Abstract
Background
Measuring malaria transmission intensity using the traditional entomological inoculation rate is difficult. Antibody responses to mosquito salivary proteins like SG6 have been used as biomarkers of exposure to Anopheles mosquito bites. Here, we investigate 4 mosquito salivary proteins as potential biomarkers of human exposure to mosquitoes infected with Plasmodium falciparum: mosGILT, SAMSP1, AgSAP, and AgTRIO.
Methods
We tested population-level human immune responses in longitudinal and cross-sectional plasma from individuals with known P falciparum infection from low- and moderate-transmission areas in Senegal using a multiplexed magnetic bead–based assay.
Results
AgSAP and AgTRIO were the best indicators of recent exposure to infected mosquitoes. Antibody responses to AgSAP, in a moderate-endemicity area, and to AgTRIO in both low- and moderate-endemicity areas, were significantly higher than nonendemic controls. No antibody responses significantly differed between low- and moderate-transmission areas, or between equivalent groups during and outside the malaria transmission seasons. AgSAP and AgTRIO reactivity peaked 2–4 weeks after clinical P falciparum infection and declined 3 months after infection.
Conclusions
Reactivity to AgSAP and AgTRIO reflects exposure to infectious mosquitoes or recent bites rather than general mosquito exposure, highlighting their promise for incorporation into multiplexed assays for serosurveillance of population-level changes in P falciparum–infected mosquito exposure.
Keywords: serosurveillance, Plasmodium falciparum, Anopheles, salivary antigens, transmission
We investigate mosquito salivary proteins as potential biomarkers of human exposure to Plasmodium falciparum–infected mosquitoes: mosGILT, SAMSP1, AgSAP, and AgTRIO. AgSAP and AgTRIO are promising candidates to incorporate into multiplexed assays for serosurveillance of changes in P falciparum–infected mosquito exposure.
Despite control efforts, the global malaria burden remains high. Plasmodium falciparum, transmitted by Anopheles mosquitoes, causes the most severe disease manifestations [1]. Accurate transmission measures can estimate burden and observe changes in transmission risk. Common metrics of transmission include the human biting rate (HBR; the rate at which humans are bitten by mosquitoes) and sporozoite rate (the proportion of mosquito salivary glands with sporozoites), whose product is the entomological inoculation rate (EIR), the number of infectious mosquito bites a person receives over time [2]. Though EIR is considered the gold-standard malaria transmission metric, EIR is costly and difficult to routinely and systematically estimate (especially in low-transmission areas), may be imprecise, and differs based on mosquito catch methods [2].
Human antibodies to Plasmodium antigens have been used as biomarkers for malaria burden and changes in transmission, especially in low-transmission settings [2–7]. Antibodies to mosquito salivary gland extracts (SGEs) have also been investigated as biomarkers of exposure to mosquito bites. Immunoglobulin G (IgG) responses to SGEs have correlated with mosquito density [8] and host malaria infection status [9, 10]. Beyond whole SGEs, specific mosquito salivary antigens (MSAs) have been used as biomarkers for Anopheles HBR and malaria transmission [11]. Individual proteins investigated include SG6 [7, 12–18] and the D7 protein family [19]. Whole SGEs and specific MSAs have been investigated in Anopheles (An. gambiae [7, 9, 10, 12–18], An. funestus [16, 17], An. albimanus [9, 20, 21], and An. darlingi [9] mosquitoes), Aedes [8, 15], and Culex [8] mosquitoes. Antibodies to SG6 are associated with human exposure to malaria vector bites [18], including the heterogeneity of mosquito biting in different geographical settings and between rainy and dry seasons [7, 12], and can show changes in response to control interventions [13].
This study investigates 4 additional MSAs with potential to serve as biomarkers specifically for infected mosquito exposure based on their reported characteristics in pathogenesis studies in human and/or mice Plasmodium transmission studies: An. gambiae sporozoite-associated protein (AgSAP, AGAP004803), An. gambiae TRIO (AgTRIO, AGAP001374) [22, 23], mosquito gamma interferon–inducible lysosomal thiol reductase (mosGILT, AGAP004551), and sporozoite-associated mosquito saliva protein 1 (SAMSP1, AGAP013726) [24, 25]. Three of these proteins, AgSAP, mosGILT, and SAMSP1, were found to be directly associated with Plasmodium sporozoites during transmission and are expressed in salivary glands [24, 26, 27]. AgSAP modulates the human immune response and facilitates the transmission of Plasmodium [26]. SAMSP1 facilitates sporozoite movement and decreases neutrophil chemotaxis, which facilitates Plasmodium sporozoite transmission [27]. Unlike AgSAP or SAMSP1, incubation with mosGILT reduced sporozoite cell traversal and lowered Plasmodium transmission [24]. In humans, people in a malaria-endemic area in Senegal had higher IgG reactivity to AgSAP and SAMSP1 than people in a nonendemic area [26, 28]. AgTRIO was identified from SGE antiserum when probing with a cDNA yeast surface display library encoding An. gambiae proteins [25]. AgTRIO is highly expressed in saliva, and Plasmodium berghei–infected female An. gambiae had higher expression of AgTRIO than uninfected mosquitoes [25]. Interestingly, people from a malaria-endemic area in Senegal did not have significantly higher IgG AgTRIO reactivity IgG than people in a nonendemic area [25].
Because AgSAP and AgTRIO show increased expression in Plasmodium-infected mosquitoes, we hypothesize that serological responses to these proteins may be higher after exposure to infected versus uninfected mosquitoes [25, 26].
Here, we leverage a multiplex assay to investigate the human humoral IgG response to sporozoite-associated MSAs as quantitative biomarkers for Plasmodium-infected mosquito exposure. Multiplex approaches are highly sensitive, allow for detection of multiple serological markers using small sample volumes [29], and are useful for high-throughput serological analyses for disease surveillance [4, 30, 31]. We investigate serological responses in longitudinal and cross-sectional cohorts from areas of low (Thiès, Senegal, with an EIR <5 infectious bites per person per year [32–34]) and moderate endemicity (Kédougou, Senegal, with an EIR ∼250 [35]), and among people with and without documented recent malaria infection. We specifically sought to determine whether people with active or recent infection have higher serological responses, which could indicate that these antigens could serve as biomarkers of exposure to Plasmodium-infected mosquitoes.
MATERIALS AND METHODS
Cohorts
Details from all cohorts used in the study are contained in the Supplementary Methods.
Recombinant Protein Expression and Peptide Synthesis
Expression of recombinant mosGILT and SAMSP1 has been described previously [24, 27]. The AgTRIO and AgSAP sequences were designed for optimal expression using baculovirus and then subcloned into pFastBac1 [28]. The SG6-P1 peptide was synthesized by Genscript. The protein and peptide expression and purification are presented in detail in the Supplementary Methods, and sizes of proteins verified by sodium dodecyl sulphate–polyacrylamide gel electrophoresis are shown in Supplementary Figure 3.
Coupling Antigens to Beads
MagPlex COOH-microspheres (Luminex, Austin, Texas) were conjugated to each purified antigen using standard coupling conditions outlined in the xMAP cookbook (see details in the Supplementary Methods).
Multiplex Magnetic Bead Assay
We tested IgG responses to SG6, AgSAP, AgTRIO, mosGILT, and SAMSP1 in a magnetic bead-based assay on the Bio-Rad Multiplex using the Centers for Disease Control and Prevention multiplex assay procedure [36], presented in detail in the Supplementary Methods.
Enzyme-Linked Immunosorbent Assay
On a subset of samples, a standard enzyme-linked immunosorbent assay (ELISA) with AgSAP and AgTRIO IgG was conducted as described previously [25, 26] for comparison with the multiplex bead assay and is presented in detail in the Supplementary Methods.
Proteins Homologous Study
Homology of saliva proteins in different mosquito species were shown using BLASTp (details in the Supplementary Methods).
Statistical Analysis
Details of statistical analysis are described in the Supplementary Methods and in corresponding figure legends. Analysis was conducted and figures were made in R 4.2.2 and GraphPad Prism 9.3.0 software.
RESULTS
Demographics
Reactivity in 164 Senegalese individuals were tested for AgSAP, AgTRIO, mosGILT, and SAMSP1: 33 from low-transmission Thiès (EIR <5 [32]), all malaria-infected at enrollment; and 131 from moderate-transmission Kédougou (EIR ∼250 [35]), 72 infected and 59 uninfected at enrollment (Table 1). The Thiès cohort was followed for 2 years and includes an average of 6.8 samples per individual. All samples from Kédougou were collected during the malaria transmission season. Uninfected samples came from a cross-sectional study. For comparison, 80 healthy healthcare workers (HCWs) from a nonendemic area were tested. Experiments were repeated to allow SG6 to be compared to the 4 proteins (using samples tested described in Supplementary Table 1), with SG6 results presented in Figure 1 and data from the 4 proteins presented in Supplementary Figures 1 and 2.
Table 1.
Demographics of Cohorts for Samples Tested for AgSAP, AgTRIO, mosGILT, and SAMSP1
| Cohort | HCW (Cross-sectional) | Thiès Infected (Longitudinal) | Kédougou Infected (Cross-sectional and Longitudinal)a | Kédougou Transmission (Cross-sectional) |
|---|---|---|---|---|
| Malaria transmission intensity of region | ||||
| EIR (infectious bites per person per year) | No transmission | <5 (32) | ∼250 (33) | ∼250 (33) |
| Demographics | ||||
| No. of subjects | 80 | 33 | 72a | 59 |
| Age, y, mean (range) | 34.0 (22–74) | 10.6 (5–15) | 20.4 (2–74) | 21.3 (1–74) |
| Sex | ||||
| Female | 65 (81.3) | 0 (0) | 32 (44.4) | 30 (50.8) |
| Male | 15 (18.8) | 33 (100) | 38 (52.8) | 29 (49.2) |
| Unknown | 0 (0) | 0 (0) | 2 (2.8) | 0 (0) |
| Samples by timepoint | ||||
| Uninfected cross-sectional | 80 (100) | … | … | 59 (100) |
| Day 0 | … | 33 (14.6) | 71 (74.0) | … |
| Week 2 | … | 31 (13.7) | 13 (13.5) | … |
| Week 4 | … | 32 (14.2) | 12 (12.5) | … |
| Month 3 | … | 30 (13.3) | … | … |
| Month 6 | … | 12 (5.3) | … | … |
| Month 12 | … | 32 (14.2) | … | … |
| Month 18 | … | 29 (12.8) | … | … |
| Month 24 | … | 23 (10.2) | … | … |
| Reinfection | … | 4 (1.8) | … | … |
Data are presented as No. (%) unless otherwise indicated.
Abbreviations: AgSAP, Anopheles gambiae sporozoite-associated protein; AgTRIO, Anopheles gambiae TRIO; EIR, entomological inoculation rate; HCW, healthcare worker; mosGILT, mosquito gamma interferon–inducible lysosomal thiol reductase; SAMSP1, sporozoite-associated mosquito saliva protein 1.
aThis group includes 16 longitudinal subjects, each sampled at a minimum of 2 timepoints, and 56 cross-sectional subjects.
Figure 1.
A, Mean fluorescence intensity (MFI) of immunoglobulin G (IgG) responses separated by cohort to SG6-P1, with geometric means shown by bars, 95% confidence intervals (CIs) shown by whiskers, and individual responses shown by dots. Cohorts are split by malaria transmission (Thiès is a low-transmission area with an entomological inoculation rate [EIR] <5 infectious bites per person per year, and Kédougou is an area of moderate malaria transmission with an EIR ∼250 infectious bites per person per year) and time relative to malaria infection (acute are samples from the time of malaria infection, recent are samples collected 2–4 weeks after malaria infection, transmission are samples collected from uninfected individuals during the malaria transmission season, and non-transmission are samples collected from uninfected individuals outside of the malaria transmission season). HCW is a cohort of healthcare workers from the United States who have not been exposed to malaria-infected mosquitoes. Heatmaps above each bar plot show significant differences between groups of linear mixed models of log-transformed MFI of post hoc pairwise comparisons adjusting P values with the Tukey method. X's indicate that comparisons were not performed. B, IgG MFI shown longitudinally by individual from Thiès and Kédougou for SG6-P1. Timepoints with infection are shown in red (day 0 and reinfection for Thiès, and day 0 for Kédougou), and timepoints without infection are shown in black. Malaria transmission seasons are shown in green. Blue line shows the geometric mean at each timepoint, and light blue shows the 95% CIs for the geometric mean (calculated excluding reinfection timepoints for Thiès).
Antibodies to AgSAP and AgTRIO Are Higher in Malaria-Endemic Areas
Human antibody responses to AgSAP and AgTRIO were significantly higher among people with acute and recent infection from both Kédougou and Thiès compared to the HCW cohort (for AgSAP: all P < .025; for AgTRIO: all P < .001; Supplementary Table 2). Additionally, AgTRIO mean fluorescence intensity (MFI) was higher among Thiès nontransmission season samples compared to HCW MFI (P = .0007), as was AgSAP MFI in uninfected people from Kédougou during the malaria transmission season (P = .0245) and AgTRIO MFI in uninfected individuals in both Thiès and Kédougou during the malaria transmission season (both P < .001) (Figure 2A and 2B). Overall, AgSAP and AgTRIO MFI was moderately correlated (Spearman correlation ρ = 0.636; P < .0001; Supplementary Figure 5).
Figure 2.
Mean fluorescence intensity (MFI) of immunoglobulin G responses separated by cohorts to the mosquito salivary antigens Anopheles gambiae sporozoite-associated protein (AgSAP; A), Anopheles gambiae TRIO (AgTRIO; B), mosquito gamma interferon–inducible lysosomal thiol reductase (mosGILT; C), and sporozoite-associated mosquito saliva protein 1 (SAMSP1; D), with geometric means shown by bars, 95% confidence intervals shown by whiskers, and individual responses shown by dots. Cohorts are split by malaria transmission (Thiès is a low-transmission area with an entomological inoculation rate [EIR] <5 infectious bites per person per year, and Kédougou is an area of moderate malaria transmission with an EIR ∼250 infectious bites per person per year) and time relative to malaria infection (acute are samples from the time of malaria infection, recent are samples collected 2–4 weeks after malaria infection, transmission are samples collected from uninfected individuals during the malaria transmission season, and non-transmission are samples collected from uninfected individuals outside of the malaria transmission season). HCW is a cohort of healthcare workers from the United States who have not been exposed to malaria-infected mosquitoes. Heatmaps above each bar plot show significant differences between groups of linear mixed models of log-transformed MFI of post hoc pairwise comparisons adjusting P values with the Tukey method. X's indicate that comparisons were not performed.
The HCW cohort had a relatively high reactivity for mosGILT and SAMSP1, resulting in no significant differences between HCWs and uninfected individuals in the transmission season in either Thiès or Kédougou (Figure 2C and 2D). For SAMSP1, only the Thiès recent infection cohort MFI was significantly higher than HCW MFI (P = .0182). For SG6, the results also did not differ between the HCW cohort and the cohorts in malaria-endemic areas (Figure 1A).
Antibody Responses to All Proteins Did Not Differ Either by Transmission Season or Intensity
None of the antibody responses distinguished seasonality from uninfected people in Thiès (Figures 1A and 2). MFI also did not differ between those in low-transmission Thiès versus moderate-transmission Kédougou (ie, between acute infection, recent infection, and uninfected transmission season samples in Thiès and Kédougou) (Figure 2).
Antibodies to AgSAP and AgTRIO Were Associated With Recent Infectious Mosquito Exposure
We next evaluated whether responses were associated with recent infectious mosquito bites using acute and recent infection. Immune reactivity to AgSAP and AgTRIO distinguished between cohorts with and without recent infectious mosquito exposure. For both AgSAP and AgTRIO in Thiès, people with acute or recent (2–4 weeks postdiagnosis) malaria infection had significantly higher reactivity than uninfected people in the transmission season (both P < .001; Figure 2A and 2B). However, in Kédougou, only those with recent infection had significantly higher AgSAP and AgTRIO reactivity than those uninfected from the transmission season (both P < .005); differences between those with acute infection compared to uninfected people in the transmission season in Kédougou were not significant.
For SAMSP1, in Thiès, recent infection increased IgG reactivity compared to uninfected individuals in the transmission season (P = .0043), but other comparisons were not significant (Figure 2D). For mosGILT, no differences were significant (Figure 2C). For SG6, people with acute infection in Thiès had higher IgG reactivity than uninfected individuals in the transmission season in Kédougou (P = .028; Figure 1A).
Longitudinally, Antibodies to AgSAP and AgTRIO Peaked at 2–4 Weeks Postinfection and Decreased by 3 Months After Infection
Peak AgSAP reactivity in the Thiès cohort was among individuals who had malaria infection 2 weeks prior (significantly higher than all timepoints other than week 4 and reinfection, P < .001) and 4 weeks prior (significantly higher reactivity than timepoints 3, 12, 18, and 24 months after infection, all P < .001; Figure 3A and Supplementary Table 3). Among the Kédougou cohort, reactivity 2 weeks and 4 weeks after infection were also significantly higher than at acute infection (P ≤ .01).
Figure 3.
Immunoglobulin G mean fluorescence intensity (MFI) shown longitudinally, by individual, from Thiès and Kédougou for each mosquito salivary antigen: Anopheles gambiae sporozoite-associated protein (AgSAP; A), Anopheles gambiae TRIO (AgTRIO; B), mosquito gamma interferon–inducible lysosomal thiol reductase (mosGILT; C), and sporozoite-associated mosquito saliva protein 1 (SAMSP1; D). Timepoints with infection are shown in red (day 0 and reinfection for Thiès, and day 0 for Kédougou), and timepoints without infection are shown in black. Malaria transmission seasons are shown in green. Blue line shows the geometric mean at each timepoint, and light blue shows the 95% confidence intervals for the geometric mean (calculated excluding reinfection timepoints for Thiès).
Similar to AgSAP, AgTRIO antibody responses in the Thiès cohort peaked among individuals who had malaria infection 2 weeks prior (significantly higher reactivity than all timepoints other than week 4 and reinfection, P < .01) and 4 weeks prior (significantly higher reactivity than at 3 months [P = .03] and 12, 18, and 24 months after infection [all P < .0001]; Figure 3B). AgTRIO reactivity among the Kédougou cohort was also highest at 2 weeks and 4 weeks after infection but was not significantly higher than at acute infection (P = .08 compared to week 2, and P = .09 compared to week 4).
For mosGILT, there were no significant antibody reactivity differences between timepoints in either cohort (Figure 3C). For SAMSP1 antibodies, in the Thiès cohort, reactivity among individuals who had malaria 2 weeks prior was significantly higher than reactivity at 3, 6, 12, 18, and 24 months after infection (all P < .05, Figure 3D). For SG6, people in the Thiès cohort with acute malaria infection and those with infection 2 weeks prior had significantly higher reactivity than people at 3, 12, and 18 months after infection (all P < .05; Figure 1B). No differences were significant between timepoints in the Kédougou cohort for SAMSP1 or SG6 reactivity (Figures 1B and 3D).
Using Cutoffs for Seropositivity, AgSAP had a High Specificity and AgTRIO had a High Sensitivity
AgSAP did not demonstrate a bimodal distribution and as such, using the negative HCW cohort, the positive cutoff value for AgSAP was an MFI of 1645 (calculated from mean + 2 standard deviations of log-transformed results; Supplementary Figure 4A). Specificity was 97.4% among HCWs (75/77). Sensitivity of AgSAP was 18.9% (7/37) and 22.5% (16/71) among acutely infection individuals in Thiès and Kédougou, respectively. Sensitivity of AgSAP was 47.6% (30/63) and 44.0% (11/25) among individuals in Thiès and Kédougou, respectively, with infection 2–4 weeks prior.
In contrast, AgTRIO MFI values exhibited a mildly bimodal distribution, and using a finite mixture model, the positive cutoff was set to an MFI of 500 (Supplementary Figure 4B). HCW MFI values for AgTRIO were too high to obtain a usable cutoff. With a cutoff of 500, specificity for HCWs was 51.9% (40/77). Sensitivity of AgTRIO was 83.8% (31/37) and 94.4% (67/71) among acutely infected individuals in Thiès and Kédougou, respectively. Sensitivity was 90.5% (57/63) and 100% (25/25) among recently infected individuals in Thiès and Kédougou, respectively. Seropositivity rates are shown in Figure 4 and described in the Supplementary Results.
Figure 4.
Seropositivity for each cohort for mosquito salivary antigens Anopheles gambiae sporozoite-associated protein (AgSAP) and Anopheles gambiae TRIO (AgTRIO). Cohorts are split by malaria transmission (Thiès is a low-transmission area with an entomological inoculation rate [EIR] <5 infectious bites per person per year, and Kédougou has a moderate malaria transmission with an EIR ∼250 infectious bites per person per year) and time relative to malaria infection (acute are samples from the time of malaria infection, recent are samples collected 2–4 weeks after infection, transmission are samples collected from uninfected individuals during the transmission season, and non-transmission are samples collected from uninfected individuals outside of the transmission season). HCW is a cohort of healthcare workers from the United States who are not expected to have exposure to malaria-infected mosquitoes. For AgSAP, cutoffs for positivity were calculated using values from the negative HCW cohort. Values were log-transformed for normality, and the mean + 2 standard deviations was calculated and then exponentiated to determine a cutoff for positivity. For AgTRIO, cutoffs for positivity were determined from a finite mixture model from a bimodal distribution [37].
IgG Responses to MSAs as Measured by Sample Type and Testing Platform Were Moderately Correlated
Comparisons by sample type (plasma vs dried blood spot) and by testing platform (multiplex vs ELISA) both showed moderate correlation (Supplementary Figures 6 and 7; described in detail in the Supplementary Results).
Homologous Proteins in Other Anopheles Species
Orthologous proteins of AgSAP, AgTRIO, mosGILT, SAMSP1, and SG6 are found in different mosquito species. The homology and percentage identity for these orthologs, particularly for other Anopheles species, are described in Supplementary Tables 4 and 5.
DISCUSSION
We determined that serological responses to MSAs can effectively indicate exposure to infectious mosquitoes, potentially serving as an alternative to EIR as a measure of P falciparum transmission intensity. Unlike EIR estimates, which are limited by sampling issues (eg, seasonal variation), precision (eg, variation in measuring HBR and contribution of indoor/outdoor biting), and accuracy (eg, spatial and temporal heterogeneity in vector density) [2], serological measurements are a more direct measure of the effect of mosquito bites on humans and may be less prone to such issues. Among the 4 MSAs, humoral responses to AgSAP and AgTRIO were the best indicators of recent infectious mosquito bites.
Higher IgG Reactivity to AgSAP and AgTRIO Is Associated With Malaria Endemicity
We show here that reactivity to AgSAP is higher in a moderate-transmission area than in healthy controls from a nonendemic area (Figure 2A), consistent with a previous study [26]. Similarly, AgTRIO responses were higher in both low- and moderate-transmission areas relative to a nonendemic area (Figure 2B). While a previous study found no significant difference [25], our study may have had greater power to detect this effect. Neither AgSAP nor AgTRIO could discriminate between populations in areas of different malaria endemicity (Figure 2A and 2B).
Although mosGILT and SAMSP1 are directly associated with Plasmodium sporozoites during transmission [24] and individuals in malaria-endemic Senegal were previously found to have higher SAMSP1 IgG reactivity than individuals living in nonendemic France [27], serological responses to mosGILT and SAMSP1 did not show clear patterns in relation to infectious mosquito exposure. The high reactivity of mosGILT, SAMSP1, and SG6 for the negative HCW cohort limited our ability to distinguish between populations.
In the Absence of Infection, AgSAP, AgTRIO, mosGILT, and SAMSP1 Are Not Biomarkers of General Mosquito Exposure or Malaria Transmission Season
The lack of differences between uninfected people in the transmission and nontransmission seasons for Thiès (comparison not available for Kédougou) and the lack of difference in cohorts in low- and moderate-transmission areas suggests that these MSAs are unable to determine seasonal differences in malaria exposure or intensity in the absence of infection. SG6 also did not show these differences, contrary to previous studies [7, 10–14, 18]. It is possible that SG6 differences would be observed in a high-transmission area rather than the low- and moderate-transmission areas tested here. SG6 reactivity was significantly higher in Kédougou in the transmission season than in Thiès outside of the transmission season, which could be due to higher exposure to mosquito bites due both to seasonal and regional differences.
AgSAP and AgTRIO Are Promising Indicators of Recent or Infectious Mosquito Exposure
AgSAP and AgTRIO both show similar trends in serological responses after exposure to an infectious mosquito. In a low-transmission area, where infectious mosquito bites are less common, IgG responses to both AgSAP and AgTRIO peak after infectious mosquito exposure (indicated by malaria infection) and decline by 3 months after infection. AgSAP [26] and AgTRIO [25] are both upregulated in infectious mosquitoes, and AgSAP and AgTRIO reactivity were highest among individuals with recent infection, suggesting that AgSAP and AgTRIO serological responses peak 2–4 weeks after clinical malaria, which is itself at least 7–10 days after an infectious bite [26, 38]. In both the low-transmission area of Thiès and the moderate-transmission area of Kédougou, AgSAP responses with recent infection were significantly higher than during acute infection (Figure 2A). AgTRIO responses were significantly higher during recent infection than acute infection in Thiès but not in Kédougou (Figure 2B). For both AgSAP and AgTRIO, responses decline after recent infection and are significantly lower 3 months after initial infection (Figure 3A and 3B), suggesting that the kinetics of AgSAP build a couple of weeks after infection and then decline relatively quickly.
SAMSP1 IgG responses are higher in individuals in Thiès with acute and recent malaria infection relative to responses in uninfected people in and outside of the transmission season and to those in a non-malaria-endemic area (Figure 2D). SAMSP1 reactivity may also peak 2–4 weeks after infection and decrease by 3 months after infection, although differences were not observed in Kédougou.
Seropositivity Could Show Differences Between Populations
A benefit of AgSAP and AgTRIO responses may lie in their ability to discriminate among changes in populations, for example, before and after a community-based intervention. Bimodal distributions can be an effective way to determine cutoffs. The current distribution of AgTRIO does not strongly fit a bimodal distribution, which limits this method's efficacy, but this method could still be explored in a disease surveillance context where the inclusion of more samples could improve the fit of the distribution [37]. Here, AgSAP had a relatively low sensitivity and a high specificity, while AgTRIO had high sensitivity and a low specificity. Both AgSAP and AgTRIO had higher sensitivity among recently infected than acutely infected individuals.
As multiple P falciparum antigens have been used to estimate malaria prevalence [4, 5, 39], we investigated whether AgSAP and AgTRIO results together could increase sensitivity and specificity. Results between AgSAP and AgTRIO were highly correlated (Spearman ρ = 0.636), and their response kinetics were similar, indicating no substantial benefit in combining these antigens.
Cohorts in Context
This study tested samples from 2 different regions of Senegal: Thiès, a low-transmission area with an EIR <5 infectious bites per person per year [32–34], and Kédougou, a moderate-transmission area with an EIR of approximately 250 [35]. The low malaria transmission in Thiès has been tracked over time [40, 41], and in the longitudinal cohort tested here, only 4 of 70 individuals had malaria reinfections over 2 years of follow-up. Conversely, since malaria transmission is higher in Kédougou, we expect this cohort to have been exposed to infectious mosquitoes more frequently than the Thiès cohort. This could result in different MSA immune responses between these areas, particularly for antigens upregulated in Plasmodium-infected mosquitoes. Therefore, individuals from Kédougou may be more likely to get a “boosting” effect from exposure to infected mosquitoes.
As such, longitudinal changes in responses from Thiès may demonstrate a more direct indicator of responses after exposure to Plasmodium-infected mosquito antigens. The longitudinal data from Thiès show clear peaks in AgSAP and AgTRIO responses 2–4 weeks after subjects had symptomatic malaria, and then tend to reduce during the subsequent 2 years of follow-up. Longitudinal data from moderate transmission Kédougou likewise show high AgSAP and AgTRIO reactivity up to 4 weeks after symptomatic malaria, but without longer follow-up, we cannot compare long-term kinetics of antibody responses in Kédougou to those in Thiès.
Limitations
We tested samples from a low- and moderate-transmission area of Senegal, but we cannot exclude the possibility that serological results may differ in an extremely high-transmission area. Although the kinetics of AgSAP and AgTRIO suggest that serological responses diminish within a couple of months, a person in a high-transmission area could be exposed to infectious mosquitoes more frequently. Our Kédougou cohorts aimed to model high transmission, but transmission in Kédougou is lower than transmission in intense, hyperendemic regions. A high-transmission area with more infectious mosquito exposure could cause serological responses to remain high, obscuring changes in malaria transmission. Additionally, we cannot distinguish between recent exposure to infectious mosquito bites that do or do not result in malaria infection. This could be evaluated, along with more exact timing of responses after infectious mosquito bites, through controlled human malaria infections.
There are homologous proteins to the proteins tested here in other Anopheles species, so reactivity to certain proteins may result from exposure to similar proteins in other mosquitoes (Supplementary Tables 4 and 5). If so, the distribution of mosquito species in a region could affect serological results. Cross-reactivity across Anopheles species has yet to be empirically tested. Most P falciparum transmission in West Africa is driven by An. gambiae, An. arabiensis, An. coluzzii, An. melas, An. funestus, An. stephensi, and An. nili [1]. In Senegal, An. gambiae is the predominant vector, though An. nili plays a relatively larger role in southeastern Senegal, including Kédougou [42].
CONCLUSIONS
This is a promising first study investigating human reactivity to 4 novel MSAs in areas of low and moderate malaria transmission. These salivary antigens represent potential complimentary antigens to SG6, an MSA that is a marker of exposure to mosquito bites, and P falciparum antigens that are markers of malaria parasite exposure to different stages of the parasite life cycle, as antigens that uniquely represent potential markers of recent exposure to infectious mosquito bites. AgSAP and AgTRIO are promising markers of exposure to infectious mosquito bites in low- and moderate-transmission areas. Further studies could investigate their utility in highly endemic areas. The multiplex approach used here is a flexible method that allows for the possibility of adding additional markers, which could make tests of these antigens in high-transmission areas and more geographical regions more feasible. This approach could efficiently track population-level changes in P falciparum exposure, particularly in regions already using multiplex surveillance, and before and after vector control interventions.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.
Supplementary Material
Contributor Information
Sarah Lapidus, Department of Epidemiology of Microbial Diseases, Yale School of Public Health.
Morgan M Goheen, Department of Epidemiology of Microbial Diseases, Yale School of Public Health; Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Mouhamad Sy, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Awa B Deme, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Ibrahima Mbaye Ndiaye, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Younous Diedhiou, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Amadou Moctar Mbaye, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Kelly A Hagadorn, Department of Epidemiology of Microbial Diseases, Yale School of Public Health.
Seynabou Diouf Sene, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Mariama Nicole Pouye, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Laty Gaye Thiam, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Aboubacar Ba, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Noemi Guerra, Department of Epidemiology of Microbial Diseases, Yale School of Public Health.
Alassane Mbengue, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Hamidah Raduwan, Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Jacqueline Gagnon, L2 Diagnostics, LLC, Department of Research and Development, New Haven, Connecticut.
Inés Vigan-Womas, G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Sunil Parikh, Department of Epidemiology of Microbial Diseases, Yale School of Public Health; Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Albert I Ko, Department of Epidemiology of Microbial Diseases, Yale School of Public Health; Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Daouda Ndiaye, Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; International Research and Training Center for Applied Genomics and Health Surveillance (CIGASS), Cheikh Anta Diop University, Dakar, Senegal.
Erol Fikrig, Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Yu-Min Chuang, Section of Infectious Diseases, Department of Internal Medicine, Yale School of Medicine, New Haven, Connecticut.
Amy K Bei, Department of Epidemiology of Microbial Diseases, Yale School of Public Health; Laboratory of Parasitology and Mycology, Cheikh Anta Diop University, Aristide le Dantec Hospital; G4—Malaria Experimental Genetic Approaches and Vaccines, Pôle Immunophysiopathologie et Maladies Infectieuses, Institut Pasteur de Dakar, Senegal.
Notes
Author contributions. A. K. B., Y.-M. C., and E. F. conceived and directed the project. S. L. and M. M. G. performed the experiments. H. R. and J. G. assisted with bead coupling. S. L. analyzed the data. M. S., A. B. D., Y. D., A. M. M., K. A. H., S. D. S., M. N. P., L. G. T., A. B., N. G., A. M., D. N., S. P., A. I. K., and A. K. B. performed data acquisition, cohort initiation, follow-up, and maintenance. A. K. B., M. G., and S. L. wrote the manuscript. All authors have read, provided feedback, and approved the submitted version of the manuscript.
Acknowledgments. We acknowledge all study participants and communities in Thiès and Kédougou, Senegal, for their participation in this project. We further acknowledge the healthcare workers from Yale New Haven Hospital who participated in this project. We wish to thank Alioune Wade for assistance with sample processing in Kédougou. We acknowledge Dyann F. Wirth for her mentorship on the Fogarty International Center K01, and continued support.
Data availability. Data associated with this manuscript can be found at: doi:10.5061/dryad.cz8w9gjbj.
Financial support. This work was supported by the Fogarty International Center of the National Institutes of Health (NIH) (award number K01 TW010496); the National Institute of Allergy and Infectious Diseases of the NIH (grant number R01 AI168238); the Ambrose Monell Foundation; and G4 group funding (G45267, Malaria Experimental Genetic Approaches and Vaccines) from the Institut Pasteur de Paris and Agence Universitaire de la Francophonie to A. K. B. This work was partly supported by the Howard Hughes Medical Institute Emerging Pathogens Initiative and NIH grant AI58615 (to E. F. and Y.-M. C.). S. L. is supported by the Yale Clinical and Translational Science Award (grant number UL1 TR001863) from the National Center for Advancing Translational Sciences, a component of the National Institutes of Health (NIH). M. M. G. is supported by a T32 training grant in Infectious Diseases (grant number 5T32AI007517–23). M. S. and A. B. D. are supported by NIH Global Health Equity Scholars Fellowships (award number FIC TW010540). L. G. T. is supported by an ARISE grant jointly funded from the African Academy of Sciences, the African Union, and the European Union. Additional support was provided by the Rotary Foundation.
References
- 1. World Health Organization (WHO) . World malaria report 2023. Geneva, Switzerland: WHO, 2023.
- 2. Tusting LS, Bousema T, Smith DL, Drakeley C. Measuring changes in Plasmodium falciparum transmission: precision, accuracy and costs of metrics. Adv Parasitol 2014; 84:151–208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wu L, Mwesigwa J, Affara M, et al. Sero-epidemiological evaluation of malaria transmission in The Gambia before and after mass drug administration. BMC Med 2020; 18:331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Rogier E, van den Hoogen L, Herman C, et al. High-throughput malaria serosurveillance using a one-step multiplex bead assay. Malar J 2019; 18:402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Helb DA, Tetteh KKA, Felgner PL, et al. Novel serologic biomarkers provide accurate estimates of recent Plasmodium falciparum exposure for individuals and communities. Proc Natl Acad Sci U S A 2015; 112:E4438–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Achan J, Reuling IJ, Yap XZ, et al. Serologic markers of previous malaria exposure and functional antibodies inhibiting parasite growth are associated with parasite kinetics following a Plasmodium falciparum controlled human infection. Clin Infect Dis 2020; 70:2544–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Badu K, Siangla J, Larbi J, et al. Variation in exposure to Anopheles gambiae salivary gland peptide (gSG6-P1) across different malaria transmission settings in the western Kenya highlands. Malar J 2012; 11:318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Fontaine A, Pascual A, Orlandi-Pradines E, et al. Relationship between exposure to vector bites and antibody responses to mosquito salivary gland extracts. PLoS One 2011; 6:e29107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Montiel J, Carbal LF, Tobón-Castaño A, Vásquez GM, Fisher ML, Londono-Rentería B. IgG antibody response against Anopheles salivary gland proteins in asymptomatic Plasmodium infections in Narino, Colombia. Malar J 2020; 19:42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Remoue F, Cisse B, Ba F, et al. Evaluation of the antibody response to Anopheles salivary antigens as a potential marker of risk of malaria. Trans R Soc Trop Med Hyg 2006; 100:363–70. [DOI] [PubMed] [Google Scholar]
- 11. Kearney EA, Agius PA, Chaumeau V, Cutts JC, Simpson JA, Fowkes FJI. Anopheles salivary antigens as serological biomarkers of vector exposure and malaria transmission: a systematic review with multilevel modelling. eLife 2021; 10:e73080. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Traoré DF, Sagna AB, Adja AM, et al. Exploring the heterogeneity of human exposure to malaria vectors in an urban setting, Bouaké, Côte d’Ivoire, using an immuno-epidemiological biomarker. Malar J 2019; 18:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Drame PM, Poinsignon A, Besnard P, et al. Human antibody responses to the Anopheles salivary gSG6-P1 peptide: a novel tool for evaluating the efficacy of ITNs in malaria vector control. PLoS One 2010; 5:e15596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sadia-Kacou CAM, Yobo CM, Adja MA, et al. Use of Anopheles salivary biomarker to assess seasonal variation of human exposure to Anopheles bites in children living near rubber and oil palm cultivations in Côte d’Ivoire. Parasite Epidemiol Control 2019; 5:e00102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Sagna AB, Kassié D, Couvray A, et al. Spatial assessment of contact between humans and Anopheles and Aedes mosquitoes in a medium-sized African urban setting, using salivary antibody-based biomarkers. J Infect Dis 2019; 220:1199–208. [DOI] [PubMed] [Google Scholar]
- 16. Ali ZMI, Bakli M, Fontaine A, et al. Assessment of Anopheles salivary antigens as individual exposure biomarkers to species-specific malaria vector bites. Malar J 2012; 11:439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Rizzo C, Ronca R, Fiorentino G, et al. Wide cross-reactivity between Anopheles gambiae and Anopheles funestus SG6 salivary proteins supports exploitation of gSG6 as a marker of human exposure to major malaria vectors in tropical Africa. Malar J 2011; 10:206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Cheteug G, Elanga-Ndille E, Donkeu C, et al. Preliminary validation of the use of IgG antibody response to Anopheles gSG6-p1 salivary peptide to assess human exposure to malaria vector bites in two endemic areas of Cameroon in Central Africa. PLoS One 2020; 15:e0242510. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Oseno B, Marura F, Ogwang R, et al. Characterization of Anopheles gambiae D7 salivary proteins as markers of human-mosquito bite contact. Parasit Vectors 2022; 15:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Jaramillo-Underwood A, Impoinvil D, Sutcliff A, et al. Factors associated with human IgG antibody response to Anopheles albimanus salivary gland extract, artibonite department, Haiti, 2017. J Infect Dis 2022; 226:1461–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Jaramillo-Underwood A, Herman C, Impoinvil D, et al. Spatial, environmental, and individual associations with Anopheles albimanus salivary antigen IgG in Haitian children. Front Cell Infect Microbiol 2022; 12:1033917. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Chuang YM, Tang XD, Fikrig E. A mosquito AgTRIO monoclonal antibody reduces early Plasmodium infection of mice. Infect Immun 2022; 90:e0035921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Chuang YM, Alameh MG, Abouneameh S, et al. A mosquito AgTRIO mRNA vaccine contributes to immunity against malaria. NPJ Vaccines 2023; 8:88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Schleicher TR, Yang J, Freudzon M, et al. A mosquito salivary gland protein partially inhibits Plasmodium sporozoite cell traversal and transmission. Nat Commun 2018; 9:2908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Dragovic SM, Agunbiade TA, Freudzon M, et al. Immunization with AgTRIO, a protein in Anopheles saliva, contributes to protection against Plasmodium infection in mice. Cell Host Microbe 2018; 23:523–35.e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Arora G, Sajid A, Chuang YM, et al. Immunomodulation by mosquito salivary protein AgSAP contributes to early host infection by Plasmodium. mBio 2021; 12:e0309121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Chuang YM, Agunbiade TA, Tang XD, Freudzon M, Almeras L, Fikrig E. The effects of a mosquito salivary protein on sporozoite traversal of host cells. J Infect Dis 2021; 224:544–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Chuang YM, Freudzon M, Yang J, Dong Y, Dimopoulos G, Fikrig E. Anopheles gambiae lacking AgTRIO inefficiently transmits Plasmodium berghei to mice. Infect Immun 2019; 87:e00326-19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. de Figueiredo AM, Glória JC, Chaves YO, Neves WLL, Mariúba LAM. Diagnostic applications of microsphere-based flow cytometry: a review. Exp Biol Med 2022; 247:1852–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Rogier EW, Giorgi E, Tetteh K, Sepúlveda N. Editorial: current research on serological analyses of infectious diseases. Front Med 2023; 10:1154584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Chan Y, Martin D, Mace KE, et al. Multiplex serology for measurement of IgG antibodies against eleven infectious diseases in a national serosurvey: Haiti 2014–2015. Front Public Health 2022; 10:897013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Wong W, Griggs AD, Daniels RF, et al. Genetic relatedness analysis reveals the cotransmission of genetically related Plasmodium falciparum parasites in Thiès, Senegal. Genome Med 2017; 9:5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Ndiaye YD, Wong W, Thwing J, et al. Two decades of molecular surveillance in Senegal reveal rapid changes in known drug resistance mutations over time. Malar J 2024; 23:205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Programme National de Lutte contre le Paludisme . Bulletin epidémiologique annuel 2022 du paludisme au Sénégal: Ministère de la Santé et de l’Action Sociale. 2023. Available from: https://www.sante.gouv.sn/programmes-et-projets/programme-national-de-lutte-contre-le-paludisme%C2%A0-pnlp. Accessed October 1, 2024. [Google Scholar]
- 35. Dia I, Diop T, Rakotoarivony I, Kengne P, Fontenille D. Bionomics of Anopheles gambiae Giles, An. arabiensis Patton, An. funestus Giles and An. nili (Theobald) (Diptera: Culicidae) and transmission of Plasmodium falciparum in a Sudano-Guinean zone (Ngari, Senegal). J Med Entomol 2003; 40:279–83. [DOI] [PubMed] [Google Scholar]
- 36. Oviedo A, Knipes A, Worrell C, et al. Combination of serological, antigen detection, and DNA data for Plasmodium falciparum provides robust geospatial estimates for malaria transmission in Haiti. Sci Rep 2020; 10:8443. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Rogier E, Wiegand R, Moss D, et al. Multiple comparisons analysis of serological data from an area of low Plasmodium falciparum transmission. Malar J 2015; 14:436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Cowman AF, Healer J, Marapana D, Marsh K. Malaria: biology and disease. Cell 2016; 167:610–24. [DOI] [PubMed] [Google Scholar]
- 39. Rogier E, Nace D, Dimbu PR, et al. Antibody dynamics in children with first or repeat Plasmodium falciparum infections. Front Med 2022; 9:869028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Sy M, Deme AB, Warren JL, et al. Plasmodium falciparum genomic surveillance reveals spatial and temporal trends, association of genetic and physical distance, and household clustering. Sci Rep 2022; 12:938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Daniels R, Chang HH, Séne PD, et al. Genetic surveillance detects both clonal and epidemic transmission of malaria following enhanced intervention in Senegal. PLoS One 2013; 8:e60780. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. US President’s Malaria Initiative . The PMI VectorLink Project annual report: October 1, 2021–September 30, 2022. 2022. https://pmivectorlink.org/results/project-results/. Accessed October 1, 2024. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.