Seroconversion to Lutzomyia intermedia LinB-13 as a biomarker for developing cutaneous leishmaniasis

Sand flies inject saliva while feeding in the vertebrate host and anti-saliva antibodies can be used as biomarkers of exposure to Leishmania vectors. We expressed recombinant salivary proteins from Lutzomyia intermedia, a vector of Leishmania braziliensis, and evaluated the seroreactivity in exposed individuals in search for exposure markers. We found a strong correlation among positive serology to recombinant proteins LinB-13, 26, 15, 21 and to salivary proteins: rLinB-13 was the top performing molecule; IgG4 was the most predominant antibody subclass and antibodies to rLinB-13 did not cross react with Lu. longipalpis salivary proteins. By evaluating a cohort of contacts of CL patients, we confirmed that rLinB-13, an antigen 5-related protein, is a marker of exposure to Lu. intermedia with high degree of accuracy. In a 5-year follow up, we determined that individuals who developed CL presented higher anti-rLinB13 IgG responses, before the appearance of clinical symptoms. They also presented a lower frequency of cellular responses to the parasite (DTH). Our results show that seroconversion to a salivary molecule, rLinB-13, is a marker of risk for CL development caused by Leishmania braziliensis. This highlight the possibility of developing tools based on vector molecules to manage the disease in endemic areas.

In the Americas, cutaneous leishmaniasis (CL) is caused mainly by Leishmania braziliensis, a parasitic protozoan transmitted by Lutzomyia intermedia sand flies. During Leishmania sp. transmission, the sand fly inoculates the vertebrate host with both Leishmania parasites and with salivary molecules. These molecules aid in blood feeding and some of these salivary proteins are immunogenic (rev. in ref. 1).
The humoral and cellular immune response to sand fly salivary proteins in animals including humans has been well documented (rev. in refs 2, 3) and the idea or concept of using sand fly salivary proteins as markers of vector exposure has been reported and used successfully for Lutzomyia longipalpis [4][5][6] and Phlebotomus papatasi [7][8][9][10] . Therefore, the antibody response to salivary antigens can be used to probe for exposure to other sand flies and, hence, to assess the possibility of developing leishmaniasis associated to that sand fly species. However, the use of salivary gland sonicate (SGS) in large-scale epidemiologic assays is challenging due the difficulty inherent to sand fly capture or rearing, both necessary for salivary gland antigen preparation. Furthermore, SGS is a combination of various antigens, limiting therefore the purity or specificity of the antigen. To overcome this obstacle, transcriptomics and proteomics approaches have been effectively used in identifying and characterizing a number of sand fly salivary proteins (rev. in ref. 11). Sand fly salivary transcripomic approaches have been followed by the production of recombinant sand fly salivary proteins in heterologous systems and the use of these recombinant proteins for serologic testing. Sand fly recombinant salivary proteins LJM17 and LJM11 have been validated in the field as markers of exposure to Lu. longipalpis in Brazil 6,12 and the recombinant salivary protein PpSP32 has been successfully used as a marker for Ph. papatasi exposure in Tunisia and Saudi Arabia 13,14 . These works underline the feasibility of using sand fly salivary recombinant proteins as epidemiological tools for vector exposure.
In addition to using anti-saliva antibodies to probe for sand fly exposure in animal reservoirs and humans, the anti-saliva immune response has also been associated with the risk of developing leishmaniasis upon infection with an infected sand fly 9,15 . In Brazil, we observed that patients with CL caused by Leishmania braziliensis presented a significantly higher IgG response to Lu. intermedia SGS compared to individuals with subclinical L. braziliensis infection 8 and natural exposure to Lu. intermedia bites indeed increases the risk of developing CL caused by L. braziliensis 16 .
Based on the availability of the Lu. intermedia salivary gland transcriptome 17 , we expressed the most abundant salivary proteins from this sand fly species and searched for antigens that could act as exposure markers to Lu. intermedia in an area where CL is caused by L. braziliensis. Besides validating the top-performing antigen in a large cohort of household contacts, we also performed a 5-year follow up study of exposed individuals. We probed for serology to rLinB-13 and CL development and we show that individuals seropositive to a defined Lu. intermedia salivary recombinant protein are at increased risk of developing disease.
LinB-13 is a 28 kDa salivary protein, member of the family of antigen 5-related proteins, with ubiquitous distribution in sand flies [rev. in ref . 11]. Although its function is still unknown, it belongs to the cysteine rich secretory proteins, present in the saliva of most blood-sucking insects and in hookworms 18 Fig. 1A). Phylogenetic analysis confirmed the clustering of LinB-13 and its homologues in Lu. longipalpis and Lu. ayacuchensis sand flies confirming the similarity between these molecules (Supplemental Fig. 1B). Validation of rLinB-13 as an exposure marker in a L. braziliensis transmission area. Next, we validated rLinB-13 as marker of exposure to Lu. intermedia by, evaluating the IgG response against this protein and against Lu. intermedia SGS. Among the household contacts (n = 264) of CL patients tested 136 (51.5%) had positive serology to rLinB-13 and 150 individuals (56.8%) had positive serology to SGS, indicating a similar degree of reactivity (Fig. 3A). Dot Blot analysis confirmed that rLinB-13 and Lu. intermedia SGS were equally recognized by individuals seropositive for both SGS and rLinB-13 but not by seronegative subjects (Fig. 3B). We also did not observe significant differences among seropositive or seronegative individuals to rLinB-13 regarding epidemiological characteristics such as age, gender, occupation, etc, (Supplemental Table 2). One exception was that subjects seropositive to rLinB-13 arrived more frequently at home after 4 pm. Furthermore, we found a strong correlation between IgG to rLinB-13 and to Lu. intermedia SGS (P < 0.0001, r = 0.74) (Fig. 3C) and ROC analysis confirmed, in this larger cohort, that antibodies against rLinB-13 discriminate a positive response to SGS (AUC = 0.86, P < 0.0001) (Fig. 3D). Therefore, rLinB-13 can predict exposure to Lu. intermedia with 90% sensitivity and 74.5% specificity (cut-off value of 0.018) (Fig. 3D). With regards to IgG subclasses, IgG4 was the most predominant isotype followed by IgG3 (Fig. 4A) and a significant correlation was found between anti-LinB-13 total IgG and anti-LinB-13 IgG4 (P < 0.001, r = 0.47) (Fig. 4B).
To evaluate cross-reactivity between anti-LinB-13 antibodies and IgG to salivary antigens from a related sand fly, we probed sera from individuals (n = 20) exposed to Lu. longipalpis against rLinB-13 and against Lu. intermedia SGS. Seventeen individuals (85%) presented positive serology to Lu. longipalpis SGS and one individual was seropositive to Lu. intermedia SGS but none of the sera reacted against rLinB-13 (Fig. 5A). On the contrary, 20 (100%) control sera (obtained from a CL endemic area) reacted against both Lu. intermedia SGS and rLinB-13 whereas two sera (10%) reacted against Lu. longipalpis SGS (Fig. 5A). Hierarchical clustering confirmed that the seroreactivity pattern of individuals from a CL endemic area is similar for both Lu. intermedia SGS and rLinB-13 whereas it is clearly distinct from that observed for individuals from a VL endemic area (Fig. 5B). These results highlight the potential of rLinB-13 to act as a marker of vector exposure in an area of L. braziliensis transmission.
Antibodies to rLinB-13 as a risk marker of CL development. We previously showed that individuals with CL caused by L. braziliensis have a higher IgG response to Lu. intermedia SGS 8 and this increases the risk of developing CL 16 . However, specific salivary molecule(s) within SGS that may play a role in this outcome remained unknown and we hypothesized that LinB-13 was a likely candidate based on its performance in predicting exposure to Lu. intermedia. We performed a 5-year follow up study in which 264 household contacts of CL patients were evaluated for presence of IgG to LinB-13. These evaluation were performed at the beginning of the study (Baseline) and at the 2 nd and 4 th years of follow-up. In parallel, individuals were evaluated yearly for the appearance of clinical signs of CL. At the time of enrollment in the study, we observed that 136/264 (51.5%) individuals developed anti-LinB13 antibodies whereas 128/264 (48.5%) did not (Fig. 6A). During the 5-year follow up, 38/264 (14.3%) subjects developed CL and, among these, 28/38 (73.7%) were seropositive to rLinB-13 whereas 10/38 (26.3%) were seronegative to rLinB-13 (Fig. 6A). In fact, the subjects who developed CL displayed a higher (p < 0.0001) anti-LinB-13 IgG response compared to those who remained free of disease. Also, the elevated anti-LinB-13 IgG response was detected before the development signs of the disease (Fig. 6B). Thus, individuals seropositive to LinB-13 are 2.6 times more likely to develop CL caused by L. braziliensis (RR = 2.6; 95% confidence interval = 1.3-5.2; P < 0.01) compared to seronegative subjects. Of note, seropositivity against LinB-13 did not change over time: individuals who presented IgG to LinB-13 at the time of enrollment in the study (Baseline) continued to display such response during the follow up period (Supplemental Fig. 2).
Presence of a delayed type hypersensitivity (DTH) reaction to Leishmania antigen is associated with protection against infection 19 . Among those individuals who seroconverted to rLinB-13, we detected a lower frequency (p < 0.05) of DTH + subjects (Fig. 7A). A positive DTH is a marker of subclinical infection in areas of L. braziliensis transmission 20 but we observed that DTH + individuals displayed lower levels (p < 0.01) of anti-rLinB-13 IgG (Fig. 7B). Therefore, we show that development of antibodies to rLinB-13 is not concurrent with development of a cellular response to Leishmania.

Discussion
Antibody responses to antigens present in arthropod saliva have been explored as vector exposure monitoring tool [21][22][23] , including vectors of leishmaniasis 4,5,7,24 . This task has been enormously facilitated by the characterization and production of recombinant salivary proteins. In the present work we searched for immunodominant  antigens in Lu. intermedia saliva using a panel of sand fly recombinant salivary proteins. We further validated the top-performing molecule in household contacts of CL patients and determined whether exposure to rLinB-13 could be used as a risk assessment tool.
Following the initial screening to identify the immunodominant molecules within Lu. intermedia salivary proteins, we found a strong correlation among anti-LinB-13 and anti-Lu. intermedia SGS antibodies, a result corroborated by ROC curve analysis. Similarly, antibody responses to LJM11 and LJM17 12 also correlated with reactivity to Lu. longipalpis SGS, albeit with an inferior performance compared to rLinB-13. Besides LJM11 and LJM17 present in Lu. longipalpis, PpSP32 present in Ph. papatasi SP32 13-15 , SP03B present in Ph. perniciosus 25,26 and rPorSP24 present in Ph. orientalis (rPorSP24) 27 were identified as immunodominant salivary antigens and validated as markers of vector exposure in different endemic areas. In this work we identified rLinB-13 as a marker of Lu. intermedia exposure, a salivary antigen different in sequence and size to any other marker of exposure described so far. rLinB-13 was validated as a marker of sand fly exposure in a large cohort of individuals residing in a L. braziliensis transmission area.
As seen with Lu. intermedia SGS 16 and in a study conducted with Ph. papatasi-exposed individuals in Egypt and Jordan, authors also found IgG4 to be the dominant subclass 28 . In the later, there was a positive correlation between concentrations of IgG4 and IgE, in agreement with reports showing concurrent IgE and IgG4 responses following exposure to Aedes sp. 29,30 . Chronic exposure to Aedes sp. bites induces higher IgG4 antibodies against salivary proteins compared to other IgG subclasses 30 and it has been suggested that IgG4 is a marker of intense exposure to arthropod vectors 31 . B-cell class switching to IgG4 and IgE is induced by IL-13 32 but IL-10, in the presence of IL-4, can also drive IgG4 class switch in the absence of IgE 33 . Herein we have not addressed the presence of anti-LinB13 IgE but both IL-10 and IL-13 were detected upon stimulation of peripheral blood cells from Lu. intermedia exposed individuals with SGS 16 , suggesting a mechanistic explanation for the presence of IgG4 antibodies in Lu. intermedia-exposed individuals.
Besides the possibility of using sand fly salivary proteins to measure vector exposure, these molecules have also been employed to access the risk of disease development. In Tunisia, individuals who developed CL caused by L. major displayed significantly higher antibodies levels against Ph. papatasi saliva compared to subjects that remained disease-free 9 . In Turkey, patients with L. tropica lesions also showed higher levels of IgG against Ph. sergenti saliva compared to control individuals 7 . We corroborated these findings in Brazil: patients with CL caused by L. braziliensis presented a higher IgG response to Lu. intermedia SGS compared to individuals with subclinical L. braziliensis infection 8 , indicated by a positive DTH to Leishmania, in the absence of clinical signs. Indeed, higher serology to Lu. intermedia SGS increases the risk of developing CL 16 . These studies were performed using SGS and the specific salivary molecule or molecules associated with such outcome remained unknown. By following up individuals living in a CL endemic area and constantly exposed to Lu. intermedia we detected seroconversion to rLinB-13 before the appearance of clinical signs of CL. Importantly, positive serology to rLinb-13 was higher individuals who later developed CL, showing that seroconversion to rLinB-13 increases the risk of developing disease. This increased risk of disease was paralleled by a significantly lower frequency of a positive DTH response to Leishmania in seropositive subjects, suggesting that exposure to Lu. intermedia LinB-13 may negatively impact a cellular responses to the parasite. We previously showed that individuals exposed to Lu. intermedia present a higher frequency of CD4 + CD25 + IL-10-secreting cells and, in vitro, such response enabled parasite replication 16 . We speculate that LinB-13 may induce a similar response thereby promoting a parasite-friendly environment at the moment of transmission. However, we cannot rule out the hypothesis that a higher antibody response to LinB-13 reflects an increased exposure to Lu. intermedia and, in consequence, and increased probability of contact with infected sand flies and increased risk of developing CL. Since we do not know when infection occurred in subjects who presented CL, the latter is difficult to be addressed in human studies.
Experimental studies addressing how exposure to sand fly saliva could promote protection from leishmaniasis 34-37 enabled the formulation of the following hypothesis: exposure to sand fly salivary proteins primes the development of CD4 + IFN-γ-secreting T cells that, upon the bite of an infected sand fly, are rapidly mobilized to the inoculation site leading to macrophage activation and parasite killing (rev. in ref. 2). Our results, however, suggest that this response may not occur as such in areas of L. braziliensis transmission by Lu. intermedia, especially given our finding that a positive serology to LinB-13 is dissociated from DTH conversion to Leishmania. In Saudi Arabia 15 and in Tunisia 38 , higher serology to salivary antigens has been reported in CL patients also suggesting a possible detrimental effect of exposure to sand fly saliva on human cutaneous leishmaniasis.
Sand fly salivary proteins are useful and practical tool to understand the dynamics of leishmaniasis transmission. Herein, we identified a single salivary molecule in Lu. intermedia saliva, LinB-13, that accurately predicts exposure to sand flies in a CL endemic area. Moreover, antibodies to rLinB-13 are a biomarker of risk of developing leishmaniasis caused by L. braziliensis. Host factors have been associated with susceptibility to CL [39][40][41] and our study now identifies a sand fly salivary component that may also be decisive in the outcome of disease. We thus envisage a translational approach in which clusters of individuals with high titers of LinB-13 could be targeted for Insecticide Residual Spraying interventions, especially in areas where L. braziliensis transmission is peridomiciliar.

Study design and selection of individuals.
This study was conducted in Corte de Pedra, municipality of Tancredo Neves, BA, Brazil, an area of L. braziliensis transmission by Lu. intermedia and it consisted of three phases (Supplemental Fig. 3). In Phase 1, we screened Lu. intermedia recombinant salivary proteins (n = 11) using sera (n = 89) from residents of the endemic area. In Phase 2, we validated the top-performing Lu. intermedia salivary molecule, as indicated by ROC analysis, in a larger cohort of household contacts of CL patients (n = 264). We also evaluated cross-reactivity of LinB-13 against sera from individuals residing in a VL endemic area (n = 20) (Vila Nova and Bom Viver), municipality of Raposa, São Luis, MA, Brazil 24 . In Phase 3, we evaluated whether exposure to LinB-13 is a risk marker for CL development. We tested sera from the cohort of household contacts of CL patients (n = 264). The cohort was established in 2010 and followed up to January 2015 42 . Inclusion criteria consisted of a negative history of any type of Leishmania infection, established after a medical interview and physical examination for signs consistent with active or previous CL or ML. At the moment of enrollment, blood was collected for serological assays against LinB-13. At the same time, the Leishmania Skin Test (DTH) was performed, as described below. The serological surveys were repeated at the 2 nd and 4 th year of follow up. Among the 264 household contacts, 237 were evaluated at the 2 nd year and 224 were evaluated at the 4 th year of follow up. The individuals not evaluated either developed CL or refused to participate. During follow-up, one visit was performed annually for active monitoring of CL development and subjects suspected of having CL were diagnosed by parasite isolation or by a positive PCR for L. braziliensis. This research was conducted with the approval of the Ethical Committee of the Hospital Prof. Edgard Santos (Salvador, Bahia, Brazil; 240/2009) and Comissão Nacional de Ética em Pesquisa (CEP, Brazilian National Ethics Committee, Brazil). Informed consent was obtained from each participant. All methods were performed in accordance with the guidelines and regulations determined by CEP.
Leishmania Skin Test. The Leishmania Skin Test was performed at the moment of enrollment in the study and at the 2 nd and 4 th years of follow up. For skin test, 0,1 mL (15 μg/mL) of Soluble Leishmania Antigen, prepared as described elsewhere 43 , was injected intracutaneously on the volar surface of the forearm, and the greater diameter of induration was measured 48-72 h later. Induration of ≥ 5 mm was defined as a positive reaction.
Sand flies and preparation of SGS. Adult Lu. intermedia or Lu. longipalpis sand flies were captured in Corte de Pedra and Cavunge, BA, Brazil, respectively. Sand fly identification and Salivary Gland Sonicate (SGS) preparation were performed as described elsewhere 44 . Cloning and expression of Lu. intermedia salivary transcripts. Eleven Lu. intermedia cDNA coding for secreted proteins (LinB-8, LinB-13, LinB-14, LinB-15, LinB-17, LinB-21, LinB-26, LinB-28, LinB-37, LinB-38, LinB-44) ( Table 1) 17 were amplified by PCR, cloned into VR2010-TOPO and sequenced as described 45 . Recombinant proteins were produced by transfection of 293-F cells (Invitrogen) with VR2010-TOPO plasmids coding for different Lu. intermedia salivary proteins by the Protein Expression Laboratory at the Frederick National Laboratory for Cancer Research (Frederick, MD). High-performance liquid chromatography purification of His-tagged Lu. intermedia salivary proteins was performed as described 6 using an NGC Chromatography System (BioRad). Eluted proteins were detected at 280 nm and collected every minute on a 96-well microtiter plate using a BioFrac Fraction Collector (BioRad). Recombinant proteins were tested for endotoxin using a QCL-1000 LAL assay (Lonza). Endotoxin-free recombinant proteins were used in all assays. Aliquots of eluted proteins were checked by SDS-PAGE developed by silver staining. Edman sequencing of the N-terminus region of the purified recombinant proteins was performed at the Protein Chemistry Section, Research Technology Branch, NIAID, NIH (Rockville, MD). Proteins were stored at −70 °C until use.

Analysis of anti-Lu. intermedia SGS antibodies and anti-Lu. intermedia salivary proteins antibodies.
Humoral (IgG and IgG subclasses) response to Lu. intermedia SGS and to recombinant salivary proteins was determined as described 16,46 . Recombinant proteins were used at 1 µg/ml. In all ELISA assays, the cut-off value was established employing sera from healthy volunteers (n = 30) from a non-endemic area and was determined as the mean optical density (OD) value plus 2.5 standard deviations. Dot Blot Analysis. Dot blot analysis was performed by spotting 1 µg of rLinB-13 or Lu. intermedia SGS onto nitrocellulose membrane (Invitrogen). Membranes were blocked with 5% BSA in PBS-T (0.1% Tween 20) for two hours. Serum samples (1/100) from individuals (n = 5) with positive serology for both LinB-13 and Lu. intermedia SGS or from non-exposed subjects (n = 4) were diluted in PBS-T with 0.1% BSA and incubated with spotted membranes for 45 minutes. Following another round of washing, anti-human HRP-conjugated IgG (Sigma) was added (1/20000) and membranes were incubated for 30 minutes. Signal detection was carried out using ECL system (Amersham) on an Image analyzer (ImageQuant Las 4000).

Statistical analysis.
Comparisons between 2 groups were performed by Mann-Whitney test and among 3 or more groups by Kruskal-Wallis test followed by Dunn multiple comparison tests. Categorical data were compared using the Fisher exact test. The matrix based on non-parametric Spearman correlation was performed using R version 3.2.1 (R Foundation for Statistical Computing, Vienna, Austria). ROC analysis was used to establish sensitivity and specificity of recombinant proteins in predicting anti-Lu. intermedia SGS positivity. Correlation between IgG to LinB-13 and Lu. intermedia SGS or between IgG and IgG4 to LinB-13 were checked using non-parametric Spearman test. Hierarchical cluster analysis of Z-score normalized serology response to Lu. intermedia SGS, Lu. longipalpis SGS and rLinB-13 using the Ward's method with bootstrap (100x) was performed using JMP Statistical Discovery (V. 12). Statistical analyses were conducted using Prism (V. 5.0) (GraphPad Software) and differences were considered significant when P < 0.05. The relative risk (RR) was calculated using the following formula: RR = I 1 /I 0 , where I 1 is the incidence of CL in exposed individuals (seropositive to LinB-13 salivary protein), and I 0 the incidence of CL in non-exposed individuals (seronegative to LinB-13 salivary protein).