Our data demonstrate that nutritional stress experienced by the mosquito Ae. aegypti at both larval and adult stages negatively affects key life history traits and anti-DENV immune responses, with adult nutritional stress enhancing mosquito susceptibility to DENV infection. To the best of our knowledge, our results provide the first empirical evidence that nutritional stress simultaneously modulate life history traits, immunity and infection susceptibility, key components in the transmission of vector-borne pathogens. These findings improve our understanding of environmentally-mediated effects on vectorial capacity and emphasize the importance of accounting for nutritional variations in vector habitats for the surveillance and control of vector-borne diseases.
There was an overall negative impact of nutritional stress on life history traits. Nutritional stress experienced during the larval stage negatively impacted adult size, in accordance with previous studies29,30,31,32. Furthermore, nutritional stress experienced by larvae and adults, along with the ingestion of an infectious blood meal, significantly reduced fecundity. Similar results were found by Vantaux et al31 and Yan et al32, where mosquitoes reared at lower levels of larval and adult diets laid fewer eggs. DENV infection has previously been shown to negatively affect fecundity in Ae. aegypti 33, 34. This suggests that being challenged with DENV may result in a trade-off between the energy demands associated with mounting an immune response and fecundity. Here we find that larval and adult nutritional stress and ingestion of an infectious bloodmeal individually negatively affect fecundity, and there is also a significantly negative effect of their three-way interaction, but a positive effect of the interaction between larval and adult nutritional stress on fecundity. This indicates that egg production responded to the antagonistic outcome between nutritional stress and DENV ingestion, but to the synergistic outcome of energy reserves from both larval and adult stages. In addition, wing length showed a positive effect, indicating that larger-size individuals tend to lay more eggs. Interestingly, the presence/absence of eggs was negatively affected by adult nutritional stress, as shown in Fig. 2B, indicating that whether mosquitoes can lay eggs or not exclusively depends on adult energy reserve. Surprisingly, neither larval and adult nutritional stress nor ingestion of an infectious bloodmeal independently increased death risk of mosquitoes, but all interactions that contain the term adult nutritional stress (i.e., LA and LL, LA and IB, LA, LL and IB) did. This suggests adult energy supply played a central role in regulating mosquito mortality. Likewise, adult nutritional stress significantly shortened adult survival as shown in survival curves, emphasizing the importance of the adult daily energy supply in longevity. This is consistent with Briegel et al35 and Yan et al32 where lower concentrations of sucrose solution reduce the longevity of Ae. aegypti. Although both positive and negative effects of larval nutritional stress on Ae. aegypti survival have been previously reported8, 36, no significant effect was found here. While a negative effect of DENV infection on Ae. aegypti survival was reported previously33, 37, a number of studies demonstrated a non-significant effect38, 39, 40, and no independent effect of the ingestion of an infectious blood meal was found here. These contradictory results may be, at least in part, due to differences in methodologies (e.g., larval diet41, 42), genetic background (e.g., different An. gambiae populations30, 43) or mating history44 of experimental mosquitoes. It may also be that negative effects of dengue infection on survival become apparent only under conditions of stress. Here, the significant increase in the hazard ratio associated with either larval or adult nutritional stress in interaction with ingestion of DENV bloodmeal supports this interpretation.
Our results indicate that adult nutritional stress (LA) increased the susceptibility to DENV infection (Fig. 5A), while it overall negatively influenced viral replication in both bodies and legs of Ae. aegypti (Fig. 5C &D), compared to normal adult nutritional level (NA). However, such a negative effect on viral replication seems overturned after certain survival time points (approx. 22 dpbf and 34 dpbf for the bodies and legs; see Fig. 5C & D), as a result of a faster increase of viral titers at LA than at NA with time since the infectious blood meal. The modeling of dissemination status, an outcome of the higher viral replication, reflected this survival time-dependent but weakly positive effect of LA in its weakly positive relationship with the interaction between LA and survival (see Fig. 5B). These findings indicate that adult energy reserve may influence the susceptibility to DENV infection as well as subsequent viral replication. Although low adult nutrition impacts viral titers per mosquito, with time as the infection continues, dissemination is enhanced in those nutritionally stressed individuals compared to mosquitoes that had a normal nutrition. This could have implications on transmission potential, with nutritionally stressed individuals more prone to potentially transmitting DENV. This is, to the best of our knowledge, the first experimental evidence that adult nutritional stress derived from low sugar concentration (1%) can affect susceptibilities to DENV infection and viral replication in mosquitoes. Similarly, Almire et al19 demonstrated that 10% sucrose feeding, compared to non-sugar feeding, in Ae. aegypti seems to block initial gut infection and dissemination and lower viral prevalence and intensity of Semliki Forest virus. This may explain why individuals cultured with a low-quality diet (1% sucrose solution) were more susceptible to infection following an infectious blood meal. However, these stressed mosquitoes harbored lower mean viral titer (intensity) in bodies and legs than did those reared at normal diet (10% sucrose solution), indicating that there might be a trade-off between infection prevalence and intensity when facing nutritional limitation (e.g., at 1% sucrose solution). That is, low energy supply in stressed mosquitoes could be more susceptible to initial infection due to a lack of inhibition effect activated/supported by normal-sugar feeding, but the shortage of energy reserve in those stressed and infected individuals may also become the bottleneck for a high-level replication of DENV, although this hypothesis warrants further study. Alternatively, survival might have played a major role in viral replication, if more mosquitoes reared at 1% sugar died too soon to allow a sufficient increase of viral titer with time (see Fig. 3C), resulting in less mosquitoes reared at 1% sugar surviving longer and developing higher titers. Larval nutritional-stressed mosquitoes (LL) showed a lower viral prevalence in both bodies and legs than their counterparts from normal larval nutritional level (NL), but such difference between larval nutritional levels for viral prevalence was not observed in our statistical models. The effects of larval nutritional stress on arbovirus infection and dissemination susceptibility in mosquitoes have been well documented, with both significant effects45, 46, 47 and a lack of effects reported48. In this study, the food amount used to create larval nutritional stress is half of that for the normal nutritional level, which might not be low enough to produce significant infection-related phenotypic differences. This hypothesis warrants further study testing a greater range of nutritional levels. Alternatively, given that wing length, a measure of adult size and close correlate of larval nutrition, is negatively associated with DENV infection prevalence, the potential effect of larval nutrition on viral prevalence might have operated through the effect of body size. This indicates that smaller-sized (or larval nutritional-stressed) individuals are more susceptible to DENV infection, in accordance with previous studies15, 49, 50.
Our gene expression analysis of components and effector genes belonging to the three immune pathways indicated downregulation of 9 out of 14 genes due to larval nutritional stress. For instance, suppression of the Toll pathway was due to significant downregulation of upstream components such as Spaetzle and Toll. Similar effects were observed with the Imd pathway, with PGRP-LC, Imd and the transcription factor Rel2 significantly downregulated in mosquitoes that experienced larval nutritional stress. Corroborating these results, cecropin expression, an important antimicrobial peptide under the regulation of these two immune pathways, in addition to lysozyme, were also downregulated. Further immune suppression was observed with the significant downregulation of Domeless and Hopscotch, upstream components of the JAK-STAT pathway. Together, these results indicate a lower immune responsive potential in mosquitoes that endured nutritional stress during larval development. Although, the impact was not as stark, our study also showed that adult nutritional stress downregulated the expression of Imd and Lysozyme.
Considering DENV infection as a predictor, and evaluating mosquitoes that presented either a positive or negative DENV infection status post-feeding on an infectious blood meal, results showed that those with positive DENV infection had significant downregulation of critical antiviral immune pathways. For instance, we observed the significant suppression of genes Domeless and Hopscotch (JAK-STAT pathway) and Rel2 and Imd (Imd pathway). Immune genes in JAK-STAT and Imd pathway are known to be elicited by DENV infection when compared to uninfected mosquitoes that fed on non-infectious blood meals. In our study, however, all mosquitoes were challenged with an infectious blood meal and those remaining uninfected (i.e., those exposed but with negative DENV infection) are potentially mosquitoes with a more robust immune defense that were successful in limiting DENV infection. This would potentially lead to graphs showing downregulation in DENV-infected mosquitoes during a relative-quantification analysis.
Evaluating the interactions of three main factors (i.e., larval and adult nutrition, DENV infection), indicated that the nutritional stress experienced at both larval and adult stages, significantly induced the expression of Cactus, the negative regulator of the Toll pathway. This might suggest that the nutritional stress experienced during the larval stage, that is further compounded by nutritional stress at the adult stage, leads to Toll pathway suppression. However, further studies are needed to corroborate this finding given that no other Toll pathway component was significantly regulated. The interaction of larval nutritional stress and DENV infection was much stronger than that of adult nutritional stress and DENV infection, with an upregulation of Hopscotch (JAK-STAT pathway) and Imd (Imd pathway) in mosquitoes that experienced larval nutritional stress and that were later challenged with DENV infection. No significant effect was observed in the interaction between adult nutritional stress and DENV infection.
Interestingly, our analysis did not indicate any significant effect of DENV infection outcomes on Toll pathway activation. This most likely indicates that the effects of nutritional stress, or the role that the Toll pathway plays in mosquito development, are stronger factors and obfuscate the DENV infection-responsive induction of this immune pathway. However, additional studies are needed to functionally determine a potential interaction between nutrition and Toll pathway-based immunity with regards to DENV infection. Antimicrobial effectors such as cecropin and defensin are critical components of the antimicrobial response that are under the regulation of the Toll and Imd pathways. However, our data show no induction of gene expression in these effectors by DENV infection nor its interaction with nutritional stress. A possible explanation is that these effectors are upregulated in response to DENV exposure, but that infection outcomes are a result of broader interactions that we were not able to distinguish here.
Toll, JAK-STAT and Imd are well-known immune pathways involved in the antiviral defense repertoire of mosquitoes, including Ae. aegypti 24, 25, 26, 29, 51. Our results align with what has been found in other insects, in that nutrition plays an important role in mounting an adequate immune response52, 53, 54, 55. Furthermore, our results appear to suggest that mosquito immunity is largely shaped by larval nutrition rather than by adult nutrition and/or DENV infection alone, which highlights the “carry-over effects” of larval ecology on mosquito immunity.
Understanding how nutrition influences phenotypic traits in medically-important species is of great significance. Overall, our study unveils the nutritional regulations of life history traits, DENV infection susceptibility, and immune gene expression in Ae aegypti, which are related to the vectorial capacity of this important arbovirus vector. It is noteworthy that vector competence is often measured in laboratories, where females often have access to constant levels of larval and adult nutrition. In the field, however, larvae often experience stressful environments with poor nutrition and/or intense competition and consequently emerge as smaller adults56. Therefore, our data shed light on neglected factors that could generate variations in the outcome of laboratory-based experiments. Although the degree to which Ae. aegypti rely on plant sugar as their energy reserves remains controversial and may depends on their habitats57, 58, 59, 60, our data highlight the role of low sugar-feeding in DENV infection susceptibility, which may help understand why those rarely sugar-feeding Ae. aegypti that live in human dwelling become such an efficient vector for DENV. Furthermore, whether and how larval and adult nutrition affect the vectorial capacity and pathogen transmission rate remains unclear. Our results contribute to this body of knowledge and increase our understanding of factors that influence vectorial capacity. Ultimately, our findings may have important implications for the practice of vector-borne disease control through environmental management (e.g., of invasive plant species or larval habitats), which could mitigate the transmission of arboviruses.