Salmonella enterica is a foodborne pathogen that can cause enterocolitis with diarrhea, fever and abdominal pain in humans, as major clinical manifestations (1). Typically, S. enterica causes a self-limited digestive disease, although extra-intestinal infections of medical importance have also been reported, especially in patients of the high-risk group (2–4). Thus, the severity of human salmonellosis varies according to the serotype involved and the health status of the host (4, 5).

The S. enterica subsp. enterica has multiple reservoirs in domestic animals and wildlife, and its transmission to humans is usually by consumption of contaminated foodstuff, such as eggs or meat (6, 7). Colonized chickens and pigs, as well as contaminated products derived from those animals, are major sources of human salmonellosis in most countries. Thus, it is important to investigate and survey the epidemiology of this pathogen in those animal species for public health and poultry and pig husbandry (8). In a high proportion of cases, the meat is contaminated with feces during slaughter or dressing line, which constitutes the main risk factor for Salmonella to enter into the food chain (9). However, the food chain is not the only transmission pathway, being the direct animal-human contact a route that has been increasingly reported worldwide (10–12).

The expression of several virulence genes during the host-pathogen interaction allows the activation of adhesion and invasion mechanisms which determines the course of infection. These factors include flagella, capsule, plasmids, adhesins, and type 3 secretion systems (T3SS), among others. These virulence factors could enhance the adaptive capacity of some serotypes, representing Salmonella pathogenic mechanisms to infect, survive and establish the infection in the host (13, 14).

An additional factor of current concern is the emergence of antimicrobial resistance (AMR) in different S. enterica serotypes. In general, self-limited salmonellosis does not require antibiotic treatment; however, in patients with the invasive disease, it may be necessary to use antibiotics, especially cephalosporins and fluoroquinolones. Thus, AMR represent a risk for patients and for public health because therapies are less effective and the genes encoding resistance are positively selected and disseminated through mobilizable elements between bacterial strains of self and different species and sources (15, 16).

There are several zoonotic Salmonella serotypes that have a wide host range and cause infection and disease in humans, wildlife and domestic animals (4, 5). Thus, reporting the Salmonella serotypes involved in an outbreak turns essential for conducting an outbreak investigation and epidemiological surveillance (17). The traditional method of serotyping Salmonella isolates follow the Kauffmann-White scheme, based on the typification by serological discrimination, through the agglutination of somatic (O) and flagellar (H) antigens, and the capsular Vi antigen; describing more than 2,650 serotypes with diverse epidemiology and geographic distribution (18, 19). Depending on the serotype, this method could be slow, laborious and costly, which justifies the study of other faster and cost-effective S. enterica typing methodologies, such as the multiplex PCR (20, 21).

In South America, there is limited surveillance data on non-typhoidal Salmonella strains in food producing animals. Therefore, the objective of this study was to identify S. enterica serotypes in pigs and chickens from commercials farms located in central Chile and to characterize their AMR phenotypes, serotypes and virulotypes in order to provide relevant data for surveillance programs at the national and international level.

Understanding the epidemiology of S. enterica in different animal reservoirs is essential to achieve sanitary control of this pathogen and reduce its transmission to humans. Our results show several differences between strains isolated from two food-producing animal species, such as pigs and chickens. First, the rate of S. enterica isolation in feces of pigs was 9% and 19% in chickens. At national level, the isolation rate from pigs is higher than that reported previously by San Martín et al. (41), but lower than that reported by Lapierre et al. (42), with rates of 4.5 and 11.9% respectively. In the case of isolation rates from poultry, and as far as we know, to date there are no published studies. On the other hand, and at international level, our isolation rates differ from those of other authors. In this context, Ishihara et al. (43) reported a prevalence of Salmonella spp. of 2.9% and 36.1% in pigs and broiler chickens, respectively. More recently, Chen et al. (44) in China registered isolation rates of 44 and 9% in pork and chickens, respectively. These differences in detection rates may be due to variations in biosecurity measures for controlling bacterial contamination during the productive cycle of animals and processing of their carcasses. Moreover, the increased use of low-risk and certified feed ingredients in Chile could also contribute to lower residual contamination in animals, facilitating the traceability for detection of other sources of infection. However, differences between studies can also be attributed to specific sample sizes and isolation methods, the local environmental conditions and husbandry practices, among other factors. Regarding methodologies for the isolation of Salmonella from animal feces, the ability of the laboratory method to recover Salmonella from a sample containing a low concentration of this bacterium can be affected by biological factors such as competing microbiota and technical factors such as the culture media used, the volume of the feces and the pre-enrichment of the samples. We performed the isolation based on that described by Jensen et al. (24). These authors propose to add novobiocin to the BPW in the pre-enrichment stage. The authors point out that this method is more sensitive and that it increases the growth of Salmonella in the samples, although they also point out that 13 positive samples were lost. The conclusion of the authors is that the increased growth of Salmonella in novobiocin medium is due to a reduction of competitive microorganisms. So, it is possible that with the addition of novobiocin to BPW, we were not able to isolate Salmonella from all contaminated samples, and probably the actual positivity rate in our samples was higher. However, it is important to note that the animals sampled were healthy animals and therefore could excrete a low concentration of Salmonella, which might not be isolated from a sample with competitive microbiota (45). Thus, there are advantages and disadvantages to the addition of novobiocin in BPW for Salmonella isolation that should be evaluated by laboratories according to their objectives.

On the other hand, the most frequent serotype found in pig farms was S. Typhimurium, while in chicken farms was S. Infantis, in agreement with other reports from United States and the European Union (46–50). In particular, S. Infantis is considered an emerging serotype in the poultry industry across the world, associated with the acquisition of a pESI-like megaplasmid conferring MDR phenotypes (9, 51–54). Furthermore, both S. Infantis and S. Typhimurium have been ranked among the top 10 most frequently Salmonella serotypes involved in human infections worldwide, representing a food safety and public health hazard (55). In Chile, these serotypes occupy the 2^nd and 3^rd position in frequency, respectively, after S. Enteritidis (56); representing in aggregate around 75% of all clinical cases during last years.

The agglutination techniques for Salmonella serotyping does not have the required discriminative power to describe transmission chains during outbreaks investigations. Therefore, other quick, cost-effective resolution procedures have been studied, such as the multiplex PCR technique. In this work, this assay was performed for genotyping Salmonella isolates, detecting 21 different genotypic patterns in 46 strains isolated from pigs, and 31 distinct genotypic patterns in 57 strains from poultry. This means that genomic polymorphisms harbored by circulating S. enterica serotypes are giving rise to a high genotypic diversity of Salmonella isolates in both hosts, with Gini coefficients lower than 0.4 in all of them (Table 1). Apparently, there are few Salmonella serotypes very well adapted for survival into industrial environments, which infect animals after having been transmitted from various sources. In this context, industry-adapted S. enterica strains show high biofilm production, abiotic stress adaptation and extracellular adhesion; as well as adaptation to their poultry hosts (57–61). This scenario contrasts with the previous results of Kim et al. (17), Beaubrun et al. (20), and Beaubrun et al. (21), which suggested a unique genotype-serotype relationship. The genetic variability observed in our strains might be associated with several contributing factors, such as the origin of the animals and their food (62), the contamination in the premises by direct and/or indirect contact with humans (62), domestic animals or wild birds (63), or the presence of other carriers acting as vehicles of S. enterica strains (64, 65). For this reason, the inclusion of routine testing and categorization of animal herds on its S. enterica infection status are important components of integrated programs focused at reducing its prevalence in the food chain and the risks for animal and public health. Additionally, the observed diversity establishes the need of future analyses to determine the capability of multiplex PCR for supporting studies like that, and its potentiality for complementing the traditional serotyping through its discriminatory ability at a sub-serotype level. In this regard, it is important to highlight that the same genotype was not shared between serotypes, maintaining the specificity of the method. Furthermore, this molecular typing method suggests that pig and chicken industries are not sharing isolates (Figures 1, 2), being affected by independent infection sources which would require specific efforts to characterize Salmonella transmission chains and the implementation of their control methods.

A variety of virulence factors have been shown to play different roles in the pathogenesis of Salmonella infections. To perform virulotyping in Salmonella it is important to choose genes that are not present in 100% of the strains, based on the literature, in order to determine the variability of the isolated strains. Regarding the detection of virulence genes, we found 25 different virulotypes in chicken and 12 different virulotypes in pig. The most detected genes were pagK (68%) and SEN1417 (42.1%). In fact, the pagK gene, as the only amplified gene, was the most frequent virulotype (n = 12; 21%), corresponding to serotypes S. Infantis (n = 10) and S. Enteritidis (n = 2). The high frequency of pagK in pigs (95.6%) and chicken (68.4%) suggests its association with an important function during bacterial colonization of animal hosts, codifying a translocated factor into the cytosol of host cells (66), which also participates in biofilm formation (9).

A striking result is the difference in the number of antibiotic resistant strains in both food animal industries. The strains isolated from pigs showed low AMR levels and no critical drug resistances were detected. On the contrary, S. enterica strains isolated from poultry were much more resistant, including resistance against critical antibiotics, especially to cephalosporins. In Chile, the commercial pig and chicken farms have developed a high technological, biosecurity and sanitary related standards. They are intensive, vertically integrated, with restrictions on the use of antibiotics and homogeneity on their sanitary and productive management. Then, the contrasting findings in the AMR profiles could suggest a serotype effect, especially associated to S. Infantis, or a management effect, related to differences in the antibiotic usage in both type of productions. In this regard, emerging strains of S. Infantis carrying the megaplasmid pESI show a MDR phenotype; and have been described in several countries, including Chile (67–69). It is therefore very likely that the S. Infantis strains, mostly isolated from chickens, contains this megaplasmid and, as a consequence, express a MDR phenotype and develop a higher virulence than non-MDR strains (70). In the case of pigs, four strains were MDR, one of which showed simultaneous resistance against 10 antibiotics. In the case of S. enterica strains isolated from chickens, 33 strains presented a MDR phenotype, with two of them resistant to 10 antibiotics, two to 11 antibiotics, and even two to 12 antibiotics. In this study, a high MAR rate (>0.50) was observed in 9 (15.8%) strains isolated from chickens, indicating that Salmonella strains isolated from chickens express higher levels of multi-resistance compared to Salmonella strains isolated from pigs.

In Chile, despite regulations for the use of antibiotics in food-producing animals have been implemented (71), the most common antimicrobials used in pig or broiler chicken production are unknown. Moreover, ceftiofur, which is a third-generation cephalosporin, is permitted in these species. Consequently, bacteria isolated from chickens expressed resistance against critical antibiotics indicated by the WHO and OIE, such as cephalosporins, especially ceftiofur (39%), and fosfomycin (33%) (36, 72). In this context, Lai et al. (73) in China, also described high (42%) ceftiofur resistance rates in Salmonella strains isolated from chickens, pigs and ducks. The importance of monitoring the resistance to ceftiofur relies on its similar mechanism of action with ceftriaxone, a cephalosporin often used to treat Salmonella infections in children (74). Therefore, the increased rate of ceftiofur resistant Salmonella strains of animal origin may has important public health implications, limiting therapeutic alternatives in human medicine. In South America, there is limited information about AMR associated to food-producing animals. Among the scarce information available, Vinueza-Burgos et al. (75) in Ecuador reported AMR to critical drugs, including 5.8% to ceftazidime, 98% to nalidixic acid, and 94% to ciprofloxacin. In our study, among the Salmonella strains isolated from chickens, we found 15% of enrofloxacin resistant strains, and similarly to Vinueza-Burgos et al. (75), our isolates were also highly resistant to nalidix acid (80%). The high percentage of resistance to this antibiotic could be reflected in the future with higher AMR to fluoroquinolones, such as enrofloxacin, a risk that deserves a close surveillance. Additionally, 33% of isolates from chickens were resistant to fosfomycin, which is a recently reintroduced antibiotic for the treatment of acute urinary tract infections in humans (76). Several studies in Enterobacteriales have reported the simultaneous presence of fosfomycin resistance genes with β-lactamases encoding genes (77, 78). In this work 17.5% of isolates showed both phenotypes of resistance, a situation that can also be considered of concern for public health.

In summary, this study suggests that in pig farms located in central Chile the predominant S. enterica serotype is S. Typhimurium and in chicken farms is S. Infantis, with a genotypic diversity implying different sources and transmission chains. Isolates from poultry showed more resistance phenotypes against antibiotics, with critical importance those associated with cephalosporins, fosfomycin and fluoroquinolones. The presence of genes, related with virulence and phenotypic antibiotic resistance, indicate a convergence of virulence and resistance determinants, highlighting that efforts should be focused on implementing strict farm-to-table surveillance programs for this pathogen in industrial pig and poultry farms, in order to prevent outbreaks in the population.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author/s.

Ethical review and approval was not required for the animal study because Study was performed with resistant bacteria. Samples where taken by authorized veterinarians working in the farms. Researchers do not perform any animal studies.

PR and LL contributed to the conception, design of the study, and wrote the first draft of the manuscript. PR, LL, LS, and JC contributed with resources to the study. JG, MB, and TA performed the laboratory analyses. PR, LL, and JC performed the statistical analysis. LL, NG, LS, and JC wrote sections of the manuscript. All authors contributed to the article and approved the submitted version.

This work was self-founded by JC and LL.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.