166 TRANSACTIONS OF THE ROYAL SOCIETY OFTROPICAL MEDICINE ANDHYGIENE (1992) 86, 166169
Enzyme-linked immunosorbent assay based on monoclonal and polyclonal antibodies for the detection of Entamoeba histolytica antigens in faecal specimens Ratri Wonsit’,
Prayong Radomyos2 and Danai Bunnag
‘Department of Microbiology and Immunology; 2Department of Tropical Paediatrics and 3Department of Clinical Tropical Medicine and Hospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol University, 42016 Rajvithi Road, Bangkok 10400, Thailand
Abstract A murine monoclonal antibody (MAb) (Eh208C2-2) raised against crude lysate of the pathogenic HM1:IMSS strain of Entamoeba histolytica was usedin an enzyme-linked immunosorbent assayfor the detection of E. histolyticaantigensin faecalspecimens.The detection limit of the assaywas 110and 280 amoebae/mlof the HM-l:IMSS and HK-9 strains in phosphate-buffered saline, respectively. The assaywas applied to single stool samplesfrom 3 groups of individuals comprising 40 patients whose stools were positive for E. histolytica trophozoites and/or cysts (group I), 48 patients whose stools were negative for E. histolytica but positive for other parasites(group II), and 36 parasitologically-negativehealthy controls (group III). Positivity rates of 77.5%, 2.1% and 2.7% were found in samplesfrom groups I, II and III respectively. Specificity, positive and negativepredictive values, and efficiency of the assaywere 97*6%, 93.9%, 90.1% and 91.1% respectively. When group I sampleswere further divided into a trophozoite-positive subgroup IA (13 samples) and a cyst-positive subgroup IB (27 samples), the positive rates were 100% and 66.7%, respectively (FYO.025).
Introduction Amoebiasis is still one of the major tropical diseasees of public health importance, affecting approximately 48 million people (WALSH, 1986). Diagnosis of intestinal amoebiasisis usually madeby microscopicalexamination of the stool for the presenceof amoebic trophozoites or cysts or both. However, patients with asymptomatic amoebiasisare not easily diagnosed, mainly becauseof the uneven distribution of the parasitesin the faecalmaterial and their erratic excretion, as the result of which several stool examinations have to be made to obtain a correct diagnosis (STAMM, 1957). In such cases,detection of amoebic antigens in the stool may supplement microscopical examination. Several antigen detection techniaues involvine nolvclonal antibodies (PAbs) and monoclonal antibodyes(MAbs) have been reported including a PAb-based enzyme-linked immunosorbent assay YELISA)on polystyrene plates (GRUNDY, 1982; GRUNDYet al.. 1987). or on micronore membranesattached to polystyrene’slides (ROOT-et al., 1978), and a MAb-based ELISA in which both PAb and MAb were used (UNGARet al., 1985; MERINOet al., 1990). We report here the development of a double antibody sandwich ELISAinvolving both MAb and PAb similar to that of UNGARet al. (1985)to detect amoebicantigensin stool samplesfrom patients in Thailand. 1
Materials and Methods Subjects
Three groups of human subjects, comprising 40 patients with symptomatic and asymptomatic amoebiasis whose stools were positive for Entamoeba histolytica trophozoitesand/or cysts (group I), 48 patients whosestools were negative for E. histolytica but positive for other parasites, e.g. E. coli, Giardia lamblia, Trichomonas hominis, Opisthorchis viverrini, Trichuris tnchiura, Enterobius vermicularis, Ascaris, hook-worm and E. hartmanni
(group II), and 36 parasitologically negativehealthy controls (erouo III). Groun I uatients were subdivided into subgroup IA, comprising i3 patients whose stools were positive for amoebic trophozoites with or without amoebic cysts, and subgroup IB, comprising 27 patients in whose stoolsonly amoebiccystswere demonstrated.
Single stool samples were collected and examined *Authorfor correspondence.
within 2 h. Two preparations, one stained with Lugol’s iodine and the other subjected to formalin-ether concentration (RITCHIE, 1948) were examined under a lierht microscope. Each fresh stool samplewas diluted to niv’ea 10% suspension (w/v) in phosphate-buffered saline, DH 7.2. containina 0.05% Tween [email protected]
(PBST). Coarse debris was eliminited by centrifugatioh at 200 g for 15 min, the clear supernatant was dispensed in 1 ml aliquots to which proteaseinhibitors were added to give a final concentratonof 1 mM phenylmethylsulphonylfluoride (PMSF) and 1 mM N-l-p-tosyl-L-lysinechloromethylketone (TLCK), and kept frozen at -70°C until testing. Preparation of E. histolytica antigen The HM-l:IMSS strain of E. histolytica was cultured
axenically in 125 ml screw-cappedErlenmeyer flasks in TPS-1 medium according to the technique described by DIAMOND (1968), except that heat-inactivated bovine serum was used in place of horse serum. Trophozoites from 48-72 h cultures were chilled to dislodae the uarasites, followed by washing 3 times with O.f5 M NaCl solution (NSS) by centrifugation at 120 R at room temperature for 5 min. The cell sediment was adjusted in NSS to give annroximatelv lo7 cells/ml and disrunted at 4°C with an ultrasonic disintegrator (MSE, London) at 20 kilocyclesimin for 2 min. Complete disruption of cells was-ascertainedby examination under a light microscone. The cell lvsate was centrifuged at 13 000 P for 20 min at 4°C to remove associa;d cell residue”and small particles. The opalescent supernantant was collected, dispensed in aliquots, lyophilized and stored at -20°C until used. The nrotein content was determined by the method of LOWRY et al. (1951) using bovine serum albumin (Sigma Chemical Company, St Louis, USA) asstandard.
Preparation of antigens from the E. histolytica-like Laredo strain, 0. viverrini and G. lamblia The E. histolytica-like Laredo strain was obtained
from the American Type Culture Collection, Rockville, Maryland, USA, cultured in TPS-1 in the presence of heat-inactivated bovine serum at room temperature as described by DIAMOND(1968), and the amoebic cell lysatewas prepared as described above. Antigens from 0. viverrini and G. lamblia were prepared as previously described (THAMMAPALERD et al., 1988; NACAPUNCHAI et al., 1986).
167 Preparation of rabbit anti-E. histolytica polyclonal antibodies
Two healthy rabbits were immunized intramuscularly 5 times with 0.5 ml of crude amoebic whole cell lysate equivalent to approximately 5~ lo6 trophozoites of the HM-l:IMSS strain of E. histolytica emulsified with an equal volume of Freund’s complete adjuvant initially and Freund’s incomplete adjuvant subsequently at intervals of 2-3 weeks. The rabbits were bled by cardiacpuncture 9 d after the last immunizing dose, and the serum indirect fluorescent antibody (IFA) titres against HMl:IMSS were both 1:640. Serawere divided into aliquots and kept at -20°C until used. Serum immunoglobulin (Ig) G was purified by affinity chromatography on protein A [email protected]
to the technique of SEPPALLAet al. (1981). The purified polyclonal IgG (PIgG) at a concentration of 5 ygiml gavean IFA titre of 1:1280. The PIgG was divided into aliquots, lyophilized and stored at -70°C until used. Anti-E. histolytica monoclonal antibodies
Murine MAbs were raised against the HM-l:IMSS strain of E. histolytica in Sp2/0 myeloma cells according to the method of GALFRE& MILSTEIN,1982(seeTHAMMAPALERD& THARAVANIJ,1991). Several antibody-secreting hybridoma clones were obtained; clone Eh208C2-2 was used in this study. This MAb, of IgGl isotype, showedgeneralizedstaining by IFA of the granules, cytosol and membrane of trophozoites of the HMl:IMSS strain, recognized 4 proteins of 122, 115, 111 and 65 kDa, and did not react with amoebic cysts (N. Thammapalerd, unpublished observations).Ascites fluid from pristane-primed BALBic mice was collected and centrifuged at 900 g for 5 min and the supernatant was kept at -20°C. The ascitesfrom mice injected with Sp2iO myeloma cells served as the control. The IgG fraction of the ascites(MIgG) was purified by protein A Sepharose [email protected]
affinity chromatography as described above. MAb-PAb-ELIs.4 for the detection of E. histolytica antigen in a model systemand in stool samples
The technique used was a modification of the double antibody sandwich ELISA previously described by UNGARet al. (1985). In a model system, a 96-well polystyrene micro ELBA plate (Dynatech Laboratories, Inc., Virginia, USA) was coated with 100 yl of 10 pg/ml of MIgG at 37°C for 1 h and at 4°C overnight. The unbound components were twice washed for 3 min with distilled water. The non-reactive sites were blocked with PBST containing 5% non-fat dry milk ([email protected]
)at 37°C for 1 h. After washing 4 times with PBST, the wells were loaded with 100 ul of varying concentrations of the standard antigen of the HM-l:IMSS or HK-9 strains of E. histolytica suspendedin either PBST alone or in a 10% suspensionin PBST of parasitologically-negative stool. The plate was incubated at 37°C for 2 h, washed with PBST, 100 yl of goat alkaline phosphatase (AI’)-conjugated anti-rabbit IgG (Zymed Laboratories Inc., South San Francisco, California, USA) diluted 1:lOOOin PBST containing 1% bovine serum albumin were added to each well, followed by incubation at 37°C for 1 h. The plate was washed and 100 ul of substrate solution (p-nitrophenyl phosphate in diethanolamine buffer, pH9.8) were added. The enzyme-substratecolour reaction was allowed to develop at room temperature and then terminated with 3 N NaOH after 40 min. The optical density (OD) of each well was measured at 405 nm with an ELBA plate reader (Titertek [email protected]
,MCCi340, Flow laboratories). When applied to faecal samples, a similar procedure was adopted except that a 10% faecal suspensionin PBST was used in place of the standard antigen. Sensitivity, specificity, positive predictive value, negative predictive value, and efriciency of the assaywere assessedas described by GALEN & GAMBINO (1975).
Results Sensitivity and specificity of the MAb-PAb-ELrsA
Result of the assay performed on stools from 36 healthy individuals gavea mean OD + standarddeviation (SD) of 0.078t0.069. Test samples with ODaO.266 were consideredpositive; this value was equivalent to the mean OD+ZSD+an estimate of error of the ELISA reader of 0.05. Based on this cut-off value, the sensitivities of the assayfor the HM-l:IMSS and HK-9 strains of E. histolytica in PBST were 110 and 280 amoebae/ml or 11 and 28 amoebae/well, respectively (Fig. 1). In a 10% suspension in PBST of parasitologically-negative stool, the assaywas lesssensitive, with detection limits of 1200and 1600amoeabeiml,respectively (Fig. 1). The specificity of the assaywas assessedby testing a 10% suspensionof stool from a healthy individual containing various concentrationsof whole cell lysatesof the E. histolytica-like Laredo strain, 0. viverrini, and G. lamblia (Fig. 2). The OD was below the cut-off level with all concentrationsof the parasitestested. Application to clinical specimens
The seropositive rate of group I samples(77.5%) was significantly higher than those of group II (2.1%) and group III (2.7%) (x*=53.6 and 43.4, respectively, de0-m Al .-. n ‘I
HM-1:IMSS In PBST HK-9 In PBST HM-1:IMSS In faecer HK-9 In fosces PEST alone Foecer alone
’ + I: :::::i: 102
Number of E. histolytica/ml Fig. 1. Sensitivity of the MAb-PAb ELISAin the detection of amoebic antigensin lysatesof the HM-1:IMSS and HK-9 strainsofEntamoebahistolyfica suspendedin phosphate-bufferedsaline plus 0.05% Tween 2Oa (PBST) or in a 10%suspensionof parasite-freestool. With a cut-off optical density (O.D.) of 0,266, sensitivity of the assaywas 110and 280amoebae/ml of HM-1:IMSS and HK-9 in PBST, and 1200 and 1600 amoebae/mlof HM-1:IMSS and HK-9 suspendedin the parasite-free stool,respectively. 0-O U-W 1.2
HM-1:IYSS In PBST HY-1:IYSS In normal faeces HK-9
-lamblia p. yhrrinl
‘0 ‘. A.
l --b \
n. . A.
Fig. 2. Specificityof the MAb-PAb EUSAassessed againstvariousconcentrations of the antigensof the HM-l:IMSS and the HK-9 strainsof Entamoebahtsrolyrica and antigensfrom other intestinal parasites,Opisrhmhis viverrini, the E. hisrolytica-likeLaredo strain and Giardia lamblia addedto the supernatantof normal stool suspension.No binding activity was observedin the presenceof cell lysatesof E. bsrolytica-like Laredo strain, 0. vivetini and G. land&a.
168 10/27 (66.7%)
2.4 2.0 7
0.0 -0.4 I Subgroup
Fig. 3. Detectionof Emamoebahistolyticaantigen in stoolsamplesfrom subjectsin groupsI (infectedwith E. hiszolytica), II (withoutE. histolytica but with otherintestinalparasites) andIII (with negativestools).GroupI sampleswere more frequently positive than the other fwo groups (FcO~OOl).Of the groupI patients,thosewho harbouredamoebictrophozoiteswith or withoutcysts(subgroupIA) weremorefrequentlypositivethanthosewhoharbouredonlyamoebiccysts(subgroupIB) (Fisher’s exacttest,PiO.025).
grees of freedom=l, P