For phage propagation, isolated from diseased chickens Salmonella Enteritidis (S. Enteritidis) strain A28 from the Strain Collection of the Department of Epizootiology and Clinic of Bird and Exotic Animals, Wrocław University of Environmental and Life Sciences was used in this study. For experimental infections, S. Enteritidis 327 lux was used. This strain was originally collected from broiler chicken (Kisiela et al., 2006), modified according to Riedel et al. (2007), and described elsewhere (Kuźmińska-Bajor et al., 2015). Briefly, S. Enteritidis 327 lux strain was modified with the chromosomal integration vector p16Slux containing the lux operon from Photorhabdus luminescens and the erythromycin resistance gene. S. Enteritidis 327 lux exhibits resistance to erythromycin and the capacity for bioluminescence, and these modifications did not affect invasiveness in mice and chickens (Kuźmińska-Bajor et al., 2012, 2015). All S. Enteritidis strains used in this study were stored with 20% (v/v) glycerol at − 80°C. To reactivate them, strains were cultivated in Luria-Bertani (LB) broth (Sigma-Aldrich, Germany), under aerobic conditions at 37°C overnight with shaking at 150 rpm. S. Enteritidis 327 lux culture was supplemented with erythromycin (0.2 mg/ml).

The bacteriophages used in this study, UPWr_S1, UPWr_S3 and UPWr_S4, were isolated from samples of sewage. Taxonomically, UPWr_S1, UPWr_S3 and UPWr_S4 belong to the genus Jerseyvirus within the Siphoviridae family and were described elsewhere (Kuźmińska-Bajor et al., 2021). In vitro tests revealed lysis effectiveness rates of 41, 82, and 55% for UPWr_S1, UPWr_S3 and UPWr_S4 phages on 67 Salmonella strains belonging to 6 serovars. In silico analysis showed that all gene products of these phages had no relevance to any known virulence, toxin, or pathogen-associated protein family or gene products of Salmonella strains or any other pathogens. Due to the characteristics of these phages, including their activity against different serovars, as well as great antibiofilm potential (Korzeniowski et al., 2022), they were considered a promising tool in combating Salmonella and hence were chosen to conduct in vivo phage efficacy experiments in mice and chickens. The amplification of UPWr_S1, UPWr_S3 and UPWr_S4 phages was performed on the S. Enteritidis A28 host strain. Phages were amplified using a method described elsewhere (Kuźmińska-Bajor et al., 2021). Briefly, bacterial cultures were prepared by inoculation of 10 ml of LB broth with a single colony following overnight incubation at 37°C with shaking at 150 rpm. After incubation, 0.5 ml of ml of phage lysate was added and the culture was continued overnight at 37°C. In the next step, the bacterial culture was centrifuged for 10 min at 5000 × g to remove any remaining cell debris and then filtered through 0.22 μm pore size syringe filters. The 5 ml of resulting phage lysate, from the first step of propagation, was added to 150 ml of host culture (OD_600nm = 0.2) and incubated overnight at 37°C. Then, the centrifugation and filtration steps were repeated. Bacteriophage titer was determined using the routine test dilution method (Adams, 1959). As a phage mixture, cocktail UPWr_S134 containing phages UPWr_S1, UPWr_S3, and UPWr_S4 was formulated by combining in an equal ratio of 1:1:1.

The stability of the UPWr_S134 phage cocktail at titer 10⁸ PFU/ml was tested to determine phage survival at temperatures (4–42°C) representing temperatures of those storage conditions, broiler handling and the chicken body and over a wide pH (2–13).

For stability at different temperature conditions, the working stock of the UPWr_S134 phage cocktail was incubated for 2 weeks at 4 different temperatures 4, 20, 36 and 42°C. 4°C corresponds to storage condition, 20°C corresponds to the temperature of broiler handling during at least half of the broiler production cycle, 36°C to slightly higher temperature of post-hatch handling and 42°C to broilers’ deep body temperature, respectively. Phage titer at each temperature was determined on the 7th and 14th days of incubation by using the double-layer agar method (Adams, 1959).

Stability studies for acidic and alkaline conditions were conducted according to the methods described by Lu et al. (2020). Briefly, the phage cocktail was adjusted with 1 M NaOH or HCl (Sigma-Aldrich), to yield a pH range of 2–13, and incubated at 37°C for 1 h to determine phage survival using the double-layer agar method. All experiments were performed in triplicate.

Simulated gastric fluid (SGF) was used to test phage sensitivity to intragastric conditions according to a method described previously (Tang et al., 2013). The experiment was performed providing three distinct conditions: (i) control SGF representing natural stomach acidity, (ii) neutralized low pH of SGF by the addition of 14% CaCO₃ as an antacid, and (iii) SGF containing 30% m/v chicken feed Brojler Grower I (Tasomix, Poland) as a simulation of the chicken stomach contents during the digestion process. 1 ml of the 10⁸ PFU/ml UPWr_134 phage cocktail was added to 9 ml of pre-warmed SGF and incubated at 42°C for 2 h. 100 μl were taken every 15 min to measure the survival rate of phages. To measure the phage titer, samples were tenfold serial diluted in LB medium and spotted on the lawn of S. Enteritidis A28. The plates were incubated at 37°C overnight and observed for the formation of plaques. This experiment was carried out in triplicate.

The potential for reduction of Salmonella by phage cocktail UPWr_S134 was examined in BALB/c female mice (Mossakowski Medical Research Centre Polish Academy of Sciences, Warsaw, Poland) experimentally infected with 10⁷ colony-forming units (CFU) of S. Enteritidis 327 lux strain. All experimental work applying the mouse model of infection was approved by the Local Ethical Committee for Animal Experimentation (protocol code 115/2015; Wrocław, Poland) and performed according to the legal requirements. In our previous research, it was demonstrated that all mice infected with this particular dose of S. Enteritidis 327 lux strain developed a typhoid-like disease course always resulting in death, preceded by luminescent signals recorded within 24 h, in a period of time ranging from 8 to 11 days post-infection (Kuźmińska-Bajor et al., 2015). Based on these findings, mice that developed luminescence and clinical symptoms of Salmonella infection including lethargy, ruffled fur, or ataxia were sacrificed. Bacteria and phages were delivered by oral gavage via a feeding needle, which allows the deposition of the microorganism directly into the mouse’s stomach. Mice at age 6–8 weeks were divided into 12 experimental groups, as shown in Table 1, and 7 mice were assigned to each group. In this study, four groups not treated with phage cocktail were prepared along with one infected with 10⁷ CFU of S. Enteritidis 327 lux strain in 100 μl of phosphate buffer saline (PBS; group 1), one treated with 10⁷ PFU/animal in 100 μl of PBS phage cocktail UPWr_S134 (group 2), one with phage cocktail UPWr_S134 at a dose of 10¹⁴ PFU in 100 μl of PBS per animal (group 3) and the uninfected and untreated animals (group 12). Eight Salmonella groups were infected with the dose of 10⁷ CFU per mouse and treated with the UPWr_S134 phage cocktail. Four of the Salmonella-infected groups were treated with the dose of 10⁷ PFU per mouse (groups 4, 6, 8, 10). Group 4 was treated with UPWr_S134 phage cocktail on day 7 post infection (p.i.). Group 6 was gavaged once a day starting on the 7th day p.i. to developed a typhoid-like disease course. Group 8 was firstly phage-gavaged on the day of infection, following the treatment from the 7th day post infection to the day of euthanasia of each mouse. Group 10 was treated with phage cocktail UPWr_S134, each day post infection with the first dose given 1 h after inoculation with S. Enteritidis 327 lux. Four groups infected with Salmonella were phage-treated with the dose of 10¹⁴ PFU per mouse (groups 5, 7, 9, 11) similarly differing in terms of the phage therapy scheme. Group 5 was treated with phages on day 7 p.i. Group 7 was gavaged once a day starting on the 7th day p.i. to the death of particular mice. Group 9 was firstly phage-inoculated on the day of infection, following the treatment from the 7th day post infection to the day of euthanasia. Group 11 was treated with the phage cocktail UPWr_S134 daily starting on the day of S. Enteritidis 327 lux infection. The plan of the phage cocktail administration is shown in Table 1. On day 21, all surviving animals were weighed and humanely killed by cervical dislocation, and all test animals were subjected to gross necropsy and dissected.

Animals were monitored every morning for bioluminescent signals using the NightOwl 983 imaging system (Berthold, Bad Wildbad, Germany) as described elsewhere (Kuźmińska-Bajor et al., 2015). Briefly, to confirm that the in vivo bioluminescence signals and changes in physical appearance accurately represent the presence of the bacterial load, as the bioluminescent signals, accompanied by clinical symptoms were observed, the mice were sacrificed and their livers, spleens and lungs were analyzed for the presence of CFU. The organs were homogenized in PBS, properly diluted, and plated on LB agar with 0.2 mg/ml erythromycin for bacterial number determination. The bioluminescent signals were always accompanied by the presence of S. Enteritidis 327 lux strain CFU in analyzed tissue samples, and CFU were not observed in livers, spleens and lungs isolated from mice that did not develop such signals.

For the in vivo experiments on experimentally infected chickens, 7-day-old, Salmonella-free ROSS 308 broilers obtained from a local farm were divided into four groups (10 birds/group) and housed separately in wire cages at an ambient temperature of 30°C. Sterile chicken feed Brojler Grower II (Tasomix, Poland) and water were provided ad libitum. All experimental work involving birds was approved by the Local Ethics Committee for Animal Experimentation (protocol code 114/2015; Wrocław, Poland). Group 1 was a positive control for Salmonella infection and birds were challenged directly to the crop with 10⁵ CFU of S. Enteritidis 327 lux strain in 1 ml of PBS per bird. Group 2 received only the UPWr_S134 phage cocktail at 1, 4, 7, 10 and 13 days of the experiment. Group 3 was infected with 10⁵ CFU/ml of S. Enteritidis 327 lux strain per bird and inoculated with a pool of 1 × 10⁷ PFU of each phage from the UPWr_S134 phage cocktail. At 1, 4, 7, 10 and 13 days post-infection (d.p.i.) birds from group 3 were treated with the UPWr_S134 phage cocktail. Group 4 was kept as an uninfected control (Table 2). On the 14th d.p.i., all birds were necropsied and attempts were made to isolate S. Enteritidis 327 lux strain and bacteriophages from the internal tissues of every bird. The spleen, liver, bursa of Fabricius and cecal tonsils were removed aseptically and the colonization level was assessed by estimating the number of CFU and PFU. The cecal tonsils and cecal contents were weighed, and homogenized in 0.2 ml of PBS. The bursa of Fabricius, spleen and liver were weighed after washing with PBS, and each organ was homogenized with 5, 5, and 50 ml of cold PBS, respectively. Ten-fold dilutions of homogenates were plated onto LB agar containing erythromycin (0.2 mg/ml) and plates were incubated overnight at 37°C. Identification of Salmonella was carried out by bioluminescence imaging and luminescent colonies were counted using bioluminescence imaging. For the PFU estimation tissue homogenates were centrifuged at 4000 × g to remove solid particles and then supernatants were filtered through 0.22 μm filters. The phage titration was estimated using a double agar layer plaque assay method. The mean CFU and PFU per 1 g of each tissue were calculated.

All statistical tests were performed on log₁₀ transformed data using STATISTICA version 13 software (TIBCO Software Inc.). To examine the variance in the titer of UPWr_S134 phage cocktail incubated in different pH and different temperature points and between treatments, the analysis of variance (ANOVA) followed by Tukey’s post-hoc test was applied. Student’s t-test was employed for statistical analysis of data obtained from in vitro assay in simulated gastric juice. To confirm that the UPWr_S134 phage cocktail treatment is a factor decreasing the colonization potential of S. Enteritidis, we compared the infection-free times of mice inoculated with S. Enteritidis 327 lux with or without phage treatment. Infection-free curves were obtained using the method of Kaplan and Meier, and the significance of differences was determined by the Gehan–Breslow–Wilcoxon test. Colonization of internal organs in both mouse and chicken models was analyzed using the Mann–Whitney U test with the non-Gaussian distribution. A value of p of < 0.05 was considered to be significant. All data were analyzed using two-tailed tests. To compare bacterial load (CFU) and phage load (PFU) recovered from chickens’ internal organs the nonparametric Mann–Whitney U-test was employed.

The UPWr_S134 phage cocktail was tested for 2-week stability at four temperatures, 4, 20, 36 and 42°C, and found to be stable at all tested conditions (Figure 1). Storage at 4°C had a negligible effect on the titer of phages (p < 0.05). The influence of 20°C within 2 weeks, and 36 and 42°C within 1 week on the titer of the phage cocktail was also negligible, but the phage titer showed a slight decrease after 2 weeks of incubation at temperatures of 36 and 42°C estimated to be 0.18 (p < 0.05) and 0.41 PFU/ml (p < 0.05), respectively.

The UPWr_S134 phage cocktail was tested against a wide pH range (2–13) during 1-h incubation, to determine its stability under acidic and alkaline conditions (Figure 2). In the acidic pH range (2–5), phages exhibited the most stability at pH 5, with minimal loss in viability of only 0.2 logs PFU/ml after 1 h of incubation (p > 0.01). Phages included in the UPWr_S134 cocktail survived at pH 3 and 4 for 1 h maintaining high activity estimated to be 5.7 and 6.6 log₁₀ PFU/ml, respectively, which corresponds to 72 and 82.4%, respectively, compared to the control. However, they did not survive at pH 2 over 1 h (Figure 2). Phages survived well at pH 7, 8, and 9 with no significant loss in viability after incubation (p > 0.01). Phages included in the UPWr_S134 cocktail were stable at pH 9, 10, and 11 with slight reductions in the population of about 4, 6, and 6%, respectively, during the incubation (p > 0.01), respectively. A significant phage titer reduction was observed for pH 12 and the number of phages was calculated to be about 19% of the control phage cocktail (p < 0.01).

UPWr_S134 phage cocktail’s ability to endure intragastric conditions was tested in vitro. The phages were very sensitive to low pH and the pepsin activity of control SGF and completely inactivated following 15 min of incubation at 42°C corresponding to the chicken deep body temperature (Figure 3). However, the addition of the antacid 14% CaCO₃ resulted in the complete neutralization of the negative influence of SGF conditions on phages (p < 0.01). We also performed the experiment of suppressing stomach acidity and pepsin activity in relation to phage survival by adding 30% chicken feed and we found that the adverse effect of SGF on the UPWr_S134 phage cocktail was completely abolished (p < 0.01; Figure 3).

To evaluate the in vivo effectiveness of the UPWr_S134 phage cocktail, a mouse model of acute typhoid fever caused by the S. Enteritidis 327 lux strain was applied. According to all the results presented in Table 3 and Figure 4, we found significant differences in colonization patterns between Salmonella-infected mice and mice from all groups treated with the UPWr_S134 phage cocktail. Oral administration of a dose of 10⁷ of S. Enteritidis CFU caused development of clinical symptoms in animals from control group 1, indicating murine typhoid like-disease by days 8 and 9. Mice administered with a single dose of 10⁷ or 10¹⁴ PFU of the phage cocktail UPWr_S134 on the 7th day post-infection with S. Enteritidis (groups 4 and 5, respectively) survived significantly longer than mice in the control group (p < 0.001). In group 4, 3 mice died on day 10, 1 mouse on day 11 and 3 mice on day 12, in group 5, 3 mice died on day 9, another 3 at 10 and 1 on day 11. A similar effect was seen in mice treated daily starting from the 7th day after S. Enteritidis 327 lux infection with both of the phage cocktail doses and with a close treatment schedule additionally treated 1 h post-infection (groups 8 and 9). In group 8, 1 mouse, 5 mice and 1 mouse died on days 10, 11 and 12, respectively. Whereas in group 9, on days 9, 10, 11 and 13 died 2, 3, 2 mice and 1 mouse, respectively. The difference in mice lifespan between the phage-treated groups 8 and 9 and control group 1 was also significant with p < 0.001. For groups 10 and 11 in which mice received phages daily starting on the 1st day post-infection at doses 10⁷ and 10¹⁴ PFU/mouse, respectively, the levels of statistical significance were set as p < 0.001 and p < 0.05, respectively. In the comparison of groups receiving different doses of UPWr_S134 phage cocktail according to the same treatment program, a significant difference was found only for the single-dose treatment schedule of groups 4 and 5 treated with 10⁷ and 10¹⁴ PFU per animal with p > 0.001, indicating greater effectiveness of the lower dose of the UPWr_S134 cocktail (Figure 5A). For the remaining 3 types of treatment schedules, no significant differences were found between 10⁷ and 10¹⁴ PFU/mouse doses (Figures 5B–D).

All of the challenged mice were sacrificed between bioluminescence signal detection and developing typhoid-like symptoms and the concentrations of strain S. Enteritidis 327 lux were determined in their livers, spleens, and lungs. The CFU of the Salmonella strain showed no statistically significant difference and were estimated to be 8.9 ± 2.82 log₁₀CFU in lungs, 10.1 ± 1.67 log₁₀CFU in spleens and 8.1 ± 2.11 log₁₀CFU in livers (p > 0.05).

To evaluate the effectiveness of the UPWr_S134 phage cocktail in poultry we used chickens experimentally infected with Salmonella. We examined the S. Enteritidis 327 lux load in the liver and lymphoid organs such as the spleen, the bursa of Fabricius, and the cecal tonsils at 14 d.p.i. Significant differences in Salmonella number between chickens infected with S. Enteritidis 327 lux from phage-treated group 3 and untreated birds (group 1) were observed in all lymphoid organs p < 0.05 (Figure 6A). No significant difference between group 1 and phage-treated group 3 was recorded for birds colonized in liver samples taken during necropsy (p > 0.05).

In comparison to group 2, the number of phages isolated from group 3 was significantly higher only in cecal tonsils (p < 0.05). In the bursa of Fabricius and liver samples of birds from Salmonella-infected group 3, in comparison to group 2, there was a higher number of organs positive for phage presence, with median values of 2.31 and 1.42, respectively. These differences were not statistically significant (p > 0.05; Figure 6B). Likewise, for liver samples median values of 1.21 and 1.42, for group 2 and 3, respectively were statistically insignificant (p > 0.05). No bacteriophages were detected in spleen samples taken from birds among both groups treated with phages (groups 2 and 3). Internal organ samples of control birds from group 4 were negative for both bacteriophages and Salmonella.