**3. Discussion**

*Salmonella* is the second human bacterial zoonosis delivered especially by chicken and pork meats, but also milk, eggs, and seafood. The WHO estimates 550 million people (including 220 million children under the age of 5 years) fall ill each year due to diarrhoeal diseases due to unsafe food [26]. In 2018, 91,857 confirmed cases of salmonellosis in humans were reported with an EU notification (EFSA, 2019), while the U.S. Center for Disease Control and Prevention (CDC) estimates that *Salmonella* bacteria cause about 1.35 million infections per year in the United States, of which only 41,930 in 2011 were laboratory confirmed [27,28]. Along with the world population increase, the consumption of meat is also increasing. Foley et al. [19] reported that since the early 1900s, the consumption of chicken in the U.S. has increased about sixfold, while pork consumption by about 20%. Whereas, the European Union data show that in 2018, Europe increased its chicken meat production by a quarter, and 70% of this production was in six member states: Poland (16.8%), the United Kingdom (12.9%), France (11.4%), Spain (10.7%), Germany (10.4%), and Italy (8.5%) [18]. An upward trend, although less steep than in the case of poultry meat, was recorded for pork meat whose consumption, in Europe, has increased by about 3.5% per person in 10 years [19]. In order to meet consumer demands, unavoidable changes in animal production were necessary. The introduction of intensive animal husbandry practices has on the one hand increased the exposure of consumers to zoonosis, and, on the other hand, has probably modified the characteristics of *Salmonella* spp. colonization in farms by selecting strains resistant to antibiotics. In animals, *Salmonella* infection can cause fever, diarrhea, prostration, and mortality. Most of the animals that survive this infection remain asymptomatic carriers, posing a threat to human health as, during slaughtering, their carcasses can contaminate others [20,21]. Within *Salmonella* serotypes, *S.* Typhimurium, *S.* Enteritidis, *S.* Heidelberg, *S.* Montevideo, and *S.* Infantis are among the major pig and poultry serotypes most frequently associated with human infections [1]. Strains of *Salmonella* spp. with antimicrobial drug resistance acquired in the animal host are now widespread in all countries [22]. Resistance to ciprofloxacin, which belongs to the group of fluoroquinolones and was, until the last decade, the treatment of choice, and to cephalosporins is increasingly being documented [22–25]. Therefore, the WHO listed resistant *Salmonella* spp. among priority pathogens for which new antibiotics were urgently needed, and several countries have established *Salmonella* surveillance and control programmes. Our data agree with the above-reported concern, because the presence of widespread resistance to ciprofloxacin is confirmed by the circumstance that 21 of the 29 analyzed strains (72.4%) were resistant to fluoroquinolone, and highlight the resistance or a reduced sensitivity to cephalosporins (cefotaxime, ceftazidime, and cefepime), especially in the *S.* Infantis serotype. Amoxicillin/clavulanic acid is another drug showing decreased efficacy especially against *S.* Infantis and monophasic *S.* Typhimurium strains. Although the combination of amoxicillin with clavulanic acid overcomes the intrinsic resistance of beta-lactamase-producing strains, and therefore makes it one of the main antimicrobial substances in swine medicine for the treatment and control of infections, the fair percentage of resistance (55.2% of strains) supports the choice of the European Medicine Agency [26] to classify this association in category C. This category includes antibiotics that are approved for use in livestock and pet animals, but which must be used with caution, only when there are few or no alternatives belonging to category D [26,27]. Natural substances represent a valid resource in the search for alternatives to current antibiotics. Thanks to their high antimicrobial potential, EOs are widely studied to counteract the development of antibiotic resistance and respond to the growing demand of consumers for antibiotic-free foods [12,28]. As noted in the introduction, the *O. vulgare* EO was found to be active against a broad spectrum of microorganisms. The antimicrobial activity is essentially mediated by the main chemicals carvacrol and thymol, which, because of their amphipathic nature, interact with the bacterial and fungal cell membrane. In particular, carvacrol is able to accumulate in the cell membrane of *Salmonella* spp and other bacteria strains, where it can bind to hydrogen by altering the cell membrane potential and inducing a conformational and metabolic modification (decrease of ATP production) up to the time of cell death [20]. This antimicrobial activity of the *O. vulgare* EO on bacterial and fungal membranes is common to many EOs caracterised by the same amphypathic chemical compounds. Despite their strong antimicrobial action, the use of EOs in farms is limited by their poor water solubility. This characteristic makes it necessary to convey them with suitable surfactants or through biotechnological processes. The Italian product GR-OLI is a water-soluble mixture of EOs emulsified in an inert carrier additive, which is regularly authorized as additive for use in animal feed. This mixture has been compared with the activity of the *O. vulgare* OE that recently received a positive opinion from the EFSA for use in animal production. The chemical analysis of both products shows that the *O. vulgare* OE and GR-OLI have respectively three (carvacrol, p-cymene, and γ-terpinene) and eight (limonene, carvacrol, 1-8 cineol, p-cimene, linalool, terpinen-4-ol, and thymol) chemicals with a concentration >5%. Furthermore, if compared to the *O. vulgare* EO, the GR-OLI has a lower concentration of carvacrol and a higher concentration of the other terpenic molecules with known antimicrobial action. If, on the one hand, the antimicrobial action of carvacrol is well known [29,30], on the other hand, this phenolic compound is acknowledged to be potentially toxic, depending on the concentration of use [31]. For this reason, a preliminary in vitro comparison between the antimicrobial properties of *O. vulgare* EO and this commercial aromatic mixture was needed. Data show that the MIC90 of the *O. vulgare* EO is slightly lower than that of GR-OLI against the different *Salmonella* strains tested, and that the sub-MIC of *O. vulgare* EO inhibits over time the *S.* Typhimurium growth more effectively than GR-OLI. However, while the *O. vulgare* EO is only capable of disaggregating a formed biofilm, GR-OLI is simultaneously capable of inhibiting the formation of the biofilm and disaggregating the formed one at minimal concentrations potentially compatible with animal palatability. The ability to prevent the early stages of bacterial adhesion to intestinal cells is critical for the establishment of chronic

colonization in animals, which are the reservoir for acute events. In this regard, data obtained from the cell adhesion assay confirmed that GR-OLI, at very low concentrations, is actually able to inhibit bacterial adhesion to the intestinal cell line Caco-2. Inhibition occurs in di fferent ways depending on the serotype. Specifically, the monophasic *S.* Typhimurium and *S.* Infantis strains showing the greatest resistance to antibiotics were sensitive only to the higher concentration tested, while the other strains tested were sensitive to both concentrations. These activities could be useful also with animals carrying *Salmonella* spp. asymptomatically. In these animals, it is important to inhibit both the adhesion and the formation of the biofilm to prevent contamination of the carcasses at the time of slaughtering. Furthermore, data obtained from the checkerboard test indicate that GR-OLI has synergistic action with ciprofloxacin at concentrations much lower than MIC. This data identifies a possible new resource in the fight against antibiotic resistances, as it indicates the possibility of reactivating the sensitivity to ciprofloxacin with low doses of natural compounds mixed with commercial antibiotics. Moreover, given the heterogeneity of the phytocomplex of each EO, the use of concentrations lower than MIC is not currently correlated with the development of resistance [32]. This makes the use of sub-MIC of the EOs mixtures safer against the development of potential resistances.

#### **4. Materials and Methods**

#### *4.1. Natural Substances, Antibiotics, and Reagents*

*O. vulgare* L. EO and GR-OLI (by APA-CT, Forlì, Italy), a confidential solution (under patent processing) containing the 25% *v*/*v* of nine EOs (*Eucalyptus globulus*, *Satureja hortensis*, *Citrus aurantium* var. *dulcis*, *Thymus vulgaris*, *Melaleuca alternifolia*, *Citrus limon*, *Lavandula hybrida*, *Melaleuca cajeputi*, *Thymus capitatus*) dispersed in a surfactant (Glyceryl polyethyleneglycol ricinoleate cod. E484), admitted in animal feed, were tested against *Salmonella* spp. No preservatives or other substances were added to the mixture.

Amoxicillin/clavulanic acid, cefotaxime, and ciprofloxacin (Sigma Aldrich, St. Louis, MO, USA) were used to test their interaction with GR-OLI. C8-C40 n-alkanes mixture, p-cymene, limonene, 1,8-cineol, thymol, carvacrol, and n-hexane were purchased from Sigma-Aldrich (Milan, Italy) and used as standards. All reference standards used for GC analysis, chromatographic-grade organic solvents, and reagents were purchased from Sigma-Aldrich (Milan, Italy).

#### *4.2. Bacterial Strains and Growth Media*

To study the e ffectiveness of the natural products, 29 isolates of *Salmonella enterica* subsp. *enterica* (specifically, 17 *S. enterica* subsp. *enterica* serovar Typhimurium, of which 4 monophasic, and 12 *S.* Infantis). *Salmonella* spp. strains were isolated, during 2017, from swine and poultry intensive farms with no epidemiological correlation and provided by Istituto Zooprofilattico of Forlì (Italy). *Salmonella* spp. strains were isolated from environmental samples (faeces, boot swabs) as part of monitoring plans for the reduction of the most important public health-related *Salmonella* serovars (*S.* Typhimurium, including monophasic variants, *S.* Enteritidis, *S.* Infantis, *S.* Virchow, and *S.* Hadar) in poultry and swine farms. The detection of *Salmonella* spp. was carried out using a culture method according to Amendment 1: Annex D of EN/ISO 6579:2002 [29]. Based on this method, colonies of presumptive *Salmonella* were subcultured and their identiy was confirmed by means of biochemical tests. The pure colonies showing typical biochemical reactions for Salmonella were also tested for the presence of *Salmonella* somatic antigens (O-antigens) and flagellar antigens (H-antigens) by slide-agglutination using polyvalent antisera (BD Difco ™—Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Serotyping of *Salmonella* spp. strains was carried out using a slide-agglutination test following the White-Kau ffmann-Le Minor scheme according to the part 3 of ISO/TR 6579-3:2014 [30]. For this purpose, a colony from a pure culture of each *Salmonella* spp. strain was cultured on nutrient agar and incubated at 37 ◦C ± 1 ◦C overnight. After the incubation, each strain was investigated for auto-agglutination by the slide-agglutination test using a 3.5% solution of sodium chloride. Once auto-agglutination was

excluded, each strain was submitted to the agglutination test for serotyping the most important public health-related *Salmonella* serovars: *S.* Typhimurium (including monophasic variants), *S*. Enteritidis, *S.* Infantis, *S.* Virchow, and *S.* Hadar. For this purpose, the following somatic antisera (O-antisera) were used: O:4, O:5, O:6, O:7, O:8, O:9, and O:46 (BD Difco™—Becton, Dickinson and Company, Franklin Lakes, NJ, USA); after agglutination with the O-antisera, the agglutination with flagellar antisera (H-antisera) was performed using the following flagellar H-antisera: H:i, H:2, H:g, H:m, H:q, H:s, H:t, H:r, H:5, H:z10, and H:x (BD Difco™—Becton, Dickinson and Company, Franklin Lakes, NJ, USA). For biphasic H-antigens strains (e.g., *S.* Typhimurium), if one H-phase was negative, a phase inversion was carried out using the Sven Gard method according to the part 3 of ISO/TR 6579-3:2014. Based on their antigenic formula, the *Salmonella* spp. strains were identified according to the White-Kauffmann-Le Minor scheme [31]. The antigenic formula of the *Salmonella* spp. strains used in this study is summarized in Table 5. Muller Hinton medium (MH, Sigma Aldrich, St. Louis, MO, USA) was used to grow the strains at 37 ◦C ± 1 ◦C for 24 h.



Note. Numbers underlined and in square brackets are in accordance with the international nomenclature for *Salmonella* spp.

#### *4.3. GC-MS Analysis*

Analyses were performed on a 7890A gas chromatograph coupled with a 5975C network mass spectrometer (GC-MS) (Agilent Technologies, Waldbronn, Germany). Compounds were separated on an Agilent Technologies HP-5 MS cross-linked poly–5% diphenyl–95% dimethyl polysiloxane (30 m × 0.25 mm i.d., 0.25 μm film thickness) capillary column. The column temperature was initially set at 45 ◦C, then increased at a rate of 2 ◦C/min up to 100 ◦C, then raised to 250 ◦C at a rate of 5 ◦C/min, and finally held for 5 min. The injection volume was 0.1 μL, with a split ratio 1:20. Helium was used as the carrier gas, at a flow rate of 0.7 mL/min. The injector, transfer line, and ion-source temperature was 250, 280, and 230 ◦C, respectively. MS detection was performed with electron ionization (EI) at 70 eV, operating in the full-scan acquisition mode in the m/z range 40–400. The EOs were diluted 1:20 (*v*/*v*) with n-hexane before GC-MS analysis.

#### *4.4. GC-FID Analysis*

Analyses were carried out on a gas chromatograph coupled with a flame ionization detector (FID) Agilent Technologies 7820A. Compounds were separated on an Agilent Technologies HP-5 cross-linked poly–5% diphenyl–95% dimethyl polysiloxane (30 m × 0.32 mm i.d., 0.25 mm film thickness) capillary column. The temperature programme was the same as described in Section 4.3. The injection volume was 0.1 μL in the split mode 1:20. Helium was used as the carrier gas at a flow rate of 1.0 mL/min. The injector and detector temperature were set at 250 and 300 ◦C, respectively. The EOs and the reference standards were diluted 1:20 (*v*/*v*) with n-hexane before GC-FID analysis. The analyses were performed in triplicate.

#### *4.5. Qualitative and Semi-Quantitative Analysis*

Compounds were identified by comparing the retention times of the chromatographic peaks with those of authentic reference standards run under the same conditions, and by comparing the linear retention indices (LRIs) relative to C8-C40 n-alkanes obtained on the HP-5 column under the above-mentioned conditions with the literature [32]. Peak enrichment by co-injection with authentic reference compounds was also carried out. Comparison of the MS-fragmentation pattern of the target analytes with those of pure components was performed, by using the National Institute of Standards and Technology (NIST version 2.0d, 2005) mass-spectral database. Semi-quantification was calculated as the relative percentage amount of each analyte; in particular, the values were expressed as the percentage peak area relative to the total composition of each EO obtained by GC-FID analysis.

#### *4.6. Antimicrobial Susceptibility Testing against Antibiotics*

To investigate the antimicrobial susceptibility to amicacin/clavulonic acid (AMC), piperacillin/tazobactam (TZP), cefotaxime (CTX), ceftazidime (CAZ), cefepime (FEP), ertapenem (ETP), imiprenem (IPM), meropenem (MEM), amikacin (AMK), gentamicin (GEN), ciprofloxacin (CIP), and trimethoprim/sulfamethoxazole (SXT), we performed antimicrobial susceptibility testing (AST) with the VITEK ® 2 system according to the manufacturer's instructions, using the software version 7.01 and the AST-N379 cards for Gram-negative bacteria. To test the antimicrobial susceptibility against ciprofloxacin, we performed AST by the Broth Micro Dilution method according to the 2006 ISO 20776-1 procedure. MIC results were categorized as susceptible (S), susceptible by increased exposure (I), and resistant (R) according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints (version 10.0) [33].

#### *4.7. Broth Microdilution Susceptibility Testing against Natural Products*

The broth microdilution (BMD) susceptibility test according to the European Committee on Antimicrobial Susceptibility Testing (EUCAST) international guidelines was performed. Muller Hilton broth (Oxoid, Basingstoke, Hampshire, UK) was used to test the antimicrobial activity of GR-OLI and *O. vulgare* EO against the *Salmonella* spp. strains. The BMD test was performed on a 96-well plate by adding 100 μL of a cell suspension equal to 5 × 10<sup>5</sup> CFU/mL to a final volume of 200 μL. Scalar dilutions, between 16% *v*/*v* (equal to 40 μL of EOs content/ mL) and 0.125% *v*/*v* (equal to 0.3 μL of EOs content/mL) of GR-OLI and between 4% (40 μL /mL) and 0.03% (0.3 μL /mL) of *O. vulgare* EO, were tested. A concentration surfactant (Tween 80, Sigma Aldrich, Saint Louis, MO, USA) corresponding to that contained in the GR-OLI was tested together with *O. vulgare* EO to facilitate its solubilization in the hydrophilic medium. Plates were incubated overnight at 37 ◦C. After this period, MIC values were determined by spectrophotometric reading at 450 nm (EL808, Biotek, Winooski, VT, USA). To evaluate the MBC, 5 μL of the content of each well were seeded on standard medium agar plates, which were incubated for 24 h at 37 ◦C. Surfactants were tested separately. The MIC is defined as the lowest concentration that completely inhibits the growth of a given organism compared with the growth in the substance-free control; whereas the MBC is defined as the lowest concentration determining the death of 99.9% or more of the initial inoculum. Each test was performed in triplicate, and in each experiment suitable positive controls and blank were added. Surfactants were tested separately.

## *4.8. Biofilm Assay*

All isolates were grown overnight in MH broth (Sigma Aldrich, Saint Louis, MO). To allow the formation of biofilm, cells were diluted in Luria Bertani broth (LB, Sigma Aldrich, Saint Louis, MO, USA) to a turbidity of 0.5 McFarland, corresponding to 5 × 108. To study the activity of both GR-OLI and *O. vulgare* EO on the biofilm formation, both natural compounds were added in triplicate at the maximum concentration of 0.5% and 0.125% and at minimum concentration of 0.125% and 0.03%, respectively. The suspension was inoculated in polystyrene 96-well plates (Thermo Fisher Scientific, Waltham, MA, USA) and incubated at 37 ◦C for 48 h. No treated cells were added as a positive control in triplicate. Wells were then washed three times with PBS and the resultant biofilms were stained with crystal violet (CV) staining (Sigma-Aldrich, Saint Louis, MO, USA) for 30 min. The stained biofilms were washed in PBS and 100 μL of ethanol were added to each well for one minute to completely dissolve the CV. Then, the ethanol was transferred into a new 96-well plate to determine the absorbance at 560 nm. To test for disaggregation of biofilm, bacterial cells were prepared as described before without adding substances. After 48 h of incubation, biofilm was washed three times with PBS and

cells fixated in acetone for 10 min. GR-OLI and *O. vulgare* were diluted in PBS at the maximum and minimum concentration aforementioned and added to the biofilm for another 24 h at 37 ◦C. Biofilm was then quantified as already described. Both tests were conducted in triplicate and repeated twice.

#### *4.9. Cell Adhesion Assay*

Two *Salmonella* spp. strains sensitive to almost all antibiotics (S33 and S42 resistant to only ciprofloxacin) and six multi-resistant strains (S24 and S32 *S.* Typhimurium, S19 and S29 monophasic *S.* Typhimurium and S40 and S41 S. Infantis) were randomly selected and used to study their adhesive capacity on human Caucasian colon adenocarcinoma cells (Caco-2) in the presence or absence of two concentrations of GR-OLI or *O. vulgare*.

The CACO-2 cell line was cultured in Dulbecco's modified Eagle medium (DMEM) (Gibco, Grand Island, NY, USA) supplemented with 10% heat-inactivated fetal calf serum (Integro B.V., Zaandam, The Netherlands), 1% nonessential amino acids (Gibco, Grand Island, NY, USA), and 1 mM glutamine (Gibco, Grand Island, NY, USA) and incubated at 37 ◦C with 5% CO2. Di fferentiated CACO-2 cells were prepared by seeding cells 5 to 10 time in 250-mL flasks (Costar, Oneonta, NY, USA) at 1.6 × 10<sup>7</sup> cells/mL in DMEM, with all supplements and then transferred to 24-well tissue culture plates at 1.6 × 10<sup>5</sup> cells/mL. The culture medium was replaced every three days. Overnight grown cultures of *Salmonella* isolates were diluted (1:100) in the presence of 0.125% *v*/*v* and 0.5% *v*/*v* of GR-OIL (APA-CT, Forlì, Italy) and grown at 37 ◦C to an OD660 of 0.8. For each strain, an inoculum was added without adding the formulate as the control. Bacteria were harvested by centrifugation and suspended in Dulbeccos Modified Eagles Medium to a final concentration of 1 × 10<sup>7</sup> CFU/mL. Then, a 1-mL bacterial suspension of each strain was added to the wells (1:100 MOI). Plates were incubated for 4 h at 37 ◦C. After incubation, monolayers were rinsed three times with PBS (phosphate bu ffer solution) and cells were gently scraped with a cell scraper (Falcon, Reynosa, Tamaulipas, Mexico) and harvested with PBS and washed by centrifugation twice. The adherent bacteria were quantified by plating serial dilutions on LB agar plates and counting CFU. The inoculum was plated to determine viable counts. The assay was performed in triplicate and repeated twice.

#### *4.10. Growth Curves*

As described in the broth microdilution susceptibility testing method, a suspension of 5 × 10<sup>5</sup> cfu/mL of the same strains used for the cell adhesion assay was seeded in a 96-well plate together with *O. vulgare* EO or GR-OLI at the MIC and sub-MIC concentrations or only with culture medium (growth control). Strains were incubated at 37 ◦C and monitored overnight by detecting OD450 every 30 min for 20 h. A statistical comparison between the OD450 detected at 10, 15, and 20 h of treated and untreated samples was made to quantify the extent of growth inhibition for each treatment.

#### *4.11. Checkerboard Titration Method*

Four strains of *S.* Infantis (S26, S35, S36, S37), all multi-resistant to amoxicillin/clavulanic acid, cefotaxime, cefepime, and ciprofloxacin, were tested using the checkerboard titration method. Then, 96-well microplates were used, each one containing MH broth with concentrations ranging from 12.5% *v*/*v* (equal to 31 μL of EOs content /mL) to 0.19% *v*/*v* (equal to 0.5 μL of EOs content /mL) for GR-OLI and from 128 to 0.125 μg/mL for amoxicillin/clavulanic acid or from 16 to 0.03 μg/mL for cefotaxime or from 4 to 0.005 μg/mL for ciprofloxacin and a combination of GR-OLI and one of the aforementioned antibiotics in a checkerboard style. The final inoculum was 5 × 10<sup>5</sup> cfu/well. The microplates were incubated for 24 h at 37 ◦C. After the incubation period, the MBCs were evaluated by sowing 5 μL of the contents of each well on nutrient agar and incubating it at 37 ◦C for 24 h. The FIC value could not be evaluated as due to the turbidity of the contents of the wells, it was not possible to define the MIC values, while the FBC index were calculated in compliance with international guidelines (EUCAST, 2000). Synergism was defined as FBC index <0.5; additivity FBC index between 0.5 and 1; indifference FBC index between 1 and 2; and antagonism FBC index > 2 [33] (EUCAST, 2000). Each experiment was performed in triplicate, independently.

#### *4.12. Statistical Analysis*

Relative data of biofilm inhibition and disaggregation (OD values) and of cell adhesion assays (CFU values) were plotted as means ± standard errors (SE). Means whose SE bars did not overlap were considered significantly different. Relative data of growth curves at 10, 15, and 20 h after treatment with *O. vulgare* EO and GR-OLI, which were shown to satisfy the conditions for ANOVA, and were subjected to one-way ANOVA within each *Salmonella* strain and time after treatment. The lowest significant difference (LSD) test at *p* < 0.05 was used to separate levels in strain/time combinations significant at the ANOVA.
