- Email: [email protected]
Studies on immunodiagnosis of dracunculiasis. II. Search for circulating antigens
Acta Tropica 70 (1998) 303 – 315
Studies on immunodiagnosis of dracunculiasis. II. Search for circulating antigens Paul Bloch *, Birgitte J. Vennervald, Paul E. Simonsen Danish Bilharziasis Laboratory, Jaegersborg Alle 1D, DK-2920 Charlottenlund, Denmark Received 20 August 1997; received in revised form 11 February 1998; accepted 9 April 1998
Abstract Sera from individuals living in a dracunculiasis endemic area of northern Ghana were examined for circulating Dracunculus medinensis antigens by applying protocols previously developed for detection of circulating antigens in other helminth infections. Antisera from rabbits immunised with homogenized first stage D. medinensis larvae were used for antigen capture and detection in three different forms, namely non-treated, biotinylated and horseradish peroxidase (HRP)-labelled. Three different preparations of human sera were examined, namely non-treated, pre-treated with polyethylene glycol/ethylenediaminetetra-acetic acid (PEG/EDTA) for analysis of precipitated immune complexes, and pre-treated with trichloroacetic acid (TCA) for analysis of isolated glycoproteins. In both SDS-PAGE/Western blotting and ELISA, significant reactivity was observed between non-treated and treated rabbit-antisera on the one hand and non-treated and treated human sera on the other. However, no significant response differences were observed between sera obtained from individuals with dracunculiasis and non-endemic controls. The reasons are analysed and possible explanations presented. The study provided no evidence that D. medinensis-specific circulating antigens, detectable by relatively simple means, occur in infected individuals. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Circulating antigens; Immunodiagnosis; Dracunculiasis; Dracunculus medinensis
* Corresponding author. Tel.: +45 39 626168; fax: + 45 39 626121, e-mail: [email protected] 0001-706X/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S0001-706X(98)00039-4
304
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
1. Introduction The prepatent period in human Dracunculus medinensis infections, i.e. the period from ingestion of infective third stage larvae to emergence of the mature female worm on the skin, is 10 – 12 months (Muller, 1971). With exception of the last 1–2 weeks, when the worm becomes palpable, the prepatent period is asymptomatic. Currently, no reliable method for diagnosis of D. medinensis infection during the asymptomatic part of the prepatent period exists. By providing means for proper parasitological characterisation of people living in endemic areas, such a method might be useful in dracunculiasis research and control. Attempts have previously been made to develop a diagnostic method for prepatent infection with D. medinensis by detection of specific serum antibodies, but with different degrees of success (Fagbemi and Hillyer, 1990; Garate et al., 1990; Bloch et al., 1993; Bloch and Simonsen, 1998). Another approach would be detection of specific circulating antigens in serum from infected individuals. With the purpose of investigating the potential for developing a reliable method for diagnosis of dracunculiasis through detection of specific circulating antigens, this study carried out a search for such antigens in non-treated, PEG/EDTA-treated and TCA-treated sera from individuals with dracunculiasis. Protocols developed for detection of circulating antigens in individuals with bancroftian filariasis (Lunde et al., 1988) and onchocerciasis (Chandrashekar et al., 1990; Thambiah et al., 1991) were applied. Antibodies for antigen capture and detection were obtained from antiserum from rabbits immunised with a D. medinensis first stage larval homogenate.
2. Subjects and methods
2.1. Study indi6iduals Three categories of individuals were included in the study. Seven individuals (four males, three females; mean age 31 years, range 22–45 years) with a prepatent D. medinensis infection (category a) were from an area of northern Ghana which is highly endemic for dracunculiasis. The infection status of this study group was determined by closely following individuals for 15 months after blood sampling during which period emergence of worms on the skin was recorded. Seventeen individuals (six males, 11 females; mean age 31 years, range 8–50 years) with a patent D. medinensis infection (category b) were from the same endemic area of northern Ghana. Five Danish individuals (three males, two females; mean age 26 years, range 21 – 33 years) who had never been to tropical countries (category c) were used as non-endemic controls.
2.2. Preparation and pre-treatment of human serum Serum was recovered from venous blood samples after clotting and centrifugation, and sodium azide (NaN3) was added to a concentration of 15 mM prior to
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
305
freezing of the samples at − 80°C until use. In part of the serum, immune complexes were precipitated with polyethylene glycol/ethylenediaminetetraacetic acid (PEG/EDTA) (Lunde et al., 1988; Chandrashekar et al., 1990; Thambiah et al., 1991). This was done by mixing 200 ml serum, 200 ml 0.2 M Na2EDTA pH 7.5 and 100 ml 12% PEG 8000 (Sigma) pH 8.3. After overnight incubation at 4°C the solution was centrifuged at 15000× g for 30 min. The precipitate was washed in 500 ml cold 2.4% PEG in 0.1 M borate buffer (0.1 M Na2B4O7 · 10 H2O solution added to 0.1 M boric acid solution until pH 8.3 was obtained). The precipitate was dissociated by resuspending it in 100 ml 0.1 M Tris–HCl pH 2.6 and the solution was neutralised with 20 ml 1.0 M Tris pH 8.0. Another part of the serum was pre-treated with trichloroacetic acid (TCA) to dissociate immune complexes and to isolate the glycoproteins (De Jonge et al., 1987). This was done at room temperature by incubating 100 ml serum mixed with 100 ml 4% TCA solution for 5 min on a rocking table followed by centrifugation at 7000×g for 30 min. The supernatant was neutralised with 100 ml carbonate buffer pH 9.6 (0.174 M NaHCO3 and 0.070 M Na2CO3).
2.3. Antigens D. medinensis adult female worms and first stage larvae obtained from infected individuals from northern Ghana were the sources of antigen. The methods for recovery of parasites, and for preparation of crude homogenates of adult worms (ADGW) and larvae (LVGW) has been described previously (Bloch and Simonsen, 1998). For one experiment a pure adult worm antigen (puADGW) was made by carefully removing larvae from the uterus of female worms by milking and dissection before homogenizing the worms. Crude homogenates of adult Onchocerca 6ol6ulus (OVAG) and Brugia pahangi (BPAG), and human albumin (Sigma, St. Louis, MO, USA) prepared in various concentrations, were used as control antigens.
2.4. Rabbit-antiserum Two naive 3 – 4 months old rabbits were immunized sub-cutaneously with LVGW in Freund’s incomplete adjuvant (1:1) six times with 2-weeks interval. Antiserum was prepared from blood obtained 12 days after the last immunization. The IgG fraction was isolated by affinity purification on Protein A Sepharose CL-4B (Kabi Pharmacia Diagnostics, Sweden) on a Automated Econo System (Bio-Rad). The IgG eluate was immediately neutralised in 1 M Tris pH 8.0 in the proportion 1:5 before being buffer exchanged by overnight dialysis against PBS. The IgG fraction was concentrated in a Sartorius Ultrasart Cell 50 (Sartorius, Germany) to a protein concentration of 5.0 mg/ml. Biotinylation was carried out according to Hudson and Hay (1989) with the modification that 5.0 mg N-hydroxy-succinimidobiotin (Sigma, St. Louis, MO) was dissolved in 125 ml dimethyl sulfoxide (DMSO, Merck, Germany). While stirring,
306
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
10 ml of the biotin solution was added to 3.0 ml IgG fraction of the antiserum. Following dialysis against PBS overnight at 4°C, the protein content was measured to be 800 mg/ml. Sodium azide was added to 15 mM and the solution was kept frozen at − 20°C until use. HRP-labelling of the IgG fraction was carried out as previously described (Avrameas and Ternynck, 1971). Following centrifugation of the HRP-labelled conjugate at 10000×g for 30 min, the supernatant was filtered through a 0.22-mm sterile filter and the protein content was measured to be 4.4 mg/ml. Sodium azide was added to a concentration of 15 mM, and the solution was kept frozen at −20°C until use.
2.5. SDS-PAGE and Western blotting Normal resolution sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) as well as high resolution tricine-based SDS-PAGE, and Western blotting, were performed according to procedures described previously (Bloch et al., 1993; Bloch and Simonsen, 1998).
2.6. ELISA Circulating antigens were measured either by indirect enzyme linked immunosorbent assay (ELISA) (Bloch et al., 1993; Bloch and Simonsen, 1998) or by double antibody sandwich ELISA. For the sandwich ELISA, the wells of the microtitre plates (Maxisorp Immunoplate 442404, Nunc, Denmark) were coated with IgG rabbit-antiserum by overnight incubation at 4°C with 100 ml antiserum diluted in coating buffer (2.5 mM NaH2PO4 · 1 H2O, 7.5 mM Na2HPO4 · 12 H2O, 0.145 M NaCl, pH 7.2). After washing and blocking in PBS-Tween containing 0.5% bovine serum albumin, the wells were incubated overnight at 4°C with 100 ml diluted sample (pre-treated or non-treated human serum) applied in triplicate. All other parts of the sandwich ELISA procedure were similar to the indirect ELISA procedure.
3. Results
3.1. Reacti6ity of rabbit-antisera The reactivity of antisera from the two LVGW-immunised rabbits was analysed by SDS-PAGE/Western blotting. IgG antibodies in non-purified sera from both rabbits reacted to numerous and almost similar LVGW and ADGW antigens in the weight range 10 – 200 kDa. Limited and no reactivity was observed to BPAG and OVAG, respectively, with these sera. Pooled non-labelled, biotinylated or HRPlabelled IgG-purified rabbit antisera reacted to similar antigens in LVGW as the non-purified antisera. However, higher background reactivities were found for the non-labelled and biotinylated as compared with the HRP-labelled rabbit antiserum.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
307
Limited reactivity was observed against OVAG whereas essentially no reactivity was observed against BPAG. Both rabbits therefore responded to the D. medinensis antigens, with only little cross-reactivity to the tested non-D. medinensis antigens. The ability of the rabbit antisera to detect D. medinensis antigens in human serum was assessed in a sandwich ELISA. IgG-purified rabbit antiserum and HRP-labelled IgG-purified rabbit antiserum was used as capture and detecting antiserum, respectively, and test samples consisted of pooled serum from the non-endemic control individuals to which different concentrations of D. medinensis antigens were added (Fig. 1). For both untreated (Fig. 1A) and TCA pre-treated (Fig. 1B) samples to which LVGW, ADGW (containing uterus dwelling first stage larvae) or puADGW (purified for uterus dwelling first stage larvae) was added, the responses were rather similar and markedly stronger than the responses observed when OVAG, BPAG or no antigen was added.
3.2. Analyses for circulating antigens in untreated human sera Sandwich ELISA was first carried out to detect circulating D. medinensis antigens in untreated human sera using biotinylated rabbit antiserum for antigen detection. Relatively high absorbance values were observed for sera from patently D. medinensis infected individuals, but similar or even higher reactivities were seen for sera from non-endemic controls and for conjugate controls. No major improvements
Fig. 1. Sandwich ELISA showing the absorbance values of pooled non-treated (A) and TCA pre-treated (B) sera diluted 1:10 in PBS from five non-endemic control individuals to which LVGW (square), ADGW (triangle up), puADGW (triangle down), OVAG (diamond), BPAG (star) or no antigen (circle) was added in concentrations from 0.5 to 100 mg/ml. Samples were further diluted 1:10 in PBS prior to application. Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12 mg/ml and HRP-labelled rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration of 5 mg/ml.
308
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
Fig. 2. Indirect ELISA showing the absorbance values of non-absorbed (solid bar) or preabsorbed (/) biotinylated rabbit-IgG-anti-LVGW antiserum, and of HRP-labelled rabbit-IgG-anti-LVGW antiserum (¯), and avidin-HRP alone (open bar), against LVGW (5 mg/ml), rabbit-IgG-anti-LVGW antiserum (5 mg/ml), normal rabbit serum (5 mg/ml) and blocking agent alone, each of which were coated to the wells. All rabbit-antisera were used at a concentration of 1 mg/ml. All wells were blocked with 0.5% bovine serum albumin and 0.1% Tween 20. The asterisk represents an experiment not performed.
were observed when varying the dilution of serum (1:4, 1:10, 1:100 and 1:1000) or the concentration of capture (0.1, 1.0, 5.0 and 12.0 mg/ml) or detecting (0.1, 1.0 and 5.0 mg/ml) antiserum. Therefore, the reactivity of HRP-labelled and biotinylated rabbit-IgG-anti-LVGW antiserum to LVGW, non-labelled rabbit-IgG-anti-LVGW and normal rabbit serum was compared (Fig. 2). The biotinylated antiserum was tested without pre-treatment as well as preabsorbed in a Sepharose column coated with rabbit-IgG-anti-LVGW (to extract potential anti-idiotype antibodies). Nonabsorbed biotinylated rabbit-IgG-anti-LVGW antiserum reacted extensively with both LVGW and rabbit-IgG-anti-LVGW, and it reacted to a limited extent with normal rabbit serum. When using preabsorbed biotinylated antiserum, the reactivity of the eluate remained equally high against LVGW, but reduced to almost half the value against itself (i.e. rabbit-IgG-anti-LVGW), reaching a level equivalent to the level of reactivity of avidin-HRP alone. Non-specific binding to the wells blocked with bovine serum albumin was not observed. HRP-labelled rabbit-IgGanti-LVGW reacted extensively with LVGW, to a very limited extent with rabbitIgG-anti-LVGW and not at all with normal rabbit serum. Thus, in contrast to the rather extensive reactivities observed when using the biotinylated rabbit antiserum, only limited reactivities were observed when testing the HRP-labelled rabbit antiserum against normal rabbit antiserum or rabbit-IgG-anti-LVGW antiserum.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
309
In SDS-PAGE/Western blotting, using biotinylated rabbit-IgG-anti-LVGW antiserum as detecting antiserum, numerous bands were observed in the molecular weight range from 40 to 200 kDa. However, in accordance with the ELISA findings, no major response differences were observed between any of the three infection categories. To determine the specific factors contributing to the high background reactivity observed when testing the reactivity of biotinylated rabbit antiserum to LVGW, the individual assay reagents were applied in different combinations in ELISA (Table 1). A major conclusion from this experiment was that avidin-HRP alone strongly react with capture antiserum and the surface of the microtitre plate well. Whether detecting antiserum alone reacted non-specifically with capture antiserum could not be determined definitively. It therefore appeared impossible to use biotinylated IgG rabbit-anti-LVGW antiserum as detecting antiserum. An attempt was then made by ELISA to identify D. medinensis-specific circulating antigens in non-treated human sera by using HRP-labelled rabbit-IgG-antiLVGW antiserum as detecting antiserum. The rabbit-antiserum reacted extensively with LVGW, but the absorbance values for sera from prepatent and patent dracunculiasis patients were almost as low as those from non-endemic controls (Fig. 3). Thus, the experiment provided no evidence for the occurrence of specific circulating antigens in the sera.
Fig. 3. Absorbance values of a sandwich ELISA for detection of circulating D. medinensis antigens in non-treated pooled sera diluted 1:10 obtained from five individuals from each of the following groups: prepatent D. medinensis infection (a), patent D. medinensis infection (b), and non-endemic control (c). A conjugate control group prepared as a complete ELISA but without addition of test serum was included. Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12 mg/ml and HRP-labelled rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration of 5 mg/ml. The reactivity of HRP-labelled rabbit-IgG-anti-LVGW antiserum was tested against LVGW. Each bar represents the mean of tests carried out in triplicate.
− + − − − − + + + + + + − −
1 2 3 4 5 6 7 8 9 10 11 12 13 14 − − − − − +b − − − − +d − +e +d
Test serum − − + − + + + + − + − + + +
Detecting antiserum + + + + + + − + +c +c + − +c +
Avidin-HRP + + + + + + + + + + + + + +
Substrate
High (1000–2500) High (1000–2500) High (1000–2500) Zero Zero Zero Zero High (1000–2500) Medium (200–1000) Medium (200–1000) Medium (200–1000) Low (1–200) Low (1–200) High (1000–2500)
Absorbance level (value range ×103)
The test sera were either from individuals with a patent D. medinensis infection or from non-endemic controls (these were untreated and diluted 1:10). Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12.0 mg/ml and biotinylated rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration 1.0 mg/ml. a Bovine serum albumin (0.5 and 2.0%) and 0.1% Tween 20 used as blocking agents all gave the same results. b Two serum samples from individual with patent D. medinensis infection were tested with the same results. c Streptavidin-HRP was also tested with the same result. d Two serum samples from individual with patent D. medinensis infection as well as two Danish controls were tested with the same results. e The serum sample was added before blocking agent; all samples (i.e. three from individuals with D. medinensis infection and three non-endemic controls) gave the same results.
− − − +a +a + − − + + + + + −
Capture antiserum Blocking
Test no.
Table 1 Summary of the results of a multi-factor ELISA experiment designed to show the effect on the absorbance value when excluding (−) or including (+) the different reagents of the assay
310 P. Bloch et al. / Acta Tropica 70 (1998) 303–315
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
311
Fig. 4. Absorbance values of an indirect ELISA for detection of circulating D. medinensis antigens in PEG/EDTA-treated pooled sera obtained from five individuals from each of the following groups: prepatent D. medinensis infection (a), patent D. medinensis infection (b), and non-endemic control (c). The dissociated immune complexes were diluted 1:1 and coated directly to the well of the microtitre plate. Rabbit-IgG-anti-LVGW antiserum in non-labelled (solid bar) and HRP-labelled (open bar) form was used as detecting antiserum at a concentration of 4.4 mg/ml. LVGW at a concentration of 5 mg/ml was included as a positive antigen control. HRP goat-anti-rabbit IgG conjugate diluted 1:2000 was used to detect non-labelled rabbit antiserum. Each bar represents the mean of tests carried out in triplicate.
3.3. Analyses for circulating antigens in PEG/EDTA precipitated immune complexes In a further attempt to search for D. medinensis-specific circulating antigens in the human sera, PEG/EDTA precipitates were dissociated and analysed by ELISA and SDS-PAGE/Western blotting. In ELISA, the dissociated precipitates were coated directly to microtitre plate wells following application of non-labelled or HRP-labelled rabbit-IgG-anti-LVGW antiserum for detection of circulating antigens (Fig. 4). Both rabbit-antisera reacted strongly with LVGW confirming their content of antibodies with specificity to the LVGW antigen homogenate. Furthermore, the HRP-labelled rabbit-antiserum reacted moderately and the non-labelled antiserum weakly with pre-treated sera from each study category. The experiment provided no evidence for the occurrence of D. medinensis circulating antigens in the sera. In SDS-PAGE/Western blotting, using HRP-labelled rabbit-IgG-anti-LVGW antiserum as detecting antiserum, numerous distinct bands were observed in the weight range from 20 to 150 kDa. However, in accordance with the ELISA findings, no major response differences were observed between the three study groups.
312
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
3.4. Analysis for circulating antigens in TCA pre-treated human sera TCA-isolated glycoproteins from sera from the three study groups were analysed for the occurrence of D. medinensis specific antigens in SDS-PAGE/Western blotting. The results showed that for sera from each of the study groups, three distinct and one faint band occurred on the strips, indicating that among the purified antigenic glycoproteins, none were specific to D. medinensis. Thus, purifying glycoproteins from the test sera provided no means for detecting D. medinensis specific circulating antigens. 4. Discussion By applying methodologies already developed for detection of specific circulating antigens in sera from individuals infected with O. 6ol6ulus or W. bancrofti (Lunde et al., 1988; Chandrashekar et al., 1990; Thambiah et al., 1991), this study aimed at detecting D. medinensis specific circulating antigens in the blood of infected individuals in either free form or in antigen-antibody complexes. To limit the extent of unspecific reactions with the capture and detecting antisera, the rabbits were immunised with D. medinensis first stage larvae homogenate (LVGW) and not adult female worm homogenate (ADGW) since we have previously observed epitopes resembling human albumins and immunoglobulins in ADGW (unpublished data). The rabbit antisera obtained were shown by SDS-PAGE/Western blotting to react extensively (in terms of staining intensity and number of bands) with LVGW no matter in which form they were tested (i.e. pooled or non-pooled, labelled or non-labelled). Their reactivity with dissolved human albumins and immunoglobulins was minimal (data not shown). In order to remove serum proteins which could interfere with the circulating antigen capture and detection process (e.g. low affinity rabbit IgM antibodies or rabbit serum albumin), the IgG fraction of pooled rabbit-antisera was extracted prior to use. Previous observations that human IgG antibodies react extensively with LVGW and ADGW justified the use of specific IgG antibodies for detection of circulating D. medinensis antigens in human sera (Bloch and Simonsen, 1998). The pooled IgG-purified rabbit-antiserum was tested in three different forms: non-labelled, labelled with biotin, and labelled with HRP. A high degree of background reactivity was associated with the use of biotinylated antiserum. This appeared to be attributable to the binding of avidin-HRP to the wells of microtitre plates and to capture antiserum in ELISA, and to the non-specific binding of biotinylated rabbit-IgG-anti-LVGW antiserum to the nitrocellulose membrane in Western blots. A similar property of avidin has been described previously (Langley and Hillyer, 1989; Bayer and Wilchek, 1990; Ahluwalia et al., 1991). Therefore, biotinylated rabbit antiserum was replaced with HRP-labelled rabbit antiserum. This resulted in lower and uniform absorbance levels in ELISA when testing untreated sera or precipitates from PEG/EDTA treated sera, reflecting a significant decline in background reactivity, while maintaining a high binding affinity to LVGW and very little reactivity to OVAG and BPAG.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
313
The human sera were tested for specific circulating antigens in three different forms: without pre-treatment (to detect freely circulating antigens), and pretreated with PEG/EDTA (to precipitate and dissociate immune complexes) or TCA (to isolate glycoprotein antigens). Pre-treatment with PEG/EDTA (or EDTA alone) has previously successfully been used for detection of specific circulating antigens in people with onchocerciasis (Chandrashekar et al., 1990; Thambiah et al., 1991) and bancroftian filariasis (Weil and Liftis, 1987; Weil et al., 1987; More and Copeman, 1990), and specific circulating glycoprotein antigens have been detected in serum and urine of individuals with urinary schistosomiasis after pre-treatment of the samples with TCA (De Jonge et al., 1987; Krijger et al., 1994). Since circulating antigens have not been reported previously from dracunculiasis patients, the risk was realised from the beginning of the study that antigens released as excretory/secretory products from live worms or liberated from dissolving dead worms, were either not specific or would not reach the circulation in detectable amounts. Also, it was realised that antigenic release into the circulation would perhaps not take place throughout the entire prepatent and patent period and therefore could be missed. Study individuals representing both prepatent and patent infection categories were therefore included to provide optimal possibilities for detecting specific circulating antigens, provided they were there. However, by none of the methods applied, and in none of the D. medinensis infection categories examined, were circulating antigens detected. The possible explanations for the negative results may be that excreted/secreted products from D. medinensis, (1) are not antigenic, (2) are not specific, (3) are discharged in quantities too low to be detected, (4) are degraded in the host tissue and rapidly removed from the circulation, or (5) are discharged in amounts which vary highly between infected individuals and/or within each infected individual over time. A distinct difference in natural history also exist between infections with D. medinensis on the one hand, and O. 6ol6ulus and W. bancrofti on the other which may contribute to the differences in detectable antigenemia, namely that the first releases larvae into the external environment (i.e. water) whereas the latter release microfilariae into the tissue and lymph of their hosts. It has been estimated that between 21000 and 362000 O. 6ol6ulus microfilariae die each day per infected individual (Duke, 1993), leading to the release of huge amounts of a variety of proteins, many of which are antigenic and different from host proteins. Thus, at least for onchocerciasis, a major proportion of the circulating antigens probably originates from microfilariae. For dracunculiasis, offspring are not released in the host, and therefore mainly the adults will be potential candidates for antigenic discharge. The lack of success in the present study, using technical approaches similar to those used for the filarial nematodes, indicates that specific circulating D. medinensis antigens may not occur in detectable quantities in the serum of infected individuals. Development of a relatively simple diagnostic test based on detection of parasite antigen may therefore be unrealistic.
314
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
Acknowledgements The present study was carried out as part of an agreement between the Danish Bilharziasis Laboratory and the Ministry of Health of Ghana on the strengthening of the Guinea Worm Eradication Programme in the Northern Region of Ghana. We thank the Ministry of Health of Ghana for its logistic and technical support. Furthermore we are grateful to Dr S. Meredith, Royal Tropical Institute, The Netherlands, and Dr D.A. Denham, London School of Hygiene and Tropical Medicine, UK, for providing 0. 6ol6ulus and B. pahangi parasite material, respectively; Mette Lund for excellent technical assistance in the laboratory in Denmark. The study was sponsored by the Danish Bilharziasis Laboratory and the Danish International Development Agency (Danida).
References Ahluwalia, A., de Rossi, D., Ristori, C., Schirone, A., Serra, G., 1991. A comparative study of protein immunobilization techniques for optical immunosensors. Biosens. Bioelectron. 7, 207 – 214. Avrameas, S., Ternynck, T., 1971. Peroxidase labelled antibody and Fab% conjugates with enhanced intracellular penetration. Immunochemistry 8, 1175. Bayer, E.A., Wilchek, M., 1990. Application of avidin-biotin technology to affinity-based separations. J. Chromatogr. 510, 3–11. Bloch P., Simonsen P.E., 1998. Studies on immunodiagnosis of dracunculiasis. I. Detection of specific serum antibodies. Acta Trop. (In press). Bloch, P., Simonsen, P.E., Vennervald, B.J., 1993. The antibody response to Dracunculus medinensis in an endemic human population of Northern Ghana. J. Helminthol. 67, 37 – 48. Chandrashekar, R., Ogunrinade, A.F., Alvarez, R.M., Kale, O.O., Weil, G.J., 1990. Circulating immune complex-associated parasite antigens in human onchocerciasis. J. Infect. Dis. 162, 1159 – 1164. De Jonge, P., Fillie, Y.E., Deelder, A.M., 1987. A simple and rapid treatment (trichloroacetic acid precipitation) of serum samples to prevent non-specific reactions in the immunoassay of a protoglycan. J. Immunol. Methods 99, 195–197. Duke, B.O.L., 1993. The population dynamics of Onchocerca 6ol6ulus in the human host. Trop. Med. Parasitol. 44, 61–68. Fagbemi, B.O., Hillyer, G.V., 1990. Immunodiagnosis of dracunculiasis by falcon assay screening test-enzyme-linked immunosorbent assay (FAST-ELISA) and by enzyme-linked immunoelectrotransfer blot (EITB) technique. Am. J. Trop. Med. Hyg. 43, 665 – 668. Garate, T., Kliks, M.M., Cabrera, Z., Parkhouse, R.M.E., 1990. Specific and cross-reacting antibodies in human responses to Onchocerca 6ol6ulus and Dracunculus medinensis infections. Am. J. Trop. Med. Hyg. 42, 140–147. Hudson L., Hay F.C., 1989. Practical Immunology, 3rd ed. Blackwell Scientific, Oxford. Krijger, F.W., van Lieshout, L., Deelder, A.M., 1994. A simple technique to pretreat urine and serum samples for quantitation of schistosome circulating anodic and cathodic antigen. Acta Trop. 56, 55–63. Langley, R.J., Hillyer, G.V., 1989. Detection of circulating parasite antigen in murine fascioliasis by two-site enzyme-linked immunosorbent assays. Am. J. Trop. Med. Hyg. 41, 472 – 478. Lunde, M.N., Paranjape, R., Lawley, T.J., Ottesen, E.A., 1988. Filarial antigen in circulating immune complexes from patients with Wuchereria bancrofti filariasis. Am. J. Trop. Med. Hyg. 38, 366 – 371. More, S.J., Copeman, D.B., 1990. A highly specific and sensitive monoclonal antibody-based ELISA for the detection of circulating antigen in bancroftian filariasis. Trop. Med. Parasitol. 41, 403 – 406. Muller, R., 1971. Dracunculus and dracunculiasis. Adv. Parasitol. 9, 73 – 151.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
315
Thambiah, G., Whitworth, J., Hommel, M., Devaney, E., 1991. Identification and characterization of a parasite antigen in the circulating immune complexes of Onchocerca 6ol6ulus infected patients. Trop. Med. Parasitol. 43, 271–276. Weil, G.J., Liftis, F., 1987. Identification and partial characterization of a parasite antigen in sera from humans infected with Wuchereria bancrofti. J. Immunol. 138, 3035 – 3041. Weil, G.J., Jain, D.C., Santhanam, S., Malhotra, A., Kumar, H., Sethumadhavan, K.V.P., Liftis, F., Ghosh, T.K., 1987. A monoclonal antibody-based enzyme immunoassay for detecting parasite antigenemia in bancroftian filariasis. J. Infect. Dis. 156, 350 – 355.
.
Studies on immunodiagnosis of dracunculiasis. II. Search for circulating antigens Paul Bloch *, Birgitte J. Vennervald, Paul E. Simonsen Danish Bilharziasis Laboratory, Jaegersborg Alle 1D, DK-2920 Charlottenlund, Denmark Received 20 August 1997; received in revised form 11 February 1998; accepted 9 April 1998
Abstract Sera from individuals living in a dracunculiasis endemic area of northern Ghana were examined for circulating Dracunculus medinensis antigens by applying protocols previously developed for detection of circulating antigens in other helminth infections. Antisera from rabbits immunised with homogenized first stage D. medinensis larvae were used for antigen capture and detection in three different forms, namely non-treated, biotinylated and horseradish peroxidase (HRP)-labelled. Three different preparations of human sera were examined, namely non-treated, pre-treated with polyethylene glycol/ethylenediaminetetra-acetic acid (PEG/EDTA) for analysis of precipitated immune complexes, and pre-treated with trichloroacetic acid (TCA) for analysis of isolated glycoproteins. In both SDS-PAGE/Western blotting and ELISA, significant reactivity was observed between non-treated and treated rabbit-antisera on the one hand and non-treated and treated human sera on the other. However, no significant response differences were observed between sera obtained from individuals with dracunculiasis and non-endemic controls. The reasons are analysed and possible explanations presented. The study provided no evidence that D. medinensis-specific circulating antigens, detectable by relatively simple means, occur in infected individuals. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Circulating antigens; Immunodiagnosis; Dracunculiasis; Dracunculus medinensis
* Corresponding author. Tel.: +45 39 626168; fax: + 45 39 626121, e-mail: [email protected] 0001-706X/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S0001-706X(98)00039-4
304
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
1. Introduction The prepatent period in human Dracunculus medinensis infections, i.e. the period from ingestion of infective third stage larvae to emergence of the mature female worm on the skin, is 10 – 12 months (Muller, 1971). With exception of the last 1–2 weeks, when the worm becomes palpable, the prepatent period is asymptomatic. Currently, no reliable method for diagnosis of D. medinensis infection during the asymptomatic part of the prepatent period exists. By providing means for proper parasitological characterisation of people living in endemic areas, such a method might be useful in dracunculiasis research and control. Attempts have previously been made to develop a diagnostic method for prepatent infection with D. medinensis by detection of specific serum antibodies, but with different degrees of success (Fagbemi and Hillyer, 1990; Garate et al., 1990; Bloch et al., 1993; Bloch and Simonsen, 1998). Another approach would be detection of specific circulating antigens in serum from infected individuals. With the purpose of investigating the potential for developing a reliable method for diagnosis of dracunculiasis through detection of specific circulating antigens, this study carried out a search for such antigens in non-treated, PEG/EDTA-treated and TCA-treated sera from individuals with dracunculiasis. Protocols developed for detection of circulating antigens in individuals with bancroftian filariasis (Lunde et al., 1988) and onchocerciasis (Chandrashekar et al., 1990; Thambiah et al., 1991) were applied. Antibodies for antigen capture and detection were obtained from antiserum from rabbits immunised with a D. medinensis first stage larval homogenate.
2. Subjects and methods
2.1. Study indi6iduals Three categories of individuals were included in the study. Seven individuals (four males, three females; mean age 31 years, range 22–45 years) with a prepatent D. medinensis infection (category a) were from an area of northern Ghana which is highly endemic for dracunculiasis. The infection status of this study group was determined by closely following individuals for 15 months after blood sampling during which period emergence of worms on the skin was recorded. Seventeen individuals (six males, 11 females; mean age 31 years, range 8–50 years) with a patent D. medinensis infection (category b) were from the same endemic area of northern Ghana. Five Danish individuals (three males, two females; mean age 26 years, range 21 – 33 years) who had never been to tropical countries (category c) were used as non-endemic controls.
2.2. Preparation and pre-treatment of human serum Serum was recovered from venous blood samples after clotting and centrifugation, and sodium azide (NaN3) was added to a concentration of 15 mM prior to
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
305
freezing of the samples at − 80°C until use. In part of the serum, immune complexes were precipitated with polyethylene glycol/ethylenediaminetetraacetic acid (PEG/EDTA) (Lunde et al., 1988; Chandrashekar et al., 1990; Thambiah et al., 1991). This was done by mixing 200 ml serum, 200 ml 0.2 M Na2EDTA pH 7.5 and 100 ml 12% PEG 8000 (Sigma) pH 8.3. After overnight incubation at 4°C the solution was centrifuged at 15000× g for 30 min. The precipitate was washed in 500 ml cold 2.4% PEG in 0.1 M borate buffer (0.1 M Na2B4O7 · 10 H2O solution added to 0.1 M boric acid solution until pH 8.3 was obtained). The precipitate was dissociated by resuspending it in 100 ml 0.1 M Tris–HCl pH 2.6 and the solution was neutralised with 20 ml 1.0 M Tris pH 8.0. Another part of the serum was pre-treated with trichloroacetic acid (TCA) to dissociate immune complexes and to isolate the glycoproteins (De Jonge et al., 1987). This was done at room temperature by incubating 100 ml serum mixed with 100 ml 4% TCA solution for 5 min on a rocking table followed by centrifugation at 7000×g for 30 min. The supernatant was neutralised with 100 ml carbonate buffer pH 9.6 (0.174 M NaHCO3 and 0.070 M Na2CO3).
2.3. Antigens D. medinensis adult female worms and first stage larvae obtained from infected individuals from northern Ghana were the sources of antigen. The methods for recovery of parasites, and for preparation of crude homogenates of adult worms (ADGW) and larvae (LVGW) has been described previously (Bloch and Simonsen, 1998). For one experiment a pure adult worm antigen (puADGW) was made by carefully removing larvae from the uterus of female worms by milking and dissection before homogenizing the worms. Crude homogenates of adult Onchocerca 6ol6ulus (OVAG) and Brugia pahangi (BPAG), and human albumin (Sigma, St. Louis, MO, USA) prepared in various concentrations, were used as control antigens.
2.4. Rabbit-antiserum Two naive 3 – 4 months old rabbits were immunized sub-cutaneously with LVGW in Freund’s incomplete adjuvant (1:1) six times with 2-weeks interval. Antiserum was prepared from blood obtained 12 days after the last immunization. The IgG fraction was isolated by affinity purification on Protein A Sepharose CL-4B (Kabi Pharmacia Diagnostics, Sweden) on a Automated Econo System (Bio-Rad). The IgG eluate was immediately neutralised in 1 M Tris pH 8.0 in the proportion 1:5 before being buffer exchanged by overnight dialysis against PBS. The IgG fraction was concentrated in a Sartorius Ultrasart Cell 50 (Sartorius, Germany) to a protein concentration of 5.0 mg/ml. Biotinylation was carried out according to Hudson and Hay (1989) with the modification that 5.0 mg N-hydroxy-succinimidobiotin (Sigma, St. Louis, MO) was dissolved in 125 ml dimethyl sulfoxide (DMSO, Merck, Germany). While stirring,
306
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
10 ml of the biotin solution was added to 3.0 ml IgG fraction of the antiserum. Following dialysis against PBS overnight at 4°C, the protein content was measured to be 800 mg/ml. Sodium azide was added to 15 mM and the solution was kept frozen at − 20°C until use. HRP-labelling of the IgG fraction was carried out as previously described (Avrameas and Ternynck, 1971). Following centrifugation of the HRP-labelled conjugate at 10000×g for 30 min, the supernatant was filtered through a 0.22-mm sterile filter and the protein content was measured to be 4.4 mg/ml. Sodium azide was added to a concentration of 15 mM, and the solution was kept frozen at −20°C until use.
2.5. SDS-PAGE and Western blotting Normal resolution sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) as well as high resolution tricine-based SDS-PAGE, and Western blotting, were performed according to procedures described previously (Bloch et al., 1993; Bloch and Simonsen, 1998).
2.6. ELISA Circulating antigens were measured either by indirect enzyme linked immunosorbent assay (ELISA) (Bloch et al., 1993; Bloch and Simonsen, 1998) or by double antibody sandwich ELISA. For the sandwich ELISA, the wells of the microtitre plates (Maxisorp Immunoplate 442404, Nunc, Denmark) were coated with IgG rabbit-antiserum by overnight incubation at 4°C with 100 ml antiserum diluted in coating buffer (2.5 mM NaH2PO4 · 1 H2O, 7.5 mM Na2HPO4 · 12 H2O, 0.145 M NaCl, pH 7.2). After washing and blocking in PBS-Tween containing 0.5% bovine serum albumin, the wells were incubated overnight at 4°C with 100 ml diluted sample (pre-treated or non-treated human serum) applied in triplicate. All other parts of the sandwich ELISA procedure were similar to the indirect ELISA procedure.
3. Results
3.1. Reacti6ity of rabbit-antisera The reactivity of antisera from the two LVGW-immunised rabbits was analysed by SDS-PAGE/Western blotting. IgG antibodies in non-purified sera from both rabbits reacted to numerous and almost similar LVGW and ADGW antigens in the weight range 10 – 200 kDa. Limited and no reactivity was observed to BPAG and OVAG, respectively, with these sera. Pooled non-labelled, biotinylated or HRPlabelled IgG-purified rabbit antisera reacted to similar antigens in LVGW as the non-purified antisera. However, higher background reactivities were found for the non-labelled and biotinylated as compared with the HRP-labelled rabbit antiserum.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
307
Limited reactivity was observed against OVAG whereas essentially no reactivity was observed against BPAG. Both rabbits therefore responded to the D. medinensis antigens, with only little cross-reactivity to the tested non-D. medinensis antigens. The ability of the rabbit antisera to detect D. medinensis antigens in human serum was assessed in a sandwich ELISA. IgG-purified rabbit antiserum and HRP-labelled IgG-purified rabbit antiserum was used as capture and detecting antiserum, respectively, and test samples consisted of pooled serum from the non-endemic control individuals to which different concentrations of D. medinensis antigens were added (Fig. 1). For both untreated (Fig. 1A) and TCA pre-treated (Fig. 1B) samples to which LVGW, ADGW (containing uterus dwelling first stage larvae) or puADGW (purified for uterus dwelling first stage larvae) was added, the responses were rather similar and markedly stronger than the responses observed when OVAG, BPAG or no antigen was added.
3.2. Analyses for circulating antigens in untreated human sera Sandwich ELISA was first carried out to detect circulating D. medinensis antigens in untreated human sera using biotinylated rabbit antiserum for antigen detection. Relatively high absorbance values were observed for sera from patently D. medinensis infected individuals, but similar or even higher reactivities were seen for sera from non-endemic controls and for conjugate controls. No major improvements
Fig. 1. Sandwich ELISA showing the absorbance values of pooled non-treated (A) and TCA pre-treated (B) sera diluted 1:10 in PBS from five non-endemic control individuals to which LVGW (square), ADGW (triangle up), puADGW (triangle down), OVAG (diamond), BPAG (star) or no antigen (circle) was added in concentrations from 0.5 to 100 mg/ml. Samples were further diluted 1:10 in PBS prior to application. Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12 mg/ml and HRP-labelled rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration of 5 mg/ml.
308
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
Fig. 2. Indirect ELISA showing the absorbance values of non-absorbed (solid bar) or preabsorbed (/) biotinylated rabbit-IgG-anti-LVGW antiserum, and of HRP-labelled rabbit-IgG-anti-LVGW antiserum (¯), and avidin-HRP alone (open bar), against LVGW (5 mg/ml), rabbit-IgG-anti-LVGW antiserum (5 mg/ml), normal rabbit serum (5 mg/ml) and blocking agent alone, each of which were coated to the wells. All rabbit-antisera were used at a concentration of 1 mg/ml. All wells were blocked with 0.5% bovine serum albumin and 0.1% Tween 20. The asterisk represents an experiment not performed.
were observed when varying the dilution of serum (1:4, 1:10, 1:100 and 1:1000) or the concentration of capture (0.1, 1.0, 5.0 and 12.0 mg/ml) or detecting (0.1, 1.0 and 5.0 mg/ml) antiserum. Therefore, the reactivity of HRP-labelled and biotinylated rabbit-IgG-anti-LVGW antiserum to LVGW, non-labelled rabbit-IgG-anti-LVGW and normal rabbit serum was compared (Fig. 2). The biotinylated antiserum was tested without pre-treatment as well as preabsorbed in a Sepharose column coated with rabbit-IgG-anti-LVGW (to extract potential anti-idiotype antibodies). Nonabsorbed biotinylated rabbit-IgG-anti-LVGW antiserum reacted extensively with both LVGW and rabbit-IgG-anti-LVGW, and it reacted to a limited extent with normal rabbit serum. When using preabsorbed biotinylated antiserum, the reactivity of the eluate remained equally high against LVGW, but reduced to almost half the value against itself (i.e. rabbit-IgG-anti-LVGW), reaching a level equivalent to the level of reactivity of avidin-HRP alone. Non-specific binding to the wells blocked with bovine serum albumin was not observed. HRP-labelled rabbit-IgGanti-LVGW reacted extensively with LVGW, to a very limited extent with rabbitIgG-anti-LVGW and not at all with normal rabbit serum. Thus, in contrast to the rather extensive reactivities observed when using the biotinylated rabbit antiserum, only limited reactivities were observed when testing the HRP-labelled rabbit antiserum against normal rabbit antiserum or rabbit-IgG-anti-LVGW antiserum.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
309
In SDS-PAGE/Western blotting, using biotinylated rabbit-IgG-anti-LVGW antiserum as detecting antiserum, numerous bands were observed in the molecular weight range from 40 to 200 kDa. However, in accordance with the ELISA findings, no major response differences were observed between any of the three infection categories. To determine the specific factors contributing to the high background reactivity observed when testing the reactivity of biotinylated rabbit antiserum to LVGW, the individual assay reagents were applied in different combinations in ELISA (Table 1). A major conclusion from this experiment was that avidin-HRP alone strongly react with capture antiserum and the surface of the microtitre plate well. Whether detecting antiserum alone reacted non-specifically with capture antiserum could not be determined definitively. It therefore appeared impossible to use biotinylated IgG rabbit-anti-LVGW antiserum as detecting antiserum. An attempt was then made by ELISA to identify D. medinensis-specific circulating antigens in non-treated human sera by using HRP-labelled rabbit-IgG-antiLVGW antiserum as detecting antiserum. The rabbit-antiserum reacted extensively with LVGW, but the absorbance values for sera from prepatent and patent dracunculiasis patients were almost as low as those from non-endemic controls (Fig. 3). Thus, the experiment provided no evidence for the occurrence of specific circulating antigens in the sera.
Fig. 3. Absorbance values of a sandwich ELISA for detection of circulating D. medinensis antigens in non-treated pooled sera diluted 1:10 obtained from five individuals from each of the following groups: prepatent D. medinensis infection (a), patent D. medinensis infection (b), and non-endemic control (c). A conjugate control group prepared as a complete ELISA but without addition of test serum was included. Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12 mg/ml and HRP-labelled rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration of 5 mg/ml. The reactivity of HRP-labelled rabbit-IgG-anti-LVGW antiserum was tested against LVGW. Each bar represents the mean of tests carried out in triplicate.
− + − − − − + + + + + + − −
1 2 3 4 5 6 7 8 9 10 11 12 13 14 − − − − − +b − − − − +d − +e +d
Test serum − − + − + + + + − + − + + +
Detecting antiserum + + + + + + − + +c +c + − +c +
Avidin-HRP + + + + + + + + + + + + + +
Substrate
High (1000–2500) High (1000–2500) High (1000–2500) Zero Zero Zero Zero High (1000–2500) Medium (200–1000) Medium (200–1000) Medium (200–1000) Low (1–200) Low (1–200) High (1000–2500)
Absorbance level (value range ×103)
The test sera were either from individuals with a patent D. medinensis infection or from non-endemic controls (these were untreated and diluted 1:10). Rabbit-IgG-anti-LVGW antiserum was used as capture antiserum at a concentration of 12.0 mg/ml and biotinylated rabbit-IgG-anti-LVGW antiserum was used as detecting antiserum at a concentration 1.0 mg/ml. a Bovine serum albumin (0.5 and 2.0%) and 0.1% Tween 20 used as blocking agents all gave the same results. b Two serum samples from individual with patent D. medinensis infection were tested with the same results. c Streptavidin-HRP was also tested with the same result. d Two serum samples from individual with patent D. medinensis infection as well as two Danish controls were tested with the same results. e The serum sample was added before blocking agent; all samples (i.e. three from individuals with D. medinensis infection and three non-endemic controls) gave the same results.
− − − +a +a + − − + + + + + −
Capture antiserum Blocking
Test no.
Table 1 Summary of the results of a multi-factor ELISA experiment designed to show the effect on the absorbance value when excluding (−) or including (+) the different reagents of the assay
310 P. Bloch et al. / Acta Tropica 70 (1998) 303–315
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
311
Fig. 4. Absorbance values of an indirect ELISA for detection of circulating D. medinensis antigens in PEG/EDTA-treated pooled sera obtained from five individuals from each of the following groups: prepatent D. medinensis infection (a), patent D. medinensis infection (b), and non-endemic control (c). The dissociated immune complexes were diluted 1:1 and coated directly to the well of the microtitre plate. Rabbit-IgG-anti-LVGW antiserum in non-labelled (solid bar) and HRP-labelled (open bar) form was used as detecting antiserum at a concentration of 4.4 mg/ml. LVGW at a concentration of 5 mg/ml was included as a positive antigen control. HRP goat-anti-rabbit IgG conjugate diluted 1:2000 was used to detect non-labelled rabbit antiserum. Each bar represents the mean of tests carried out in triplicate.
3.3. Analyses for circulating antigens in PEG/EDTA precipitated immune complexes In a further attempt to search for D. medinensis-specific circulating antigens in the human sera, PEG/EDTA precipitates were dissociated and analysed by ELISA and SDS-PAGE/Western blotting. In ELISA, the dissociated precipitates were coated directly to microtitre plate wells following application of non-labelled or HRP-labelled rabbit-IgG-anti-LVGW antiserum for detection of circulating antigens (Fig. 4). Both rabbit-antisera reacted strongly with LVGW confirming their content of antibodies with specificity to the LVGW antigen homogenate. Furthermore, the HRP-labelled rabbit-antiserum reacted moderately and the non-labelled antiserum weakly with pre-treated sera from each study category. The experiment provided no evidence for the occurrence of D. medinensis circulating antigens in the sera. In SDS-PAGE/Western blotting, using HRP-labelled rabbit-IgG-anti-LVGW antiserum as detecting antiserum, numerous distinct bands were observed in the weight range from 20 to 150 kDa. However, in accordance with the ELISA findings, no major response differences were observed between the three study groups.
312
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
3.4. Analysis for circulating antigens in TCA pre-treated human sera TCA-isolated glycoproteins from sera from the three study groups were analysed for the occurrence of D. medinensis specific antigens in SDS-PAGE/Western blotting. The results showed that for sera from each of the study groups, three distinct and one faint band occurred on the strips, indicating that among the purified antigenic glycoproteins, none were specific to D. medinensis. Thus, purifying glycoproteins from the test sera provided no means for detecting D. medinensis specific circulating antigens. 4. Discussion By applying methodologies already developed for detection of specific circulating antigens in sera from individuals infected with O. 6ol6ulus or W. bancrofti (Lunde et al., 1988; Chandrashekar et al., 1990; Thambiah et al., 1991), this study aimed at detecting D. medinensis specific circulating antigens in the blood of infected individuals in either free form or in antigen-antibody complexes. To limit the extent of unspecific reactions with the capture and detecting antisera, the rabbits were immunised with D. medinensis first stage larvae homogenate (LVGW) and not adult female worm homogenate (ADGW) since we have previously observed epitopes resembling human albumins and immunoglobulins in ADGW (unpublished data). The rabbit antisera obtained were shown by SDS-PAGE/Western blotting to react extensively (in terms of staining intensity and number of bands) with LVGW no matter in which form they were tested (i.e. pooled or non-pooled, labelled or non-labelled). Their reactivity with dissolved human albumins and immunoglobulins was minimal (data not shown). In order to remove serum proteins which could interfere with the circulating antigen capture and detection process (e.g. low affinity rabbit IgM antibodies or rabbit serum albumin), the IgG fraction of pooled rabbit-antisera was extracted prior to use. Previous observations that human IgG antibodies react extensively with LVGW and ADGW justified the use of specific IgG antibodies for detection of circulating D. medinensis antigens in human sera (Bloch and Simonsen, 1998). The pooled IgG-purified rabbit-antiserum was tested in three different forms: non-labelled, labelled with biotin, and labelled with HRP. A high degree of background reactivity was associated with the use of biotinylated antiserum. This appeared to be attributable to the binding of avidin-HRP to the wells of microtitre plates and to capture antiserum in ELISA, and to the non-specific binding of biotinylated rabbit-IgG-anti-LVGW antiserum to the nitrocellulose membrane in Western blots. A similar property of avidin has been described previously (Langley and Hillyer, 1989; Bayer and Wilchek, 1990; Ahluwalia et al., 1991). Therefore, biotinylated rabbit antiserum was replaced with HRP-labelled rabbit antiserum. This resulted in lower and uniform absorbance levels in ELISA when testing untreated sera or precipitates from PEG/EDTA treated sera, reflecting a significant decline in background reactivity, while maintaining a high binding affinity to LVGW and very little reactivity to OVAG and BPAG.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
313
The human sera were tested for specific circulating antigens in three different forms: without pre-treatment (to detect freely circulating antigens), and pretreated with PEG/EDTA (to precipitate and dissociate immune complexes) or TCA (to isolate glycoprotein antigens). Pre-treatment with PEG/EDTA (or EDTA alone) has previously successfully been used for detection of specific circulating antigens in people with onchocerciasis (Chandrashekar et al., 1990; Thambiah et al., 1991) and bancroftian filariasis (Weil and Liftis, 1987; Weil et al., 1987; More and Copeman, 1990), and specific circulating glycoprotein antigens have been detected in serum and urine of individuals with urinary schistosomiasis after pre-treatment of the samples with TCA (De Jonge et al., 1987; Krijger et al., 1994). Since circulating antigens have not been reported previously from dracunculiasis patients, the risk was realised from the beginning of the study that antigens released as excretory/secretory products from live worms or liberated from dissolving dead worms, were either not specific or would not reach the circulation in detectable amounts. Also, it was realised that antigenic release into the circulation would perhaps not take place throughout the entire prepatent and patent period and therefore could be missed. Study individuals representing both prepatent and patent infection categories were therefore included to provide optimal possibilities for detecting specific circulating antigens, provided they were there. However, by none of the methods applied, and in none of the D. medinensis infection categories examined, were circulating antigens detected. The possible explanations for the negative results may be that excreted/secreted products from D. medinensis, (1) are not antigenic, (2) are not specific, (3) are discharged in quantities too low to be detected, (4) are degraded in the host tissue and rapidly removed from the circulation, or (5) are discharged in amounts which vary highly between infected individuals and/or within each infected individual over time. A distinct difference in natural history also exist between infections with D. medinensis on the one hand, and O. 6ol6ulus and W. bancrofti on the other which may contribute to the differences in detectable antigenemia, namely that the first releases larvae into the external environment (i.e. water) whereas the latter release microfilariae into the tissue and lymph of their hosts. It has been estimated that between 21000 and 362000 O. 6ol6ulus microfilariae die each day per infected individual (Duke, 1993), leading to the release of huge amounts of a variety of proteins, many of which are antigenic and different from host proteins. Thus, at least for onchocerciasis, a major proportion of the circulating antigens probably originates from microfilariae. For dracunculiasis, offspring are not released in the host, and therefore mainly the adults will be potential candidates for antigenic discharge. The lack of success in the present study, using technical approaches similar to those used for the filarial nematodes, indicates that specific circulating D. medinensis antigens may not occur in detectable quantities in the serum of infected individuals. Development of a relatively simple diagnostic test based on detection of parasite antigen may therefore be unrealistic.
314
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
Acknowledgements The present study was carried out as part of an agreement between the Danish Bilharziasis Laboratory and the Ministry of Health of Ghana on the strengthening of the Guinea Worm Eradication Programme in the Northern Region of Ghana. We thank the Ministry of Health of Ghana for its logistic and technical support. Furthermore we are grateful to Dr S. Meredith, Royal Tropical Institute, The Netherlands, and Dr D.A. Denham, London School of Hygiene and Tropical Medicine, UK, for providing 0. 6ol6ulus and B. pahangi parasite material, respectively; Mette Lund for excellent technical assistance in the laboratory in Denmark. The study was sponsored by the Danish Bilharziasis Laboratory and the Danish International Development Agency (Danida).
References Ahluwalia, A., de Rossi, D., Ristori, C., Schirone, A., Serra, G., 1991. A comparative study of protein immunobilization techniques for optical immunosensors. Biosens. Bioelectron. 7, 207 – 214. Avrameas, S., Ternynck, T., 1971. Peroxidase labelled antibody and Fab% conjugates with enhanced intracellular penetration. Immunochemistry 8, 1175. Bayer, E.A., Wilchek, M., 1990. Application of avidin-biotin technology to affinity-based separations. J. Chromatogr. 510, 3–11. Bloch P., Simonsen P.E., 1998. Studies on immunodiagnosis of dracunculiasis. I. Detection of specific serum antibodies. Acta Trop. (In press). Bloch, P., Simonsen, P.E., Vennervald, B.J., 1993. The antibody response to Dracunculus medinensis in an endemic human population of Northern Ghana. J. Helminthol. 67, 37 – 48. Chandrashekar, R., Ogunrinade, A.F., Alvarez, R.M., Kale, O.O., Weil, G.J., 1990. Circulating immune complex-associated parasite antigens in human onchocerciasis. J. Infect. Dis. 162, 1159 – 1164. De Jonge, P., Fillie, Y.E., Deelder, A.M., 1987. A simple and rapid treatment (trichloroacetic acid precipitation) of serum samples to prevent non-specific reactions in the immunoassay of a protoglycan. J. Immunol. Methods 99, 195–197. Duke, B.O.L., 1993. The population dynamics of Onchocerca 6ol6ulus in the human host. Trop. Med. Parasitol. 44, 61–68. Fagbemi, B.O., Hillyer, G.V., 1990. Immunodiagnosis of dracunculiasis by falcon assay screening test-enzyme-linked immunosorbent assay (FAST-ELISA) and by enzyme-linked immunoelectrotransfer blot (EITB) technique. Am. J. Trop. Med. Hyg. 43, 665 – 668. Garate, T., Kliks, M.M., Cabrera, Z., Parkhouse, R.M.E., 1990. Specific and cross-reacting antibodies in human responses to Onchocerca 6ol6ulus and Dracunculus medinensis infections. Am. J. Trop. Med. Hyg. 42, 140–147. Hudson L., Hay F.C., 1989. Practical Immunology, 3rd ed. Blackwell Scientific, Oxford. Krijger, F.W., van Lieshout, L., Deelder, A.M., 1994. A simple technique to pretreat urine and serum samples for quantitation of schistosome circulating anodic and cathodic antigen. Acta Trop. 56, 55–63. Langley, R.J., Hillyer, G.V., 1989. Detection of circulating parasite antigen in murine fascioliasis by two-site enzyme-linked immunosorbent assays. Am. J. Trop. Med. Hyg. 41, 472 – 478. Lunde, M.N., Paranjape, R., Lawley, T.J., Ottesen, E.A., 1988. Filarial antigen in circulating immune complexes from patients with Wuchereria bancrofti filariasis. Am. J. Trop. Med. Hyg. 38, 366 – 371. More, S.J., Copeman, D.B., 1990. A highly specific and sensitive monoclonal antibody-based ELISA for the detection of circulating antigen in bancroftian filariasis. Trop. Med. Parasitol. 41, 403 – 406. Muller, R., 1971. Dracunculus and dracunculiasis. Adv. Parasitol. 9, 73 – 151.
P. Bloch et al. / Acta Tropica 70 (1998) 303–315
315
Thambiah, G., Whitworth, J., Hommel, M., Devaney, E., 1991. Identification and characterization of a parasite antigen in the circulating immune complexes of Onchocerca 6ol6ulus infected patients. Trop. Med. Parasitol. 43, 271–276. Weil, G.J., Liftis, F., 1987. Identification and partial characterization of a parasite antigen in sera from humans infected with Wuchereria bancrofti. J. Immunol. 138, 3035 – 3041. Weil, G.J., Jain, D.C., Santhanam, S., Malhotra, A., Kumar, H., Sethumadhavan, K.V.P., Liftis, F., Ghosh, T.K., 1987. A monoclonal antibody-based enzyme immunoassay for detecting parasite antigenemia in bancroftian filariasis. J. Infect. Dis. 156, 350 – 355.
.