ISOLATION OF ARBOVIRUSES FROM MOSQUITOES COLLECTED IN NORTHERN VIETNAM

INTRODUCTION

Every year there are approximately 1,000 cases of acute encephalitis syndrome (AES) in the 28 provinces of northern Vietnam. The majority of cases occur among children less than 15 years of age, with mortality rates of approximately 8–9%, and approximately half of survivors are left with severe neurologic sequelae.^1,2 One of the leading causes of pediatric encephalitis in Asia is the mosquito-borne pathogen Japanese encephalitis virus (JEV). Prior to the introduction of JEV vaccine in 1995, approximately 70% of all AES cases in Vietnam were JEV positive. However, data from the last four years suggest that JEV now accounts for less than 40% of AES cases nationwide.¹ Encephalitis in northern Vietnam is highly seasonal. The majority of cases occur from May through July, with a marked peak in June (Figure 1). Importantly, trends for both JEV-confirmed and JEV-negative cases follow the same seasonal pattern.

Beginning in 2001, increases in non-JEV AES cases were noted in several provinces in northern Vietnam. In May 2003, admissions to the encephalitis unit of the National Pediatric Hospital in Hanoi exceeded capacity, causing shortages of beds and significant alarm within the community. In April 2003, 11 infants between the ages of 6 and 18 months died within 48 hours of hospitalization. Etiologies could not be determined for the majority of cases. In response to these events, and to heightened awareness regarding the risks associated with emergence of novel pathogens, national health authorities in Vietnam initiated improved surveillance activities for JEV, as well as surveys for arbovirus activity in regions where AES cases had occurred. Here we report the results of attempted virus isolations from mosquitoes collected in four provinces of northern Vietnam during June of 2002 and July–August 2004.

Mosquitoes were collected at eight sites in metropolitan/ suburban communities surrounding the city of Hanoi, within the provinces of Ha Tay, Bac Giang, Ha Noi, and Ha Nam. The sites were characterized by moderately dense human population (e.g., communes of 6,000–12,000 persons) with variable numbers of domestic animals (pigs, cattle, buffalo; approximately 3,000–8,000 per commune), and mixed rice paddy and irrigated agriculture. Adult mosquitoes were obtained using Centers for Disease Control (CDC) carbon dioxide (dry ice–baited) light traps. Specimens were sorted by species using morphologic keys developed by Stojanovich and Scott³ and Reuben and others,⁴ and pooled into groups of blood fed or non-blooded specimens, approximately 50 mosquitoes/pool. Mosquitoes pools were triturated in BA-1 medium (M199 medium with Hank’s salts, 1% bovine serum albumin, 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 1 μL/mL of fungizone) by using a Mixer Mill apparatus (Qiagen, Valencia, CA), and suspensions were clarified by centrifugation. Clarified supernatant was assayed for virus by plaque formation in Vero cell monolayers in six-well plates using a double-overlay method with SeaPlaque low-melting temperature agarose (BioWhittaker, Rockland, MD).⁵ Second overlays containing neutral red were added at four days post-infection. For all samples that demonstrated plaque formation, the agarose overlay was removed, and cell monolayers were recovered by scraping and dilution in 1 mL of M199 medium with Hank’s salts, 100 units/mL of penicillin, 100 μg/mL of streptomycin, and 1 μL/mL of fungizone. The infected cell suspension was stored at −70°C. One additional passage in Vero cells was conducted to provide infected cells for immunofluorescent antibody (IFA) testing and to amplify virus for extraction of viral RNA.

Virus isolates were identified by using the indirect IFA test and by reverse transcription–polymerase chain reaction (RT-PCR), followed by nucleic acid sequencing. The IFA test was conducted with polyvalent mouse hyperimmune ascitic fluids (Arbovirus Grouping Fluids; National Institutes of Health, Bethesda, MD) with fluorescein isothiocyanate–conjugated goat-anti-mouse IgG as the secondary antibody.⁶ Extraction of RNA was done with cell culture supernatants using the QIAmp Viral RNA kit (Qiagen) and the RT-PCR was conducted using the Titan One Tube RT-PCR kit (Roche, Indianapolis, IN) with primers designed for generic amplification of flaviviruses (FU2/CFD3),⁷ alphaviruses (α 10247A/T₂₅V-Mlu⁸; α6533f [5′-TTGCAGGAGATACCAATGGA]/α6999c [5-AACATTCCGGATTTCATCAT-3′], R. Lanciotti, Division of Vector-Borne Infectious Diseases, CDC, Fort Collins, CO), and bunyaviruses (primers obtained from S. Tong, CDC, Atlanta, GA, unpublished data), as well as species-specific primers for identification of Simbu group viruses (Simbu: Simbu 3F/Simbu 2R⁹; Oya: Oya-S-21f [5′-GACGCCGAGGCCCAATGTTA-3′]/Oya-S-334r [5′-CAAGCGGGAAGAAGGTGAATGTC-3′]; Akabane: AKA-M-1389f [5′-TTATTATTGCCGTCACTCCTGCTTT-TAG-3′]/AKA-M-1841r [5′-GTTCGCCATCATCCA CTTTTGTAGAG-3′]). Amplified DNA fragments were visualized by electrophoresis on agarose gels. Positive DNA fragments were extracted using the QIAquick Gel Extraction Kit (Qiagen) and stored at −20°C. The RT-PCR fragments were sequenced by using the CEQ2000 Dye Terminator Cycle Sequencing with the Quick Start kit (Beckman Coulter, Inc., Fullerton, CA) with listed primers and analyzed using a CEQ 8000 automated sequencer (Beckman Coulter). Both strands of the DNA were sequenced. Nucleic acid sequences were compared with the GenBank database using the BLAST program.

A total of 653 and 469 mosquito pools from 2002 and 2004, respectively, containing 20,615 mosquitoes were processed for virus isolation. Tables 1 and 2 shows the species distribution of collections in each year, representing a total of 19 species from 7 genera. Blood-engorged specimens constituted 39 and 44% of mosquitoes collected in 2002 and 2004, respectively. Despite similar ecologic conditions among the four collection sites sampled in 2002, species distribution varied markedly. A larger percentage of Armigeres subalbatus (29.1%) was obtained from Ha Nam province than from the other sites, and the dominant Culex species collected from Cat Que commune in 2002 was Cx. gelidus (37%), rather than Cx. tritaeniorhynchus, which was the dominant species collected from all other sites in both years. In general, the 2004 collections exhibited more uniform species distribution. The dominant species in 2004 were Cx. tritaeniorhynchus (51–81%), followed by Cx. vishnui (9–38%). Other Culex, Anopheles, Mansonia, and Aedes species were surprisingly rare at all sites.

A total of 43 virus isolates were obtained from mosquitoes collected in 2002 and 2004. Twenty-eight of the mosquito pools (4.4%) from 2002 were found to contain viruses based on plaque formation. Eleven isolates of Sagiyama virus (SAGV) were obtained from pools of Cx. gelidus (2), Cx. tritaeniorhynchus (2), Cx. vishnui subgroup (1), An. vagus (2), An. sinensis (1), Anopheles spp. (1), Ma. annulifera (1), and Ar. subalbatus (1). Thirteen isolates of Oya virus (OYAV) were obtained from Cx. tritaeniorhychus (1), Cx. vishnui subgroup (2), Cx. gelidus (4), Culex spp. (3), An. sinensis (1), An. vagus (1), and Ma. indiana (1). Four isolates of Akabane virus (AKAV) were obtained from Cx. tritaeniorhynchus (1), Cx. vishnui subgroup (1), An. vagus (1), and Ochlerotatus spp. (1). Although the majority of isolates were obtained from pools of blood fed mosquitoes, all three viruses were also isolated from pools of non-blooded mosquitoes.

From the 2004 collections, 15 mosquito pools (3%) were found to contain virus, all of which originated from the collection site at Chan Ly commune in Ha Nam province. All 15 isolates were identified as Getah virus (GETV). The GETV isolates were obtained from pools of both blood fed and non-blooded mosquitoes, from An. vagus (3), Cx. bitaeniorhynchus (1), Cx. tritaeniorhynchus (4), Cx. vishnui (6), and Ma. indiana (1). Maximum likelihood estimates of mosquito infection rates were calculated using the software program PooledInfRate, version 2.0 (http://www.cdc.gov/ncidod/dvbid/westnile/software.htm) (Table 3).

Data from 2002 and 2004 indicated that the most abundant species (Culex spp.) were associated with relatively low infection rates, whereas higher infectivity was observed for Ochlerotatus, Anopheles, and Mansonia spp.

Although SAGV and GETV are currently classified as distinct species in the Family Togaviridae, genus Alphavirus, a recent comparative study of alphaviruses suggested that SAGV and GETV should be considered strains of the same virus species.¹⁰ The first isolation of GETV was from Cx. gelidus in Kuala Lumpur, Malaysia in 1955; SAGV was isolated the following year in Japan from Cx. tritaeniorhynchus. Many isolates of both SAGV and GETV have been obtained since that time from mosquitoes, swine, and equines. Both viruses appear to be widely distributed throughout southeast Asia, and are serologically and biologically similar. Neither virus has been associated with human illness; however, analysis of full-length virus genomes shows that SAGV shares 86% nucleic acid identity with Ross River virus, the most common vector-borne disease of humans in Australia.¹¹ Given the geographic distribution of SAGV and GETV along the Asian and Australian sides of the Pacific Rim, it has been suggested that migrating birds are involved in the transmission of these viruses, as is the case for several related alphaviruses.¹⁰

Akabane and Oya viruses are members of the Simbu sero-group of arthropod-borne viruses in the Family Bunyaviridae, genus Bunyavirus. Although AKAV was first isolated in 1968 from mosquitoes in Japan, it has more frequently been associated with Culicoides imicola (biting midges).¹² Similar to other Simbu viruses, AKAV is associated with still birth, abortions, hydroencephaly, and tetraogenicity in ruminants. Studies have shown that AKAV is endemic in livestock throughout the Middle East and Asia, and has caused sporadic outbreaks with major economic impact in Japan and Isreal.^13,14 The closely related OYAV was only recently discovered as a result of the National Swine Surveillance Programs for Nipah virus in Malaysia.⁹ Serosurveys of pig farms indicate that AKAV and OYAV may be among the most prevalent viruses in domestic animal populations in Asia: 75% seropositivity of AKA among pigs in Taiwan, and 93% seropositivity to Oya virus in a survey of Malaysian pig breeding districts.^9,12 To date, no human pathology has been documented for either of these viruses.

This report provides the first documented evidence of SAGV/GETV, AKAV, and OYAV in Vietnam. Pigs are among the primary mammalian hosts of these virus species^10,12; this survey clearly underscores the significant role of swine in arbovirus transmission within these regions. In Vietnam, as in many parts of Asia, domestic animals are housed in the immediate vicinity of human dwellings. Whether presence of these animals significantly enhances human exposure and infection to disease remains unclear at this time.

Data available to date suggest that there are large numbers of non-JEV encephalitis cases each year. Incidence of AES cases is strongly correlated with seasonal fluctuations in precipitation, and informal observations from community health officials have suggested a correlation between AES and annual harvests of orchard crops (e.g., longan and lychee nut). As pediatric encephalitis continues to be a major disease problem in Vietnam and throughout southeast Asia, the identification of novel etiologies of AES, including testing and screening for a wider variety of encephalopathic infectious agents, will be necessary to improve prevention and control efforts.