The insect-specific Palm Creek virus modulates West Nile virus infection in and transmission by Australian mosquitoes

In this study we assessed modes of transmission of PCV, the prototype insect-specific flavivirus in Australia, and its ability to modulate replication and transmission of WNV. Although Cq. xanthogaster has yielded all isolates of PCV to date [4, 7], this species is difficult to colonize in the lab and could not be collected in sufficient numbers from the field to undertake laboratory experiments. Therefore, Cx. annulirostris, Ae. aegypti and Ae. vigilax, which are common Australian species and known vectors of a range of arboviruses, were used for laboratory infection studies. We predominantly focused on Cx. annulirostris, because this species is the key vector of the encephalitic flaviviruses WNV and MVEV, and Hobson-Peters et al. [4] conducted their in vitro experiments with these viruses. Furthermore, PCV RNA was detected in Cx. annulirostris collected during the original study when the virus was first identified, although the virus was not subsequently isolated from these mosquitoes (R. A. Hall and J. Hobson-Peters, unpublished data).

Our experiments demonstrated that PCV could not infect Cx. annulirostris by the oral route using a biologically relevant dose. However, the virus could replicate efficiently in this species, as well as in Ae. aegypti and Ae. vigilax, if injected intrathoracically. Post-infection, the virus could not be detected in the saliva or progeny collected from any of the PCV-infected mosquitoes. This suggests that the virus is not transmitted horizontally or vertically by these species. Even though the sample sizes of mosquitoes are comparable to previous studies [3, 19], the lack of evidence of horizontal or vertical transmission in the current study may highlight the narrow host range of PCV, in that infection is restricted primarily to Cq. xanthogaster. Alternatively, it could indicate that the mode of infection of the female parent mosquito is important. For instance, Cx. pipiens inoculated intrathoracically with CxFV could not transmit the virus vertically, whereas 100 % of naturally infected females transmitted the virus to their progeny [19].

The study by Hobson-Peters et al. [4] showed that mosquito cells previously infected with PCV were less permissive than uninfected cells to replication of WNV or MVEV, indicating an ISF-induced mechanism of interference or super infection exclusion. Indeed, these findings and results of other studies [3, 6] suggested that ISFs may suppress the replication of heterologous flaviviruses in mosquitoes and potentially regulate their transmission. In the present study we further explored this hypothesis in vivo, using sequential infection of laboratory mosquitoes. As predicted by the in vitro data, the mosquitoes previously infected with PCV were less susceptible to oral infection with WNV and less competent to transmit the virus when compared to uninfected mosquitoes. However, this phenomenon was dependent on the mode of the secondary infection, with WNV infection and transmission rates after intrathoracic injection relatively unaffected by prior PCV infection of the mosquito.

The vector competence of mosquitoes for arboviruses is influenced by a number of factors, including genetics of the mosquito population, virus strain, environmental conditions or the presence of endosymbionts which can impact infection [20, 21]. Despite being a pest species in some areas [9], Cq. xanthogaster is not considered to be an important arbovirus vector in Australia and has not yielded any isolates of WNV or MVEV. This is despite a large number of pools of this species being processed during periods when these viruses have been isolated from recognized vectors [22, 23]. Thus, the minor role that Cq. xanthogaster serves as an arbovirus vector may be explained by inhibition of pathogenic viruses caused by systemic infection with PCV, which can be prevalent in some populations of this species [4]. However, similar laboratory studies to the experiments described herein need to be conducted with Cq. xanthogaster to confirm this hypothesis.

The above findings are also consistent with our IHC data that showed specific localization of PCV to the epithelial cells lining the midgut of Cx. annulirostris after intrathoracic inoculation. Indeed, these cells are the site of entry of WNV infection following ingestion of an infectious blood meal [24]. This suggests that prior infection of the midgut epithelial cells with PCV may inhibit their infection with WNV. The precise mechanism for this phenomenon could involve competition between PCV and WNV for cellular resources required for efficient replication. Indeed, our in vitro growth kinetics data suggest that in some mosquito cells, PCV replicates more efficiently than WNV during the early stages of infection. This may provide an advantage to PCV over WNV for replication in co-infected cells of the midgut. Alternatively, the upregulation of antiviral responses in the mosquito by PCV infection, including RNAi pathways (siRNA and piRNA) and Janus kinase (Jak)-signal transducer and activator of transcription (STAT) (Jak-STAT) activation via Vago, may also affect WNV replication upon subsequent infection [25–29].