Antimicrobial Resistance and Biofilm Formation Capacity of Salmonella enterica Serovar Enteritidis Strains Isolated from Poultry and Humans in Poland

Salmonella enterica ser. Enteritidis (S. enterica ser. Enteritidis) is the most frequently detected serovar in human salmonellosis, and its ability to produce a biofilm and the risk of transmission from animals and food of animal origin to humans are significant. The main aim of the present work was to compare S. enterica ser. Enteritidis strains isolated from poultry and human feces in terms of resistance profiles, prevalence of selected resistance genes, and their potential for biofilm formation, by assessing their biofilm growth intensity, the prevalence and expression of selected genes associated with this phenomenon, and the correlation between increased antimicrobial resistance and biofilm formation ability of the two tested groups of S. enterica ser. Enteritidis. This study showed a difference in antimicrobial resistance (minimal inhibitory concentration value) between S. enterica ser. Enteritidis groups; however, the majority of multidrug-resistant (MDR) strains were isolated from poultry (environmental samples from chicken broilers, turkey broilers, and laying hens). Differences in the prevalence of resistance genes were observed; the most common gene among poultry strains was floR, and that among strains from humans was blaTEM. S. enterica ser. Enteritidis strains isolated from poultry under the tested incubation conditions exhibited better biofilm growth than strains isolated from humans. A higher level of gene expression associated with the production of cellulose was only detected in the S48 strain isolated from poultry. On the other hand, increased expression of genes associated with quorum sensing was observed in two strains isolated from poultry farms and one strain isolated from human feces.


Introduction
According to the World Health Organization (WHO) [1], foodborne diseases remain a significant problem. Foodborne disease can be severe, especially among children, the elderly, and patients with immunosuppression. According to the latest European Food Safety Authority (EFSA) reports, Salmonella spp. continue to be the second most common cause of human infections and food poisoning associated with contaminated food [2]. Salmonella infections are a serious epidemiological and economic problem practically all over the world [3]. Human salmonellosis is an infectious disease that occurs in a variety of clinical forms Table 1. Antimicrobial dilution range for 95 Salmonella enterica ser. Enteritidis strains isolated from poultry sources and human feces according to minimal inhibitory concentration (MIC) values. 5

of 22
Pathogens 2020, 9, x FOR PEER REVIEW 5 of 23   The most common resistance gene among strains from poultry was floR (51.0%), which conditions phenicol resistance ( Figure 2). The prevalence of this gene among strains isolated from human feces was 20.5%. The most common resistance gene among strains isolated from human feces was blaTEM (63.6%), which confers beta-lactam resistance. The same gene was found in 35.3% of poultry strains. In addition, the blaPSE gene was present only in 5.9% of isolates from poultry. The prevalence of tetracycline resistance genes in both groups was as follows: the tetA gene was found in 37.3% of poultry strains and 38.6% of human strains, and tetB in 17.6 and 36.4% of strains, respectively. In contrast, tetC and tetG were detected only in isolates from poultry, and their prevalence was low-2.0 and 3.9%, respectively. Other resistance genes tested were cat1, found only in 11.8% of isolates from poultry, and cat3, found only in 29.5% of strains from humans. In both groups of S. enterica ser. Enteritidis strains, a very low prevalence of quinolone resistance genes was found, and no aminoglycoside or polymyxin resistance genes were found. It should be emphasized that 13.7% of S. enterica ser. Enteritidis strains isolated from poultry had resistance genes to three groups of tested antimicrobials, as much as 39.2% for two groups and 19.6% for one group. In contrast, 22.7, 36.4, and 25.0% of human isolate genes conferred resistance to three, two, and one group of antibiotics, respectively.

Prevalence of Biofilm-Related Genes
In all tested S. enterica ser. Enteritidis strains from the poultry farm environment (n = 51) and human feces (n = 44), csgD and sdiA genes were detected. The adrA gene was found in 100.0% of poultry strains. Only in one S. enterica ser. Enteritidis human strain (H43) was the gene encoding the cellulose synthesis protein (adrA) not present.

Biofilm Formation Assay
The results of the biofilm assay are presented in Table 3. The highest percentage of biofilm formation was in TBS medium at 28 °C (66.7% of poultry strains and 68.2% of strains isolated from humans, respectively). In turn, at 37 °C, most strains from both groups were characterized as weak biofilm formers, including 86.3% of isolates from poultry and 88.6% of isolates from humans. At 6 °C, it was observed that weak-producing EPS strains constituted the vast majority-78.4% and 65.9%, respectively. Most strains isolated from poultry (60.8%) at 28 °C were characterized by average biofilm growth intensity, while the highest percentage of isolates from humans (56.8%) formed 39

Biofilm Formation Assay
The results of the biofilm assay are presented in Table 3. The highest percentage of biofilm formation was in TBS medium at 28 • C (66.7% of poultry strains and 68.2% of strains isolated from humans, respectively). In turn, at 37 • C, most strains from both groups were characterized as weak biofilm formers, including 86.3% of isolates from poultry and 88.6% of isolates from humans. At 6 • C, it was observed that weak-producing EPS strains constituted the vast majority-78.4% and 65.9%, respectively. Most strains isolated from poultry (60.8%) at 28 • C were characterized by average biofilm growth intensity, while the highest percentage of isolates from humans (56.8%) formed biofilms poorly. At the temperature associated with the growth of human pathogenic bacteria (37 • C), only strains isolated from poultry showed medium (19.6%) or weak (80.4%) biofilm production. In contrast, 100.0% of strains isolated from humans were characterized by weak biofilm growth. At 6 • C, the vast majority of strains among both groups did not form biofilm or formed it weakly. Table 3. The number and percentage of Salmonella enterica ser. Enteritidis strains isolated from poultry (n = 51) and humans (n = 44) and biofilm growth intensity in TSB and M9 media at 28, 37, and 6 • C.

Expression of Genes Associated with Biofilm Production
The 12 S. enterica ser. Enteritidis strains selected for gene expression analysis were characterized by weak (n = 6) or strong (n = 6) biofilm growth in assay at 28 • C. All relative expression values ≤1.0 were interpreted as non-expression and >1.0 as increased expression. The high expression of genes responsible for regulating cellulose production in Salmonella spp., adrA and csgD, was observed only in strain S48 derived from poultry. A high expression level of luxS, involved in quorum sensing, was observed in A12 and A19 isolates from poultry. The expression of sdiA, a gene responsible for communication between bacterial cells, was high in A12 and H38 strains isolated from humans. It is worth noting that high expression levels of adrA and csgD were found in strain S48, showing a strong biofilm production capacity. In the case of luxS and sdiA, high expression levels were observed in strains H38, A12, and A19, which poorly formed a biofilm ( Figure 4).
None of the gene transcripts responsible for cellulose production and quorum sensing were significantly different between strong and weak biofilm producers ( Figure 5).
No significant difference in expression levels between the genes responsible for cellulose formation (adrA p = 0.3939, csgD p = 0.8410) and quorum sensing (sdiA p = 0.8182) was found. However, a significantly higher expression level (p = 0.0381) of the luxS gene was observed in strains isolated from poultry ( Figure 6).

Statistical Analysis
Statistical analyses carried out in this study aimed to assess the occurrence of a statistically significant relationship between antibiotic resistance, morphotype, presence of resistance genes, and the ability to form biofilm by S. enterica ser. Enteritidis strains under the conditions selected for testing. The results showed there was no statistically significant relationship between the occurrence of the MDR phenotype and the ability to form biofilm under the tested conditions in either group of S. enterica ser. Enteritidis strains. A statistically significant relationship was observed only between the morphotypes (RDAR, PDAR, BDAR, and SAW) of S. enterica ser. Enteritidis strains isolated from humans and biofilm formation in TSB medium at 6 • C. For strains isolated from poultry farms, relationships between the type of morphotype and the culture in M9 medium at 28 and 37 • C as well as between the presence of resistance genes and the ability of biofilm formation by these strains in M9 medium at 6 • C were noted.   None of the gene transcripts responsible for cellulose production and quorum sensing were significantly different between strong and weak biofilm producers ( Figure 5). No significant difference in expression levels between the genes responsible for cellulose formation (adrA p = 0.3939, csgD p = 0.8410) and quorum sensing (sdiA p = 0,8182) was found. However, a significantly higher expression level (p = 0.0381) of the luxS gene was observed in strains isolated from poultry ( Figure 6).

Statistical Analysis
Statistical analyses carried out in this study aimed to assess the occurrence of a statistically significant relationship between antibiotic resistance, morphotype, presence of resistance genes, and the ability to form biofilm by S. enterica ser. Enteritidis strains under the conditions selected for

Discussion
In the European Union, salmonellosis is the second most common zoonotic disease, after campylobacteriosis, and the most commonly detected serovar among humans is S. enterica ser. Enteritidis. Antimicrobial resistance is currently a global problem [3], and, according to the WHO [1, 28], it is extremely important to determine the prevalence of resistance genes in order to accurately describe the antibiotic resistance of microorganisms isolated from humans, animals, and food. Moreover, in the latest EFSA-ECDC antimicrobial resistance report [3], among S. enterica ser. Enteritidis strains recovered from humans in 2018, the most common were resistant to ampicillin, sulfonamides, tetracyclines, and ciprofloxacin. Therefore, one of the aims of our research was to determine and compare the antimicrobial resistance of S. enterica ser. Enteritidis strains from two sources. Moreover, the collected strains were analyzed for the discussed above antimicrobials.
Our results indicated that 11.4% of S. enterica ser. Enteritidis strains isolated from humans were susceptible to all tested antimicrobials, and in the case of poultry isolates only 2.0%. By contrast, Wasyl and Hoszowski [23] noted that 10.8% of S. enterica ser. Enteritidis strains isolated from the poultry farm environment, feed, and food were susceptible to antimicrobials.
Our research showed that all S. enterica ser. Enteritidis strains isolated from poultry and human feces were susceptible to gentamycin. Similar results were presented by Harb et al. [29], while other authors recorded strains being resistant to gentamycin from 7.1 to 12.5% [30][31][32]. In addition, the presence of aadB and aacC genes was not detected in our research, which was consistent with the results of other authors [33,34].
In our study, low resistance to beta-lactams was noted, which was previously demonstrated by Mąka et al. [24] among the isolates from food. On the other hand, Zhu et al. [32] noted resistance to ampicillin at 80% among S. enterica ser. Enteritidis strains from meat, but a low percentage of strains resistant to other beta-lactams. The high prevalence of the blaTEM gene (63.6% in our study) also corresponded with Zhu et al. [32]. It should be remembered that beta-lactam resistance may also be conditioned by other beta-lactamases in Salmonella spp. [35].
Tetracycline resistance, in our research, was demonstrated only in S. enterica ser. Enteritidis strains isolated from poultry (17.6%). In contrast, studies by Oliveira et al. [31] showed that approximately 12.5% of S. enterica ser. Enteritidis strains isolated from humans were resistant. S. enterica ser. Enteritidis resistance to tetracyclines was also found in studies by other authors, at rates of 1.0-100.0% [23,25,29]. It is worth pointing out that despite the lack of tetracycline-resistant strains among isolates from human feces, detection of resistance genes to this antimicrobial was noted among them. Comparable results were presented by Adesij et al. [36]. Similarly, the prevalence of beta-lactam resistance genes may have been related to the inactivity of the detected genes, but also to the potential for resistance under appropriate environmental conditions.
In the case of sulfamethoxazole, a large difference was observed in the percentage of resistant strains. These results were confirmed by Chen et al. [37] and Duffy et al. [38]. However, the prevalence rates presented by other authors [23,24,30,39] generally indicate a lower prevalence (0.0-10.0%) of sulfonamide resistance. It is worth noting that strains resistant to older generations of antimicrobials, such as tetracyclines and sulfamethoxazole, are most commonly isolated among Salmonella spp. [3,40].
Despite the high percentage of strains possessing phenolic resistance genes (poultry strains: floR 51.0% and cat1 11.8%; strains from humans: cat3 29.5% and floR 20.5%), no significant prevalence was observed among chloramphenicol-resistant strains. Similar results were presented by Zhu et al. [32]. The results herein and from other researchers may confirm the assumption that some genes are "turned off" in bacteria in vitro, but still have the potential to spread to other bacteria or to be involved in vivo, especially under the pressure of antibiotic selection [41].
It is worth noting that the majority of strains isolated from humans (56.8%) and poultry (41.2%) were resistant to nalidixic acid. Similar results (40.0%) were shown by Skarżyńska et al. [26]. Other authors noted a wide variety of nalidixic acid resistance (25.0-39.8%) in isolates from humans and food [25,29,31,42]. We noted that the prevalence of selected genes turned out to be very low, with a high percentage of nalidixic acid-resistant Salmonella strains. Other authors [10] did not record the presence of qnrA and qnrC genes among S. enterica ser. Enteritidis strains isolated from humans and from food samples-they found aac-(6')-Ib. In summary, PMQR resistance, although more commonly reported in other serovars than S. enterica ser. Enteritidis [10,43], may have serious public health consequences, as fluoroquinolones are commonly used to treat adults with invasive diseases or other serious infections caused by bacteria of the genus Salmonella.
Colistin is not an antibiotic commonly used in medicine; nevertheless, its use in large-scale breeding of animals, including poultry, may lead to selection pressure, which may result in the creation of resistant strains. In this aspect, the results obtained in our research indicate a high prevalence and should be considered significant. The prevalence of colistin-resistant S. enterica ser. Enteritidis strains isolated from poultry (84.3%) and from humans (79.5%) was high compared to other national reports [14,25] and reports from other countries [30,44]. The cited authors did not note colistin-resistant strains in their studies. This major difference may be due to changes in CLSI and EUCAST recommendations [45,46] for colistin resistance reporting (currently MIC > 2.0 mg/µl). In this paper, mcr-1 and mcr-2 genes were not detected; however, colistin resistance genes (also mcr3, mcr-4, and mcr5) were previously identified in various S. enterica serovars, but not in S. enterica ser. Enteritidis [47][48][49][50].
It is worth considering that in human salmonellosis, ampicillin, fluoroquinolones, co-trimoxazole, third-generation cephalosporins and chloramphenicol are commonly used, while antibiotic therapy is prohibited and not implemented in the case of salmonellosis of poultry in the European Union. However, antimicrobials are used for other bacterial infections in poultry, and their presence in the environment may cause selective pressure for strains, including Salmonella spp. In recent years, among others in the United States, Europe and Asian countries, a decrease in ciprofloxacin susceptibility was observed among Salmonella enterica strains [3]. However, fluoroquinolones are still the most used chemotherapeutic agents in the treatment of salmonellosis in adults in Poland [51].
Currently, not only is resistance to individual antimicrobials monitored, but it is extremely important to determine the occurrence of multidrug-resistant bacterial strains. Literature analysis showed that the presence of multidrug-resistant S. enterica ser. Enteritidis strains is not as common compared to other Salmonella serovars [52]. In the results of our study (27.5% of MDR strains isolated from poultry and 9.1% of strains from humans), a lower percentage of MDR strains was noted compared to the report presented by Wasyl and Hoszowski [23] and Oliveira et al. [31]. However, the analysis of these results considered resistance to two or more groups of antimicrobials (not to three or more, as in the case of research in this study), which could generate such a high result. Comparable to the results of Oliveira et al. [31], the percentage of MDR strains (57.1%) was noted in the United States among isolates from imported food [53]. It is worth noting that in our own study, the A09 strain was described, isolated from poultry with an MDR phenotype, and was resistant to nalidixic acid, chloramphenicol, colistin, tetracycline, and sulfamethoxazole, which is a wide range of resistance (five groups of antimicrobials). Such a wide range of resistance in poultry isolates may increase the threat of S. enterica ser. Enteritidis zoonosis. In recent years, an increase in isolation of MDR Salmonella spp. (resistance to ampicillin, chloramphenicol, co-trimoxazole and tetracycline) the in European Union and other regions was observed [1,3]. For this purpose, the program "Communication from The Commission on a Community Strategy Against Antimicrobial Resistance" [54] implemented in European Union countries has aimed at controlling the use of antibiotics in animal husbandry as well as the rational use of antibiotics in treatment.
In our initial research assumptions, the tested S. enterica ser. Enteritidis strains might be similar in the field of antimicrobial resistance. The differences described above may result from un-matching sample collection dates. The poultry isolates were collected in the period 2009-2016 while the human feces isolates were collected in the period 2015-2017. In addition, poultry strains came from laboratories from two different regions of Poland. Moreover, by receiving strains isolated from human feces, it could not obtain information about the patient's region (protection of personal data). These time-geographical discrepancies may have affected the differences in the results obtained.
Salmonella enterica as well as other species from the Enterobacteriaceae family are capable of adhering and forming biofilm on different kinds of materials during their life cycle [55]. Biofilms might be crucial to the survival of Salmonella in adverse environmental conditions, such as the poultry farm environment [56]. In the present research, the prevalence of adrA and csgD (cellulose production) as well as adrA and sdiA (quorum-sensing) genes was evaluated (100.0% of each gene). Only one strain isolated from human feces (H43) did not present adrA. The obtained results were consistent with Halatsi et al. [57] and Bhowmick et al. [58], who detected the genes described above not only in S. enterica ser. Enteritidis, but also in other Salmonella serovars. The presence of these genes and their expression might be useful in Salmonella biofilm production, via involvement in synthesis of EPS components and quorum-sensing communication.
The occurrence of morphotypes is dependent on the temperature and below 30 • C, production of crucial Salmonella biofilm elements is stimulated [16,19]. Solomon et al. [59] showed that 80% of S. enterica ser. Enteritidis isolates were characterized by the RDAR morphotype and the remainder by the SAW morphotype. On the other hand, the results by Lamas et al. [60] regarding Salmonella spp. from poultry farms showed a higher percentage of SAW morphotypes at 28 • C (77.1%) compared to our results, with a simultaneous 19.5% prevalence of the RDAR morphotype. Our results obtained at 37 • C (100.0% SAW morphotype among strains isolated from humans and 88.3% from poultry) were similar to those of Lamas et al. [60] and Römling et al. [19].
In compliance with data [61] describing that minimal media stimulates biofilm growth, TSB medium and M9 minimal medium were used for the present experiment. In the study by Lamas et al. [60], most strains from poultry farms (60.3%) were moderate biofilm formers at 28 • C in TSB, which was similar to the strains in our work (66.7%). On the other hand, Nair et al. [62] showed that S. enterica ser. Enteritidis (41.4%) isolated from poultry meat and human clinical samples produced a low level biofilm at 37 • C in TSB. Nonetheless, 14.3% of tested strains did not produce biofilm, which was not observed in our research. The obtained results of biofilm growth in M9 were significantly comparable to that of Lamas et al. [60], in which, at 28 • C, the majority (56.2%) of S. enterica strains from poultry showed moderate biofilm production. Moreover, incubation at 6 • C showed non-biofilm-forming strains (31.5%), as in our results, although 48.0% of tested strains by Lamas et al. [60] in these conditions showed weak biofilm formation. Research by Paz-Mendez et al. [63], contrary to our results, showed that all tested S. enterica strains produced biofilm in minimal media at 6 • C.
In the study of genes associated with biofilm formation expression, an increase was noted mainly in poultry strains. Gene expression was checked after 48 h incubation, which is in the earlier stages of biofilm formation; therefore, the relatively low gene expression associated with biofilm production may have been due to the initial phase of biofilm formation, when the ability to adhere to surfaces mainly is due to the cell surface characteristics, their hydrophobicity, and not to the EPS structure [64]. Transcription of biofilm-related genes is influenced by environmental factors such as acidic or low-nutrient conditions [65] as well as aerobic conditions during the incubation of Salmonella strains [66]. It is worth noting that among strains isolated from poultry (A12, A19, and S48), the increase in gene expression occurred in pairs: adrA + csgD and sdiA + luxS. Lamas et al. [67] showed a positive correlation in the expressions of csgD, luxS, and sdiA in anaerobic conditions. These results suggest a link between the expression of genes associated with biofilm detection and quorum sensing. Similarly, Wang et al. [21] achieved similar results in different culture media, suggesting a common expression of these genes. Confirming our results, Wang at al. [65] showed that in the early biofilm (72 h), investigated genes were downregulated. Upregulated expression of genes responsible for biofilm formation was assessed on the seventh day, i.e., in the mature biofilm [65]. Expression levels of csgB and csgD in TSB significantly increased after prolonged incubation, in line with the development of a mature biological layer [67], so it is worth expanding our own research in the future to analyze changes in expressions of selected genes after varying incubation times as well as in different media and temperature ranges. It is worth noting that many authors determine only the presence of a gene associated with biofilm formation, without checking the transcription level of individual genes [66]. The statistical analysis carried out showed no significant differences between genes responsible for either cellulose production or quorum sensing at the transcription level among the strains; therefore, it can be concluded that, among the randomly selected strains characterized by a strong biofilm growth, there was a high level of diversity relative to the genes tested.

Bacterial Strains
A total of 95 Salmonella enterica ser. Enteritidis strains were used in the present study and were isolated from two sources: (1) Poultry S. enterica ser. Enteritidis strains (n = 51) isolated from the environment of chicken broilers, turkey broilers and laying hen farms. Isolates were obtained from two veterinary laboratories: AGRO-VET Laboratory in Wroclaw and SLW BIOLAB in Ostróda. These laboratories collected strains in the period 2009-2016. For our research, strains isolated from poultry are designated as "A" or "S". (2) Human S. enterica ser. Enteritidis strains (n = 44) isolated from human fecal samples were collected in the period 2015-2017 by Dialab Laboratory in Wroclaw. Strains isolated from humans are designated as "H".
All collected strains were protected for further analysis and stored at −70 • C using Microbank ® (BioMaxima, Lublin, Poland).

Biofilm Formation Assay
The dye crystal violet combines with the negatively charged components in biofilm extracellular polymeric substances (EPSs) and can, therefore, be used to determine the biofilm mass [61]. The experimental procedure, with modifications regarding the temperature and incubation time, was performed with the protocol developed by Lamas et al. [60]. Strains were precultured on LB (BIOCORP, Warszawa, Poland), and a single colony was used for inoculation. Incubation at 28 • C (stimulating cellulose and curli expression) and at 37 • C (optimal temperature for the growth of pathogenic bacteria) continued for 48 h and at 6 • C (using to store food in refrigerators) for 168 h (7 days). LB and M9 minimal media were used. Optical densities were measured at 590 nm (OD590) with a Spark microtiter reader (TECAN, Männedorf, Switzerland). Each strain was tested in quadruplicate in two independent experiments. Results were interpreted using classification by Stepanović et al. [61] as follows: OD ≤ ODc non-biofilm former; ODc < OD ≤ 2ODc weak biofilm former; 2ODc < OD ≤ 4ODc moderate biofilm former; 4ODc < OD strong biofilm former.

Expression of Genes Associated with Biofilm Production
In order to compare the expression levels of csgD, adrA, sdiA, and luxS (the primers listed in Table 5), genes in 12 strains, which included weak biofilm formers ((n = 6) H30, H38, H44, A11, A12, and A19) and strong biofilm formers ((n = 6) H02, H08, H12, A34, S48, and S51) at 28 • C were used. Both groups included representative human and poultry strains (six human isolates and an equal number of isolates from poultry). Each strain was cultured in duplicate in a volume of 40 mL TSB medium (condition of bacterial culture-the density of the culture and general condition corresponded to the biofilm formation assay) for 48 h at 28 • C in Greiner Cellstar Dish plates (100 × 20 mm) (MERCK, Warszawa, Poland), which are made of the same type of plastic as a plates used in the biofilm formation assay. After incubation, the plates were washed twice, and the adhering bacterial cells were recovered by scraping and suspension in 1 mL of sterile PBS (Fisher BioReagents™, Waltham, MA, USA). After centrifugation (2000 g, 5 min), pellets were used for RNA isolation. RNA was extracted using RNA Mini Plus (A&A Biotechnology, Poland) following the manufacturer's recommendations. The RNA concentration and purity (A260 nm/A280 nm ratios) were obtained spectrophotometrically. (The integrity after isolation was checked by 1.0% agarose gel electrophoresis.) Contaminated DNA was removed using DNase I (Sigma-Aldrich, St. Louis, MO, USA) in a volume of 50 µL at 37 • C for 2 h. Total cDNA was amplified from 1 µg of RNA with SuperScript III (Invitrogen, Waltham, MA, USA) according to the manufacturer's instructions. Synthesis was performed with Random Hexamers in the following parameters: 5 min at 25 • C, 50 min at 50 • C, and 15 min at 70 • C. Real-time PCR reactions were performed in a thermocycler CFX-Connect™ Real-Time System (Bio-Rad, Irvine, CA, USA) according to the protocol developed by Lamas et al. [66]. After optimization, the reaction efficiency was higher than 97% for every tested gene. The reaction mixture contained 1 µL of template cDNA, 0.5 µM of each primer (listed in Table 5.), 10 µL of SsoFast EvaGreen Supermix, and water up to 20 µL, and the protocol was as follows: 95 • C for 30 s, followed by 35 cycles of 95 • C for 10 s and 60 • C for 15 s. Specificity of PCR was evaluated by melt-curve analysis in a temperature range 95-58 • C, and analysis of PCR products on agarose gel was performed for each reaction. The mRNA levels of all target genes were normalized to the 16S rRNA gene as the most stable in biofilm production in Salmonella [66]. Relative transcript levels were calculated according to Pfaffl [70] using Bio-Rad CFX Manager 3.0, and statistical calculations were made using Statistica 12.5 software (Stat Soft Polska Sp. z.o.o.). The non-parametric Mann-Whitney test was used for comparisons between biofilm-forming groups. Table 5. Primers used to assess the associated genes with biofilm expression in Salmonella enterica.

Statistical Analysis
The analysis was performed at a significance level of 5% using PQStat ver. 1.6.2 for Windows (PQStat Software © 2016), using Fisher's exact test, and TIBCO Software Inc. (2017), Statistica (data analysis software system), version 13.3, for Spearman rank correlation.

Conclusions
S. enterica ser. Enteritidis strains isolated from poultry and humans were similar in terms of the high prevalence of resistance to nalidixic acid. Similarly, a high colistin resistance was described, while no mcr-1 and mcr-2 genes could be detected. However, a high percentage of sulfamethoxazole-resistant strains was noted among isolates from poultry, and tetracycline resistance was described only within this group. We noted the presence of MDR strains, but it is worth emphasizing that most MDR phenotypes were described among S. enterica ser. Enteritidis strains isolated from poultry. The most common MDR profiles in this group were COL NAL SMX and COL TET SMX. In our study, the differences in the prevalence of resistance genes to beta-lactam, tetracycline, and phenicol between isolates from two sources were described; the most frequent gene among strains isolated from poultry was floR, and among strains from humans, blaTEM. To summarize biofilm production research, it should be noted that S. enterica ser. Enteritidis strains isolated from poultry under the tested conditions in vitro were characterized phenotypically as better biofilm formers.
This study did not show statistically significant differences in the transcriptomic levels between the two S. enterica ser. Enteritidis groups in terms of biofilm formation and the expression of genes responsible for cellulose production and quorum sensing. A similar observation on the expression of genes responsible for cellulose production as well as quorum-sensing genes in poultry and human strains was made. Moreover, a statistically significant correlation was not found between the MDR phenotype and the ability to form biofilm in the tested in vitro conditions in both analyzed S. enterica ser. Enteritidis strains.