All animal procedures were carried out following European and national and regional regulations on animals used for experimentation and other scientific purposes. The protocols were evaluated and approved by the Ethics Committee at NEIKER (NEIKER-OEBA-2018–0001) and authorized by the Regional Council (BFA-38012).

All bacteria used for ex vivo and in vivo challenge assays were grown to the exponential phase at 37°C under aerobic conditions.

A schematic representation of the experimental design is shown in Figure 1 . New Zealand White female rabbits at 7 weeks of age were purchased from authorized experimental animal dealers (Granja Cunícola San Bernardo, Tulebras, Spain) and left on a 15-day acclimatization period. A total of 25 rabbits were divided into five groups of five animals each: non-vaccinated and non-challenged group (NC), challenged group (CC), subcutaneously vaccinated group with the commercially available vaccine Silirum^® (CV), a combined treatment group consisting of vaccination with the commercially available vaccine Silirum^® and probiotic administration (CV-PR), and finally an only-probiotic administration group (PR). The inactivated commercial vaccine was administered subcutaneously as a single dose (12.5 mg of antigen in 1 mL) on day 1. The probiotic was administered orally 3 days/week on alternate days during a total period of 4 weeks at a dose of 1 × 10⁷ CFU/day. Map oral challenge was performed on groups CC, CV, CV-PR, and PR on days 43, 44, and 45 by administering 1 × 10⁹ CFU/day. Blood sampling was performed monthly throughout the study to carry out different techniques depending on the sampling point: PMN isolation for bactericidal activity evaluation, reactive oxygen species (ROS) activity of PMNs and monocytes after stimulation with Map antigens, lactate quantification in plasma, IgG and IgA levels in plasma to Map protoplasmic antigen (PPA-3), and IgG and IgA levels in plasma to Map antigens. At the endpoint, 112 days (16 weeks) after the start of the experiment, all animals were euthanized by intracardiac administration of pentobarbital (200 mg/kg) (Vetoquinol^®) after deep sedation with xylazine (5 mg/kg) (Calier^®) and ketamine (35 mg/kg) (Merial^®). Tissue samples from the sacculus rotundus (SR), vermiform appendix (VA), mesenteric lymph nodes (LN), ileum (IL), liver (L), and spleen (S) were then collected and stored at −20°C for Map isolation. Moreover, samples from SR were fixed in neutral-buffered formalin for immunohistochemical analysis.

Peripheral polymorphonuclear neutrophils (PMNs) were isolated from 20 mL of blood passed through Histopaque^® 1119 as described previously (Ladero-Auñon et al., 2021). After the peripheral blood mononuclear cell (PBMC) layer was aspirated, and the entire volume of Histopaque^® 1119 was eliminated. PMNs localized in the pellet with remaining RBCs were aspirated to a new tube where hypotonic lysis was performed, and the samples were centrifuged at 300 g for 10 min. PMNs were suspended in rabbit leukocyte buffer and counted on a Bio-Rad TC20™ counter. This isolation protocol yields PMNs with >95% purity and >95% viability (Ladero-Auñon et al., 2021).

Cattle field Map 764 was sonicated and used for in vitro stimulation as described previously (Ladero-Auñon et al., 2021). Briefly, Map was grown in 7H9 OADC MJ broth to exponential phase at 37°C for 3 weeks. The culture was centrifugated at 10,000 g for 20 min, and after two washes with cold PBS, the pellet was suspended in PBS and sonicated for three cycles of 5 min at 18 W on ice. Protein concentration was determined on a spectrophotometer and was adjusted to 100 µg/mL.

Oxidative burst of PMNs and monocytes was evaluated using whole blood using the Respiratory Burst Assay kit (Abcam, ab236210) following the manufacturer´s instructions. Briefly, whole blood extracted with EDTA was mixed with 123 dihydrorhodamine (DHR) and incubated at 37°C in water bath for 15 min. Blood was mixed with PBS (non-stimulation), phorbol 12-myristate 13-acetate (PMA) (positive stimulus) and Map sonicate (specific stimulus), and incubated for 45 min at 37°C in a water bath. Afterward, RBC lysis was performed by adding the kit lysis buffer and incubating at 37°C in a water bath for 10 min. The samples were centrifuged at 500 g for 10 min at RT. The supernatant was aspirated, and the pellets were suspended in an assay buffer. Cell suspensions were analyzed using a CytoFLEX Flow Cytometer (Beckman Coulter). PMNs and monocytes were identified according to their specific FSC/SSC patterns. PMNs and monocytes producing reactive oxygen species (ROS) and converting DHR into rhodamine 123 were identified based on their fluorescence in the FITC channel. Data from the experiments were depicted as percentages of FITC-positive PMNs and monocytes of at least 10,000 events. Flow cytometry data were analyzed with CytExpert v2.3 software (Beckman Coulter). ROS percentages were expressed as the number of FITC-positive PMNs and monocytes divided by the total number PMNs and monocytes, respectively.

Isolated PMNs were seeded in tubes at 10⁷cells/mL in a total volume of 150 μL of RPMI 1640 (1×) with 15 μL of autologous serum and incubated for 15 min at 37°C. S. aureus was seeded at a multiplicity of infection (MOI) of 1 in a total volume of 165 μL. Immediately, 50 μL of volume (bacteria and PMNs) was removed to a new tube with 950 μL of cold PBS to stop the reaction and time zero was seeded. The rest was incubated in agitation for 30 min at 37°C. Afterward, tubes were put on ice to stop the reaction, and the samples were centrifuged at 150 g at 4°C for 10 min. The supernatant was removed and transferred to a new tube. The pellet was washed with 300 μL of cold PBS and centrifuged at 150 g for 10 min at 4°C. The obtained second supernatant was added to the first one, and 400 μL of PBS was added. The pellet was lysed with 150 μL Triton 0.1% disaggregated and incubated for 10 min at RT, and afterwards 850 μL of PBS was added to the lysed pellet. Finally, the extracellular bacteria from the supernatant and the intracellular bacteria from the lysed pellet were seeded. S. aureus CFUs were counted and recorded 18 h after seeding. The killing rate was calculated as 100 – [(extracellular bacteria CFU + intracellular bacteria CFU)/CFU grown in time zero) * 100].

Lactate levels were analyzed in plasma samples taken before treatment (BT) and at 1 month after treatment (PT1). Lactate levels were quantified using the Lactate-GloTM assay (Promega) following the manufacturer’s instructions. Briefly, 50 μL plasma was diluted 1:100, and 10 twofold serial dilutions of a lactate standard were run in duplicate in 96-well plates. The assay was performed with the addition of 50 μL lactate detection reagent (0.25 μL reductase substrate, 0.25 μL reductase, 0.25 μL lactate dehydrogenase, and 0.25 μL NADH) and 50 μL luciferin detection solution. After incubation in the dark for 1 h at RT, luminescence was recorded using a plate-reading luminometer (Promega). The amount of lactate was determined by extrapolating the sample luminescence values from the standard curve, and the increase in lactate concentration (PT1-BT) was analyzed.

Homemade indirect ELISA was performed using Map protoplasmatic antigen 3 (PPA-3) (Allied Monitor) as previously described (Arrazuria et al., 2016). Anti-rabbit IgA peroxidase (Abcam) and Protein G peroxidase (Sigma-Aldrich) were used as secondary reagents. Absorbance was measured at 405 and 450 nm using an automated ELISA plate reader (Multiskan EX^®). The reading obtained at 450 nm was subtracted from the reading at 405 nm to reduce optical imperfections in the plate. The results were expressed as a relative absorbance index (quotient obtained of the division of the mean absorbance of the sample by the mean absorbance of the negative control sample).

Map strain 832 was grown for 4 weeks in 7H9 OADC MJ medium at 37°C. At 8 OD, culture was washed with PBS 0.1 M, and full extraction buffer was added [buffer extraction, IP25 (protease inhibitor 25×), DTT (dithiothreitol)] and was placed in the Tissuelyser (Qiagen) with glass beads for Map protein extraction. Furthermore, 60 μg of protein extract was transferred to a polyvinylidene fluoride (PVDF) microporous membrane Immobilon^® (Millipore Ibérica, Madrid, Spain) by Trans-Blot^® Semi-Dry electrophoretic transfer cell (Bio-Rad Laboratories, Spain) at 15 V (Arrazuria et al., 2016). The PVDF membrane was cut in 0.5-cm large strips that were blocked overnight at 4°C with 0.1 M Tris-buffered saline (TBS) supplemented with 5% non-fat dry milk (TBS-M). Plasmas were diluted 1:50 in TBS-M before incubation with the membrane for 1 h at 37°C in slow agitation. After four 5-min washes with TBS, the membranes were incubated with recombinant protein G peroxidase (Sigma-Aldrich) 1:8,000 and Goat Anti-Rabbit IgA alpha chain peroxidase 1:5,000 (Abcam) in TBS-M for 1 h at 37°C. The membrane was washed four times more with TBS, and chemiluminescence reagent (ECL) was added. The immunoreactive bands were visualized by autoradiography, and after scanning the autoradiography, reactivity intensity was measured with Image J software (Arrazuria et al., 2016).

For Map isolation, 0.5 g of tissues was collected and cultured on Herrold’s Egg Yolk Medium (HEYM) supplemented with MJ as described by Arrazuria et al. (2015). Mesenteric lymph node was spliced in tiny pieces and weighed, whereas VA, SR, L, SPL, and IL were scraped for the collection of mucosae and weighed, and the samples were treated as described previously (Arrazuria et al., 2015). Briefly, hexadecylpyridinium chloride (Sigma-Aldrich) 0.76% decontaminated suspensions were centrifuged at 2,885 g for 10 min. The supernatant was discarded, and the pellet was washed once with sterile water. After a last centrifugation step in the same conditions, the pellet was resuspended in 2 mL of water, and four drops/tube were seeded. All seeded tubes were incubated at 37°C. Map growth was examined at 8, 12, and 16 weeks.

Isolated colonies were confirmed by a real-time multiplex PCR detecting IS900 and ISMap02 Map sequences (Sevilla et al., 2014).

Samples of SR were used for immunohistochemical analysis to characterize macrophage subpopulations and proliferation. For this purpose, different monoclonal antibodies against macrophage-expressed proteins were used ( Table 1 ) to identify M1 and M2 subpopulations and macrophage proliferation. The SR samples fixed in formalin were conventionally processed for histological examination. Deparaffination and antigen retrieval were performed at 96°C with citrate buffer pH 6 or pH 9 solutions using a PT Link system (Dako, Agilent Technologies, Glostrup, Denmark). Afterward, endogenous peroxidase was inactivated by incubating in 3% hydrogen peroxide in methanol for 30 min. Subsequently, the slides were incubated with different primary antibodies overnight at 4°C in a dark and humidified chamber. After washing in PBS, the slides were incubated with EnVision + Dual Link System-HRP (Dako, Agilent Technologies, Santa Clara, CA, USA). After washing with PBS, antibody localization was determined using 3,3-diaminobenzidine (Sigma-Aldrich Corp., Madrid, Spain). Sections were counterstained with hematoxylin briefly and dehydrated through graded alcohol series. Slides were mounted with DPX (dibutyl phthalate xylene) and observed under a light microscope.

Only immunolabeled macrophages present inside granulomas were considered and evaluated. For analysis, 20 representative fields through the whole slide containing granulomatous lesions were selected, photographed, and analyzed. The granulomas’ whole area and immunolabeled area were measured using Qupath software. For group comparison, the labeled areas were scored using an immunohistochemistry score (H-score) that incorporates both the staining intensity (i) and the percentage of stained cells at each intensity level (Pi), where i values range from 0 to 3 (0, none; 1, weak; 2, moderate; and 3, intense) and Pi values range from 0% to 100%. The H-score is ultimately calculated by summing up i multiplied by Pi as follows: (0xP₀) + (1xP₁) + (2xP₂) + (3xP₃). The H-score falls within the range of 0 to 300. The H-score was calculated by summing all areas of each intensity and multiplying by 1, 2, or 3 depending on the intensity as described before and dividing by the sum of the whole area of granulomas.

Significance of the differences among groups for all variables: PMN and monocyte producing ROS percentage, PMN killing activity, lactate level increase (the difference between BT and PT1), ELISA OD levels, Map tissue PCR C _t s, Map CFUs in tissues, and H-score were assessed using analysis of variance (ANOVA). When ANOVA showed significant differences, Tukey’s post hoc testing was used to make paired comparisons.

All statistical analyses were performed using Graph Pad Prism 9 statistical software (9.5.1), and differences among groups for all variables were stated at p < 0.05.

For the evaluation of the training of the innate response, ROS production of phagocytes after stimulation with Map antigens was evaluated in whole blood. The ROS production results at PT1 (1 month after treatment) and PT4 (3 months after treatment) are shown on Figure 2 . An increase of ROS production by PMNs and monocytes was observed only in the CV group against Map sonicate at 30 days post-treatment (PT1), although significant differences were only detected in the PMNs (p < 0.0001). This effect was maintained as a tendency at 50 days post-treatment PT2 (data not shown) and at 105 days post-treatment (PT4).

Heterologous protection against another pathogen that affects ruminant livestock was evaluated by measuring the bactericidal capacity of PMNs against S. aureus. The killing rate of PMNs at PT1 (1 month after treatment initiation) is shown on Figure 3 . The CV group showed the highest killing rate with significant differences compared to NC (p < 0.01) and CV-PR (p < 0.05).

In order to evaluate the immune activation produced by the different treatments, the lactate levels at BT (before treatment) and PT1 (1 month after treatment initiation) were measured ( Figure 4 ). The lactate levels in plasma were increased by vaccination, although significant differences were only detected in the CV-PR group compared to the NC group (p < 0.05).

Humoral immune response was assessed by PPA-3 ELISA and by immunoblot against Map 832 strain extract. The anti-IgG and anti-IgA reactivities against PPA-3 are presented in Figures 5A–F , respectively. It was observed that both vaccinated groups, CV and CV-PR, presented a significant increase of both IgG ( Figures 5A–C ) and IgA ( Figures 5D–F ) antibody levels, in contrast with the NC, CC, and PR groups that did not show reactivity against PPA-3. Both CV and CV-PR animals presented similar levels throughout the experimental time points, suggesting that the administration of the probiotic and/or the challenge with Map did not interfere in the humoral response generated by vaccination.

To further analyze the humoral response, anti-IgG and anti-IgA against Map antigens extracted from strain 832 were evaluated by immunoblot ( Supplementary Figure S1 ), and reactivity intensity was measured with Image J software and plotted ( Figures 5G–I ). Vaccination stimulated both IgG and IgA production in both vaccinated groups (CV and CV-PR), and the antibody levels increased throughout the experiment. The CV group showed the highest reactivity against Map antigens, followed by the CV-PR group. In the case of the CC and PR groups, the anti-IgG reactivity was increased after challenge and thereafter. It is worth highlighting that the PR group showed stronger anti-IgG and anti-IgA reactivity compared to the CC group at time PV4, indicating that the administration of the probiotic prior to challenge increased the antibody levels against Map.

A lower Map presence detected in tissues by PCR and/or isolation in agar was used as an indicator of protection of the treatments. The tissues that contained high levels of Map were VA ( Figures 6A, C ) and SR ( Figures 6B, D ), although Map was also detected in SPL, LN, IL, and L (data not shown). The group treated with the probiotic (PR) presented the highest bacterial burden for both techniques. The highest amount of Map DNA (lowest C _t ) was detected in this group in both VA ( Figure 6A ) and in SR ( Figure 6B ). The same behavior was observed in VA ( Figure 6C ) and SR culture ( Figure 6D ), showing the highest CFUs. The CC group presented higher CFU levels in VA ( Figure 6C ) and similar CFUs as CV and CV-PR in SR ( Figure 6D ). Significant differences were observed in C _t s in VA between CV and PR (p < 0.001) and between CV-PR and PR (p < 0.0001) and in SR between CV and PR (p < 0.05). Significant differences were observed in tissue culture, in VA (p < 0.05) and in SR (p < 0.05) between CV and PR. These results indicate that the probiotic alone administered in the conditions of the present study can be detrimental for Map elimination, even exacerbating bacterial burden. However, when administered in combination with the vaccine, it does not fully interfere with the protective effect as seen by similar bacterial burden levels observed in the CV-PR group compared to CV.

Macrophage polarization and proliferation were analyzed for CD163, IFNγ, calprotectin, and TNF-α in the granulomatous lesions of SR by immunohistochemistry ( Supplementary Figure S2 ). H-scores representing the intensity and percentage of immunolabeled areas for these macrophage markers in SR are presented in Figure 7 .

The high individual variability has impeded the observation of differences between experimental groups. The only clearly observed differences have been the decrease in IFNγ in the CV-PR group (p < 0.05) ( Figure 7B ) and the decrease in calprotectin in the CV group (p < 0.05) ( Figure 7C ). The immunolabeling with the rest of the markers suggests that both vaccination and probiotic administration alter macrophage profiles, but not significantly in the settings of this experiment, 3 months after challenge. The H-score for CD163 was higher in the PR and CC groups, although significant differences were not observed between groups ( Figure 7A ). The H-score of TNF-α ( Figure 7D ) was highest in the CV group, although significant differences were not found between groups.