The human immune response to saliva of Phlebotomus alexandri, the vector of visceral leishmaniasis in Iraq, and its relationship to sand fly exposure and infection

Background Sand fly saliva exposure plays an important role in immunity against leishmaniasis where it has mostly been associated with protection. Phlebotomus (Ph.) alexandri transmits Leishmania (L.) infantum, the causative agent of visceral leishmaniasis (VL), in Iraq. Our group recently demonstrated that 20% of Operation Iraqi Freedom (OIF) deployers had asymptomatic VL (AVL) indicative of prior infection by the parasite L. infantum. Little is known about Ph. alexandri saliva, and the human immune response to it has never been investigated. Here, we characterize the humoral and cellular immune response to vector saliva in OIF deployers naturally exposed to bites of Ph. alexandri and characterize their immunological profiles in association to AVL. Methodology/Principal findings The humoral response to Ph. alexandri salivary gland homogenate (SGH) showed that 64% of 200 OIF deployers developed an antibody response. To assess the cellular immune response to saliva, we selected a subcohort of subjects based on their post-travel (median 4 months; range 1–22 months) antibody response (SGH Antibody [Ab] positive or negative) as well as their AVL status; ten never-traveled controls were also included. Banked peripheral blood mononuclear cells (PBMC), collected ~10 years after end of deployment, were stimulated with SGH for 96 hours. The levels of IFN- γ, IL-6, IL-10, IL-13 and IL-17 were determined by ELISA. Our findings indicate that OIF deployers mounted a cellular response to SGH where the anti-SGH+ asymptomatic subjects developed the highest cytokine levels. Further, stimulation with SGH produced a mixture of pro-inflammatory and anti-inflammatory cytokines. Contrary to our hypothesis, we observed no correlation between the cellular immune response to Ph. alexandri SGH and prevention from asymptomatic infection with L. infantum. Conclusions/Significance As we found, although all infected deployers demonstrated persistent disease control years after deployment, this did not correlate with anti-saliva systemic cellular response. More exposure to this vector may facilitate transmission of the L. infantum parasite. Since exposure to saliva of Ph. alexandri may alter the human immune response to bites of this vector, this parameter should be taken into consideration when considering the VL risk.


Introduction
Visceral leishmaniasis (VL), due to Leishmania infantum, is a persistent intracellular parasitic infection transmitted by the bite of infected sand flies. VL is endemic in central Iraq and has expanded to southeastern regions after the Gulf War [1]. In Iraq, the Leishmania (L). infantum parasite is transmitted by the bite of an infected Phlebotomus (Ph). alexandri sand fly [2,3]. In 2003-4, a 0.09% L. infantum infection rate was reported for Ph. alexandri at Tallil Air Base, Iraq [4]. Asymptomatic visceral leishmaniasis (AVL), indicative of infection with L. infantum, showed a substantial prevalence in Operation Iraqi Freedom (OIF) deployers. Using molecular and immunological assays on blood samples, we showed that 20% of OIF deployers (n = 200) had asymptomatic L. infantum infection defined by a positive cellular/humoral response or nucleic acid testing result [5]. Although Ph. alexandri is anthropophilic, it is not peridomiciliary and is mostly found in outdoor environments [6,7]. In mountainous areas, the density of sand flies is decreased and restricted to places with sporadic cottages or stables [8]. A shift in peak activity from early evening in April and October to later in the night during the hotter months of May and June was observed for several sand fly species in Iraq, including Ph. alexandri, likely increasing the rate of bites during sleep [9].
Compared to several sand fly species such as Lutzomyia (Lu.). longipalpis, the New World VL vector [10], and Ph. papatasi and Ph. perniciosus, Old World vectors of cutaneous leishmaniasis (CL) and VL respectively [11,12], little is known about Ph. alexandri and the immunogenicity of its saliva in humans. Exposure to Phlebotomus species bites or salivary proteins results in strong cellular and/or humoral immunity specific to some of these components [13,14]. Experimental exposure of naïve hosts (including mice [15], dogs [16] and humans [17]) to sand fly bites shows that antibody responses to saliva are acquired rapidly, usually within a few weeks of exposure and that the magnitude of the antibody response increases with the number of sand fly bites [16]. Field studies have also demonstrated the presence of anti-saliva antibodies in humans naturally exposed to sand fly bites [15,18]. Antibodies against sand fly saliva or against salivary recombinant proteins may serve as markers of vector exposure [19,20] as well as potential surrogate markers for the risk of Leishmania infection [21].
Regarding human cellular immune responses to sand fly saliva, human volunteers experimentally exposed to Lu. longipalpis bites displayed an increased frequency of CD4 + and CD8 + T cells as well as an increase in IFN-γ and IL-10 production upon in vitro PBMC stimulation with vector salivary gland homogenate (SGH) [17]. Analysis of cellular immune responses against Ph. papatasi saliva in PBMC from individuals naturally exposed to bites of this vector showed low proliferation, absence of IFN-γ production but significant IL-10 levels, which could favor establishment of infection with L. major [22]. Mali individuals exposed to bites of the CL vector Ph. duboscqi displayed a systemic immune response to its saliva involving the production of Th1 and Th2 cytokines; in contrast an abundant local expression of IFN-γ in the absence of a Th2 response was observed in skin biopsies of bite sites from individuals with a positive delayed-type hypersensitivity reaction [23].
Cellular and humoral responses to Ph. alexandri saliva have not been reported to date. This vector species has not been raised in insectaries, necessitating use of wild-caught sand flies for our work. Here, we report an analysis of cellular and humoral responses directed against Ph. alexandri saliva in Iraq-deployed U.S. personnel, presumed immunologically-naïve to vector saliva prior to deployment. We also investigate saliva-specific responses in individuals with or without AVL; notably in this cohort there were no participants with active VL [5].

Ethics statement
This research protocol was approved by the Uniformed Services University (USU), Walter Reed National Military Medical Center (WRNMMC), and the William Beaumont Army Medical Center (WBAMC) Institutional Review Boards (IRB). All participants provided written informed consent.

Characterization of sample and data collection sources
The AVL study cohort, further described in [5], included 200 OIF deployed asymptomatic individuals and 46 never-traveled healthy control subjects (HC, living in US, and have never been to a leishmaniasis endemic area lifelong). To study the long-lasting cellular responses, peripheral blood mononuclear (PBMC) cells were isolated from blood collected on average a decade after deployment; however the humoral response to SGH was assessed using banked sera from the same 200 OIF deployers, collected a median of 4 months (range 1-22 months) after return from deployment [5]. These sera were obtained from the Department of Defense Serum Repository, The Armed Forces Health Surveillance Branch, U.S. Department of Defense, Silver Spring, MD; release dates 2015-2017. To study the cellular immune responses to Ph. alexandri SGH, we selected a subcohort of 82 individuals (41% of deployers) and 10 never traveled controls based on residual cell availability (banked cells) to perform all required testing, reactivity in SGH ELISA, and AVL status. Using PBMC, AVL status was predetermined [5] as having either an elevated interferon gamma level post-stimulation with soluble Leishmania antigens (SLA), a positive SLA antibody response, or a positive L. infantum PCR (REPL target) [24]. For this study, we classified the selected subcohort into four groups ( Table 1): AVL -SGHAb + (Ab: Antibody) (n = 35) (AVL -: deployers who did not have evidence of exposure to Leishmania), AVL + SGHAb + (n = 28) (SGHAb + : presence of antibody response to SGH), AVL + SGHAb -(n = 9), AVL -SGHAb -(n = 10) and included never traveled controls (n = 10). Sand fly exposure information was abstracted from the AVL leishmaniasis risk factor survey [5].

Ph. alexandri collection and SGH preparation
Ph. alexandri sand flies were collected using CDC light traps (John W. Hock Company) during the summers of 2016 and 2017 at a site on the outskirts of Waqqas, Jordan located in the northern Jordan Valley, GPS coordinates 32.547852 N, 35.614674 E. Waqqas is a small city in the Irbid governorate with an average annual rainfall of 250 ml. Ph. alexandri females were identified using pharyngeal and spermathecal characteristics [25]. Salivary glands from female Ph. alexandri were dissected and transferred to polypropylene vials, usually in groups of 20 pairs of glands in 20 μl of HEPES-saline. Salivary glands were stored at −80˚C until needed, when they were disrupted by sonication and prepared as previously described [5].

Determination of anti-SGH antibody response
Banked sera were analyzed using an indirect enzyme-linked immunosorbent assay (ELISA) that measures anti-Ph. alexandri SGH total IgG as previously reported [5]. The ELISA cutoff was defined as average OD of unexposed controls + 3SD (0.108).

Cell stimulation and cytokine quantification
Banked PBMC were thawed and rested overnight at 37˚C, 5% CO 2 . One million cells were plated in triplicate and stimulated with 1ug/ml Ph. alexandri SGH, 10% phytohemagglutinin (PHA, Life Technologies) or left unstimulated. Ninety-six hours later, supernatants were collected and stored at -80˚C for cytokine quantification. Levels of IFN-γ, IL-4, IL-6, IL-10, IL-12p70, IL-13 and IL-17 in the supernatants were determined by ELISA using commercial cytokine kits (Uncoated human ELISA kits, Invitrogen) as specified by the manufacturer.

Statistical analysis
Cytokine levels were compared using a nonparametric Wilcoxon test and a Pearson correlation analysis. All tests were performed using Prism software (GraphPad Software, Inc.). Differences were considered significant if p value < 0.05. For the analysis of risk factors for sand fly exposure, differences in categorical variables were analyzed using Fisher's exact or Chi squared tests; continuous variables were assessed using Student's t-test or analysis of variance (ANOVA). Odds ratios are stated as point estimates with 95% Confidence Intervals (95% C.I.)

Individuals deployed to L. infantum endemic Iraq developed antibodies against Ph. alexandri saliva
We examined the antibody levels to Ph. alexandri SGH in serum samples of 200 Iraq deployed individuals and 46 control individuals [5]. Sixty four percent of OIF deployers had an antibody response to Ph. alexandri saliva ( Fig 1A) (accounting for 77% AVL + and 61% AVL -) (p = 0.06, Odds Ratio 2.14 [95% CI 0.95, 4.81]). We also evaluated the persistence of this humoral response by measuring antibody reactivity to SGH after OIF deployers had returned to the United States where Ph. alexandri is absent. Interestingly, antibodies to sand fly saliva could be detected in a few subjects as late as 18 months post exposure to bites of Ph. alexandri (Fig 1B).

The systemic immune response to Ph. alexandri SGH is skewed toward a Th2 profile
To assess the cellular immune response to Ph. alexandri saliva, we selected a subcohort of 82 OIF deployers and 10 controls ( Table 1). The selected subcohort consisted mainly of male   (Table 1). These demographics were closely aligned with the overall cohort characteristics except the age of the control group is significantly different than each of the deployed groups (ANOVA p < .0001, Tukey's HSD p < .01) [5]. The OIF deployers were sorted into AVL -SGHAb + , AVL + SGHAb + , AVL + SGHAband AVL -SGHAbgroups based on their antibody reactivity to vector saliva and AVL status. Stimulation of banked PBMCs from 82 OIF deployers with Ph. alexandri SGH produced variable and higher amounts of IFN-γ, IL-6, IL-13, IL-10 and IL-17 compared to control subjects (Fig 2A). Though not statistically significant, potentially due to the variability of the response in OIF deployers, this indicates prior contact with Ph. alexandri. No IL-4 or IL-12p70 were detected. Next, we compared the cellular response in OIF deployers with or without a anti-SGH antibodies(likely indicative of a significant or weak/absent exposure to bites by Ph. alexandri, respectively). The SGHAb + group consisted of 35 individuals selected for the highest serum total IgG values against Ph. alexandri SGH (group average OD = 0.23). Levels of most SGH-stimulated cytokines were higher in those who were seropositive to Ph. alexandri saliva compared to those who were seronegative (SGHAb -), with the exception of IL-10 and IL-17 ( Fig 2B). The fold increase in the mean cytokine levels over controls were 2.62, 1.23 and 1.75-fold higher for IFN-γ, IL-13 and IL-6, respectively, in SGHAb + compared to SGHAbgroup. The mean levels of IL-10 were similar in both groups while IL-17 levels were 2-fold lower in the SGHAb + group (Fig 2B).
After defining the profile based on reactivity to SGHAb, we further compared the cytokine response in terms of AVL status (AVL + or AVLindividuals) (Fig 2C). Among the studied groups, the AVL + SGHAbhad a relatively poor response, while the magnitude of the cellular immune response against saliva was highest in the AVL + SGHAb + group and consisted mainly of IL-6, IL-10 and IL-13. AVL + SGHAb + produced a significantly higher level of IL-13 (p = 0.023) compared to HC (Fig 2C). Though, a statistically significant difference was observed for IL-13 for AVL + SGHAbversus AVL -SGHAb + (p = 0.0039) and AVL + SGHAbversus HC (p = 0.02), this was caused by an outlier responder and is not considered representative of the group response ( Fig 2C). Moreover, we saw a 4-fold increase in IFN-γ for AVL + S-GHAb + (mean 40 pg/ml) compared to AVL -SGHAb + group (mean 10 pg/ml), though the difference was not statistically significant.
Based on the proinflammatory (IFN-γ and IL-17) or anti-inflammatory (IL-10 and IL-13) nature of the secreted cytokines, the AVL + SGHAb + individuals were more likely to show a mixed response while the AVL -SGHAb+ group was more polarized towards either a proinflammatory or anti-inflammatory response ( Table 2). Analyzing the ratio of IFN-γ to IL-10 or IL-13 indicates that the response is overall biased towards an anti-inflammatory profile (S1 Table). Independent of the four analyzed subgroups, 62% (51/82) of deployed individuals developed a cellular response to SGH, of which 50% (26/51) were AVL + ( Table 2). Individual responses as shown for Table 2 are masked by the group mean.
Having defined the cytokine profile after stimulation with Ph. alexandri SGH, we speculated whether it may be affected by, or may affect, the response to Leishmania. To address this, we analyzed the overall response to the SGH in the context of infection with L. infantum based on the historical cellular response of PBMC stimulated with soluble Leishmania infantum antigen (SLA) as demonstrated in a previous study [5]. Twenty of the 26 (76.9%) AVL + individuals with a cellular response to SGH and 6/11(54.5%) AVL + but no cellular SGH response were also Leishmania infantum IFN-γ release assay-positive (IGRA + ).

Risk factors for sand fly exposure
Risk characteristics for sand fly exposure were assessed in the AVL-SGHAb + , AVL + SGHAb -, AVL + SGHAb + and AVL -SGHAbgroups (Table 3). A statistical difference in terms of increased mean Ph. alexandri total IgG antibody levels was found between these groups, (p < .0001) ( Table 3). No statistical association was found comparing the use, the frequency of use, of an insect repellent, or permethrin-treated uniform usage with either AVL or SGHAb status (Table 3). Though not statistically significant, the AVL -SGHAbgroup appears to have slept fewer nights in open environment without shelter (9.2 nights) compared to all other groups (34.9-54.1 nights), suggesting that spending less time outdoors had the most protective effect against sand fly bites and therefore the risk of contracting VL.

Discussion
The sand fly bite is a critical event in Leishmania transmission and vector saliva is a determining factor for Leishmania infection. The large array of pharmacological substances in sand fly  saliva is essential for efficient blood feeding by the sand fly, promoting blood flow to the bite site and modulating the local immune response to the injected molecules, thus interfering with both the host hemostatic and immune responses [14,26]. Sand fly saliva has immunomodulatory properties due to antigenic components, mainly proteins [10]. The composition of sand fly saliva differs between vector species, but sometimes differences can also be detected among similar species from distinct geographical areas [27]. In this study, we analyzed the humoral and cellular immune responses directed against the saliva of wild-caught Ph. alexandri, a vector of VL, in persons deployed to L. infantum endemic regions in Iraq. We investigated how natural exposure to Ph. alexandri shapes the human immune response to salivary gland antigens and whether such immune responses are associated with AVL. Despite the study limitation of not knowing when AVL+ individuals were exposed to a L. infantum-infected bite amid recurrent uninfected bites, we achieved our objective by analyzing the cellular response to wild caught Ph. alexandri saliva in well-defined groups (AVL -SGHAb + , AVL + SGHAb + , AVL + S-GHAb -, AVL -SGHAb -, and HC) and presented our findings in the context of self-reported risk factors for sand fly exposure. A seropositive response to Ph. alexandri SGH was detected in 64% of 200 Iraq-deployed individuals, mirroring earlier findings regarding the immunogenicity of saliva from other sand fly vectors in endemic populations, including Lu. intermedia (59%, mainly IgG1 and IgG4) [28], Ph. papatasi (40%) [29] and Lu. longipalpis [10]. The antibody response elicited by sand fly salivary proteins has been shown to be specific for several vector species [27] but can be found in common for closely related species that share high levels of similarity in their salivary proteins like Ph. duboscqi and Ph. papatasi [30]. Additionally, a high titer of anti-saliva antibodies has been associated with an increased risk of contracting leishmaniasis due to more frequent contact with sand flies, thus increasing the probability to encounter infected sand fly bites [15,19].
In endemic areas, the sand fly population fluctuates seasonally which may also influence host anti-saliva antibody or cellular levels [6,31]. Importantly, after a six-month or one-year bite-free period, boosting by re-exposure with Lu. longipalpis or Ph. argentipes bites, respectively, caused significant increases of antibody levels in humans indicating an antibody memory response to SGH for both sand fly species [27]. In our study, antibodies to SGH could be detected up to 18 months after last exposure in at least five individuals (among 128) including three with AVL. Those subjects had more than a one-time deployment suggesting that reexposure may be associated with longer persistence in antibody levels.
In terms of cellular responses to Ph. alexandri saliva, we demonstrated that a high amount of IL-10 was produced following SGH stimulation in the deployed groups, regardless of their antibody response to SGH. We hypothesize that some study subjects may have received a low number of sand fly bites that are insufficient to induce a robust adaptive immune response. In naïve individuals lacking an adaptive immune response to sand fly saliva, immunomodulatory salivary proteins have been shown to be mostly anti-inflammatory [31]. Although no data on the composition of Ph. alexandri saliva, and no 'omics' studies exist yet, we speculate that IL-10 secretion possibly results from the presence of adenosine and adenosine monophosphate in sand fly saliva [32,33]. Adenosine increases IL-10 and decreases IL-12 production [34]. Alternatively, individuals with a negative history to any type of Leishmania infection but that were exposed to Ph. papatasi or Lu. intermedia also developed an IL-10 dominant response [22,28], which inhibits the proliferation of lymphocytes producing IFN-γ. Following Ph. papatasi bites, CD8 + T cells were the primary source of IL-10 [22]. In contrast, the source of IL-10 after Lu. intermedia exposure was both CD4 + CD25 + Foxp3 + and Foxp3 − cells [28]. IL-10 has suppressive roles that could help limit the immune-mediated VL pathology, especially in the liver, and promote parasite replication and disease progression [35].
In addition to IL-10, both IL-6 and IL-13 were produced in response to Ph. alexandri SGH stimulation of PBMC from our volunteers who deployed to a VL endemic area nearly a decade ago. Interleukin-6 is a multifunctional cytokine produced by several cell types and has a systemic effect on the immune system (primarily on B lymphocytes). It also stimulates hematopoiesis [36]. We speculate that saliva of Ph. alexandri has an inhibitory effect on lymphocyte proliferation suppressing early production of IL-2, IL4, and IFN-γ and increasing production of IL-6, similar to what was reported for Lu. longipalpis and Ph. papatasi saliva [30,37]. Lu. longipalpis saliva induced an increase in IL-6, IL-8, and IL-12p40 production by PMBCs but decreased TNF-α and IL-10 production [38]. However, increased levels of IL-10 mRNA were observed in the ear tissues of mice co-injected with parasites and SGH of Lu. longipalpis that was associated with higher levels of IL-10 in supernatants of restimulated draining lymph node (LN) cells [39].
In terms of the cellular response profile to Ph. alexandri saliva in the context of AVL status, the AVL + SGHAb + showed the highest response, producing a mix of proinflammatory and antiinflammatory responses. In comparison, AVL -SGHAb + individuals were more polarized towards either a proinflammatory or anti-inflammatory response that may influence their infection outcome upon future exposure to a Leishmania-infected bite. The mixed or dichotomous response of human PBMC upon stimulation with vector SGH was also observed in individuals chronically exposed to Ph. duboscqi saliva who were categorized as Th1, Th2, or mixed responders [23]. Additionally, individuals experimentally exposed to Lu. longipalpis bites produced systemic IFN-γ, IL-10, and TNF-α [17]. The majority of the SGHAb + subgroup was also Leishmania IGRA + [5] which raises the possibility that immunity to saliva may produce an environment that helps the parasites survive or persist. Alternatively, another explanation could be that multiple bites over time are needed to develop SGH antibody responses (based on controlled human challenge studies) and that the SGHAb + group was selected for a higher degree of vector saliva exposure and potentially greater likelihood for encountering a Leishmania infected bite.
Collectively, our findings highlight differences in the bite-induced immune responses to saliva of wild Ph. alexandri in this American military cohort. We observed mixed Th1 and Th2-biased responses to saliva that may result in different individual outcomes following exposure to an infected bite. This observation is further complicated by the fact that we do not know the discrete timing/order of the uninfected sand fly exposure versus Leishmania inoculation in our subjects and also that we analyzed the systemic and not the cellular immune response at the skin level. A local immune analysis at the skin bite site may provide additional clues and could be more appropriate for associating saliva-specific immunity to protection after sand fly exposure. Moreover, expecting that Th1-biased or mixed immune responses to vector saliva can prevent infection is a high bar, it might have contributed to an appropriate immunologic response which prevented active VL. Future studies can better address the effect of varied immune responses to vector saliva on the outcome of VL in a larger cohort that includes symptomatic VL cases. Finally, our convenience samples came from banked sera collected close in time to end of deployment (since humoral responses to SGH wane) while PBMC were collected on average about a decade later. We assume that persistent cellular memory immune responses against salivary homogenate will be the most impactful in preventing VL activation.
It was shown that mounting a humoral immune response to Lu. longipalpis saliva paralleled the development of a cell-mediated immune response to Leishmania [6]. In our study, no such correlation was observed. The absence of correlation between proliferative and humoral response to Ph. papatasi saliva has also been reported [11]. On the other hand, our results suggest that more intense or longer exposure to bites amplify the chances of getting an infected bite thus favoring AVL development.
Taken together, our study highlights the complex nature of the human immune response to sand fly saliva and its influence on the resulting immunity against Leishmania. Future studies should further explore the relevant antigens in Ph. alexandri saliva and their contribution to immunogenicity and susceptibility to, or protection from, Leishmania infection.