The 20-hydroxyecdysone agonist, halofenozide, promotes anti-Plasmodium immunity in Anopheles gambiae via the ecdysone receptor

Mosquito physiology and immunity are integral determinants of malaria vector competence. This includes the principal role of hormonal signaling in Anopheles gambiae initiated shortly after blood-feeding, which stimulates immune induction and promotes vitellogenesis through the function of 20-hydroxyecdysone (20E). Previous studies demonstrated that manipulating 20E signaling through the direct injection of 20E or the application of a 20E agonist can significantly impact Plasmodium infection outcomes, reducing oocyst numbers and the potential for malaria transmission. In support of these findings, we demonstrate that a 20E agonist, halofenozide, is able to induce anti-Plasmodium immune responses that limit Plasmodium ookinetes. We demonstrate that halofenozide requires the function of ultraspiracle (USP), a component of the canonical heterodimeric ecdysone receptor, to induce malaria parasite killing responses. Additional experiments suggest that the effects of halofenozide treatment are temporal, such that its application only limits malaria parasites when applied prior to infection. Unlike 20E, halofenozide does not influence cellular immune function or AMP production. Together, our results further demonstrate the potential of targeting 20E signaling pathways to reduce malaria parasite infection in the mosquito vector and provide new insight into the mechanisms of halofenozide-mediated immune activation that differ from 20E.

www.nature.com/scientificreports/ coli and Plasmodium berghei survival in the mosquito host 8 . 20E agonists, which were initially developed as insecticides to reduce larval populations, have shown promise for use on adult An. gambiae mosquitoes 10 . Methoxyfenozide (DBH) reduced mosquito lifespan, decreased fecundity, and significantly lowered the prevalence of P. falciparum parasite infection in the mosquito host 10 . However, the mechanisms by which 20E agonists prime anti-Plasmodium immunity remain unexplored.
Here, we examine the 20E agonist, halofenozide, to better understand the mechanisms of anti-Plasmodium immunity stimulated by the direct application to the mosquito cuticle. In agreement with previous work on another 20E agonist 10 , we demonstrate that halofenozide significantly reduces P. berghei parasite numbers when applied to mosquitoes prior to infection. Furthermore, we demonstrate that halofenozide treatment influences the success of Plasmodium ookinete invasion and requires the function of ultraspiracle (USP) as part of the heterodimeric ecdysone receptor (USP/EcR). When compared to earlier studies of 20E immune induction 8 , we find that halofenozide and 20E differentially influence mosquito cellular immunity. These findings are supported by a recent study in Drosophila, which demonstrate differences in the physiological effects between 20E, a steroid hormone, and chromafenozide, a non-steroidal agonist 11 . In summary, our study provides new detail into the mechanisms by which 20E agonists promote mosquito immunity and limit Plasmodium parasite survival.

Halofenozide application reduces P. berghei survival and infection prevalence. Previous studies
have demonstrated the ability of 20E and the 20E agonist, DBH, to induce mosquito responses that limit malaria parasite infection in An. gambiae 8,10 . While recent evidence provides insight into the effects of 20E immune induction 8 , the manner in which 20E agonists influence mosquito immunity have not been previously examined. To explore this question, we performed experiments with halofenozide, a similar 20E agonist, and examined its influence on malaria parasite infection. Halofenozide was topically applied to adult female mosquitoes without significant effect on mosquito survival (Fig. S1) and challenged with P. berghei ~ 24 h post-application. Compared to control (acetone-treated) mosquitoes, halofenozide treatment with either 0.25 µg/µL or 0.5 µg/µL significantly reduced the intensity of P. berghei oocysts 8 days post infection (Fig. 1A). Moreover, halofenozide reduced the prevalence of infection from 85% in control mosquitoes to < 50% in halofenozide-treated mosqui- Figure 1. Halofenozide priming reduces P. berghei survival and infection prevalence. (A) 100% acetone (control) or halofenozide dissolved in 100% acetone (0.25 µg/µL and 0.5 µg/µL) were applied to Anopheles gambiae adult females 24 h prior to challenge with a P. berghei-infected blood meal. Eight days post-infection, oocyst numbers were examined from dissected midguts. The red bar delineates the median number of oocysts from pooled data from three independent experiments. Data were analyzed by Kruskal-Wallis with a Dunn's post-test using GraphPad Prism 6.0. n = the number of mosquitoes examined for each condition. (B) The prevalence of infection (% infected/total) is depicted for mosquitoes under each treatment and examined by Χ 2 analysis to determine significance. www.nature.com/scientificreports/ toes (Fig. 1B). Given the low mortality rate and the strong anti-Plasmodium effect, the remainder of the halofenozide experiments were conducted using the 0.5 µg/µL concentration.
Halofenozide promotes the killing of P. berghei ookinetes. To better determine how and when halofenozide limits parasite numbers, we examined the effects of halofenozide application on distinct stages of malaria parasite infection. When oocyst numbers were examined 2 days post-P. berghei infection, halofenozidetreated mosquitoes displayed a significant reduction in early oocyst survival ( Fig. 2A), suggesting that halofenozide application may promote ookinete killing. To further validate this point and determine if halofenozide application also influenced other stages of parasite development, halofenozide applications were performed approximately 24 h post-P. berghei infection. Halofenozide treatment did not influence parasite survival when applied after an established P. berghei infection (Fig. 2B), suggesting that the effects of halofenozide treatment are only effective against Plasmodium ookinetes when applied prior to infection. To determine if halofenozide had an inhibitory effect on ookinete development or viability, we performed in vitro P. berghei ookinete cultures in the presence of 0.5 µg/µL of halofenozide. Treatment with halofenozide did not change the morphology or the number of ookinetes when compared to acetone controls (Fig. 3A), supporting that halofenozide does not influence Plasmodium sexual stage development. Moreover, both control-and halofenozide-treated ookinetes were able to glide normally in Matrigel motility assays (Movie S1 and S2), suggesting that halofenozide treatment did not influence ookinete viability in vitro. Similar experiments were also performed in vivo, where Plasmodium sexual development was evaluated in the mosquito blood bolus 18 h after P. berghei infection. Fluorescent mCherry parasites were evaluated by morphology in control-and halofenozidetreated mosquitoes, displaying no differences in sexual stage development between treatments (Fig. 3B). These data provide strong support that the reduced parasite numbers associated with halofenozide treatment are not the result of direct interactions on Plasmodium parasites.
Since the mosquito gut microbiota are also major determinants of mosquito vector competence 12,13 , we examined the potential that halofenozide could impact the microbiota. Levels of 16 s rRNA expression, which serves as a proxy to assess bacteria numbers 14,15 , were not significantly altered following halofenozide application (Fig. S2A). Moreover, halofenozide had no impact on bacterial growth in vitro (Fig. S2B), arguing that the anti-Plasmodium effects of halofenozide application do not involve alterations to the mosquito microbiota.
Additional experiments ookinete midgut invasion demonstrate that the number of live ookinetes (based on mCherry fluorescence) is significantly reduced following halofenozide treatment (Fig. 3C). This suggests that the reduced numbers of early oocysts following halofenozide treatment ( Fig. 2A) are mediated by a yet undescribed mechanism of mosquito anti-Plasmodium immunity that limits the success of ookinete invasion. This also explains why halofenozide treatment after an established infection ( Fig. 2B; post-ookinete invasion) had no effect on parasite survival. To determine if the heterodimeric ecdysone receptor comprised of the ecdysone receptor (EcR) and ultraspiracle (USP) is critical for the induction of anti-Plasmodium immunity following halofenozide treatment, we used RNAi to examine the role of the respective components in halofenozide immune activation (Fig. S4). In RNAi experiments, the injection of dsRNA targeting the heterodimeric receptor had no effect on EcR, yet significantly depleted levels of USP (Fig. S4). As a result, efforts to ascertain the function of the ecdysone receptor were evaluated with USP-silencing. In control, dsGFP-silenced mosquitoes, halofenozide application prior to P. berghei challenge resulted in a significant reduction in parasite numbers ( Fig. 4C), similar to previous results (Figs. 1 and 2). However, the topical application of halofenozide in the dsUSP-silenced background ( Fig. S4) did not influence P. berghei survival (Fig. 4C). Together, this suggests that halofenozide activation of anti-Plasmodium immunity requires the function of canonical 20E signaling through the heterodimeric ecdysone receptor (EcR/USP).

Mechanisms of halofenozide immune induction are distinct from 20E.
We previously examined the influence of 20E on cellular immunity and the effects of 20E priming that limit malaria parasite infection 8 . Similarly, we wanted to determine if halofenozide application promotes similar changes to 20E in gene expression, cellular immunity, and anti-microbial defense. To compare gene expression, we examined the expression of cecropin 1 (CEC 1) and cecropin 3 (CEC 3), which previously were significantly up-regulated in response to 20E injection in naïve mosquitoes 8 . However, halofenozide did not stimulate the expression of CEC1 or CEC3 (Fig. 5A,B), suggesting that halofenozide initiates a different transcriptional repertoire than that of 20E 8 . This led us to question whether halofenozide application primed cellular immunity similar to 20E in naïve mosquitoes 8 . Contrary to previous work with 20E, halofenozide treatment did not influence the phagocytic activity of mosquito immune cells (Fig. 5C). Together, these results suggest that the mechanisms of halofenozide immune induction are distinct from that of 20E in limiting malaria parasites.

Discussion
To date, the most effective malaria control strategies target the mosquito host by reducing mosquito habitats, preventing transmission through the use of bed nets, or by killing adult mosquitoes through the use of insecticides 19 . However, due to increasing insecticide resistance in many malaria-endemic regions of the world, it is critical to develop new methods or improve on existing tools to interrupt malaria transmission. This includes the use of commonly used insecticides to manipulate mosquito host physiology, influencing fitness or rendering them less likely to acquire and transmit mosquito-borne pathogens. Childs et al. demonstrated that the application of a 20E agonist, DBH, negatively influenced mosquito survival, reproduction, and reduced P. falciparum survival in the www.nature.com/scientificreports/ mosquito host 10 , supporting that 20E agonists have the potential to improve existing malaria control strategies 10,20 . This is further supported by recent studies demonstrating that 20E confers anti-Plasmodium immunity 8 , yet prior to our presented work herein, the manner in which 20E agonists influence Plasmodium survival have remained unknown.
In agreement with previous work with DBH 10 , we found the topical application of halofenozide, a similar 20E agonist, significantly impaired P. berghei oocyst survival and reduced infection prevalence. Additionally, we provide new details into the temporal effects of halofenozide immune activation, demonstrating that the anti-Plasmodium properties of halofenozide treatment only function to reduce Plasmodium numbers when applied , and compared to 20E as a positive control. Statistical significance was determined by Mann-Whitney analysis (**P < 0.01; ns, not significant) from three or more independent biological samples. (C) To determine the involvement of the heterodimeric EcR/USP receptor in mediating the effects of halofenozide priming, we examined USP function by RNAi in GFP (control)-or USP-silenced mosquitoes. Two days post dsRNA injection, we topically applied acetone (−, control) or halofenozide (+, 0.5 µg/µL) and challenged mosquitoes with P. berghei ~ 24 h post-halofenozide topical application. Oocyst numbers were evaluated 8 days for parasite development. The red bar delineates the median number of oocysts from pooled data from three independent experiments. Data were analyzed by Mann-Whitney analysis using GraphPad Prism 6.0. n = the number of mosquitoes examined for each condition.

Figure 5.
Halofenozide does not influence AMP production or cellular immunity. The effects of halofenozide topical application and 20E injection were compared by examining their respective influence on gene expression using the 20E-responsive anti-microbial peptides (AMPs), cecropin 1 (CEC1) (A) and cecropin 3 (CEC3) (B) Statistical significance was determined by Mann-Whitney analysis (*P < 0.05) from three or more independent biological samples. The influence of halofenozide was also examined on the phagocytic activity of mosquito immune cells, evaluating the percentage of phagocytic cells and the phagocytic index (number of beads per cell) (C) Data was analyzed by Mann-Whitney analysis using GraphPad Prism 6.0 to determine significance. ns not significant.

Scientific Reports
| (2020) 10:21084 | https://doi.org/10.1038/s41598-020-78280-8 www.nature.com/scientificreports/ prior to malaria parasite challenge. Therefore, if 20E agonists such as halofenozide are integrated into bed nets or employed by indoor residual spraying (IRS) as previously proposed 10,20 , mosquitoes must to come into contact prior to taking an infectious blood meal to influence parasite survival. However, further studies are required to determine if other 20E agonists, such as DBH, share similar mechanisms of limiting malaria parasites. Several lines of evidence suggest that halofenozide application promotes "early-phase" components of anti-Plasmodium immunity that influence the success of ookinete invasion 5,15,21 . This includes the reduction in the number of invading ookinetes, as well as early oocyst number. These data are further supported by our results demonstrating that halofenozide application ~ 24 h post-infection, and after ookinete invasion, did not influence malaria parasite survival. As a result, halofenozide (and potentially other 20E agonists) may influence complement function [22][23][24] , or a yet undescribed mechanism to limit Plasmodium ookinete survival. However, further experiments are required to define the direct mechanisms by which halofenozide promotes malaria parasite killing.
The integral role of USP in conferring the effects of halofenozide application provide strong support that halofenozide activates ecdysone signaling, which is in agreement with previous evidence that 20E agonists competitively bind to the EcR/USP heterodimer 25 . Although we were unable to knock-down EcR, EcR alone is incapable of high affinity binding to 20E to promote the activation of downstream 20E-regulated genes 26 . This is further supported by the activation of Vg and cathepsin B, two highly responsive 20E-regulated genes 18,27 , following halofenozide application at comparable levels to the injection of 20E. Therefore, our data support that halofenozide-mediated immune induction occurs through the heterodimeric ecdysone receptor.
Previous work has identified the downstream targets of 20E signaling in An. gambiae, as well as demonstrated the influence of 20E on cellular immunity and anti-pathogen defense responses that limit bacterial and parasite survival 8 . Since halofenozide functions through the canonical 20E receptor, we originally hypothesized that halofenozide and 20E would similarly influence mosquito physiology and immunity 8 . While both 20E and halofenozide stimulate Vg and cathepsin B, halofenozide application does not stimulate immune gene expression or cellular immunity similar to 20E 8 . Together, these results suggest that halofenozide and 20E work through similar, yet distinct mechanisms to influence mosquito physiology and promote anti-Plasmodium immunity. This may potentially be explained by differences between steroid hormones such as 20E and their non-steroidal agonist to enter target cells and stimulate cellular function 28 . In support of this hypothesis, previous work in Drosophila demonstrates that 20E and chromafenozide, a nonsteroidal 20E agonist, differentially enter cells to stimulate cell function 11 . 20E requires the ecdysone importer (EcI) to gain cell entry, while chromafenozide can stimulate cellular responses independent of EcI 11 . Based on these notable differences between steroid hormones and non-steroidal agonists, additional studies are required to further delineate the mechanisms of 20E and halofenozide immune induction in the mosquito host.
In summary, we demonstrate that the 20E agonist, halofenozide, is an effective tool to promote physiological responses that reduce the prevalence and intensity of malaria parasite infection, similar to previous studies with the 20E agonist DBH 10 . Moreover, we define the temporal requirements of halofenozide application, where Plasmodium infection intensity is limited only in mosquitoes treated pre-infection, consistent with a role in halofenozide mediating physiological responses that limit the success of ookinete invasion. We demonstrate that halofenozide requires the function of USP, implicating ecdysone signaling in mediating the physiological responses that limit Plasmodium infection. However, through comparative analysis of gene expression and cellular immune function, we establish that halofenozide application produces physiological responses not directly comparable to the effects of 20E 8 . From these data, we provide new fundamental insight into the mechanisms of halofenozide immune induction to better understand 20E agonist function. Together, the evidence presented here and in previous studies 10,20 suggest that 20E agonists could be an effective tool to help reduce the transmission of malaria.

Methods
Ethics statement. The protocols and procedures used in this study were approved by the Animal Care and Use Committee at Iowa State University (IACUC-18-228). All experiments were performed in accordance with the relevant guidelines and regulations of the approved study.

Mosquito rearing and Plasmodium infections.
Anopheles gambiae (G3 strain) were maintained at 27 °C and 80% relative humidity, with a 14/10-h light/dark cycle. Larvae were maintained on a ground fish food diet (Tetramin). Pupae were isolated using a pupal separator (John W. Hock Company) and placed in containers of ~ 50 mosquitoes where they were allowed to eclose in mixed populations of male and female mosquitoes. Adult mosquitoes were maintained on a 10% sucrose solution.
For mosquito infections with P. berghei, a mCherry P. berghei strain 29 was passaged into female Swiss Webster mice as previously described 8,15,30,31 . After the confirmation of an active infection by the presence of exflagellation, mice were anesthetized and placed on mosquito cages for feeding. Infected mosquitoes were sorted and then maintained at 19 °C. Mosquito midguts were dissected 8 days post-infection in 1× PBS, and oocyst numbers were measured by fluorescence using a Nikon Eclipse 50i microscope. www.nature.com/scientificreports/ using a repeating syringe dispenser (PB600-1 Hamilton syringe with a 10-µL syringe). Surviving mosquitoes were challenged (E. coli or P. berghei) 24 h post-topical application or were used to collect samples for RNA isolation to determine the effects of halofenozide treatment on gene expression.

Gene expression analysis.
Total RNA was isolated from whole mosquitoes (~ 10 mosquitoes) using TRIzol (Thermo Fisher Scientific) according to the manufacturer's protocol. RNA samples were quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific) and ~ 2 µg of total RNA was used as a template for cDNA synthesis using the RevertAid First Strand cDNA Synthesis kit (Thermo Fisher). Gene expression was measured by qRT-PCR using gene-specific primers (Table S1) and PowerSYBR Green (Invitrogen). Samples were run in triplicate and normalized to an S7 reference gene and quantified using the 2 −ΔΔCT method as previously 32 .
dsRNA synthesis and gene-silencing. T7 primers for GFP and ultraspiricle (USP) were used to amplify cDNA prepared from whole An. gambiae mosquitoes and cloned into a pJET1.2 vector using a CloneJET PCR Cloning Kit (Thermo Fisher). The resulting plasmids were used as a template for amplification using the corresponding T7 primers (Table S2). PCR products were purified using the DNA Clean and Concentrator kit (Zymo Research) and used as a template for dsRNA synthesis as previously described 8,30 . The resulting dsRNA was resuspended in RNase-free water to a concentration of 3 µg/uL. For gene-silencing experiments, 69nL of dsRNA was injected per mosquito and evaluated by qRT-PCR to establish gene knockdowns at 1-5 days post-injection. Time points with the highest efficiency of gene-silencing were chosen for downstream experiments.
Phagocytosis assays. Phagocytosis assays were performed in vivo as previously described 8 Evaluation of halofenozide impacts on bacteria. The potential impacts of halofenozide on the mosquito microbiota were evaluated using both in vivo and in vitro experiments. For in vivo experiments, we analyzed the relative quantification of 16 s rRNA expression 14,15 as a proxy for mosquito microbiota titers between acetone (control) and halofenozide (0.5 µg/µL) treatments. cDNA was prepared from whole mosquitoes receiving topical applications as described above, with 16 s rRNA expression examined by qRT-PCR using universal 16S bacterial primers listed in Table S1. For in vitro experiments, E. coli was cultured in Luria Bertani (LB) broth over night at 37 °C, then used to seed a bacterial suspension (OD 600 = 0.4) in 2 mL of LB broth. To determine the potential impacts of halofenozide on bacterial growth, 2 µL of either halofenozide (0.25 µg/µL and 0.5 µg/µL) or 100% acetone were added to bacterial suspensions and continually cultured at 37 °C with shaking at 210 rpm. Bacterial cultures were measured by optical density (OD) at 600 nM every 2 h up to 6 h post-challenge to monitor bacterial growth.
Mosquito ookinete invasion assay. The influence of halofenozide on Plasmodium sexual development and ookinete invasion was examined in vivo as previously described 33 . Mosquitoes were topically treated with acetone and halofenozide (0.5 µg/µL) as described above, then challenged with P. berghei infection. At 18 h post infection, mosquito midguts were dissected dissociating the blood bolus and midgut. For each mosquito, the proportion of each parasite developmental stage (zygote or ookinete) was determined by counting approximately 50 parasites from random chosen fields. In addition, the number of live ookinetes (as determined by mCherry fluorescence of the transgenic P. berghei strain) was determined for each of the dissociated midgut samples to assess parasite invasion.
Ookinete culture. In vitro culture of P. berghei ookinetes were performed as previously described 34 . Briefly, mice were injected i.p. with 200 μL of 10 mg/mL phenylhydrazine/1× PBS (Sigma). Three days later, mice were infected i.p. with 10 8 P. berghei-GFP iRBCs. Three days later, the P. berghei gametocyte-enriched blood was collected by cardiac puncture and cultured in 25 mL flasks with ookinete culture medium containing 0.5 µg/µL of halofenozide in 100% acetone (1:1000 dilution) or an equal volume of 100% acetone, for 24 h at 19 °C with gentle agitation. Parasites were fixed with 4% w/v paraformaldehyde and stained with an anti-Pbs21 antibody 35 . Pbs21-positive cells including gametes and zygotes (round forms), retorts and ookinetes were counted under a fluorescence microscope.
Ookinete motility assays. Plasmodium berghei-GFP ookinetes cultured in the presence of 0.5 µg/µL of halofenozide in 100% acetone (1:1000 dilution) or an equal volume of 100% acetone were mixed with an equal volume of Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix (Thermo Fisher Scientific). The mixture was applied onto a glass slide, then covered with a round coverslip and incubated at room temperature for 5 min to allow the matrix to solidify. Ookinete motility was immediately monitored for 15 min at room temperature with frames taken every 15 s on an Axio Imager M2 fluorescence microscope using a 20× objective Scientific Reports | (2020) 10:21084 | https://doi.org/10.1038/s41598-020-78280-8 www.nature.com/scientificreports/ and an Axiocam 506 mono camera (Zeiss). Videos were acquired and processed using the Zen 2.5 Software (Zeiss).