Culture of M. hyosynoviae was carried out using previously described protocols [1820]. The two clinical isolates used herein were isolated from clinical specimens known to be devoid of any other major swine bacterial and viral pathogens by the ISU-VDL (i.e., Haemophilus parasuis, Streptococcus suis, M. hyorhinis, Salmonella spp., Erysipelothrix rhusiopathiae, Porcine circovirus, and Porcine reproductive and respiratory syndrome virus). For isolation, a 1 : 10 ratio of filtered joint fluids from selected clinical specimens were inoculated into Difco pleuropneumonia like organism (PPLO) (Difco; BD Biosciences, USA) based media with turkey serum (15%) and mucin from porcine stomach (5%) broth and incubated at 37℃ for 3 to 7 days. After the initial culture, an aliquot of culture fluid was tested for the presence of M. hyosynoviae by qPCR [3]. To confirm the absence of contaminating bacteria, routine aerobic culture for non-mycoplasma species was also performed using both chocolate agar and blood agar plates. To develop challenge material, 1 mL of frozen stock of each isolate was adapted to Difco PPLO media supplemented with 15% heat-inactivated horse serum (Donor Horse Serum; Atlanta Biologicals, USA), termed Difco horse sera (DHS). All sera utilized throughout the study were heat inactivated at 56℃ (via water bath) for 30 min prior to use. Media was mixed and pH adjusted to 7.4 (via 1N HCl or 1N NaOH) before being 0.22 µm filtered and placed at 37℃ for 48 h to check for contamination by visual inspection for turbidity. Each media was then aliquoted into 1.8 mL in 30 tubes to set up a ten-fold serial dilution series in triplicate. Frozen stock M. hyosynoviae cultures at passage 2 (p2), 3491 and 3496, were thawed at room temperature, after which 0.5 mL was used to inoculate 5 mL of D-TS. Prior to DHS inoculation, Difco Turkey Serum was used for the initial passage to ensure growth from frozen stock. Following 24 h of incubation at 37℃, 0.2 mL of p3 were serially diluted into DHS and incubated. The challenge inocula were prepared in Difco PPLO media supplemented with 15% heat-inactivated horse serum (Donor Horse Serum; Atlanta Biologicals) to prevent anaphylaxis. To prepare challenge material, 1 mL of each DHS-adapted isolate was passed into 5 mL of fresh DHS, after which a final inoculum was passed into 25 mL of media for challenge administration. A 100 µL (10⁻¹ dilution) sample was subsequently spread plated onto blood and chocolate agar plates for purity verification. Next, 10 µL (10⁻² dilution) was plated onto brain heart infusion (BHI) agar plates to confirm mycoplasma growth by morphology. All plating was performed in triplicate throughout the experiment. Quantification of all inocula, as per color-changing unit-CCU per mL, was performed by dilution plating on DHS agar plates and inoculating broth [20]. Color-changing units per mL were calculated using arginine and phenol red, and a total of 10¹⁰ CCU per mL was used for IV and IN inoculation upon restraining all animals as previously described [720]. For inoculation, animals were restrained with a snare as previously described [10]. Despite the route and volume of inoculum administered, a 10¹⁰ CCU per mL of either isolate was used. IV inoculation was carried out using a butterfly catheter in an ear vein; whereas, for IN administration half of the inoculum was placed into each nostril using a sterile syringe [718].

Pigs were clinically monitored prior to and after inoculation for any clinical signs, particularly lameness. A single evaluator collected all observations throughout the study to ensure consistency. Lameness was scored from 0 to 3 (from slightly to severe clinical signs) as previously described [12]. Whole blood, nasal swabs, and tonsil scraping samples were collected prior to inoculation, and every 2 to 3 days post-inoculation (dpi) until necropsy. A fraction of all blood samples was then centrifuged at 2,000 × g for 10 min, after which serum was aliquoted into 5 mL tubes (BD Falcon; BD Biosciences, USA). Whole blood was collected using a 10 mL EDTA (K₂) tube (BD Vacutainer; BD Diagnostics, USA). Nasal swabs were collected using a single 15 cm sterile nylon flocked swab (Puritan; Puritan Medical Products, USA) sampling both nares each time, and placed into 5 mL polystyrene round bottom snap-cap tubes (BD Falcon; BD Biosciences) containing 2 mL of 1× sterile phosphate buffer saline solution (PBS; Invitrogen, USA). Tonsil scrapings were collected by holding the mouth open with a sterilized oral speculum and scraping the surface of the tonsil using a sterilized, blunt stainless steel spoon with an elongated handle [23]. Tonsil scraping contents were then collected from the steel spoon using a 15 cm sterile nylon swab (Puritan; Puritan Medical Products) and placed into 5 mL tubes (BD Falcon; BD Biosciences) containing 3 mL of sterile 1× PBS. All samples were stored at −20℃ until assayed.

DNA extraction was performed using magnetic beads (MagMAX Total Nucleic Acid Isolation Kit; Applied Biosystems, USA) in a semi-automatic system (KingFisher 96 Magnetic Particle Processor; Thermo Fisher Scientific, USA) according to the manufacturer's instructions, after which samples were frozen at −20℃ until assayed for M. hyosynoviae by qPCR [3]. All clinical specimens including nasal swab, whole blood, serum, tonsil scraping, and synovial fluids or swabs were tested in duplicate for M. hyosynoviae by qPCR [3519]. qPCR was also used to detect the presence (1) or absence (0), and to determine the M. hyosynoviae load per 2.5 µL of blood, serum, and joint samples [3].

Euthanasia was performed by administering a barbiturate overdose (Fatal-Plus; Vortech Pharmaceuticals, USA) via IV injection. At 12, 14 or 15 dpi, animals were necropsied, and the following joint cavities were sampled: scapula-humeral, humero-radial, coxi-femoral, femoro-tibial, and tibio-tarsal from the right and left side of each animal. Joint fluids or swabs were aseptically collected for M. hyosynoviae qPCR. Synovium tissues were placed into a 10% buffered formalin solution for histopathology evaluation. In addition, tonsil tissues were collected and ground in 10 mL of 1× PBS for M. hyosynoviae qPCR detection. Joint fluids were also plated on blood agar, chocolate agar, and BHI, and were cultivated aerobically at 37℃ for 3 to 5 days to check for extraneous bacterial growth. The histopathology scoring used was modified from a previously described method [6]. Histopathology scores for each sample were combined to provide a composite inflammatory score ranging from 0 to 15. Scoring was conducted by a board certified veterinary pathologist who was blinded to all treatments.

Statistical analysis of the data was performed using the GraphPad Prism 6 (GraphPad Software, USA) and MedCalc 15.6 (MedCalc Software, Belgium). Descriptive analysis was initially conducted to assess the overall quality of the data. The proportion of clinical cases, degree of lameness cases, and qPCR positive joints per group, were analyzed using a Chi-squared test (GraphPad Prism 6). One-way ANOVA was used to compare the percentage of clinical cases across treatments over time. Histopathology scoring was analyzed using a non-parametric one-way ANOVA (Kruskal-Wallis) followed by multiple group comparison using Dunn's test. The significant level across all analyses was based on a p < 0.05. Receiving operating characteristic curve analysis was used to determine the cut-off point for the histopathology score to differentiate diseased vs. non-diseased animals using MedCalc 15.6. After establishing the cut-off, the histopathology results were converted to a binary variable, with 0 indicating joints below the threshold histopathology score threshold of 1, and 1 indicating all other values. It should be noted that the binary value of 1 is different than the histological score of 1. It simply happens that the threshold established for the microscopic alteration equals 1. This was justified based on the microscopic score of 0 for all joints in the control group and the score of 1 in the disease group being a morphological alteration typically found in M. hyosynoviae infected animals (i.e., synovial cell proliferation). The qPCR test results were also interpreted as negative (0) or positive (1) [3]. The combination of both metrics (i.e., histopathology and qPCR) was then used to verify the odds ratio-OR to predict clinical disease. The OR was subsequently analyzed using a 95% confidence interval, which implies that if the value of 1 is included in this interval, there is absence of difference for disease vs. no disease occurrence given this diagnostic criterion.

All pigs used in the study were clinically healthy prior to challenge. Anaphylaxis did not occur upon inoculation. Lameness was clinically evident post challenge; however, it varied in severity in pigs from each inoculation group, regardless of isolate, volume, or route of inoculation (Fig. 1). Overall, a higher proportion of cases was observed in animals inoculated with isolate 3491 vs. 3496, with disease developing more rapidly upon IV inoculation. IN inoculation also led to robustness in the development clinical disease when using 10 mL of inocula. A significantly higher proportion of cases and longer duration of lameness were observed for isolate 3491 vs. 3496. Due to limited availability of CDCD-SPF pigs, the 5 mL volume for IN was only assessed for isolate 3491. A significantly higher mean lameness score was observed for animals inoculated IN with 10 mL of isolate 3491 compared to 5 mL, and this was greater than IV administration (Fig. 2). Overall, only mild lameness was observed for isolate 3496, despite the inoculation route.

At least 30% of the joints tested positive for M. hyosynoviae by qPCR at necropsy, regardless of the route of inoculation and isolate used in the challenged groups. Approximately 60% of all samples tested from two joints (humero-radial and coxi-femoral) were qPCR positive when combining data from all routes of inoculation for each isolate separately or together (Fig. 3). No significant differences were found when comparing either the proportion of M. hyosynoviae positive joints among anatomical location were observed within each inoculated group, or in the log10 bacterial load/2.5 µL of sample among positive samples across groups (data not shown). Conversely, significant differences in the proportion of M. hyosynoviae positive joints were found when inoculated groups were compared within each anatomical location. Specifically, a higher proportion of positive samples was found for the following groups given in three joints: 5 mL IN 3491, 10 mL IN 3491, and 10 mL IN 3496 for the humero-radial joint (p = 0.0238); 2 mL IV 3491 and 3496 for the coxi-femoral joint (p = 0.025); and 2 mL IV 3496 for the femoro-tibial joint (p = 0.0115) compared to other challenged and control groups (Fig. 4). The tonsils of all inoculated animals were consistently positive for M. hyosynoviae; however, bacteremia was very rarely detected in whole blood or serum (data not shown). All control animals were negative for M. hyosynoviae in the joint cavity, tonsils, blood, serum, and nasal cavity. No other bacteria were found in the joints of inoculated animals.

All challenge groups had at least one joint that was significantly more affected than the control group. The coxi-femoral joint was most consistently more affected in the challenge group than the control group. Specifically, with the exception of the group receiving 10 mL IN of 3496, all challenged groups had significantly higher histopathological alterations in the coxi-femoral joint than the non-challenge group (Fig. 5). When combining the histopathological and qPCR findings into a diagnostic metric, a significantly higher likely of developing clinical disease vs. none was associated with being qPCR positive and having histopathological lesions in joints except for the femoro-tibial joint (Table 1).