Use of multivariate analysis to evaluate antigenic relationships between US BVDV vaccine strains and non-US genetically divergent isolates

Authors: MOSENA, ANA CRISTINA - Federal University Of Rio Grande Do Sul; Falkenberg, Shollie; Ma, Hao; Casas, Eduardo; Dassanayake, Rohana; BOOTH, RICHARD - Royal Veterinary College; DE MIA, GIAN MARIO - The Istituto Zooprofilattico Sperimentale Delle Venezie (IZSVE); SCHWEIZER, MATTHIAS - University Of Bern; CANAL, CLAUDIO - Federal University Of Rio Grande Do Sul; Neill, John

Submitted to: Journal of Virological Methods
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/9/2021
Publication Date: 10/8/2021
Citation: Mosena, A.S., Falkenberg, S.M., Ma, H., Casas, E., Dassanayake, R.P., Booth, R., De Mia, G., Schweizer, M., Canal, C., Neill, J.D. 2021. Use of multivariate analysis to evaluate antigenic relationships between US BVDV vaccine strains and non-US genetically divergent isolates. Journal of Virological Methods. 299. Article 114328. https://doi.org/10.1016/j.jviromet.2021.114328.
DOI: https://doi.org/10.1016/j.jviromet.2021.114328

Interpretive Summary: Providing protection against bovine viral diarrhea virus (BVDV) is challenging due to the ability of BVDV to infect the fetus, and the increased level of protection needed to protect the fetus. Typically, virus neutralization (VN) assay and the subsequent VN titer is used determine antigenic differences among BVDV isolates, but interpretation of the data can be difficult. Interpretation of the data currently relies on comparison or VNT values among each isolate and interpretation of what the different values mean. Data from this study utilized an analysis that generates graphical scatter plots for visualization of VN results to determine antigenic relationships between vaccine strains and BVDV field isolates. Calves were exposed to six BVDV strains currently contained in vaccine formulations, and serum was obtained from these calves and used for VN to measure the neutralizing antibody titers against 15 U.S. and 23 non-US (Switzerland, Italy, Brazil, and the UK) BVDV field isolates. The results demonstrated spatial patterns in the graphs that were suggestive of antigenic differences among the BVDV isolates used in the assay. Some BVDV isolates had very distinct spatial patterns in the graphs that could suggest extremely antigenically divergent isolates. This analysis and graphs provides an alternative and more efficient means to analyze large VN datasets to visualize antigenic relationships between BVDV isolates.

Technical Abstract: Bovine viral diarrhea virus (BVDV) comprises two species, BVDV-1 and BVDV-2, but given the genetic diversity among pestiviruses, at least 21 subgenotypes are described for BVDV-1 and 4 for BVDV-2. Genetic characterization is generally accomplished through complete or partial sequencing and phylogeny, but genetic characterization is not a reliable method to define antigenic relationships. The traditional method for evaluating antigenic relationships between pestivirus isolates is the virus neutralization (VN) assay, but interpretation of the data to define antigenic relatedness can be difficult to discern for BVDV isolates within the same BVDV species. Data from this study utilized a multivariate analysis for visualization of VN results to analyze the antigenic relationships between vaccine strains and field isolates from the US, Switzerland, Italy, Brazil, and the UK. Polyclonal sera were generated against 6 BVDV strains currently contained in vaccine formulations, and each serum was used in VN’s to measure the VN titers against 31 BVDV field isolates and 7 vaccine strains. Principal component analysis (PCA) was performed using VN titers, and results were interpreted from PCA clustering within the PCA dendrogram and scatter plot. The results demonstrated clustering patterns among isolates suggestive of antigenic differences. While expected, the BVDV-1 and BVDV-2 isolates did not cluster together and had the greatest spatial distribution. Other BVDV isolates had distinct spatial patterns suggesting antigenically divergent isolates. Additionally, clusters of BVDV-1 isolates could be visualized and may represent BVDV-1 antigenically related subgroups. The multivariate analysis may be a method to better characterize antigenic relationships among BVDV strains that belong to the same BVDV species and do not have distinct antigenic differences.