PRODUCTION AND CHARACTERIZATION OF MONOCLONAL ANTIBODIES AGAINST THE EXCRETORY-SECRETORY ANTIGEN THE LIVER FLUKE (OPrST~O~C~IS VrVE~~~~I) S. AMORNPUNT,* *Department

S. SARASOMBATH~

OF

and S. SIRISINHA*$

of Microbiology, Faculty of Science and TDepartment of Microbiology, Siriraj Hospital, Mahidol University, Bangkok, Thailand

Faculty

of Medicine,

(Received 8 August 1990; accepted 27 December 1990) S. and SIRISINHA S. 1991. Production and characterization of against the excretory-secretory antigen of the liver fluke (Opisthorchis viverrini). InternationaiJournaifor Parasitology 21: 421428. Monoclonal antibodies (MoAb) were produced against a major soluble metabolic product (excretory-secretory, ES) of Upisrhorchis viverrini. The latter was obtained in a form of spent cufture medium in which the adult flukes had been maintained in vitro. The MoAbproduced wereexclusively associated with either IgG or IgM isotypes. When screened against a panel of parasite antigens by indirect ELISA, these MoAb exhibited three patterns of reactivity. Approximately 50% of the MoAb were highly specific for 0. viverrini and another 25% cross-reacted only with Ctonorchis sinensis. The remaining 25% cross-reacted extensively with other parasites. Results from radioimmunoprecipitation and immunoblotting experiments showed all MoAb to react with the 89-kDa glycoprotein. By indirect immunofluorescence, these MoAb reacted almost exclusively with the tegumental surface, tegumental cells, cecum and developing miracidium. Different lines of evidence suggest that these MoAb reacted with different epitopes on the same 89-kDa ~l~ptide carrier. AbStr#d--AMORNPUNT

monoclonal

INDEX product;

S., SARASOMBATH

antibodies

KEY WORDS: Opisthorchis viverrini; liver fluke; monoclonal excretory-secretory antigen; opisthorchiasis.

ES antigen;

metabolic

tegument was recently shown to be highly speciesspecific and gave a satisfactory result for the diagnosis of opisthorchiasis (Poopyruchpong, Viyanant, Upatham & Srivatanakul, 1990). The availability of monoclonal antibodies against the antigens of this liver fluke is particularly useful in the characterization of these antigens. We recently described the production of MoAb against the 90-kDa antigen associated with the muscular system and the I6-kDa antigen of the tegument (Billings, Utsakhit & Sirisinha, 1990). In the present study, we described MoAb against the major soluble metabolic product of the fluke and attempted to identify the reactive epitopes by radioimmunoprecipitation, immunoblotting and immunofluorescence.

INTRODUCTION

fluke infection caused by ~~isth~rch~~viverrini is still an important public health problem in many parts of the world. Current methods for the diagnosis of the disease are based on stool examination for characteristic eggs. However, the method is highly variable, particularly in cases with light infection, demands highly trained technicians, and is invariably time cons~ing. False positive diagnoses are not uncommon, particularly in patients concomitantly infected with small intestinal flukes whose eggs very much resemble those of 0. viverrini. A number of serodiagnostic methods have been developed but none has been found to be entirely satisfactory in terms of specificity and sensitivity (Sirisinha, 1986). Knowledge of indi~dual antigens, their abundance, and their immunogenicity is needed if one is to develop a proper immunodiagnostic method for routine use in diagnostic laboratories. We have previously demonstrated the presence of an 89-kDa metabolic antigen that appeared to have potential for further investigation (Wongratanacheewin, Chawengkirttikul, Bunnag& Sirisinha, 1988; Wongratan~h~~n, Bunnag, Vaeusorn & Sirisinha, 1988; Sirisinha, Sahassananda, Bunnag & Rim, 1990). An extract of the surface LIVER

$ To whom all correspondence

antibody;

MATERIALS

AND METHODS

Antigenpreparation. Both somatic and metabolic (excretorysecretory, ES) antigens of 0. viverrini and other parasites were prepared as previously described (Wongratanacheewin, Chawengkirttikul, Bunnag& Sirisinha, 1988). The immunogen used for monclonal antibody production was prepared from a spent culture medium in which the adult flukes had been m~ntain~ in vitro. The fluid was concentrated by ultrafiltration using a PM-10 membrane filter and then dialyzed against several changes of 0.85% NaCl. Production of monoclonal antibodies (MoAb). BALB/c mice, 68 weeks of age, used for MoAb production were

should be addressed. 421

S. AMORNPUNT.S.SARASOMBATH~~~

422

kindly provided by the Division of Veterinary Medicine, Armed Forces Research Institute of Medical Sciences, Bangkok, Thailand. The animals were immunized by an intraperitoneal injection of 25 pg of concentrated ES mixed with an equal volume of complete Freund’s adjuvant. One month after the first injection, the animals were bled from the tail vein and the serum obtained was tested for antibody by enzyme-linked immunosorbent assay (Wongratanacheewin, Bunnag, Vaeusom & Sirisinha, 1988). Three days prior to cell fusion, the high responders received a booster injection of 50 fig of antigen in normal saline by an intraperitoneal route. The poor responders were injected intraperitoneally with the antigen mixed with an equal volume of incomplete Freund’s adjuvant approximately 1 month after the first injection, subsequently bled and the serum again tested for anti-ES. Those showing satisfactory responses were finally given a pre-fusion booster as before and sacrificed 3 days later. Splenic cells taken from these animals were fused with the P3 x 63Ag8.653 mouse myeloma cells using polyethylene glycol PEG-4000 (Accurate Chemical and Scientific Co., NY, U.S.A.) as previously described (Sarasombath, Lertmongkolchai & Banchuin, 1988). The hybrids were cultured, identified, cloned by a single dilution method and selected clones were expanded in 50 ml tissue culture flasks. Immunological techniques. Radioimmunoprecipitation, immunoblotting, indirect enzyme-linked immunosorbent assay (ELISA) and indirect immunofluorescence were performed essentially as described in previous communications (Wongratanacheewin, Bunnag, Vaeusorn & Sirisinha, 1988: Wongratanacheewin, Chawengkirttikul, Bunnag & Sirisinha, 1988; Sarasombath et al., 1988; Billings et al., 1990). Other techniques. Protein concentrations of the various parasite preparations were determined by a Folin method using bovine serum albumin as standard (Lowry, Rosebrough, Farr & Randall, 1951). RESULTS

Recovery of’antibody-secreting hybridomas A large number of hybridomas secreting antibody reactive with 0. viverrini were obtained from five identical fusions (Table I). Supernatant fluids from hybridomas grown in 24-well plates were screened for antibody activity by indirect ELISA, using crude somatic extract of adult 0. viverrini at a concentration of IO pg ml-‘. An optical density (O.D.) reading of at least 0.2 units (492 nm) above the background was taken as a criterion for positive antibody secretion. Results presented in Table 1 showed that the percentage of antibody secreting wells varied considerably from one fusion to another. Positive clones were subsequently tested for activity against the ES antigen also by indirect ELISA. Most of these clones

TABLE 2-ISOTYPESOF 0.

Reactive

TABLE

S.SIRISINHA I-FUSION

Fusion no.

64/240 (26) Fibroblast overgrowth 288/288 (100)

ES-IV ES-V

100 100

571264 (22) 138/168 (82)

* Screened by indirect ELISA using crude somatic 0. viverriniantigens. An optical density value (292 nm) of > 0.3 was used as a cut-off criterion.

were found to give a much higher O.D. reading with the ES antigen than with the somatic antigen. Clones with an O.D. reading higher than 0.5 against the 0. viverrini ES antigen were analyzed for specificity by an indirect ELISA using a panel of parasite antigens prepared from Clonorchis sinensis, Schistosoma mansoni, Fasciola gigantica, Gigantocotyl siamensis and Gnathostoma spinigerum. Results from this experiment showed that the 23 clones from the 23 original wells that were screened exhibited three patterns of reactivity (Table 2). While 13 of the 23 clones were found to be highly specific for 0. viverrini antigen (Group I), the remaining IO clones crossreacted to varying degrees with the antigen from its close relative C. sinensis (Groups 2 and 3). Of these 10 cross-reactive clones, four were found to cross-react also with the antigens prepared from other parasites (Group 3). It should be mentioned that one of these four clones reacted equally strongly with all antigens used for specificity testing. One interesting observation when these monoclonal antibodies were typed was the distribution of the heavy chain isotypes which were found to exclusively belong to the IgG, and IgM classes (Table 2). The presence of IgM in approximately 40% of the MoAb was unexpected in view of the fact that these were taken from animals that had been hyperimmunized with the parasite antigen prior to the time of fusion. One additional unusual finding was that with the exception of the one Igh4-producing clone in Group I, all other IgM producers belonged to the cross-reactive clones in Groups 2 and 3 (Table 2). On the other hand, regardless of their specificity and heavy chain isotype, all of these MoAb were found to possess kappa light chains.

ANTIBODYSECRETING HYBR~DOMAS

from Isotypes

Opisthorchis

Clonorchis

+ + +

+ +

_

* Schistosoma, Fasciola, Gigantocotyl,

No. of secreting hybridomas*/total (%)

100 100 100

Group I 2 3

% Fusion efficiency

ES-I ES-II ES-III

viverrini-specwlc ANDCROSS-REACTIVE with antigen

EFFICIENCYAND PERCENTAGE OF ANTIBODY SEcRETlNGHyBRlDOMAS

Total

Others* _ _ + Gnathostoma

12 W,(k),

1 IgM (kappa)

2 IgG, (k), 4 IgM (kappa) 4 IgM (kappa)

13 6 4

Monoclonal

antibodies

against

IdentiJcation of immunoreactive component by radioimmunoprecipitation and immunoblotting Monoclonal antibodies with strong ELISA activity for 0. viverrini ES antigen from seven original clones (Table 3) were characterized further by radioimmunoprecipitation and immunoblotting. The autoradiographic patterns of labelled 0. viverrini ES antigen showed that both the polyclonal pre-fusion antiserum and the seven MoAb immunoprecipitated the same antigenic component that migrated to the 89kDa position in SDS-PAGE (Fig. 1, Table 3). A similar radioimmunoprecipitating pattern was observed when these MoAb were ailowed to react with the crude somatic antigen previousiy shown to contain trace quantities of the 89-kDa component (data not TABLE

3---RADlOlMMUNOl’REClPlTATION

Category 0. viverrini-specific

Cross-reactive

Opisihorc~~ ES antigen

shown). However, with the somatic antigen preparation, the polyclonal antibodies in the pre-fusion antiserum reacted with three additional components that migrated at the 52-, 20- and 1617-kDA positions. Unlike the radioimmunoprecipitating patterns, the immunoblotting patterns developed with these seven MoAb appeared to be more diverse (Fig. 2, Table 3). The one major difference in the results obtained by these two techniques was that while with the former an identical pattern was noted with all seven MoAb, with the latter technique some MoAb failed to react (Table 3). Four of the five immunoblot-positive MoAb gave a single discrete band also at the 89-kDa position. The 4A, MoAb, however, gave a definite immunoblot staining that smeared from approximately 40-kDa to

AND IMMUNOBLOTTlNG

Original clones

Isotype

6B,

IgG,

4A, 7C5 4D5

IgG, IgG, IgM

6’2,

IgM

3A, 14

IgM IgM

423

PATTERNS OF SEVEN MONOCLONAL

Radioimmunoprecipitation

Single discrete band at 89 kDa

1

ANTIBODIES

Immunoblotting Single discrete band at 89 kDa Smearing between 40 and 89 kDa Not detected Single discrete band at 89 kDa Not detected Single discrete band at 89 kDa Singte discrete band at 89 kDa

mol .wt

-89

24.018.4 14.3 Es

IMSFBSA

BCDEFG

Ro. 1. Autoradiographic patterns of ‘2SI-labelled0. viverrini ES antigen before (ES*) and after being immunoprecipitated with monoclonal antibodies (A-G). Immune mouse serum (IMS) and PBS served as positive and negative controls, respectively. Molecular

weight markersare

shown on theleft.

Lane A = 6B,, B = 4A,, C = 7C,, D = 4D,, E = 6C,, F = 3A, and G = lD,.

424

S. AMORNPUNT,S.SARASOMBATH~~~S.SIRISINHA

89-kDa positions (lane B, Fig. 2), regardless of whether the somatic or the ES antigens were used for analysis. Moreover, with the somatic antigen, the lDz MoAb reacted not only with the 89-kDa component, but to a lesser extent also with the components migrating at the 85 and 82-kDa positions (data not shown). Localization of the immunoreactive 89-kDa component by indirect immunojluorescence All seven MoAb were used for the localization of the immunoreactive 89-kDa component by the indirect

immunofluorescent technique. Frozen sections of adult 0. viverrini taken from experimentally infected hamsters were acetone-fixed at room temperature for 5 s, exposed to appropriate concentrations of MoAb and then processed as previously described (Billings et al., 1990). Representative staining patterns of 0. viverrini-specific MoAb (6B,, 7C, and 4D,) and crossreactive MoAb (6D,, 3A, and ID,) are shown in Figs. 3 and 4, respectively. Unlike the radioimmunoprecipitating and immunoblotting patterns, the immunofluorescentstaining patterns developed with 0. viverrini-specific

TABLE ~IMMUNOLOGICALCHARACTER~ST~CS~FMONOCLONALANTIBODIESTO 0. viverrini ES ANTIGEN

Reactive with 89 kDa

Immunofluorescent staining

Monoclonal antibodies

0. viverrini-specific

Cross-reactive

C, Cecum and its content; on external surface.

6B, 4A, 7C5 4D5 6C, 3A, lD, T, tegument;

Radioimmunoprecipitation

Immunoblotting

C

T

TC

M

Mi

E

+ + + + + + +

+ Smearing _

+ -

+ + + + + +

+ + + + + +

+ + +

+ + + + + + +

f f + + +

TC, tegumental

+ _ + +

cell; M, muscle; Mi, miracidium;

E, egg shell/adhering material

mol .wt

. -73.40

66.0 -

24.0 18.4 14.3 ES

IMS MM

A

B

D

F

G

FIG. 2. Immunoblotting patterns of 0. viverrini ES antigen electrophoresed in the presence of SDS, electroblotted onto a nitrocellulose membrane, allowed to react with monoclonal antibodies (A-G) and developed with phosphate-conjugated goat anti-mouse IgG. Immune mouse serum (IMS) served as the positive control and myeloma condition medium (MM) served as the negative control. See legend to Fig. 1 for other explanations.

Monoclonal

antibodies

agai .nst Opisthorchis ES antigen

MoAb (Fig. 3) were not only markedly different from those developed with cross-reactive MoAb (Fig. 4), but also showed some variation within the group. For example, the staining with 6B, was markedly different from that produced by the remaining three 0. viverrini-specific MoAb (4A,, 7C, and 4DJ. With the 6B,, the fluorescence was noted exclusively in the cecum and its contents (panels 1 and 2, Fig. 3). On the other hand, the remaining three MoAb in this group

425

stained the parasite surface brightly and uniformly throughout the entire thickness of the tegument (panels 4-9, Fig. 3) and failed to stain the cecum (Table 4). With these MoAb, the tegumental cells underneath the unstained muscular layer also fluoresced brightly. In addition to these structures, thin fluorescent streaks traversing the muscular layer could be noted on occasions (panel 4, Fig. 3). The latter most likely represented the cytoplasmic stalks connecting the

FIG. 3. Indirect immunofluorescence localization of immunoreactive 89-kDa component on freshly frozen sections of adult 0. viverrini stained with species-specific MoAb 6B, (panels l-3), 7C, (panels 4-6) and 4D, (panels 7-9). Cecum (C, panel 1) and secretion in lumen (lm, panel 2) stained brightly with 6B,. Surface tegument (t) and tegumental cells (tc) also fluoresced with 7C, and 4D, (panels 4,5,7-9). Thin fluorescence streaks traversing the unstained muscular layer(m) should be noted (panel 4). Panels 3 and 6 showed miracidium staining inside the autofluorescence of egg shell. OS, oral sucker. Scale bar = 100 nm.

426

S. AMORNPUNT,~.

SARASOMBATH~~~S.SIRISINHA

FIG. 4. Indirect immunofluorescence locatization of immunoreactive 89-kDa component

on freshly frozen sections of adult 0. viverrini stained with cross-reacting MoAb 6C, (panels 1 and 2), 3A, (panels 3 and 4) and ID, (panels 5 and 6). The 6C, stained brightly tegumental surface (t), tcgumental cells (tc) and muscle (m). Fluorescent staining ofdeveloping miracidium (panel 2) together with egg shell and/or gelatinous material adhering to the external surface (panels 3 and 4) should be noted. Scale bar = IOOnm.

tegumental cell proper with the superficial tegumental surface layer. Unlike the 0. viverrini-specificMoAb, all three crossreactive MoAb exhibited similar immunofluorescent staining (Fig. 4). As shown in the figure, in addition to

the staining of the surface tegument and tegumental cell, these MoAb also reacted with epitopes associated with peripheral muscle fibers (panels 1,5 and 6, Fig. 4). However, the relative intensity of staining of these structures varied from one MoAb to another.

Monodonal

antibodies against Opisthorchis ES antigen

In addition to staining the somatic components, all seven MoAb stained, to varying degrees of intensity, the miracidium developing inside the eggs (Figs. 3 and 4, Table 4). However, in addition to the miracidium, there was a tendency for cross-reactive MoAb to stain the shell and/or gelatinous materials adhering to its external surface (panels 2 and 4, Fig. 4). DISCUSSION

Results presented in this study show that the monoclonal antibodies obtained from mice immunized with soluble ES antigen of 0. viverrini recognized specific and cross-reactive epitopes present on an immunoreactive 89-kDa glycoprotein component secreted by the parasite. Although based on molecular size, the 89-kDa component recognized by the MoAb in this study would be difficult to distinguish from the 90-kDa component recognized by the MoAb reported previously (Billings et al., 1990), different lines of evidence presented herein suggest that these two components represent two separate entities. Firstly, with the exception of the 6B,, all other MoAb described in the present study stained components present almost exclusively in the tegument and tegumental cells and not in the peripheral muscular layers recognized by the MoAb in the previous study (Billings et al., 1990). Secondly, the developing miracidium could be stained only with the MoAb against this soluble metabolic product. It should be recalled that the polyclonal rabbit anti-ES was previously shown to react also with the miracidium (Wongratanacheewin & Sirisinha, 1987). Thirdly, the MoAb from clones, 6B,, 4A,, 7C, and 4D, reacted exclusively with 0. viverrini while all anti-90 kDa muscular-associated MoAb cross-reacted with C. sinensis. The weaker muscular staining of the crossreactive MoAb from clones 6C,, 3A, and ID,, on the other hand, somewhat resembled but was not identical with that of the anti-90 kDa MoAb in the previous study (Billings ef al., 1990). The immunofluorescent pattern of 0. viverrinispecific MoAb 4A,, 7C, and 4D, was clearly different from the anti-16 kDa MoAb reported previously although both reacted positively with the tegumental layer. The MoAb which immunoprecipitated the 16kDa somatic antigen described earlier gave a bright immunofluorescent staining at the very surface of the tegument but with decreasing intensity through the bulk of the tegument (Billings et al., 1990). The 0. viverrini-specific MoAb in the present study not only reacted uniformly throughout the entire tegumental layer, but also reacted with tegumental cells underneath. This staining reaction gave it a characteristic pattern of the unstained muscular layer sandwiched in between the brightly stained surface tegument and tegumental cell layer. These results suggest that these two groups of tegumental reactive MoAb recognized two distinct parasite antigens. Moreover, the staining pattern noted with the anti-16 kDa tegumental somatic antigen suggests that the 16-kDa immuno-

427

reactive component is not sy.lrl-rsizcd in situ but is probably absorbed from the outside via the surface tegument. The staining pattern of the anti-89 kDa metabolic product on the other hand suggests that the immunoreactive component is synthesized by the tegumental cells. The thin streaks crossing the unstained muscular layer noted on occasions (panel 4, Fig. 3) is suggestive of the intracellular transport of this glycoprotein from the tegumental cell to the tegumental surface. The results from radioimmunoprecipitation and immunoblotting showed that, regardless of the epitope specificity, the MoAb to the ES antigen reacted with the same 89-kDa glycoprotein. The inability to detect even a trace amount of any second labelled component by radioimmunoprecipitation supports the above conclusion. Although a direct inhibition study to more precisely identify the reactive epitopes has yet to be performed, circumstantial evidence from the immunofluorescent experiment suggests that these MoAb recognized different epitopes. Epitopes on the 89-kDa glycoprotein can be associated with either the carbohydrate side chain or with the polypeptide carrier. With regard to the latter, one can visualize them as being in the form of conformational or sequential types. It is tempting to predict from the species-specificity pattern of these MoAb that the epitopes recognized by the three crossreactive MoAb (6A,, 3A, and ID,) are associated with the carbohydrate moiety as the latter can be expected to be more ubiquitous among the various organisms. Moreover, the IgM nature of these cross-reactive MoAb is also consistent with the general nature of the antibodies to polysaccharide antigens. The epitopes recognized by 0. viverrini-specific MoAb (6B,, 4A,, 7C, and 4D,), on the other hand, should be logically suspected to be associated with the protein moiety. From the immunofluorescent pattern, these epitopes were associated almost exclusively with the tegumental component(s) of the parasite. Species specificity of the surface tegument has also been reported very recently (Poopyruchpong et al., 1990). The failure of the 7C, MoAb to give a positive immunoblot suggests that it is associated with a conformation epitope. On the other hand, it is more difficult to explain why the 6B, stained almost exclusively the cecum and not the tegument. It is possible that the epitope recognized by this MoAb is hidden in the native molecule and is somehow exposed in the cecal environment. The staining of miracidium by all anti-89 kDa MoAb indicates that this secreted 89-kDa component is present during the early stage of development. This is in marked contrast to the MoAb against the muscleassociated 90-kDa antigen described in the previous study (Billings et al., 1990) which appeared not to react with this immature stage. The data present in the current study as well as in the previous reports clearly indicate that the 89-kDa glycoprotein is an integral component of the tegument and is also being secreted

428

S. AMORNPUNT,S.SARASOMBATH~~~S.SIRISINHA

in large quantities into its surrounding (Billings et al., 1990; Wongratanacheewin & Sirisinha, 1987). The latter may act as a decoy released to protect the parasite against attack by the host immune defences, thus allowing it to survive for a long time in the biliary system of a definitive host. The presence of antibodies against the 89-kDa glycoprotein in the bile has been reported previously (Wongratanacheewin, Bunnag, Vaeusorn & Sirisinha, 1988). We are now in the process of using these MoAb for the quantitation of the amount of 89-kDa glycoprotein secreted by adult worms maintained in vitro as well as that released in the bile of experimentally infected hamsters. This information should provide us with a more detailed understanding of host-parasite interaction for this infection. An attempt is also being made to use these MoAb to develop a simplified and specific serological method for the detection of 0. viverrini infection in humans (Sirisinha, Chawengkirttikul, Sermswan, Amornpant, Mongkolsuk & Panyim, unpublished). authors are grateful to Mrs Pranee Sithisarn and Dr H. J. Rim (Korea University, Seoul, Korea) Acknowledgements-The

for some technical advice and for making Clonorchis sinensis available for this study, respectively. This investigation received financial support from the U.S. Agency for International Development Grant No. 936-5542-G-00-6027-00.

REFERENCES BILLINGSP. B., UTSAKHITN. & SIRISINHAS. 1990. Monoclonal antibodies against Opisthorchis viverrini antigens. Parasite

Immunology12:545-557. LOWRY 0. H., ROSEBROUGHN. J., FARR A. L. & ILZNDALLR. S. 1951. Protein measurement with the Folin phenol reagent. Journal of Biological Chemistry 193: 265-275. POOPYRUCHPONG N., VIYANANT V., UPATHAM E. S. & SRIVATANAKULP. 1990. Diagnosis of opisthorchiasis by enzyme-linked immunosorbent assay using partially purified antigens. Asian Pa@ Journal of Allergy and Immunology 8: 27-3 1. SARASOMBATH S., LERTMONGKOLCHAI G. & BANCHUINN. 1988. Characterization of monoclonal antibodies to protein antigen of Salmonella typhi. Journal of Clinical Microbiology 26: 508-5 12. SIRISINHA S. 1986. Immunodiagnosis of human liver fluke infections. Asian Pacific Journal of Allergy and Immunology 4: 81-88. SIRISINHAS., SAHASSANANDA D., BUNNAGD. & RIM H. J. 1990. Immunological analysis of Opisthorchis and Clonorchis antigens. Journal of Helminthology 64: 133-138. WONGRATANACHEEWIN S. & SIRISINHA S. 1987. Analysis of Opisthorchis viverrini antigens: physicochemical characterization and antigen localization. Southeast Asian Journal of Trouical Medicine and Public Health 18: 5 1I520. ” . WONGRATANACHEEWINS., BUNNAG D., VAEUSORN N. & SIRISINHA S. 1988. Characterization of humoral immune response in the serum and bile of patients with opisthorchiasis and its application in immunodiagnosis. American Journal of Tropical Medicine and Hygiene 38: 356-362. WONGRATANACHEEWIN S., CHAWENGKIR~IKUL R., BUNNAGD. & SIRISINHA S. 1988. Analysis of Opisthorchis viverrini antigens by immunoprecipitation and polyacrylamide gel electrophoresis. Parasitology 96: 119-128.