Malaria is one of the most important parasitic diseases of the tropics and subtropics. Despite global decreases in malaria cases, Latin America has experienced an increase in cases since 2015. Colombia is among the top five Latin American countries with the highest numbers of malaria cases reported annually (Chaparro-Narvaez et al., 2016). In 2019 the number of malaria cases officially reported was 78,513 with Chocó, Nariño, Cordoba, and Antioquia, the departments with the highest number of malaria cases (31.4, 20.4, 11.6, and 9.3%, respectively) (Instituto Nacional de Salud, 2019). Several factors contribute to this phenomenon. First, more than 80% of the Colombian territory is suitable for the transmission of malaria due to the presence of a diverse number of mosquito species capable of serving as vectors of Plasmodium spp. (Rodriguez et al., 2011). Second, at least three species of Plasmodium cause infection in Colombia, with Plasmodium vivax as the main species, followed by Plasmodium falciparum and Plasmodium malariae (Arevalo-Herrera et al., 2012). Forced migration and precarious socioeconomic conditions also favor the spread and prevalence of Plasmodium spp. (Martens and Hall, 2000; Jitthai, 2013; Rodrigues et al., 2018).

Globally, most efforts to prevent malaria transmission focus on P. falciparum, the deadliest malaria parasite species. However, P. vivax is also a significant cause of morbidity and mortality (Chaparro-Narvaez et al., 2016; Geleta and Ketema, 2016; Zain Ul et al., 2016), and improved methods of monitoring its transmission are needed. Malaria transmission in Colombia occurs year-round, with peaks typically occurring between February and June (Ruiz et al., 2006). Malaria incidence is highly associated with changes in climatic conditions (i.e., temperature, elevation, and humidity) (Ruiz et al., 2006; Cassab et al., 2011). More importantly, ~87% of the reported cases come from people living in rural areas with limited access to health care (Rodriguez et al., 2011). Frequently, research about malaria transmission in these areas is restricted due to the lack of infrastructure to support research, including limited access to electricity.

Currently, the gold standard to determine malaria transmission intensity is the Entomological Inoculation Rate (EIR), which represents the number of infectious mosquito bites a person receives per night. This provides an indirect estimation of the risk of acquiring the disease. One of the challenges of this technique is the need to detect parasites in mosquitoes that are collected in the field. Mosquito trapping systems can be biased and overrepresent the most abundant species in a specific area. In addition, current methods are unable to pinpoint the exact individuals who have been bitten or whether there are differences in the number of bites different people receive. Thus, current surveillance efforts lack individualized, deployable, and cost-effective tools to measure human-vector contact and improve measurement of malaria transmission risk. Such tools would be particularly valuable in assessing the efficacy of vector control interventions in protecting people from malaria vector bites.

Our previous studies demonstrated that IgG antibodies against salivary proteins of major disease vectors correlate with the clinical presentation of dengue fever and malaria (Londono-Renteria et al., 2010; Londoño-Rentería et al., 2015a; Montiel et al., 2020a), and can be used to evaluate vector control interventions (Londono-Renteria et al., 2015c). In fact, IgG antibodies against a salivary peptide (gSG6-P1) from a major African malaria vector has been validated as a reliable marker to measure exposure to several Anopheles spp. (Poinsignon et al., 2009; Badu et al., 2012; Sagna et al., 2013). However, recent studies suggest that the gSG6 gene is not present in the subgenus Nyssorhynchus, which is the most prevalent American Anopheles subgenus (Arca et al., 2017). Although, previous studies have shown that the gSG6-P1 may be useful in areas of the New World where species belonging to subgenus other than Nyssorhynchus (Londono-Renteria et al., 2015b; Montiel et al., 2020a), there is a need to detect and validate novel biomarkers for exposure to these Anopheles spp. There are at least 40 Anopheles species in Colombia (Montoya-Lerma et al., 2011). The main vectors in the country are Anopheles nuneztovari, Anopheles albimanus and Anopheles darlingi—all three species within the subgenus Nyssorhynchus (Montoya-Lerma et al., 2011). Recently, we described several peptides that could be used to determine exposure to Anopheles albimanus bites, with antibodies against transferrin and salivary peroxidase proteins associated with the clinical presentation of malaria (Montiel et al., 2020a).

In this study, we evaluated the immunogenicity of salivary proteins in An. darlingi salivary gland antigen (SGA) and designed four peptides to measure IgG responses in people living in malaria endemic areas. The objective was to evaluate whether IgG antibodies against these peptides could be reliable markers to detect the risk of disease. An. darlingi is one of the most prolific mosquito vectors of malaria in Central and South America (Mirabello and Conn, 2006; Villarreal-Trevino et al., 2015; Prado et al., 2019), known for its behavioral plasticity and highly anthropophilic behavior. This mosquito species bite indoors and outdoors (endophagic/exophagic) and display exophilic/endophilic opportunistic behaviors (Moutinho et al., 2011; Reinbold-Wasson et al., 2012; Moreno et al., 2015). Our study suggests that IgG antibodies against An. darlingi salivary proteins are useful to assess malaria risk in areas where this species is endemic. This is the first study describing An. darlingi salivary peptides as potential tools to evaluate immune responses against mosquito saliva with the aim to improve accurate estimation of malaria transmission intensity, which is imperative for directing control efforts and predicting the effects of such interventions.

An. darlingi is one of the most important malaria vectors in the Americas, and is associated with transmission of both P. falciparum and P. vivax (de Arruda et al., 1986; Flores-Mendoza et al., 2004). Some data suggest that infected An. darlingi tend to bite more outdoors and in the early evening, which reduces the effectiveness of bed nets (Moreno et al., 2015; Newell et al., 2018). The (EIR) is the current gold standard to estimate malaria transmission intensity. However, this technique is labor intensive and requires sensitive methods to accurately detect Plasmodium infection in field-captured mosquitoes. In addition, it is impossible to track individual human exposure to infective bites and estimation of risk that can be later used to determine local transmission dynamics. Tools capable of monitoring human-vector contact are needed to more accurately estimate disease risk. Here, we evaluated the IgG responses against An. darlingi salivary proteins and showed through ELISA that peptides against a salivary apyrase and peroxidase are suitable antigens to estimate the risk of suffering a malaria infection.

In an effort to identify candidate salivary peptides from An. darlingi salivary proteins, an immunogenic ~65 kDa protein band was identified by antibodies in serum collected from study participants. LC-MS revealed 17 salivary proteins, with apyrase and a peroxidase as the most predominant proteins. Most of the identified proteins have previously been associated with blood feeding. Salivary apyrases hydrolyze ATP and ADP inhibiting platelet aggregation. They are associated with host-seeking behavior and the ability of mosquitoes to find blood vessels (Calvo et al., 2006). Infection status with human pathogens influences mosquito behavior and changes the content of salivary gland and salivary secretions (Chisenhall et al., 2014a,b). A previous study suggested that apyrase activity decreases as sporozoite load increases, rendering the infected mosquitoes more likely to probe and re-feed after an infectious first blood meal (Thievent et al., 2019). Each probing event increases the likelihood of injecting sporozoites into the skin, which increases the chance of infection. Salivary proteins are also injected during each probing event (Ribeiro, 2000; Chisenhall et al., 2014a), which may explain the high antibody responses against salivary proteins in people with active malaria infection. Interestingly, immunization with mosquito saliva leads to the development of anti-apyrase antibodies, which can block its enzymatic activity. However, a decrease in feeding or probing time was not observed in immunized mice even in the presence of higher titers of anti-apyrase antibodies (Mathews et al., 1996). Previous studies have shown changes in salivary gland physiology of infected mosquitoes (Cotama et al., 2013; Pinheiro-Silva et al., 2015). It is unclear if infection with either P. vivax or P. falciparum changes the expression profile of apyrase or peroxidase proteins in the salivary gland of An. darlingi mosquitoes. This information would be helpful to determine a potential function of these proteins in malaria transmission by An. darlingi mosquitoes.

Besides the apyrase and the salivary peroxidase described above, other important salivary proteins that had been previously reported were also detected in this study. First, two members of the SG1 family were detected: the putative salivary protein SG-1B along with a SG1-like salivary protein. The SG-1B appears to be uniquely expressed in the female salivary glands, while the SG1-like can also be expressed in males (Arca et al., 2005). Recently, a member of the SG1 family, TRIO protein, was evaluated as a potential vaccine candidate against Plasmodium infection since immunization against these proteins induced partial protection against Plasmodium berghei infection. Protection was further improved when animals were co-immunized with saliva and parasite proteins (Dragovic et al., 2018). Second, a long form D7 protein was also detected. Our previous studies demonstrated the immunogenicity of this family of proteins in Ae. aegypti saliva (Londono-Renteria et al., 2018). The D7 protein family has been identified as a member of an odorant binding protein (OBP) superfamily known to bind biogenic amines such as histamine, playing an important role as mediators of inflammation and vascular permeability, and facilitating blood feeding (Calvo et al., 2006). Finally, a 30 kDa salivary antigen family protein was identified. These proteins are exclusively found in the salivary glands of adult female mosquitoes and previous work suggested that they may inhibit platelet aggregation (Yoshida and Watanabe, 2006; Calvo et al., 2007b). A venom allergen 5 was also identified, and this family of proteins is found ubiquitously in animals and plants. Their specific function in Anopheles mosquitoes is still unknown, but it is thought to function either in the suppression of the host immune system, or as an anticoagulant in a wide range of hematophagous arthropods (Ribeiro and Francischetti, 2003; Calvo et al., 2007a).

It is important to note that whole salivary gland proteins contained in the SGA have been successfully used as antigen to detect IgG antibodies in vertebrate serum (Waitayakul et al., 2006; Badu et al., 2012). However, salivary gland dissection is a tedious process requiring skilled personnel (Coleman et al., 2007) and a continuous source of mosquito salivary glands. Not all mosquitoes are easily colonizable, and although using salivary glands from field mosquitoes can offer a better approximation of the antigens that people are actually exposed to (Andrade et al., 2009), it can be challenging to collect them in sufficient numbers and maintain a proper cold chain and protein stabilization before they are processed for antigen preparation (Fontaine et al., 2011a). Thus, the use of recombinant proteins or peptides is advantageous, particularly when consistency and reproducibility are desired. In that regard, we selected four peptides from the two main immunogenic proteins found after sequencing of the ~65 kDa band, an apyrase and a salivary peroxidase. The ELISA results revealed that when using the AnDar_Apy1, AnDar_Apy2, and AnDar_PeroX1 peptides as antigens, the IgG antibody levels were significantly higher in samples with active malaria infection as compared to uninfected samples. These results are consistent with our recent study where we observed that malaria-infected samples had significantly higher antibody levels against An. albimanus salivary peptides (Londono-Renteria et al., 2020).

Previous studies associated this increase in antibodies with a potentially higher exposure to mosquito bites (Fontaine et al., 2011b). Moreover, this increase in IgG antibodies may also be the result of saliva-dependent activation of immune cells leading to the production of cytokines such as IL4 and IL10 (Vogt et al., 2018). IL4 is involved in the activation of B cells and their differentiation to plasma cells (Granato et al., 2014) and recent studies suggest that IL10 promotes IgG4 antibodies (Jeannin et al., 1998), which is the main antibody subtype that recognizes arthropod salivary antigens (Brummer-Korvenkontio et al., 1994; Cardenas et al., 2019). Further studies are needed to determine the mechanisms involved in the differential antibody profiles between people residing under the same conditions but with different malaria infection status.

In contrast to a previous study in Brazil (Andrade et al., 2009), we did not find significant differences between malaria-infected and uninfected samples when using SGA. One of our previous studies showed the possibility of discrete differences in the responses against Anopheles SGA from two different mosquito colony strains, with those differences potentially associated with differences in geographical origins as well as the time of colonization (Londono-Renteria et al., 2020; Montiel et al., 2020a). Another study detected important differences in salivary gland content from wild and colony maintained arthropods from the same species (Maldonado-Ruiz et al., 2019). Consequently, it is possible that the difference between this current study and the one performed in Brazil may reside in the source of the SGA and its components. Specifically, in this study, we used SGA from a recently colonized strain of An. darlingi in Iquitos, Peru, while the previous study used SGA from field-caught mosquitoes in Brazil (Andrade et al., 2009).

Regardless, analysis of the immunogenic proteins among mosquito strains may reveal shared immunogenic proteins that represent suitable candidates to measure risk of disease irrespective of the geographical area or the strains. Thus, it was important to identify the evolutionarily conserved regions within immunogenic proteins as we seek to develop widely applicable biomarkers of malaria risk. Only AnDar_Apy2 was significantly higher in El Bagre, the area with historically higher abundance of An. darlingi relative to Turbo (Gutierrez et al., 2009, 2010; Rosero et al., 2013); all other peptides did not show differences between the two sites. The lack of difference in antibody levels against most of the peptides between the two study sites may be explained by the 56 and 80% identity with other major salivary proteins representing a source of cross-reactivity. However, since we did not include concurrent mosquito collection data, we cannot confirm the concurrent An. darlingi population density at the time of this study. Thus, the results presented here are more sensitive at correlating antibodies against salivary proteins with presence of malaria infection rather than An. darlingi bite intensity.

Previous studies suggested that there is a significant association between the level of IgG antibodies against the most prominent Plasmodium antigens and the development of human immunity against malaria symptoms (Doolan et al., 2009; Rodriguez-Barraquer et al., 2018). Among these antigens, MSP-1 is considered as one of the more important in the development of immunity against malaria and has been investigated as a potential vaccine candidate. Immunity against malaria is thought to be acquired through chronic and sustained exposure to Plasmodium antigens (Rodriguez-Barraquer et al., 2018). However, recent evidence suggests that immune responses against salivary proteins may also contribute to the development of immunity against disease (Manning et al., 2018). Our current study revealed that the level of IgG antibodies against AnDar_Apy 2 and the AnDar_PeroX2 were significantly associated with IgG levels against both PvMSP1 and PfMSP1 in all study groups, suggesting a positive association between intensity of exposure to malaria antigens and both peroxidase and apyrase salivary proteins. Furthermore, when measuring the risk of malaria infection through odd ratio calculations, a significant increase in risk was only observed in antibodies against the apyrase peptides. Previous study showed an association between the level of antibodies against PfMSP1 and the only currently validated salivary biomarker from the An. gambiae salivary peptide, gSG6-P1 (Badu et al., 2012; Ya-Umphan et al., 2018). Our results suggest that AnDarApy1 and AnDarApy2 may be more suitable biomarkers to estimate malaria risk in areas where An. darlingi mosquitoes are endemic. Also interesting is that the correlation between Plasmodium antigens and the apyrase peptides salivary was also significant in the uninfected population. It is important to remember that infected individuals present significantly higher IgG antibodies. People in endemic areas may encounter a significant number of bites containing infective sporozoites throughout their lives. Furthermore, in endemic areas for P. vivax, as the one where this study was carried out, is very difficult to differentiate a recent infection from a recrudescence, and people may suffer several malaria infections in a year. Since this study did not included follow up of patients, we cannot pinpoint the time of the most recent infection. However, our study showed a significant higher antibody levels against both apyrases peptides and the AnDar_Perox1 in P. vivax infected populations suggesting the relevance of these peptides to measure risk of infection with the most prevalent species circulating in this area. Further studies are urgently needed to determine the persistence of antibodies against the salivary proteins and correlate with previous/recent malaria episodes.

Although further studies including a greater number of serum samples from different geographical areas are needed to validate these peptides as biomarkers of both malaria risk and human-vector contact, the current study represents an important advance in evaluating immune responses against mosquito salivary peptides in Latin America.

Peptides derived from the An. darlingi salivary apyrase and peroxidases are suitable candidate markers for measuring human IgG antibody responses that may be associated with the risk of malaria infections.

All datasets generated for this study in the sequencing data are included in the article and the Supplementary Material. Other raw data will be available upon request sent to blondono@ksu.edu.

The studies involving human participants were reviewed and approved by Kansas State University and University of Antioquia. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

BL-R, AL, HV, GV, EC, MF, and MC: study design and data analysis, writing, and reviewing. JM and AT-C: sample collection. KE-V, LR, and MB: mosquito rearing and salivary gland dissection. BL-R and JM: ELISA testing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.