BVDV is typically a pathogen of cattle but is not restricted to this host. However, BVDV was generally thought not to infect non-artiodactyls. Experimental intra-venous exposure of a range of species suggested that rabbits could propagate the virus[19, 20]. In addition, serological studies have suggested that natural infection of rabbits by BVDV may occur. However, experimental exposure by a natural route has never been tested and the role of the rabbit as a natural host for BVDV infection remained controversial. Nevertheless, the abundance of rabbits on livestock pastures in countries and nations that are trying to eradicate BVDV, such as the Republic of Ireland and Scotland, make an improved understanding of BVDV infection of rabbits advisable.
Therefore, rabbits were challenged with BVDV by different routes and the development of viraemia and virus-specific antibody responses were monitored (Figure 1). Intravenous inoculation of virus was used as a positive control because this had been reported previously as a successful route of BVDV infection of rabbits, whereas oro-nasal administration of virus by nebulisation represented a more natural route of virus entry, which allowed some degree of control over the dose of virus administered to each animal. Following previous studies which indicated that repeated pestivirus exposure facilitated interspecies transmissions[29–31], we re-challenged the IV and N group animals after 14 days. A third group of animals was exposed to BVDV by a potentially natural cattle-to-rabbit route, virus-contaminated hay, which was replenished daily for the first two weeks of the experiment. Since the rabbits were fed on dry pellets, the hay was primarily used as a source of long fibre and as bedding material.
The virus used for all exposures was not an established, tissue culture adapted, laboratory strain but a Scottish field isolate that was minimally passaged in order to maintain quasispecies diversity of the infecting virus. In addition to providing better representation of the natural situation of infection from a PI animal, using a virus with a degree of quasispecies diversity may be important for interspecies transmission as indicated by in vitro and in vivo experiments[32, 33].
None of the exposed animals showed any clinical signs or elevated body temperature upon infection, in accordance with the previous description. This seems remarkable in view of the viraemia and depleted GALT observed in some animals. Viral RNA was not detected in whole blood samples but could be detected in several buffy coat samples isolated from animals in the IV and N groups at day five (Table 1). This suggests that BVDV load during viraemia is low and primarily cell-associated. This is comparable to the situation in transiently infected cattle where viraemia occurs at a similar time after infection to that which we observed for these rabbits and is usually short lived. Notably, viraemia in cattle is not easily detectable in all cases and virus shedding is very limited. It is therefore not surprising that BVDV viral RNA was not detected in the oral swab samples from the rabbits. In contrast, BVDV RNA was detected in several organ samples at day five (Table 1) and in the spleen samples of all BVDV exposed rabbits in the IV and N groups. As the viraemia appears to be primarily associated with leukocytes, frequent detection in the spleen is not surprising and explains why intravenous virus transmission using spleen homogenate was successfully used in previous experiments[19, 20]. It is worth noting that the lowest Ct values (and thus the highest viral load) was detected in the sacculus rotundus part of the ileum (data not shown), which is the rabbit equivalent to the Peyer’s patches in cattle. This tissue was therefore used for virus isolation and, in all cases analysed, infectious BVDV was detected (Table 1). In PI cattle, the virus is typically found in the Peyer’s patches, which can suffer severe lymphoid depletion following the onset of mucosal disease. Mild to moderate depletion of the Peyer’s patches was also reported in transient infection of cattle with BVDV. Similarly, mild to severe depletion of the GALT in ileum and appendix was observed in all BVDV exposed rabbits where tissue quality was adequate (Figure 3; Table 1).
Following the detection of viraemia at day five, an antibody response was detectable in all BVDV exposed rabbits of IV and N groups by day 14 but not in the mock exposed animals (Figure 2). Since there is no validated BVDV ELISA for rabbit sera available, definition of seroconversion was difficult due to the lack of data to allow cut-off values to be established. However, the modified ELISA used here was a biphasic assay, so that false positive reactions due to non-specific antibody binding to components of the coating antigen or the reaction plate can be largely ruled out. Furthermore, no BVDV-specific reactivity was detected in any pre-exposure samples and no increase in S/P values was observed in plasma from mock-exposed rabbits. Even though we cannot exactly determine the time point of seroconversion in the rabbits, the onset of the antibody increase was observed between days five and 14. The time course of the antibody response measurable by ELISA therefore appears similar to that observed in cattle, where seroconversion is reported to occur between two to three weeks after infection. While the in-house ELISA used here is thought to detect mainly antibodies against the conserved non-structural NS3 (p80) protein, produced only during active virus replication, we cannot exclude a contribution from antibodies against structural proteins. However, retesting of plasma samples from day 28 in a modified commercial blocking ELISA specific for the detection of anti-NS3 antibodies (BVDV/MD/BDV p80 Protein Antibody Test Kit, Idexx) confirmed the results of the in-house ELISA (data not shown) and showed that the antibody reaction measured was not simply due to a hyperimmune reaction against virus particles. In contrast to ELISA, SNT detects mainly antibodies against the structural E2 envelope protein. Interestingly, while the ELISA results from rabbits seemed to be similar to the time course and degree of antibody response in transiently infected cattle, the SNT titres after 4 weeks were low (Figure 2). In cattle sera from field cases, titres are reported to be about ten times higher. However, levelling off of the SNT titre in cattle is only reached 10–12 weeks after infection. Thus, the neutralising titres in rabbits might have reached higher titres if the animals had been maintained for a longer time. The virus strain used for the SNT was of the same BVDV-1a subgroup as the virus used for challenging the rabbits, antigenic differences should therefore not be the reason for the low titres. The E2 protein is known to be the main determinant of host species selection in pestiviruses and the E2 coding region is known to be highly variable within and between BVDV isolates[33, 41]. Sequencing of virus recovered from the experimentally infected rabbits would show if changes in the E2 coding region upon interspecies transmission may contribute to the low neutralising titres. However, the virus neutralising antibody titres described previously in wild rabbits were very similar to those observed here in the experimentally infected rabbits. It is unlikely that these low titres would confer protection against re-infection, particularly with a different BVDV isolate. However, in addition to neutralising antibodies, cellular immunity is known to be important for protection against re-infection in cattle. Further experiments are necessary to determine the quality and duration of immune protection in rabbits.
In summary, our results indicate that rabbits are susceptible to infection by BVDV and that infection does not cause clinically apparent disease. The evidence of viraemia and the detection of anti-NS3 antibodies strongly suggest the virus can be propagated in rabbits. Importantly, evidence for virus propagation was found both after intravenous infection and also when the animals were exposed to the virus oro-nasally. Although infection was less successful in the rabbits exposed to virus-contaminated hay, even this indirect route of transmission led to an anti-BVDV immune response, as measured by ELISA, in four out of six animals (Figure 2). While some aspects of infection seem to be similar to infection in cattle, such as the time course of antibody development and the targeting of GALT organs, others are clearly different, such as the poor production of neutralising antibodies and the lack of clinical signs. To our knowledge this is the first report of BVDV infection by a natural route of animals other than even-toed ungulates and it highlights the flexibility of BVDV with regard to host range. However, in order to determine the role of rabbits as a potential reservoir for BVDV, further experiments are necessary to show whether persistently infected rabbits can be produced to generate cattle-independent chains of infection. Furthermore, BVDV ELISA analyses of sera from wild rabbits from regions with different cattle densities may provide additional information on the epidemiology of BVDV in rabbits in the field.