Synthetic Peptide-Based Enzyme-Linked Immunosorbent Assay for ...

3 downloads 0 Views 263KB Size Report
amoebiasis, and 8 patients with malaria from areas (Brazil or Sudan) where leishmaniasis is endemic, were also tested. Seventeen serum specimens from.
JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1996, p. 241–248 0095-1137/96/$04.0010 Copyright q 1996, American Society for Microbiology

Vol. 34, No. 2

Synthetic Peptide-Based Enzyme-Linked Immunosorbent Assay for Serodiagnosis of Visceral Leishmaniasis CHRISTEL FARGEAS,1 MARCEL HOMMEL,2 RHAIZA MAINGON,2 CIBELE DOURADO,2 MICHEL MONSIGNY,1 AND ROGER MAYER1* Glycobiologie, Centre de Biophysique Mole´culaire, Centre National de la Recherche Scientifique, Orle´ans Cedex 2, France,1 and Liverpool School of Tropical Medicine, Liverpool L3 5QA, United Kingdom2 Received 6 July 1995/Returned for modification 23 August 1995/Accepted 30 October 1995

Synthetic peptides, derived from the amino acid sequence of a Leishmania donovani clone, were used to develop an enzyme-linked immunosorbent assay (ELISA) for detecting antibodies against L. donovani. For this purpose, five peptides were conjugated to a protein carrier, human serum albumin (HSA), by using a heterobifunctional reagent, «-maleimidocaproic acid N-hydroxysuccinimide ester, to obtain a well-defined product. The sensitivity and the specificity of the peptide-specific ELISA were determined with a panel of 106 serum samples from individuals living in areas where visceral leishmaniasis is endemic; sera from post–kala azar dermal leishmaniasis-infected patients and from individuals suffering from other infectious diseases were also included. ELISAs were performed with either a single peptide-HSA conjugate or a mixture of two peptide-HSA conjugates. Ninety-seven percent of the serum samples from patients with visceral leishmaniasis had detectable antibodies to one or more of the single synthetic peptides. ELISA with a single peptide-HSA conjugate proved to be less sensitive (less than 71%) but more specific (up to 93%) than ELISA with crude promastigote antigens (80% sensitivity and 79% specificity); when a combination of two different peptide-HSA conjugates was used, the test increased both in sensitivity and in specificity. Chemically defined peptide-protein conjugates improve the reproducibility and reliability of ELISA for the serodiagnosis of L. donovani infection. oping VL. The standard treatment for VL uses a 20- to 28-day chemotherapy regimen of injectable pentavalent antimonial agents, but this long, very expensive, and rather toxic treatment is increasingly ineffectual, particularly in immunosuppressed patients. Post–kala azar dermal leishmaniasis (PKDL) is one of the better-known forms of VL recrudescence following inadequate treatment; in India it occurs in 20% of cases, but it is rarer in other locations. The clinical suspicion of VL may be confirmed by finding the parasite in its visceral location by invasive procedures, including those in which splenic aspirates or bone marrow or lymph node biopsy specimens are obtained (30); the detection of parasites may be achieved either by microscopic examination of stained biopsy samples or, when feasible, by using in vitro cultivation or PCR with these specimens to increase the chance of parasite detection (43). These invasive procedures are, however, difficult to perform under field conditions, may be hazardous to the patient, require the skills of experienced laboratory technicians, and do not provide a positive diagnosis in all cases. In VL, high titers of antileishmanial antibodies may be detected, and this feature has been exploited for the development of various serological tests such as indirect immunofluorescent antibody test (4, 5), enzyme-linked immunosorbent assay (ELISA) (3, 14, 23, 27, 29, 38, 40, 46), or direct agglutination test (DAT) (21, 24, 25, 39, 50). Although the various techniques described in the literature are sensitive enough, none has as yet been proven to be reliable enough to obviate the need for microscopy as the ‘‘gold standard’’ for diagnosis. One of the problems of serological assays with crude leishmanial antigens is the existence of cross-reactivity with other pathogens including Trypanosoma cruzi, mycobacteria, malaria parasites, or amoeba, which are coendemic in some parts of the world (3, 23, 25, 29, 38, 40, 46). Improved specificity may be achieved by the use of purified leishmanial antigens, and several candidates either in the form of recombinant proteins or

Visceral leishmaniasis (VL), also known as kala azar, is a severe, often fatal, disease which, having no pathognomonic symptoms, is usually identified clinically as a combination of prolonged unexplained fever, hepatosplenomegaly, and pancytopenia. Leishmania donovani, the causative agent of VL, is a protozoan parasite that is a member of a complex group of species in the family Trypanosomatidae (17). The life cycle of the organism goes through two morphologically different stages: the amastigote, which is intracellular in cells of the reticuloendothelial system, and the promastigote, which is an extracellular flagellated form found in the gut of the sandfly vector. L. donovani has a wide geographical distribution and is found in many tropical and subtropical countries, where it is normally restricted to its sylvatic and peridomestic transmission cycle (including wild canids, dogs, and sandflies) and where it only occasionally spills over to infect humans (26, 33). The normal pattern of infection in areas where the disease is endemic is one in which the infection rate in dogs may be high (reflecting a high degree of transmission) but in which the disease in healthy humans is relatively rare. In contrast, in economically depressed areas (e.g., ‘‘favellas’’ of Jacobina in Brazil) (2) or areas of political upheaval (e.g., the war situation of southern Sudan) (42), a combination of risk factors including malnutrition (16) may transform a higher proportion of subclinical or asymptomatic infections into severe disease; this feature makes VL a true disease of poverty. In addition, the spread of the human immunodeficiency virus pandemic to areas where leishmaniasis is endemic may substantially increase the incidence of human disease; this interaction has been particularly well described in Spain (36) and southern France (45), where it was estimated that AIDS increased by 100-fold the risk of devel* Corresponding author. Mailing address: Glycobiologie, Centre de Biophysique Mole´culaire, C.N.R.S, rue Charles-Sadron, F-45071 Orle´ans Cedex 2, France. Phone: (33) 38 51 55 62. Fax: (33) 38 69 00 94. Electronic mail address: [email protected]. 241

242

FARGEAS ET AL.

J. CLIN. MICROBIOL. TABLE 1. HPLC and mass spectrometric analysis of the synthesized peptides

Peptide

P1 P2 P3 P4 P5

Sequencea

TEEYVQTSRESS DSPPLSSPTLFSSL RGGYLGRGDSPPLSS GRGDSG GGARGD

Average Retention Calculated Measured no. of time b mass (Da) mass (Da) coupled (min)c peptides

1,414.6 1,414.6 1,446.7 1,447.8 1,517.7 1,518.9 547.5 547.9 531.2 532.2

5 4 4 5 6

18.8 21.5 19.4 3.0d 3.3d

a

Amino acids in boldface type indicate the conserved RGD motif. Masses were assessed by matrix-assisted laser desorption ionization mass spectrometry. c Peptides were analyzed by reverse-phase chromatography on a LiChrospher 300WP RP-18 column (250 by 4.6 mm). Mobile phase A was H2O plus 0.1% trifluoroacetic acid, and mobile phase B was acetonitrile plus 0.1% trifluoroacetic acid; the elution was achieved by using a gradient from 5 to 80% mobile phase B over 20 min; the flow rate was 1 ml/min. d Peptides P4 and P5 were eluted with an isocratic mixture of 5% acetonitrile in water containing 0.1% trifluoroacetic acid. b

FIG. 1. DNA sequence and deduced amino acid sequence around the ArgGly-Asp (RGD) site in the gp63-like gene (clone 10). The RGD extended domain is indicated in boldface type. Numbers are arbitrary and refer to the amino acid residues. Peptide P1, P2, and P3 sequences are underlined.

in the form of antigens purified from parasite extracts have been proposed (12, 29, 38, 46). Here we report on an evaluation of the anti-Leishmania antibodies present in the sera of patients by using synthetic peptides, the sequence of which derives from the sequence of an L. donovani gene homologous to genes of the gp63 multigene family (35), the immunodominant surface glycoprotein of Leishmania promastigotes (9). MATERIALS AND METHODS Collection of human sera. Thirty-five serum specimens from kala azar patients were divided into the following different categories on the basis of clinical findings, DAT, and/or skin test results: VL1, 10 serum specimens from kala azar patients with positive parasitology; parasites were identified in spleen or lymph aspirates; VL2, 7 serum specimens from presumably previously exposed patients with a positive skin test (.5 mm) and a positive DAT result; VL3, 4 serum specimens from kala azar patients; they were admitted because of a positive DAT result (i.e., .1:3,200) and/or a strong clinical suspicion of kala azar; no spleen or lymph aspirations were done; VL4, 5 serum specimens from presumably previously exposed patients with a positive skin test (.5 mm) and a negative DAT result; and VL5, 9 serum specimens from patients with symptoms of illness but a negative DAT result; no other data were available for these patients. All sera from patients with VL were from the Sudan except six VL1 samples, which were from Brazil. Sudanese sera had been included in the different operational groups (VL1 to VL5) on the basis of clinical, parasitological, and DAT performed in the field, as reported by Goris (22). Nine serum specimens from Sudanese patients with PKDL admitted on clinical grounds (hypopigmented and erythematous patches, mostly on the face) were included in the present study. In order to point out a putative cross-reactivity, sera from patients suffering from other infectious diseases, namely, 6 patients with cutaneous leishmaniasis, 14 patients with trypanosomiasis, 10 patients with strongyloidiasis, 10 patients with amoebiasis, and 8 patients with malaria from areas (Brazil or Sudan) where leishmaniasis is endemic, were also tested. Seventeen serum specimens from healthy donors obtained from the Centre de Transfusion Sanguine (Orle´ans, France) were used as controls from areas where leishmaniasis is not endemic. All sera were stored at 2208C until use. Synthetic peptides. Three different peptides derived from the sequence of a gp63-like gene (clone 10) (35) were used. This gene was isolated from a genomic library of L. donovani amastigote and showed 33% identity with the published sequence of a promastigote gp63 gene in L. major (13). The positions of the three peptides in the cloned sequence are indicated in Fig. 1. Peptides P1, P2, and P3 were selected on the basis of their uniqueness, accessibility, hydrophilicity (32), and hydropathy (28), determined by using the algorithm DNASTAR. Peptide P3 contains a RGDS integrin ligand motif. Two other short, peptides P4 and P5, containing the RGDS motif but unrelated to the gp63 sequence were also included. Peptides were synthesized according to the solid-phase procedure of Merrifield (37) by the 9-fluorenylmethoxycarbonyl (Fmoc) strategy with 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate activation (Fastmoc; 0.1-mmol small-scale cycles) and a 4-hydroxymethyl phenoxymethyl resin

(0.96 meq/g; Applied Biosystems, Foster City, Calif.) on a peptide synthesizer (model 433A; Applied Biosystems) with conductivity monitoring. Fmoc derivatives were from Neosystem (Strasbourg, France). All amino acids were of the L configuration. Upon synthesis, the peptide-resin bond and the side chain-protecting groups were removed by treatment with trifluoroacetic acid in the presence of scavengers (31). Crude peptides were purified by reverse-phase highpressure liquid chromatography (HPLC) with a C18 column (250 by 10 mm; Vydac, Hesperia, Calif.). The purity of the peptide was confirmed by the presence of a single peak by analytical HPLC on a LiChrospher 300WP RP-18 column (250 by 4.6 mm; Merck, Darmstadt, Germany). The amino acid analysis was performed with the Picotag system (Waters, Bedford, Mass.). Analysis of the peptides and the conjugates was performed by matrix-assisted laser desorption ionization mass spectrometry on a Lasermat 2000 (Finnigan MAT, San Jose, Calif.) time-of-flight mass spectrometer (Table 1). Preparation of antigens. (i) Crude promastigote antigens extract. L. donovani (MHOM/ET/67/Hu3; LV9 strain) promastigotes were grown to the stationary phase in HO-minimal essential medium (HO-MEM) medium supplemented with 20% sterile fetal calf serum (6). Parasites were collected by centrifugation and were washed three times in phosphate-buffered saline (PBS; pH 7.4), and the final pellet was adjusted to 108 parasites per ml. A soluble promastigote extract was prepared by disrupting the parasites by using three rapid freeze-thaw cycles (liquid nitrogen, 378C); this was followed by sonication on an ice bath. The extract was then centrifuged at 11,500 3 g for 30 min to remove particulate debris. The protein concentration was assessed with Coomassie blue as described by Bradford (7), and the supernatant was divided into aliquots and stored at 2808C. (ii) Synthesis of peptide-HSA conjugates through a thioether linkage. Peptides were conjugated to human serum albumin (HSA; Sigma, St. Louis, Mo.) through a thioether bond (34) by using 6-maleimidocaproic acid N-hydroxysuccinimide ester (MCS; Sigma) as a heterobifunctional reagent. In a first step, the carrier was thiolated by acylation of amino groups with the heterobifunctional reagent N-succinimidyl-3-(2-pyridyldithio)propionate (SPDP) from Pierce (Rockford, Ill.). Briefly, 9 mmol (2.6 mg) of SPDP in 400 ml of ethanol was added to a solution of 0.45 mmol (30 mg) of HSA in 3 ml of 0.1 M borate buffer (pH 8.6). Upon stirring for 3 h at room temperature, the thiolactivated carrier was purified by gel filtration on a Bio-Gel P2 column (Bio-Rad, Paris, France), eluted with n-butanol–water (5:95; vol/vol), and lyophilized. The amount of SPDP coupled was determined by measuring the amount of pyridine2-thione released by measuring its A343 nm (15) upon reduction of the disulfide bond with Tris-(2-carboxyethyl)phosphine (TCEP) (11) from Molecular Probes (La Jolla, Calif.). The modified carrier contained approximately 10 thiopyridyl groups per macromolecule. The maleimidylated peptide was prepared by incubating 7 mmol of MCS in 50 ml of dimethylformamide and 6 mmol of peptides in 1 ml of 0.1 M sodium phosphate buffer (pH 7.0) for 90 min at room temperature. The maleimidylated peptide was separated from MCS by gel filtration on Bio-Gel P2 equilibrated with 1 mM sodium phosphate buffer (pH 7.0). The maleimidylated peptide was coupled to the modified carrier. The thiolactivated HSA (0.22 mmol, 15 mg) was dissolved in 1.5 ml of nitrogen-saturated 0.1 M sodium phosphate buffer (pH 7.0) and was reduced with 1 eq of TCEP for 1 h at room temperature. To this solution, an amount of dry maleimidylated peptide, equivalent to threefold the molar equivalent of thiol groups on the carrier, was added. The reaction mixture was stirred under nitrogen for 18 h at 208C. To block any remaining thiol groups, 10 mg (54 mmol) of iodoacetamide (Aldrich, Milwaukee, Wis.) was added and the reaction mixture was stirred for 1

VOL. 34, 1996

SERODIAGNOSIS OF VISCERAL LEISHMANIASIS WITH PEPTIDES

h at 208C. The conjugate was separated from the unreacted peptide by gel filtration through a Bio-Gel P100 column equilibrated in n-butanol–water (5:95; vol/vol). The conjugate was lyophilized, and the number of peptides linked per HSA molecule was determined from mass spectrometry data. Peptide ELISA. Because free peptides did not give any response with sera, in preliminary experiments (19) we selected a peptide-carrier construct in order to obtain a better epitope presentation; HSA proved to be the carrier of choice compared with ovalbumin, bovine serum albumin, thyroglobulin, and myoglobulin because of its lowest background reactivity. Peptides were initially coupled to HSA by means of a glutaraldehyde (1, 44). These conjugates appeared to be recognized nonspecifically by many patient serum specimens. To overcome this problem, we used a heterobifunctional reagent in order to reproducibly couple the peptide to the carrier in a well-defined way. Peptide-HSA conjugates were incubated in 96-well flat-bottom polystyrene microtiters plates (Nunc Maxisorp Immuno plates; Poly Labo, Strasbourg, France) at a concentration of 50 mg/ml in 0.1 M sodium carbonate buffer (pH 9.6) for 90 min at 378C (100 ml per well). The plates were washed three times with 1 M PBS (pH 7.4) containing 0.1% (vol/vol) Tween 20 and 5 mg of bovine hemoglobin (Hb; Sigma) per ml. In order to avoid nonspecific absorption of the antisera, the wells were further treated with 200 ml of PBS containing 0.1% (vol/vol) Tween 20 and 30 mg of Hb per ml for 2 h at room temperature. After three washes as described above, the sera (50 ml) to be tested were added after a 60-fold dilution in PBS containing 0.1% (vol/vol) Tween 20 and 5 mg of Hb per ml. The plates were kept at room temperature for 2 h. Following three washes, alkaline phosphatase-conjugated goat anti-human immunoglobulin G (whole molecule) antibody from Sigma diluted 400-fold in the same buffer (100 ml) was added. The plates were incubated for 1 h at room temperature and were washed three times as described above; 1 mg of para-nitrophenyl phosphate per ml (100 ml) in 1 mM MgCl2–1 mM ZnCl2–0.1 M glycine buffer (pH 10.4) was used as the substrate. Upon 1 h of incubation at 258C in the dark, the A405 was read with a Titertek Multiskan Plus MKII instrument (Labsystem, Helsinki, Finland). In the case of crude antigens, the procedure described above was used, with the following modifications. The plates were coated with 12 mg of antigens per ml in a 0.1 M sodium carbonate buffer (pH 9.6), and at the last step the substrate was incubated for 30 min. All sera were tested in duplicate. HSA-coated wells were used as specificity controls. A cutoff value was calculated for each conjugate by using the data obtained from 17 crude L. donovani promastigote extract-negative sera from areas where leishmaniasis is not endemic (negative controls). The mean A405 was determined for each peptide and each combination of peptides. The mean value plus 2 standard deviations was chosen as the cutoff value (see Fig. 2 and 3). Any value above the determined cutoff value was considered positive and statistically significant.

RESULTS The present study was carried out with serum samples from 106 individuals including 17 from healthy donors from areas where leishmaniasis is not endemic and 35 patients suffering from kala azar, 6 patients suffering from PKDL, 6 patients suffering from cutaneous leishmaniasis, and 42 patients suffering from other infectious diseases. Patients with kala azar were divided into the five groups described above. Individual results for crude antigens and peptide-HSA conjugates are provided as scattergraphs in Fig. 2 and 3. Detection of antibodies specific for L. donovani crude antigens. In order to determine the presence of anti-L. donovani antibodies, sera were tested at a dilution of 1:60 by ELISA with crude promastigote antigens. The data in Table 2 indicate that 28 of the 35 (80%) serum samples from patients with kala azar had antibodies detectable by an ELISA with crude leishmanial antigens, while only 16 (46%) were positive by DAT. All (100%) DAT-negative serum samples from patients in groups VL1 and VL4 and 4 (45%) of the 9 DAT-negative serum samples patients in the VL5 group were positive in the ELISA with crude promastigote antigens. To assess the assay specificity, the cross-reactions with 48 serum samples from patients suffering from diseases other than visceral leishmaniasis, namely, cutaneous leishmaniasis, malaria, trypanosomiasis, strongyloidiasis, and amoebiasis, and 17 serum samples from healthy individuals from a country where leishmaniasis is not endemic were tested. Antibody binding to whole soluble leishmanial antigens was not detected by ELISA

243

with sera from healthy subjects and patients with strongyloidiasis and amoebiasis, but 50% of the samples from patients with cutaneous leishmaniasis, 50% of the samples from patients with malaria, and 50% of the samples from patients with trypanosomiasis (Chagas’ disease) contained antibodies that cross-reacted with crude L. donovani antigens (Table 2). Overall, ELISA conducted with crude parasite antigen led to 80% sensitivity and 79% specificity. Detection of antibodies specific for peptides linked to HSA through a thioether linkage. The presence of antibodies in the same set of sera was further studied by determining their ability to react with synthetic peptides in an ELISA system. Peptides were coupled to HSA by their a-amino group through a thioether linkage with a heterobifunctional reagent, MCS. Among the 35 serum samples from patients with kala azar, 34 (97%) recognized one or more synthetic peptides: 25 (71%) reacted with P1, 18 (51%) reacted with P2, 17 (49%) reacted with P3, 22 (63%) reacted with P4, and 17 (49%) reacted with P5 (Table 3). The binding of antisera to peptides in sera from the different groups of patients was heterogeneous, but peptide P1 or P4 appeared to be the most antigenic toward sera from any group of patients with VL. In patients with confirmed kala azar (group VL1), both conjugates P4 and P5 appeared to be the most antigenic (80%) and conjugate P1 was the least active. In addition, conjugates P1, P4, and P5 allowed the antibody detection in 67, 83, and 83% of the sera from patients with PKDL, respectively. The specificity reached 77% with the five peptides together, but individual peptides reached 93% of specificity with P1, P2, or P3 and 90% with P4 or P5 (Table 3). None of the six serum specimens from patients with cutaneous leishmaniasis gave a positive response with P4 and 1 (10%) of 10 serum specimens from patients with strongyloidiasis gave a positive response with P5; 2 (25%) of 8 serum samples from patients with malaria reacted more frequently with P2, 3 (21%) of 14 samples from patients with trypanosomiasis reacted with P1 or P4, 2 (20%) of 10 serum samples from patients suffering from amoebiasis reacted with P3, P4, or P5, and 1 (5%) of 17 serum samples from healthy donors reacted with each conjugate. According to the results obtained with the single peptideHSA conjugates, it appears that a judicious combination of two conjugates could improve the sensitivity of the assay. Accordingly, 25 mg of each of the two conjugates per ml in various combinations were coated onto a single well and were tested against sera diluted 120-fold in order to decrease the background due to nonspecific reactions and thus increase the sensitivities of the assays. The same 35 serum samples were tested against the following six mixtures: P1-P2, P1-P3, P1-P4, P2-P3, P2-P4, and P2-P5. The results, summarized in Table 4, show that the combination of P1 and P4 was the least reactive (63%) with any group of sera. Any other combination allowed the detection of between 80 and 88% of the sera from patients with kala azar. The specificities of these assays with a mixture of two conjugates were also estimated (Table 4). None of the sera from patients suffering from cutaneous leishmaniasis reacted with any of the combinations. While 1 (12%) serum sample from a patient with malaria cross-reacted with P2-P5, samples from patients with trypanosomiasis reacted more often with P1-P4 (28%), those from patients with strongyloidiasis reacted more often with P1-P3 (30%), those from patients with amoebiasis reacted more often with P1-P3 (40%), and 2 (12%) serum samples from healthy subjects reacted more often with P2-P5.

244

FARGEAS ET AL.

J. CLIN. MICROBIOL.

VOL. 34, 1996

SERODIAGNOSIS OF VISCERAL LEISHMANIASIS WITH PEPTIDES

245

FIG. 2. ELISA of sera from patients with leishmaniasis and from healthy and diseased controls tested against crude promastigote antigen or different HSA-peptide conjugates. Each symbol represents the absorbance obtained with a single serum; F and E, serum reacting with none of the conjugates; å and Ç, serum reacting with every conjugate; ■ and h, serum reacting with one conjugate; , and