The M. hyorhinis isolate used during this study was isolated from the pericardium of a pig affected by pericarditis, originating from Hungary in 2019. The initial isolation was carried out using the filter cloning technique in Mycoplasma Experience Medium (Mycoplasma Experience Ltd., Bletchingley, UK), and the isolate was identified by the partial sequencing of the 16S-23S rRNA intergenic spacer region, using Mycoplasma genus-specific primers (Lauerman et al., 1995); the sequencing was performed on an ABI Prism 3100 automated DNA sequencer (Applied Biosystems, Waltham, MA, USA), followed by sequence analysis and BLASTN search (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The challenge material was prepared freshly for each challenge day by inoculating the fourth passage of the isolate 48 h prior challenge to Mycoplasma Experience Medium (Mycoplasma Experience) and incubating at 37°C. The color-changed broth was directly used for the challenge. The viable cell count determination was carried out on the day of the challenge. The number of color-changing units (CCU/ml) was calculated by broth micro-dilution from the highest dilution showing color change (red to yellow shift; Hannan, 2000).

Sixteen, 4-week-old Choice hybrid piglets were transported to the animal house of the Veterinary Medical Research Institute 6 days prior infection. The animals were obtained from a farm with low M. hyorhinis prevalence and high health status (free from brucellosis, leptospirosis, Aujeszky's disease, porcine reproductive and respiratory syndrome, swine dysentery, atrophic rhinitis, Actinobacillus pleuropneumoniae, Mycoplasma hyopneumoniae, lice, and mange). The M. hyorhinis-free status of the piglets was checked before the challenge by real-time PCR testing and Mycoplasma culture of nasal swabs.

Upon arrival, the animals were weighed and randomly divided into three groups with similar average weights. The animals were of mixed gender, which was not a factor when the groups were formed. The groups were housed in separate pens of 6 m² each. Humidity and air quality were regulated by the built-in ventilation system, and the age-appropriate temperatures were provided by heating panels. Feed and water were provided ad libitum. The experiment was approved by the National Scientific Ethical Committee on Animal Experimentation under reference number: PE/EA/746-7/2021.

Group IV-IV (n = 6) was inoculated by intravenous (IV) route on 0 and 1 day post-infection (0 DPI, 1 DPI) with 10 ml challenge material. Group IV-IP (n = 6) was challenged IV on 0 DPI with a 10-ml challenge material and intraperitoneal (IP) route on 1 DPI with a 20-ml challenge material. For the IV inoculation, the external jugular vein was used, while for the IP route, the challenge material was injected into the abdominal cavity in the left lower abdominal quadrant. On 0 DPI, the viable cell count of the challenge strain was 4.6 × 10⁵ CCU/ml, while on 1 DPI, it was 1.2 × 10⁶ CCU/ml. The total challenge dose was 1.66 × 10⁷ CCU/pig and 2.86 × 10⁷ CCU/pig in Group IV-IV and IV-IP, respectively. The controls (n = 4) were inoculated by IV route on 0 DPI. Two of these animals were inoculated by the IV route and the remaining two pigs by the IP route on 1 DPI. Animals in the control group received only sterile liquid media in the same volume as the challenged groups.

The animals were observed daily from settlement until the end of the study, at 28 DPI. Clinical signs of arthritis [swollen joints (mostly detected visually, complemented with palpation) and lameness] and respiratory disease (coughing or labored breath) were recorded. Body temperatures were measured daily from 2 days prior challenge. Body weight measurement, blood, and nasal swab sampling were carried out twice a week. Cotton swabs were used for collecting nasal swabs (Swab in tube Alum + Cotton, Deltalab S.L., Barcelona, Spain), and blood was collected from the external jugular vein. The schedule of events is summarized in Table 1. Average daily weight gain (ADWG) was calculated by subtracting the weight measured at arrival (6 days prior challenge, −6 DPI) from the weight measured at 27 DPI and dividing it by the number of days past (n = 33).

Nasal swabs for Mycoplasma isolation and PCR were taken twice a week from all animals throughout the study. Separate swab samples for Mycoplasma isolation and PCR were collected during necropsy as well (see below). For Mycoplasma isolation, swabs were cut into Mycoplasma liquid media (Mycoplasma Experience Ltd.), and vortex mixed to aid the release of bacteria into the media, which was then filtered using 0.45 μm pore size filters and incubated at 37°C until color change.

DNA extraction from the swabs and color-changed broths was performed using a ReliaPrep gDNA Tissue Miniprep System (Promega Inc., Madison, USA) according to the manufacturer's instructions. For the M. hyorhinis species-specific real-time PCR, previously published (Resende et al., 2019) primers targeting the 16S rRNA gene were optimized. Primer and probe sequences were the following: Forward primer 5′- CGT ACC TAA CCT ACC TTT AAG−3′, Reverse primer 5′- TAA TGT TCC GCA CCC C−3′, Probe 5′- FAM-CCG GAT ATA GTT ATT TAT CGC ATG AG-BHQ−3′. The PCR was performed using a Bio-Rad C1000 Touch™ Thermal Cycler, CFX96™ Real-Time System (Bio-Rad Laboratories Inc., USA). The PCR master mix consisted of 6 μl 2 × qPCRBIO Probe Mix No-ROX (PCR Biosystems Ltd., UK), 0.4 μl of each primer (10 μM), 0.2 μl probe, and 2 μl DNA in the final volume of 12 μl. PCR conditions were the following: 95°C for 2 min, 45 cycles of 95°C for 5 s, and 60°C for 20 s. In order to test the sensitivity of the developed assays, 10-fold dilutions of the DNA of the type strain (NCTC 10130) were used in the range of 10⁶-10⁰ template copy number/μl. Template copy number was calculated with the help of an online tool (http://cels.uri.edu/gsc/cndna.html) by measuring the concentration of DNA of pure M. hyorhinis culture using a Nanodrop 2000 Spectrophotometer (Thermo Fisher Scientific Inc., USA). The lowest DNA concentration giving a specific signal was considered the detection limit of the assay. The specificity was tested by including M. hyopneumoniae, Mycoplasma hyosynoviae, and Mycoplasma flocculare in the analyses.

Necropsy samples were also tested for the presence of M. hyopneumoniae (Wu et al., 2019) and M. hyosynoviae (Martinson et al., 2018a) using PCR. M. hyorhinis positive isolates were genetically characterized by multi-locus sequence typing (MLST: used as a costly but robust genotyping system) and multiple-locus variable-number tandem-repeat analysis (MLVA: used as a rapid and cheap genotyping system with high-resolution), according to previously published assays (Földi et al., 2020).

Joints of the carpus, elbow, tarsus, and stifle on both sides were opened and examined for signs of arthritis. The thoracic and abdominal cavities (pleura, pericardium, and peritoneum) were checked for serositis. Body condition, skin, subcutaneous tissues, musculoskeletal system, eyes and conjunctiva, nasal, and oral cavity, trachea, lungs, heart, lymph nodes, gastrointestinal system, liver, spleen, kidney, and brain were also checked for lesions. The scoring system of the gross pathological examination is detailed in Supplementary Table 1. Lesions of joints and serosa were scored to reflect severity based on previously described criteria (Martinson et al., 2018b). Total scores were calculated by summarizing all organ scores.

Swab samples for bacterial culture, M. hyorhinis isolation, and PCR were taken from the conjunctiva, lung, serosa, the four examined joints, and the brain. Joints on both sides were sampled with the same swab. During necropsy, viscose swabs were used (Swab in tube PS + Viscose, Deltalab).

Samples for histopathology were collected from the conjunctiva, choana, tonsilla, trachea, lungs (seven lobes), pericardium, heart, mediastinal and mesenteric lymph nodes, liver, spleen, kidney, joints, and brain (cerebrum, cerebellum, and brain stem). Tissue samples were fixed in 10% formaldehyde and embedded in paraffin, and then, 4 μm thick sections were cut and stained with hematoxylin and eosin (H&E) and examined by light microscope. Given the limited number of examined animals in the present study, the establishment of a general scoring system was not possible. Therefore, lesions were categorized based on the comparison of the severity of the histopathological changes with each other. The criteria of the scoring system used in this study are detailed in Supplementary Table 2.

The presence of bacterial pathogens other than Mycoplasma sp. was tested by culturing the necropsy samples on Columbia sheep blood agar (Biolab Inc., Hungary) and sheep blood agar supplemented with nicotinamide adenine dinucleotide (Sigma-Aldrich Co., USA) at the final concentration of 20 μg/ml. The agar plates were incubated in the presence of 5% CO₂ at 37°C for 48 h.

Sera were tested in duplicates by an in-house ELISA, using an antigen prepared according to the sarcosyl assay described previously (Stipkovits et al., 1993). In brief, to prepare the antigen, six clinical isolates of M. hyorhinis were propagated (Supplementary Table 3). After the color changes, the isolates were mixed, washed, and treated with 0.5% sarcosyl. The protein content of the antigen was determined using a Coomassie (Bradford) Protein assay kit (Thermo Fisher Scientific Inc.) according to the manufacturer's instructions.

Furthermore, 96-well ELISA plates were coated with the antigen diluted to the concentration of 1.25 μg/ml in phosphate-buffered saline (PBS, pH 7.4). After blocking with 1% gelatin from cold water fish skin (Sigma-Aldrich Co.), each well was incubated with a serum sample diluted to 1:100 in PBS, followed by a horseradish peroxidase-conjugated rabbit anti-swine immunoglobulin (Dako A/S, Denmark) diluted to 0.125 μg/ml in PBS. The reaction was visualized with tetramethylbenzidine substrate (TMB, Diavet Ltd., Hungary), and the optical density of the solution was measured at 450 nm using a Multiscan FC reader (Thermo Fisher Scientific Inc.).

Blood samples were centrifuged after collection, and the sera were maintained at −70°C. Each serum sample was thawed only once. Each plate contained a negative control (mix of the sera of each control animal taken at 28 DPI from this study), a positive control (mix of the sera of each animal in Group IV-IV taken at 28 DPI from this study), and a background control, where PBS was measured instead of the serum sample. The mean OD value of the background control was subtracted from the mean OD values of the samples and the controls (Terato et al., 2016). Samples were considered positive when the OD values were higher than three times the mean OD values of the negative controls (Lardeux et al., 2016). The mean of all negative controls used in the ELISAs was used to calculate this value.

Statistical analyses were accomplished using the R program (R Core Team, 2021). To compare the effect of the different challenge routes statistical analysis of the ADWG, pathological scores (separately for the joints, serosa of the pericardium, pleura, and peritoneum, and summary of scores), histopathological scores (separately for the joints, serosa of the pericardium, pleura, and peritoneum, and summary of scores), and ELISA results from the last sampling were performed. In the case of the pathological and histopathological scores, first, a Kruskal–Wallis non-parametric ANOVA test was carried out to determine whether the differences among the medians of the three study groups are statistically significant or not. If the results of the Kruskal–Wallis test were significant, Dunn's test was performed to determine exactly which groups are different by making pairwise comparisons between each group. Since multiple groups were considered at the same time, p-values were adjusted for multiple comparisons by the Bonferroni method. In the case of the ADWG and the ELISA results, instead of the non-parametric test, a one-way ANOVA followed by Tukey's multiple comparisons of means was performed after the normal distribution of the data was tested using the Shapiro–Wilk normality test.

No clinical alterations were detected in the control group throughout the study. No body temperature higher than 40.3°C was recorded during the study. One pig in Group IV-IP had a body temperature higher than 40°C on 3 consecutive days (4–6 DPI, Supplementary Table 4). No respiratory signs were recorded in the challenge groups.

Swollen joints were detected as early as 6 DPI in Group IV-IP and 8 DPI in Group IV-IV. Typically, the first swollen joint was one of the tarsal joints. By 16 DPI, all pigs in Group IV-IP had at least one swollen joint, three out of six pigs had two swollen tarsi, and in one animal, carpal joints were also affected. In Group IV-IV, by 20 DPI, swollen tarsal joint was observed in four out of six pigs (one side only), and in one animal, both tarsi were affected, while no swollen joints were detected in one pig (Supplementary Table 5).

Weight gain dynamics of the different groups are shown in Figure 1 and detailed in Supplementary Table 6. The average starting weights of the groups were 10.5 kg (SD 1.2), 10.3 kg (SD 1.2), and 10.3 kg (SD 1.1) in Groups IV-IV, IV-IP, and control, while at the end of the study average body weights of the groups were 18.0, 16.0, and 22.0 kg, respectively. Mean ADWG was 223, 170, and 350 g in Groups IV-IV, IV-IP, and control, respectively. Significant differences in ADWG were detected between control and IV-IV groups (p = 0.05) and control and IV-IP groups (p < 0.01; Supplementary Data 1).

The sensitivity of the reaction with the optimized primers was 10¹ copies/reaction, and no cross-reactions were detected for M. hyopneumoniae, M. hyosynoviae, and M. flocculare.

Nasal swabs of all animals were negative for M. hyorhinis by PCR and isolation at the beginning of the study (2 days prior challenge). After the inoculation of the pigs, one sample from each challenged group was positive by isolation which was also positive by PCR either at the same time or at different sampling times. These animals remained PCR positive for 2–4 consecutive sampling points. Furthermore, two animals in Group IV-IV and one animal in Group IV-IP were PCR-positive as well at one sampling point. All nasal samples of the control group were negative by PCR and isolation for M. hyorhinis throughout the study (Supplementary Table 7).

Samples collected from the conjunctiva and meninx during necropsy were negative for the tested mycoplasmas in all animals, while one lung sample in Group IV-IV was positive for M. hyorhinis by PCR. Three samples from different serosa (pleura, pericardium, and peritoneum) were positive by PCR as well in Group IV-IV. A high number of joint samples were positive by PCR in both challenged groups. In Group IV-IV, two out of six stifle, four out of six elbow, five out of six tarsus, and four out of six carpus samples were positive for M. hyorhinis by PCR. However, in Group IV-IP, two out of six stifle, four out of six elbow, four out of six tarsus, and one out of six carpus samples were positive. All samples from the control group were negative for M. hyorhinis (Supplementary Table 8). M. hyopneumoniae or M. hyosynoviae were not detected in any samples collected during necropsy.

During the challenge study, two nasal isolates and isolates from six necropsy samples (tarsal, carpal, elbow, and stifle joints) were collected and their genotypes were first determined by MLVA. Two re-isolates in Group IV-IP differed from the challenge strain on one allele (MHR444; Supplementary Table 9). They were micro-variants due to within-host evolution. The sequence types of these two isolates, two other isolates from the same animals, and one isolate from Group IV-IV were also determined by MLST. All the re-isolated strains showed the same sequence type (ST) with MLST as the challenge strain (GenBank IDs: OR250721-OR250756). MLST and MLVA trees are shown in Supplementary Figure 1.

None of the cultures of the necropsy samples showed growth of pathogenic bacteria that could also be associated with the lesions, other than M. hyorhinis.

Arthritis of at least one joint was observed in all pigs in the challenge groups. Mild-to-severe arthritis was found in all joints examined in one pig and in three joints examined in another pig in Group IV-IV. A single joint was affected in the remaining four animals in this group. Mild-to-severe arthritis was found in three, two, or one joints of two-two pigs in Group IV-IP. Arthritis manifested as serous or purulent inflammation (Figures 2A, B) and was detected most often in the tarsus (8/12) followed by the elbow (6/12), stifle (5/12), and carpus (4/12) on one or on both sides.

Diffuse, severe, chronic pericarditis presenting a large amount of connective tissue was detected in two animals in both challenge groups (Figure 2C). Additionally, mild or moderate chronic pleuritis presenting filaments of connective tissues were detected in two animals (Figure 2D) and mild chronic peritonitis presenting filaments of connective tissues occurred in one other animal in Group IV-IV.

Macroscopic scores of lesions in the affected organs are demonstrated in Figure 3. No gross pathological alterations were found in the remaining organs examined. No gross pathological lesions were detected in any examined organs in the control group. Body condition in all groups was normal. Detailed pathological scores are given in Supplementary Table 1.

Significant differences in pathological scores were detected when scores of joint lesions and total scores of groups were compared. Pathological scores in both challenge groups differed significantly from the control group (p = 0.03 and p = 0.02 regarding joint lesions, p = 0.02 and p = 0.03 regarding total scores for Group IV-IV–control group and Group IV-IP–control group, respectively) but not from each other in both cases. No significant difference was found when scores of serosa lesions were compared (Supplementary Data 1).

The results of the histological examination are summarized in Supplementary Table 2. The main alterations were detected in the joints and in the serosa of parenchymal organs in the thoracic and peritoneal cavities. In the joints, lesions were detected in six out of six animals in Group IV-IV and four out of six animals in Group IV-IP. Joint lesions were evident in several cases only with histological examination. A total of 24 joints presented histological lesions (Supplementary Table 2), and most of these lesions (16 cases) were given a score of 3 (Figure 4; Supplementary Table 2). In the joints, inflammation was characterized by infiltration of mononuclear cells and, in more severe cases, hyperplasia of the synoviocytes (Supplementary Table 2). Lymphoid follicles formed around the blood vessels (four out of six animals in both Group IV-IV and IV-IP; Figure 5A) and fibrin exudates in the joint cavity were observed in several infected animals (Figure 5B; Supplementary Table 2). One animal in Group IV-IP presented acute-subacute erosive synovitis associated with acute hemorrhages and frequent occurrence of fibrin exudates, blood cells, and neutrophil granulocytes in the joint cavity (Figure 5B). Lesions of the serosal membranes (pleura, pericardium, and peritoneum) were characterized by the presence of filamentous projections consisting of connective tissue and by different severity of serosal thickening caused by proliferating connective tissue (Figures 4, 6; Supplementary Table 2). Alterations of the pleura were detected in three out of six pigs in Group IV-IV and four out of six pigs in Group IV-IP (Figures 4, 6B). However, lesions of the epi- and pericardium were detected in three out of six animals in Group IV-IV and two out of six animals in Group IV-IP (Figures 4, 6A). Finally, lesions of the peritoneum were presented in the capsule of the liver (one out of six animals in Group IV-IV) and the spleen (one animal in both challenge groups). Multinucleated giant cells in the joints in two cases and in the pleura in one of those cases were detected in the animals from Group IV-IV (Supplementary Tables 2, 10). Moreover, in Group IV-IV, one animal showed mild-to-moderate acute rhinitis, and in another case, acute ulcerative conjunctivitis was detected (Supplementary Tables 2, 10). No lesions were detected in the other organs. Combined macroscopic and histopathologic lesion scores with further details of the histopathologic lesions are given in Supplementary Table 10.

Scores of histological lesions of affected organs are shown in Figure 4. Based on the statistical analysis, scores of joints and total scores differed significantly between groups. In both cases, significant differences were detected between Group IV-IV and the control group (p = 0.04 regarding joint lesions, p = 0.04 regarding total score; Supplementary Data 1).

All animals were serologically negative for M. hyorhinis at the beginning of the study. The positive serological response appeared in Group IV-IP on 5 DPI, and the OD value of one out of six pigs was higher than the set threshold. The positive serological response was recorded in Group IV-IV at the next sampling on 8 DPI. By 28 DPI, all challenged animals were ELISA-positive (Supplementary Table 11). The mean OD values of the groups throughout the study are demonstrated in Figure 7. Significant differences in OD of 28 DPI were detected between the control group and Group IV-IV (p < 0.01) and the control group and Group IV-IP (p < 0.01; Supplementary Data 1). Animals from the control group remained negative throughout the study.