of the abattoirs, and farm owners for sample collection. Tissue samples from hog tonsils and jejunum were collected post-slaughter and under the supervision of a veterinarian. Sample collection was performed as previously described [30]. Brie y, tissues were collected from each hog upon evisceration using sterile forceps and scissors, and then immediately transferred into sterile bags. All samples were kept chilled upon collection and during transport, and were immediately processed in the laboratory.

Bacterial Isolation And Identi cation
Bacteria were rst enriched prior to isolation as previously described [30]. Brie y, 25 g of each sample was transferred to 225 mL buffered peptone water (BPW), and incubated overnight at 35°C. Afterwards, 100 µL of pre-enriched bacterial culture in BPW was inoculated into 10 mL Rappaport-Vassiliadis broth (RVB), and then incubated overnight at 42°C for selective enrichment of S. enterica. RVB cultures were then inoculated onto brilliant green agar (BGA) and xylose lysine deoxycholate agar (XLD), and then incubated overnight at 35°C for isolation. Presumptive S. enterica were then inoculated onto nutrient agar and incubated overnight at 35°C for subsequent total DNA extraction.
Total DNA was extracted by harvesting colonies using a sterile 1 µL loop and suspending these in 100 µL TE buffer (10 mM Tris, 1 mM EDTA at pH 8.0). The suspension was boiled for 10 min, and pelleted at 6000 rpm for 5 min. The supernatant was collected and then stored at -20 °C until use. These DNA extracts were used in both PCR-based identi cation of S. enterica and detection of ARG.
Each PCR reaction for S. enterica identi cation contained 2 µL DNA, 10 pmol each of forward and reverse primers, and HiPi PCR Premix (Elpis Biotech, Daejeon, South Korea) in a nal volume of 20 µL. Ampli cation of a 244-bp region in the species-speci c invA gene was performed as previously described [31]. PCR products were subsequently analyzed via capillary electrophoresis. Salmonella enterica KCTC 2421 was used as a positive control.
Antimicrobial susceptibility testing Vitek® 2 AST was used to generate the antimicrobial susceptibility pro les of the isolates. It automatically classi es isolates into susceptible, intermediate, or resistant to a particular antimicrobial agent based on the latest breakpoints provided by the Clinical and Laboratory Standards Institute (CLSI). Multidrug resistance was de ned as non-susceptibility to at least one antimicrobial agent in three or more antimicrobial categories as recommended by Magiorakos et al. [24].
Inoculum preparation for the automated AST was followed as previously described [30]. Detection of bla and qnr genes S. enterica isolates were screened for ß-lactamase-encoding genes (bla CTX−M , bla SHV , and bla TEM ) and quinolone resistance genes (qnrA, qnrB, and qnrS) using monoplex PCR assays. The primers used are listed in Table 3  S. enterica isolates carrying bla CTX−M were subjected to further PCR assays to identify CTX-M variants.
The primers used in CTX-M variant typing are listed in Table 3. Each reaction contained 2 µL DNA, 10 pmol each of forward and reverse primers, and 6.25 µL GoTaq® Green Master Mix (Promega) in a nal volume of 12.5 µL. Ampli cation was performed as previously described. [6].
Amplicons were analyzed in 1.5% agarose gels stained either with GelRed™ Nucleic Acid Gel Stain or SYBR® Safe DNA Gel Stain (1:10,000). Amplicons were allowed to separate at 100 V for 20-30 min, and then viewed in a gel documentation system. KAPA™ Universal Ladder was used to estimate the molecular weights of the products. Informed consent was obtained from the Philippine National Meat Inspection Service, hence, ethics approval was waived for this particular study. Animal slaughter and evisceration were performed according to national regulations. Informed consent was also obtained from veterinarians in charge of the abattoirs, and farm owners for sample collection.