Indian wild-type strain A. stephensi were maintained as previously described (Rodriguez et al., 2021; Briggs et al., 2022). All mosquito life stages were maintained at 28°C, 80% humidity, and under 12 h light–dark cycles (0800–2000). At a minimum, 24 h prior to the start of an experiment, adult female mosquitoes were moved and housed in 2 L polypropylene containers with mesh screening.

To test the impacts of ingested histamine and 5-HT in combination on A. stephensi, 4–6 day old adult female mosquitoes were provisioned with the following treatments delivered as described below: control (water only), malaria-associated histamine (10 nM histamine), malaria-associated 5-HT (0.15 μM 5-HT), healthy-associated histamine (1 nM histamine), healthy-associated 5-HT (1.5 µM 5-HT), malaria-associated combination (0.15 µM 5-HT + 10 nM histamine) and healthy-associated combination (1.5 µM 5-HT + 1 nM histamine).

As described previously, (Rodriguez et al., 2021; Briggs et al., 2022), a blood meal consisting of washed human red blood cells (RBCs) and heat inactivated human serum (1:1, vol:vol) was provided via glass bell feeders to A. stephensi for 15 min. Time of blood feeding has been shown to alter some reproductive variables, but infection success is not impacted (O’Donnell et al., 2019), so blood meals were offered between 0800–1,100 to maintain consistency across all studies. Sucrose-soaked cotton balls (10%) were used to maintain mosquitoes between blood meals.

The treatments described in Section 2.2 were provisioned via artificial blood meal (Section 2.3) to mosquitoes, after which experiments on behavior, reproduction, lifespan and infection (described below in Sections 2.8–2.13) were conducted. Histamine and 5-HT were diluted in water then diluted as 3 μL aliquots in 3 mL of blood for the final concentrations indicated in Section 2.2.

The treatments described in Section 2.2 were provisioned to mosquitoes by delivery of treatment-soaked cotton balls prior to the first blood meal, previously described as priming (Rodriguez et al., 2021; Briggs et al., 2022). Priming was used to provision treatments prior to feeding on infected mice. A subset of behavioral studies was completed with primed, uninfected A. stephensi for direct comparison to studies with primed, infected A. stephensi. Female mosquitoes received a sugar cube and treatment-soaked cotton balls with the treatments from Section 2.2. The treatment solutions were made fresh daily, and treatment-soaked cotton balls were changed twice per day between 0830–1,100 and 1,630–1800 for 3 days prior to the first blood meal. Prior to the first blood meal, the sugar cube and treatment-soaked cotton balls were removed within 30 min to 1 h.

Each treatment (Section 2.2) was offered in a blood meal to groups of 120 female mosquitoes as described in Section 2.4. Following feeding, 40 blood-fed mosquitoes from each treatment group were placed into individual housing (50 mL conical tubes) for 48 h to oviposit as described (Rodriguez et al., 2021; Briggs et al., 2022). Following oviposition, female mosquitoes were transferred back to group housing and eggs were counted and recorded. The process was repeated weekly for 3 weeks or three gonotrophic cycles with blood fed mosquitoes. Separate cohorts of mosquitoes were used to perform five biological replicates.

Each combination treatment (Section 2.2) was offered in a blood meal to groups of 120 female mosquitoes as described in Section 2.4. Mosquitoes were offered a treated blood meal once weekly until no mosquitoes remained alive as described (Rodriguez et al., 2021; Briggs et al., 2022). After the first blood meal only, non-fed mosquitoes were discarded. A 250 mL polypropylene cup lined with moistened filtration paper was provided for oviposition after each blood meal. Each week the numbers of blood fed and non-fed mosquitoes were counted and recorded. Separate cohorts of mosquitoes were used to perform three biological replicates.

Female mosquitoes were treated as described in Section 2.4 with delivery of treatments in the first blood meal or as described in Section 2.5 with priming prior to the first blood meal (Rodriguez et al., 2021; Briggs et al., 2022). For these studies, 80–120 female mosquitoes were moved and housed in 2 L polypropylene containers with mesh screening. Following the first blood meal, non-fed mosquitoes were discarded. At 4 days or 14 days after the first blood meal, a second blood meal was offered, and blood fed and non-fed mosquitoes were recorded. Separate cohorts of mosquitoes were used for 4 days and 14 days feeding studies. Five biological replicates were performed for both priming and blood meal delivery of the treatments.

Flight behavior experiments were conducted in flight cages under ambient environmental conditions (<10 cm/s air velocities, 22.5°C and 410 ppm CO₂) which allowed rapid testing of different treatment conditions. The flight assay used 30 cm × 30 cm × 30 cm cages (Bioquip), in a manner similar to the experiments by Rodriguez et al. (2021), which were conducted in a wind tunnel. Briefly, a Photonic Fence Monitoring Device (PFMD), a 3D insect tracking system (Photonic Sentry, Bellevue, WA), was used to track mosquito trajectories. The system is comprised of two cameras, each associated with a ring of infrared LEDs, positioned on the side of the cage to record mosquito trajectories at 100 frames/s. A retroreflective material (3M Scotchlite 8910) was installed at the back of the cages to provide backlighting of the flying mosquitoes. The ambient temperature and CO₂ outside of the tunnel were 22.5°C and 410 ppm (ca. 0.041%), respectively. For each assay, approximately 25 female mosquitoes were transferred into the arena and allowed to acclimate for 15 min. In these experiments, filtered air was released into the cage for 20 min (pre-CO₂), followed by 5% CO₂ for 20 min, after which filtered air was returned to the cage for 20 min (post-CO₂). The 5% CO₂ plume or filtered air was delivered using two mass flow controllers (MC-200SCCM-D, Alicat Scientific, Tucson, AZ) controlled by a Python script that allowed synchronizing odor and filtered air delivery with the trajectory behaviors. To examine mosquito responses to visual stimuli, we randomly placed 3 cm diameter white and black paper circles (Color-aid Corp., Hudson Falls, NY, United States) 20 cm apart on the floor of the cage. The tracking system cannot maintain mosquito identities for extended periods of time, but individual trajectories were assumed to be independent for statistical analysis. Trajectories that were captured for at least 60 frames (1 s) were included in the analyses. Flight velocities for each mosquito trajectory were recorded and analyzed based on their 3D trajectory for the pre-CO₂ (filtered air), and CO₂ conditions. Attraction of mosquitoes to the visual cues was examined by summing the total number of mosquito trajectories that came within 3 cm of the cues. Experiments were performed with mosquitoes provisioned with histamine and 5-HT in combination at healthy and malaria-associated levels in an artificial blood meal (Section 2.4).

The mouse malaria parasite P. y. yoelii 17XNL (hereafter, P. yoelii) was maintained in 9–11-week-old CD-1 mice (Envigo, Indianapolis, IN) as previously described (Rodriguez et al., 2021; Briggs et al., 2022). All procedures were approved by the Institutional Animal Care and Use Committee of the University of Idaho (protocol IACUC-2023-8, renewed 2/27/2023). Mice were infected by injection of P. yoelii-infected RBCs and monitored daily for parasitemia by Giemsa staining starting 2 days after infection. Exflagellation was evaluated using wet prep slides with a drop of blood at 3 days post-infection. A total of 4–10 fields were counted for exflagellation events to identify mice with similar exflagellation rates for mosquito infection. Mice were anesthetized using ketamine (50 mg/kg) and xylazine (5 mg/kg) in sterile saline and placed on top of 2 L polypropylene cartons with 100–120 female mosquitoes. Mosquitoes were allowed access to mice for 30 min, after which non-fed mosquitoes were removed and discarded. After mosquito feeding, mice were euthanized via CO₂ asphyxiation followed by cervical dislocation.

For these studies, 100–120 female mosquitoes were housed in 2 L polypropylene containers with mesh screening. The mosquitoes were primed for 3 days prior to infection as described in Section 2.5, and then infected as described in Section 2.10. At 10 days post-feeding, midguts were dissected from 15 to 20 mosquitoes and stained with mercurochrome to count oocysts. At 14–15 days post-feeding, salivary glands were dissected and scored from 10 to 15 mosquitoes using the following scale: 1 for 1–99 sporozoites, 2 for 100–999 sporozoites, 3 for 1,000–9,999 sporozoites, and 4 for 10,000+ sporozoites. Oocyst and sporozoite infections were assessed with separate cohorts of mosquitoes with four biological replicates. Salivary gland infections were scored by the same person for all replicates.

Infected female mosquitoes from Section 2.11 were also used for feeding behavior studies. A cohort of 15–20 female mosquitoes from each treatment group was moved and housed in 0.95 L polypropylene containers with mesh screening and were provisioned a second blood meal at 4 days following infection. An additional cohort of 15–20 female mosquitoes was moved and provisioned a second blood meal at 11 days following infection. Each time point used different cohorts of mosquitoes where the numbers of blood fed and non-blood fed mosquitoes were recorded, and then discarded. Four biological replicates were performed.

Average numbers of eggs per female per gonotrophic cycle were analyzed using two-way analysis of variance (ANOVA). Main effects were biogenic amine treatment and gonotrophic cycle. Because both main effects and the interaction between them showed significant differences, means were separated using LS Means Differences Student’s t-test. The percentages of mosquitoes ovipositing were analyzed using a Chi-square test between treatments. Kaplan-Meier analysis followed by the log-rank test of Mantel-Cox and Gehan-Breslow-Wilcoxon were used to analyze the individual replicates of lifespan to compare the treatments, in both survivorship and the duration of feeding behavior. The median survival in days and the median duration of feeding in weeks were compared among treatments using the Kruskal-Wallis ANOVA for non-normal data. Means were separated using Dunn’s multiple comparisons test where appropriate. Percentages of uninfected and infected A. stephensi taking a second blood meal and infection prevalence of oocysts and sporozoites were analyzed using Chi-square to compare treatments. Oocyst and sporozoite infection intensities were analyzed using one-way ANOVA and means were separated using the Newman-Keuls multiple comparisons test. For flight assays, the mean activity index pre- and during- CO₂, mosquito attraction to visual objects during CO₂, and the mean velocity of mosquito trajectories during CO₂ were compared among treatments using the Kruskal-Wallis test. Means were separated post hoc using Dunn’s multiple comparisons test.

We constructed a mathematical model to project potential population level effects on human hosts of biogenic amines delivered by priming to uninfected mosquitoes. This scenario was selected to model the appearance of infection in exposed mosquitoes with different levels of biogenic amines and the associated potential changes to malaria outbreak parameters in human populations. Accordingly, the model is a host-vector epidemiological compartmental model comprised of discrete time equations. The general model considers a human population divided into compartments by infection status: Susceptible, Exposed, Infectious, and Recovered. The mosquito population is divided similarly: Susceptible, Exposed, Infectious. In both populations, susceptible individuals are not infected but can become infected. Exposed individuals have been infected but are not yet infectious; in this stage, the pathogen is undergoing an incubation period. Infectious individuals are infected and capable of transmitting the pathogen. Finally, humans enter a recovered class when they have cleared an infection. The mosquito population is divided by age classes: eggs, larvae and pupae, and adult classes from ages 1, 2, … N_A, where N_A is the maximum number of age classes included in the model. Rates of transitions among the compartments were determined by parameters related to disease processes. To compare the impacts of biogenic amines, we used different parameterizations of the model. We considered three different scenarios with the model: biogenic amines impacting the tendency to take a second blood meal 4 and 14 days later, fecundity, or the combination of tendency to take a second blood meal and fecundity. We opted not to explore all experimental results in this work as this is planned for a future study with this model. We projected model outputs of human malaria such as daily incidence, time and size of the peak daily incidence, total cases, endemic prevalence, and daily prevalence. A full description of the model and parameters are included in Supplementary Material.

A two-way ANOVA was performed to evaluate the effects of gonotrophic cycle and biogenic amine treatment on clutch size. Means and standard errors for all combinations are provided in Supplementary Table S1. There were significant main effects for both gonotrophic cycle [F (2,1711) = 28.82, p < 0.001] and biogenic amine treatment [F (6,1711) = 2.17, p = 0.043]. We also found a significant interaction between the two main effects [F (12,1731) = 1.81, p < 0.04]. Across all treatments, the average clutch size in the first gonotrophic cycle (39.65 ± 0.69) was significantly smaller than those in the second (45.74 ± 0.70) and third (47.04 ± 0.88) cycles. Across gonotrophic cycles, mosquitoes provisioned with low histamine produced significantly more eggs (44.66 ± 1.06) than control mosquitoes (41.63 ± 1.16) and more than mosquitoes treated with high 5-HT (42.44 ± 1.29) or high histamine (42.00 ± 1.20). The significant interaction between main effects showed that treatment with histamine and 5-HT have differing effects on patterns of egg production across gonotrophic cycles (Supplementary Table S1). Control mosquitoes and mosquitoes treated with high and low histamine peaked in clutch size in the third gonotrophic cycle whereas other treatments peaked during the second gonotrophic cycle (Supplementary Table S1). There were also significant differences in the percentages of females laying eggs in the first and second gonotrophic cycles, but not in the third gonotrophic cycle (Figures 1B–D for raw numbers, Figures 1E–G for corresponding percent data). In the first gonotrophic cycle, oviposition percentages were increased in groups treated with low histamine and low 5-HT compared to the respective high treatments (Figures 1B, E). The effect of low 5-HT was moderated by 10 nM histamine, the combination which represents malaria-associated levels of biogenic amines (Figures 1B, E). In the second gonotrophic cycle, the percentage of mosquitoes ovipositing was lowest in the group treated with high 5-HT and this was significantly different from the group treated with the malaria-associated combination of histamine and 5-HT (Figures 1C, F).

Treatment with the malaria-associated combination of histamine and 5-HT was associated with a significant increase in mosquito survival relative to control only in the first replicate (Figures 2A–C; Supplementary Table S2), but there were no treatment effects on median lifespan in any replicate (Supplementary Figure S1). Treatment had variable effects on the percentages of mosquitoes feeding weekly over lifespan in two of three replicates (Figures 3A–C; Supplementary Table S3), but there were no effects of treatment on the median week of blood feeding cessation for any replicate (Supplementary Figure S2).

In previous studies, provisioning of malaria-associated histamine in blood was associated with increased percentages of uninfected A. stephensi taking a second blood meal at both 4 days and 14 days, while there were no effects of histamine delivered by priming at either time point (Rodriguez et al., 2021). Conversely, provisioning of healthy-associated 5-HT by priming and blood was associated with increased percentages of uninfected A. stephensi taking a second blood meal at 4 days, but there were no effects of 5-HT delivered by priming or blood on tendency to take a second blood meal at 14 days (Briggs et al., 2022). Based on these observations, we sought to directly compare treatments with histamine and 5-HT with the effects of malaria-associated and healthy-associated combinations delivered via blood or priming on the tendency of uninfected A. stephensi to take second blood meals at 4 days (Figures 4A–D) and 14 days (Figures 5A–D).

At 4 days, treatment effects on the tendency to take a second blood meal were observed only for mosquitoes provisioned by priming (Figures 4C, D). Histamine treatment was associated with increased feeding tendency relative to 5-HT treatment (1 nM histamine vs. 1.5 µM 5-HT, 10 nM histamine vs. 1.5 µM 5-HT, 10 nM histamine vs. 0.15 µM 5-HT; p < 0.05). Further, the addition of malaria-associated or healthy-associated histamine to 5-HT increased the tendency to feed relative to the same concentration of 5-HT (0.15 µM 5-HT vs. 0.15 µM 5-HT + 10 nM histamine, 1.5 µM 5-HT vs. 1.5 µM 5-HT + 1 nM histamine; p < 0.05). Hence, at 4 days, the tendency to take a second blood was enhanced by histamine, but this effect was limited to provisioning by priming.

At 14 days, treatment effects on the tendency to take a second blood meal were observed for mosquitoes provisioned with histamine and 5-HT in blood (Figures 5A, B) and by priming (Figures 5C, D). The effects of provisioning by priming and blood, however, were distinct. Specifically, for provisioning in blood, in each case where there was a difference between malaria-associated and healthy-associated treatments (1 nM vs. 10 nM histamine, 10 nM histamine vs. 1.5 µM 5-HT, 10 nM histamine vs. 1.5 µM 5-HT + 1 nM histamine, 0.15 µM 5-HT vs. 1.5 µM 5-HT + 1 nM histamine, 0.15 µM 5-HT + 10 nM histamine vs. 1.5 µM 5-HT + 1 nM histamine; p < 0.05), malaria-associated treatments yielded higher percentages of fed females relative to healthy-associated treatments (Figures 5A, B). For priming, in each case where there was a difference between malaria-associated and healthy-associated treatments (1 nM vs. 10 nM histamine, 1 nM histamine vs. 0.15 µM 5-HT + 10 nM histamine, 10 nM histamine vs. 1.5 µM 5-HT + 1 nM histamine, 1.5 µM 5-HT vs. 0.15 µM 5-HT + 10 nM histamine; p < 0.05), malaria-associated treatments yielded lower percentages of fed females relative to healthy-associated treatments (Figures 5C, D). Similarly, the percentage of fed mosquitoes treated with the malaria-associated combination trended lower than the percentage of fed mosquitoes treated with the healthy-associated combination (Figure 5D).

We previously reported that provisioning of malaria-associated histamine by priming and in blood increased flight activities of uninfected A. stephensi relative to controls in response to CO₂ but did not increase investigation of visual objects or flight velocities (Rodriguez et al., 2021). In contrast, we observed that provisioning of 5-HT in blood had no effect on flight activities at either healthy- or malaria-associated levels, but flight velocities and object visitation were highest in mosquitoes provisioned with 0.15 µM 5-HT relative to 1.5 µM 5-HT or controls (Briggs et al., 2022).

To evaluate the combined effects of blood-derived histamine and 5-HT on flight, we tested A. stephensi behavior at 3 days following provisioning in 30 cm × 30 cm x 30 cm × 30 cm arenas that allowed fine-scale control of concentration and timing of olfactory stimuli. These arenas also allowed exploratory behavior, which has been demonstrated to provide neuromodulatory feedback during flight to brain regions associated with olfactory and visual responses (Van Breugel et al., 2015; Chapman et al., 2017; Vinauger et al., 2019; Wolff et al., 2023). During filtered air release, few mosquitoes flew in the arenas (across all replicates, mean = 133 ± 26 SEM) and there were no significant differences between treatments and controls that received no histamine and 5-HT (Kruskal-Wallis test, p > 0.37; Figure 6A). However, treatment with histamine and 5-HT more than doubled the numbers of flying mosquitoes (mean = 340 ± 38 SEM; Figure 6A), with the malaria-associated combination treatment trending towards higher activity compared to the healthy-associated combination treatment and controls (Figure 6A).

During CO₂ exposure, the numbers of trajectories associated with mosquitoes visiting the white and black visual cues were significantly different across treatment groups (Figure 6B; Kruskal-Wallis test, p = 0.027). Mosquitoes provisioned with the healthy-associated combination of biogenic amines showed a tendency to inspect the visual objects more than control mosquitoes although this difference was not significant (Figure 6B; Dunn’s test, p > 0.1). Mosquitoes provisioned with the malaria-associated combination treatment, however, displayed a significantly higher tendency to inspect the visual objects than control mosquitoes (Figure 6B; Dunn’s test: p = 0.007).

As with the numbers of flying mosquitoes, flight velocities were significantly different across treatments (Figure 6C; Kruskal-Wallis test, p = 0.03). Although the flight velocities of mosquitoes provisioned with the malaria-associated combination treatment were not significantly different from the control (Dunn’s test, p > 0.9), mosquitoes provisioned with the healthy-associated treatment had significantly higher flight velocities than mosquitoes provisioned with the malaria-associated treatment (Dunn’s test, p = 0.025). Increased activity and inspection of visual objects during CO₂ exposure following combination treatments, with trends toward higher activity in the malaria-associated combination treatment, are qualitatively similar to previous observations of increased activity with provisioning of malaria-associated histamine (Rodriguez et al., 2021). By contrast, we previously observed that flight velocities were not affected by malaria-associated histamine, but they were increased by malaria-associated 5-HT, which suggests that healthy levels of histamine and 5-HT may interact to increase flight velocity (Briggs et al., 2022).

We previously reported that priming with malaria-associated histamine increased the percentage of mosquitoes with at least one P. yoelii oocyst relative to both control and healthy-associated histamine treated mosquitoes, with increased mean midgut oocysts in both primed groups compared to controls (Rodriguez et al., 2021). Histamine priming had no effect on percentages of mosquitoes with salivary gland infections, but P. yoelii sporozoite density was higher in mosquitoes primed with malaria-associated histamine relative to controls (Rodriguez et al., 2021). In contrast, priming with healthy-associated 5-HT increased the percentage of mosquitoes with at least one P. yoelii oocyst relative to controls, with increased mean midgut oocysts in both groups of primed mosquitoes relative to control and a trend towards higher mean oocysts in the malaria-associated 5-HT treatment group (Briggs et al., 2022). In contrast to the effects of histamine, priming with 5-HT had no effect on salivary gland infections with P. yoelii sporozoites (Briggs et al., 2022). Effects of histamine provisioning on P. falciparum oocyst and sporozoite infection of A. stephensi were qualitatively similar to those of histamine priming on P. yoelii infection (Rodriguez et al., 2021). In contrast, the only effect of 5-HT provisioning on P. falciparum infection was a decrease in percentage of mosquitoes with at least one oocyst in the healthy-associated 5-HT treatment group relative to controls (Briggs et al., 2022). Histamine and 5-HT had no direct effects on parasite growth in vitro, so effects on infection were interpreted as indirect and due to biogenic amine-induced changes in the mosquito host (Rodriguez et al., 2021; Briggs et al., 2022). Taken together, malaria-associated histamine enhanced both P. yoelii and P. falciparum oocyst and sporozoite infection of A. stephensi, with the most epidemiologically relevant parameter of salivary gland infection unaffected by 5-HT for either parasite species. To understand the potential for interacting effects of histamine and 5-HT on sporogony, we primed mosquitoes with these biogenic amines alone and in combination prior to P. yoelii infection.

There were no treatment effects on the percentages of mosquitoes with P. yoelii oocysts (Figures 7A, B), but mosquitoes primed with the malaria-associated combination of histamine and 5-HT had higher mean numbers of oocysts than mosquitoes primed with healthy-associated histamine and 5-HT and the healthy-associated combination of these biogenic amines (Figure 7C). Treatment effects on salivary gland infection with P. yoelii sporozoites were more striking. Specifically, the malaria-associated histamine and 5-HT combination treatment was associated with higher percentages of infected mosquitoes relative to healthy-associated histamine and healthy-associated 5-HT, respectively (Figures 8A, B). Further, this pattern appeared to be driven by malaria-associated histamine (Figures 8A, B). In contrast to differences in percentages of infected mosquitoes, mean sporozoite infection scores in infected mosquitoes did not vary by treatment (Figure 8C).

The tendency of uninfected A. stephensi provisioned by priming to take a second blood meal was enhanced by histamine, with histamine treatment associated with increased feeding tendency relative to treatment with 5-HT (Figures 4C, D). While the tendency of P. yoelii-infected A. stephensi to take a second blood meal at 4 days later was modestly altered by treatment (Figures 9A, B), these effects contrasted with those in uninfected A. stephensi. Specifically, the tendency of infected A. stephensi to re-feed at 4 days was enhanced by malaria-associated 5-HT relative to malaria-associated histamine (10 nM histamine vs. 0.15 µM 5-HT, p < 0.05) or by the addition of malaria-associated 5-HT to histamine (10 nM histamine vs. 0.15 µM 5-HT + 10 nM histamine, p < 0.05; Figures 9A, B).

At 14 days, the tendency of uninfected A. stephensi to take a second blood meal trended lower for mosquitoes primed with malaria-associated treatments relative to females primed with healthy-associated treatments (Figures 5C, D). Similarly, the tendency of infected A. stephensi to take a second blood meal at 11 days was lower for mosquitoes primed with malaria-associated treatments relative to females primed with healthy-associated treatments (Figures 9C, D). With respect to individual treatments, the tendency of infected A. stephensi to take a second blood meal at 11 days was higher following histamine treatment alone at either dose relative to malaria-associated 5-HT or the malaria-associated combination (10 nM histamine vs. 0.15 µM 5-HT + 10 nM histamine, 10 nM histamine vs. 0.15 µM 5-HT, 1 nM histamine vs. 0.15 µM 5-HT, 1 nM histamine vs. 0.15 µM 5-HT + 10 nM histamine, p < 0.05; Figures 9C, D). At the same time, treatments with healthy 5-HT alone was associated with an increased percentage of fed, infected mosquitoes relative to those treated with malaria-associated 5-HT (1.5 µM 5-HT vs. 0.15 µM 5-HTp < 0.05; Figures 9C, D). In sum, priming with low 5-HT, alone or in combination with histamine, was associated with lower percentages of fed, infected mosquitoes at 11 days relative to other treatment groups. Taken together, feeding patterns of primed uninfected and primed P. yoelii-infected A. stephensi, which also exhibited large differences in salivary gland infection by treatment (Figures 8A, B), suggested an interaction between treatment and infection on the tendency to re-feed at 11 days post-infection.

Results in Figure 10; Supplementary Figures S7, S8 show modeling projections of the effects of feeding behavior, fecundity or these two parameters on malaria outbreak parameters. Timing and duration of outbreaks and peak numbers of human cases occurring during outbreaks were impacted by biogenic amines in A. stephensi; each of these short-term outcomes differed across mosquito groups in the present study. However, endemic prevalence (as well as total number of human cases) did not differ substantially across treatments (Figures 10G–I). All three predicted measures of outbreak dynamics were similar or nearly identical for both the low histamine and the healthy combination mosquito groups (also with low histamine), suggesting that the presence of mosquitoes with healthy levels of histamine does not lead to substantial changes in outbreak parameters.

Interestingly, the impacts on outbreak dynamics in human populations differed when considering mosquito fecundity versus the tendency to take a second blood meal. When differences in the tendency to take a second blood meal alone were considered, the projected peaks in daily incidence of human infections for the healthy combination and low histamine mosquito groups were earlier and higher than for all other mosquito groups (Figures 10A, D). In the presence of mosquitoes with malaria-associated levels of both biogenic amines or low 5-HT alone, human cases peaked later and lower than cases associated with other groups except high 5-HT alone. This suggests that for blood feeding behavior, the presence of malaria-associated levels of histamine in A. stephensi at the initiation of an outbreak does not lead to substantial differences in human infection dynamics from that associated with A. stephensi with malaria-associated levels of 5-HT.

When differences in fecundity associated with biogenic amine levels were considered alone, the differences in outbreak timing and peak size across mosquito groups were greater than was observed when considering the tendency to take a second blood meal alone. The differences in fecundity observed in the high 5-HT, high histamine, and malaria-associated combination groups seemed to lead to higher and faster peaks in human malaria incidence relative to the other treatments (Figures 10B, E), and these three groups were also associated with the greatest outbreak size (Figure 10H). Outbreak parameters associated with mosquitoes with the healthy combination or low histamine alone were similar to the control group, whereas outbreak parameters associated with mosquitoes with low 5-HT alone were lower and peaked later than with all other mosquito groups.

When considering the combined impacts of both fecundity and tendency to take a second blood meal, the time to human outbreak peak was highest and earliest with the high histamine mosquito group, suggesting that the impact on fecundity in the high histamine group drives outbreak dynamics with the malaria-associated combination mosquito group (Figures 10D, F). Time to outbreak peaks associated with the low histamine and the healthy combination mosquito groups followed in size and timing, but were earlier and higher than were observed for all other groups in this scenario. Interestingly, human outbreak dynamics associated with the high 5-HT alone mosquito group were almost identical to control, suggesting that the influence of 5-HT on feeding behavior substantially tempers the impacts on fecundity that were observed (Figures 10D–F). Across all mosquito groups, the outbreak peak associated with the low 5-HT group was the latest and lowest, emphasizing that the impacts on fecundity of high histamine greatly influence malaria outbreak dynamics.

Although differences were observed across mosquito groups for all three scenarios for human outbreak dynamics, the long-term prevalence of human infection was not substantially affected by biogenic amine-associated differences in mosquito fecundity or tendency to take a second blood meal (Supplementary Figures 7G–I).