Salmonella Derby from pig production chain over a 10-year period: antimicrobial resistance, biofilm formation, and genetic relatedness

Abstract

The aim of this study was to evaluate 140 Salmonella Derby isolates collected over a 10-year period from porcine origins (environment, pig carcass, lymph nodes, intestinal content, and pork) for their phenotypic and genotypic antimicrobial resistance, their ability to produce biofilm, and their genetic relatedness. The minimum inhibitory concentration (MIC) was determined using microdilution broth method and antimicrobial resistance genes were investigated by PCR. The quantification of biofilm formation was performed in sterile polystyrene microtiter plates. Genetic relatedness was determined by Xba-I macrorestriction analysis. The highest frequencies of non-wildtype (nWT) populations were observed against tetracycline (75.7%), streptomycin (70%), and colistin (11.4%), whereas wildtype populations were observed against ciprofloxacin, ceftazidime, and gentamicin. The resistance genes found were bla_TEM (ampicillin), aadA variant (streptomycin/spectinomycin), tetA (tetracycline), and floR (florfenicol). On 96-well polystyrene microtiter plate, 68.6% of the isolates proved to be biofilm producers. Among 36 S. Derby isolates selected to PFGE analysis, 22 were clustered with 83.6% of similarity. Additionally, 27 isolates were clustered in 11 pulsotypes, which presented more than one strain with 100% of similarity. Most of S. Derby isolates were able to form biofilm and were classified as nWT or resistant to tetracycline, streptomycin, and colistin. PFGE allowed the identification of closely related S. Derby isolates that circulated in pig slaughterhouses and pork derived products along a decade.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42770-022-00846-7.

Introduction

Salmonella enterica is one of the most common foodborne pathogens known, and it is considered one of the leading causes of gastroenteritis in humans in several countries, including Brazil. According to the Brazilian Ministry of Health, Salmonella was responsible for 14.6% of the foodborne outbreaks in 2017 [1]. Salmonellosis is caused by the ingestion of contaminated food, and it has been established that consumption of pork may be responsible for up to 31.1% of all Salmonella infections in humans [2]. In Southern Europe, 43.6% of the confirmed cases of human salmonellosis were related to the ingestion of contaminated pork products [3]. Pigs can harbor Salmonella with no apparent symptoms of any disease. Such clinically unapparent carriers can excrete Salmonella over a long period of time, thereby contaminating their environment and acting as sources of infection for other animals and humans through contamination of pig carcasses on the slaughter line [4].

Salmonella enterica subsp. enterica serovar Derby (S. Derby) is the most frequently serovar reported from pigs and pig carcasses and the sixth most detected in humans in the European Union (EU) [5]. In China, S. Derby isolates represent the third most common serovar isolated from patients with diarrhea after S. Enteritidis and S. Typhimurium. Also, those are typically one of the most commonly isolated from pork [6, 7]. Salmonella Derby has also been found in samples from swine finishing herds, slaughterhouses, and pork in Southern Brazil [8–11].

Antimicrobial resistance has been a worldwide problem since the discovery of antimicrobial agents. The interpretative criterion for antimicrobial resistance generally is based on clinical breakpoint values, which divide bacteria into three categories: susceptible (associated with therapeutic success), intermediate (associated with an indeterminate or uncertain therapeutic effect), and resistant (associated with a high risk of therapeutic failure). However, this categorization is traditionally a clinical one and it is made irrespective of whether or not the organism harbor resistance mechanisms. In order to monitor the resistance development, the concept of “epidemiological cut-off value” (ECOFF) was raised. The purpose of ECOFF values is to separate the wildtype (WT) population (microorganisms without acquired resistance mechanisms) and nWT population (microorganisms with acquired resistance mechanisms) [12, 13].

The ability of Salmonella to form biofilms is seen as an important factor that contributes to its resistance and persistence in both host and non-host environments and is especially important when related to food processing environments [14]. Biofilms are defined as structured communities of bacterial cells embedded in a self-produced polymeric matrix attached to living or non-living surfaces [15, 16]. Bacteria in biofilms are generally well protected against environmental stresses, antimicrobials, disinfectants, and the host immune system, and as a consequence are extremely difficult to eradicate [17, 18]. Several reports have demonstrated the ability of Salmonella from porcine origins to form biofilms under unfavorable conditions, such as pig farms or slaughterhouses and food-processing plants [19–21].

Thus, the aim of this study was to evaluate Salmonella Derby isolates collected over a 10-year period from various porcine origins (environment, pig carcass, lymph nodes, intestinal content, and pork) for (i) their phenotypic and genotypic antimicrobial resistance profile, (ii) their ability to produce biofilm, and (iii) their molecular relationships by macrorestriction analysis.

Material and methods

Origin of the isolates

A total of 140 isolates of Salmonella enterica subsp. enterica serovar Derby were selected from the culture collection of the Laboratory of Preventive Veterinary Medicine (FAVET-UFRGS), Porto Alegre, Southern Brazil. These isolates were found in samples collected from pig slaughterhouses located in Santa Catarina (SC) and Rio Grande do Sul (RS) states, and from pork derived products obtained from supermarkets and butcher shops in Porto Alegre, Brazil, during the period of 1999 to 2010 (Table 1).

Table 1.

Salmonella Derby isolates included in this study

Sampling	Origin	Number of isolates	Sampling period	Reference
I	Mesenteric lymph node	12	1999–2000	Bessa et al. [48]
	Intestine fragment	12
II	Submandibular lymph nodes and tonsils	13	2001–2002	Ferraz et al. [49] Ferraz et al. [50]
	Intestinal content	4
	Ground meat	7
III	Mesenteric lymph node	28	2005	Schwarz et al. [51]
IV	Fresh pork sausage	9	2005	Mürmann et al. [10]
V	Pig carcass	21	2007–2009	Silva et al. [52]
	Intestinal content	10
	Lairage environment	2
	Floor of slaughterhouse	1
VI	Pig carcass	16	2010	Pissetti et al. [53]
	Working table for offal	2
	Lairage environment	3
Total		140

Minimum inhibitory concentration (MIC)

The MIC was determined using microdilution broth method according to Clinical and Laboratory Standard Institute [22, 23]. Thus, the following antimicrobials were used: nalidixic acid (Sigma-Aldrich, USA), ampicillin (Sigma-Aldrich, USA), cefotaxime (Sigma-Aldrich, USA), ceftazidime (Sigma-Aldrich, USA), ciprofloxacin (Sigma-Aldrich, USA), colistin (European Pharmacopeia, UK), streptomycin (Carl Roch, Germany), florfenicol (Sigma-Aldrich, USA), gentamicin (Sigma-Aldrich, USA), and tetracycline (Sigma-Aldrich, USA). The established set of tested antimicrobials, concentration range, and ECOFF and breakpoint values were carried out in accordance with EUCAST recommendations [24–26]. In the absence of EUCAST interpretative criteria, CLSI breakpoint values were adopted [23] (Table S1). Escherichia coli ATCC 25,922 was used as a reference strain for quality purposes.

Antimicrobial resistance genes profiling

Salmonella Derby isolates that displayed phenotypic resistance profile against ampicillin (n = 8), florfenicol (n = 1), streptomycin (n = 98), and tetracycline (n = 106) were investigated for the presence of resistance genes previously reported in this genus. Genomic DNA was prepared using the NucleoSpin Tissue Kits (Macherey–Nagel; Düren, Germany). Genes encoding resistance to β-lactams (bla_TEM and bla_PSE-1), phenicols (floR and catA1), aminoglycosides (strA, strB, aadA, aadB), and tetracyclines (tetA and tetB) were investigated by PCR assays, using the primers described by Lopes et al. [27].

Evaluation of biofilm-forming ability on polystyrene microtiter plates

The quantification of biofilm formation was performed in 230 µL of tryptic soya broth (TSB) with no glucose (Becton Dickinson & Company, USA) in sterile 96-well flat-bottomed polystyrene microtiter plates (Techno Plastic Products, Germany). A quantity of 20 µL of overnight bacterial culture, adjusted at 0.5 on the MacFarland scale, was added into each well. The plates were incubated aerobically at 37 °C for 24 h and at 28 °C for 96 h [28]. After incubation, the content of the plate was drained and the wells washed three times with sterile distilled water. During the washing process, the plates were vigorously shaken for removal of all non-adherent cells. The remaining attached bacteria were fixed with 250 µL of methanol per well. After 15 min, each plate was emptied and its air dried. The plates were stained with 250 µL per well of 2% Crystal Violet for 5 min. Any stain excess was rinsed off using distilled water. Subsequently, the dye bound to adherent cells was resolubilized with 250 µL of 33% (v/v) glacial acetic acid per well. The optical density (O.D.) of each well was measured at 570 nm using a spectrophotometer Strip Reader (EL301, BioTek, USA). Each isolate was tested in triplicate. Staphylococcus epidermidis ATCC 35,984 and Salmonella Typhimurium ATCC 14,028 were used as positive control for biofilm formation, while Salmonella Enteritidis ATCC 13,076 was used as a negative control for biofilm formation. The wells with no inoculum were used as quality control for the medium.

Determination of genetic relatedness using pulsed-field gel electrophoresis (PFGE)

The PFGE was carried out according to PulseNet standardized laboratory protocol for molecular subtyping of Salmonella (https://www.cdc.gov/pulsenet/pathogens/pfge.html). The genomic DNA was digested with 50 units of XbaI (Thermo Scientifics, USA) at 37 °C for 18 h. The respective fragments were separated by pulsed-field gel electrophoresis in certified agarose 1% (BioRad, CA, USA) in a CHEF DR II system (BioRad, CA, USA) at 6 V cm⁻¹ with 0.5 × Tris–borate-EDTA as the running buffer. The pulse times were increased from 2.2 to 63.8 s during 19–20 h period. The XbaI fragments of Salmonella Braenderup H9812 served as size standards. The gel was stained with ethidium bromide (1 μg mL⁻¹, Sigma, St. Louis, USA) and photographed under UV light. The images were recorded for analysis.

Data analysis

For each biofilm microtiter plate, the cut-off O.D. (O.D.c) was defined as three standard deviations above the mean O.D. of the negative control. The isolates were classified into four categories: O.D. ≤ O.D.c = no biofilm producer; O.D.c < O.D. ≤ (2 × O.D.c) = weak biofilm producer; (2 × O.D.c) < O.D. ≤ (4 × O.D.c) = moderate biofilm producer; and (4 × O.D.c) < O.D. = strong biofilm producer. The macrorestriction patterns were analyzed via GelCompar II software (Applied Maths, Belgium), and the similarities between patterns were determined in accordance with the Dice correlation coefficient, with a maximal position tolerance of 1.0% and optimization of 1.0%. The patterns were clustered using the unweighted pair group method with arithmetic averages (UPGMA). Any isolates with one band of difference were considered as different pulsotypes.

Results

Antimicrobial resistance profiling

Out of the 140 Salmonella Derby isolates included in this study, 123 (87.8%) were classified as nWT population and only 17 (12.1%) isolates were classified as WT population, according to epidemiological cut-off values. The highest frequencies of nWT populations were observed against tetracycline (75.7%), streptomycin (70%), and colistin (11.4%). There were no detectable nWT population to ciprofloxacin, ceftazidime, and gentamicin (Table 2). The highest values of MIC₅₀ and MIC₉₀ were observed for streptomycin and tetracycline, which exceeded ECOFF values. The antimicrobials cefotaxime and ciprofloxacin exhibited the lowest values of MIC₅₀ and MIC₉₀.

Table 2.

MIC distributions of 10 antimicrobial agents against 140 Salmonella Derby isolates, non-wildtype population, and resistance

^aAntimicrobial agents tested: ampicillin (AsMP), cefotaxime (CTX), ceftazidime (CTZ), ciprofloxacin (CIP), colistin (CL), florfenicol (FLF), gentamicin (GEN), nalidixic acid (NAL), streptomycin (STR), and tetracycline (TET)

^bNon-wildtype (nWT) isolates, according to ECOFF values

^cResistant isolates (R), according to clinical breakpoints

Black vertical lines indicate the ECOFF values and dotted vertical lines indicate the clinical breakpoints, when they are different

Concentrations not included in the test panel are shaded grey

Clinical breakpoints were used to establish the phenotypic profile of the S. Derby isolates. In this study, 123 (87.8%) isolates proved to be resistant to at least one antimicrobial tested, while only 17 (12.1%) isolates were classified as susceptible to all antimicrobials tested. There was no resistance detected for cefotaxime, ceftazidime, ciprofloxacin, and gentamicin. Resistance occurred most frequently to tetracycline (75.7%), streptomycin (70.0%), colistin (11.4%), nalidixic acid (9.3%), ampicillin (5.7%), and florfenicol (0.7%). The resistance profile STR-TET was the most frequent, which was found in 50.0% of the isolates (Table 3). Also, multidrug resistance (resistance to three or more classes of antimicrobial agents) was detected in 12.9% of the isolates. Resistance genes were detected among the resistant S. Derby isolates submitted to PCR analysis. The resistance genes found were bla_TEM for ampicillin resistance, aadA variant for streptomycin/spectinomycin resistance, tetA for tetracycline resistance, and floR for florfenicol resistance.

Table 3.

Antimicrobial resistance profiles of Salmonella Derby from pig production chain

Resistance profilea	No. of isolates	%
Susceptible	17	12.1
NAL	2	1.5
STR	10	7.1
TET	14	10.0
CL-NAL	2	1.5
CL-TET	2	1.5
NAL-STR	2	1.5
NAL-TET	3	2.1
STR-TET	70	50.0
AMP-CL-NAL	1	0.7
AMP-NAL-TET	1	0.7
AMP-STR-TET	2	1.5
CL-STR-TET	9	6.4
NAL-STR-TET	1	0.7
AMP-CL-STR-TET	2	1.5
AMP-FLF-STR-TET	1	0.7
AMP-NAL-STR-TET	1	0.7
Total	140	100

^aAMP, ampicillin; CL, colistin; FLF, florfenicol; NAL, nalidixic acid; STR, streptomycin; TET, tetracycline

Biofilm-forming ability of Salmonella Derby

Biofilm formation on polystyrene microtiter plate was observed in 68.6% (96/140) of the S. Derby isolates, where the plates were incubated at 28 °C (Fig. 1). Among biofilm producer isolates, 37.9% (53/140) were classified as weakly adherent, while 30.7% (43/140) were classified as moderately adherent. No strongly adherent isolates were observed at 28 °C. The values of optic density (O.D.) ranged from 0.1203 to 0.8110, with O.D. means of 0.2156 for cut-off (D.O.c), 0.1902 for non-adherent isolates (D.O. < D.O.c), 0.3538 for weakly adherent isolates (D.O.c > D.O. < 2 × D.O.c), and 0.4841 for moderately adherent isolates (2 × D.O.c > D.O. < 4 × D.O.c). It was proven that the temperature set at 37 °C did affect the biofilm formation. Only 5.7% (8/140) of S. Derby isolates were weakly adherent on polystyrene plate when incubated under this condition. The values of O.D. ranged from 0.1233 to 0.3647; the O.D. means of 0.1870 for cut-off (D.O.c), 0.1614 for non-adherent isolates (D.O. < D.O.c), and 0.2082 for weakly adherent isolates (D.O.c > D.O. < 2 × D.O.c). No moderately or strongly adherent isolates were observed at 37 °C.

Fig. 1.

Biofilm-forming ability of Salmonella Derby isolates on polystyrene microtiter plate

Genetic relatedness of Salmonella Derby

Through the biofilm and antimicrobial resistance profiles observed, 36 isolates were selected for macrorestriction analysis followed by PFGE, and at least one isolate from each phenotypic profile was included. XbaI-macrorestriction generated 20 pulsotypes with a similarity index ranging from 44.8 to 100%, nine of them were unique (Fig. 2). Among the 36 S. Derby isolates, 22 were clustered reporting 83.6% of similarity. Additionally, 27 isolates were clustered in 11 pulsotypes, which presented more than one strain with 100% of similarity (Table 4). The pulsotypes P10, P13, and P16 were found circulating in different sampling periods and from different origins (Table 4).

Fig. 2.

Dendrogram of Salmonella Derby isolates based on XbaI macrorestriction (PFGE) patterns. The similarity analysis was performed using the Dice coefficient and UPGMA method (tolerance 1%)

Table 4.

Characteristics of Salmonella Derby isolates belonged to the same PFGE pulsotype

PFGE pulsotype	Biofilm	Resistance profilea	No. of isolates	Sampling	Origin
P3	+	TET	2	II	Submandibular lymph node
					Ground meat
P6	+	AMP-STR-TET	2	II	Submandibular lymph node
					Ground meat
P7	+	Susceptible	2	V	Intestinal content
P8	+	TET	2	VI	Pig carcass
P10	+	CL-TET	2	I	Mesenteric lymph node
				II	Tonsil
		TET	2	I	Intestine fragment
				II	Submandibular lymph node
P11	+	TET	2	IV	Fresh pork sausage
		CL-STR-TET
P12	+	NAL-TET	2	III	Mesenteric lymph node
P13	-	STR-TET	2	VI	Working table for offal
	+	Susceptible	1	II	Ground meat
P15	-	STR-TET	2	II	Intestine fragment
P16	-	STR-TET	2	IV	Fresh pork sausage
			1	V	Pig carcass
		CL-STR-TET	1	VI	Pig carcass
P18	+	TET	2	I	Mesenteric lymph node

^aAMP, ampicillin; CL, colistin; NAL, nalidixic acid; STR, streptomycin; TET, tetracycline

Discussion

In this study, the concept of ECOFF value was used for determining nWT populations among 140 S. Derby isolates from pig production chain in a 10-year period. Using the MIC distribution, it was possible to identify 12.1% of the isolates as WT for all antimicrobial agents tested. Among the 140 S. Derby isolates, the highest frequencies of nWT were observed against tetracycline (75.7%), followed by streptomycin (70%), colistin (11.4%), nalidixic acid (9.3%), and ampicillin (5.7%). For ciprofloxacin, ceftazidime, and gentamicin, all of the isolates were classified as WT.

Due to international regulations on the use of antimicrobials in animal production by countries that import meat from Brazil, the use of tetracycline in Brazilian pig farms has decreased considerably lately. Even so, there are still occurrences of high rates of resistance to tetracycline in Salmonella remains [8, 29]. The maintenance of resistance to tetracycline is associated to the presence of tet genes in mobile genetic elements. When other resistance genes are collocated with a tet gene on the same plasmid, such a plasmid can be acquired under the selective pressure imposed by the use of antimicrobial agents other than tetracycline [30]. In the present study, the tet(A) gene was found in all tetracycline resistant isolates (n = 106), whereas the gene tet(B) was not detected. The gene tet(A) is part of a small non-conjugative transposon Tn1721, which are often integrated into conjugative and non-conjugative resistance plasmids in Enterobacteriaceae, and this may contribute to their spread among bacteria populations and environment [31].

A high frequency of nWT was observed against streptomycin (70%). Streptomycin is an aminoglycoside antimicrobial commonly used for treating enteric diseases in cattle, pigs, sheep, and poultry. The most common mechanisms associated with streptomycin resistance are the phosphotransferases APH(6)-Ia and APH(6)-Id, encoded by the strA and strB genes, respectively [31]. The maintenance of resistance to streptomycin is also associated to the presence of aadA genes coding for aminoglycoside-3″-O-adenyltransferases that confer resistance to streptomycin and spectinomycin [32]. In the present study, the aadA variant was found in all streptomycin resistant isolates (n = 98). Gene cassettes carrying variants of the aadA have been constantly found inserted into class 1 or class 2 integrons, or being part of the SGI1- or SGI2-associated multiresistance gene clusters, widely disseminated among Enterobacteriaceae [27, 33]. The other resistance genes such as bla_TEM for ampicillin resistance and floR for florfenicol resistance were also detected. The bla_TEM gene is usually part of transposon Tn3 in Salmonella isolates [31; 32]. The floR gene can be found in the chromosome or in plasmids of multiresistant Salmonella enterica serovars, including S. Typhimurium DT104 and S. Newport [31]. It is clear that these resistance genes, specially tet(A), are spread between environment, pig carcass, lymph nodes, intestinal content, and pork.

Our study also included colistin in the panel, which is widely used in veterinary medicine for controlling Enterobacteriaceae infections, especially post-weaning diarrhea in pigs due to Escherichia coli, and for prophylaxis purposes [24, 34]. Although the frequency of nWT isolates against colistin was 11.4%, most of isolates (75%) had MIC of 2 mg L⁻¹, very close to the clinical breakpoint and ECOFF value (> 2 mg L⁻¹), classified as borderline susceptible. It has been previously shown that colistin resistance in Enterobacteriaceae is resulting of chromosomal mutations, including mutations in the pmrA/pmrB two-component system that regulate the synthesis and structure of the lipopolysaccharide [35]. Additionally, the plasmid-mediated colistin resistance gene mcr-1 was identified in E. coli strain from a pig in China [36]. The detection of mcr-1 gene in one E. coli isolate classified as wildtype (susceptible) with MIC of 2 mg L⁻¹ was recently reported [37]. These findings indicate that there is a need to review the established epidemiological cut-off value for colistin.

Overall, a low frequency (0.7%) of nWT isolates to third-generation cephalosporins (cefotaxime and ceftazidime) was observed among S. Derby, corroborating previous studies [38, 39]. Also, resistance to fluoroquinolones such as ciprofloxacin was not found. These results represent a favorable situation for the public health panorama, since fluoroquinolones and extended-spectrum cephalosporins are the agents of choice for treating complicated Salmonella infection in humans. The emergence of extended-spectrum β-lactamases (ESBLs) in Enterobacteriaceae is an increasing problem throughout the world, compromising the use of these drugs for treating invasive gastroenteritis in humans. The increasing percentage of Salmonella resistant to fluoroquinolones and cephalosporins is a matter of concern and must be monitored carefully to assess potential risks to human patients, as well as any direct risk to veterinary husbandry [24].

Biofilm formation is one of the processes that may contribute to Salmonella persistence in different environments. The ability to adhere to surfaces and equipment, especially those commonly encountered in the food industry, may offer considerable risks of recurrent contamination affecting the quality and safety of food [40, 41]. The persistence of pathogenic microorganisms in these environments is a major concern due to the increased resistance to sanitizers and antimicrobials, plus conventional cleaning and sanitation carelessness to eradicate them from such surfaces. Studies have highlighted the presence of Salmonella biofilms and its resistance to the most commonly used sanitizers in the food industry [17, 42, 43].

It is interesting to observe that temperature has influenced the biofilm formation on polystyrene microtiter plates. When incubated at 28 °C, 68.6% (96/140) of the S. Derby isolates were positive for biofilm formation on polystyrene. Out of these, 37.9% were classified as weakly biofilm producer and 30.7% as moderately biofilm producer. Our results agree with the literature, which reported Salmonella isolates with biofilm-forming ability classified as weakly adherent on polystyrene microtiter plates [20, 44, 45]. When the test was performed at 37 °C, only 5.7% (8/140) of the isolates were positive for biofilm formation on polystyrene microtiter plate, being categorized as weakly biofilm producers. The biofilm formation behavior of Salmonella serovars from beef processing plants was investigated under a range of temperatures (4 °C, 10 °C, 25 °C, 37 °C, and 42 °C), and the strongest biofilm formation occurring at 25 °C [46]. Our data are also consistent with the research results of Yang et al. [43] who found Salmonella Enteritidis strains forming denser biofilms at 25 °C. In our study, the optimal temperature for the Salmonella isolates to produce biofilm was 28 °C, which is close to ambient temperature on farms, as well as during transportation from the farm to the slaughterhouse and in lairage areas.

PFGE, multiple locus variable number tandem repeat analysis (MLVA), and whole genome sequencing (WGS) are PulseNet’s main subtyping tolls. Based on biofilm and antimicrobial resistance profiles, 36 S. Derby isolates were selected for PFGE analysis. XbaI-macrorestriction of the 36 isolates generated 20 pulsotypes, nine of them being unique (P1, P2, P4, P5, P9, P14, P17, P19, and P20) (Fig. 2). A great cluster of 22 isolates showing 83.6% of similarity was identified. Three pulsotypes (P10, P13, and P16) were found circulating in different sampling periods and from different origins. Four isolates from P10 group showed biofilm-forming ability, tet(A) gene, and were originated from pigs at slaughter, been isolated from mesenteric lymph node, tonsil, intestine fragment, and submandibular lymph node, during the first and second sampling periods (I and II). In contrast to the P10 group, the isolates from P13 were more diverse. One isolate within the P13 group was biofilm producer, susceptible to antimicrobials and originated from ground meat in 2001/2002, while two isolates were non-biofilm producers, harbored aadA and tet(A) resistance genes, and were collected from working table for offal in 2010. A similar degree of homogeneity in the strain characteristics was seen among four isolates assigned to P16. The P16 isolates were from fresh pork sausage and pig carcasses of three different sampling periods (IV, V, and VI), non-biofilm producers, harboring aadA and tet(A) resistance genes. Therefore, closely related of S. Derby isolates were widely distributed and persistent among slaughter-age pigs, environment, and meat products along a decade in Southern Brazil. An integrated production system is characterized by closed herds in multi-site production units, which can provide several opportunities for Salmonella transmission. In this study, animals are originated from integrated production systems and the contamination with the same S. Derby clonal group might have occurred at the farm level, and spread during transport to slaughterhouse, during slaughter process, or in the handling of pork products. Identical Salmonella Derby isolates were also obtained from samples recovered from finishing farms and slaughterhouses in a study performed in Paraná state, Brazil, indicating that genetically close S. Derby isolates are spread in pig production chain in this region [8]. Salmonella isolates with the same PFGE pattern are more likely to have similar antimicrobial resistance profile; however, a correlation among PFGE, biofilm, and antimicrobial resistance profiles was not always observed. In this study, the P13 isolates were biofilm producer and non-biofilm producers with different resistance profiles. Han et al. [47] also reported that a correlation between PFGE clusters with resistant profiles and virulotypes was not observed among Salmonella isolates from a duck slaughterhouse.

Conclusion

Most of S. Derby isolates were able to form biofilm and the highest frequencies of nWT and resistant isolates were observed against tetracycline, streptomycin, and colistin. Through the macrorestriction by PFGE, it was possible to identify closely related S. Derby isolates between different collection periods and distributed among the different origins. In fact, considering the resistance towards antimicrobial agents of various classes and with different mechanisms of action, along with biofilm-forming ability and persistence in the pig production chain, the S. Derby isolates may pose a potential risk to human health when such isolates enter in the food chain.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by CS, VL, TA, and GVL. The first draft of the manuscript was written by CS and GVL, and all authors commented on previous versions of the manuscript. The manuscript was reviewed by GVL and MC. All authors read and approved the final manuscript.

Funding

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior — Brasil (CAPES) — Finance Code 001.

Declarations

Conflict of interest

The authors declare no competing interests.