Sugar feeding protects against arboviral infection by enhancing gut immunity in the mosquito vector Aedes aegypti

Abstract

As mosquito females require a blood meal to reproduce, they can act as vectors of numerous pathogens, such as arboviruses (e.g. Zika, dengue and chikungunya viruses), which constitute a substantial worldwide public health burden. In addition to blood meals, mosquito females can also take sugar meals to get carbohydrates for their energy reserves. It is now recognised that diet is a key regulator of health and disease outcome through interactions with the immune system. However, this has been mostly studied in humans and model organisms. So far, the impact of sugar feeding on mosquito immunity and in turn, how this could affect vector competence for arboviruses has not been explored. Here, we show that sugar feeding increases and maintains antiviral immunity in the digestive tract of the main arbovirus vector Aedes aegypti. Our data demonstrate that the gut microbiota does not mediate the sugar-induced immunity but partly inhibits it. Importantly, sugar intake prior to an arbovirus-infected blood meal further protects females against infection with arboviruses from different families. Sugar feeding blocks arbovirus initial infection and dissemination from the gut and lowers infection prevalence and intensity, thereby decreasing the transmission potential of female mosquitoes. Finally, we show that the antiviral role of sugar is mediated by sugar-induced immunity. Overall, our findings uncover a crucial role of sugar feeding in mosquito antiviral immunity which in turn decreases vector competence for arboviruses. Since Ae. aegypti almost exclusively feed on blood in some natural settings, our findings suggest that this lack of sugar intake could increase the spread of mosquito-borne arboviral diseases.

Fig 1. Sucrose feeding increases the expression levels of antiviral genes in the digestive tract of the female Ae. aegypti.

(A) Schematic of experimental design. Females that did not have access to sucrose for 48 h were fed or not with 10% sucrose solution. Digestive tracts from both populations were dissected just before sucrose feeding (BSF) and at 2 and 16 h post sucrose feeding time, and at 20, 24 and 48 h for the sucrose fed females. (B to F) RNA transcript levels of (B) p400, (C) ago2, (D) vir1, (E) piwi4 and (F) ppo8 (black empty dots and black squares, non-sucrose fed; blue squares, sucrose fed). Box plots display the minimum, first quartile, median, third quartile, and maximum relative expression levels. Data were analysed by Mann-Whitney test (non-sucrose vs sucrose fed at each time point, BSF, 2 and 16 h) and by Kruskal-Wallis test (all times, for non-sucrose fed and sucrose fed, black and blue bars respectively showing p value summary). N = 5 pools of 5 digestive tracts per condition. Only p values < 0.05 are shown. *, p value < 0.05; **, p value < 0.01; ***, p value <0.001.

Fig 2. Sucrose feeding-mediated increase of antiviral genes expression level is specific to sucrose.

Females were treated as in Fig 1A except that females were (A) either not fed (NSF), fed with 2% sucrose or 10% sucrose, (B) either not fed (NSF), fed with 10% sucrose (SF) or given a blood meal (BF). Digestive tracts were dissected 16 h post sucrose or blood feeding time. RNA transcript levels of p400, piwi4 and ppo8 were analysed by RT-qPCR. Box plots display the minimum, first quartile, median, third quartile, and maximum relative expression levels. Statistical significance was assessed with an analysis of variance followed by a Fisher’s multiple comparison test. N = 5 pools of 5 digestive tracts per condition. ns, p value > 0.05; *, p value < 0.05; **, p value < 0.01; ***, p value <0.001; ****, p value <0.0001.

Fig 3. Sucrose, glucose and fructose increase the expression levels of antiviral genes in the digestive tract of the female Ae. aegypti.

(A) Schematic of experimental design. Females that did not have access to sucrose for 48 h were either not sugar fed or fed with 10% sucrose, 10% glucose or 10% fructose solution. Digestive tracts were dissected 16 h post sugar feeding time. (B to D) RNA transcript levels of (B) p400, (C) piwi4 and (D) ppo8. Box plots display the minimum, first quartile, median, third quartile, and maximum relative expression levels. Data from three separate experiments were combined after verifying the lack of a detectable experiment effect. N = 15 pools of 5 digestive tracts per condition. Statistical significance of the sugar feeding effect was assessed with an analysis of variance followed by a Fisher’s multiple comparison test (Sugar vs No sugar). ***, p value < 0.001; ****, p value <0.0001.

Fig 4. Microbiota partly inhibits sugar-induced immunity in the digestive tract of the female Ae. aegypti.

(A) Schematic of experimental design. Females, previously treated or not with antibiotics, did not have access to sucrose for 48 h and were either not fed or fed with 10% sucrose. Digestive tracts were dissected 16 h post sugar feeding time. (B to H) RNA transcript levels of (B) p400, (C) ago2, (D) vir1, (E) piwi4, (F) ppo8, (G) cecD and (H) defE. Box plots display the minimum, first quartile, median, third quartile, and maximum relative expression levels. Data from two separate experiments were combined after verifying the lack of a detectable experiment effect. N = 10 pools of 5 digestive tracts per condition. Statistical significance of the treatments effect was assessed with a two-way ANOVA (statistical analysis summary of treatment effect and interaction between treatments given in S4 Fig). Pair-wise comparisons shown on the graphs were obtained with a post hoc Fisher’s multiple comparison test (No sucrose vs Sucrose, Bacteria vs No bacteria). ns, p value > 0.05; *, p value < 0.05; **, p value < 0.01; ***, p value <0.001; ****, p value <0.0001.

Fig 5. Sugar feeding prior to a blood meal increases antiviral gene expression in blood fed females.

(A) Schematic of experimental design. Females, previously treated or not with antibiotics, did not have access to sucrose for 48 h and were either not fed or fed with 10% sucrose. Females received an SFV-infected blood meal at a titre of 7.8 x 10⁷ PFU/mL 16 h post sugar feeding time and digestive tracts were dissected 6 h after the blood meal. PBM, post blood meal; PSF, post sugar feeding. (B to H) RNA transcript levels of (B) p400, (C) ago2, (D) vir1, (E) piwi4, (F) ppo8, (G) cecD and (H) defE. Box plots display the minimum, first quartile, median, third quartile, and maximum relative expression levels. N = 5 pools of 5 digestive tracts per condition. Statistical significance of the treatments effect was assessed with a two-way ANOVA (statistical analysis summary of treatment effect and interaction between treatments given in S4 Fig). Pair-wise comparisons shown on the graphs were obtained with a post hoc Fisher’s multiple comparison test (No sucrose vs Sucrose, Bacteria vs No bacteria). ns, p value > 0.05; *, p value < 0.05; **, p value < 0.01; ***, p value <0.001; ****, p value <0.0001.

Fig 6. Sugar feeding protects the female mosquito Ae. aegypti against SFV infection.

Females were treated as in Fig 5A. All females were given access to sucrose solution 30 h after the infectious blood meal. (A) Virus titration by plaque assays of individual females sacrificed straight after blood meal to assess infectious particles ingested. N = 5 per condition. (B to D) Virus titration by plaque assays of individual (B) guts, (C) bodies and (D) heads sampled four days post infection. N = 30 tissues per condition. Only infected samples (with 5 PFU or more per tissue, i.e. at least one plaque in one of the two replicate wells at the first dilution) were plotted on the graphs. Box plots display the minimum, first quartile, median, third quartile, and maximum. Data were analysed using two-way ANOVA (statistical analysis summary of treatment effect and interaction between treatments given in S4 Fig). Pair-wise comparisons shown on the graphs were obtained with a post hoc Fisher’s multiple comparison test (No sucrose vs Sucrose, Bacteria vs No bacteria). (E) SFV infection, dissemination and transmission potential prevalence (in percentage). Lower and upper limits of the 95% confidence interval were calculated using the Wilson score interval method with a correction for continuity. Statistical significance of the treatments effect on prevalence were assessed with a Chi-square test (compared to No bacteria—No sucrose group). ns, p value > 0.05; *, p value < 0.05; ***, p value <0.001; ****, p value <0.0001. Numbers of infected females and total number of females per group detailed in S5 Fig.

Fig 7. Sugar feeding protects the female mosquito Ae. aegypti against ZIKV infection.

Females were treated as in Fig 5A except that females received a ZIKV-infected blood meal at a titre of 5.3 x 10⁶ PFU/mL. All females were given access to sucrose solution 30 h after infectious blood meal. (A) ZIKV RNA levels, relative to S7 ribosomal protein transcript, in individual whole females assessed by RT-qPCR. Data were analysed as described (88) so as the RQ geomean of the group ‘No antibiotics-No sucrose’ is 1. N = 27 to 30 females per condition. Only ‘highly infected’ samples (with Ct less than 30, samples with Ct superior to 30 were considered as weakly infected or not infected) were plotted on the graphs. Box plots display the minimum, first quartile, median, third quartile, and maximum RQ values. Log2-transformed values of RQ values were analysed using two-way ANOVA (statistical analysis summary of treatment effect and interaction between treatments given in S4 Fig). Pair-wise comparisons shown on the graphs were obtained with a post hoc Fisher’s multiple comparison test (No sucrose vs Sucrose, Bacteria vs No bacteria). (B) ZIKV infection prevalence (in percentage). Lower and upper limits of the 95% confidence interval were calculated using the Wilson score interval method with a correction for continuity. Statistical significance of the treatment effect on prevalence were assessed with a Chi-square test (compared to No bacteria- No sucrose group). ns, p value > 0.05; **, p value < 0.01; ***, p value <0.001; ****, p value <0.0001. Numbers of infected females and total number of females per group detailed in S6 Fig.

Fig 8. Sugar-mediated protection against arboviral infection is mediated by sugar-enhanced immunity.

Females treated with antibiotics, were injected with dsRNA targeting either luciferase (dsLuc, control) or targeting runx4, piwi4, dcr2 and myd88 simultaneously (dsImm). Two days later, females were starved before being fed with 10% sucrose (dsLuc-SF and dsImm-SF) or not (dsLuc-NSF). Females received an SFV-infected blood meal at a titre of 7.8 x 10⁷ PFU/mL 16 h post sugar feeding time. All females were given access to sucrose solution 30 h after the infectious blood meal. (A) Virus titration by plaque assays of whole females four days post infection. N = 15–24 females per condition. Only infected samples (with 5 PFU or more per tissue, i.e. at least one plaque in one of the two replicate wells at the first dilution) were plotted on the graphs. Box plots display the minimum, first quartile, median, third quartile, and maximum. Data were analysed using Kruskal-Wallis with Dunn’s multiple comparison test. (B) SFV infection prevalence (in percentage). Lower and upper limits of the 95% confidence interval were calculated using the Wilson score interval method with a correction for continuity. Statistical significance of the treatments effect on prevalence were assessed with a Chi-square test (compared to dsLuc-NSF group). ns, p value > 0.05; **, p value < 0.01. Numbers of infected females and total number of females per group detailed in S8 Fig.