For this study, bacilli with antioxidant and antimutagenic properties isolated from chickens' feces were used. Fecal samples were collected from three cage poultry farms, a one-floor poultry farm, and three free-range poultry farms in the Rostov and Krasnodar regions of the Russian Federation. For free-range chickens, fecal samples were only taken from portions of the feces that were not in direct contact with the soil. Samples were collected in sterile containers and were transferred to the laboratory within 24 h.

Luria-Bertani (LB) medium was used to isolate intestinal bacilli. Bacilli were selected among other bacteria according to the morphology of the colonies. This was further confirmed using microscopy for the selection of spore-forming rods (PrimoStar, Zeiss, Jena, Germany). Individual strains were identified by MALDI-TOF mass spectrometry at the Rostov Research Institute of Microbiology and Parasitology, Rostov-on-Don. The study was performed on a Microflex LT instrument (Bruker Daltonics GmbH, Leipzig, Germany) using Biotyper (version 3.0) software (Bruker Daltonics GmbH). Overall, nineteen bacilli strains were selected from the litter.

To determine the safety of the selected strains, each was assessed for several potentially harmful properties, including antibiotic resistance. To determine hemolytic activity, all strains were plated on blood agar and incubated at 42°C for 48 h. To determine the sensitivity to antibiotics, the strains were seeded on appropriate solid nutrient media to a final concentration of 10⁶ CFU/g. Next, standard antibiotic disks (HiMedia, Mumbai, India) were placed onto the inoculated media. Amoxicillin (30 μg), ciprofloxacin (5 μg), ceftriaxone (30 μg), azithromycin (30 μg), erythromycin (15 μg), tetracycline (30 μg), and gentamicin (10 μg) (all HiMedia) were used, with 4 disks per plate with 2 replicates. The results were determined after 48 h of incubation at 42°C. Bacterial biosensors were used to estimate the DNA-protective and antioxidant activities of bacterial fermentates. We used E.coli MG 1655 carrying a plasmid with luminescence genes controlled by a stress-inducible promoter as biosensors. E.coli MG 1655 pRecA-lux reacts to DNA damage and E. coli MG1655 pKatG-lux reacts to oxidative stress. The details of this method are described in Zavilgelsky et al. and Prazdnova et al. (22, 23).

After completing the above-mentioned experiments, three strains were selected for feed supplementation as potential probiotics: B. subtilis KB16, B. subtilis KB41, and B. amyloliquefaciens KB54. The fermented milk product isolate, B. subtilis KATMIRA1933 [tested in our previous studies (24–28)], was obtained from the Russian National Collection of Industrial Microorganisms (RNCIM, Moscow, Russia) and used as a positive control.

The feed additives for these strains were prepared using solid-phase fermentation (29). Briefly, soybeans were inoculated with an overnight culture of the studied Bacillus strains and grown for 2 days at 42 °C. The fermented substrate was then milled and dried. The milling equipment used to prepare the solid-phase materials was sanitized with 95% ethanol. The spore content of each feed additive was 10⁸ CFU/g with a final spore count of 10⁵ CFU/g in the poultry feed.

A total of 200 newly hatched Cobb 500 broiler chickens were randomly divided into five groups, including the control group, with four replicates per group and ten broilers for each replicate. The control group was fed without probiotic supplementation to the diet (CON). The four treatment groups were labeled as T1, T2, T3, and T4, and their feed was supplemented with B. subtilis KATMIRA1933 (0.1%), B. subtilis KB16 (0.1%), B. subtilis KB41 (0.1%), and with B. amyloliquefaciens KB54 (0.1%), respectively.

This study was carried out for 42-day at the Center for Agrobiotechnology of Don State Technical University. Ten birds per cage were housed under the following light cycles: 24 h light for day 1, 23 h light/1 h dark for day 2, 18 h light/6 h dark for days 3–9, 15 h light/9 h dark for days 10–20, 12 h light/12 h dark for days 21–35, 23 h light/1 h dark for days 36–42. Cage sizes were: height – 50 cm, width - 95 cm, and depth - 70 cm. The planting density per bird was - 0.066 m². There was one vacuum drinker with a volume of 3 liters per cage. The feeding front was 9 cm with a total length of the feeder of 90 cm. The initial room temperature was fixed at 32°C on day 1 and gradually lowered until it reached 21°C on day 21. For days 22–42, the temperature was constantly held at 21°C. The air humidity was 60–65% for the duration of the experiment. The temperature and humidity regimens were maintained throughout the poultry house. Ventilation in the poultry house was natural supply and exhaust. The compound feed manufactured by the “BEST” company was used for feeding during three phases: STARTER (days 1–14), GROWER (days 15–28), and FINISHER (days 29–42) (Table 1). The feed and feeding of the animals corresponded to the GOST P 51899-2002 “Granulated mixed feeds. General specifications”¹ The feed compositions for all three phases included wheat, corn, soybean meal, wheat gluten, corn gluten, corn cake, soybean cake, sunflower cake, sunflower oil, tricalcium phosphate, lysine, methionine, threonine, protein-vitamin-mineral concentrates. The temperature, humidity, and light conditions were constant in all groups. Water and feed were provided ad libitum during the 42-day study period. No lethal cases were observed.

Total RNA was isolated from the spleen of one chicken for each replicate with four chickens in total per group (31). Isolation of total RNA from the samples was carried out by the phenol-chloroform method using the ExtractRNA reagent (Evrogen, Moscow, Russia). Spleen samples (100 mg) were homogenized in a mortar with liquid nitrogen. Total RNA was purified using the CleanRNA Standard kit (Evrogen) following the provided protocol. All actions were carried out following the instructions for the corresponding kits from the manufacturer. Primers targeting IL-10, IL-6, and reference (β-actin) genes were previously described in the literature (32).

The reverse transcription reaction was carried out using the MMLV RT kit protocol (Evrogen). qPCR from the obtained cDNA was performed using the qPCRmix-HS SYBR kit (Evrogen) on a Bio-Rad CFX96 amplifier. Experiments were conducted following the protocols of the kit manufacturers. The results were analyzed using the Bio-Rad CFX Manager software (Hercules, CA, USA).

Expression levels were normalized to β-actin, which was used as a reference gene. The change in the expression level of the target genes was calculated using the ΔΔCt method as a fold change in gene expression in experimental samples relative to the control sample. The difference was considered statistically significant at p < 0.00625, taking the Bonferroni correction into account (33).

IBM SPSS 26.0 (SPSS, Inc., Chicago, IL, USA) was used for data analysis. Data were not normally distributed according to the Shapiro–Wilk test. The Kruskal-Wallis test, followed by the Dunn-Bonferroni post-hoc test, was used to evaluate different medians among treatments. Results were presented as mean ± standard deviation. The statistical significance was determined as p < 0.05.

The study was approved by the Ethics Committee of the Don State Technical University, Rostov-on-Don, Russia (Protocol No. 67-43-2).

All nineteen selected strains were examined for safety. Hemolytic activity, mutagenicity (by lux-biosensors), and antibiotic resistance were examined. None of the strains were promutagenic, hemolytic, or resistant to the tested antibiotics, however, some of them showed prooxidant activity.

The strains were examined for antioxidant and antimutagenic activity using lux-biosensors. Figure 1 demonstrates the difference in the properties of Bacillus strains. For further studies, three strains out of nineteen isolated bacilli with the highest pooled antimutagenic and antioxidant activities were selected, which were B. subtilis KB, B. subtilis KB41, and B. amyloliquefaciens KB54.

There were no significant differences in basophils, rod neutrophils, segmented neutrophils, and lymphocytes counts among the groups. However, there was a significant increase in eosinophils in the T4 group and a significant decrease in monocytes in the T2, T3, and T4 groups (Figure 2).

Also, there were no significant differences in most of the studied biochemical parameters. However, there were some significant changes in glucose levels among the groups (Figure 3), which do not go beyond the values described in healthy chickens (34).

There were significant differences observed in BW and ADWG among the groups during the experiment period, with the highest values in the T4 group and the least in the CON group. FCR was significantly lower in the T4 group than in the CON group in almost every studied period. It was noticed that on the 35th day of the experiment, there were no significant differences in BW, ADWG, FCR, and FI. A significant difference in FI was observed on the 42nd day in the T3 group compared to CON. While there were no significant differences in FI for the T4 group, this group showed the highest values of BW and ADWG among all groups (Table 2).

Table 3 illustrates the number of microorganisms in the small intestine and cecum of the birds. Given the high variability in the number of microorganisms, the differences between the groups cannot be considered reliable. Chickens in all groups had values of bifidobacteria and lactobacilli reaching about 10⁶ CFU/g and 10⁷ CFU/g respectively.

A small number of Bacillus spores and vegetative cells were detected in the crop and proventriculus of the control group that did not receive probiotics. Most likely, they got there from the bedding. Already in the gizzard, they were not detected.

In the gut of group T1, trace amounts of spores were detected only in the proventriculus and gizzard. Vegetative cells were identified at very low numbers from the crop to the cecum.

In the T2 group, all bacilli cells were in a vegetative form. Their numbers increased rapidly from the small intestine, nearly reaching 10⁵ CFU/g in the colon compared to smaller amounts in the crop, proventriculus, and gizzard.

In the T3 group, vegetative cells were observed in the crop and proventriculus, reaching 1.1 × 10⁴ CFU/g. Their numbers dropped in the gizzard and then increased again, reaching a maximum of 1.1 × 10⁴ CFU/g in the cecum. The concentration of spores was up to 8.0 × 10² CFU/g isolated from the cecum and colon.

In the gut of the T4 group, vegetative cells were found in small numbers (9.0 × 10² CFU/g) but were not identified in the cecum. However, Bacillus spores were detected in the small intestine, reaching 3.8 × 10⁴ CFU/g in the rectum (Figure 4). It can be noticed that the number of both spores and vegetative cells differs from strain to strain, although the initial number of spores obtained with food was the same for all birds.

Potential probiotic in the T2 group had no significant effect on the expression of IL-10 (p = 0.016) and IL-6 (p = 0.02). The potential probiotic of the T1 group also did not significantly affect the expression of IL-10 (p = 0.2). However, they decreased the expression of IL-6 by 3.45 times compared to the control (p = 0.0016). Potential probiotics from T3 and T4 groups significantly increased the expression of both pro and anti-inflammatory cytokines. In the T3 group, there was a significant increase in IL-6 expression by 85.21 times (p = 0.0000001) and IL-10 by 29.02 times (p = 0.002). There was also a significant increase in the expression of IL-6 by 67.75-fold (p = 0.0000001) and IL-10 - by 43.51-fold (p = 0.002) in the T4 group (Figure 5).