JOURNAL OF CLINICAL MICROBIOLOGY, May 1986, p. 847-851 95-1137/86/5847-5$2./ Copyright D 1986, American Society for Microbiology Vol. 23, No. 5 Enzyme-Linked Immunosorbent Assay for Detection of Antibody to Gnathostoma Antigen in Patients with Intermittent Cutaneous Migratory Swelling A. DHARMKRONG-AT,' S. MIGASENA,2 P. SUNTHARASAMAI,2 D. BUNNAG,2 R. PRIWAN,1 AND S. SIRISINHA1* Department of Microbiology, Faculty of Science,' and Department of Clinical Tropical Medicine and Hospital for Tropical Diseases, Faculty of Tropical Medicine,2 Mahidol University, Bangkok 14, Thailand Received 13 September 1985/Accepted 22 January 1986 A sensitive enzyme-linked immunosorbent assay is described for the detection of immunoglobulin G antibody to Gnathostoma antigen in the sera of patients with intermittent cutaneous migratory swelling who were suspected of being infected by the tissue nematode Gnathostoma spinigerum. The antigen used was a crude somatic aqueous extract of the third larval stage obtained from naturally infected eels. The 1:32 dilution of all sera from 46 patients gave enzyme-linked immunosorbent assay values above the mean (+2 standard deviations) of normal healthy controls, which was arbitrarily used as a cutoff point. The sera from patients suspected of being infected by another tissue nematode, Angiostrongylus cantonensis, as well as from patients with intestinal roundworm infections, also reacted weakly in the test system used in this study. Nonetheless, when the assay was carried out using both Gnathostoma and Angiostrongylus antigens simultaneously, it appeared to be a reliable laboratory test that can be used to support a clinical diagnosis of gnathostomiasis in patients with intermittent cutaneous migratory swelling. Intermittent cutaneous and subcutaneous migratory swellings that are experienced by the people of many Asian countries are largely attributed to infection caused by Gnathostoma spinigerum, a nematode of cats and dogs (2). However, a few other parasitic infections may also cause migratory swelling that cannot be readily distinguished from that caused by G. spinigerum. Moreover, the parasite can also migrate to the central nervous system, causing eosinophilic meningitis, which is often confused with that caused by Angiostrongylus cantonensis, another tissue nematode that is not uncommon in many Asian countries and Pacific islands (1). Epidemiological data and information regarding eating habits are sometimes helpful in distinguishing between these parasitic infections. The diagnosis of human gnathostomiasis is only presumptive unless the parasite can be recovered and identified. The diagnosis is currently based upon clinical manifestations, nonspecific laboratory findings (e.g., eosinophilia), a history of eating improperly cooked fish and other intermediate hosts, and by the exclusion of other diseases. The confirmed diagnosis can only be made following a spontaneous expulsion of the parasite, which is possible in only a few cases. On other occasions, the larvae can be surgically removed from the affected skin sites or recovered in the brain or spinal cord sections of persons who died of infection involving the central nervous system. There is thus a need for a reliable laboratory test that can be used alone or in conjunction with clinical manifestations for the diagnosis of gnathostomiasis. A battery of immunological tests has been used in an attempt to develop a reliable immunodiagnostic method for gnathostomiasis (2). However, like the immunodiagnosis of other parasitic infections, the sensitivity and specificity of these tests leave much to be desired. We recently reported * Corresponding author. 847 on the use of an enzyme-linked immunosorbent assay (ELISA) for gnathostomiasis but, due to its poor sensitivity, found it to be unsuitable for general diagnostic purposes (6). We describe in the present study an improved ELISA method which is highly reliable and useful in supporting the clinical diagnosis of gnathostomiasis in patients with intermittent cutaneous migratory swellings. MATERIALS AND METHODS Patients. Sera used in this study were obtained from patients attending the outpatient clinic of the Hospital for Tropical Diseases (Faculty of Tropical Medicine, Mahidol University, Bangkok, Thailand). Examinations of stool samples for intestinal parasitic infections, together with blood and urine analyses, were routinely performed for all patients. Forty-six blood specimens were obtained from patients with intermittent cutaneous migratory swellings suspected of having been infected with G. spinigerum. Blood was taken from these patients from as early as 2 days after such a swelling was noted to as long as 6 years thereafter. About 15% of these patients experienced repeated swellings before coming in for diagnosis and treatment. Sera were also available from seven patients with neurological signs and symptoms suggestive of eosinophilic meningoencephalitis. The clinical features and a history of consuming poorly cooked Pila snails, which resulted in an outbreak of the illness, are typical for infection caused by the larval stages of A. cantonensis. A few patients with hookworm and other intestinal roundworm infections were also available. Reference sera. (i) Positive serum. The positive reference serum used throughout this study was pooled from eight patients with chief complaints of intermittent cutaneous migratory swellings and with other clinical manifestations, laboratory findings, and histories typical for gnatho-
848 DHARMKRONG-AT ET AL. stomiasis. However, the diagnosis of these patients was only presumptive as no parasite could be recovered. Physical examinations and laboratory findings ruled out other systemic disorders exhibiting cutaneous swellings that may be confused with those caused by G. spinigerum infection. These sera exhibited high levels of immunoglobulin G (IgG) antibody to G. spinigerum by ELISA. (ii) Negative serum. The negative reference serum was obtained from several apparently healthy adults with no previous history of migratory swellings or other parasitic infections. Stool examination by the Formalin-ether concentration technique gave no indication of the presence of intestinal parasites at the time of blood collection. Results from blood and urine analyses of specimens from these individuals were within the normal ranges. Antigen. Infective third-stage larvae of G. spinigerum (L3G) were collected from the liver digest prepared from infected eels bought from local markets. Liver samples prepared from several eels were pooled and digested at 37 C overnight with a ph 2. pepsin solution (1:3, wt/vol). Living larvae were identified under a dissecting microscope using morphological criteria as described by Daengsvang (2). The larvae were individually picked and separated from the digested liver tissue. They were washed twice with.85% NaCl and kept at -2 C before proceeding further for antigen extraction. Crude somatic aqueous extract was prepared by homogenization in saline with a ground-glass tissue grinder followed by brief sonication essentially as described previously (3, 7). The protein concentration of the L3G antigen preparation was determined by the Folin method (4) using bovine serum albumin as the standard. The Gnathostoma antigen prepared as described gave a positive hemagglutination reaction with sera from mice experimentally infected with G. spinigerum larvae. It also gave a strong positive ELISA result in a preliminary test using serum from a patient with confirmed diagnosis for gnathostomiasis (kindly supplied by W. Rattanasiriwilai, Department of Parasitology, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand). The crude somatic extract of the infective third-stage larvae of A. cantonensis (L3A) was prepared from infected Biomphalaria glabratus snails and was characterized as previously described by Dharmkrong-at and Sirisinha (3). The antigen gave positive hemagglutination and precipitation reactions with sera of rats experimentally infected with A. cantonensis larvae. ELISA procedure. The procedure used was modified from that described by McLaren et al. (5). Several parameters affecting ELISA reaction were defined, resulting in an optimal assay condition used throughout this study. Wells of the microtiter plates (Immulon II; Dynatech Laboratories, Inc., Alexandria, Va.) were coated with 5,ul of an appropriate concentration of antigens diluted in.5 M carbonate buffer (ph 9.6). After incubation overnight at 4 C, the plates were washed once with ELISA buffer containing.9% NaCl,.5% Tween 2, 1% gelatin, and.2% bovine serum albumin. A 1% bovine serum albumin solution (2,ul, crystallized; Sigma Chemical Co., St. Louis, Mo.) in ELISA buffer was subsequently added to each well. The plates were left at room temperature for 3 min and then washed three times with the ELISA buffer. Positive and negative reference sera (5,ul each) together with samples to be tested, appropriately diluted in the buffer described above, were added and allowed to react at 37 C for 1 h. The plates were then washed again as described above. Horseradish peroxidaseconjugated rabbit anti-human IgG (5,ul; Dakopatts A/S, Copenhagen, Denmark), appropriately diluted in ELISA buffer, was then added to each well. After 1 h of incubation at 37 C, the plates were washed as described above. Enzyme assay solution (1,ul;.1% o-phenylenediamine in.25 M phosphate citrate buffer [ph 5.5] containing.3% H22) was added, and the enzymatic reaction was allowed to proceed in the dark at room temperature. After 2 to 3 min, the reaction was terminated by adding 25,ul of 8 N H2SO4 and the optical density (OD) was determined in a Titertek Multiskan spectrophotometer at. Variations which occurred from plate to plate on different occasions were corrected by standardizing the OD of positive reference serum, which was included in every plate. RESULTS J. CLIN. MICROBIOL. Determination of optimal ELISA conditions. A checkerboard titration was performed to obtain the optimal antigen concentration and a serum dilution that would best discriminate between positive and negative samples. Figure 1 shows that for both Gnathostoma (L3G) and Angiostrongylus (L3A) antigens, the concentration that could discriminate positive and negative reactions varied from 4 to as low as 1.25,ug/ml. By using these concentrations, a near-linear slope was obtained at a serum dilution from 1:16 to 1:256 (Fig. 2). In subsequent studies, the antigen concentration which was twice the minimum value that could discriminate positive and negative reactions was used, i.e., 2.5,ug/ml. By using this antigen concentration, the best cutoff serum dilution was determined to be 1:32. Table 1 summarizes the results of a 1:32 dilution of individual positive and negative sera before pooling. The pooled positive and negative sera were thereafter divided into small samples, kept at -2 C, and used as reference sera throughout this study. Detection of antibody to Gnathostoma antigen in patients with migratory swellings. The levels of IgG antibody to Gnathostoma antigen (L3G) were determined in the 46 serum samples from patients with intermittent cutaneous migratory swellings. The ELISA values, i.e., the OD reading at, of these sera were markedly higher than those of 22 normal healthy controls (Fig. 3). The mean value for the patients was significantly higher than that of the controls (P <.5). When a mean (+ 2 standard deviations) of the control group was used as the cutoff value, any reading with an OD higher than.143 would be considered ELISA positive for IgG anti-g. spinigerum. By this criterion, all samples from the 46 patients with clinical manifestations suggestive of gnathostomiasis were found to be positive for G. spinigerum antibody, giving a sensitivity of the test of 1%. Serum from the one patient with confirmed diagnosis for gnathostomiasis gave the OD that fell within this range. Fourteen additional patients with migratory swelling and whose stools were also positive for other intestinal parasites did not give ELISA values that were significantly different from those of the 46 patients without intestinal parasites. The mean ELISA value of patients with other roundworm infections was significantly lower than that of the migratory swelling group (Fig. 4). Attempts were made to determine whether there was any correlation between other laboratory findings and the ELISA value. It was found that patients with high ELISA values tended to have higher blood eosinophil counts, giving a correlation coefficient of.3. No correlation was noted with other parameters, including the chronicity nature of the disease, i.e., the time interval before coming in for diagnosis and treatment.
VOL. 23, 1986 DETECTION OF ANTI-GNATHOSTOMA ANTIBODY BY ELISA 849 A 6.l,:32-1:64 1:12 B D -- Positive serum 1.61 *---*Normal serum IA- 1.2-1.-.6-.6-.4- ex I I26.o -1: 6 2. -.. 1:352 -= ~.. =.5 --.5 2325. * I' 4o 2 io 5 2.5 1.25.62.nsY L3G (pgq/mi) L3A (pug/ml) FIG. 1. Checkerboard titration of Gnathostoma (L3G) (A) and Angiostrongylus (L3A) (B) antigens against corresponding sera. The positive serum for Gnathostoma antigen was a pooled reference serum obtained from patients with migratory swelling suspected of having been infected with G. spinigerum. The positive serum for Angiostrongylus antigen was a pooled reference serum from patients with eosinophilic meningoencephalitis. The normal serum used for both antigens was a pooled serum from normal healthy individuals with no previous history of infections. Cross-reactivity with Angiostrongylus antigen. The results diagramed in Fig. 4 show that the sera from patients with migratory swelling also reacted with the antigen prepared from A. cantonensis larvae. However, the mean ELISA value for Angiostrongylus antigen was significantly lower than that for Gnathostoma antigen (P <.1; Fig. 4). Similarly, when the sera from patients with eosinophilic meningoencephalitis were tested, the mean value for Angiostrongylus antigen was significantly higher than that for Gnathostoma antigen (P <.1; Fig. 4). The results were more obvious when individual serum was titrated simultaneously against both antigens (P <.5 by paired t test, Fig. 5). Unlike these two groups of patients, the limited data available for sera from patients with other roundworm i.4- K B L21 LO O.6 6 La. KIU 2.5p/mi 1.25mg/wm..-:Cirftalm. I l?too 1i32 1-4 1* 125.2 Serum dilution Serum dilution FIG. 2. Titration curves of positive and negative sera diluted from 1:16 to 1:2,56 against Gnathostoma (A) and Angiostrongylus (B) antigens. For other explanations, see the legend to Fig. 1.
85 DHARMKRONG-AT ET AL. TABLE 1. ELISA values of reference positive and negative sera Reference OD492 (mean ± SE) for a: serum L3G L3A Negative.55 ±.1.63 ±.1 control Migratory 1.23 ±.92 NA swelling Eosinophilic NA.836 ±.115 meningoencephalitis " L3G, Gnathostoma third-larval-stage antigen; L3A, Angiostrongyluis third-larval-stage antigen; NA, not applicable. infections and from normal individuals indicated that they reacted equally with both antigens but to a lesser extent than did sera from individuals suspected of having been infected with the respective parasites. DISCUSSION The results presented here clearly demonstrate that IgG antibody reactive with G. spinigerum antigens is present in the sera of all patients with intermittent cutaneous migratory 1.61 1.4 1.2. O- swelling, giving a 1% sensitivity of the test in this series of patients. By using the assay condition and the criteria described here, the antibody levels in all patients, expressed as the ELISA value at at a serum dilution of 1:32, were determined to be significantly higher than those of normal healthy individuals. The ELISA values of a few serum specimens from patients from whom G. spinigerum was recovered were indistinguishable from those reported in this study (unpublished observations). It appears therefore that the ELISA is a reliable laboratory test that can be used to support the clinical diagnosis of gnathostomiasis in patients with a chief complaint of intermittent cutaneous migratory swelling and in whom no parasite can be recovered. The simultaneous presence of other roundworm infections in these patients had no significant effect on the results of the test described above. This method therefore represents the only reliable quantitative test that is potentially useful in confirming the diagnosis of human gnathostomiasis. The present improved ELISA method which could detect antibody to G. spinigerum in all patients with migratory swelling is considerably more sensitive than the one used in the previous report, which gave positive results in only 56% of the patients (6). Although the improvement could have been attributed to various minor technical alterations, the major factor probably involved the sources of third-stage larvae used in antigen extraction. In the previous report (6), we used larvae obtained from experimentally infected mice, whereas in the present study, the larvae were obtained from naturally infected eels, which are one of the common sources of infected meat consumed by humans. It is possible that both the somatic components and the metabolic prod- 1.- J. CLIN. MICROBIOL. 1.'.8 :..6'.4.2' ~~~~~~~~ % _,_ %O,e MIGRATORY SWELLINGS NORMAL CONTROLS FIG. 3. ODs of the ELISA reaction for IgG antibody to Gnathostoma (L3G) antigen in the sera of patients with migratory swelling and of normal controls. Also shown in the figure are means + 2 standard deviations for each group. The ODs for all patients are higher than the mean + 2 standard deviations for the normal control group (----). Normol(22) MS(46) EME(7) Others( 3) FIG. 4. Means and standard errors of the ELISA values of IgG antibody in the sera from different groups of patients and normal controls tested against Gnathostoma () and Angiostrongylus (O) antigens. Value in parentheses represents the number of specimens in each group: normal, healthy controls; MS, migratory swelling; EME, eosinophilic meningoencephalitis; others, other roundworm infections.
VOL. 23, 1986 DETECTION OF ANTI-GNATHOSTOMA ANTIBODY BY ELISA 851 A OD BOD Cl-41 I.2\ 1.2-1..8.8.6 Q4 2 `.6.41.2 L3A L3G L3A L3G L3A FIG. 5. ODs for IgG antibody to Gnathostoma (L3G) and Angiostrongylus (L3A) antigens in the sera of normal individuals (A) and of patients with either migratory swelling (B) or eosinophilic meningoencephalitis (C). The paired t test shows a significant difference (P <.5) in the ODs developed against the two antigens by the sera in these two groups of patients but not in the normal control. ucts of larvae obtained from infected eels are different from those of infected mice, and these products may represent the major antigenic constituents that stimulate immune response in humans. Therefore, when these components are included in the antigen preparation used here a higher percentage of positive serum could be detected. Physicochemical and immunological comparisons between these two antigen preparations should clarify this point. Although it cannot be concluded from the data presented here whether the weak cross-reactivity among different species of roundworms noted in this study was due to a true cross-reaction or to unrecognized previous infections, our preliminary results with the sera from animals experimentally infected with these different parasites suggest a true cross-reaction (unpublished data). It was found by imrnunoprecipitation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis that the antigens used were highly complex and that a more refined antigen preparation is needed to minimize cross-reactivities among these parasites. It is also possible to use a recombinant DNA approach to search for specific epitopes that would be diagnostically useful. Such an approach is now being explored. Alternatively, it is also possible to develop a more specific serological test using monoclonal antibody to detect circulating parasite antigen that might be present in these patients. The latter idea is especially attractive for the detection of antigen(s) in the cerebrospinal fluid in those with central nervous system involvement. Until these specific reagents are available, the crude antigen that is currently available and that can be obtained in realistic quantities will have to be used. The data presented in this study nevertheless show that even with this highly complex antigen, the results of the test can be used with a high degree of reliability to support the clinical diagnosis of the disease, particularly when a panel of other parasite antigens are also available for simultaneous assays. ACKNOWLEDGMENTS This work was supported in part by a grant from Mahidol University. We thank T. Harinasuta for her valuable suggestions during the course of this study and Peter B. Billings for his comments and criticisms during the preparation of the manuscript. LITERATURE CITED 1. Cross, J. H. 1979. Studies on angiostrongyliasis in Eastern Asia and Australia. NAMRU-2, Taipei, Taiwan. 2. Daengsvang, S. 198. A monograph on the genus Gnathostoma and Gnathostomiasis in Thailand. SEAMIC/IMFJ, Tokyo, Japan. 3. Dharmkrong-at, A., and S. Sirisinha. 1983. Analysis of antigens from different developmental stages of Angiostrongylus cantonensis. Southeast Asian J. Trop. Med. Public Health 14:154-162. 4. Lowry,. H., N. J. Rosebrough, A. J. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. 5. McLaren, M. L., J. E. Lilywhite, D. W. Dunne, and M. J. Doenhoff. 1981. Serodiagnosis of human Schistosoma mansoni infections: enhanced sensitivity and specificity in ELISA using a fraction containing S. mansoni egg antigens w1 and a,. Trans. Roy. Soc. Trop. Med. Hyg. 75:72-79. 6. Suntharasamai, P., V. Desakorn, S. Migasena, D. Bunnag, and T. Harinasuta. 1985. ELISA for immunodiagnosis of human gnathostomiasis. Southeast Asian J. Trop. Med. Public Health 16:274-279. 7. Techasoponmani, R., and S. Sirisinha. 198. Use of excretory and secretory products from adult female worms to immunize rats and mice against Angiostrongylus cantonensis. Parasitology 8:457-469.