Canine milk and colostrum were extracted aseptically from female dogs by a veterinary professional. Maternal milk and colostrum samples were collected aseptically by washing the nipple area with a soap and water solution to avoid contamination with other microorganisms from the skin. The veterinary wear appropriate sterilized gloves during the procedures. After collection, the samples were transferred to the laboratory of Food Microbiology, of the Faculty of Veterinary Sciences from the Universidad de Concepción (Chillán city, Chile). The samples were inoculated in Man Rogosa and Sharpe (MRS) broth and incubated in microaerophilia, at 37°C, for 48 h. The bacteria were then streaked on MRS agar and the plates were incubated under the mentioned conditions. As control strains, the probiotic Lacticaseibacillus casei Shirota and the strain Lactobacillus acidophilus of the commercial probiotic BIOPOWER (CPB) were used. In addition, an immunomodulatory strain from pig's breast milk (TUCO-4) belonging to the lactobacilli group was also used (23).

Canine LAB strains were studied based on their macroscopic and microscopic characteristics according to Gram stain, colony morphology, and catalase test. Their functional characteristics were also determined according to their resistance to bile salts (Oxgall 0.5 and 5.0% w/v), NaCl (2.0, 6.0 and 9.0% w/v), pancreatin (0.5, 1.0 and 2.0% w/v) and to acid (pH 3), as described previously (23, 24). Briefly, the strains were cultured in MRS broth with bile salts, NaCl, pancreatin and at pH 3 (adjusted with HCl 1M) and incubated in microaerophilia, at 37°C for 24 h. Viability was tested on MRS agar and incubated in the same mentioned conditions.

The effect of canine LAB strains against pathogens was analyzed in soft agar or semi solid agar (24). The bacterial pathogens used in these experiments were E. coli ATCC 25922, an E. coli enterotoxigenic strain (ETEC), Salmonella enterica ATCC 13076, a clinical isolate of Salmonella sp., and Clostridium perfringens NCTC 13170. LAB strains were washed twice with Butterfield's Buffer (5,000 rpm for 5 min), and adjusted to 0.5 McFarland. Drops were spotted on the surface of MRS agar and were allowed to dry for 45 min. The plates were incubated in microaerophilia, at 37°C for 48 h. Each pathogen strain was adjusted to 0.5 McFarland and 1 mL was added to 9 mL of soft agar (75%) or thioglycolate medium (for C. perfringens). Plates were incubated aerobically at 37°C for 24 h and anaerobically for C. perfringens cultures. The diameters of the inhibition zones were measured.

Canine LAB strains were tested according to presumptive safety, on sheep blood agar and in gelatin medium (23). The strains were streaked on sheep blood agar and incubated at 37°C for 48 h in microaerophilic condition. For gelatinase detection, the strains were streaked deep in gelatin medium and incubated in the same conditions mentioned above. After this period, the tubes with the cultures in gelatin medium were placed at 4°C for 2 h to verify liquefaction action. In addition, canine LAB strains were analyzed for their antibiotic resistance profile toward antibiotics discs (amikacin 30 μg, erythromycin 15 μg, vancomycin 30 μg, ciprofloxacin 5 μg, gentamicin 20 μg, amoxicillin 25 μg, tetracycline 30 μg and ampicillin 10 μg) that were placed on MRS agar surface with the tested LAB strain adjusted to 0.5 McFarland concentration. The plates were incubated at 37°C for 48 h in microaerophilia. The diameters of the inhibition zones were measured.

The selected canine strains (TUCO-16 and TUCO-17) were classified by conventional PCR to determine their membership in the actual lactobacilli group using the LbG primers (forward 5′AGAAGAGGACAGTGGAAC and reverse 5′TTACAAACTCTCATGGTGTG). The DNA was extracted according to the instructions of the supplier of the Mo bio commercial kit (Carlsbad, CA USA). L. casei Shirota was used as positive control and the strain Staphylococcus aureus ATCC 29213 as negative control (25).

The immunomodulatory activity of the canine LAB strains was evaluated in the canine macrophage cell line DH82. The porcine strain TUCO-4 and the probiotic CPB were used for comparison. DH82 cells were seeded in 24-well plates until 80% confluence. Canine macrophages were stimulated for 12 h with the different bacterial suspensions (10⁶, 10⁷, 10⁸ or 10⁹ CFU/mL) of TUCO-16, TUCO-17, TUCO-4 or CPB strains. Negative controls without treatment for 12 and 24 h were included. All conditions were performed in triplicate. Bacterial resuspensions were prepared with DMEM culture medium supplemented with 10 % fetal bovine serum (FBS), 2 mM L-glutamine, and non-essential amino acid (NEAA) solution. Cells were grown at 37°C with 5% CO₂. After the stimulations, RNA extraction was performed using the Trizol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA), according to the manufacturer's instructions. RNA samples were quantified with the Synergy HTX microplate reader (BioTek Instruments, USA), integrity was verified by agarose gel electrophoresis and the samples were stored at −80°C. Real-time PCR (RT-PCR) assays were performed to assess the relative expression of the pro-inflammatory cytokines TNF-α and IL-8, and the pattern recognition receptors (PRRs) TLR2 and NOD2. GAPDH expression was used as a normalizing gene (Table 1). For the synthesis of complementary DNA and real-time PCR, the Brilliant II SYBR^® Green QRT-PCR Master Mix, 1-Step commercial kit (Agilent, USA) was used in the AriaMx Real-Time PCR System (Agilent, USA). The results were analyzed using the comparative method of Ct (2^−ΔΔCt).

Female 5-week-old BALB/c mice were obtained from the closed colony kept at CERELA-CONICET (Tucumán, Argentina). Animals were housed in plastic cages in a controlled room (22 ± 2°C temperature, 55 ± 2% humidity) with a 12 h light/dark cycle. Mice were housed in plastic cages and environmental conditions were kept constant, in agreement with the standards for animal housing. Animal welfare was in charge of researchers and special staff trained in animal care and handling at CERELA. The minimal number of mice required for an appropriate statistical analysis was calculated with the help of the Biostatistics Laboratory of CERELA. Mice health and behavior were monitored twice a day. Animals were euthanized immediately after the time point was reached by using xylazine and ketamine. No signs of discomfort or pain were observed before mice reached the endpoints. No deaths were observed before mice reached the endpoints.

All experiments were carried out in compliance with the Guide for Care and Use of Laboratory Animals and approved by the Ethical Committee of Animal Care at CERELA, Argentina (protocol numbers BIOT-CRL/14 and BIOT-CRL/11).

The strains TUCO-16 and TUCO-17 were orally administered to different groups of mice for five consecutive days at a dose of 10⁸ cells/mouse/day (Supplementary Figure 1). The LAB-treated groups and the untreated control mice were fed a conventional balanced diet ad libitum. One day after the last LAB administration (day 6) animals were challenged with pathogenic E. coli or Salmonella.

In the first set of experiments, animals were orally infected with a mouse adapted enterotoxigenic E. coli (ETEC) K88 strain (1 × 10⁹ cells) diluted with 0.1 M carbonate buffer (pH 9.0) (26). Two days after the infection, the mice were sacrificed to collect the jejunum, ileum, spleen, and liver samples. The collected tissues were weighed and homogenized in BHI broth. Homogenates were plated on the kanamycin resistant MAC agar medium for ETEC counts. Results were expressed as log of colony-forming units (CFU) per gram of organ. Serum biochemical markers of injury as well as intestinal cytokines concentrations were also evaluated 2 days after ETEC challenge as described below.

In the second set of experiments, treated and control mice were challenged with 50 μl of 10⁷ cells/mouse of Salmonella typhimurium (20LD50) by oral administration (27). An aliquot (200 μL) of the intestinal pathogen from an overnight culture was placed in 5 mL of sterile BHI broth and incubated for 4 more hours (37°C, aerobiosis). The concentration of Salmonella was adjusted to 1 × 10⁷ CFU in PBS. Two days after the infection, the mice were sacrificed to collect the jejunum, ileum, spleen, and liver samples. The collected tissues were weighed and homogenized in BHI broth. Serum biochemical markers of injury as well as intestinal cytokines concentrations were also evaluated 2 days after Salmonella challenge as described below.

Lactate dehydrogenase (LDH) and aspartate aminotransferase (AST) activities were determined in the serum to evaluate gastrointestinal injury indirectly. Blood samples were obtained through cardiac puncture under anesthesia. LDH and AST activities, expressed as units per liter of serum, were determined by measuring the formation of the reduced form of nicotinamide adenine dinucleotide (NAD) using the Wiener reagents and procedures (Wiener Lab, Buenos Aires, Argentina) (26).

Intestinal fluid samples were obtained as described before (26). Briefly, the small intestine was flushed with 5 mL of PBS and the fluid was centrifuged (10,000 g, 4°C 10 min) to separate particulate material. The intestinal supernatant samples were kept frozen at −80°C until use. Tumor necrosis factor (TNF)-α, IL-6, IL-10, IL-15, interferon (IFN)-β and IFN-γ, chemokine KC (or CXCL1), and MCP-1 concentrations in intestinal fluid a were measured with commercially available enzyme-linked immunosorbent assay (ELISA) technique kits following the manufacturer's recommendations (R&D Systems, MN, USA).

The following programs for statistical data analysis were combined and used: software GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA, USA) and Excel (Microsoft 365). Statistical significance of the results was evaluated using one-way ANOVA and Dunnett's multiple comparison test. The level of statistical significance was defined as p < 0.05.

Five strains from canine breast milk (TUCO-17, TUCO-2P1, TUCO-2P2, TUCO-3P1, and TUCO-3P2), and one strain from canine colostrum (TUCO-16) were isolated and characterized for their functional properties (Table 2). Probiotic strains TUCO-4, Shirota and CPB were used for comparisons. All the isolated strains were Gram positive, of which only two strains presented bacillary morphology: TUCO-16, and TUCO-17. The functional characteristics of the strains showed that all of them were resistant to pH 3 at 24 h (Table 2). The TUCO-16 and TUCO-17 strains stand out, presenting resistance between 0.5 and 5.0% w/v of bile salts, resistance in the range of 2.0 to 9.0% w/v of NaCl, and were resistant at all concentrations of pancreatin used in this study, similar to the control strains (Table 2). Thus, the results showed that these two canine strains could resist gastric conditions. The strains TUCO-2P1 and TUCO-3P2 were resistant at 5% w/v of bile salts and in the rage of 2.0–9.0% w/v of NaCl, while the strain TUCO-2P2 did not growth in the presence of bile salts neither pancreatin concentrations but did grow at 2% w/v of NaCl. The strains TUCO-2P1 and TUCO-3P1 were resistant to pancreatin in the rage of 0.5–1.0% w/v and the strain TUCO-3P2 did not resist the pancreatin concentrations (Table 2).

The ability of canine LAB to inhibit the growth of intestinal pathogens was analyzed (Table 3). All the studied strains, including TUCO-16 and TUCO-17, showed inhibitory effect against C. perfringens, with inhibition halos > 30 mm of diameter. The TUCO-16 and TUCO-17 strains showed also inhibitory effect against the pathogenic E. coli and Salmonella strains (Table 3), and they were as effective as the Shirota and CBP strains. However, the canine strains were less efficient than the porcine TUCO-4 strain to limit the growth of pathogenic E. coli and Salmonella (Table 3). The strains TUCO-2P1, TUCO-2P2, TUCO-3P1, and TUCO-3P2 were also evaluated in their abilities to inhibit intestinal pathogens but no inhibition zones were found (data not shown). Considering these results, the strains TUCO-16 and TUCO-17 were selected for further experiments. The TUCO-16 and TUCO-17 strains were identified to belong to the actual lactobacilli group by conventional PCR.

The presumptive safety characteristics of the selected strains are summarized in Table 4. TUCO-16 and TUCO-17 strains did not show hemolytic or gelatinase activity. The canine strains presented antibiotic resistance profiles that were similar to the observed for the Shirota and CPB controls. TUCO-16 and TUCO-17 strains were sensitive to amoxicillin, erythromycin, ampicillin, and tetracycline.

For the evaluation of the immunomodulatory potential of TUCO-16 and TUCO-17 strains, DH82 canine macrophages were used. A significant increase in the expression of TNF-α was observed for the treatments with TUCO-16 and TUCO-17 at 12 h post-stimulation with the highest dose (Figure 1). Similarly, the highest doses of CPB augmented the expression of TNF-α. Of note, the two lower doses of TUCO-4 increased the expression of this inflammatory cytokine while no effect was observed with the two higher doses. A clearer dose dependent effect was observed when the expression of IL-8 was analyzed in canine macrophages after the stimulation with TUCO-16 and TUCO-17 strains (Figure 1). The TUCO-16 and TUCO-17 strains were as effective as the probiotic control TUCO-4 to increase IL-8. However, the canine LAB strains were less efficient than the CPB to enhance the inflammatory chemokine.

The expression of the pattern recognition receptors TLR2 and NOD2 were also evaluated in canine macrophages after the stimulation with TUCO-16 and TUCO-17 strains (Figure 1). The highest dose (1 × 10⁹ CFU/mL) of the canine strains significantly reduced the expression of TLR2. The same effect was observed for the probiotic controls TUCO-4 and CPB. In contrast, the doses of canine LAB strains inferior to 1 × 10⁹ CFU/mL upregulated the expression of TLR2. No significant differences were observed in the expression of NOD2 when basal levels were compared with those in macrophages stimulated with TUCO-16 and TUCO-17 strains (Figure 1). In contrast, NOD2 was significantly increased in cells treated with probiotic controls TUCO-4 and CPB.

The ability of the canine TUCO-16 and TUCO-17 strains to improve the resistance against ETEC infection was evaluated. A model of ETEC infection in mice (26) was used for this purpose. The challenge of mice allowed the colonization of ETEC in jejunum, ileum, liver, and spleen as shown by the bacterial cell counts in these tissues (Figure 2). The intestinal pathogen was not detected in blood samples in this time point post-infection. Mice preventively treated with TUCO-16 or TUCO-17 strains had significantly lower ETEC counts in jejunum and ileum. Furthermore, animals that received the canine LAB strain had undetectable levels of the pathogen in liver and spleen (Figure 2). In line with the improved resistance against ETEC challenge, mice treated with TUCO-16 or TUCO-17 strains showed significantly lower loss of body weight and reduced levels of the biochemical markers of injury LDH and AST when compared to controls (Figure 3). As reported previously (26), ETEC infection in mice induced an increase in the levels of proinflammatory cytokines IFN-β, IFN-γ, IL-6, and TNF-α (Figure 4) and chemokines KC, MCP-1, and IL-15 (Figure 5) in the intestine. In addition, an increase in the regulatory cytokine IL-10 was observed in the intestinal fluid of ETEC-challenged animals. Of note, mice preventively treated with TUCO-16 or TUCO-17 strains had significantly higher levels of intestinal IFN-β, IFN-γ, and IL-10 than controls. Canine LAB-treated animals also had lower levels of intestinal IL-6, TNF-α (Figure 4), KC, MCP-1, and IL-15 (Figure 5) than controls.

Finally, the capacity of the canine TUCO-16 and TUCO-17 strains to improve the resistance against Salmonella infection was evaluated in a mice model (27). The pathogen was detected in blood, liver, and spleen samples in control mice (Figure 6). Mice preventively treated with TUCO-16 or TUCO-17 strains had significantly lower Salmonella counts in liver and spleen. Furthermore, animals that received the canine LAB strain had undetectable levels of the pathogen in blood samples. The challenge with the pathogen augmented the levels of cytokines TNF-α, IL-1β, IL-6, IFN-γ (Figure 7) and the chemokines KC, MCP-1, and IL-15 (Figure 8) in the intestine. Significantly lower levels of the proinflammatory factors TNF-α, KC, and MCP-1 were found in mice preventively treated with the canine TUCO-16 or TUCO-17 strains. In addition, canine LAB stimulated a higher production of IL-1β, IL-6, IFN-γ, and IL-10 in the intestine of Salmonella-challenged animals. Of note, no differences were found in the levels of intestinal IL-15 when control and TUCO-16- and TUCO-17-treated mice were compared (Figure 8).