Materials and Methods

Animals—Sixty 2-year-old crossbred beef heifers were enrolled in the study. Prior to enrollment, the heifers were vaccinated against 7 strains of clostridia,^a treated with a pour-on anthelminic,^b and tested and found to have negative results for antibodies against BVDV types 1 and 2 (SVN titer, < 1:2) and persistent infection with BVDV by immunohistochemical analysis of skin biopsy specimens.

In addition, 4 BVDV PI cows (2 with BVDV genotype 1b and 2 with BVDV genotype 2a), each from a separate herd of origin, were procured for the study. Persistent infection with BVDV was determined by serial isolation of virus from serum samples at 3-week intervals. Viral genotype was determined from derived nucleic acid sequences of approximately 1,000 bases from the 5′ end of the RNA from each virus. The nucleic acid sequences of the viruses isolated from the PI cows were distinct and readily allowed identification of each virus. A nested-multiplex PCR assay was used to genotype the isolated viruses.

All procedures involving cattle were approved by the Michigan State University Institutional Animal Care and Use Committee.

Housing—During the early phase of the study, the 60 heifers were housed in a single enclosed pen at a commercial laboratory animal procurement facility^c located in the state of Maryland. Heifers chosen for the challenge-exposure phase of the study (n = 30) were transported to Michigan State University and housed in a 1-hectare pasture at the College of Veterinary Medicine Veterinary Research Farm. The 4 PI cows were also housed in that same pasture. The pasture was isolated (separated by a minimum of 185 m) from all other cattle.

Vaccination—Prior to start of the study, heifers were randomly allotted to 2 treatment groups (30 heifers/group). On days 0 and 21, heifers assigned to group 1 (control cattle) were injected with saline (0.9% NaCl) solution, whereas heifers assigned to group 2 were vaccinated with a commercially available multivalent vaccine containing an inactivated genotype 1a and 2 BVDV fraction.^d In addition to BVDV, the vaccine contained temperature-sensitive mutants of infectious bovine rhinotracheitis and parainfluenza-3 virus, modified-live bovine respiratory syncytial virus, and a pentavalent leptospiral bacterin.

Breeding—On days 24 and 35, each heifer was administered a 5-mL injection of prostaglandin F_2α.^e Heifers were observed for signs of estrus from days 35 through 44, and those observed in estrus were bred by artificial insemination. In addition, all heifers were bred by artificial insemination on day 38 (approx 72 and 80 hours after the prostaglandin F_2α injection). Semen used for insemination was certified to be free of BVDV and bovine herpesvirus-1. On day 84 (approx day 46 of gestation), 20 of 30 control cattle and 19 of 30 vaccinated cattle were confirmed pregnant by transrectal palpation. Fifteen pregnant control cattle and 15 pregnant vaccinated cattle were randomly selected and transported to Michigan State University for use in the challenge-exposure phase of the study.

Viral challenge exposure and clinical assessment—On day 90 of the study (approx day 52 of gestation), the 4 BVDV PI cows were commingled with the 30 pregnant heifers to provide a continual viral challenge. The PI cows were allowed to remain in the pasture with the heifers through day 188 (approx day 150 of gestation). Cattle were monitored twice daily to evaluate general health and to detect signs of estrous activity or abortion. On days 84, 90, 120, 150, and 180 (approx days 46, 52, 82, 112, and 142 of gestation, respectively), pregnancy status and fetal viability of each pregnant heifer was determined by transrectal ultrasonography.

Cesarean section—On days 188 to 190 (approx day 150 of gestation), heifers were moved to the veterinary medical teaching hospital. Fetuses were obtained aseptically by caesarean section. Immediately after delivery, fetuses were euthanized by barbiturate overdose. Fetal carcasses were moved to a separate room and prepared for sample collection.

Sample collections and assays—Serum samples were obtained from the heifers on days 0, 21, 35, 63, 80, 90, 120, and 150 and from fetuses immediately following cesarean section and euthanasia. Serum samples were analyzed for SVN antibody titers against BVDV types 1 and 2 by use of standard laboratory procedures.¹⁶ The BVDV strains 5960 and 125C were used as reference strains in the assays for BVDV types 1 and 2, respectively. Blood samples were collected from the heifers on days 90, 97, 99, and 101 and analyzed for BVDV by use of virus isolation from peripheral blood mononuclear cells.¹⁶ Samples of the brain (cerebellum), liver, lungs, and spleen were aseptically collected from each fetus. Samples were collected within 10 minutes after fetuses were euthanized. Fetal tissue samples were analyzed for BVDV by use of virus isolation.¹⁷ In addition, the right ear pinna of each fetus was collected for analysis by use of an antigen-detection ELISA.¹⁸,^f Blood samples and nasal swab specimens were collected from PI cows on days 90, 120, and 150 for use in virus isolation and virus titration.¹⁹

Genotypic analysis—After viral isolation and biological cloning of viral isolates from the fetuses, a nested multiplex PCR assay was used to determine the genotype of multiple viral clones from each fetus.²⁰ The PCR amplification products were electrophoresed on a 2% agarose gel and stained with ethidium bromide. Amplified bands at 360 bp corresponded with type 1 BVDV isolates, whereas bands at 640 bp corresponded with type 2 BVDV.

Data analysis—All data were analyzed by use of a commercial statistics program.^g Serologic data for the heifers and fetuses were transformed (log base 2) prior to analysis. Titer values < 1:2 were assigned a value of 1 prior to transformation. A mixed linear model for repeated measures that included the random effects of block and animal and fixed effects of treatment, time point, and interaction between treatment and time point was used to analyze serologic results for the heifers. Serologic results for the fetuses were analyzed by use of a general linear mixed model that included the random effect of block and the fixed effect of treatment. Results of virus isolation from peripheral blood mononuclear cells of heifers and from fetal tissues were analyzed by use of the Fisher exact test. A value of P ≤ 0.05 was used to assess significant differences.

Results

All heifers remained healthy throughout the study. On day 171 (81 days after initiation of viral challenge exposure), it was confirmed that 1 control heifer aborted. No fetus was recovered. All other heifers maintained pregnancy, and the fetuses were judged to be viable by transrectal ultrasonography until day 180 (90 days after initiation of viral challenge exposure). One heifer in the vaccinated group had twin fetuses, both of which were delivered by cesarean section. One of the PI animals lost weight, developed diarrhea, and became lethargic 15 days after initial commingling with the pregnant heifers. That PI animal was euthanized 32 days after initial commingling with the heifers.

The SVN antibody responses of the heifers to BVDV type 1 and 2 were summarized (Table 1). All heifers were seronegative (SVN titer, < 1:2) on day 0 for BVDV types 1 and 2. Nonvaccinated control heifers remained seronegative through day 90 (day of initial viral challenge exposure). Geometric mean antibody titers for BVDV types 1 and 2 (determined by SVN) of vaccinated heifers were significantly higher than titers for control heifers on days 21, 35, 63, 80, 90, and 120. In addition, geometric mean antibody titers for BVDV type 1 (determined by SVN) of vaccinated heifers were significantly higher than titers for control heifers on days 150 and at the time of cesarean section (day 188, 189, or 190).

Table 1—

Serologic responses against BVDV type 1 and 2 on various days of the study for 15 nonvaccinated (control) and 15 vaccinated heifers.

View Table

Virus was isolated from nasal swab specimens obtained from PI cows on all sample days. Virus titers for nasal swab specimens determined for each PI cow on each sample day ranged from 1.7 to 4.5 log₁₀ CCID₅₀, and the mean titer for each sample day was 3.4, 2.9, and 4.1 log₁₀ CCID₅₀ on days 90, 120, and 150, respectively.

Results of isolation of BVDV from heifers were summarized (Table 2). Peripheral blood mononuclear cells from all heifers yielded negative results for BVDV by use of virus isolation on the day that PI cows were initially commingled (day 90). The proportion of heifers in which viremia was detected on at least 1 day was lower, but not significantly (P = 0.12) different, between vaccinated (3/15) and nonvaccinated (8/15) heifers. Genotype 2 BVDV was isolated 15 times, whereas genotype 1 BVDV was isolated only twice. Only genotype 2 BVDV was isolated from vaccinated heifers. Interestingly, both genotypes 1 and 2 BVDV were identified in 1 nonvaccinated heifer on day 99 of the study (9 days after initial viral challenge exposure with PI cows).

Table 2—

Number of heifers from which BVDV genotype 1 or 2 (or both) was isolated from peripheral blood mononuclear cells after initiation of viral challenge exposure on day 90.

View Table

Because of the difficulty of obtaining fetal blood samples, there was a low serum volume for several fetuses. These low volumes were inadequate to achieve the standard 1:2 dilution and necessitated that some samples be initially tested at higher dilutions (ie, < 1:3 or < 1:7). Three of 14 control fetuses (SVN reciprocal titers of < 3, < 7, and 10, respectively) and 3 of 15 vaccinated fetuses (SVN reciprocal titers of < 7, 2, and 16, respectively) had BVDV type 1 SVN titers ≥ 1:2, whereas all other fetuses had BVDV type 1 SVN titers < 1:2 (Table 3). Three of 14 control fetuses (SVN reciprocal titers of < 3, < 7, and < 7, respectively) and 2 of 15 vaccinated fetuses (SVN reciprocal titers of < 7 and 5, respectively) had BVDV type 2 SVN titers ≥ 1:2, whereas all other fetuses had BVDV type 2 SVN titers < 1:2. There was no statistical difference in serologic results between groups of fetuses.

Table 3—

Serum antibody titers, results of virus isolation, and antigen detection in tissues obtained from fetuses of vaccinated and nonvaccinated (control) heifers for BVDV genotype 1 or 2 (or both).

View Table

Results for isolation of BVDV from tissues obtained from fetuses were summarized (Table 3). One vaccinated heifer had twin fetuses. For statistical purposes, these fetuses were considered as 1 experimental unit. Bovine viral diarrhea virus was isolated from fetal tissues for 4 of 15 (26.7%) vaccinated heifers and 14 of 14 (100%) nonvaccinated heifers. In all virus-positive fetuses, BVDV was isolated from all 4 tissues evaluated. Similarly, an antigen-detection ELISA performed on skin biopsy specimens obtained from the fetuses revealed viral antigen in all fetuses that had positive results for viral isolation. The proportion of fetal infection was significantly lower in vaccinated heifers. Genotype 1 or 2 BVDV was isolated from 8 and 11 infected fetuses from the 14 nonvaccinated heifers, respectively. Genotype 1 BVDV was isolated from all 4 infected fetuses from vaccinated heifers, whereas genotype 2 BVDV was isolated from 2 of the infected fetuses from vaccinated heifers.

Discussion

Cattle that have persistent infection with BVDV are the major source of virus for susceptible herdmates. In field settings, exposure to BVDV is likely to be continuous as long as a PI animal has contact with the susceptible population. For field conditions, vaccines must be capable of stimulating an immune response that can withstand such a challenge exposure. In other studies^9–11,13,14 on the efficacy of BVDV vaccines to provide fetal protection, BVDV exposure typically was at a single time point by use of an artificial inoculation technique such as intranasal or IV installation of the virus. Although these studies are important and may in fact provide a challenge exposure that is more severe than natural challenge exposure, they do not simulate the prolonged exposure that would be evident in field settings. To our knowledge, the study reported here is the first to provide information on fetal protection by a BVDV vaccine with long-term (100 days) constant exposure to PI cattle as the virus challenge source during parts of the first and second trimesters of gestation. Other studies that used PI cattle as the virus challenge source were of short duration (14 days)¹² or during a period of gestation at which the fetus is not susceptible to becoming a BVDV PI animal.¹⁵ The challenge exposure method reported here that used long-term continual exposure to BVDV during the first trimester of gestation is likely a more realistic measure of vaccine efficacy.

In our study, all criteria for a challenge-exposure method were met. The control heifers remained seronegative for both BVDV types 1 and 2 until after the day of challenge exposure, and BVDV was detected in all tissue samples from all fetuses from those heifers. In contrast to intranasal or IV administration, in which a consistent and known amount of virus is dispensed directly into each animal, the challenge exposure method used in this study depended on exposure to PI cows that were shedding BVDV during the study period. After commingling with the PI cows, all nonvaccinated heifers seroconverted to both BVDV types 1 and 2, and their fetuses were infected with BVDV. Similarly, there was a substantial increase in antibody titer against BVDV types 1 and 2 in all vaccinated heifers. These data support the conclusion that all heifers in the study were exposed to BVDV.

Virus was isolated from a greater number of control heifers during the 11 days after commingling with PI cows. This is consistent with other experimental challenge studies^14,21–23 that have revealed a reduction in viremia in cattle vaccinated against BVDV. Viremia is likely a precursor to fetal infection, and so a reduction in viremic animals as a result of vaccination is beneficial. A possible weakness of the study reported here was the infrequent collection of samples to detect viremia in heifers after exposure to PI cows. This sample frequency may have missed a viremia of short duration. However, the primary outcome measure of the study was to evaluate fetal exposure to virus, so frequent collection of samples from heifers was not considered a necessity. Genotype 2 virus was isolated more frequently than genotype 1 from vaccinated and nonvaccinated heifers. This may have been attributable to a higher natural exposure rate to BVDV type 2 or a higher titer of virus in blood, which has been reported¹⁷ with BVDV type 2.

When compared with infection in control fetuses, fetal infection with BVDV was reduced by 73% in vaccinated heifers. This significant reduction in fetal exposure revealed that the vaccine was efficacious for constant exposure to BVDV through contact with PI cows. Fetal protection by administration of killed BVDV vaccines has been evaluated.^11,24–27 Those studies included the use of a monovalent vaccine and a homologous challenge virus administered intranasally or orally. In 1 study¹² conducted in the United Kingdom, investigators used a natural method for challenge exposure, similar to the one reported here. Efficacy of the vaccine in that study¹² was 100%; however, the viruses used for challenge exposure were the same genotype as the vaccine virus, and exposure to PI cattle lasted for only 14 days. Because of our study design, all that can be stated is that BVDV was detected in the fetuses on the day they were collected by cesarean section. To determine persistence of BVDV would necessitate serial isolation of the virus over time. However, given the timing of the challenge exposure and fetal collection by cesarean section, in combination with the distribution of virus in the infected fetuses, it is likely they would have been PI animals had gestation been allowed to continue to term.

An interesting finding was the detection of dual infection with BVDV types 1 and 2 in several fetuses. This has been reported after experimental infection in another study.¹⁰ However, to our knowledge, this is the first description of dual fetal infection following natural infection through exposure to PI cattle. Although this is likely a rare event in nature, it has important implications because multiple strains of virus circulating in the same animal could lead to genetic rearrangements with subsequent changes in virulence or antigenicity. This phenomenon deserves further study.

A commercially available multivalent vaccine containing an inactivated BVDV fraction administered before breeding significantly reduced the risk of fetal infection with BVDV for heifers continually exposed to BVDV PI cattle. However, protection was not 100%, which emphasizes the need to combine vaccination with identification and elimination of BVDV PI cattle and biosecurity measures as part of a comprehensive control program.