Animals and Experimental Design. This study and this paper are consistent with ARRIVE guidelines 2.0. This report adheres to both the ARRIVE essential 10 and the Recommended Set of additional reporting (https://arriveguidelines.org/). All local and federal laws and international standards were met or exceeded in this work. The protocol was approved by Texas Tech University Institutional Animal Care and Use Committee (#22020-03 March 7, 2022). Texas Tech University’s animal care program is fully accredited by AAALAC International, and the university program is inspected by the USDA for compliance. The IACUC reviews the protocol which must be approved before the work started. We do not have a local mechanism to register our protocol. Because this work is novel, no existing protocol for the animal work has been reported. All data will be available in supplementary materials linked to this paper.
The work was conducted at the University Swine Unit which is a smaller-scale modern commercial farm with slotted concrete floors and mechanical ventilation. Each pen has hanging chains for environmental enrichment (EE). However, for pigs with the sprayer, they would have had additional EE because they interact with the sprayer willingly. Because of the nature of the vaccine, pigs did not become ill. No pig illness or deaths were observed. No adverse events were noted. Several weeks after the end of the study, pigs were marketed as commercial pigs, that is, they were sold to a broker who sold them to a slaughter plant.
The animal work was performed at Texas Tech University (TTU, Lubbock, TX USA). The laboratory assays were performed at Iowa State University (ISU, Ames, IA USA). It was not possible to blind TTU scientists because of the nature of the study and the obvious treatments pig in each treatment group received. Each treatment group was obvious to even a casual observer. However, the samples were sent to ISU with unique labels that allowed ISU scientists to be blinded to treatment groups. All laboratory analyses were performed with collaborators blinded to treatment groups.
All pigs were PIC genetics with a white-line sow and a Duroc boar. This is a common genetic line of commercial pigs. Pigs had access to a corn-soy based commercial feed and water ad libitum. Three pens per treatment, each individual pen housed 2 castrated males and 2 females (12 pigs per treatment and 36 pigs in total). Pigs were 15 weeks of age at the start of the study. Each pig was identified at birth with a unique alphanumeric colored ear tag. Pigs were randomly assigned to pens and pens to treatment group. Average starting weights were different between treatment groups necessitating inclusion of weight as a covariate in the analysis (S3). All pens were in the same barn. Pens assigned to the same treatment group had fence-line contact. An empty pen or aisle was between treatment blocks or contemporaries not enrolled in the study (S2). Animals selected for this work were pigs at an age and weight about 2 months before slaughter. No pigs were excluded and a random sample of pigs was included.
The pig was the Experimental Unit because treatments were applied to the individual pigs. Twelve pigs were sampled per treatment. The sample size could not be estimated based on past work because such studies have not been reported. We used professional judgment to come up with a sample size of 12 pigs per treatment group for the 3 treatment groups. Because the vaccine caused no harm and might have prevented disease, there is no negative associated with evaluating too many pigs with the common commercial vaccine. The risk on sample size was that we might have used too few pigs to detect differences. We now know that our sample size was adequate to detect the differences that were observed.
The study was designed as a completely random design with a split plot over time (3 time points: day − 7, 14 and 21). Groups had an equal number of pigs (no morbidity or mortality were recorded) in each treatment. An examination of the data showed heterogeneity of variance among time points, making the full model not valid. Therefore, we analyzed treatment effects at each time point. At time zero (7 days prior to vaccination), some pigs had background antibody to Salmonella. Thus, for time points 14- and 21-day post vaccination (dpv), the time zero antibody levels were included as a covariate to equalize background antibody at time zero when evaluating 14 and 21 dpv data. A few oral fluid samples had inadequate volume for analyses. Thus, General Linear Models was used to evaluate treatments. The statistical model included three treatments (Control, Hand-drenched, and Self-vaccinated) at time zero and then the same model with the added time zero values and starting body weight as covariates. Within each time point, Least Squares means were separated using the Predicted Difference test within SAS Studio software (Release 3.8, 2020). The raw data and an example statistical analysis is provided in supplementary materials (S5).
Vaccine handling and administration. The commercial, lyophilized, bi-valent avirulent live Salmonella typhimurium-Salmonella choleraesuis vaccine culture was stored and handled per manufacturer recommendations (supplemental information, S1). Once reconstituted, the vaccine was then added to distilled water containing sodium thiosulfate and blue dye (Reload Pack) mixed at a ratio of 1 bottle to 3.8 L to aid in the administration of oral vaccine. To maintain vaccine culture viability, vaccine was added to drinking and stock solution water treated with Reload Pack and consumption complete within 6 hours of reconstitution. Each pig received at least 1 full dose (2 mL) of vaccine. For pigs in the sprayer pens, the maternal pheromone was added at 10 ppm which we have shown increased pig interest (18).
Three consecutive pens with fence-line contact within but not between treatments were assigned. Pigs were randomly assigned to pens. Animals spent 12 weeks with their pen cohort and mixing would result in their need to re-establish social hierarchy (so the stress of mixing was avoided). Because the avirulent live vaccine can be shed for a minimum of 4 weeks following vaccination leading to bias in the outcome measures. To minimize the possibility of cross-contamination of the non-vaccinated group, three pens housing 4 pigs each assigned to the non-vaccinated control group across the aisle from the vaccinated groups. No sham vaccination was performed.
Three pens housing 4 pigs each were assigned to receive vaccine via an adjustable height pen-mounted prototype EE device that pigs could operate by pressing a panel with their snout (S4). When pressed hard enough the panel triggered a spray up to 4 mL in volume to be delivered from the sprayer reservoir. Based on previous observations, 15-week-old pigs exposed to the sprayer for the first time would visit 2 times per hour. Therefore, a 200 mL total volume was placed in each sprayer reservoir at the ratio of 10 mL of vaccine (equivalent to 1.25 doses per pig), 1 mL of maternal pheromone to encourage active engagement with the sprayer and 189 mL of sodium thiosulfate treated water. One sprayer was placed in each of the 3 pens for a maximum of 5 h. After consumption, the reservoir was filled with water and left in the pen for an additional 18 h.
The pigs in the remaining 3 pens received a 5 mL total volume at the ratio of 2 mL vaccine and 3 mL sodium thiosulfate treated water. The specified volume per pig was aseptically drawn into a plastic dose syringe. Without restraining pigs, the syringe was inserted into the corner of each pig's mouth and administered per os.
Feces, serum, and oral fluid collection. Individual fecal samples were collected from individual pigs as it was passed per rectum or using a gloved finger to obtain at least 1 g feces per pig. Restraint was not required in order to collect feces. Individual pig specimens were submitted 7 days prior to vaccination (time zero) from all pigs. Samples from the sprayer group were tested individually on days 1, 3, and 7 dpv. Samples from the drench and control groups were pooled by pen for testing on days 1, 3 and 7 dpv. At 14 dpv, feces were collected from individual pigs then pooled at the farm by pen for each pen. Finally, a fecal composite collected from each pen’s dunging area was collected at 14 and 21 dpv. Feces was placed into a sterile leak proof container for transport. Each sample was culture for Salmonella under selective enrichment media (tetrathionate) broth and selective/differential agar plates following standard protocols at the AAVLD-accredited Iowa State University Veterinary Diagnostic Laboratory (14).
Individual blood samples were collected from the cranial vena cava. Pigs were restrained as appropriate for their age of production for blood collection using a hand-held hog snare. Blood was collected 7 days prior to vaccination (time zero) then 14 and 21 dpv. Blood was spun at 2500 rpm for 10 min in a bench-top centrifuge. Serum (3–5 mL) was transferred to leak-proof tubes for transport to the AAVLD-accredited veterinary diagnostic laboratory for antibody testing.
Individual oral fluids were collected using cotton balls tied to a string. Individual pigs were allowed to chew on cotton balls until moistened. Samples were collected 7 days prior to vaccination (time zero) then 14 and 21 dpv. Moist cotton balls were wrung out into plastic bags to obtain 1–2 mL then transferred to plastic screw cap tubes (1–2 mL) for transport to the AAVLD-accredited diagnostic laboratory serology section for antibody testing.
All samples were placed on ice or ice packs immediately after collection. The diagnostic laboratory best suited to perform testing was in another state. Therefore fecal, serum and oral fluid samples were shipped overnight on ice packs in a Styrofoam cooler to minimize temperature insult. Feces was shipped to the diagnostic laboratory within 18 hours of collection. Serum or oral fluids not shipped within 12 hours of collection were stored in an upright freezer at -20°C.
During descriptive assessment of the data, a non-vaccinated control pig (R261) housed in Pen 20 was identified as an outlier (supplemental information). Although inclusion didn’t result in within pen or treatment differences (P > 0.40).
Salmonella IgG/IgA indirect ELISA. A commercial Salmonella serum blocking ELISA (IDEXX Swine Salmonella Ab Test) was adapted into an indirect ELISA format (i.e., sample dilution, assay conditions, conjugate, substrate, and stop solution) for isotype-specific antibody (IgG, IgA) detection in serum and oral fluids. In brief, serum samples were tested at 1:20, while oral fluid samples were tested at a 1:2 dilution using kit sample diluent (100 µL final reaction volume). After 1h incubation at 37°C, plates were washed 5 times (350 µL/well) with kit washing solution, and 100 µL of corresponding conjugate was added to each well and incubated at 37°C for 30 min. Specifically, the kit conjugate was replaced with peroxidase-conjugated goat anti-pig (Fc) IgG (15) at 1:40,000 (serum) or 1/3,000 (oral fluid); or peroxidase conjugated goat anti-pig IgA (15) at 1/5,000 (serum/oral fluid) in conjugate diluent (20% fetal bovine serum, 0.05% Tween 20, phosphate buffered saline, pH 7.4). After another washing step, the reaction was visualized after 5 min incubation with 100 µL of tetramethylbenzidine-hydrogen peroxide (TMB) substrate solution per well (16) and stopped with 100 µL of stop solution per well (16). Optical density was measured at 450 nm using an ELISA plate reader and SoftMax Pro7 software (17). Antibody responses were expressed as sample-to-positive (S:P) ratios.