Standardized quantitative enzyme linked immunoassay for antibodies to Toxoplasma gondii

0095-1137/80/06-0675/07$02.00/0

Standardized

Quantitative Enzyme-Linked Immunoassay for

Antibodies

to

Toxoplasma

gondii

T. M.LIN,' S. P.HALBERT,`* ANDG. R.O'CONNOR2

CordisLaboratories, Miami, Florida331271; andFrancis I. ProctorFoundationfor Research in

Ophthalmology, University

of California,

SanFrancisco,

California

941432

A quantitative and standardized enzyme-linked immunosorbent assay is

de-scribed which uses lyophilized antigen-coated disks for the detection of human

antibodies to Toxoplasma gondii. It doesnotrequire serial dilution of thetest

specimen, and theobjective absorbance readingsareconverted into international

unitspermillilitertraceabletothe World Health Organization's reference

stan-dardpreparation oftoxoplasma antibodies. It was shon.-<m tobe highly specific,

reproducible, and sensitive. The reagents were stable, without biohazard, and ready foruse. Studies of variousparametersin the assayindicated that thetest

conditions were relatively flexible. The enzyme-linked immunosorbent assay

results correlated satisfactorily with the methylene blue dye test, the indirect immunofluorescence test,and the passivehemagglutinationtest.

The diagnosis of human toxoplasmosis has depended heavily on serological methods for

detecting antibodies to Toxoplasma gondii. These principally include the classical

Sabin-Feldman methylene blue dye test (MBD), the indirect immunofluorescence test (IIF),and the

passive hemagglutination test (PHA). The cur-rentmethodsofchoiceforclinicalpurposes are

the MBD andIIFtests(8). However, the former

uses liveorganisms andascarce"accessory fac-tor,"andIIFrequiresafluorescent microscope and well-standardized fluorescein-labeled

anti-body preparations.

Recently, enzyme-linked immunosorbent as-says(ELISA) have been appliedto the detection

oftoxoplasma antibodies, particularlyin micro-titerplates (1, 2, 16-18). The microtiter ELISA is generally semiquantitative and may give

er-ratic results due to the nonuniform absorbent

quality of the plastic microtiter plates (3). In our laboratories, an ELISA for detecting

hepatitis B surface antigeninhuman blood

sam-ples has been developed which utilizes small

plastic disks for covalent binding of the

solid-phase reagents (6). This procedure was

incor-poratedinto adiagnostic kit (Cordia H,Cordis

Laboratories, Miami, Fla.). As reported here, this format was modified for the quantitative determination oftoxoplasma antibodies in hu-manspecimens. Inaddition, to establish a uni-form basis ofreference in thereporting of results, we measured thetoxoplasma antibody levels in international units permilliliter, providing val-ues directlytraceable to the World Health

Or-ganization's (WHO) firststandard human

anti-toxoplasmaserumpreparation (19). In the

pres-ent ELISA technique, serial dilution titrations and pretreatment of the test samples are not required;thetotal incubationtime is short (1h and 50min); and either serumorplasmacanbe used. The communication describesseveral as-pects of thegeneralprocedure and comparative results withMBD, IIF,and PHA assays.

MATERIALS AND METHODS T. gondii organisms of the RH strain were har-vested by differential centrifugation from peritoneal

exudatesof3-dayinfected mice with

heparin-contain-ingsaline (1 mg/nm). The suspensions were centri-fuged and washed three timesat 4°C (3,000 xg, 10

min). Afterbeingfrozenat-20°Catleastovernight,

theywerethawed,sonicated,andcentrifugedathigh

speed (45,000xg,30min).The supernatant (soluble

antigen) was collected and diluted with 0.1 M

NaHCO3, pH 9.6, to an appropriate concentration.

Special plasticdiskscontainingsurfaceisothiocyanate

groups(13-mmdiameter)weremixed with thisantigen solution, agitatedat 4°C overnight, washed, and ly-ophilized. These dried antigen-coateddisks

(contain-ing3 to 5ytg ofproteineach)werestoredat4°Cin the

presenceofadesiccant untiluse.Antibodiestohuman

immunoglobulinsG and Mwereseparately produced ingoats,salted outwithammoniumsulfate, and im-munospecifically purified. Theywerethenconjugated to calfalkaline phosphatase (PL Biochemicals, Mil-waukee, Wis.)withglutaraldehyde(Polysciences,Inc.,

Warrington, Pa.) (5). Thedialyzedantibody-enzyme

conjugatesweredilutedtoworkingconcentrationsin

buffer at pH 8.0 (0.0025 M

tris(hydroxymethyl)-aminomethane [Tris], 0.2% bovine serum albumin,

0.001 M MgCl2, 0.15 M NaCl, and 0.03 M sodium

phosphate)containing0.02%sodiumazide.Theywere

mixedproportionallyforuse andkeptsterile at 2 to

8°Cwhen not in use.

675

on February 7, 2020 by guest

http://jcm.asm.org/

676 LIN, HALBERT, AND O'CONNOR

The human positive control was a pool of eight recalcified plasmas containing relatively high levels of Toxoplasma antibodies. It served as the secondary standard and was calibrated by the present ELISA method against the WHO's international standard for anti-toxoplasma serum, human, first standard, 1959 (19). The humannegative control was a pool of recal-cified humanplasma with negligible amounts of toxo-plasmaantibodies (less than 20 IU/ml). Both controls werepreserved with 0.1% sodium azide, made sterile by filtration, andstored in portions at 4°C until use.

Foreachtest, one of the antigen-coated disks was incubated with 10

Id

ofsample diluted in 500 Ml of specimen diluent (containing 0.01 M phosphate-buffered saline [pH 8.0], 6% bovine serum albumin, 0.05% Tween 20, and 0.1% sodium azide). After incu-bation at37°C in a shaking water bath for 45 min, the samplesolution was removed by aspiration, and the

antigendisk (with antibody attached if present) was

washed five times each with 2.5 ml of 0.002 M Tris-bufferedsaline, pH 7.2 to 8.0. Antibody-enzyme con-jugate(0.5ml)wasthenaddedtothedisk, which was incubated as above for 45 min. The disk was again washed five times as above, transferred to anunused vial, and incubated as above for 20 min with 1 ml of theenzymesubstrate,p-nitrophenylphosphate (1 mg/ ml in 0.028 M sodium bicarbonate buffer [pH 9.8], containing 0.01 M MgCl2).The enzyme activity was stopped with 0.1 ml of3N NaOH and read with a spectrophotometer at a wavelength of 405 nm. To facilitate the assays,equipmentused for theCordia H

procedure (Cordis Laboratories) was employed, i.e., devices fortransferring and washing the disks.

The positive control and three fivefold serial dilu-tionsof it inspecimendiluent were tested simultane-ously in each assay. The absorbance values and anti-toxoplasmaactivity(internationalunits permilliliter) wereplotted onlog-log paper to establish a standard

curvefrom which theanti-toxoplasma activity of each

testsamplewasconvertedtointernational units per

milliliter,based on its absorbancereading. Tests with the WHO standard, as well as with the secondary

positive control, revealed that thestraight-line rela-tionship between absorbance readings and the anti-toxoplasma activity covered a range from 2 to 1,000 IU/ml. Because of the linearity of the relationship, the conversionfrom absorbance valuetointernational

units per milliliter could be made easily and rapidly with minimum errorwith an inexpensive hand-held calculator (e.g., Texas Instruments TI-55).

Inaddition to bloodsamples from local volunteers

or blood donors, we obtained serum samples from three laboratories doing routinetoxoplasma antibody immunoassays for clinical purposes and assayed them

ina codedmanner (after they had been tested in the laboratories supplying them) by the PHA, MBD, or IIFprocedures. Comparisons of titers obtained by the present ELISA test and these other tests were ana-lyzed bythecorrelation coefficient ofPearson and Lee

(R) (10) after logarithmic transformation (15), with

thenegative results assuming the titers to be either 1,

or oneor moredilutionlevel(s) below the lowest level

determinedinthetitrationsystem (15).The nonpar-ametricrankcorrelation coefficient ofSpearman (R.)

(10) wasalso usedto avoid the bias due tothis

as-sumption for negativeresults, especiallywherethey coveredawide range of titers (e.g.,less than16).For the same reason, where the sample size was large enough forameaningful conclusion,results from

sam-ples with titersof16orgreaterwere alsoseparately

analyzedby R. In additiontotheconventional arith-metic calculation for standard deviation and coeffi-cient ofvariation, the geometric standard deviation andgeometricmean werealso used forcomparisonof

reproducibilityof thepresentELISA with othertest

results, asrecomme'nded bythe Center for Disease Control (15).

RESULTS

Atypical standardcurve establishedwith

se-rialfivefold dilutions of the positive control is

shown inFig. 1. It canbe seenthat a straight-linerelationshipwasobtainedon alog-logscale

and that the coefficient of correlation of the

international unitspermilliliter andabsorbance

values in the example shown was 0.998. This

coefficientwasgenerallygreater than 0.995.

Sim-ilarstraight-line relationshipswere found with

the WHO primary standard, as well as with

severalindividual high-titeredserumspecimens. The

slopes

of these linestendedtobesimilar.

Numerousparametersinthe assay were

var-ied singly to study their effects on the final

absorbance values, as well as on the linearity andslopesof thestandardcurves.Thevariables included such factorsastheamount ofantigen used for coating the disks, incubation times, volumes oftestsample, volumesofconjugates, etc.Theresultsaresummarized in Fig.2.When variousamountsofantigenwereusedfor

coat-ing, the absorbance achieved reachedaplateau

at 5 ug/disk. However, it was found that the

absorbance values for the negative control and background decreased when the diskwascoated

with moreantigen,resultingingreater slopes for

CONTROLIU/ML 405NM)

s .JUNDILUTEC340 2.865

z 1:5 68 1.135

1:25 13 0.340 1:125 2.70. 145

R=O.998

z/

a

0.5-ct'

0

C,)

C,)

3 Oj

U/ML

FIG. 1. Example ofa typical standard curve of standardizedquantitativeELISAfortoxoplasma

an-tibodies. Experimental data indicate that this

straight-linerelationshipcoversarangeatleastfrom 2 to1,000IU/ml.

J. CLIN. MICROBIOL.

on February 7, 2020 by guest

http://jcm.asm.org/

QUANTITATIVE ELISA FOR TOXOPLASMOSIS 677

the standardcurves.When the concentration of

substrate wasincreased, the slope dropped

ap-preciably, although the absorbance increased slightly. Both absorbance and slope decreased

06_

~051 _ 05

2.5 5 10 20

pGANTIGN/DS

0.6_

Z0.5_

04

2 4

° CONJUGATECONC,

0

0.67

405-0.4

200

'100

\

_{9 ~~~Lo}

Z

5 10 25 50 30 45 60 90 pL SAMPLE PANUTE SAMPLE

INCUBATION

200

g~~~~~~~~~~~~~

-iooS

~

0-25 05 'I 2 30 45 6;0 9`0 ° ML CONJUGATE TE CO*T

300

Is

-200U. 0

*100 §

z r

0

0 5 2 4 0.5 2 4 10 20 40 80

MG SUBSTRATE/ML ML SUBSTRAE' MNUTE SUBSTRATE

INCUBATION

FIG. 2. Effect ofsomeparameterson the

absorb-ancevaluesofthepositivecontrolandonthe

regres-sion coefficient (slope) ofthe standard curve. The

latter was established with the readings from the

positivecontrol anditsthreeserialfivefold dilutions. These data represent averages from two or three

experiments eachruninduplicate. All of these

stan-dard curvesclosely fit with correlation coefficients

rangingfrom0.995to0.999,exceptfor thetwomarked

with a check. The standardprocedure used inthe

presentstudy employed thesecondcondition listed in allcases.Theconjugate concentrationwaspresented

asfoldofthatused inthestandardprocedure(#).

substantially as the volume of substrate (at 1

mg/ml)wasincreased.Increased volumes oftest

sample increased the absorbance valuesslightly

but decreased the linearity and slope of the standard curve. Varying the incubation times

with thetestsample,orwith theconjugate,had

relatively little effect onthesevalues, whereas

increasing the substrate incubation time in-creased the absorbance, but had only a minor

effect on the slope. Increasing the conjugate

volume orits concentrationtended toincrease

theabsorbance and theslope,butalso increased thebackgroundabsorbance

unacceptably.

Back-ground absorbance decreasedto a steady level asthe concentration of bovineserumalbumin in

the specimen diluentwas increased to 4%and

above. Incubation at 370C gave increased

ab-sorbance and provided better standard curves

thanatroomtemperature.In general, these data

indicate that many of the conditions for the assay are not critical and could be tailored to

meetparticular needs withinmostof the

param-etersstudied.

Assays of specimens derived from a single

bleeding from the same donors indicated that

either serum or plasma could be used in the

ELISA assay, with equivalent results. Fifteen

volunteers donated blood, and the anticoagu-lants usedwereheparin, citrate, dextrose citrate,

andethylenediaminetetraacetic acid. Five of the donors showed toxoplasma antibodies 200IU/ mlorhigher, and three showed values of 10IU/

mlorless.After correction foranydilutionfactor

duetotheanticoagulantused, itwasfoundthat thetoxoplasmaantibody levels in theserumand

variousplasmas ofanyindividual determinedby

ELISA were equivalent, i.e., they were within

theassay error(averagecoefficient ofvariation,

15%).

The reproducibility of the ELISAprocedure

wasquitesatisfactory,asshown inTable1.The "within-run"reproducibility of theassayshowed

TABLE 1. Within-run reproducibility of ELISA for Toxoplasma antibodies

ELISAobservedvalues

Testspecimens (no. of repli- Absorbance IU/mi

cates)

Arithmetic Geometric mean Arithmetic mean Geometric mean

mean (% C.

V.)X

x/. SD" (%C. V.) x/-* SD

A(9) 0.42(9.2) 0.42 x/ 1.10 14(15.5) 14x/+ 1.19

B(9) 2.80(6.0) 2.80

x/÷

1.06 420(10.7) 418 X/+ 1.11

C(9) 1.58 (5.2) 1.58X/÷ 1.05 152(9.2) 151

x/+

1.10

D(8) 0.14 (3.3) 0.14X/. 1.03 2.1(6.0) 2.1x/. 1.06

E(8) 0.84 (5.0) 0.84x/. 1.05 49(8.9) 49x/. 1.09

F(8) 2.60 (7.0) 2.59x/. 1.07 367(12.9) 356x/+ 1.15

G(8) 0.19 (3.0) 0.19 x/ 1.03 3.5(5.3) 3.5 X/ * 1.05

Avg % C.V.or SD 5.5 1.06 9.8 1.1

aC.V.,Coefficientofvariation.

bSD, Standard deviation.

VOL. 11,1980

I

I

I

1 4

on February 7, 2020 by guest

http://jcm.asm.org/

678 LIN, HALBERT, AND O'CONNOR

an averagecoefficientof variation of 5.5% and a geometric standard deviation of 1.06 in the ab-sorbance values. These values were 9.8% and 1.11, respectively, after conversion to

interna-tional units per milliliter. Tests of day-to-day interlaboratoryreproducibility showed approxi-mately equivalent results.

The specificity of the ELISA values for

toxo-plasma antibodieswasconfirmedby absorption

experiments. In theexamples shown in Fig. 3,

six human serum specimens (5 positive and 1

negative)werepreincubatedwith T. gondii

an-tigen, Entamoeba histolytica anan-tigen,

Trichi-nella spiralisantigen, or no antigen. The latter twoparasitic antigenswerecrude extracts incor-poratedintodiagnostickitsusing immunodiffu-sion (Cordis Laboratories) and were known to possessrelativelyhigh specific antigenic activity,

i.e., sufficient for gel difflusion techniques. The T. spiralis was derived from infected mouse tissue. The preincubated specimens were then assayed as above to determine whether

neutral-ization of the antibody in the sample had

oc-curred bytheparticular antigen in question. It is clear from the results that the toxoplasma antigen was the only preparation capable of

significantly reducing the ELISA values ofthe

positive specimens, indicating the immunologi-cal specificity of the present ELISA for

toxo-plasma antibodies. Reciprocally,when the disks were coated with E. histolytica antigens and

usedin similar neutralizationexperimentswith

humanseracontaining anti-amoeba antibodies,

600.

500.

400.

300.

2200. (I)

100 w

T.G.=T.GONDII T.S. =TSPIRAUS

E.=E.HISTOLYTICA N.=NONE

N3

R

_El

onlyE.histolytica antigen (not toxoplasma an-tigen) was shown to reduce significantly the anti-amoeba antibody activity (unpublished data).

Comparisonsbetween thepresent ELISA and three other immunoassays (MBD, IIF, and PHA) are summarized in Fig. 4-7. High and

significant correlationcoefficients wereobserved

in all instances. These data also indicated that specimens showing up to 30 IU/ml by ELISA weremostly negative by the other three assays. Inaddition,almost allspecimens positive bythe

otherassaysshowed greater than 10IU/ml,and mostof themwereabove 30IU/ml.Four of the 206specimens shown in Fig. 4 were nonreactive or reactive undiluted in the MBD test, but

showed high ELISA values (>50 IU/ml). Ab-sorption experiments as described above re-vealed that onlythe Toxoplasma antigen (but riot nornalmouseplasma,oramoebaor Trichi-nella antigens) wasable tosignificantly reduce thehigh ELISA valuesin thesespecimens.

TheCenterfor DiseaseControl(Atlanta,

Ga.)

distributesreferencehuman serumcontrols for the assay oftoxoplasma antibodies. One lot of

t,hese referenceserawasassayed by

ELISA,

with

goodagreement in the resultsasshown in Table 2.These valuesareinaccordwith the distribu-tionof values shown inFig. 7.

DISCUSSION

TheELISA procedure describedhere has

sev-eral

advantages

overother

commonly

used

as-MA0

04

E;LW

u

wwE

QlX

E;L~

sa fS 15B m

BLOCKING 0O2 dW OJL2CIIiZ (.(/LiZ0

o(/uWzf

dOXi)W

ANTIGEN H-- F-H- HI- HI- .aFs:

-HUMAN ELISA E LISA U-36-1I LI-36-12 U-36-17 LI-36-18

SERUM POSITIVE NEGATIVE

SAMPLE CONTROL CONTROL

FIG. 3. Demonstrationoftheimmunological specificity of ELISAreactions for toxoplasma antibodies by

absorption studies. Six humanserumspecimens (5 positiveand1negative)werepreincubated at37°C for30 minwith solubleantigens derived from: (i) T. gondii, (ii)E.histolytica, (iii) T. spiralis,or(iv)none.After this preincubation,eachspecimenwasassayed bytheELISAprocedure. It can be seen that the solubletoxoplasma antigensdramatically blockedthe ELISA reactions in eachof the positiveserumspecimens, whereas the

otherantigens hadnosignificant effect.

J. CLIN. MICROBIOL.

ni

on February 7, 2020 by guest

http://jcm.asm.org/

VOL. 11,1980 QUANTITATIVE ELISA FOR

679

t

S

I

o

9I

30-- 10- a 0

5-LLI ~~~~~~~~~~RsO0.80(N 206;P<0.001

R,=0.87(N=206. P<0.001 R2z-0.64 (N= 87iP<0.00)

NEGATIVE 4 16 64 256 1024 4096

(I4) M B D TITER

FIG. 4. Scatter diagram and regression lines ofELISA compared to theMBD results on 206serum specimens (bothassaysweredone in thelaboratory ofG.R.O.).R, the rank correlationcoefficient ofSpearman.

RlandR2, the correlationcoefficient ofPearson and Leeforall260specimensand those withMBD titerof

16 orgreater,respectively. TheregressionlinesareforRl andR2.

Rs=090(N=20,P<OOoi

R,=0,92(N=20,P<OO01

NEGATIVE 4

M 8 D TITER

16 64 256 1024

FIG. 5. Comparison between the ELISA and the

MBD tests on 20 human serum samples for

Toxo-plasma antibodies. MBD, run at the University of

California, San Francisco; the ELISAs were run

independently and codedattheCordisLaboratories.

RYJthe rankcorrelationcoefficientof Spearman. Rl,

thecorrelationcoefficients ofPearsonand Lee for all

20specimens. The regression line is for Rl.

saysfordetecting andmeasuring antibodiesto T.gondii.TheMBD,IIF,PHA,and most of the micro-ELISAtestsrequireserial dilutions of the

test samples for establishing the level of

anti-bodies,whereasourELISAyieldssuch

quanti-tative values on the basis of a determination

with asingle dilution oftestsample.Moreover,

t*ie

presentassaywas standardized against the

WHO reference preparation, as the antibody

concentrations were reported in international

units permilliliter. Suchasinglebasis of

refer-enceshouldminimize technical variations in

dif-ferent laboratories and permit more uniform

measurements of

toxoplasma

antibodies in any

laboratory using this procedure. This problem with the other assays is underscored by the description submitted with the WHO interna-tional standard for

anti-toxoplasma

serum.This

single

serum

preparation

wasassayedina

num-berofqualified

laboratories,

and the titersfound rangedfrom 400 to10,000in theMBDtest,from

25 to 500in the complement fixation test, and between 250and6,000 intheIIF test.

The ELISA described here was satisfactory

for serum or plasma, and sodium azide as a preservative did not interfere with the results. No pretreatment of the test specimens is re-quired, the assay is relatively rapid, and the reagents, which arereadyforuse, revealed

con-siderable stability without any biohazard. Pre-liminary data indicated that all reagents could be kept refrigerated at 2 to8°C forat least 2.0 years without noticeable effect on the test re-sults. Theantigenonthesolidphasecould even bekeptatroomtemperature(23°C) for atleast 10months without loss ofactivity.

000-, 30--n

-i ,o.

0

0

on February 7, 2020 by guest

http://jcm.asm.org/

680 LIN, HALBERT, AND O'CONNOR J. CLIN. MICROBIOL.

0

Rs 0.90(N-95; P<0.00I) Ri=O2 (N=95,P<O.OOI) R2 =0.60(N=23;P<0.OI)

_

I_v

&- P H A TITER

FIG. 6. Comparison between the ELISA and the PHA tests on human serum samples for Toxoplasma antibodies. PHAwasrunby the Bio-sciences Laboratories, and the ELISA were run independently and coded

atthe Cordis Laboratories. R., the rank correlation coefficient ofSpearnan. Rl and R2, the correlation coefficientsof Pearson and Lee for all 95 samples and those with titers of 16 or greater, respectively, assuming the PHA titerof1for those less than16.

o 0

8

0 0

8

0 0 0

0

0

Rs=.79(N=22;P<0.001) R,=084(N=22;P<0.001)

(<16) 16 64 256 1024 4096 16384

NEGATIVE

F TITER

FIG. 7. Comparison between theELISA and theIIFtestson22humanserumsamples for Toxoplasma

antibodies.IIFwas runbytheParasiticSerology Laboratories,CenterforDiseaseControl, Atlanta;and the

ELISAwererunindependentlyandcodedattheCordisLaboratories. R.,the rank correlationcoefficient of

Spearman.Rl, the correlationcoefficient ofPearson andLee, assumingtheIIFtitersof4forthenegative

results. Theregressionline isforRl.

Arathergood correlationwasfound between MBD and IIF. In a separate, ongoing clinical

the ELISA values(international unitspermili- research study with 34 specimens with MBD

liter) and titers of toxoplasma antibodies found titers between 1,024 and 8,192, there was no

in several panelsoftest specimens assayed by indicationofaplateauof theELISAtestresults

I1

-i

wL

0 0 0

8

0 0

0 0 0

8

(GI6}

NEGATIVE 16 32 64

1000_

looi

30-

10-Uf)

-J

w

0

0

on February 7, 2020 by guest

http://jcm.asm.org/

QUANTITATIVE ELISA FOR TOXOPLASMOSIS

TABLE 2. ELISAassay results with Centerfor DiseaseControl toxoplasmosis control human seraa

Center for ELISA observed(IU/ml)b Serum DiseaseCon- Arithmetic

trol-stated mean Geometrc ter, IIF m

iS

meanx/+SD

Negative <32 2.1±0.3 2.1 x/+ 1.17

Lowtiter 32-128 62±11 62x/+ 1.19

High titer 512-2,048 203±30 202x/÷ 1.17 aCatalog no.

KK0501,

lot 76-0339K.

b Measured in six experimentsatdifferent

times.

'SD, standard deviation.

with these high-titered samples (unpublished data). Failure to achieve even higher correla-tionsmaybe duetotwofactors. The MBD and IIF procedures allow a plus-or-minus fourfold errorintheendpoint titers (12). Thus,asample witha"truetiter"of 1:64 may reveal titers of 1: 16 or 1:256 inreplicateassays.The reproducibil-ity of theELISAisconsiderably better, showing

ageometric standard deviation of1.11in inter-national units per milliliter, which is smaller thanmostofthosereported for immunoglobulin quantitation (13, 14). The latter, in turn, has beenfoundtoprovide appreciably better

repro-ducibility than those serologicaltests using serial

titrations, which allow accuracy to within one

twofold dilution (14). In addition, the ELISA

determination yieldsresults which are presum-ablyasummationofalltheantibodies directed

against the severaltoxoplasma antigens (at least

seven) knownto induce antibody responses in

humans (4, 7, 9, 11; unpublished data, G.R.O.), since the preparations used to coat the disks weretotalextracts of the organisms prepared by verymildsonication procedures. In contrast, the

MBD and theIIF tests undoubtedly measure a morelimitedsetof antibodyresponses,

presum-ably those situatedon thesurface membrane of

thetoxoplasma organisms (9).

ACKNOWLEDGMENT

Thisworkwassupported in part by grant EY-01204 from the National Institutes ofHealth, Bethesda, Md.

LITERATURE CITED

1. Camargo,M.E., A. W.Ferreira,J. R.Mineo, C. K.

Takiguti, and0.S.Nakahara. 1978.Immunoglobulin G andimmunoglobulinMenzyme-linked immunosor-bentassaysand definedtoxoplasmosis serological pat-terns.Infect. Immun.21:55-58.

2. Capron, A.,D.Bout, andJ.C.Dugimont.1975.Apport

des methodesimmunoenzymologiques utilisantdes

an-tigenes purifiesaudiagnosticspecifiqueetautomatise

des infectionsparasitaires.Ann. Soc.BelgeMed.Trop. 55:443-454.

3. Chessum, B.S., andJ.R.Denmark.1978.Inconstant ELISA. Lancet i:161.

4. Chordi, A., K.W.Walls,and I. G.Kagan.1964. Anal-ysis ofToxoplasmagondii antigensby agar diffusion methods. J.Immunol.93:1034-1044.

5. Engvall, E., and P. Perlmann. 1971. Enzyme-linked

immunosorbent assay(ELISA).Quantitative assayfor

immunoglobulinG.Immunochemistry 8:871-874.

6. Halbert, S.P.,and M.Anken.1977.Detection of hepa-titisB surface antigen (HB,Ag)withuse ofalkaline phosphatase-labelledantibodytoHB.Ag.J. Infect. Dis. 136:S318-S323.

7. O'Connor, G.R. 1957. Precipitating antibodyto

toxo-plasma.Am.J.Ophthalmol.44:75-83.

8. Palmer,D.F.,J. J.Cavallaro,K.Walls,A.Sulzer,

and M. Wilson.1976.Aprocedural guidetothe per-formance of theserologyoftoxoplasmosis.Center for DiseaseControl, Atlanta,Ga.

9.Remington,J.S.,and J. LKrahenbuhl.1976.

Immu-nology oftoxoplasma infection, p. 236-267. In S. Cohen and E. Sadun(ed.),Immunologyof Parasitic Infections.

BlackwellScientificPublications, Philadelphia,Pa.

10. Snedecor,G.W.,and W.G.Cochran.1967.Statistical

methods,6th ed. IowaStateUniversity Press,Ames.

11. Strannegard,0.1962.Studies oftoxoplasma

precipitin-ogen and theircorrespondingantibodiesbymeansof

diffusion-in-gel methods. Br. J. Exp. Pathol.

43:600-613.

12. Sulzer, A. J., and E. C. Hall.1967.Indirect fluorescent antibody test forparasitic diseases. IV. Statisticalstudy ofvariation in the indirectfluorescentantibody (IFA)

testfortoxoplasmosis.Am.J.Epidemiol.86:401-407.

13. Taylor,R.N.,K. M.Fulford,and A. Y.Huong.1978.

Comparison of kinetic andend-pointdiffusionmethods for quantitating human serum immunoglobulins. J. Clin. Microbiol.8:23-27.

14. Taylor, R. N., K.M.Fuilford,and V. A.Przybyszeski.

1977.Proficiencytestingsummaryanalysis: diagnostic immunology1977I: rubella,carcinoembryonic antigen,

immunoglobulin quantitation, C-reactive protein,

sal-monella (typhoid0) agglutination.Center forDisease

Control, Atlanta, Ga.

15. Taylor, R.N.,A.Y.Huong,and K. M. Fulford.1975. Measurementof variation inserologictests.Center for DiseaseControl, Atlanta, Ga.

16. Voller, A., D. E.Bidwell,A.Bartlett,D.G.Fleck,M.

Perkins, and B. Oladehin. 1976. A microplate

en-zyme-immunoassay fortoxoplasma antibody. J.Clin. Pathol. 29:150-153.

17.Walls,K.W.1978.Serodiagnosis oftoxoplasmosis.Lab. Manage.16:26-31.

18. Walls,K.W.,S. I. Bullock,and D. K.English. 1977.

Use of the enzyme-linked immunosorbent assay (ELISA) and itsmicroadaptation for the serodiagnosis

oftoxoplasmosis.J.Clin. Microbiol.5:273-277.

19. World Health Organization. 1977. Biological sub-stances. International standards, reference prepara-tions,andreferencereagents, p.20-21.World Health Organization,Geneva.

681

VOL. 11,1980

on February 7, 2020 by guest

http://jcm.asm.org/