Wolbachia Blocks Currently Circulating Zika Virus Isolates in Brazilian Aedes aegypti Mosquitoes

1The mosquito Aedes aegypti, typically linked with dengue (Flaviviridae) and chikungunya (Togaviridae) transmission, is also associated with the alarming spread of Zika virus (ZIKV) (Flaviviridae), a previously obscure arbovirus that has recently gone global. Since 2007, ZIKV infection has been reported in 39 countries worldwide, including Brazil, where infection was first linked to cases of microcephaly during a large outbreak in 2015. Combined with the implication of the virus in cases of the autoimmune disorder Guillain-Barré syndrome, ZIKV has ballooned into a public health crisis.

2In the absence of a vaccine, current effective control options are limited to reducing the abundance of mosquito vector populations. However, there is a clear need for novel efficacious approaches, given that existing strategies such as insecticides and larval biological control have proven unsustainable and ineffective at halting disease spread.

3After decades of being proposed as a potential means of vector control, the endosymbiotic bacterium Wolbachia, present in an estimated 40% of all known terrestrial insect species, is currently being utilized around the world as part of an innovative approach to control the transmission of dengue (http://www.eliminatedengue.com) and other pathogens. This is possible because the reproductive parasitism associated with Wolbachia infection, typified by cytoplasmic incompatibility, gives the bacterium the ability to quickly and stably invade host populations. Critically, the bacterium also blocks the transmission of many important human pathogens in mosquitoes, including Plasmodium and chikungunya, giving it great utility as a control agent.

4As many different strains of the bacterium cause this inhibition, we hypothesized that the wMel Wolbachia strain (wMel_Br), currently being utilized as part of dengue control efforts in Brazil, might be able to restrict ZIKV infection and transmission in Ae. aegypti. To that end, we performed experimental infections with two currently circulating ZIKV isolates and used a qRT-PCR-based assay to a quantify ZIKV levels in mosquito tissues and saliva, in order to assess whether Wolbachia could potentially be used to combat the emerging Zika pandemic.

5Through experimental infection and transmission assays using two currently circulating Brazilian ZIKV isolates (BRPE243/2015 [BRPE] and SPH/2015 [SPH]), we compared ZIKV infection in wMel-infected mosquitoes (wMel_Br) with Wolbachia-uninfected mosquitoes collected in Urca, Rio de Janeiro, Brazil in early 2016 (Br). Due to the regular introduction of F1 Br males (the eggs of field-collected Br mosquitoes) in wMel_Br colony cages over 2 years, both lines had a similar genetic background (see Supplemental Experimental Procedures).

6The ZIKVs were isolated in the field in late 2015 and maintained in cell culture, and viral titers were quantified via plaque-forming assay prior to experimental infection (Table 1). In two separate experiments, fresh ZIKV-infected supernatant was harvested from culture, mixed with human blood, and used to orally infect wMel_Br and Br mosquitoes. ZIKV levels were quantified in mosquito heads/thoraces and in abdomens at 7 and 14 days post-infection (dpi) using a TaqMan-based qRT-PCR assay (Figure 1).

7The prevalence of ZIKV infection was significantly reduced among Wolbachia-infected mosquitoes (Table 1, analysis via Fisher’s exact test, p < 0.0001 unless stated). For the BRPE isolate (Figure 1A), Wolbachia decreased ZIKV prevalence by 35% in abdomens, although there was no significant difference for this tissue (p > 0.05), by 100% in head/thoraces at 7 dpi, and by 65% and 90% at 14 dpi, respectively. For the SPH isolate (Figure 1B), Wolbachia reduced prevalence by 95% and 67% in head/thoraces and abdomens (p = 0.0002), respectively, at 7 dpi, and by 74% and 68% in head/thoraces and abdomens, respectively, at 14 dpi.

8Likewise, the intensity of ZIKV infection was greatly reduced in wMel_Br mosquitoes for both tissues and time points (Mann-Whitney U tests, p < 0.0001). Additionally, we observed that median ZIKV titers in the head/thoraces of Br mosquitoes increased over time for both isolates (Mann-Whitney U test; BRPE, p < 0.0001; SPH, p = 0.0094), while there was no such effect in wMel_Br mosquitoes.

9Saliva was collected from Br and wMel_Br mosquitoes at 14 dpi, after the 5- to 10-day ZIKV extrinsic incubation period was likely completed, in order to determine if Wolbachia infection also inhibited ZIKV transmission (Figure 1C). We used mosquitoes infected with the BRPE isolate as it had a higher titer in culture (Table 1). ZIKV levels were quantified directly for individual saliva samples using the same qRT-PCR assay. We observed that Wolbachia infection reduced ZIKV prevalence in individual saliva samples by 55% (Fisher’s exact test, p < 0.0001) and median ZIKV copies by approximately 5 logs (Mann-Whitney U test, p < 0.0001).

10To determine if the virus in these samples was infectious, a further ten wMel_Br and ten Br saliva samples, from the samples described above, were intrathoracically injected into 8–14 naive Br mosquitoes each (Figure 1D), using a previously described method. The overall mortality rate among injected mosquitoes was 11.93%. The presence or absence of ZIKV infection was determined at 5 dpi in eight mosquitoes injected with each saliva, amounting to a mean proportion sampled of 0.68. Of the 80 mosquitoes injected with Br saliva, 68 (85%) became infected with ZIKV, with all Br saliva samples producing at least one infected mosquito. In contrast, none of the 80 mosquitoes injected with wMel_Br saliva became infected (Fisher’s exact test, p < 0.0001; odds ratio 882.3, 95% CI, 51.3–15187), indicating that while some of the wMel_Br saliva samples did contain detectable ZIKV, we saw no evidence that the saliva contained infectious virus.

11There is a clear correlation between the inhibition of pathogens by Wolbachia and bacterial density in insect tissues. In order to determine if there was a link between Wolbachia density and ZIKV prevalence and intensity, we measured total Wolbachia RNA levels in the wMel_Br mosquitoes used in the ZIKV infection assays, using qRT-PCR as described above. We saw that ZIKV infection explained less than 5% of the variance in Wolbachia density that was observed between ZIKV-infected and -uninfected wMel_Br mosquitoes at either 7 dpi or 14 dpi and was not a significant predictor (PERMANOVA; p > 0.05). Furthermore, we observed no relationship between Wolbachia density and ZIKV load among wMel_Br mosquitoes that became infected with the virus (Spearman correlation; heads/thoraces, r = 0.5952, p = 0.1323; abdomens, r = −0.01891, p = 0.9210). This suggests that there may not be a direct link between Wolbachia density in individual mosquitoes and ZIKV infection, indicating that the inhibition of ZIKV may arise through other means, indirectly due to the presence of the bacterium.

12Our results indicate that the ability of Wolbachia infection to greatly reduce the capacity of mosquitoes to harbor and transmit a range of medically important pathogens, including the dengue and chikungunya viruses also extends to ZIKV. While wMel did not completely inhibit ZIKV infection, we observed a similar decrease in prevalence and intensity of infection to that of wMel-infected Ae. aegypti challenged with viremic blood from dengue patients, which was considered sufficient to drastically decrease viral transmission. Additionally, the fact that we did not observe an increase in disseminated ZIKV infection over time, and that ZIKV prevalence and infectivity in wMel_Br mosquito saliva was significantly decreased, may indicate that, as for dengue, wMel extends the ZIKV extrinsic incubation period. This in turn would likely further decrease overall ZIKV transmission rates, given the small decrease in lifespan associated with wMel infection.

13We observed that the wMel Wolbachia infection in Ae. aegypti greatly inhibited ZIKV infection in mosquito abdomens, and it reduced disseminated infection in heads and thoraces and ZIKV prevalence in mosquito saliva. Most critically, our results suggest that saliva from wMel-infected mosquitoes did not contain infectious virus. That this inhibition occurred for two ZIKV isolates that circulated in Brazil during the 2015 epidemic, and for mosquitoes with a wild-type genetic background, suggests that wMel could greatly reduce ZIKV transmission in field populations of Ae. aegypti, which in turn would likely reduce the frequency of Zika-associated pathology in humans.

14Wolbachia can invade and persist in wild mosquito populations and represents a relatively low-cost, self-sustaining form of mosquito control that is already being trialed in countries where ZIKV outbreaks have been reported and has recently been recommended by the World Health Organization as a suitable tool to control ZIKV transmission (http://migre.me/tDWVe). It is important to point out that extensive public engagement will be required before releases of Wolbachia-infected mosquitoes can be scaled up for use in other areas. However, the results presented here indicate that wMel-infected Ae. aegypti represent a realistic and effective option to combat the ZIKV burden in Brazil and potentially in other countries and should be considered as an integral part of future control efforts.

15The work reported in this paper was performed under the oversight of the Committee for Ethics in Research (CEP)/FIOCRUZ (License CEP 732.621).

16Conceptualization, H.L.C.D., M.N.R., and L.A.M.; Methodology, H.L.C.D. F.B.S.D., E.P.C., and L.A.M.; Formal analysis, H.L.C.D. and E.P.C.; Investigation, H.L.C.D.; M.N.R., F.B.S.D., S.B.M., and E.P.C.; Writing—Original Draft, H.L.C.D.; Writing—Review & Editing, H.L.C.D., E.P.C., and L.A.M.; Funding Acquisition, L.A.M; Resources, L.A.M.; Supervision, L.A.M.