Comparison of enzyme-immunoassay and radioimmunoassay for detection of human rotaviruses and adenoviruses from stool specimens

Journalof

VirologicalMethods,

@ Elsevier/North-Holland

l(1980)

Biomedical

331

331.-341

Press

COMPARISON OF ENZYME-IMMUNOASSAY DETECTION SPECIMENS

HANNU

OF HUMAN ROTAVIRUSES

K. SARKKINEN,

Department (Accepted

HANNA

immunoglobulin adenovirus

(ISA)

as indicator

antigens

from

used previously

gastroenteritis.

E. HALONEN

Turku 52, Finland

and

in our laboratory

but in each of these non-specific

EIA was found

reactions provided

rabbit

negative

of the developed

with specimens sensitive

EIA tests showed cases to prove

giving low positive

and reliable

that appropriate

rotavirus

swine of human

anti-rabbit

rotavirus

and

EIA and radioimmunoassay from children

and adenovirus them

with acute

EIAs but not

to be false negatives.

the specificity reactions

as RIA in the routine

confumatory

pig anti-rota-

or anti-adenovirus

peroxidase-conjugated for the detection

by both

in several

guinea

anti-rotavirus

was made with 250 stool specimens found

cases confirmatory

to be as specific,

gastroenteritis

horseradish

A comparison

EIA tests were also necessary

to eliminate

as the solid phase,

antibody,

has been developed

stool specimens. were

beads

as primary

antibody, antibody,

Two specimens

confirmatory

using polystyrene

immunoglobulin

as secondary

immunoglobulin

adenovirus

FOR

FROM STOOL

12 July 1980)

An enzyme-immunoassay

by RIAs,

and PEKKA

of Virology, University of Turku, SF-20520

virus or anti-adenovirus

(RIA)

TUOKKO

AND RADIOIMMUNOASSAY

AND ADENOVIRUSES

of the binding

The or

in EIA. The developed

diagnosis

of rotavirus

and

tests were included.

INTRODUCTION

Rotaviruses

are the main viral pathogens

in hospitalized

patients

with infantile

gas-

troenteritis (Middleton et al., 1977) and also adenoviruses seem to play a significant role as enteric pathogens in children (Retter et al., 1979). Several methods have been developed for the direct detection of these agents from stool specimens. Electron microscopy of negatively stained stool specimens has been the standard method (Middleton et al., 1977) but other methods such as counterimmunoelectrophoresis (Tufvesson and Johnsson, 1976; Spence et al., 1977; Mankikar et al., 1979) immunofluorescence (Peterson et al., 1976; Retter et al., 1979) radioimmunoassay (RlA) (Kalica et al., 1977; Halonen et al., 1980) and enzyme-immunoassay (EIA) (Yolken et al., 1977, 1978; Wade11 et al., 1979) have also been used. A radioimmunoassay, using polystyrene beads as the solid-phase, guinea pig anti-virus immunoglobulin as primary antibody, rabbit anti-virus immunoglobulin as secondary antibody and iodinated anti-rabbit immunoglobulin as indicator antibody, has previously been developed in our laboratory for the detection of human rotavirus and adenovirus antigens in stool specimens (Sarkkinen et al., 1979a; Halonen et al., 1980). In the present

332

communication antibody

a similar

to rabbit

comparison

enzyme-immunoassay,

immunoglobulins

of the developed

instead

using horseradish peroxidase labelled of 1251-labelled antibody, is reported. A

enzyme-immunoassay

and radioimmunoassay

was made

with 250 stool specimens, positive or negative for rotavirus and adenovirus. MATERIALS

AND METHODS

Specimens The test material

consisted

of 93 negative specimens in rotavirus and adenovirus

RIA,

118 positive specimens in rotavirus RIA and 39 positive specimens in adenovirus RIA. The specimens were from children with acute gastroenteritis and they had been stored at -20°C. Virus purification Nebraska calf diarrhea virus (NCDV), used as a control antigen in rotavirus enzymeimmunoassay (rotavirus EIA), was grown in LLC-MK2 cell cultures in the presence of trypsin and partially purified as reported earlier (Sarkkinen et al., 1979b). This semipurified virus was further purified by two subsequent centrifugations in CsCl-gradients according to Petric et al. (1975). The protein content was determined by the method of Lowry et al. (195 1). The purification of group-reacting adenovirus type 2 hexon antigen, used as the control antigen in adenovirus enzyme-immunoassay (adenovirus EIA), has been reported elsewhere (Halonen et al., 1980). Immunization of animals The purification of human rotavirus and group-reacting adenovirus type 2 hexon antigen for immunization purposes and procedures has been reported previously in detail (Sarkkinen et al., 1979a; Halonen et al., 1980). EIA reagents Guinea pig and rabbit anti-rotavirus IgG and guinea pig and rabbit anti-adenovirus immunoglobulin (Ig) fractions were prepared by precipitation of sera with an 18% (w/v) final concentration of sodium sulphate followed by chromatography on a Sephadex G-200 column for anti-rotavirus IgG and Sephadex G-25 for anti-adenovirus Ig. Horseradish peroxidase-conjugated swine antibodies against rabbit IgG were purchased from a commercial source (Orion, Espoo, Finland). The substrate solution, which consisted of 3 mg/ml of o-phenylenediamine (OPDA, Koch-Light Laboratories, Colnbrook Bucks, England) in 0.1 M citrate-NA2HP04 buffer, pH 5.5, and 10 ~1 of 30% H202 per 15 ml in the same buffer, was made freshly before use. The reaction was stopped with

333

1N HCl (Orion, Espoo, Finland). The diluent in all steps in the rota-EIA except for the substrate solution was PBS, pH 7.35, containing 20% inactivated fetal calf serum (Gibco Europe, Glasgow, Scotland), 2% Tween 20 and low4 M merthiolate. For adenovirus EIA the same buffer was used with the exception that inactivated normal sheep serum was substituted for fetal calf serum. EIA procedure Polystyrene beads (6.4 mm in diameter, Precision Plastic Ball Co., Chicago, IL) were coated with anti-rotavirus guinea pig IgG or anti-adenovirus guinea pig Ig by incubating the untreated beads overnight at room temperature in an antibody solution containing 5 pg of anti-rotavirus IgG or anti-adenovirus Ig per ml (1 pg/bead) in carbonate buffer, pH 9.6 (Voller, 1976). Beads were stored in this antibody solution at 4°C until used, usually for l-2 weeks. Stool specimens in 200 ~1 abquots in a single l/20 (v/v) dilution or serially diluted were pipetted into disposable polystyrene tubes and a polystyrene bead with adsorbed anti-rotavirus IgG or anti-adenovirus Ig was then added to each tube. After incubation at 37°C for 1 h the stool specimens were aspirated and the beads were washed twice with 5 ml of tap water. A 200 ~1 volume of rabbit anti-rotavirus IgG (3.2 pg/ml) or rabbit anti-adenovirus Ig (8 pg/ml) was then added to each tube and the beads were incubated at 37°C for 1 h followed by washing as described above. A 200 ~1 volume of horseradish peroxidase-conjugated swine anti-rabbit immunoglobulins (1 : 1000 dilution) was then added to each tube. After 1 h incubation at 37’C the beads were washed as described above, changed to new tubes, and a 500 ~1 volume of the substrate solution was added to each tube and the beads were incubated for another hour at room temperature in the dark. After incubation a 500 1.11volume of 1 N HCl was added to each tube to stop the reaction and 500 ~1 aliquots of the reaction mixtures were then transferred to clean polystyrene cuvettes and the absorbance measured at 492 nm using a FP-9 .Ana.lyzer (Labsystems, titrations absorbance

Division

of rotavirus and adenovirus

of Finnpipette control

Co., Finland).

antigens were included

Buffer blanks

and

in each assay. The

values of buffer blanks were usually about 0.2 and the cut-off value used in

both rotavirus and adenovirus or more were considered ing.

EIA was 0.5. Specimens

with an absorbance

positive provided the confirmatory

value of 0.5

tests showed specific bind-

EIA confirmatory test To test the specificity of the antigen binding a blocking test was performed on all RIA-positive specimens (both rotavirus and adenovirus) with absorbance values below 1.5 in EIA and on all RIA-negative specimens (both rotavirus and adenovirus) with absorbance values above 0.5 in EIA. The test was done in the way described above except that anti-rotavirus or anti-adenovirus guinea pig serum was added before the secondary antibody. The test was done as follows: 200 ~1 volumes of the specimens were pipetted

334

into three tubes, a bead with adsorbed

anti-rotavirus

or anti-adenovirus

antibody

was

added to each of the three tubes and the beads were incubated for 1 h at 37°C. After the washing procedure a 200 ~1 volume of anti-rotavirus or anti-adenovirus guinea pig serum (1 : 500 dilution)

was added to the first tube and the same volume of normal

guinea pig serum (pre-infection

serum, 1 : 500 dilution)

to the second; to the third only

a dilution

buffer was added. After the incubation at 37°C for 1 h, a 200 ~1 volume of Ig (16 pg/ml) was added to each rabbit anti-rotavirus IgG (6.4 pg/ml) or anti-adenovirus of the three tubes (without the removal of the previous 200 ~1 sample) giving a total volume of 400 ~1. The beads were then incubated for 1 h at 37°C. The rest of the test was performed as the assay proper. The test was considered positive if a 50% or greater decrease in absorbance values was noticed with the specimen incubated with guinea pig anti-rotavirus or anti-adenovirus hyperimmune serum as compared normal guinea pig serum in the same dilution or with dilution buffer.

to incubation

with

Radioimmunoassay (RIA) procedure Radioimmunoassays for human rotavirus and adenovirus earlier (Sarkkinen et al., 1979a; Halonen et al., 1980). The c.p.m. in both rotavirus and adenovirus RIA and specimens more were considered positive, with the proviso that the

were performed as reported cut-off value used was 500 with c.p.m. values of 500 or confirmatory test indicated

a specific binding. Electron microscopy The electron microscopy of negatively stained specimens was done on all specimens, with contradictory results in RIAs and EIAs, on 63 of the rotavirus RIA-positive specimens and on all the adenovirus

RIA-positive

specimens (Sarkkinen

et al., 1979a; Halonen

et al., 1980). Statistical methods Linear

regression

and rz (correlation

were calculated with Hewlett-Packard

coefficient)

values presented

in Figs. 2 and 3

98158 Software General Statistics, Vol. 1.

RESULTS

The sensitivities of rotavirus and adenovirus EIAs were determined by diluting the purified Nebraska calf diarrhea virus (NCDV) and adenovirus type 2 hexon antigen according to protein content and by measuring the absorbance values of each dilution. Twice the negative control was taken as the cut-off line. The sensitivities of the assays for NCDV and adenovirus type 2 hexon antigen varied from test to test between 1 and 10 ng/ml of purified viral protein (Fig. la, b).

335

._

8

5-

(a)

cb)

.

O-

15-

./ - - _ - _-CUT-00 _ -- UNE ./ 1

‘”

NCDV ANTIGEN (%/ml)

Fig. 1. The sensitivity hexon and

antigen for buffer

detection blanks

lw

curve

.

1

BUFFER

1V

BLANK HEXON ANTIGEN (w/ml)

of Nebraska

calf diarrhea

by enzyme-immunoassay. are means

of three

beads.

virus

(NCDV)

The absorbance The cut-off

(a) and

adenovirus

type

2 (b)

values for each value in the curve

lines are twice

the means

of the buffer

blanks.

The results of titrations in EIA and RIA of three adenovirus-positive and three rotavirus-positive stool specimens and three adenovirus-negative and rotavirus-negative stool specimens are presented in Tables 1 and 2. All positive stools were positive up to a dilution of l/2000 or higher in each test and the absorbance values of the negative stools were correspondingly low. Rotavirus-positive specimens and adenovirus-positive specimens were also cross-tested in adenovirus and rotavirus RIA and EIA, and they were all negative proving the specificity

of the immunoreagents

for rotavirus and adenovirus.

93 rotavirus-negative and 118 rotavirus-positive specimens previously screened in RIA were tested in EIA. Fig. 2 indicates the distribution of absorbance values in rotavirus EIA before the confirmatory test as compared to c.p.m. values in rotavirus RIA. Out of 93 rotavirus RIA-negative specimens, 9 1 were negative in EIA (absorbance values below 0.5). The two remaining specimens were considered positive in EIA (absorbance values above 0.5) with absorbance values of 0.786 and 1.243. The mean absorbance value for the rotavirus RIA-negative specimens in rotavirus EIA was 0.228 + 0.102. Out of 118 rotavirus RIA-positive specimens, 116 were positive in EIA. Two specimens had absorbance values of 0.361 and 0.471 and were considered negative. 93 adenovirus-negative specimens (the same negative ones as above) and 39 adenovirus-positive specimens were tested in adenovirus EIA (Fig. 3). Out of 93 adenovirus RIA-negative specimens 90 were negative in EIA (absorbance values below 0.5). Three were positive in adenovirus EIA with absorbance values of 0.541, 1.258 and 1.029. Two

336

TABLE

1

Representative with

three

results positive

of rotavirus

(Nos.

l-3)

antigen

detection

and three

negative

by radioimmunoassay (Nos.

4-6)

and enzyme-immunoassay

specimens

from

children

with acute

gastroenteritis Specimen

no.

Dilution

of specimen

l/20 1

l/200

288ga

2202

2.233’

1.817

2

896

296

202

0.519

0.259

0.217

2670

2065

931

330

245

2.182

1.696

0.558

0.255

0.237

232

NTC

NT

NT

NT

NT

NT

NT

NT

NT

5

0.228

222

225

0.253

0.220

6

115

177

0.109

c

not tested

TABLE

0.232

per minute at 492 nm

2

Representative assay

264 0.259

2021

154

absorbance

411 0.292

1.608

0.173

counts

996 0.578

1/200,000

2542

4

b

1/20,000

2.145 3

a

l/2000

with

results three

of adenovirus

positive

(Nos.

l-3)

antigen

detection

and three

by radioimmunoassay

negative

(Nos. 4-6)

and enzyme-immuno-

stool

specimens

from

children

with acute gastroenteritis Specimen

Dilution

no.

l/20 1 2 3

of specimen l/200

l/2000

1/20,000

2382a

2267

1612

504

2.657b

2.696

1.824

0.525

2289

1866

739

254

2.529

2.124

0.695

0.300

1/200,000 196 0.283 185 0.263

2252

1925

580

225

2.465

2.124

0.580

0.338

0.271

161

4

153

127

NTC

NT

NT

5

0.137 196

0.270 162

NT

NT

NT

6

0.299 161

0.270 153

NT

NT

NT

0.101

0.280

a

counts

b

absorbance

per minute

c

not tested

at 492 nm

337

0

0.5

1.5

1.0

2.0

2.5 ) 2.:

ABSORBANCE AT 492nm

Fig. 2. Distribution

of absorbance

values in radioimmunoassay from

children

line represents

0

with

values

acute

gastroenteritis.

the linear regression

05

1.0

in enzyme-immunoassay

(RIA) for 93 rotavirus-negative

15

liney

2.0

2.5

The dotted

(EIA) and counts

per minute

and 118 rotavirus-positive

lines indicate

= a + bx (a = -68.04,

the cut-off

b = 1.31),r2

(cpm)

stool specimens values and the solid

= 0.73.

2.5

ABSORBANCE AT 45’2 nm

Fig. 3. Distribution values mens

of absorbance

in radioimmunoassay from

children

solid line represents

with

values

(RIA) acute

in enzyme-immunoassay

for 93 adenovirus-negative

gastroenteritis.

the linear regression

The dotted

(EIA) and counts

per minute

and 39 adenovirus-positive lines indicate

line y = a + bx (a = -27.08,

the cut-off

(cpm)

stool specivalues

and the

b = 1.2), rz = 0.66.

of these three were the same specimens with positive reactions also in rotavirus EIA. The mean absorbance value for the adenovirus RIA-negative specimens in adenovirus EIA was 0.255 f 0.147. Out of 39 adenovirus RIA-positive specimens, 37 were positive in adenovirus EIA and two specimens negative in EIA had absorbance values of 0.458 and 0.450. To prove the specificity of the binding in EIA and to eliminate non-specific reactions,

338

confirmatory

tests were made on selected

specimens

indicated

in the Materials

and

Methods. Table 3 indicates the representative results of the rotavirus confirmatory test with two specimens with specific binding and two specimens with non-specific binding. In specific reactions the rotavirus guinea pig hyperimmune serum blocked more than 50% of the bound reactivity while the normal guinea pig serum (non-immune serum) had no blocking

effect as compared

to the binding in dilution

buffer control. With non-

specifically reacting specimens the blocking effect was also obtained by the normal guinea pig serum. The adenovirus confirmatory test was done in exactly the same way. All the positive reactions in both adenovirus and rotavirus EIA with the RIA-negative specimens proved to the non-specific and all the positive reactions in both adenovirus and rotavirus EIA for the RIA-positive reactions were found to be specific. Also the two rotavirus RIA-positive but rotavirus EIA-negative, and the two adenovirus RIApositive but adenovirus EIA-negative specimens were shown to be true positives by EIA confirmatory tests even though the absorbance values were low. The results of electron microscopy produced further evidence for the specificity of EIA results since the two rotavirus RIA-positive but EIA-negative and one of the two adenovirus RIA-positive but EIA-negative specimens in electron microscopy.

also contained

rotavirus

and adenovirus

particles,

respectively,

DISCUSSION

The results of the present study indicate that EIA is as sensitive and as specific as RIA in the detection of rotavirus and adenovirus in stool specimens. The detection limits of the assays were found to be between 1 and 10 ng/ml of purified viral protein which is approximately the same as reported earlier with RIA (Sarkkinen et al., 1979a; Halonen et al., 1980). Two false negative results were obtained both in adenovirus and rotavirus EIA as compared

to RIA. However,

all these four cases had absorbance

values only

slightly below the cut-off line and in each case confirmatory tests proved the specificity of th& binding. Confirmatory tests were also necessary in several cases in both rotavirus and adenovirus EIA to prove the specificity of the low positive reactions and to eliminate the non-specific

binding in some negative specimens.

The finding of several specimens with non-specific binding causing false positive reactions makes it necessary to prove the specificity of the binding with a confirmatory type of test with all specimens with absorbance values between 0.4 and 1.5 in our present EIA method. False positive reactions are, however, usually easy to detect, since by using two or more identical antigen detection systems (e.g. rotavirus and adenovirus) the nonspecific binding is almost always detectable in each test and the tests thus serve as a control for each other. Dual infections, on the other hand, which would produce true positive reactions in both tests, are rare. In our prospective study of infantile gastroenteritis we found rotaviruses and adenoviruses in stool specimens at the same time in only 0.7% of all cases studied (Vesikari et al., 1980). In spite of the fact that non-specific binding is often easily detectable by using two or

a

0.602

0.597 0.134 0.140 0.211

1.438 1.126 1.226 0.795

1.823

Absorbance when specimen incubated with hyperimmune guinea pig serum

Absorbance when specimen incubated with dilution buffer

Nebraska calf diarrhea virus

Control antigen (NCDVa) 100 ng/ml

Specimen no.

0.1792

1.606 0.982 0.115 0.196

Absorbance when specimen incubated with non-immune guinea pig serum

positive

positive positive negative negative

Interpretation of the test

Rotavirus confirmatory enzyme-immunoassay (EIA) for testing the specificity of the binding in rotavirus EIA with two positive and two negative stool specimens from children with acute gastroenteritis

TABLE 3

340

more identical should

antigen

be determined

detection

systems, we feel that positive

spectrophotometrically

rather

and negative

reactions

than by mere visual reading. By

using a semiautomated spectrophotometer, results in printed form can be obtained and evaluated almost immediately after the final incubations. Also the proper use of confirmatory

tests requires the exact measuring of the absorbance

values of the specimens.

The problem of non-spec~c binding and the necessity of conf~ato~ tests is naturally not limited to EEA but is present in RIA as well. In fact we have had the same experience of the importance of the confirmatory tests in RIA during the 1% year routine use of rotavirus and adenovirus RIA in our diagnostic unit. During this period we have tested more than 1000 stool specimens and about 10% have required confnmation and approximately 90% of these have been non-specialty reacting. The results of the present study suggest that the number of specimens which must be confirmed in EIA is approximately the same as in RIA. The exact nature of the factor causing non-specific binding in four-layer immunoassays is at present unknown but N-acetylcysteine (Chao et al., 1979) has been reported to reduce it indicating that this factor might belong to IgM-class antibodies (Yolken and Stopa, 1979). Our own prel~~a~ the specific binding in rotavirus- and unpublished results) and the use of with weakly positive stool specimens. The stool specimens included in for rotavirus and adenovirus. In spite

results suggest that ~-acetylcyste~e also decreases adenovirus-positive stools (Sarkkinen and Halonen, this chemical may thus cause false negative results this study were routinely screened earlier in RIA of the fact that several tec~ci~s had performed

the tests, only one specimen out of 251 studied was found to be reported falsely as negative due to a technical mistake (this specimen was excluded from the material). This prospective fmding further indicates that RIA and EIA are indeed very reliable and practical tools in the diagnosis of viral gastroenteritis provided that confmatory tests are included. The ~rn~o~says are also more suitable than electron microscopy for large scale routine work (Sarkkinen et al., 1979a; Yolken et al., 1977). As well as local laboratory facilities, the equipment available will determine whether RIA or EIA is more appropriate but the difficulties in handling radioactive waste products probably makes EIA a better alternative for most laboratories in large-scale routine work. ACKNOWLEDGEMENTS

The excellent

technical

assistance of Ms. Kaija Johansson

is gratefully

This study was supported by a grant from the Sigrid Juse’lius Foundation and Yrjii Eskola Foundation.

acknowledged. and the Daisy

REFERENCES Chao,

R.K., M. Fishaut,

Halonen, 614.

J.D. Schwartzman

P., H. Sarkkinen,

P. Arstila,

and K. McIntosh, E. Hjertsson

and

1979, J. Infect. E. Torfason,

Dis. 139,483.

1980,

J. Clin.

Microbial.

11,

341

Kalica,

A.R.,

R.H.

Purcell,

1977, J. Immunol. Lowry,

O.H., NJ.

Mankikar,

SD.,

Middleton,

Rosebrough,

Petric,

M., M.T. Szymanski

Retter,

M., P.J. Middleton, HX.,

Sarkkinen,

Voller, Wadell,

T., M. Maki,

R.H.,

Lancet

1975, Intervirology

and P.E. Halonen,

1979b,

Arstila

Stand.

5, 248. B 84, 225.

and P.E. Halonen,

Arch.

Dis. Child,

in press.

Med. Hyg. 70,98.

A.H. Kidd and C.R. Madely,

against

J. Med. Virol. 3,281.

1977, .I. Clin. Microbial.

P.P.

10,574.

J. Med. Virol. 4,255.

1976, Acta Path01 Microbial. Sarkkinen,

3, 376.

5,233.

1979, J. Clin. Microbial.

1979a,

193,265.

131, 733.

1976, J. Clin. Microbial.

Virus Diseases,

Munich,

H.W. Kim, T. Clem, R.G. Wyatt,

1979,

Sheraton

A.R. Kalica,

in: Abstracts Congress

Papers and Posters.

Center,

R.M.Chanock

September

and AZ.

Euro-

5- 7.

Kapikian,

1977,

2,263. R.H.,

B. Barbour,

R.G. Wyatt,

A.R. Kalica,

AZ.

Kapikian

201,259. Yolken,

and A.Z. Kapikian,

10,253.

1977, Am. J. Dis. Child

and PP. Arstila,

R. Sot. Trop.

G., M. Johansson,

Yolken, Yolken,

H.K.

1951, J. Biol. Chem.

J. Clin. Microbial.

J.S. Tom and M. Petric,

B. and T. Johnsson,

pean Association

and M. Petric,

R. Petro and S. Bloch,

A., 1976, Trans.

1979,

and P.J. Middleton,

H.K., O.H. Meurman

Vesikari,

H.W. Kim, R.M. Chanock

Farr and R.J. Randall,

and R.A. Smart,

P.E. Halonen

IL., M. Fauvel,

Tufvesson,

R.G. Wyatt,

and M. Petric,

M.W., R.S. Spendlove

Sarkkinen,

Serene,

AL.

P.J. Middleton

P J., M.T. Szymanski

Peterson,

Spence,

MM.

118,1275.

R.H.and

P.J. Stopa,

1979,

J. Clin. Microbial.

10,703.

and R.M. Chanock,

1978, Science