Background
Visceral Leishmaniasis (VL) is a sand fly-borne disease caused by infection with protozoan parasites of the
Leishmania donovani complex. The vast majority of the 300,000 estimated annual cases are reported from focal regions in Ethiopia, Sudan, South Sudan, India, Bangladesh and Brazil, though the disease is also endemic in the Mediterranean basin [
1]. VL incidence is highest in the Indian sub- continent, followed by East Africa, where the causative organism is
L. donovani and transmission is anthroponotic. In Brazil and the Mediterranean basin, the disease is caused by
L. infantum, while transmission is zoonotic, with dogs serving as an intermediary host [
1]. VL incidence has also been reported in previously non- endemic regions owing to travel to and migration from endemic regions and environmental conditions that have expanded the habitat of the sand fly [
2‐
4]. Current diagnosis of VL is based on presentation of clinical symptoms such as fever, splenomegaly and weight loss, then confirmed by parasite detection in bone marrow/splenic biopsy in Africa and South America; or detection of antibodies against the rK39 antigen in the Indian subcontinent [
5].
Multiple drug regimens are available to treat VL. Antimoniates including sodium stibogluconate (SSG) and meglumine antimoniate (MEG) are the first line drugs in Brazil and Africa, while resistance to antimoniates have phased them out in the Indian sub- continent, where amphotericin B, paromomycin and miltefosine are the drugs of choice [
6,
7]. Liposomal amphotericin B (AmBisome™) is now preferred in Europe and the Indian sub- continent [
8].
Timely diagnosis and treatment are imperative as, without treatment, VL can be fatal [
1,
6]. Drug regimens are expensive and can cause severe side effects [
5,
6,
9]. These factors have contributed to low compliance in many VL endemic regions, leading to unresponsiveness and the emergence of drug- resistant parasite strains [
7,
9]. It is important to monitor treatment and detect unresponsiveness as early as possible. Treatment failure ranges from 3–10 % in immune competent individuals and 50–60 % in immune compromised individuals [
4,
10]. Though simple tests are available to accurately confirm VL disease, none of the current diagnostic tests is suitable to monitor treatment or cure [
11]. Commonly used diagnostic tests such as DAT and the rK39 rapid detection test (RDT) cannot differentiate between past and current infections because the antibodies detected persist after clinical cure is achieved [
12]. Microscopy of splenic or bone marrow biopsies, though confirmatory for parasite clearance is not a practical tool to monitor treatment due to the painful and invasive nature of sampling.
A non-invasive, standardized test for monitoring treatment success in a clinical setting would complement tools to confirm VL, and aid in effectively managing VL. Such a test should ideally detect parasite or parasite products as a measure of infection since presence of parasite products should theoretically correlate with parasite burden. Hence, a drop in antigen levels as measured by the test would reflect a decrease in parasite burden due to anti-Leishmania treatment and eventual clearance of parasites. It must also be sensitive, specific, easy to use, quantitative and preferably non- invasive for repeated sample collection.
At present, KAtex is the only commercially available
Leishmania antigen detection test [
13]. Although highly specific, KAtex’s sensitivity has been variable, limiting its widespread use for the assessment of treatment [
14‐
17]. Knowing that
Leishmania antigens are excreted in the urine of VL patients, we developed a sensitive urine-based test to detect antigens with which to evaluate treatment [
18,
19]. We compared its performance to a similar product developed by Kalon Biological Ltd., UK. We discuss the validation of the antigen detection tests and their evaluation for determining VL cure post- treatment.
Methods
Samples
Urine samples of VL patients were collected as part of routine diagnosis and treatment. Except the post- treatment samples, all VL patient samples were collected at diagnosis prior to treatment start. Samples were collected at Gedaref Hospital, Sudan the Rajshahi Medical College Hospital, Bangladesh and at the clinic in Sergipe, Aracaju, Brazil. Urine samples from Ethiopia were collected in Southern Ethiopia during ongoing field studies. Each of the Ethics Committees of Khartoum University, Rajshahi Medical College, University of Sergipe and Addis Ababa University approved study protocols, respectively. Written or verbal informed consent was obtained from patients at the time of collection. Inclusion criteria for VL patients in Ethiopia, Sudan and Brazil were presentation of clinical symptoms and demonstration of parasites in spleen, bone marrow, or lymph node smear or positive rK39 for Bangladesh. Urine samples from patients with other diseases (OD) were kindly provided by FIND, Geneva and consisted of 10 each from patients with human African trypanosomiasis (HAT) and P. falciparum malaria from Uganda, and 10 from TB patients in Thailand. For all samples provided by FIND, written, informed consent was obtained at the time of collection. Non-endemic control (NEC) samples (n = 49) were collected from local volunteers in Seattle, USA or purchased from Equitech-Bio, Inc. (Kerrville, TX). Urine samples from 10 healthy endemic controls (EC) were also obtained from Bangladesh as evidenced by lack of symptoms and a negative rK39-RDT. All urine samples were stored frozen and transported on dry- ice to minimize any adverse effects of freeze- thawing.
Generation of anti-Leishmania antibodies
L. donovani (MHOM/SD/00/1S-2D) promastigotes were seeded in culture flasks and cultured at 25 °C in M199 medium (Sigma, St. Louis, MO) supplemented with 10 % FCS (Hyclone), 1XM199 Hanks salt, 1XMEM amino acids Solution (Invitrogen), 10 mM MEM non-essential amino acids (Sigma), 40 mM HEPES pH 7.4, 0.1 mM adenine, 5 μg/mL hemin, 1.5 uM biotin, 100 U/mL penicillin, 100 μg/mL streptomycin, 2 mM L-glutamine, 0.35 mg/mL sodium biocarbonate at final pH 7.2 for 10 days. Stationary phase promastigotes at a density of 2-4×107 parasites/mL were harvested and washed three times with cold PBS and frozen at −80 °C. To prepare lysates, the pelleted parasites were and resuspended in 10 mM Tris–HCl, 1 mM EDTA (pH 8.0) containing 1X Halt protease inhibitor (ThermoScientific) at 1×109 parasites/mL. Whole cell lysate (WCL) was prepared by freeze thawing the pellet in liquid nitrogen (6X) followed by three rounds of sonication for 30 s at 10 Hz. Soluble lysate antigen (SLA) was obtained by further centrifuging at 15,000 rpm for 45 min and discarding the insoluble pellet.
To generate antibodies, New Zealand white rabbits (NZW) with low residual reactivity to WCL were selected and immunized with 0.5 mg WCL added complete Freund’s adjuvant (Sigma) followed by 3 booster immunizations with 0.25 mg WCL with incomplete Freund’s adjuvant (Sigma) at 3 week intervals. Blood was collected 2 weeks after the final boost (R & R Research, LLC., Stanwood, WA).
Affinity purification and labeling
Total IgG was purified from the anti-sera of three rabbits with high IgG titers using Protein G Sepharose. Total IgG was further affinity purified against SLA coupled with Cyanogen bromide (CNBr)-activated-Sepharose™ 4B (GE Healthcare). Affinity purified antibodies were conjugated with horseradish peroxidase (HRP labeling kit, Thermofisher) at a 6:1 molar ratio using sodium periodate (NaIO4) and sodium cyanoborohydride (NaBH4CN) for oxidation and reduction, respectively. Protein concentration was determined using Bradford’s method and 0.77 volume of glycerol was added to the conjugated antibodies before storage at −20 °C.
Capture ELISA optimization
ELISA conditions were optimized in the context of ELISA plate selection, capture and detection antibody concentrations, urine sample dilution as well as sample incubation duration. In brief, Immulon™ 2HB plates were coated with 1 μg/mL affinity purified antibodies in 0.1 M sodium bicarbonate- carbonate buffer (pH9.6) at 4 °C overnight. After blocking with 1 % BSA-PBS/PBS-T, 50 uL of 1:1 diluted urine samples were added and allowed to incubate at room temperature for 1 h on a shaker. After 5 washes with 1X PBS/PBS-T, 100 μL of HRP-labeled anti-SLA IgG was added to each well at 1:1000 dilution and the plate was incubated at room temperature for 1 h. After five washes, 100 μL of SureBlue™TMB peroxidase substrate (KPL, Inc.) was added and incubated for 5 min before the reaction was stopped with 50 μL of 1 N H2SO4. Optical density was read at 450 nm (VersaMax microplate reader, Molecular devices) immediately. Affinity purified and labeled antibodies were then transferred to InBios International Inc., Seattle for developing the Leish Antigen Detect™ ELISA.
Leishmania Antigen Detect™ ELISA
ELISA was performed according to the manufacturer’s instructions (InBios International Inc. Seattle, WA). In brief, plate controls (positive and negative) and test samples were diluted with ELISA dilution buffer at 1:1 ratio and 50 μl added to duplicate wells for incubation at 37 °C for 30 min. After 6 washes, 50 μl of HRP-Leishmania ELISA conjugate was added to each well and allowed to incubate at 37 °C for 30 min. After washing, 75 μl of TMB solution was added to each well for 10 min. Reactions were terminated by adding 50 μl of stopping solution, and plates were read immediately at 450 nm. Each plate was assessed passed based on the discrimination between the signals observed for the in-built positive and negative controls, with criterion being a Positive control/ negative control >5. For determining the specificity of the ELISA, ROC curves were generated as described below. For determining positivity of samples, cut- offs were generated based on a panel of NEC samples run in each plate. Samples with signals over the cut- off were deemed positive.
KAtex
KAtex agglutination tests were conducted following manufacturer’s instructions (Kalon Biological Ltd., Guildford, UK), with minor modifications. In brief, 200 μl of freshly thawed urine was transferred into a collection tube, submerged in boiling water for five minutes and then cooled to room temperature. 50 μl of the boiled urine was added to the reaction zone on the glass slide and mixed with one drop of well-mixed latex. The liquids were mixed with a toothpick and spread to cover the entire surface of the reaction zone. The glass slide was tilted with a rotating action, and the degree of agglutination interpreted as instructed.
Leishmania Antigen ELISA
Leishmania antigen ELISA produced by Kalon Biologicals Ltd., UK was provided by FIND, Geneva, Switzerland. Polyclonal antibodies were prepared from antiserum of sheep immunized with 4×109 cultured promastigotes from Leishmania donovani strains from Sudan (LV9), Nepal (BPK282). Antisera were purified using an antigen affinity column prepared using concentrated BPK 282 spent cell culture. Kits were used according to the manufacturer’s instructions. Briefly, samples were diluted 1:20 in assay diluent, after which 100 ul/well of these and antigen calibrators were added to 96-well plates in triplicate or duplicate, respectively. Plates was incubated for 30 min at 37 °C, then washed 5 times before the addition of 100 uL of 1X Tracer. After 30 min of incubation at 37 °C plates were washed and 100 ul/well TMB Substrate solution was added for 30 min. Reactions were terminated by adding 100 ul/well of Stop solution. Plates were then read immediately at 450 and 620 nm (VersaMax microplate reader, Molecular Devices). Resulting OD was obtained by subtraction of OD at 620 nm from OD at 450 nm. Urinary antigen unit (UAU)/mL of samples were extrapolated from a four parameter logistic standard curve constructed in Excel using the mean values obtained for the calibrators. Samples with UAU less than the lowest calibrator (2 UAU/mL) were considered negative.
Calculations and statistical analysis
GraphPad Prism six software was used for generating Receiver-operating characteristic (ROC) curves and statistical analysis. ROC curves were generated with non-VL (NEC, OD) and VL samples to determine thresholds that afforded the best specificity. Sensitivity was calculated as the percentage of VL samples correctly assayed as positive while specificity was calculated as the percentage of non-VL samples (NEC, OD) correctly assayed as negative. Area under the curve (AUC) used to assess the accuracy of each test. One-way ANOVA (Dunnett’s multiple comparison) was performed between VL and non- VL samples with p values ≤0.05 were considered statistically significant. Cohen’s kappa values were used to determine the correlation between the ELISAs.
Discussion
Early diagnosis and efficacious treatment are the keys to VL management. VL treatment options are plagued by high costs, severe systemic side effects and unresponsiveness. Though simple and accurate diagnostic tests are widely available to confirm VL, they are not suitable to monitor treatment efficacy and cure. According to the WHO’s recommendations, tests for treatment efficacy and cure should be the “highest research priority” in VL control [
20]. At this time, microscopy and KAtex are the only available tests to follow treatment [
11]. Though attempts have been made to develop capture ELISAs to detect antigens in the urine of VL patients, they have not yet progressed to standardized tests [
21].
With the goal of developing a
Leishmania antigen detection test suitable for following treatment, we developed the
Leishmania Antigen Detect™ ELISA. As a tool for monitoring treatment, the
Leishmania Antigen Detect™ ELISA was sensitive, specific and suggested parasite clearance during clinical cure (Fig.
3). Thus, in terms of performance, the test represents a significant improvement over current options. The ELISA also has practical advantages over KAtex such as not requiring the boiling of urine, thus improving convenience of the assay. The profile of
Leishmania antigens in the urine of VL patients can be complex with not one dominating antigen detectable at all times [
18]. The
Leishmania Antigen Detect™ ELISA contains antibodies raised against a diverse panel of antigens derived from
L. donovani. The diversity of the antibody repertoire could be a reason for the high sensitivity of the
Leishmania Antigen Detect™ ELISA compared to KAtex and other previous capture ELISAs, both in- house and documented, which were developed to detect single antigens [
21,
22] (data not shown). The
Leishmania Antigen Detect™ ELISA also performed comparably to the
Leishmania Antigen ELISA, an antigen detection ELISA developed by Kalon Biological Ltd. as an improvement over their KAtex test. The Katex and Kalon VL ELISA are both based on polyclonal antibody response to whole promastigotes. However the
Leishmania Antigen ELISA utilizes affinity purified polyclonal antibodies based on an antigen similar to the whole cell lysate. The format for both ELISAs is similar, and so are their performances on urine samples from VL patients and controls (Table
1). However, the
Leishmania Antigen Detect™ ELISA was more specific for VL based on the OD samples tested in this study (although endemic controls were not tested) as well as more sensitive on VL samples from endemic regions in which
L. donovani is the causative agent, namely Sudan and Bangladesh.
Early detection of treatment non- responsiveness could help alter treatment options in time to prevent adverse outcomes such as mortality. It may be possible to refine and use the ELISAs as a means of identifying treatment non- responsiveness, if detection of antigen in urine can be considered a surrogate for parasite burden. A concern for VL patients infected with
L. donovani is the potential of developing PKDL after clinical cure [
23,
24]. As much as 50 % of treated VL patients in Sudan, South Sudan and Northern Ethiopia develop PKDL, 6 months after treatment [
24]. It is probable that antigen excretion in urine reflects parasite clearance by the drug, which occurs within 26 days for both antimoniates and Amphotericin B [
25‐
28]. None of the treated patients in this study relapsed or developed PKDL as of this time. Since these patients were from Southern Ethiopia where relapse and PKDL are rare compared to other regions of East Africa, it is not possible to draw any conclusions regarding the dynamics of antigen presence in urine and the risk for PKDL development. The testing of PKDL and Cutaneous Leishmaniasis urines would be useful to determine if both these two ELISAs are specific to Visceral Leishmaniasis.
Though originally developed as a tool to monitor treatment success, the high sensitivity of the ELISAs on VL patient samples from diverse endemic regions indicates that it also has potential as a tool for primary diagnosis. It could serve as an alternative in cases where currently used diagnostic tests are not as effective. Such a situation can arise in individuals with low antibody titers to diagnostic antigens like rK39, and immune suppression co- morbidities such as HIV [
14,
29]. Accurate diagnosis in such difficult scenarios can significantly aid case management. The
Leishmania Antigen ELISA (Kalon Biological Ltd.) was designed as a proof of cure and test for treatment failure. It was not developed to be a screening assay as not enough endemic controls or other endemic disease urines were used in the initial performance evaluation. The two ELISAs detect
Leishmania antigens but with different sensitivities based on their cut- off for positivity (Table
3). Further refinement of the cut- off for positivity will be undertaken in future based on data obtained from an expanded panel of samples.
The
Leishmania Antigen Detect™ ELISA and the
Leishmania Antigen ELISA could serve as a standardized tool to measure the effectiveness of emerging treatment regimens in clinical trials and help make policies on implementation of new drug regimens in endemic regions [
11]. Currently, splenic biopsy or PCR are used as a surrogate for parasite burdens in most trials involving new treatment regimens [
26,
30]. Splenic biopsy is invasive, and not amenable for repeated sampling. PCR is expensive and hard to standardize among laboratories. As a direct detection test that reflects parasite burdens, the ELISAs could be the alternative with non- invasive sampling, high sensitivity and a quantitative read- out being distinct advantages.
Our study provides compelling data for further refinement of the Leishmania Antigen Detect™ ELISA for deployment in endemic regions for multiple VL management purposes. Further assessments with samples from endemic regions reflecting co- infections with TB or HIV and different treatment regimens are needed to qualify the Leishmania Antigen Detect™ ELISA for clinical use, including diagnosis and test of cure. Development of lateral flow assay formats of both ELISAs, which would be more suitable for community use should also be envisioned.
Conclusions
The lack of standardized tools to monitor treatment hampers VL management in many ways. An effective tool that reflects parasite burden in VL: patients can help assess the suitability of the treatment regimen, foresee treatment failure and possibly predict post- treatment complications such as relapse. In this study, we have compared the suitability of two direct detection sandwich- ELISA based standardized tests, developed to detect Leishmania– specific antigens in the urine of patients. Both displayed high sensitivity and specificity on samples from the major endemic regions as well as, reflected parasite clearance in patients undergoing antimonial treatment in Ethiopia, a significant improvement over KATex, the only existing antigen detection test in the market. The ELISAs are user- friendly and quantitative and are suitable for deployment in routine care. They could also be adapted to a more cost- effective and point- of - care format such as lateral flow. Thus, we consider the tests a promising alternative to existing tests to monitor treatment and cure and worthy of further assessment and wide- spread deployment in endemic regions.
Acknowledgments
We would like to thank all of the patients and donors who provided the valuable samples that made this study possible. This work was supported by grants from the Bill and Melinda Gates Foundation (OPP49932 and OPP1084251) and funds from the Foundation for Innovative New Diagnostics (FIND) which supported this not involved in data analysis or interpretation was not involved in data analysis or interpretation.
We want to acknowledge InBios International, Inc., Seattle, USA and Kalon Biological Ltd for generating the ELISA kits and FIND, Geneva, Switzerland for sharing other disease urine samples with us. We would like to thank Dr. Randall Howard for project management support, Drs Gregory Ireton and Ajay Bhatia for early project guidance and Allen Casey (IDRI) and Marcel Hommel for assay development support. We also want to acknowledge Ruben Raychaudhuri and Lea Gaucherand for assistance in receiving and archiving samples.
Competing interests
The authors declare that they have no competing interests in this study.
Authors’ contributions
ACV, MSD and SGR conceived of the study and participated in manuscript writing; ACV, YLM, RM and SP designed the study, generated and interpreted data; AOA, AEA, MAS and MLA were involved in sample collection, indexing and follow-up interpretations, AH, MM, RA, DM, AA and HG were involved in co- ordination of the study, data interpretation and manuscript writing. All authors read and approved the final manuscript.