Recorded outbreaks of Japanese encephalitis (JE) in Asia date back to the late 19th century. JE is a major cause of viral encephalitis and is endemic in several regions of Asia and the Pacific, causing an estimated 50,000 infections and 10,000 deaths annually [12]. JE virus (JEV) is maintained in a natural cycle between Culex mosquitoes and water birds, and is amplified in pigs. Accidental infection occurs in dead-end hosts, including humans and horses. The most important mosquito vector in Korea is Culex tritaeniorhynchus [34]. In regions where JE is endemic, vaccination against JEV is the most effective measure to control infection in humans and domestic pigs [5].

JEV can infect humans and a several other animals. However, most animals show no overt signs of infection, except for humans and horses. In horses, most infections are inapparent, and the mortality rate of viral encephalitis is less than 5%. Pigs are affected only during pregnancy; JEV infection manifests as abortion, mummified fetuses, and stillborn or weak piglets, depending on the stage of pregnancy [6]. JEV infection is an important economic disease in the domestic swine industry.

JEV, a member of the genus Flavivirus in the family Flaviviridae, clusters in the JE serocomplex group. JEV has an 11-kb encapsidated RNA genome composed of a single large open reading frame that encodes three structural and seven non-structural viral proteins (Fig. 1) [78]. The outer membrane of the virion consists of 90 envelope (E) protein homo-dimers, which play roles in the attachment to cellular receptors and membrane fusion [8]. Although the E protein was reported to be the major target for protection, animal experiments have shown that protection from disease is also mediated by antibodies against NS1, which is secreted from infected cells [910].

From 1984 to 2009, few cases (a mean of 2.1) of JE in humans were reported annually due to the national vaccination policy for Korean infants [11]. In 2010 and 2012, however, there were human outbreaks involving 26 cases with 7 deaths and 20 cases, respectively [12]. There has been no official notification of an outbreak in the Korean pig population since 2007 (Fig. 2). Swine JE is the second notifiable disease in the Act on the Prevention of Contagious Animal Disease [13]. A JE vaccination program for the entire Korean sow population has been performed annually to prevent reproductive disorders. In addition, regular serosurveillance has been conducted in the pig population to evaluate herd immunity in sows and monitor JEV in fattening pigs.

JEV can be classified into five genotypes based on similarities in the E gene (Fig. 3A) [14]. The replacement of JEV genotype 3 (G3) by genotype 1 (G1) was first reported in 1994 in Japan [15]. Similar to many other Southeast Asian countries, G1 viruses were introduced into Korea in 1993 and subsequently completely replaced G3 viruses as the dominant circulating JEV genotype (Fig. 3B) [316]. In addition, a JE G5 virus was first reported from a Culex mosquito in 2011; it is closely related to the Muar strain found in Malaysia in 1952 [17]. The genotypes differ by 10%-20% and 2%-6% at the nucleotide and amino acid levels, respectively. It has been suggested that JEV comprises a single serotype due to its relatively limited diversity at the amino acid level [18]. A mouse experiment demonstrated that vaccine strains induced significant, but lower, cross-neutralization against heterologous viruses [819]. However, Fan et al. [5] reported recently that live attenuated G3 swine vaccine (at222 strain) was partially protective against G1 viruses with very low cross-protective antibody titers. It is widely accepted that the current JEV vaccines derived from G3 strains are also able to protect against the now predominant G1 viruses. However, it is controversial whether the vaccine strain (G3) can induce a protective immune response against the now dominant JEV G1 strain.

When the genome sequence of the vaccine strain was compared with those of other JEV strains, it was grouped with G3 viruses. At the nucleotide level, it showed high homology with other G3 viruses but differed clearly from G1 viruses [26]. At the start of the 3'-untranslated region, the genome of G3 viruses is characterized by specific nucleotide insertions right after the stop codon [16]. At the deduced amino acid level, there are several differences between G3 and G1 viruses scattered throughout the genome, including nonstructural proteins. There are also several differences from all other JEV strains, which were thought to be related to the viral attenuation or cell adaptation that occurred during the many passages in CEF cells. In the prM and E protein genes, which are important for protective immunity, there are four amino acid differences compared to the G1 virus group, especially two amino acid differences related to the induction of neutralizing antibodies in the E domain III region. These results indicate that the vaccine strain clearly has the molecular character of JEV G3, with several amino acid differences from the JEV G1 strains that now prevail in Korea (Fig. 6).

In the JE endemic region, vaccination against JEV is the most effective measure for controlling JEV infection in humans and domestic pigs. This article reviewed the history of JEV vaccine development and the molecular characterization of the vaccine strain in animals. The low immunogenicity of the current vaccine in pigs and the recent genotype shift phenomenon (G3 to G1) in Korea have increased the demand for new G1 vaccines. New inactivated and live attenuated G1 vaccines have been developed to induce more effective immunity in animals. Furthermore, novel, effective vaccines should be developed for animals using advanced technologies so that we can deal promptly with newly emerging JE genotype viruses.