Infection of anopheline mosquitoes with Wolbachia: Implications for malaria control

Africa carries a disproportionate share of the malaria burden. For example, more than 90% of the 446,000 malaria-related deaths reported in 2015 occurred in sub-Saharan Africa [1]. Anopheles gambiae, A. coluzzii, A. arabiensis, and A. funestus are some of the most important African vectors of malaria [2,3]. Other species, such as A. melas and A. merus are also efficient vectors but have a limited geographical distribution [3]. Insecticide-based strategies, mainly the distribution of insecticide-treated nets and indoor residual spraying, are efficient against a wide range of mosquitoes and are the current cornerstones of malaria control programs. However, the growing number of reports of insecticide resistance is driving the development of novel vector control strategies [4].

Wolbachia on viruses is strain and host specific. For example, native Wolbachia had no effect on DENV infection in Aedes notoscriptus [13].
Transinfection of host with Wolbachia triggers stronger antiviral protection than native Wolbachia strains [14]. The Drosophila wMel-Pop [15] and wMel [16] Wolbachia strains were adapted to infect Aedes aegypti cell lines and were used to transinfect A. aegypti mosquitoes by embryonic microinjection. For both strains, transinfected mosquitoes displayed strong vertical transmission with CI and greatly reduced DENV transmission [16]. A. aegypti transinfected with wMel were released in test sites in Australia, where they replaced natural mosquito populations and reached near fixation levels within a few months after their initial release [17]. Wolbachia prevalence has been stable for several years in these locations [17,18] and has slowly spread throughout the area [19].
A systematic survey in Thailand detected native Wolbachia infections in 23 species of mosquitoes, including species from the genera Aedes, Culex, and Mansonia [20]. None of the 19 species of Anopheles screened were found to harbor Wolbachia. The lack of Wolbachia infections in anophelines was later confirmed in screenings using European, African, and American specimens [21,22].
Thoracic microinjections of Wolbachia in A. gambiae resulted in ubiquitous infections in somatic tissues. However, germline cells were not infected and this precluded vertical transmission [23]. A stable Anopheles stephensi line infected with A. albopictus wAlbB strain was established by embryonic microinjection of A. albopictus ooplasm [24]. Wolbachia-infected A. stephensi were partially protected against Plasmodium falciparum infections, resulting in a modest decrease in oocyst numbers and a strong reduction in salivary gland sporozoites. Similar protection against Plasmodium was also observed when wMel-infected A. aegypti were challenged with P. gallinaceum [15] or when A. gambiae carrying somatic Wolbachia infections were infected with P. falciparum [23]. Other reports have suggested that some combinations of Wolbachia, host, and environmental factors could actually enhance Plasmodium infection in mosquitoes [25,26].

Identification of Wolbachia in natural African anopheline mosquito populations
Recently, traces of Wolbachia genomic DNA were identified in a microbiome survey of the reproductive organs of A. gambiae and A. coluzzii in malaria-endemic areas of Burkina Faso in West Africa ( Fig 1A) [27,28]. Wolbachia-specific PCR amplification and sequencing was used to confirm the presence of Wolbachia in these mosquito populations [27,29]. In an independent study, native Wolbachia infections were identified in the same mosquito species collected in two villages from Mali ( Fig 1A) [30]. Wolbachia infection was observed in two collections made five years apart, indicating that the symbiosis has remained stable in the population. More recently, native Wolbachia infections were identified in a broad range of African anophelines from the Democratic Republic of Congo, Guinea, Uganda, and Madagascar [31], as well as Gabon [32]. Phylogenetic analysis suggested that several independent horizontal transfers of Wolbachia infection have occurred, but whole-genome sequencing will be necessary for an indepth analysis of the evolutionary relationship between these strains. A significant reduction in Plasmodium prevalence was observed in mosquitoes carrying native Wolbachia infection in Mali. This effect was dose dependent, with a nonlinear negative correlation ( Fig 1B) [29,30]. Because confounding environmental or ecological variables could influence the analysis of field-collected mosquitoes, a colony of wAnga-Mali-infected A. coluzzii was established. Infection of colony-adapted mosquitoes with P. falciparum NF54 confirmed that Wolbachia negatively affects sporozoite infection with a similar negative correlation ( Fig 1B) [30].

Challenges for the development of Wolbachia as a tool for malaria control
The identification of native Wolbachia infections in A. gambiae that reduce malaria transmission is a remarkable finding. The fact that several species of Anopheles, including all the major malaria vectors in Africa, have been shown to harbor a variety of Wolbachia strains opens the possibility that one of these strains may confer CI and disrupt disease transmission. However, several challenges remain before Wolbachia can be proposed as a tool for malaria control. Implementation of Wolbachia-based strategies would rely on CI for Wolbachia to spread rapidly in natural populations following mosquito releases. At present, it is not clear whether native Wolbachia can induce CI in Anopheles. Induction of CI was not observed in caged experiments using wAnga-BF-infected Anopheles. However, CI has been shown to be influenced by environmental factors [33], and optimal conditions might be different from the ones typically used for rearing laboratory mosquitoes.
Wolbachia levels have also been shown to influence CI [34,35], and the lack of CI might be due to the low levels of native Wolbachia infection in Anopheles (wAnga-Mali genome copies are usually less than 0.1% of the mosquito genome). Wolbachia levels are also positively correlated with the intensity of protection against viruses. Similarly, higher Wolbachia levels may induce CI in anophelines and confer stronger protection against Plasmodium, which will be essential for Wolbachia to be developed as a tool against malaria. The mechanisms limiting Wolbachia levels are not well understood. One hypothesis is that Wolbachia might have adapted to control its replication as a strategy to hide from host immunity. Alternatively, high levels of Wolbachia could reduce mosquito fitness and be negatively selected. Another possibility is that Anopheles might not be a good host for Wolbachia. For example, the mosquito microbiota limits Wolbachia infections in A. gambiae. More specifically, the presence of bacteria from the genus Asaia in germline cells prevents Wolbachia invasion of A. gambiae ovaries [36]. Microbiome analyses of mosquitoes collected from Wolbachia-endemic areas found that the prevalence of Asaia was lower than in reports from other locations in Africa [28] or failed to find evidence of co-infections between Wolbachia and Asaia [31]. However, the level of P. falciparum infection was not significantly different between females with or without Asaia in A. gambiae s.l. females collected in Guinea [31]. It is also possible that Anopheles metabolism or immunity might limit Wolbachia levels. In Aedes, Wolbachia manipulates host lipid metabolism, and cholesterol sequestration seems to play a role in the protection against viruses [37,38]. The current understanding of Anopheles' lipid metabolism is incipient, but poor nutritional stores could explain the inability of anophelines to sustain high Wolbachia densities.
Infection of embryo-derived somatic cell lines would be an invaluable in vitro model to identify the components limiting wAnga levels. Transcriptomics and metabolomics of infected cells would also allow for a detailed analysis of the effect of Wolbachia on host homeostasis. Once potential target genes are identified, it would be possible to carry out functional screens to identify genes or nutrients limiting Wolbachia infection. However, it will still be necessary to evaluate how this information can be translated to a whole mosquito and eventually to the field. An alternative approach would be the development of directed evolution protocols that would allow progressive adaptation of Wolbachia or hosts to sustain higher bacteria levels.

Future perspective
Laboratory transinfection of Wolbachia to African vectors of malaria has been limited to somatic tissues, and Wolbachia failed to be vertically transmitted. We now know that Wolbachia strains in Africa have been able to overcome this barrier and establish infections in anophelines. The presence of this bacteria negatively correlates with Plasmodium sporozoite prevalence. This is an exciting new development in the field, but several challenges remain. It is not clear whether those strains can induce CI, which is needed for implementation of Wolbachia as a control strategy. Further studies should address this question and identify the factors limiting Wolbachia replication in anophelines. A more detailed understanding of the molecular components mediating Wolbachia establishment and host adaptation might make it possible to adapt CI-carrying Wolbachia strains to infect anophelines or to increase the levels of native Wolbachia infections and disrupt malaria transmission.