Toxkoa, Vd . 30, No. 12, pp. 1609-1620, 1992 . Printed in Goat Britain .
0041-0101/92 SS.OD + .00 © 1992 Perpmon Prea Ltd
THE USE OF ENZYME-LINKED IMMUNOSORBENT ASSAY FOR THE QUANTITATION OF CALLOSELASMA RHODOSTOMA (MALAYAN PIT VIPER) VENOM AND VENOM ANTIBODIES NGSr-Hoxc Tax, KS-HUAT Y>~ and Mot-ro ISxwxDnR Nnc Jn~~It Department of Biochemistry, University of Malaya, Kuala Lumpur, Malaysia (Received 29 April 1992; accepted 11 July 1992) N.-H. Twx, K.-H . Y0o and M. I. Nnc J~nR. The use of enzyme-linked
immunosorbent assay for the quantitation of Calloselasma rhodostoma (Malayan pit viper) venom and venom antibodies. Toxicon 30, 1609-1620, 1992.-The specificity and sensitivity of an indirect and two (an `ordinary' and a `rapid') double sandwich enzyme-linked immunosorbent assay (ELISA) procedures for the quantitation of Calloselasma rhodostoma (Malayan pit viper) venom were examined. The three assays were equally sensitive and the accuracy of the assays was not substantially affected by individual variation in the venom composition . The specificity of the assays was examined against 26 venoms from snakes of the families Viperidae and Elapidae . While the double sandwich ELISA procedures were sufficiently specific to be used in the clinical immunodiagnosis of C. rhodostoma bite in Malaysia, the indirect ELISA procedure exhibited extensive cross-reactivity with other Malaysian pit viper venoms. Attempts were made to improve the specificity of the indirect ELISA procedure for the quantitation of C. rhodostoma venom . A `low ELISA crossreactivity' venom fraction (termed VF52) was isolated from C. rhodostoma venom by repeated Sephadex G-100 gel filtration chromatography . The indirect ELISA procedure using antibodies to VF52 as immunoreagent showed an improvement in specificity . The use of the indirect ELISA procedure for the detection of C. rhodostoma antibodies was also examined and the results show that the assay was sufficiently specific to be used for retrospective diagnosis of C. rhodostoma bite in Malaysia, in particular when VF52 was used as the coating antigen .
INTRODUCTION
ENZYbIE-LINKED immunosorbent assay (ELISA) is the most widely used technique for the immunodiagnosis of snakebite (CHANDLER and HURRELL, 1982; Ho et al., 1986x; MIxTON, 1987; THE~xs~rox, 1983; TI~xsTOx et al., 1977) . The double sandwich ELISA method is usually used for detecting snake venom in tissues and body fluids, while indirect ELISA is mainly used for detecting antibodies to venom resulting from previous envenoming. Specificity of ELISA for detecting venom and venom antibody is one of the major considerations in evaluating the applicability of the method for confirmatory and retrospective immunodiagnosis of snakebite . 1609
161 0
N.-H . TAN et al.
Several authors (DHALIWAL et al., 1983; Ho et al., 1986b) have described a double sandwich ELISA procedure for detection of Calloselasma rhodostoma (Malayan pit viper) venom, but the specificity of the assay has not been fully evaluated. Calloselasma rhodostoma is the commonest cause of snake venom poisoning in Malaysia (LIM, 1982). The present study was undertaken to determine the applicability of the indirect ELISA procedure for the quantitation of both Malayan pit viper venom and antibodies. For comparison purposes, the specificity of the double sandwich ELISA for Malayan pit viper venom was also examined . MATERIALS AND METHODS Materials The Malaysian snake venoms used in this study include those of Calloselasma rhodostoma (Malayan pit viper), Trimeresurvs purpereonwcelater (shore pit viper), T. albolabris (white-Tipped tree viper), T. popeorium (Pope's tree viper), T. sronatranus (Sumatran pit viper), T. wagleri (speckled pit viper), Enhydrina schistose (common sea snake), Naja raja sputatrix (Malayan cobra), Naja kaouthia (Monocellate wbra), Ophtophagus hannah (lung cobra), Bungarus cmrdidus (Malayan krait) and B. jasciatus (banded krait) . Other venoms included in the specificity studies are venoms of Agkistrodon b. btlirteatus, A . c. contortrix, A . p. piscivorur, Bothrops riper, B. atrox, Crotohu adamantees, C. atrox, Sistrwur c. ttrgtmitees, Trimeresurar jlavoviridir, Echis carinetus, Vipere a. ammodytes, Daboia russelli siamensis, Dendroaspis angusticeps, Notechis scutatus and Hoplocephalus atephensü. The venoms were obtained from Miami Serpentarium Laboratories (U .S.A .), Latoxan (France), Ophidia (Switzerland), Sigma Chemical Company (LI .S .A) and SEA Venom Institute (Malaysia) . Rabbits were supplied by Central Animal House, Faculty of Medicine, University of Malaya. Freund's adjuvant was obtained from Difco Laboratories (U.S .A.) . Goat anti-rabbit IgG-horseradish peroxidase conjugate was from Bio-ltad Laboratories (U.S .A.) . Micro-ELISA plates were from Nunc (Denmark). Protein A-agarose was from Sigma Chemical Company (U .S.A). Other chemicals and reagents used are of analytical grade and were all purchased from Sigma Chemical Company. Preparation of antiserum to snake venom Antisera against the pit viper venoms (C. rhodostoma, T. pwpureomaeelates, T. albolabris, T. Popeorum, T. sionatranes and T. waglere) or venom Erection 52 (VF52, see below) were preparod as follows: the venom (2 mg in 0 .1 ml phosphato-buB'ered saline, pH 7.4) was mixed with an equal volume of Freund's complete adjuvant and injected i .m. into the kft and right thighs, alternatively, of rabbits (2 kg) at biwcekly intervals. The rabbits were bled 9 days after the second booster. Antisera against the elapid venoms (N. n . sputatrix, N. kaoethia, O . hannah, B. candides and B. fasciates) were prepared using a hyperimmunization scheme and the venom antigen was cross-linked by glutaraldehyde (Brune and VoaT, 1971) prior to injection . The rabbit was immunized every 2 weeks for 3 months with increasing doses (0 .5 to 4 .0 mg) . The sera obtained after clotting were stored frozen until used. Determination ojprotein concentration Protein concentration was determined according to the Lowry's method (Pe~rrrssort, 1983). Indirect ELISA procedure The indirect ELISA procedure used was modified from Hw~rorro et al. (1980) and performed at room temperature (28°C) . The micro-ELISA plates were coated overnight at 4°C with the venom and washed by PBS-Tween solution . The antiserum or normal rabbit serum (control) was added to the wells and incubated for 1 hr. The wells were washed and goat anti-rabbit IgG conjugate was then added and incubated for 1 hr. After washing, substrate solution (H=O~ and O-phenylene diamine) was added. After 10 min, the reaction was stopped by adding 12.5% sulphuric acid and the absorbance at 492 nm was determined by a miau-ELISA reader. Preliminwry checkerboard titrations were conducted to obtain the optimum operation conditions . The C. rhodartoma venom used for immunization (Latoxan, lot A223, the snake was from southern Thailand) was used as the refenenoe C. rhodostoma venom in ELISA. Isolation of IgG and non-IgG antibodies and preproation ojlgG-horseradish peroxidast conjugate Antibodies produced against a venom are predominantly immunoglobulin G (IgG) but also include non-IgG antibodies such as IgM, IgA, IgD and IgE (Cr.eaa et ol., 1972) . IgG antibodies were purified using protein A
ELISA of Malayan Pit Viper Venom
161 1
affinity chromatography (Htmsox and Hev, 1980) . Non-IgG antibodies were isolated from the unbound proteins by ammonium sulphate precipitation at SO'/e saturation (Hexi.ow and Lexe, 1988) . The preparation of the IgCr-horseradish peroxidase (HRP) conjugate was carried out according to Tussnv (198 .
Double sandwich ELISA procedure Double sandwich ELISA procedure was conducted as modified from HNrvox and P~wr.n~ (1982). In the `ordinary' procedure of double sandwich ELISA for C . rluxlostoma venom, microtitre plates were coated overnight at 4°C with either purifiod IgG antibodies to C. rhodostoma or IgG from pre-immune serum (control) . The plates were washed four times with PHS-Tween solution and C. rhodastoma venom (or other appropriate venoms) was added and incubated for 2 hr. After washing, IgCrHRP conjugate was added and incubated for another 2 hr, followed by addition of the substrate solution . After 10 min the reaction was stopped by adding sulphuric acid . Using precoated microtitre plates, the total assay duration for this `ordinary' double sandwich ELISA procedure was approximately 5 hr. In the `rapid' double sandwich ELISA of C. rhodostorna venom, the microtitre plates were coated overnight at 4°C with IgG anti-C. rJrodostomn venom. After washing, the venom was added and incubated for 15 min at 28°C . This was followed by the sequential incubation of IgG conjugate for 15 min and substrate solution for 15 mice . Using precoated microtitre plates, the total assay duration was approximately 50 min. Double sandwich ELISA of snake venom by non-IgG anti-C. rhodostoma venom was carried out in the same way as the ordinary procedure, except that the microtitre plates were coated with the non-IgG antibodies (2 pg/ml). Reaction between the non-IgG antibodies and snake venom (100 ng/ml) was detected using the IgG anti-C. rhodosroma venom-HRP conjugate, and the control consisted of equivalent concentration of the nonIgG antibodies from pre-immune serum .
Sephadex G-100 gelfiltration chro»urtography ojC. rhodostoma venom and isolation oja 'low ELISA cross-
reactivity' venom fraction Calloselasma rJrodostoma venom (Latoxan lot A223) was fractionated by a Sephadex G-100 gel filtration column . Ovalbumin (mol. wt 45,000) and lysozyme (mol . wt 14,400) were used as mol . wt markers . The antigenantibody reactions between the venom fractions and the six monospecific antibodies to the Malaysian pit viper venoms (C. rhodostoma, T. purpureomaculatus, T. albolabris, T. popeorurrr, T. sumatramas, T. walglerr) were rxnmined using indirect ELISA procedure. The venom fractions were diluted 40,000-fold and the antibodies were diluted 2000-fold . Calloselasrrra rhodostoma venom fractions which showed high indirect ELISA reactions with anti-C. rlurdostoma venom but low reactions with the five anti-Trtmeresurus venom were identified as `low ELISA cross-reactivity region' (tubes 45-60), the tubes were pooled and further fractionated by a second Sephadex G-100 gel filtration chromatography. The reactions between rechromatographed fractions (1 :10,000) and the above-mentioned six monospacific antibodies were again expmined by indirect ELISA . Tube 52 of the rechromatographed fractions (termed venom fraction 52 or VF52), which reacted strongly with the anti-C . rhodostoma venom, exhibited minimum ELISA cross-reactions with the five anti-Trbneresurus venoms.
Statistical analysis The significance in difference was determined by Student's t-test . The level of significance was P = 0 .05 .
RESULTS
Indirect ELISA jar C. rhodostoma venom The indirect ELISA procedure for C. rhodostoma venom yielded a linear dose-response curve for the venom at concentrations of 5 ng/ml (absorbance = 0.2) to 50 ng/ml (absorbance = 1 .2) when the microtitre plates were coated overnight with the venom. A total of 13 different samples of Malayan pit viper venom from different commercial suppliers were used to examine the effect of individual ditFerences in Malayan pit viper venom on the antigen antibody reaction of the indirect ELISA quantitation of the venom. While different venom samples yielded slightly different absorbance readings ranging from 80%
1612
N.-H. TAN et al.
T~at.E 1 . CROSS-REACI'IVrrY HECWBQi ANTIBODn~ TO Catiosetasma rhodostoma vENOS+ eNn ort~x sNexE vENO~rs nv rtm tNOtaECr ELISA DETA.CI'ION OF Tt~ VENOM
Coating venom
O.D. as % of C. rhodeutorrra venom (mean1S.D ., n = 9-15) Anti-C . rhoetostoma venom Anti-VF52 as as antibodies antibodies
C. rhoetostoma Family CrotaGdae Subfamily Crotalinae A. b. bitineatus A. c. contortrix A. p. piscivores B. asper B. atrox C. adammiteur C. atrox S. c. tesgeminus T. Jtavoviridis T. albolabris T. popeorhan T. sumatranus T. pwpureomaculates T. wagleri
100
100
50 .7114.1 31 .212 .9 55 .618 .7 23 .314 .8 19 .711 .1 40 .415 .4 17 .715 .3 36 .417 .9 36 .713 .9 54 .915 .4 30.016 .E 21 .812.3 30.418.5 32.416.2
28 .210 .3 16 .712 .E 28 .910 .9 6.212 .2 4.913 .4 11 .410 .8 6.810.E 1 .610.5 14.510.2 2.910.1 8.410.4 2.110.3 11 .717.1 3 .310.4
Subfamily Vipcrinae E. ceuinatus V. a. ammodytes D. r. siamensis
12.412.8 17.414.E 22.915 .5
1.410.8 4.810.2 7 .610.2
Family Elapidae B. candides B. fasciates N. n. spetatrix N. kaoethia O. hannah D. angesticeps N. seetates H. stephensü E. schistosa
10 .611 .1 11 .915 .1 9 .51 I .6 15 .112 .2 15 .916 .3 11 .311 .0 8.912 .8 18 .912 .E 11 .512 .0
1.810 .1 4.411 .3 1.310 .1 2.310 .3 2.010 .7 6.910 .4 7.110 .4 4.l 11 .4 9.910 .8
The assay mixture contained 100 pl venom (25 ng/ml), 100 tel anti-C. rhodostoma venom (1 : 2000) and 100pl conjugate (1 : 4000). The absorbance at 492 nm for C. rhodostoma venom was 1.0410.05 with anti-C . rhodostoma venom, and 0.4610.03 with anti-VF52.
to 108% of the reference venom sample, the readings were generally not statistically significant (P > 0.05) from the reference sample . Extensive cross-reactions between the anti-C. rhodostoma venom and venoms of the Crotalinae subfamily (particularly venoms from the three Agkistrodon and the genus Trimeresurus) were observed in the indirect ELISA (Table 1). On the other hand, there were far fewer cross-reactions between the anti-C. rhodostoma and venoms of snakes from the subfamily Viperinae and the family Elapidae. In an attempt to shorten the assay time for C. rhodostoma venom, the microtitre plates were coated with the venom for 1 hr at room temperature instead of overnight coating at 4°C. This also yielded a linear dose-response curve (not shown) for the venom at concentrations of 5 ng/ml (absorbance = 0.2) to 50 ng/ml (absorbance = 0.78). Further reduction in the time for coating reduced the sensitivity substantially.
ELISA of Malayan Pit Viper Venom
161 3
Teet~ 2 . Srecuacmr oF nm txntsECr ELISA rnocFnuxE t~t T~ nsrecnoN op Calioadatma rhodostotna verroa ~xneonos
Venom antibodies Anti-CaUoselasma rhodostoma Anti-T. albolabris Anti-T. popeorium Anti-T. .rronatranus Anti-T. purpereonracuJates Anti-T. wagkri Anti-B. candides Anti-B. jasciater Anti-Naja n. spetatrix Anti-Naja kaoet6ia Anti-Ophiophagua hanna6
O .D. as % of anti-C. rhodoatwna C. rhodoatoma Venom fraction 52 as antigen (VF52) as antigen 100 20 .3 16 .5 5 .8 I5 .4 12 .2 0 0 3 .2 5 .9 2 .0
100 (0.0 10.0 1 .3 6.3 0 n.d . n.d . n.d . n .d. n .d.
The assay mixture contained 100 pl venom antigen (25 ng/ml for C. rhodostoma venom, and 100 ng/ml for VF52 antigen), 100 ~l venom antibodies (1 : 2000) and 100 pl conjugate (1 : 4000) . when anti-C. rhodwstoma venom was used as antibodies, the absorbance at 492 nm for C. rhodoatwna venom antigen was 1 .04 f 0 .05, and for VF52 was 1 .0810 .02 . Value waa mean of 9-15 determinations . n .d. = not determined .
Detection of C. rhodostoma venom antibodies using indirect ELISA When the micro-ELISA plates were coated with 100 ng/ml of C. rhodostoma venom, the indirect ELISA for detection of C. rhodostoma venom antibodies yielded an exponential dose-response curve for the laboratory-prepared C. rhodostoma venom antibodies from a dilution of 1 :2000 (absorbance = 1.2) to 1 :20,000 (absorbance = 0.2) (not shown) . Table 2 shows the specificity of the indirect ELISA procedure for the detection of C. rhodostoma venom antibodies. Antibodies to venoms of the Malaysian elapids (including B . candidus, B. fasciatus, N. n. sputatrix, N. kaouthia and O. hannah) and that of T. sumatranus yielded low cross-reactions ( < 6%), whereas the anti-albolabris, anti-T. popeorium, anti-T. purpureomaculatus and anti-T. wagleri yielded 12-20% cross-reactions .
Double sandwich ELISA of C. rhodostoma venom The double sandwich ELISA procedure yielded an exponential dose-response curve for C. rhodostoma venom at concentrations from 5 n~/ml (absorbance = 0 .26) to 100 ng/ml (absorbance = 0.95). To examine the effect of individual variation in C. rhodostoma venom composition on the ELISA absorbance readings, a total of 13 different samples of C. rhodostoma venom from different commercial suppliers were used as venom antigens . In general, the venom samples yielded absorbance readings comparable to that of the reference venom sample . The specificity of the double sandwich ELISA for the detection of C. rhodostoma venom was examined by the reaction between IgG anti-C. rhodostoma and venoms from snakes of the families Viperidae and Elapidae . Table 3 shows that there were no significant crossreactions between the IgG anti-C. rhodostoma antibodies and all the 26 venoms tested . The possibility that non-IgG anti-C . rhodostoma venom may contribute to the extensive cross-reactions observed in the indirect ELISA was examined by using the double sandwich ELISA procedure with non-IgG anti-C. rhodostoma as coating antibodies. The
1614
N :H . TAN et al.
Twet.E 3. Cross-xEecnvrrv ~EEx exneonms ro Calloselasma rhodostoma vENOSr ~Nn orrme sNe¢E vENOMs uv rt~ nouaLE seNnw~cH ELISA DETA.CfION OF rt~ VENOM
Venom C. rhodostoma
O .D . as % of C. rhodostoma venom (mean f S .D., n = 9-1 ~ Ordinary procedure Rapid procedure 100
100
1 .5 f 0.1 I .6 f 0 .1 1 .9 f 0.3 0.6 t 0.2 0 0 .7 t 0 .1 1 .9 f 0.2 0.1 f 0.1 1 .9 t 0.2 0 .2 t 0 .1 0 0.1 t 0.1 0.1 f0.1 0
3 .8 f 0 .1 2 .9 f 0 . I 1 .7 f 0.3 l .2 t 0 .9 1 .6t0 .3 1 .4 t 0 .3 1 .8 f 0.0 1 .0 t 0.1 1 .4 t 0 .3 0 .3 t 0 .2 1 .4 f 0 .3 0.9 ~ 0.4 0.8i0.3 0 .8 f 0 .0
Subfamily Viperinae E. carinatus V. a . ammodytes D . r. siamerrsis
0 0 0
0.8 i 0.0 0.6 f 0.0 0
Family Elapidae B . candides B . jasciates N. n . sputatrix N. kaouthia O .hannah D. angusticeps N. scetatus If. stephensü E. schistosa
0 0 0 0 0 0 0 0 0
0 0.7 t 0.3 0 .2 f 0.2 0 0 0 0 0 0
Family Crotalidae Subfamily Crotalinae A . b . bilineatus A . c. contortrix A . p . piscivorus B. asper B. atrox C. adamanteus C. atrox S. c . tergemim~s T. Jlavoviridis T. albolabris T. popeorhen T. sumatramis T. purpureomaculatus T. wagleri
The assay mixture contained 100 pl IgG anti-C. rhodostorna venom (2 pg/ml for the ordinary assay, 10 pg/ml for the rapid assay), 1001e1 of venom (100 ng/ml), and 100 ul of conjugate (1 : 3000 for the ordinary assay, 1 : 1000 for the rapid assay) . Absorbance at 492 nm was 0.93 f 0.02 for the ordinary assay, and 0 .77 f 0 .01 for the rapid assay.
results show that the non-IgG antibodies did not react with C. rhodostoma venom, nor with any of the other venoms tested . Rapid double sandwich ELISA of C. rhodostoma venom
The `rapid' double sandwich ELISA procedure for C. rhodostoma venom yielded a linear dose-response curve ranging from 5 to 25 ng/ml venom (absorbance readings between 0 .05 to 0 .24), but an exponential dose-response curve was obtained when the venom concentration was in the range of 25 to 100 ng/ml (absorbance readings between 0.24 and 0.7~. Different samples of C. rhodostoma venom yielded only slightly different absorbance readings . Table 3 shows that the rapid double sandwich procedure was also
ELISA of Malayan Pit Viper Venom
16(5
Fto. l.a,b . Ixnu~cr ELISA rAOt+a~s of Calloaelavna rhodastoma vwoM t~cnoxs oe~r~x~ av St~n~c G-(00 0~. tru xs~nox c~neoauroaxeexv. Two-hundred milligrams of C. rl~odoatoma venom, dissolved in 12 ml of PBS (pH 7.2) was applied to the column (2 x 100cm) equilibrated with the same buffer. Tube volume = 4.5 ml . Miaotitre plates were coated at 4°C with 100pl of the venom fraction (1 :40,000) and incubated with 100 pl of the following antibodies (1 :2000) : A, anti-C. rhodartoma venom; B, anti-T. albolabris venom; C, anti-T. popeornon venom; D, anti-T. pwpureomaculatur venom; E, anti-T. smnatrmws venom; F, anti-T. wagleri venom . -, absorbante at 280 am ; , absorbante of the indirect ELISA reaction mixture at 492 nm .
very specific and that there were only minimum cross-reactions with all the 26 venoms tested . Antigenic studies of C. rhodostoma venom jractions
Sephadex G-100 gel filtration chromatography resolved the C. rhodostoma venom into three protein peaks (Fig. la), corresponding to high mol. wt ( > 45,000), medium mol. wt (14,0005,000) and low mol. wt ( < 14,000) proteins, respectively . The antigen-antibody reactions between the gel-filtration fractions of C. rhodostoma venom and the monospecific antisera to six Malaysian pit viper venoms (C. rhodostorna and the five Trimeresurus) were examined using indirect ELISA (Fig. 1b). The medium mol. wt venom fractions yielded high indirect ELISA reading (absorbance > 0.8) while the high and low mol. wt protein fractions yielded moderate (absorbance of 0.3-0.8) and low (absorbance < 0.3) ELISA readings, respectively. Comparison of the indirect ELISA profiles of the antigen-antibody reactions between the C. rhodostoma venom fractions and the six antisera to the Malaysian pit viper venoms (C . rhodostoma and the five Trimeresurus) indicated the presence of a `low ELISA cross-reactivity' region (Fig. l, tubes 45-60) .
1616
N: H. TAN et aJ.
b
3~2
B I I~ I I I II I I I
2~6 2"4 2,0 _ ~ 1,B-
I I
Fia. 1 . (Cont~~
-o~e
n
"I I i 1
1~
0 .4 ~ ' 0
F- 1~2
1
0~6 -
0
- ~ .8
Î 1 1 1
1 .2 -
Tube Number
~ 2 .0
I 1 1 1
20
40 60 Tube Number
1
-0~4
80
0
E c `N O W
` "
ELISA of Malayan Pit Viper Venom
161 7
These tubes were pooled and the pooled fractions were rechromatographed on the same Sephadex G-100 gel filtration column to yield a major protein peak (tubes 400) and a very small, minor protein peak (not shown) . Anti-C . rhodostoma venom yielded high indirect ELISA readings with the major protein peak only, while the five anti-Trimeresurus venoms all yielded low indirect ELISA readings with both the major and the minor protein peaks. Tube 52 (VF52) of this rechromatographed `low ELISA cross-reactivity' venom fraction yielded maximum indirect ELISA reading with anti-C . rhodostoma venom and very low indirect ELISA readings with the five anti-Trimeresurus (see Table 2). The fraction was used to raise antibodies (anti-VF52). Improved indirect ELISAs for C. rhodostoma venom and venom antibodies using antiVF52 and VF52 as immunoreagents
Table 1 shows that the use of anti-VF52 as immunoreagent for indirect ELISA for C.
rhodostoma venom indeed reduced substantially the cross-reactions with other venoms . In general, venoms from the same Agkistrodontini tribe yielded less than 30% crossreactions, while the cross-reactions for venoms from the genus Trimeresurus were below
15%. The elapid venoms yielded less than 10% cross-reactions . Table 2 shows that with VF52 as coating antigen, the indirect ELISA for C. rhodostoma antibodies was also more specific; the cross-reactions with other anti-Trimeresurus were less than 10% .The indirect ELISA using VF52 as coating antigen was as sensitive as the one using whole venom as antigen. DISCUSSION
Indirect ELISA for C. rhodostoma venom The indirect ELISA for C. rhodostoma was as sensitive as the double sandwich method . Individual variation in the C. rhodostoma venom composition generally did not result in substantial differences in the indirect ELISA readings . Thus, the use of any C. rhodostoma
venom available to construct the standard curve is not likely to greatly affect the accuracy of the assay. The presence of high background readings due to non-specific adsorption of immunoreagents to the microtitre wells is one of the major problems affecting specificity of the ELISA (Ho et al., 1986x) . Tween 20 and non-specific proteins such as bovine serum albumin (BSA) have been used in indirect ELISA procedure to eliminate or decrease the background readings (BoHEIt and OwxsY, 198). In the present procedure, Tween 20 was incorporated in the washings. Preliminary studies, however, showed that the use of BSA as a blocking reagent affected the reaction between venom antigen and antibody (Y$o, K. H., M.Sc. Thesis, University of Malaya, Kuala Lumpur, 1992). As the specific venomantibody reaction in the present study is sufficiently strong to allow the background to be acceptable by simply subtracting the background from specific reaction, BSA was not used as a blocking agent in the assay. Specificity is one of the major considerations in evaluating applicability of ELISA for immunodiagnosis of snakebite (Ho et al., 1986x ; MAIesxALI, and Haxx~nxx, 1984) . Specificity is generally defined as the lack of cross-reactivity between closely related species . Many authors reported that the ELISA systems which used poorly characterized snake venoms or snake venom antibodies as reagents had low specificity (Mu~rrox, 198, as common antigens occur in venoms of related and even unrelated snake species
161 8
N.-H . TAN et al.
and ICAO, 1956; BIItGER and BHATTI, 1989). Our results agree with this, indicating that the indirect ELISA for the quantitation of C. rhodostoma venom was not sufiïciently specific; in particular, extensive cross-reactions between venoms from the subfamily Crotalinae and the anti-C . rhodostoma venom were observed. The high level of cross-reactions with the Trimeresurus venoms studied has significant implications as many of the Trimeresurus snakes are also of medical importance in Malaysia. Thus, the use of indirect ELISA for immunodiagnosis of C. rhodostoma bite in Malaysia is of limited reliability if anti-C. rhodostoma venom is used as the immunoreagent. To improve the specificity of the indirect ELISA for C. rhodostoma venom, the use of more specific antibodies as immunoreagent is necessary. Sephadex G-100 gel filtration of the C. rhodostoma venom and examination of the indirect ELISA reactions between the venom fractions and the antibodies to C. rhodostoma as well as the five Malaysian Trimeresurus venoms indicated the presence of a `low indirect ELISA cross-reactivity region'. Presumably, these fractions contained antigens that exhibited low antigenic crossreactivity with the Trimeresurus venom antigens. VF52, the rechromatographed C. rhodostoma venom fraction that reacted strongly with anti-C. rhodostoma venom but showed minimum antigenic cross-reactivity with the antibodies to Trimeresurus venoms, was used to raise antibodies . Indirect ELISA with anti-VF52 as immunoreagent showed a definite improvement in specificity, in particular the cross-reactions with the Malaysian Trimeresurus have been reduced to between 2 and 12% . While the indirect ELISA is easier to set up compared to the double sandwich procedure, which requires the purification of IgG and preparation of the specific conjugate, the indirect ELISA is more time consuming. By modifying the coating procedures (e.g. 1 hr at room temperature), it is possible to complete the indirect ELISA within 2-4 hr of sampling, but preliminary experiments showed that the time of assay of the indirect ELISA of C. rhodostoma cannot be shortened to less than 2 hr without substantially reducing the sensitivity. The 2~ hr assay time after sampling is generally considered to be too slow for the clinician managing snakebite. Nevertheless, in the absence of immunoreagents for the double sandwich ELISA procedure, the indirect procedure is still a useful aid for confirmatory immunodiagnosis of the biting species, as well as a research tool in snake venom research. (K[rElcAttrn
Indirect
for C. rhodostoma venom antibodies THEAxsroN et al. (1977) proposed the use of antibody detecting indirect ELISA for retrospective diagnosis of snakebite. Some authors reported that this approach is of limited accuracy due to extensive cross-reactivity (Ho et al., 1986x). Our results, based on studies using laboratory-prepared venom antibodies, however, show that the indirect ELISA detection of C. rhodostoma antibodies is sufficiently specific for retrospective diagnosis of C. rhodostoma bite in Malaysia, in particular when VF52 was used as the coating antigen. It is, however, necessary to perform a parallel assay using T. albolabris or T. purporeomaculatus venom as coating antigen so that cross-reactions due to any Trimeresurus venom antibodies can be readily detected . ELISA
Double sandwich ELISA for C. rhodostoma venom As in the indirect ELISA, individual variation in the C. rhodostoma venom composition did not substantially affect the accuracy of quantitation for the double sandwich ELISA
ELISA of Malayan Pit Viper Venom
1619
procedures . Preliminary experiments showed that it was not necessary to use BSA as blocking reagents (YEO, K. H., M.Sc. Thesis, University of Malaya, Kuala Lumpur, 1992). Many authors reported that cross-reactivity was not a problem in venom detection assays using double sandwich ELISA (Z~iEAICSTON et al., 1977; COULTER et al., 1980; Ho et al., 1986x), although cross-reactions have also been reported occasionally (MARSHALL and HERRn1ANx, 1984). DHALIWAL et al. (1983) and Ho et al. (1986b) have also reported double sandwich ELISA procedures which were specific for C. rhodostoma venom. However, in their studies the specificity of the assay was not examined in detail . The present work shows that there were indeed no substantial cross-reactions between IgG anti-C . rhodostoma venom and the 26 other venoms from the families Viperidae and Elapidae in both ordinary and the rapid procedures. It should be noted, however, that the ELISA described here were conducted in buffers and the presence of blood in the sample may affect the results. Our results demonstrate that the double sandwich ELISA procedure for C. rhodostoma venom exhibited much higher specificity when compared to the indirect ELISA procedure . This difference in specificity may be due to: (1) double sandwich ELISA employed purified IgG antibodies while the indirect ELISA described employed antibodies that include both IgG and non-IgG antibodies; and (2) the inherent difference between the indirect ELISA, which can detect an antigen with only one epitope, and the double sandwich ELISA, which can only detect antigens with at least two epitopes . Experiments on double sandwich ELISA using the non-IgG antibodies as immunoreagent, however, rule out the first possibility. The time required for the rapid double sandwich ELISA for C. rhtxiostoma venom is approximately 50 min after sampling. This is comparable to that reported by DHALIWAL et al. (1983). We have not attempted to shorten further the time of assay required for the `rapid' double sandwich ELISA of C. rhodostoma venom, as preliminary experiments showed that this will require the use of excessive amounts of immunoreagents and will also reduce the sensitivity of the assay. However, in view of the pattern of the course and evolution of the C. rhodostoma venom poisoning (IZEm, 1968; WARRED et al., 1986), the rapid procedure described here is fast enough in particular as an aid to antivenom treatment of the bite. It has been suggested that in C. rhodostoma bite it is highly desirable to wait for clear clinical evidence of systemic poisoning before giving antivenom (REm, 1970). Aabwwledgcnunta-TThis work was supported by a research grant, IRPA-3-07-04097, from the Government of
Malaysia .
REFERENCES B~, B. J. and BHArn, A. R. (1989) Snake venom components end their cross-tnactivity: a review . Bioeltan . CeIJ. Biol. 67, 59701 . Boaifle, M. A. and Owrnnr, C. L. (1988) Rapid decline in blood antimyotoxin Ievels in the presence of myotorin A from prairie rattlesnake (Crotatus virldis vtridis) venom. CJi . Toxicot. 26, 303-312. BrtAna, V. and Voa'r, W. (1971) Immunization against cobra venom. Experientia 27, 1338-1339. CxAivntm, H. M. and Hvnx~-r., J. G. R (1982) A new enzyme immunoassay system suitable for field use and its application in snake venom detection kit. Clin . chin. Acts 121, 225-230. Cwes;, D. G., M~cMuAC~, D. D., Et.uar, E., Wot.oorr, R. G., L~rm~., A. M. and RAFrr av, M. A. (1972) Elapid neurotoxins. Purification, characterization and immunochemical studies of o-bungarotoxin. Btoclumistry 11, 1663--1668.
1620
N .-H . TAN et al.
Cotn.~t, A. R ., Hwxws, R. D. and Surt~rtt.wtvo, S . K . (1980) Enzyme immunoassay for the rapid clinical identification of snake venom . Med. J. Aunt. 1, 43335 . Dxwt .twws ., J . S ., L~tu, T . W. and SutcueswawN, K. D . (1983) A double antibody sandwich miao-ELISA kit for the rapid diagnosis of snakebite . S.E. Asian J. trop . Med. publ. Hlth 14, 367-372. Hwe®soxn, G . W., Stsrrts, S . J . and Noat~, G. R . (1980) Sensitivity and specificity of enzyme immunoassay for serodiagnosis of influenza A virus infection . J. inject. Dis. 141, 644-651 . Hwxs.ow, E . and Lwrta, D. (1988) Antibodies: a Laboratory Menial. New York: Cold Spring Harbor Laboratory . Hwxssorr, M . W . and Pwwtrtc, K . M . (1982) Enzyme immunoassay for direct detection of influenza type A and adenovirus antigens in clinical specimens. J. clip . Microbial. 15, 5-11 . Ho, M ., WwrtrtPr r , M . J WwseRer.L, D . J., BIDWELL, D. and Vot.t~at, A . (1986a) A critical reappraisal of the use of enzyme-linked immunosorbent assays in the study of snakebite . Toxicon 24, 211-221 . Ho, M ., WwttteEt.t., D. A ., Loowstsa~suwwx, S ., PHILLIPS, R. E ., C~swivTftwvwrrtcx, P., Kwxewwrrc, J., Suewlvwewrroxo, W., Vrtewvwlv, C ., Hu~I-roN, R . A. and Velcxo, S. (1986b) Clinical significance of venom antigen levels in patients envenomed by We Malayan pit viper (Calloselasrna rhodostoma) . Am. J. trop . Med. Hyg . 35, 579-587. HuosoN, L . and HwY, F. C . (1980) Practical Immunology . Oxford : Blackwell Scientific . Kuescwaxl, M . E . and Row, S. S. (1956) Antigenic composition of venons of poisonous snakes of India . In: Venons, pp. 175-182 (BUCKLEY, E . and PORGl3, N ., Eds) . Washington: American Association for the Advancement of Science. Lus, B . L. (1982) Poisonous Snakes ojPeninsular Malaysia, pp. 56-65 . Kuala Lumpur : Malayan Nature Society. lyl wQCwwr y, L. R. and Haexeswrtrr, R . P. (1984) Cross reactivity of bardick snake venom with death adder antivenom. Med. J. Aust. 140, 541-542 . MIrrroIV, S. A . (1987) Present tests for detection of snake venom : clinical applications . Ann . emergency Med. 16, 932J937 . Pt-reagoN, G . L. (1983) Determination of total protein. Meth. Enzym . 91, 95-119. RP.m, H. A . (1968) Symptomatology, pathology and treatment of land snakebite in India and Southeast Asia . In : Venomous Animals and Their Venons, Vol. 1, pp. 61142 (Buctmtes ., W ., BucKLev, E. E. and DsaLOPtav, V., Eds) . New York : Academic Press . Rr~n, H . A . (1970) The principles of snakebite treatment . Clin . Toxicol. 3, 474--082. TISewICSrox, R . D . G . (1983) The application of immunoassay techniques, including enzyme-linked immunosorbent assay (ELISA), to snake venom research. Toxicon 24, 341-352 . THFrtK.4rON, R. D. G., LLOVO-Jortss, M . J. and REIn, H . A . (1977) Micro-ELISA for detecting and assaying snake venom and venom antibody. Lmlcet U, 63942. Tirs~IV, P. (1985) Preparation of enzyme-antibody or other enzyme-macromolecule conjugates . In : Practice and Theory of Enzyme Immunoassays, pp . 221-278 (Buxoox, R. H . and Vwx Krr>PPExeP.Ito, P . H ., Eds). Amsterdam : Elsevier . WARRELL, D. A ., LOOART~UWAN, R., TIIEwICSrox, R . D . G ., PrnLLIPS, R . E., CHwxrxwvwxlcH, P., VIRAVAN, C ., SuPwrIwxwrtoxn, W ., Kwxewwxa, J ., Ho, M ., HuTTOIV, R. A . and VP.ICxo, S . (1986) Randomized comparative trial of three monospecific antivenoms for bites by the Malayan pit viper (Calloselasma rhodostoma) in southern Thailand: clinical and laboratory correlations. Am . J. trop. Med. Hyg . 36, 1235-1247.
0041-0101/92 SS.OD + .00 © 1992 Perpmon Prea Ltd
THE USE OF ENZYME-LINKED IMMUNOSORBENT ASSAY FOR THE QUANTITATION OF CALLOSELASMA RHODOSTOMA (MALAYAN PIT VIPER) VENOM AND VENOM ANTIBODIES NGSr-Hoxc Tax, KS-HUAT Y>~ and Mot-ro ISxwxDnR Nnc Jn~~It Department of Biochemistry, University of Malaya, Kuala Lumpur, Malaysia (Received 29 April 1992; accepted 11 July 1992) N.-H. Twx, K.-H . Y0o and M. I. Nnc J~nR. The use of enzyme-linked
immunosorbent assay for the quantitation of Calloselasma rhodostoma (Malayan pit viper) venom and venom antibodies. Toxicon 30, 1609-1620, 1992.-The specificity and sensitivity of an indirect and two (an `ordinary' and a `rapid') double sandwich enzyme-linked immunosorbent assay (ELISA) procedures for the quantitation of Calloselasma rhodostoma (Malayan pit viper) venom were examined. The three assays were equally sensitive and the accuracy of the assays was not substantially affected by individual variation in the venom composition . The specificity of the assays was examined against 26 venoms from snakes of the families Viperidae and Elapidae . While the double sandwich ELISA procedures were sufficiently specific to be used in the clinical immunodiagnosis of C. rhodostoma bite in Malaysia, the indirect ELISA procedure exhibited extensive cross-reactivity with other Malaysian pit viper venoms. Attempts were made to improve the specificity of the indirect ELISA procedure for the quantitation of C. rhodostoma venom . A `low ELISA crossreactivity' venom fraction (termed VF52) was isolated from C. rhodostoma venom by repeated Sephadex G-100 gel filtration chromatography . The indirect ELISA procedure using antibodies to VF52 as immunoreagent showed an improvement in specificity . The use of the indirect ELISA procedure for the detection of C. rhodostoma antibodies was also examined and the results show that the assay was sufficiently specific to be used for retrospective diagnosis of C. rhodostoma bite in Malaysia, in particular when VF52 was used as the coating antigen .
INTRODUCTION
ENZYbIE-LINKED immunosorbent assay (ELISA) is the most widely used technique for the immunodiagnosis of snakebite (CHANDLER and HURRELL, 1982; Ho et al., 1986x; MIxTON, 1987; THE~xs~rox, 1983; TI~xsTOx et al., 1977) . The double sandwich ELISA method is usually used for detecting snake venom in tissues and body fluids, while indirect ELISA is mainly used for detecting antibodies to venom resulting from previous envenoming. Specificity of ELISA for detecting venom and venom antibody is one of the major considerations in evaluating the applicability of the method for confirmatory and retrospective immunodiagnosis of snakebite . 1609
161 0
N.-H . TAN et al.
Several authors (DHALIWAL et al., 1983; Ho et al., 1986b) have described a double sandwich ELISA procedure for detection of Calloselasma rhodostoma (Malayan pit viper) venom, but the specificity of the assay has not been fully evaluated. Calloselasma rhodostoma is the commonest cause of snake venom poisoning in Malaysia (LIM, 1982). The present study was undertaken to determine the applicability of the indirect ELISA procedure for the quantitation of both Malayan pit viper venom and antibodies. For comparison purposes, the specificity of the double sandwich ELISA for Malayan pit viper venom was also examined . MATERIALS AND METHODS Materials The Malaysian snake venoms used in this study include those of Calloselasma rhodostoma (Malayan pit viper), Trimeresurvs purpereonwcelater (shore pit viper), T. albolabris (white-Tipped tree viper), T. popeorium (Pope's tree viper), T. sronatranus (Sumatran pit viper), T. wagleri (speckled pit viper), Enhydrina schistose (common sea snake), Naja raja sputatrix (Malayan cobra), Naja kaouthia (Monocellate wbra), Ophtophagus hannah (lung cobra), Bungarus cmrdidus (Malayan krait) and B. jasciatus (banded krait) . Other venoms included in the specificity studies are venoms of Agkistrodon b. btlirteatus, A . c. contortrix, A . p. piscivorur, Bothrops riper, B. atrox, Crotohu adamantees, C. atrox, Sistrwur c. ttrgtmitees, Trimeresurar jlavoviridir, Echis carinetus, Vipere a. ammodytes, Daboia russelli siamensis, Dendroaspis angusticeps, Notechis scutatus and Hoplocephalus atephensü. The venoms were obtained from Miami Serpentarium Laboratories (U .S.A .), Latoxan (France), Ophidia (Switzerland), Sigma Chemical Company (LI .S .A) and SEA Venom Institute (Malaysia) . Rabbits were supplied by Central Animal House, Faculty of Medicine, University of Malaya. Freund's adjuvant was obtained from Difco Laboratories (U.S .A.) . Goat anti-rabbit IgG-horseradish peroxidase conjugate was from Bio-ltad Laboratories (U.S .A.) . Micro-ELISA plates were from Nunc (Denmark). Protein A-agarose was from Sigma Chemical Company (U .S.A). Other chemicals and reagents used are of analytical grade and were all purchased from Sigma Chemical Company. Preparation of antiserum to snake venom Antisera against the pit viper venoms (C. rhodostoma, T. pwpureomaeelates, T. albolabris, T. Popeorum, T. sionatranes and T. waglere) or venom Erection 52 (VF52, see below) were preparod as follows: the venom (2 mg in 0 .1 ml phosphato-buB'ered saline, pH 7.4) was mixed with an equal volume of Freund's complete adjuvant and injected i .m. into the kft and right thighs, alternatively, of rabbits (2 kg) at biwcekly intervals. The rabbits were bled 9 days after the second booster. Antisera against the elapid venoms (N. n . sputatrix, N. kaoethia, O . hannah, B. candides and B. fasciates) were prepared using a hyperimmunization scheme and the venom antigen was cross-linked by glutaraldehyde (Brune and VoaT, 1971) prior to injection . The rabbit was immunized every 2 weeks for 3 months with increasing doses (0 .5 to 4 .0 mg) . The sera obtained after clotting were stored frozen until used. Determination ojprotein concentration Protein concentration was determined according to the Lowry's method (Pe~rrrssort, 1983). Indirect ELISA procedure The indirect ELISA procedure used was modified from Hw~rorro et al. (1980) and performed at room temperature (28°C) . The micro-ELISA plates were coated overnight at 4°C with the venom and washed by PBS-Tween solution . The antiserum or normal rabbit serum (control) was added to the wells and incubated for 1 hr. The wells were washed and goat anti-rabbit IgG conjugate was then added and incubated for 1 hr. After washing, substrate solution (H=O~ and O-phenylene diamine) was added. After 10 min, the reaction was stopped by adding 12.5% sulphuric acid and the absorbance at 492 nm was determined by a miau-ELISA reader. Preliminwry checkerboard titrations were conducted to obtain the optimum operation conditions . The C. rhodartoma venom used for immunization (Latoxan, lot A223, the snake was from southern Thailand) was used as the refenenoe C. rhodostoma venom in ELISA. Isolation of IgG and non-IgG antibodies and preproation ojlgG-horseradish peroxidast conjugate Antibodies produced against a venom are predominantly immunoglobulin G (IgG) but also include non-IgG antibodies such as IgM, IgA, IgD and IgE (Cr.eaa et ol., 1972) . IgG antibodies were purified using protein A
ELISA of Malayan Pit Viper Venom
161 1
affinity chromatography (Htmsox and Hev, 1980) . Non-IgG antibodies were isolated from the unbound proteins by ammonium sulphate precipitation at SO'/e saturation (Hexi.ow and Lexe, 1988) . The preparation of the IgCr-horseradish peroxidase (HRP) conjugate was carried out according to Tussnv (198 .
Double sandwich ELISA procedure Double sandwich ELISA procedure was conducted as modified from HNrvox and P~wr.n~ (1982). In the `ordinary' procedure of double sandwich ELISA for C . rluxlostoma venom, microtitre plates were coated overnight at 4°C with either purifiod IgG antibodies to C. rhodostoma or IgG from pre-immune serum (control) . The plates were washed four times with PHS-Tween solution and C. rhodastoma venom (or other appropriate venoms) was added and incubated for 2 hr. After washing, IgCrHRP conjugate was added and incubated for another 2 hr, followed by addition of the substrate solution . After 10 min the reaction was stopped by adding sulphuric acid . Using precoated microtitre plates, the total assay duration for this `ordinary' double sandwich ELISA procedure was approximately 5 hr. In the `rapid' double sandwich ELISA of C. rhodostorna venom, the microtitre plates were coated overnight at 4°C with IgG anti-C. rJrodostomn venom. After washing, the venom was added and incubated for 15 min at 28°C . This was followed by the sequential incubation of IgG conjugate for 15 min and substrate solution for 15 mice . Using precoated microtitre plates, the total assay duration was approximately 50 min. Double sandwich ELISA of snake venom by non-IgG anti-C. rhodostoma venom was carried out in the same way as the ordinary procedure, except that the microtitre plates were coated with the non-IgG antibodies (2 pg/ml). Reaction between the non-IgG antibodies and snake venom (100 ng/ml) was detected using the IgG anti-C. rhodosroma venom-HRP conjugate, and the control consisted of equivalent concentration of the nonIgG antibodies from pre-immune serum .
Sephadex G-100 gelfiltration chro»urtography ojC. rhodostoma venom and isolation oja 'low ELISA cross-
reactivity' venom fraction Calloselasma rJrodostoma venom (Latoxan lot A223) was fractionated by a Sephadex G-100 gel filtration column . Ovalbumin (mol. wt 45,000) and lysozyme (mol . wt 14,400) were used as mol . wt markers . The antigenantibody reactions between the venom fractions and the six monospecific antibodies to the Malaysian pit viper venoms (C. rhodostoma, T. purpureomaculatus, T. albolabris, T. popeorurrr, T. sumatramas, T. walglerr) were rxnmined using indirect ELISA procedure. The venom fractions were diluted 40,000-fold and the antibodies were diluted 2000-fold . Calloselasrrra rhodostoma venom fractions which showed high indirect ELISA reactions with anti-C. rlurdostoma venom but low reactions with the five anti-Trtmeresurus venom were identified as `low ELISA cross-reactivity region' (tubes 45-60), the tubes were pooled and further fractionated by a second Sephadex G-100 gel filtration chromatography. The reactions between rechromatographed fractions (1 :10,000) and the above-mentioned six monospacific antibodies were again expmined by indirect ELISA . Tube 52 of the rechromatographed fractions (termed venom fraction 52 or VF52), which reacted strongly with the anti-C . rhodostoma venom, exhibited minimum ELISA cross-reactions with the five anti-Trbneresurus venoms.
Statistical analysis The significance in difference was determined by Student's t-test . The level of significance was P = 0 .05 .
RESULTS
Indirect ELISA jar C. rhodostoma venom The indirect ELISA procedure for C. rhodostoma venom yielded a linear dose-response curve for the venom at concentrations of 5 ng/ml (absorbance = 0.2) to 50 ng/ml (absorbance = 1 .2) when the microtitre plates were coated overnight with the venom. A total of 13 different samples of Malayan pit viper venom from different commercial suppliers were used to examine the effect of individual ditFerences in Malayan pit viper venom on the antigen antibody reaction of the indirect ELISA quantitation of the venom. While different venom samples yielded slightly different absorbance readings ranging from 80%
1612
N.-H. TAN et al.
T~at.E 1 . CROSS-REACI'IVrrY HECWBQi ANTIBODn~ TO Catiosetasma rhodostoma vENOS+ eNn ort~x sNexE vENO~rs nv rtm tNOtaECr ELISA DETA.CI'ION OF Tt~ VENOM
Coating venom
O.D. as % of C. rhodeutorrra venom (mean1S.D ., n = 9-15) Anti-C . rhoetostoma venom Anti-VF52 as as antibodies antibodies
C. rhoetostoma Family CrotaGdae Subfamily Crotalinae A. b. bitineatus A. c. contortrix A. p. piscivores B. asper B. atrox C. adammiteur C. atrox S. c. tesgeminus T. Jtavoviridis T. albolabris T. popeorhan T. sumatranus T. pwpureomaculates T. wagleri
100
100
50 .7114.1 31 .212 .9 55 .618 .7 23 .314 .8 19 .711 .1 40 .415 .4 17 .715 .3 36 .417 .9 36 .713 .9 54 .915 .4 30.016 .E 21 .812.3 30.418.5 32.416.2
28 .210 .3 16 .712 .E 28 .910 .9 6.212 .2 4.913 .4 11 .410 .8 6.810.E 1 .610.5 14.510.2 2.910.1 8.410.4 2.110.3 11 .717.1 3 .310.4
Subfamily Vipcrinae E. ceuinatus V. a. ammodytes D. r. siamensis
12.412.8 17.414.E 22.915 .5
1.410.8 4.810.2 7 .610.2
Family Elapidae B. candides B. fasciates N. n. spetatrix N. kaoethia O. hannah D. angesticeps N. seetates H. stephensü E. schistosa
10 .611 .1 11 .915 .1 9 .51 I .6 15 .112 .2 15 .916 .3 11 .311 .0 8.912 .8 18 .912 .E 11 .512 .0
1.810 .1 4.411 .3 1.310 .1 2.310 .3 2.010 .7 6.910 .4 7.110 .4 4.l 11 .4 9.910 .8
The assay mixture contained 100 pl venom (25 ng/ml), 100 tel anti-C. rhodostoma venom (1 : 2000) and 100pl conjugate (1 : 4000). The absorbance at 492 nm for C. rhodostoma venom was 1.0410.05 with anti-C . rhodostoma venom, and 0.4610.03 with anti-VF52.
to 108% of the reference venom sample, the readings were generally not statistically significant (P > 0.05) from the reference sample . Extensive cross-reactions between the anti-C. rhodostoma venom and venoms of the Crotalinae subfamily (particularly venoms from the three Agkistrodon and the genus Trimeresurus) were observed in the indirect ELISA (Table 1). On the other hand, there were far fewer cross-reactions between the anti-C. rhodostoma and venoms of snakes from the subfamily Viperinae and the family Elapidae. In an attempt to shorten the assay time for C. rhodostoma venom, the microtitre plates were coated with the venom for 1 hr at room temperature instead of overnight coating at 4°C. This also yielded a linear dose-response curve (not shown) for the venom at concentrations of 5 ng/ml (absorbance = 0.2) to 50 ng/ml (absorbance = 0.78). Further reduction in the time for coating reduced the sensitivity substantially.
ELISA of Malayan Pit Viper Venom
161 3
Teet~ 2 . Srecuacmr oF nm txntsECr ELISA rnocFnuxE t~t T~ nsrecnoN op Calioadatma rhodostotna verroa ~xneonos
Venom antibodies Anti-CaUoselasma rhodostoma Anti-T. albolabris Anti-T. popeorium Anti-T. .rronatranus Anti-T. purpereonracuJates Anti-T. wagkri Anti-B. candides Anti-B. jasciater Anti-Naja n. spetatrix Anti-Naja kaoet6ia Anti-Ophiophagua hanna6
O .D. as % of anti-C. rhodoatwna C. rhodoatoma Venom fraction 52 as antigen (VF52) as antigen 100 20 .3 16 .5 5 .8 I5 .4 12 .2 0 0 3 .2 5 .9 2 .0
100 (0.0 10.0 1 .3 6.3 0 n.d . n.d . n.d . n .d. n .d.
The assay mixture contained 100 pl venom antigen (25 ng/ml for C. rhodostoma venom, and 100 ng/ml for VF52 antigen), 100 ~l venom antibodies (1 : 2000) and 100 pl conjugate (1 : 4000) . when anti-C. rhodwstoma venom was used as antibodies, the absorbance at 492 nm for C. rhodoatwna venom antigen was 1 .04 f 0 .05, and for VF52 was 1 .0810 .02 . Value waa mean of 9-15 determinations . n .d. = not determined .
Detection of C. rhodostoma venom antibodies using indirect ELISA When the micro-ELISA plates were coated with 100 ng/ml of C. rhodostoma venom, the indirect ELISA for detection of C. rhodostoma venom antibodies yielded an exponential dose-response curve for the laboratory-prepared C. rhodostoma venom antibodies from a dilution of 1 :2000 (absorbance = 1.2) to 1 :20,000 (absorbance = 0.2) (not shown) . Table 2 shows the specificity of the indirect ELISA procedure for the detection of C. rhodostoma venom antibodies. Antibodies to venoms of the Malaysian elapids (including B . candidus, B. fasciatus, N. n. sputatrix, N. kaouthia and O. hannah) and that of T. sumatranus yielded low cross-reactions ( < 6%), whereas the anti-albolabris, anti-T. popeorium, anti-T. purpureomaculatus and anti-T. wagleri yielded 12-20% cross-reactions .
Double sandwich ELISA of C. rhodostoma venom The double sandwich ELISA procedure yielded an exponential dose-response curve for C. rhodostoma venom at concentrations from 5 n~/ml (absorbance = 0 .26) to 100 ng/ml (absorbance = 0.95). To examine the effect of individual variation in C. rhodostoma venom composition on the ELISA absorbance readings, a total of 13 different samples of C. rhodostoma venom from different commercial suppliers were used as venom antigens . In general, the venom samples yielded absorbance readings comparable to that of the reference venom sample . The specificity of the double sandwich ELISA for the detection of C. rhodostoma venom was examined by the reaction between IgG anti-C. rhodostoma and venoms from snakes of the families Viperidae and Elapidae . Table 3 shows that there were no significant crossreactions between the IgG anti-C. rhodostoma antibodies and all the 26 venoms tested . The possibility that non-IgG anti-C . rhodostoma venom may contribute to the extensive cross-reactions observed in the indirect ELISA was examined by using the double sandwich ELISA procedure with non-IgG anti-C. rhodostoma as coating antibodies. The
1614
N :H . TAN et al.
Twet.E 3. Cross-xEecnvrrv ~EEx exneonms ro Calloselasma rhodostoma vENOSr ~Nn orrme sNe¢E vENOMs uv rt~ nouaLE seNnw~cH ELISA DETA.CfION OF rt~ VENOM
Venom C. rhodostoma
O .D . as % of C. rhodostoma venom (mean f S .D., n = 9-1 ~ Ordinary procedure Rapid procedure 100
100
1 .5 f 0.1 I .6 f 0 .1 1 .9 f 0.3 0.6 t 0.2 0 0 .7 t 0 .1 1 .9 f 0.2 0.1 f 0.1 1 .9 t 0.2 0 .2 t 0 .1 0 0.1 t 0.1 0.1 f0.1 0
3 .8 f 0 .1 2 .9 f 0 . I 1 .7 f 0.3 l .2 t 0 .9 1 .6t0 .3 1 .4 t 0 .3 1 .8 f 0.0 1 .0 t 0.1 1 .4 t 0 .3 0 .3 t 0 .2 1 .4 f 0 .3 0.9 ~ 0.4 0.8i0.3 0 .8 f 0 .0
Subfamily Viperinae E. carinatus V. a . ammodytes D . r. siamerrsis
0 0 0
0.8 i 0.0 0.6 f 0.0 0
Family Elapidae B . candides B . jasciates N. n . sputatrix N. kaouthia O .hannah D. angusticeps N. scetatus If. stephensü E. schistosa
0 0 0 0 0 0 0 0 0
0 0.7 t 0.3 0 .2 f 0.2 0 0 0 0 0 0
Family Crotalidae Subfamily Crotalinae A . b . bilineatus A . c. contortrix A . p . piscivorus B. asper B. atrox C. adamanteus C. atrox S. c . tergemim~s T. Jlavoviridis T. albolabris T. popeorhen T. sumatramis T. purpureomaculatus T. wagleri
The assay mixture contained 100 pl IgG anti-C. rhodostorna venom (2 pg/ml for the ordinary assay, 10 pg/ml for the rapid assay), 1001e1 of venom (100 ng/ml), and 100 ul of conjugate (1 : 3000 for the ordinary assay, 1 : 1000 for the rapid assay) . Absorbance at 492 nm was 0.93 f 0.02 for the ordinary assay, and 0 .77 f 0 .01 for the rapid assay.
results show that the non-IgG antibodies did not react with C. rhodostoma venom, nor with any of the other venoms tested . Rapid double sandwich ELISA of C. rhodostoma venom
The `rapid' double sandwich ELISA procedure for C. rhodostoma venom yielded a linear dose-response curve ranging from 5 to 25 ng/ml venom (absorbance readings between 0 .05 to 0 .24), but an exponential dose-response curve was obtained when the venom concentration was in the range of 25 to 100 ng/ml (absorbance readings between 0.24 and 0.7~. Different samples of C. rhodostoma venom yielded only slightly different absorbance readings . Table 3 shows that the rapid double sandwich procedure was also
ELISA of Malayan Pit Viper Venom
16(5
Fto. l.a,b . Ixnu~cr ELISA rAOt+a~s of Calloaelavna rhodastoma vwoM t~cnoxs oe~r~x~ av St~n~c G-(00 0~. tru xs~nox c~neoauroaxeexv. Two-hundred milligrams of C. rl~odoatoma venom, dissolved in 12 ml of PBS (pH 7.2) was applied to the column (2 x 100cm) equilibrated with the same buffer. Tube volume = 4.5 ml . Miaotitre plates were coated at 4°C with 100pl of the venom fraction (1 :40,000) and incubated with 100 pl of the following antibodies (1 :2000) : A, anti-C. rhodartoma venom; B, anti-T. albolabris venom; C, anti-T. popeornon venom; D, anti-T. pwpureomaculatur venom; E, anti-T. smnatrmws venom; F, anti-T. wagleri venom . -, absorbante at 280 am ; , absorbante of the indirect ELISA reaction mixture at 492 nm .
very specific and that there were only minimum cross-reactions with all the 26 venoms tested . Antigenic studies of C. rhodostoma venom jractions
Sephadex G-100 gel filtration chromatography resolved the C. rhodostoma venom into three protein peaks (Fig. la), corresponding to high mol. wt ( > 45,000), medium mol. wt (14,0005,000) and low mol. wt ( < 14,000) proteins, respectively . The antigen-antibody reactions between the gel-filtration fractions of C. rhodostoma venom and the monospecific antisera to six Malaysian pit viper venoms (C. rhodostorna and the five Trimeresurus) were examined using indirect ELISA (Fig. 1b). The medium mol. wt venom fractions yielded high indirect ELISA reading (absorbance > 0.8) while the high and low mol. wt protein fractions yielded moderate (absorbance of 0.3-0.8) and low (absorbance < 0.3) ELISA readings, respectively. Comparison of the indirect ELISA profiles of the antigen-antibody reactions between the C. rhodostoma venom fractions and the six antisera to the Malaysian pit viper venoms (C . rhodostoma and the five Trimeresurus) indicated the presence of a `low ELISA cross-reactivity' region (Fig. l, tubes 45-60) .
1616
N: H. TAN et aJ.
b
3~2
B I I~ I I I II I I I
2~6 2"4 2,0 _ ~ 1,B-
I I
Fia. 1 . (Cont~~
-o~e
n
"I I i 1
1~
0 .4 ~ ' 0
F- 1~2
1
0~6 -
0
- ~ .8
Î 1 1 1
1 .2 -
Tube Number
~ 2 .0
I 1 1 1
20
40 60 Tube Number
1
-0~4
80
0
E c `N O W
` "
ELISA of Malayan Pit Viper Venom
161 7
These tubes were pooled and the pooled fractions were rechromatographed on the same Sephadex G-100 gel filtration column to yield a major protein peak (tubes 400) and a very small, minor protein peak (not shown) . Anti-C . rhodostoma venom yielded high indirect ELISA readings with the major protein peak only, while the five anti-Trimeresurus venoms all yielded low indirect ELISA readings with both the major and the minor protein peaks. Tube 52 (VF52) of this rechromatographed `low ELISA cross-reactivity' venom fraction yielded maximum indirect ELISA reading with anti-C . rhodostoma venom and very low indirect ELISA readings with the five anti-Trimeresurus (see Table 2). The fraction was used to raise antibodies (anti-VF52). Improved indirect ELISAs for C. rhodostoma venom and venom antibodies using antiVF52 and VF52 as immunoreagents
Table 1 shows that the use of anti-VF52 as immunoreagent for indirect ELISA for C.
rhodostoma venom indeed reduced substantially the cross-reactions with other venoms . In general, venoms from the same Agkistrodontini tribe yielded less than 30% crossreactions, while the cross-reactions for venoms from the genus Trimeresurus were below
15%. The elapid venoms yielded less than 10% cross-reactions . Table 2 shows that with VF52 as coating antigen, the indirect ELISA for C. rhodostoma antibodies was also more specific; the cross-reactions with other anti-Trimeresurus were less than 10% .The indirect ELISA using VF52 as coating antigen was as sensitive as the one using whole venom as antigen. DISCUSSION
Indirect ELISA for C. rhodostoma venom The indirect ELISA for C. rhodostoma was as sensitive as the double sandwich method . Individual variation in the C. rhodostoma venom composition generally did not result in substantial differences in the indirect ELISA readings . Thus, the use of any C. rhodostoma
venom available to construct the standard curve is not likely to greatly affect the accuracy of the assay. The presence of high background readings due to non-specific adsorption of immunoreagents to the microtitre wells is one of the major problems affecting specificity of the ELISA (Ho et al., 1986x) . Tween 20 and non-specific proteins such as bovine serum albumin (BSA) have been used in indirect ELISA procedure to eliminate or decrease the background readings (BoHEIt and OwxsY, 198). In the present procedure, Tween 20 was incorporated in the washings. Preliminary studies, however, showed that the use of BSA as a blocking reagent affected the reaction between venom antigen and antibody (Y$o, K. H., M.Sc. Thesis, University of Malaya, Kuala Lumpur, 1992). As the specific venomantibody reaction in the present study is sufficiently strong to allow the background to be acceptable by simply subtracting the background from specific reaction, BSA was not used as a blocking agent in the assay. Specificity is one of the major considerations in evaluating applicability of ELISA for immunodiagnosis of snakebite (Ho et al., 1986x ; MAIesxALI, and Haxx~nxx, 1984) . Specificity is generally defined as the lack of cross-reactivity between closely related species . Many authors reported that the ELISA systems which used poorly characterized snake venoms or snake venom antibodies as reagents had low specificity (Mu~rrox, 198, as common antigens occur in venoms of related and even unrelated snake species
161 8
N.-H . TAN et al.
and ICAO, 1956; BIItGER and BHATTI, 1989). Our results agree with this, indicating that the indirect ELISA for the quantitation of C. rhodostoma venom was not sufiïciently specific; in particular, extensive cross-reactions between venoms from the subfamily Crotalinae and the anti-C . rhodostoma venom were observed. The high level of cross-reactions with the Trimeresurus venoms studied has significant implications as many of the Trimeresurus snakes are also of medical importance in Malaysia. Thus, the use of indirect ELISA for immunodiagnosis of C. rhodostoma bite in Malaysia is of limited reliability if anti-C. rhodostoma venom is used as the immunoreagent. To improve the specificity of the indirect ELISA for C. rhodostoma venom, the use of more specific antibodies as immunoreagent is necessary. Sephadex G-100 gel filtration of the C. rhodostoma venom and examination of the indirect ELISA reactions between the venom fractions and the antibodies to C. rhodostoma as well as the five Malaysian Trimeresurus venoms indicated the presence of a `low indirect ELISA cross-reactivity region'. Presumably, these fractions contained antigens that exhibited low antigenic crossreactivity with the Trimeresurus venom antigens. VF52, the rechromatographed C. rhodostoma venom fraction that reacted strongly with anti-C. rhodostoma venom but showed minimum antigenic cross-reactivity with the antibodies to Trimeresurus venoms, was used to raise antibodies . Indirect ELISA with anti-VF52 as immunoreagent showed a definite improvement in specificity, in particular the cross-reactions with the Malaysian Trimeresurus have been reduced to between 2 and 12% . While the indirect ELISA is easier to set up compared to the double sandwich procedure, which requires the purification of IgG and preparation of the specific conjugate, the indirect ELISA is more time consuming. By modifying the coating procedures (e.g. 1 hr at room temperature), it is possible to complete the indirect ELISA within 2-4 hr of sampling, but preliminary experiments showed that the time of assay of the indirect ELISA of C. rhodostoma cannot be shortened to less than 2 hr without substantially reducing the sensitivity. The 2~ hr assay time after sampling is generally considered to be too slow for the clinician managing snakebite. Nevertheless, in the absence of immunoreagents for the double sandwich ELISA procedure, the indirect procedure is still a useful aid for confirmatory immunodiagnosis of the biting species, as well as a research tool in snake venom research. (K[rElcAttrn
Indirect
for C. rhodostoma venom antibodies THEAxsroN et al. (1977) proposed the use of antibody detecting indirect ELISA for retrospective diagnosis of snakebite. Some authors reported that this approach is of limited accuracy due to extensive cross-reactivity (Ho et al., 1986x). Our results, based on studies using laboratory-prepared venom antibodies, however, show that the indirect ELISA detection of C. rhodostoma antibodies is sufficiently specific for retrospective diagnosis of C. rhodostoma bite in Malaysia, in particular when VF52 was used as the coating antigen. It is, however, necessary to perform a parallel assay using T. albolabris or T. purporeomaculatus venom as coating antigen so that cross-reactions due to any Trimeresurus venom antibodies can be readily detected . ELISA
Double sandwich ELISA for C. rhodostoma venom As in the indirect ELISA, individual variation in the C. rhodostoma venom composition did not substantially affect the accuracy of quantitation for the double sandwich ELISA
ELISA of Malayan Pit Viper Venom
1619
procedures . Preliminary experiments showed that it was not necessary to use BSA as blocking reagents (YEO, K. H., M.Sc. Thesis, University of Malaya, Kuala Lumpur, 1992). Many authors reported that cross-reactivity was not a problem in venom detection assays using double sandwich ELISA (Z~iEAICSTON et al., 1977; COULTER et al., 1980; Ho et al., 1986x), although cross-reactions have also been reported occasionally (MARSHALL and HERRn1ANx, 1984). DHALIWAL et al. (1983) and Ho et al. (1986b) have also reported double sandwich ELISA procedures which were specific for C. rhodostoma venom. However, in their studies the specificity of the assay was not examined in detail . The present work shows that there were indeed no substantial cross-reactions between IgG anti-C . rhodostoma venom and the 26 other venoms from the families Viperidae and Elapidae in both ordinary and the rapid procedures. It should be noted, however, that the ELISA described here were conducted in buffers and the presence of blood in the sample may affect the results. Our results demonstrate that the double sandwich ELISA procedure for C. rhodostoma venom exhibited much higher specificity when compared to the indirect ELISA procedure . This difference in specificity may be due to: (1) double sandwich ELISA employed purified IgG antibodies while the indirect ELISA described employed antibodies that include both IgG and non-IgG antibodies; and (2) the inherent difference between the indirect ELISA, which can detect an antigen with only one epitope, and the double sandwich ELISA, which can only detect antigens with at least two epitopes . Experiments on double sandwich ELISA using the non-IgG antibodies as immunoreagent, however, rule out the first possibility. The time required for the rapid double sandwich ELISA for C. rhtxiostoma venom is approximately 50 min after sampling. This is comparable to that reported by DHALIWAL et al. (1983). We have not attempted to shorten further the time of assay required for the `rapid' double sandwich ELISA of C. rhodostoma venom, as preliminary experiments showed that this will require the use of excessive amounts of immunoreagents and will also reduce the sensitivity of the assay. However, in view of the pattern of the course and evolution of the C. rhodostoma venom poisoning (IZEm, 1968; WARRED et al., 1986), the rapid procedure described here is fast enough in particular as an aid to antivenom treatment of the bite. It has been suggested that in C. rhodostoma bite it is highly desirable to wait for clear clinical evidence of systemic poisoning before giving antivenom (REm, 1970). Aabwwledgcnunta-TThis work was supported by a research grant, IRPA-3-07-04097, from the Government of
Malaysia .
REFERENCES B~, B. J. and BHArn, A. R. (1989) Snake venom components end their cross-tnactivity: a review . Bioeltan . CeIJ. Biol. 67, 59701 . Boaifle, M. A. and Owrnnr, C. L. (1988) Rapid decline in blood antimyotoxin Ievels in the presence of myotorin A from prairie rattlesnake (Crotatus virldis vtridis) venom. CJi . Toxicot. 26, 303-312. BrtAna, V. and Voa'r, W. (1971) Immunization against cobra venom. Experientia 27, 1338-1339. CxAivntm, H. M. and Hvnx~-r., J. G. R (1982) A new enzyme immunoassay system suitable for field use and its application in snake venom detection kit. Clin . chin. Acts 121, 225-230. Cwes;, D. G., M~cMuAC~, D. D., Et.uar, E., Wot.oorr, R. G., L~rm~., A. M. and RAFrr av, M. A. (1972) Elapid neurotoxins. Purification, characterization and immunochemical studies of o-bungarotoxin. Btoclumistry 11, 1663--1668.
1620
N .-H . TAN et al.
Cotn.~t, A. R ., Hwxws, R. D. and Surt~rtt.wtvo, S . K . (1980) Enzyme immunoassay for the rapid clinical identification of snake venom . Med. J. Aunt. 1, 43335 . Dxwt .twws ., J . S ., L~tu, T . W. and SutcueswawN, K. D . (1983) A double antibody sandwich miao-ELISA kit for the rapid diagnosis of snakebite . S.E. Asian J. trop . Med. publ. Hlth 14, 367-372. Hwe®soxn, G . W., Stsrrts, S . J . and Noat~, G. R . (1980) Sensitivity and specificity of enzyme immunoassay for serodiagnosis of influenza A virus infection . J. inject. Dis. 141, 644-651 . Hwxs.ow, E . and Lwrta, D. (1988) Antibodies: a Laboratory Menial. New York: Cold Spring Harbor Laboratory . Hwxssorr, M . W . and Pwwtrtc, K . M . (1982) Enzyme immunoassay for direct detection of influenza type A and adenovirus antigens in clinical specimens. J. clip . Microbial. 15, 5-11 . Ho, M ., WwrtrtPr r , M . J WwseRer.L, D . J., BIDWELL, D. and Vot.t~at, A . (1986a) A critical reappraisal of the use of enzyme-linked immunosorbent assays in the study of snakebite . Toxicon 24, 211-221 . Ho, M ., WwttteEt.t., D. A ., Loowstsa~suwwx, S ., PHILLIPS, R. E ., C~swivTftwvwrrtcx, P., Kwxewwrrc, J., Suewlvwewrroxo, W., Vrtewvwlv, C ., Hu~I-roN, R . A. and Velcxo, S. (1986b) Clinical significance of venom antigen levels in patients envenomed by We Malayan pit viper (Calloselasrna rhodostoma) . Am. J. trop . Med. Hyg . 35, 579-587. HuosoN, L . and HwY, F. C . (1980) Practical Immunology . Oxford : Blackwell Scientific . Kuescwaxl, M . E . and Row, S. S. (1956) Antigenic composition of venons of poisonous snakes of India . In: Venons, pp. 175-182 (BUCKLEY, E . and PORGl3, N ., Eds) . Washington: American Association for the Advancement of Science. Lus, B . L. (1982) Poisonous Snakes ojPeninsular Malaysia, pp. 56-65 . Kuala Lumpur : Malayan Nature Society. lyl wQCwwr y, L. R. and Haexeswrtrr, R . P. (1984) Cross reactivity of bardick snake venom with death adder antivenom. Med. J. Aust. 140, 541-542 . MIrrroIV, S. A . (1987) Present tests for detection of snake venom : clinical applications . Ann . emergency Med. 16, 932J937 . Pt-reagoN, G . L. (1983) Determination of total protein. Meth. Enzym . 91, 95-119. RP.m, H. A . (1968) Symptomatology, pathology and treatment of land snakebite in India and Southeast Asia . In : Venomous Animals and Their Venons, Vol. 1, pp. 61142 (Buctmtes ., W ., BucKLev, E. E. and DsaLOPtav, V., Eds) . New York : Academic Press . Rr~n, H . A . (1970) The principles of snakebite treatment . Clin . Toxicol. 3, 474--082. TISewICSrox, R . D . G . (1983) The application of immunoassay techniques, including enzyme-linked immunosorbent assay (ELISA), to snake venom research. Toxicon 24, 341-352 . THFrtK.4rON, R. D. G., LLOVO-Jortss, M . J. and REIn, H . A . (1977) Micro-ELISA for detecting and assaying snake venom and venom antibody. Lmlcet U, 63942. Tirs~IV, P. (1985) Preparation of enzyme-antibody or other enzyme-macromolecule conjugates . In : Practice and Theory of Enzyme Immunoassays, pp . 221-278 (Buxoox, R. H . and Vwx Krr>PPExeP.Ito, P . H ., Eds). Amsterdam : Elsevier . WARRELL, D. A ., LOOART~UWAN, R., TIIEwICSrox, R . D . G ., PrnLLIPS, R . E., CHwxrxwvwxlcH, P., VIRAVAN, C ., SuPwrIwxwrtoxn, W ., Kwxewwxa, J ., Ho, M ., HuTTOIV, R. A . and VP.ICxo, S . (1986) Randomized comparative trial of three monospecific antivenoms for bites by the Malayan pit viper (Calloselasma rhodostoma) in southern Thailand: clinical and laboratory correlations. Am . J. trop. Med. Hyg . 36, 1235-1247.
