The WEEV was responsible for large human and equine outbreaks in western North America throughout the early to mid-20th century, leading to thousands of equine and human deaths due to encephalitis. However, during the late 20th century, a marked reduction in WEEV transmission was recorded, with the last human case documented in 1998 and a decrease in infected mosquito pools identified in longitudinal surveillance programs 14–16.
A similar trend has occurred in South America in recent decades, with only sporadic cases occurring after the 1980s 22. This period of epidemiological silence contrasts with the present outbreak from 2023 to 2024, one of the largest and deadliest in South America's history.
Although the clinical data suggested an equine encephalitis virus, the outbreak was unequivocally characterized by sequencing a conserved coding region of the RNA-dependent RNA polymerase (nsP4) that distinguishes WEEV from other related alphaviruses 32. The WEEV identification allowed the deployment of more cost and time-effective qPCR methods, including the PAHO/WHO recommended assay 27,33. However, the Lambert et al. methodology 33 performed poorly in Uruguayan strains, likely due to the mismatches in probes and PCR primers (Fig. 4). In comparison, the alternative method described by Brault et al. 35 has greater homology in the target region of the 2023–2024 outbreak strains, resulting in lower Ct values and improved sensitivity (Fig. 4).
The 2023–2024 outbreak affected 1,419 equine and 58 human cases in Argentina. In Uruguay, it impacted 1086 equines, with approximately 35% of them dying from severe neurological symptoms. Considering that Argentina has eight times more horses than Uruguay 41, the outbreak has been particularly severe in Uruguay. However, it is important to note that differences in reporting may exist. The estimated mortality (horse deaths/population size = 388/405,644) was 0.09%, and the morbidity (affected horses/population size = 1,086/405,644) was 0.26%. During the outbreak, vaccination was recommended for the affected ranches and equine gathering events (Resolution Nº 282, MGAP). The Ministry of Public Health reported five human WEEV cases during the same period. These findings underscore the threat of WEEV to equine and human health and highlight the need for ongoing surveillance.
The multiplex-NGS methodology applied to the 2023–2024 South American outbreak yielded 15 complete genomes directly from deceased horse samples. These new WEEV sequence data significantly increased the number of publicly available genomes in GenBank and allowed us to perform the first large-scale genomic study of this arbovirus in South America. Our findings support the existence of two main lineages of WEEV. The North American lineage contains strains from various hosts (mosquitoes, birds, horses, and humans) collected from 1930 to 2005. These strains were relatively genetically stable, with only a maximum of 3.7% nucleotide sequence divergence 15. This lineage was extensively described and divided into groups A and B. Group B comprised three genogroups (B1 to B3) that are not necessarily monophyletic, bearing specific amino acid markers and circulation periods 14,51. Group A includes the original California strain (KJ554965.1), obtained from a horse in 1930, the human isolate McMillan (GQ287640.1) from Canada (1941), the Fleming strains (MN477208.1) from the USA, and strains from Russia (1962) and Cuba (1971). In the United States, group A might have become extinct in the 1940s and was displaced by Group B, which became predominant. Strains from ancestral groups A and B1 are generally more virulent than recent strains from groups B2 and B3. When all three group B sublineages were circulating, there was a concurrent increase in estimated viral population size between 1965 and the late 1980s. However, after the late 1980s, a reduction in estimated population size occurred when the group B3 viruses became predominant in North America 14,15.
South America has a more limited strain diversification than North America, possibly due to the fewer available strains from historical outbreaks. The South American lineage comprises horse strains from 2023 and 2024 (Uruguay and Brazil) and the CBA87 strain collected in 1958 from Argentina (Córdoba).
The two American lineages share a common ancestor with basal strains from Argentina (horse and mosquito) and Guyana (horse). The phylogenetic clustering suggests that WEEV in America shared a common origin and diversified into North and South lineages.
The 2023–2024 outbreak started in Argentina and was later detected in Uruguay and Brazil. Strains from Uruguay and Brazil are closely related and bear several clade-specific mutations. They are here denoted as a clade because they compose a monophyletic group, share a recent common ancestor, and are synchronous. This clade comprises the Uruguayan and Brazilian subclade I and the Uruguayan unique monophyletic subclades II and III (Fig. 3). The Argentine CBA87 strain collected in 1958 from a horse is basal for the 2023–2024 outbreak clade. This old Argentine strain might represent an ancestor of the outbreak's source that persisted in an enzootic transmission cycle.
This study identified distinct nucleotide synonymous changes and amino acids that define WEEV phylogenetic clustering. We analyzed the 2023–2024 clade to determine the presence of neurovirulence and transmission markers previously described. All the 2023–2024 clade strains contained the E2-214R residue previously associated with low neurovirulence in a murine model 52. Residue E2-214R was also identified as necessary for the efficient infectivity of the mosquito vector Culex tarsalis 52. Additionally, the 2023–2024 clade lacks the six amino acid mutations previously described to increase fitness in avian and mosquito hosts (nsP3-52I, nsP4-602S, C-89R, C-250W, E2-23T, and E1-374S) 14. The clade contains all six residues resembling the ancestral, less fit state (nsP3-52T, nsP4-602N, C-89K, C-250K, E2-23A, and E1-374T).
Our analysis revealed no marker linked to distinct WEEV phenotypic changes. All the unique amino acid markers detected in our analysis that have not been previously studied for their virulence or transmission characteristics should be considered subclade markers. Whether these markers have functional implications or are influenced by the mosquito or avian species involved remains unanswered, highlighting the need for further research.
Interestingly, the 2023–2024 outbreak spread rapidly in Uruguay, beginning in the region bordering Argentina and reaching the central and eastern areas. This spread pattern aligns with WEEV emerging in Argentina and spreading to Uruguay following a west-to-east expansion (Fig. 1). Similar behavior may have occurred in Brazil, where municipalities in the West detected the first cases, followed by a city from the East part of Rio Grande do Sul state.
Historically, equine encephalitis activity follows multiyear cycles, with epidemics occurring after periods of excessive rainfall starting during the preceding year 53,54. In November and December 2023, the weather in neighboring Argentina, Brazil, and Uruguay regions showed high temperatures and rain, leading to mosquito proliferation, particularly those of the Aedes albifasciatus species. The virus may have emerged in Argentina from genetically close strains that invaded Uruguay and Brazil. Genetic polymorphisms in the original Argentine viral population and different crossing-border events would explain the different subclades observed in Uruguay and Brazil (Fig. 3a).
Mosquitoes have a narrow flight range of less than a kilometer, and horse transportation between regions does not impact the transmission because the viremia in these animals is insufficient to infect new mosquitoes. Birds and, to a lesser extent, some mammals, such as lagomorphs and rodents, are the amplification hosts that transmit the virus to the local mosquito population 22. House sparrows are a natural enzootic amplification/reservoir host for WEEV in North America, developing a high-titer, short-lived viremia that peaks on day 1 post-infection without detectable morbidity 55. Therefore, wild birds with fast displacement across the country are the more likely amplification host associated with transmission. The outbreak coincided with migratory and indigenous birds arriving in areas with abundant water sources, including wells, lakes, ponds, streams, or rivers. Migratory birds remain the most plausible means of transporting the virus long distances from southern locations.
A limitation of this study is the absence of molecular diagnostics in asymptomatic and recovered horses. The lack of Argentine WEEV sequences for comparison and the unavailability of sequences from mosquitoes, amplification hosts, and humans hinders a more comprehensive description of the outbreak's epidemiological scenario.