Japanese encephalitis (JE) virus was first recognised in Australia when an outbreak was identified in the Torres Strait in 1995. JE virus activity was detected on many islands, and three human cases were serologically confirmed. Virus isolations from mosquitoes suggested that Culex annulirostris, a member of the Cx. sitiens subgroup, was the species involved in transmission. Nucleotide sequence analysis of isolates revealed high levels of homology, suggesting the JE virus activity came from a single source. Although the source could not be identified, Irian Jaya or Papua New Guinea (PNG) was considered likely, given the widespread distribution of activity and proximity to PNG. However, JE virus had not previously been reported in PNG. As a result of the outbreak, investigations were undertaken to:
1. determine the source of incursions of JE virus in the Torres Strait and northern Queensland
(QLD) and the mechanism of introduction;
2. investigate recurrences of JE virus in the Torres Strait and emergence in northern QLD;
3. ascertain the likelihood of establishment of JE virus in the Torres Strait and mainland Australia; and
4. develop an alternative to sentinel pig herds for surveillance of JE virus.
Adult mosquitoes were collected at numerous locations in Western Province of PNG between 1997 and 1999, yielding three isolates of JE virus. All isolates were from members of the Cx. sitiens subgroup collected during the late wet season in 1997 (Lake Murray) and 1998 (Balimo and Abam). Nucleotide sequence analysis revealed the isolates from PNG and the Torres Strait were almost identical, suggesting that PNG was a source of incursions of JE virus. Other flaviviruses (Murray Valley encephalitis and Sepik viruses) were also isolated from mosquitoes collected at
Balimo in 1998, providing evidence that activity of one flavivirus does not preclude activity of others.
Serological surveys of people in different locations in PNG revealed that infection with JE virus may be widespread in Western Province, that the prevalence of antibodies to JE virus was possibly increasing, and that the distribution of JE virus may also be increasing. Furthermore, three human cases of JE were serologically confirmed during 1997 and 1998. The possible increasing prevalence of antibodies to JE virus, coupled with the absence of large outbreaks of disease and low infection rate in mosquitoes suggests that JE virus is enzootic in PNG.
Backward trajectory analyses implicated windborne infected mosquitoes in the introduction of JE virus into Australia. A longitudinal study on Saibai Is., a northernmost island in the Torres Strait, was undertaken to investigate movement of potentially infected mosquitoes from PNG into
Australia at altitude. No evidence for windborne mosquitoes at altitude was found, although weather conditions were unfavourable for most of the study period. However, there was evidence of the movement of other insects, including gravid or parous Culicoides midges, from PNG over Saibai Island. There was also evidence of surface movement of mosquitoes associated with storm fronts and land breezes.
In response to recurrent activity of JE virus in the Torres Strait and its first appearance on mainland Australia in 1998, entomological investigations were undertaken between 1997 and 2000. Mosquitoes collected from Badu and Saibai islands, and western Cape York, were processed for virus isolation, yielding 43 isolates of JE virus in 1998 (from Badu Is.) and 2 isolates in 2000 (from Badu and Saibai islands). Most isolates were from members of the Cx. sitiens subgroup, however single isolates were also obtained from Ochlerotatus
vigilax from Badu Is. in 1998, and the recently introduced Culex gelidus from Badu Is. in 2000. Given that Oc. vigilax has desiccation-resistant eggs that can survive long dry periods, the potential for vertical transmission of JE virus to occur needs to be investigated. No JE virus isolates were obtained from western Cape York, and there was no evidence for establishment of JE virus on western Cape York. However, other flavivimses isolated included Murray Valley encephalitis (MVE), Kunjin (KUN), Alfuy (ALF) and Kokobera (KOK) viruses, indicating that conditions were suitable for flavivirus transmission in the area.
Nucleotide sequence analysis of isolates of JE virus from the Torres Strait revealed that JE vims activity between 1995 and 1998 was caused by isolates belonging to genotype II. However, isolates from mosquitoes and/or sentinel pigs on Badu and Saibai islands during the wet season in 2000 were most closely
related to virus strains belonging to genotype I, not previously detected in Australia. The source of strains of JE virus belonging to genotype I remains unknown, although more than one genotype of JE virus may exist in PNG. No large weather events preceded the activity in 2000 to potentially carry infected mosquitoes on the wind from PNG into the Torres Strait, suggesting that a different mechanism of introduction was involved.
Although no JE virus activity has been detected on mainland Australia since 1998, current methods of surveillance are not optimal. Sentinel pigs are expensive, dangerous, and pose an increased risk of transmission to nearby communities. A study was undertaken to investigate the potential for JE virus surveillance by detection of infected mosquitoes in the Torres Strait and Cape York. The effectiveness of the MosquitoMagnet™, a new trap powered by propane gas that also combusts to produce the bait CO2,
was investigated. When compared with three other mosquito trap devices, using a latin square design, results suggested that the MosquitoMagnet™ was in most instances equal in performance to other mosquito collection devices currently used in Australia. The MosquitoMagnet™ has the potential to collect mosquitoes continuously for up to three weeks with minimal maintenance and handling, making it a suitable mosquito collection device for remote areas.
In order for this method of surveillance to be possible, virus viability and RNA stability in dead mosquitoes held for long periods at high ambient temperature and relative humidity needed to be examined. In addition, the detection of single infected mosquitoes in large pools needed to be investigated, such that large numbers of mosquitoes may be processed for surveillance. To this end, a semi-nested PCR was developed and compared to cell culture and enzyme immunoassay (EIA) methods currently
routinely used to detect infected mosquitoes. Laboratory-reared Cx. sitiens mosquitoes were infected with JE virus and incubated for 12 days before the head and salivary glands were removed and assayed by immunofluorescence to confirm infection. Dead, infected mosquitoes (minus the head/salivary gland) were then placed in humidified chambers held at 28°C to 32°C for up to 14 days on no substrate, silica gel, thymol-soaked cotton wicks or both silica gel and thymol wicks. Mosquitoes were then assayed individually by RNA extraction of homogenate and semi-nested PCR, and inoculation of homogenate onto C6/36 cells and assay by EIA of fixed monolayers to detect infectious virus. Results showed that virus infectivity declined rapidly after mosquito death. However, JE virus RNA remained relatively stable in dead mosquitoes. Thymol-saturated cotton wicks proved the most suitable method of storing dead mosquitoes.
Single confirmed JE virus infected mosquitoes were also combined with pools of up to 200 mosquitoes and assayed by virus isolation on C6/36 cells and subsequent EIA, as well as RNA extraction and semi-nested PCR. Whilst the infectivity of pools with infected mosquitoes declined as pool size increased, JE virus RNA was detected in pool sizes up to 200 mosquitoes. Furthermore, 1/5 and 1/10 dilutions of pools of 100 mosquitoes containing single infected mosquitoes were also detected by semi-nested PCR. Thus, large pools of mosquitoes may be processed for surveillance of flaviviruses using the rapid semi-nested PCR developed during this study. By combining the MosquitoMagnet™ for continuous collection of vector mosquitoes in remote areas and the rapid, relatively inexpensive semi-nested PCR for processing large pools of mosquitoes, an alternative to sentinel pigs for JE virus surveillance is a distinct possibility.
Despite the absence of JE virus
activity on mainland Australia since 1998, confirmed and potential vector species of mosquitoes are abundant, as are potential vertebrate hosts such as feral pigs and water birds. Continued activity in the Torres Strait suggests future incursions of JE virus onto mainland Australia are inevitable, and there is a risk of establishment of JE virus in transmission cycles. Implementation of suitable surveillance methods in order to detect future incursions of JE virus and thereby implement appropriate prevention and control measures is imperative.