Evolution and ecology of vesicular stomatitis virus in the southwestern United States: 1995 to 1998

Abstract

Vesicular stomatitis (VS) is a disease of horses, pigs, and cattle that is clinically undistinguishable from foot and mouth disease that causes considerable economic loss during outbreaks in the United States (US). Cases of vesicular stomatitis virus (VSV) occur in the southwestern United States approximately every ten years and the complete epidemiological relationship among agent, host, environment and time remains unknown. This dissertation was a multi-phase approach to understanding specific essential unsolved components of the epidemiology of the disease. Both molecular and spatial analytical techniques are used in the study's approach.

The phylogenetic relationship of viruses circulating in the southwestern United States in 1995 with viruses in Mexico from 1984 to 1997 was investigated. A 750 base fragment of the glycoprotein gene (G gene) was amplified by RT-PCR and sequenced from eleven isolates of bovine origin from different regions of Mexico from 1984 to 1997. Several nucleotide changes were found with viruses from Mexico and viruses previously sequenced from the western United States. Viruses from the eastern US formed a different lineage and were not closely related to viruses causing outbreaks in the western US. Overall sequence analysis indicated relatively homogenous and distinct viral lineages for each documented outbreak in the western US, with viruses from Mexico in each lineage of the major US outbreaks from 1982/83 to 1995, suggesting a common source for these viruses. Viruses from Central America formed a diverse collection of lineages, with a marked correlation between position in the phylogenetic tree and their geographical origin.

VSV has been used as a laboratory model for viral evolution and in vitro studies due to its capacity to grow well in multiple cell lines. VSV is an RNA virus and despite its capability for rapid mutation, VSVs circulating within endemic areas remain genetically conserved over long periods of time. One possible selective factor could be the vector(s) involved in its natural cycle. In this study we compared the growth rate and rate of evolution of three genetically distinct VSV-NJ viruses from different ecological regions and different hosts in a natural (sand fly) versus non-natural (mosquito and mammalian) host cell lines. The genetic mutations and rate of evolution through serial passage of each virus was compared with viruses in nature. The virus originating from sand flies grew to higher titers in insect cells than did viruses of mammalian origin. Although all viruses readily established persistent infections both in mosquito and sand fly cells maintained in culture for up to 81 days, sustained virus yields were observed only in sand fly cells. Sequence analyses of the viruses after 0, 10 or up to 25 passages in each cell line showed no changes in the hypervariable region of the phosphoprotein (P) gene or the intergenic junction between the glycoprotein (G) and the polymerase (L). In contrast, the G gene demonstrated mutation rates between 1.39x10-4 and 6.95x10-5 in most cell-virus combinations. The majority of the mutations were non-synonymous, suggesting positive selection. These results showed that although insect cells can be a selective factor in VSV evolution in vitro, no significant differences between rates of evolution of VSV-NJ by passage in homologous (sand fly) versus heterologous (mosquito or mammalian) cells were detected.

To investigate possible ecological factors associated with regions of VSV activity, a study using geographical information systems (GIS) and subsequent spatial analysis were used to identify spatial-temporal clustering of VS premises from the 1995, 1997, and 1998 VS outbreaks. Ecological and climatic conditions were investigated in regions identified to have a high risk of VS in relation to possible vector associations. The SaTScan program was used for the spatial, temporal, and spatial-temporal analyses that identified clusters of higher risk for VSV in Arizona, Colorado, New Mexico, and Utah. Further analyses in Colorado and New Mexico indicated clusters in western and southwestern Colorado and central New Mexico along the Rio Grande River in 1995 and in north central Colorado in 1998. Descriptive statistics indicated that positive premises were located in close proximity to perennial streams and canals, at an elevation of 1377 meters to 3207 meters, and located within the dry domain ecological region in the southwestern United States. An increase in precipitation was observed from 1994 to 1999 from the 30-year average, but no substantial departures in temperature were identified from 1994 to 1999. Statistically significant associations with precipitation or temperature were not identified with VS positive premises for each month prior to the first or last month of the outbreak during 1995, 1997 or 1998. The results suggest that close proximity of VS positive premises to perennial water sources, the clustering of cases near major water sources and vegetative areas in dry ecological regions could be associated with nearby vector habitat.