Materials and Methods

Study herds and vaccine—All of the herds included in this study were vaccinated with a modified-live virus vaccine^a containing the genotype 1, genogroup 1a NADL strain of BVDV. In addition, 1 beef herd was vaccinated with genotype 1, genogroup 1a BVDV-NADL and BVDV-C24V and with genotype 2, genogroup 2a BVDV-296.^b The vaccinated herds were commercial beef (n = 13) or dairy (4; Appendix). At a minimum, all heifers and cows in the herds studied were vaccinated before breeding. Persistently infected calves identified in the herds were found as the result of routine surveillance for PI cattle or were found as the result of the diagnostic process involving cattle with clinically apparent disease.

Viral stains—After investigation of herd vaccination practices and vaccine history of cattle in question, viral stains from qualifying herds were submitted from veterinary diagnostic laboratories (n = 8), or clinical specimens (blood or tissue) from PI calves were submitted directly from veterinary practices (9) to the DCPAH at Michigan State University (Appendix). Persistently infected calves were identified as the result of routine screening for PI cattle or because of clinical condition (disease or poor performance) consistent with PI. The PI status of all calves was confirmed by the submitting veterinary diagnostic laboratories by use of 2 successive positive test results for BVDV in clinical samples obtained from a calf or was similarly confirmed at the DCPAH by use of clinical samples submitted from veterinary practitioners. The PI status of the dams of all PI calves was tested at the laboratory involved in the initial diagnostic investigation or at the DCPAH. Failure to confirm that virus was not isolated from the dam, or that BVDV antigen or genetic material was not detected in tissue from the dam, excluded the use of viruses isolated from PI calves during the course of this study. All viral stains were passaged at least twice in bovine turbinate cells that were free of endogenous BVDVs. During each passage, cell cultures were observed for effects that would indicate the presence of CP BVDV or other CP viruses. If a CP effect consistent with presence of a CP BVDV was observed, the coisolated NCP BVDV was biologically cloned free of contaminating CP BVDV by propagation at limiting dilution, followed by serial passage to confirm successful cloning and absence of CP BVDV.

Genetic analysis—Viral RNA was extracted from infected cell cultures by use of a commercially available monophasic solution of phenol and guanidine isothiocyante.^c Extraction and purification of RNA were performed in accordance with manufacturer recommendations. A set of BVDV-specific PCR primers, forwardCCAMRGCACATCTTAACC and reverse-ACCAGTTGCACCAACCATG, was used to amplify approximately 1,150 bases of the viral genome that included > 100 bases from the 5' untranslated region and all of the N^pro and Capsid proteins. To perform the PCR reactions, commercially available reagents were used according to the manufacturer's instructions.^d The PCR product was subjected to agarose gel electrophoresis, then extracted from the agarose by use of a commercially available kit.^e The extracted product was submitted to the Michigan State University Research Technology Support Facility for DNA sequencing. Additional PCR primers were designed from the initial sequences derived from viruses to allow derivation of sequence from forward and reverse directions. Chromatograms from each sequence submission were edited and assembled by use of a commercial software program.^f The sequences were aligned and trimmed to the same length (1,068 bases) by use of commercially available software.^g Phylogenetic analyses were conducted by use of integrated software as previously described.²⁷

Results

Viral origin and biotype—The precise geographic origin of all viral stains was not known, as some viruses were obtained from veterinary diagnostic laboratories and the address of the cattle owner was not provided. The inability to collect information on farm address was expected because that information can be considered confidential by veterinary diagnostic laboratories. However, the state of origin of the cattle herds is known and included Montana (n = 4 herds, within approx 200 miles of each other), Wisconsin (3 herds, locations unknown), Kansas (1), Ohio (1), South Dakota (1), and Texas (1). States of origin of BVDVs represent a large geographic area, beef and dairy herds, and some herds with multiple PI calves (Appendix). All viruses that were analyzed in this study were NCP, and all dams of PI calves had negative test results for BVDV. All dams of PI calves were vaccinated before breeding, and some dams had been vaccinated multiple times over their life span by use of the BVDV-NADL strain of virus. One dam (KS-1 viral isolate) also had been vaccinated before breeding with a modified live-virus vaccine that contained a genotype 1, genogroup 1a BVDV (BVDVC24V) and a genotype 2, genogroup 2a BVDV (BVDV-296). The age of 10 of 17 dams was available, and those cows were 2 or 3 years of age when they gave birth to PI calves. During the course of this study, additional viruses were obtained from PI calves born into herds with a history of vaccination. However, those viruses were not included in this study because the field investigation could not confirm vaccination of the dam of the PI calf or the dam was not available to test for PI status. Those 2 pieces of information were considered necessary for the assumption that the virus being analyzed had eluded a recently induced immune response stimulated by vaccination.

Phylogenetic analyses—On the basis of nucleic acid sequence, 17 strains of BVDV in this study were classified as genotype 1, genogroup 1b (n = 1), which originated from a dairy herd in Ohio, or genotype 2, genogroup 2a (16). The genogroup 2a viruses were isolated from various beef or dairy herds located in several states. In herds where multiple BVDVs were obtained (Appendix), viruses from within a herd were genetically similar²⁸ (Figure 1). For example, viral isolates MT-3 through MT-6 were from the same herd in Montana; viral isolates TX-1, TX-2, and TX-3 were from the same herd in Texas; and viral isolates SD-1 and SD-2 were from the same herd in South Dakota. Genetic variation of BVDVs from within a herd varied from 0 to 7 bases out of 1,068 bases analyzed. Those base variations resulted in the change of 1 or 2 amino acids. In each instance, the variations were found toward the amino terminus of the N^pro protein coding region. This minimal genetic variation of viruses from the same herd likely indicates viral exposure with a single BVDV, for which the vaccination did not provide a protective immune response, as opposed to multiple exposures of each herd with multiple strains of BVDV.

View Full Size

The neighbor-joining method was used to generate this phylogenetic tree.²⁸ Bovine viral diarrhea virus 890 and BVDV 1373 are representative of genotype 2, genogroup 2a BVDV. Bovine viral diarrhea virus-NADL and BVDV-Singer are representative of genotype 1, genogroup 1a BVDV. Bovine viral diarrhea virus-Osloss and BVDV-CP7 are representative of genotype 1a, genogroup 1b BVDV. All other BVDVs in this dendrogram were isolated from PI calves born to vaccinated cattle. The optimal tree with the sum of branch length = 0.90625512 is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (2,000 replicates) is shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed by use of the Jukes-Cantor method and are in the units of the number of base substitutions per site.

Citation: American Journal of Veterinary Research 70, 1; 10.2460/ajvr.70.1.86

• Download Figure
• Download figure as PowerPoint slide

In contrast to the minimal genetic differences found among BVDVs isolated from multiple PI calves within a herd, substantial genetic differences were identified among the genogroup 2a viral isolates that originated from different herds. The nucleic acid sequences for the 16 genogroup 2a viral isolates varied from 3% to 10% (25 to 106 bases), compared with the BVDV-890 reference strain of virus. This variation in base sequence resulted in a 3% to 8% (9 to 26 amino acids) difference in amino acid sequence that was distributed over the length of the genomic segment analyzed. Compared with each other, the differences in nucleic acid sequence and amino acid sequence ranged from 3% to 11% and 3% to 9%, respectively. The wide geographic distribution and small number of viral stains did not allow identification of geographically defined genetic clusters of BVDVs, with the possible exception of a group of 6 viruses from Montana (MT-1 through MT-6; Figure 1). Those viruses were isolated from PI calves born 2 years apart on 3 farms separated by 100 to 200 miles. Genetic variation among that group of viruses was 3% to 5% for nucleic acid and amino acid sequence.

Discussion

The goal of this study was to collect and partially characterize field strains of BVDVs that may be able to elude the immune response of cattle stimulated by a well-defined genotype 1, genogroup 1a vaccine virus (BVDV-NADL strain). The field investigations for this study extended over parts or all of 3 calendar years: 2003 to 2005. However, most of the PI calves that were the source for viruses analyzed here were born in 2004. Viruses that ostensibly eluded the immune response stimulated by the NADL strain of vaccine virus were primarily genotype 2, genogroup 2a viruses. This was not an unanticipated finding considering that the vaccine virus was a genotype 1 virus and that genogroup 2a BVDVs are prevalent in the United States. It is well known that genotype 1 and genotype 2 BVDVs represent distinct serotypes of virus. Thus, genotype 1 and genotype 2 BVDVs share some antigenic sites, but the viral genotypes are readily differentiated in vitro by use of viral neutralization assays performed with convalescent sera obtained from recently infected cattle.

Field stains of genotype 2 BVDV analyzed in the current study represent a genetically diverse array of viruses. This indicates that evasion of the immune response stimulated by a genotype 1 vaccine virus is not likely restricted to a genetically narrow range of genotype 2 BVDV. It is likely that, under the appropriate circumstances, a wide range of genetically divergent genotype 2 BVDV exists in the field capable of escaping the fetal protection afforded by a genotype 1 vaccine. Circumstances that allow a field virus to cross the placenta of a vaccinated cow were not investigated in the current study.

Identification of a genogroup 1b virus that was not protected against with a modified-live genogroup 1a viral vaccine is of interest. In contrast to readily identifiable serologic differences between genotype 1 and 2 of BVDV, the 1a and 1b genogroups of BVDV commonly found in the United States are not easily separated by serologic testing of convalescent serum. This indicates that the genogroup 1a and 1b viruses share many antigenic sites associated with viral neutralization. Possibly, a BVDV vaccine virus would provide better cross protection from viruses in the field of the same genotype, but of a different genogroup, than it would for viruses of a different genotype. Consequences of the high mutational capacity of BVDV include the existence of a diverse array of genetic variants of the virus, shifts in viral biotype, antigenic drift, and shifts in viral virulence.²⁹ It is possible that extensive use of a vaccine virus over time would provide sufficient selective pressure to influence the prevalence of viral populations in the field and favor emergence of viruses for which the vaccine no longer provides a protective immune response.

Previous experimental studies^30–38 that assessed fetal protection provided by modified-live vaccine virus have shown that the vaccines studied afforded limited to complete protection. Overall, fetal protection appears to be most limited when the experimental challenge virus is of a genotype different from the vaccine virus.^{31, 37} This experimental finding is supported by a recent field investigation of outbreaks of bovine viral diarrhea in vaccinated herds.³⁹ In that study, genotype 2 BVDV was identified in an aborted fetus from 1 herd and in 2 PI calves from a second herd after vaccination with a genotype 1 modified-live vaccine virus. Also in that study, cattle and calves in other vaccinated herds that had outbreaks of BVDV had serologic evidence of exposure with a genotype 2 BVDV. A nonprotective vaccine-induced immune response likely is not limited to a difference in genotype between the vaccine virus and the challenge virus, as some studies^{30,32,33,35,37,38} reported fetal protection in vaccinated animals was circumvented after challenge with BVDV of the same genotype as the vaccine virus.

Our findings are in agreement with the aforementioned field investigation³⁹ and with many of the experimental studies. In our study, the genotype of BVDV most often isolated from PI calves (15/17) born into vaccinated herds was of a viral genotype different from the vaccine virus. Although it is tempting to attribute this finding to limited protection conferred by a genotype 1 vaccine virus against a genotype 2 field virus, the disproportionately high number of genotype 2 BVDV may simply reflect the prevalence of viral genotypes in the geographic areas represented in the current study. Viruses of a genotype that is similar to a vaccine virus might elude the immune response caused by vaccination, but detection of such vaccination failures would be limited if that viral genotype rarely occurred in the field. Two strains of BVDV in the current study were of the same genotype as the vaccine virus or as one of the vaccine viruses used in the herd. One of those isolates was a genotype 1, genogroup 1b BVDV (OH-1; Appendix) of the same genotype, but a different genogroup from the BVDV-NADL (genotype 1, genogroup 1a) vaccine virus. The other isolate was a genotype 2, genogroup 2a BVDV (KS-1), which came from a cow vaccinated with 2 vaccines. One of those vaccines contained BVDV-NADL, and the other vaccine contained a genotype 1, genogroup 1a virus (BVDV-C24V) and a genotype 2, genogroup 2a virus (BVDV-296).

Finding a genogroup 1b BVDV in PI calves born to dams vaccinated with a genogroup 1a BVDV supports the concept that future vaccines for BVDV should contain viruses representing all viral genotypes and genogroups to be maximally effective. Although it may not be practical, or necessary, to produce a universal vaccine containing all of the recognized genetic groups of BVDV to accommodate all areas of the world, it may be practical to formulate vaccines for regional use that contain viruses representing the commonly encountered viral genotypes and genogroups. Finding a genogroup 2a virus (KS-1) in a PI calf born to a cow that was vaccinated with a genogroup 2a vaccine virus is perhaps more troublesome. This might indicate that, under field conditions, there already may be viruses for which the immune response stimulated by a vaccine virus of like genotype and genogroup is nonprotective. In the viral genomic segment analyzed, the KS-1 viral isolate was 90% similar in base sequence with the BVDV-890 reference strain of virus and the BVDV-296 vaccine strain of virus (data not shown). A more extensive comparison of the KS-1 viral isolate with the vaccine viruses administered to the dam of the PI calf from which the KS-1 isolate was obtained would be of interest, but those analyses were beyond the scope of this study.

The role of vaccinated but nonimmunized cattle in occurrence of these PI calves was beyond the scope of this study but cannot be ruled out as a factor that would influence birth of a PI calf. A retrospective study of vaccine failure has 3 major assumptions that cannot be verified. The first is that the vaccine was handled correctly prior to vaccination. The second is that all cattle in the herd were appropriately vaccinated. And the last is that all cattle responded with an equally protective immune response to vaccination. Failure of any of these 3 assumptions could lead to an incorrect conclusion that the field strains of BVDV eluded immune responses of vaccinated cattle. It cannot be discounted that any of these factors may have been involved in one or more of these PI calves. Clearly, more work is needed to define the circumstances of vaccine failure and the limitations of vaccine-induced fetal protection.