The RB51 vaccine was obtained from a commercial source (Colorado Serum Co, Denver, CO). For serologic and proliferation assays, RB51 and Brucella abortus strain 2308 (S2308) was obtained from the culture repository at the National Animal Disease Center, Ames, IA. After growth on tryptose agar plates for 72 hr at 37°C and 5% CO₂, RB51 and S2308 suspensions were removed from plates by washing with 0.15 sodium chloride (saline), concentrated by centrifugation, and inactivated by γ-irradiation (1.4 × 10⁶ rads). Following irradiation, suspensions were washed in saline and stored in 1 ml aliquots at −70°C until use.

Pellets were prepared by mixing 10 mg magnesium stearate (for lubrication) with lyophilized RB51 from two 25 dose bottles obtained from a commercial source (Colorado Serum Co., Denver, CO, USA). The powder mixture was compressed using a manual pellet making device into projectiles of 3 mm diameter x approximately 10 mm length with each pellet weighing between 118 and 144 mg each.

Thirty-two bison heifers approximately 6 to 9 months of age were obtained from a brucellosis-free herd. One-fourth of the heifers (n = 8/treatment) were randomly assigned to the following treatments: (1) control; (2) single vaccination with 10¹⁰ colony-forming units (CFU) of the commercial RB51 vaccine; (3) parenteral vaccination with the commercial RB51 vaccine as calves combined with booster vaccination 10 months later with a similar dosage of RB51; and (4) surgical implantation of a dry dart formulation containing RB51. The parental RB51 vaccine was prepared in accordance with manufacturer's instructions and subcutaneously administered in the cervical region. The RB51 dry dart formulation was administered by making a small incision into the hide in the cervical region between the shoulder and head, inserting the dart approximately 4 cm into subcutaneous tissue, and delivery of the vaccine materials by plunger. The incision was closed with non-absorbable suture.

The concentration of RB51 in the inoculums were determined by dilution of the vaccines in saline and standard plate counts.

Blood samples were collected by jugular venipuncture prior to vaccination, and at 4, 8, 13, 16, 21, and 24 wks post-inoculation. Blood was also collected from all animals prior to administration of the RB51 booster inoculation, and at 4, 8, 12, 15, 22, and 27 wks after vaccination. Blood was allowed to clot for 12 hr at 4°C and centrifuged at 800 × g for 30 minutes. Serum was divided into 1 ml aliquots, frozen, and stored at −70°C. Antibody titers to Brucella were determined by a previously described ELISA in which γ-irradiated RB51 is used as antigen (8).

At 4, 8, 13, 16, 21, and 24 wks after initial vaccination, and 12, 15, 22, and 27 weeks after booster vaccination, blood was obtained from the jugular vein of all bison and placed into an acid-citrate dextrose solution. Peripheral blood mononuclear cells (PBMC) were enriched by density centrifugation (Sigma Diagnostics, Inc., St. Louis, MO, USA) and adjusted to 110⁷ viable cells per ml in RPMI as determined by trypan blue dye exclusion.

Fifty μl of each cell suspension, containing 5 × 10⁵ cells, were added to each of two separate flat-bottom wells of 96-well microtiter plates that contained 100 μl of RPMI 1,640 medium only, or 1,640 medium containing γ-irradiated RB51 (10⁵ to 10⁹ bacteria per well). Cell cultures were incubated for 7 days at 37°C in 5% CO₂. After 7 days incubation, cell cultures were pulsed with 1.0 μCi of [³H]-thymidine per well for 18 hr. Cells were harvested onto glass filter mats and counted for radioactivity in a liquid scintillation counter (Perkin Elmer, Hopkinton, MA, USA).

Bison heifers were pasture bred at approximately 26 months of age and pregnancy status and stage of gestation determined by rectal palpation. Pregnant bison were transferred to a Biosafety level 3 containment facility at 2 to 3 weeks prior to challenge where they were housed for the duration of the study. Based on palpation data, pregnant bison were intraconjuctivally challenged between 170 and 180 days of gestation with approximately 1 × 10⁷ CFU of B. abortus strain 2,308 (50 μl of inoculum per eye). Concentration of bacteria within the challenge inoculum was determined by standard plate counts.

Blood samples were collected by jugular venipuncture prior to experimental challenge, at 4 weeks post-challenge, and at necropsy after parturition or abortion. Blood was prepared as described previously and antibody responses to Brucella were determined by ELISA using γ-irradiated 2,308 as antigen (8).

Bison were euthanized within 72 hr after parturition by intravenous injection of sodium pentobarbitol (Sleepaway, Ft. Dodge Labs, Ft. Dodge, IA, USA). Fetal viability was assessed, and crown-to-rump length of the fetus was recorded. Maternal samples obtained at necropsy for bacteriologic evaluation included: lymphatic tissues (bronchial, hepatic, internal iliac, mandibular, messenteric, parotid, prescapular, retropharyngeal, and suprammamary), samples of milk and mammary tissue from all four quarters, maternal placentome and/or caruncle, spleen, liver, lung and vaginal and conjunctival swabs. Samples obtained from calves included: rectal swab, lung, liver, spleen, bronchial lymph node, and abomasal contents. Fetal and maternal blood was also obtained at necropsy for bacteriologic and serologic evaluation and mixed 1:1 with tryptose broth (Difco Laboratories, Detroit, MI, USA) containing 1% sodium citrate.

For bacterial culture, tissues were triturated in 0.85% NaCl (saline) using a tissue grinder and placed on tryptose agar containing 5% bovine serum or Kuzdas and Morse plates (9) as previously described (6, 10, 11). Swabs were directly plated on tryptose agar plates containing 5% bovine serum. One ml from blood cultures of each heifer was directly plated on tryptose agar containing 5% bovine serum. Remaining blood cultures were held at −5° C for 24 hr and then placed at 37°C and 5% CO₂ with 1 ml volumes plated onto tryptose agar containing 5% bovine serum after 7, 14, 21, and 28 days incubation. Following incubation at 37° C and 5% CO₂ for 72 hr, B. abortus was identified on the basis of colony morphology, and growth characteristics (9). Isolation of B. abortus was confirmed using a polymerase chain reaction procedure designed to identify individual species of Brucella (12).

Abortion was defined as the premature birth of a Brucella-infected, non-viable fetus at any time after B. abortus challenge. Mammary infection was defined as the recovery of the B. abortus from milk, mammary gland, or supramammary lymph node, whereas uterine infection was defined as recovery from placentome, vaginal swab, or internal iliac lymph node. Maternal or fetal infection was defined as the recovery of B. abortus from any maternal or fetal sample, respectively.

Bacteriologic and proliferation data were converted to log₁₀ with any values of 0 changed to 1 before transformation. Serologic, lymphoproliferative responses, and colonization data were compared by two-way general linear model procedure using treatment and time of sampling in the model, with means separated by a least square means procedure (SAS Institute Inc., Cary, NC, USA). Fisher's exact test was used to compare the rate of infection and abortion in vaccinates and controls after experimental challenge. Data are presented as mean ± SEM. Mean responses were considered significant when the statistical model indicated that P < 0.05.

Standard plate counts indicated that initial vaccines contained 1.8 × 10¹⁰ CFU for the commercial RB51 vaccine and dry dart pellets contained 9.9 × 10⁹ CFU of RB51. In a similar manner, the booster vaccine delivered 10 months after initial vaccination contained 1.1 × 10¹⁰ CFU.

Mean concentration of B. abortus in the experimental challenge inoculum was 0.98 × 10⁷ ± 0.15.

After initial inoculation with the commercial RB51 vaccine or the RB51 dry dart, vaccinates in both treatments demonstrated greater (P < 0.0001) serum humoral responses in an ELISA assay to RB51 antigens when compared to mean responses of non-vaccinates (Figure 1). Bison vaccinated with the RB51 dry dart had greater (P < 0.0001) mean optical densities at 4 and 8 weeks when compared to responses of control bison. In comparison, parentally vaccinated bison had greater (P < 0.001) mean humoral responses at 4, 8, 12, and 16 weeks when compared to responses of non-vaccinated bison. Mean responses of parenteral vaccinates were greater (P < 0.01) at 4, 8, 12, and 16 weeks when compared to mean optical density of sera from bison inoculated with the dry dart.

Prior to booster vaccination at ten months after initial vaccination, mean humoral responses to RB51 did not differ (P > 0.05) between any treatment (Figure 2). After parenteral administration of the RB51 booster vaccination, bison in this treatment demonstrated greater (P < 0.0001) mean serum responses on the ELISA assay at all sampling times after booster vaccination when compared to all other treatments (Figure 2).

After initial vaccination, PBMC from bison parenterally vaccinated with RB51 had greater (P < 0.05) mean antigen-specific proliferative responses at 8, 12, and 16 wks when compared to responses of non-vaccinates (Representative data in Figure 3). In comparison, PBMC from bison inoculated with the RB51 dry dart did not differ (P > 0.05) in proliferative responses to RB51 antigens at any sampling time when compared to responses of non-vaccinates. After initial vaccination, mean proliferative responses of bison parenterally vaccinated with RB51 were greater (P < 0.05) only at 12 wks when compared to mean responses of animals in the dry dart treatment.

After booster vaccination, antigen-specific proliferative responses of PBMC from booster vaccinates did not differ (P > 0.05) at any sampling time from mean responses of all other treatments (Data not shown).

Rectal palpation indicated 7 bison were pregnant in the control and single RB51 vaccination groups, 5 bison in the boostered RB51 vaccine group, and 6 bison in the RB51 dry dart treatment.

All treatments demonstrated greater (P < 0.01) mean humoral responses to S2308 on the ELISA assay after experimental challenge when compared to mean responses of sera obtained prior to challenge (Figure 4). Booster vaccinates had greater (P < 0.05) mean ELISA responses at 4 weeks after challenge when compared to the other 2 vaccination treatments. However, mean humoral responses of all 4 treatments to S2308 did not differ (P > 0.05) for samples obtained at necropsy after parturition or abortion.

When compared to rates in non-vaccinated bison, bison in all 3 RB51 vaccination treatments (single or booster vaccinated) had a reduced (P < 0.05) incidence of abortion (Table 1). In addition, bison administered single or boostered parenteral RB51 vaccinations had a reduced (P < 0.05) incidence of infection in uterine (placentome, vaginal swab, and/or internal iliac lymph node) and fetal samples (fetal lung, liver, spleen, gastric contents, bronchial lymph node or rectal swab). Bison booster vaccinated with RB51 also had a reduced (P < 0.05) percentage of animals with colonization of maternal samples as compared to the incidence of infection in non-vaccinated bison. In comparison, bison in the RB51 dry dart vaccination treatment did not differ (P > 0.05) from the control group in the percentage of animals in which Brucella was recovered from uterine, mammary, fetal, or maternal tissues at necropsy.

An efficacious vaccine would also demonstrate reduced colonization in tissues after experimental challenge. In the current study, bison receiving single or boostered parenteral vaccination had reduced (P < 0.05) mean Brucella colonization (CFU/gm) in placenta, mammary gland, various lymph nodes (mandibular, prescapular, and supramammary), and fetal tissues (lung, liver, spleen, and bronchial lymph node) when compared to mean colonization in tissues obtained at necropsy from non-vaccinated bison (Representative data in Table 2). In comparison, bison inoculated with the RB51 dry dart demonstrated a mean reduction in colonization when compared to controls in mammary gland and fetal spleen samples. When comparing to single parenteral RB51 vaccination, there was a trend for boostered vaccinates to have lower mean colonization in tissues, but differences were only statistically different (P < 0.05) in mammary gland samples.

Incidence of infection (percentage of animals that were culture positive) in individual tissues tended to follow similar trends as colonization data. In many tissues, single and boostered parenteral vaccinates, but not animals inoculated with the RB51 dry dart, demonstrated reduced (P < 0.05) numbers of animals from which Brucella was recovered at necropsy when compared to non-vaccinated animals (Table 2). There was a trend (P > 0.05) for many tissues of booster vaccinates to have a reduced percentage of animals from which Brucella was recovered when compared to bison receiving a single parenteral vaccination with RB51.