Journal of Virological Methods, 8 (1984) 137-145

Elsevier

JVM 03290

137

ENZYME IMMUNOASSAY FOR ANTIBODIES TO MEMBRANE

ASSOCIATED ANTIGEN OF VARICELLA ZOSTER VIRUS

J.C. COX, M.B. MOLONEY, R.W. HERRINGTON, A.W. HAMPSON and J.G.R. HURRELL

Commonwealth Serum Laboratories, Poplar Road, Parkville, Victoria, 3052, Australia

(Accepted 17 November 1983)

An in situ enzyme immunoassay to viral membrane antigen was developed to enable the specific

estimation of antibodies to varicella zoster (VZ) virus. The technique was compared with a modified

fluorescent antibody to membrane antigen (FAMA) procedure and with the complement fixation (CF) test

by parallel assay of 352 plasma samples. The enzyme immunoassay (EIA) procedure showed very good

correlation with the modified FAMA procedure, and both were far more specific than the CF test. This

specificity was achieved by the use, in the EIA, of VZ virus-infected cells grown and fixed in situ with

glutaraldehyde. Thus the only virus antigens accessible to antibody were the VZ-specific antigens expressed

at the cell membrane, cross-reactions with herpes simplex virus antibodies thereby being avoided.

varicella zoster virus antibodies enzyme immunoassay

INTRODUCTION

Varicella zoster (VZ) virus infection of immunosuppressed or otherwise susceptible

adults and new-born children can result in a severe, often fatal illness. The clinical

effects can be prevented or alleviated if a high-titre zoster immune globulin can be

administered within 96 h from exposure to virus (Centerfor Disease Control, 1979). A

shortage of zoster immune globulin has led to restrictions being placed upon its

clinical usage (Center for Disease Control, 1979; Schiff, 1979). In an attempt to

alleviate this shortage, a rapid screening test for VZ antibodies was sought so that

outdated plasma, supplied by the Australian Red Cross Blood Transfusion Services

could be tested for possible inclusion in such a VZ immune globulin pool. Because the

donor history of such plasma would be unknown, it was essential that the test

procedure specifically detect VZ neutralising antibodies only, i.e. that the test not

react with antibodies to herpes simplex virus (HSV) or other herpes viruses. Addition-

ally, it was important that the test should be able to be performed simply in large

numbers without the need for specialist operators, and that it should have the

potential for ready automation. In this paper we report a test which combines the

specificity and accuracy of the indirect fluorescent antibody to membrane antigen

016h-09iJ/X4/S03.00 <, 19X4 Elsevicr Science Publr\hcr\ B.V.

138

(FAMA) assay (Zaia and Oxman, 1977) with the convenience of enzyme immunoas-

say technology.

MATERIALS AND METHODS

Plasma

Samples of plasma for testing were removed from blood donations from VZ

convalescent patients and from randomly selected outdated plasma donations, both

supplied routinely by the Australian Red Cross Blood Transfusion Service.

Viruses

The VZ virus was strain RCH0163, a cell associated virus strain isolated at The

Royal Children’s Hospital, Melbourne, and kindly supplied by Mr. Ian Jack. The

virus had been passaged 33 times in human diploid cell lines prior to use in the assay.

Virus was stored at -196°C as infected cell cultures following trypsinisation of cells

showing a 50 to 70% CPE.

The herpes simplex type 1 virus was kindly supplied by Mrs. M. Kennett, Fairfield

Hospital, Melbourne, Australia. The virus had been maintained in diploid human cell

lines and stored at -6O’C as cell free virus.

Preparation of in situ antigen

Both viruses were grown in a diploid foetal human tongue cell line, CSL 300,

derived at the Commonwealth Serum Laboratories (CSL), Melbourne, Australia and

used at approximately 30 populations doublings. Growth medium was Eagle’s mini-

mum essential medium with non-essential amino acids and supplemented with 10%

unheated foetal calf serum (CSL, Melbourne, Australia).

VZ virus infected cells from storage were mixed in growth medium with uninfected

cells in the ratio of 1 : 40 and grown in a gassed (5% CO, in air), humidified incubator

at 37°C. The seeding density for both slides and trays was 4 X IO4 cells/ml. Square

Petri dishes (100 mm) containing 3 prewashed glass microscope slides were seeded

with 25 ml of cell suspension and incubated for 24 h. Polystyrene 96-well microtitre

trays for cell culture (Disposable Products P/L, Adelaide, Australia) were seeded with

0.1 ml cell suspension per well and incubated for 48 h.

Slides and trays of herpes simplex type 1 virus were prepared in a similar manner to

VZ virus with the exception that a cell free virus harvest was added to an uninfected

cell suspension at a dilution of approximately 1 in 4,000. This level of infection had

been shown to produce satisfactory focal coverage of both slides and trays after 24 h

incubation.

Fixation of in situ antigen

In preliminary experiments, the effect of glutaraldehyde at various concentrations

was studied. On the basis of the work of Zaia and Oxman (1977), four levels of

139

glutaraldehyde were chosen, viz. O.Ol%, 0.025%, 0.05% and 0.075%. The experiment

was conducted on glass slides with control slides fixed in absolute ethanol at 20°C for 5

min. Slides were processed by the standard FAMA test procedure. Two batches of VZ

immune globulin were titrated on each slide. One was the standard immune globulin

which, under the conditions of the standard test procedure, showed one plus (+)

staining intensity at a dilution of 1 in 2,000. The other showed a titre of 1,000. Two

plasmas which showed weak or negative staining at a 1 in 10 dilution were also titrated

on each slide to serve as negative controls.

Based on the results of these preliminary studies, the following procedure was

adopted for routine use. At the end of the incubation period, slides and wells were

washed twice with PBS (0.01 M sodium phosphate, 0.145 M sodium chloride, pH 7.2),

flooded with freshly prepared 0.05% (w/v) glutaraldehyde (TAAB Laboratories,

England) in PBS at room temperature for 60 set, then immediately washed in PBS.

Following a second wash in PBS, slides were air dried and stored at -70°C. Microtitre

trays, following their second rinse, were dried overnight at 4°C under vacuum in the

presence of phosphorus pentoxide, then stored immediately in airtight containers at

4°C or -20°C.

Viability testing of fixed in situ antigen

Cells fixed under the above conditions were tested as follows to determine whether

viable virus was still present. Eight bottles were seeded with a mixture of infected and

uninfected cells (1 : 40) and grown for 48 h as described previously. Four bottles were

fixed in 0.05% glutaraldehyde for 60 set then rinsed several times in PBS, the other

four bottles were rinsed in PBS only. Two bottles from each group were overseeded

with uninfected CSL 300 cells. Cells from the remaining four bottles (two fixed, two

unfixed) were scraped separately into 1 ml culture medium, sonicated on a MSE

sonicator for 60 set on high setting, then added to bottles of fresh CSL 300 cells.

CF test procedure

Complement fixation tests were performed essentially according to procedures

outlined by the U.S. Department of Health, Education and Welfare, 1965. Antigen for

the CF test was obtained from Flow Laboratories Australasia P/L, Sydney, Australia.

Plasmas were titrated at doubling dilutions from 1 in 8, and those with a titre of 1 in 32

or greater from varicella zoster convalescent patients were accepted as suitable for

processing to zoster immune globulin. Comparative tests showed that this titre corres-

ponded to a titre of 1 in 100 in the FAMA.

FAMA test procedure

Microscope slides were compartmentalised into 30 separate areas (10 X 3) with a

fine line of fingernail polish. Plasma samples diluted 1 in 100 in enzyme immunoassay

(EIA) diluting buffer (PBS containing 0.5% (w/v) ovalbumin and 0.01% (v/v) Tween

20) were carefully applied to each of the first 24 compartments so that the cell sheet

140

within each compartment was covered but not scratched. The last six compartments

on each slide were reserved for the standard positive and negative controls previously

mentioned. Following a 30-min incubation in a humid chamber at 20°C slides were

washed twice for 5 min each in EIA wash buffer (PBS containing 0.01% (v/v) Tween

20) allowed to drain dry then flooded with an appropriate dilution of fluorescent

conjugate (FITC sheep IgG anti-human IgG). Slides were incubated and washed as

before then examined by narrow band blue incident light fluorescence at an overall

magnification of X 100. Experience was required to distinguish between the apple-

green colour of fluorescein and a dull to moderately bright yellow autofluorescence

induced by the glutaraldehyde fixation.

EIA test procedure

Trays from storage were flooded with diluting buffer, sealed with transparent

adhesive sheets and incubated at 37°C for 30 min then rinsed three times with wash

buffer and three times with PBS. The same plasma dilution preparations used in the

FAMA test were applied to duplicate microtitre trays, 0.1 ml per well, and the trays

sealed and incubated at 37°C for 30 min. Each tray contained 88 test plasmas and

appropriate positive and negative controls. Trays were washed three times in wash

buffer and three times in PBS, then 0.1 ml of urease-labelled sheep IgG anti-human

IgG (CSL, Melbourne, Australia) was added to each well and the trays sealed and

incubated as before. Trays were washed three times in wash buffer and five times in

0.145 M sodium chloride, drained, then 0.1 ml of modified urease substrate solution

was added. (Modified urease substrate was prepared by adding 10-I M EDTA to

standard urease substrate (CSL, Melbourne, Australia) to give a final EDTA concen-

tration of 2 X 10m3 M). The trays were sealed and incubated until the 1 in 2,OOOdilution

of the standard positive control was beginning to change from yellow to purple. Trays

were read on an EIA plate reader (Titertek Multiskan, Flow Laboratories) using a 533

nm filter. A 1 in 2,000 dilution of the standard positive VZ immune globulin was

chosen as the end-point for comparison with the 1 in 100 dilution of test plasma

because, during normal batch processing, IgG concentration and titre are increased

approximately 15 fold. It was therefore assumed that a pool of plasmas of minimum

titre of one-twentieth of the final product would, after processing, yield a product of

comparable titre to the VZ immune globulin batch used as the standard positive

control.

Stability of EIA antigen

Trays for EIA and slides for FAMA were stored dry at 4°C and -20°C. Aliquots of

positive and negative control sera and urease-antibody conjugate in 50% (v/v) glyce-

rol were stored in small glass vials at -20°C. At doubling intervals from 1 wk, slides

and trays were removed from storage and tested against freshly-thawed vials of

control sera and conjugates.

141

RESULTS

Fixation and viability of in situ antigen

Fixation with 0.01% glutaraldehyde was insufficient to completely seal the cell

membrane to penetration by antibody, intracellular staining being quite intense in

these cells. Cells fixed at concentrations of and above 0.025% glutaraldehyde showed

membrane staining only; however, a definite decrease in specific membrane staining

was apparent at the 0.075% level. The 0.05% level was chosen for routine use.

In the retention of viability studies, extensive virus growth was apparent in the

bottles of unfixed unsonicated VZV infected cells but, although the overlay cells grew

well, no foci of virus growth were observed in the bottles of unsonicated VZV-infected

cells fixed in 0.05% glutaraldehyde even after three passages. The unfixed sonicated

cells gave extensive virus growth when added to fresh CSL 300 cells. However, the

number of foci which developed with the fixed sonicated cells was less than 1% of the

number of foci present at the time of fixation. Since each focus can be assumed to

contain a large number of virus particles, this represents a substantial reduction in the

total virus titre.

Comparison of EIA, FAMA and CF tests

A total of 352 plasma samples were tested on three separate occasions by EIA and

FAMA, the former test being performed on duplicate plates on each occasion. Plasma

samples were tested at a single dilution, viz. 1 in 100. In the FAMA test, plasmas

showing the same intensity of staining at this dilution as the standard positive control

diluted to 1 in 2,000 were rated as one plus (+) reactivity. Similarly, in the EIA test, the

purple colour intensity of each well was compared with that of the 1 in 2,000 positive

control well and an equal intensity defined as a one plus (+) end-point. All the plasma

samples had been tested previously for CF antibodies to VZ virus. Comparative

results are summarised in Table 1. The plasma samples could be classified into five

groups:

TABLE 1

Comparative evaluation of in situ EIA, FAMA and CF tests for estimation of antibodies to VZ virus

Plasma samples

Group Number

Test result by

in situ

EIA

FAMA CF

A 141

B 20

C 135

D 27

E 29

Negative

Positive

Negative

Weak positive

Borderline positive

Negative

Positive

Negative

Weak positive

Borderline positive

Negative

Positive

Positive

Positive

24/29 Positive

142

Group A. Negative to trace reactivity in all three tests.

Group B. Significant antibody levels (i.e. 3 +) indicated by all three tests.

Group C. Significant antibody levels in the CF test only.

Group D. Significant antibody levels in the CF test with weak positive reactivity

(i.e. < i-) in the FAMA and EIA.

Group E. Borderline reactivity (i.e.< +) by FAMA and EIA, generally with significant

antibody levels in the CF test (24 of 29). This borderline reactivity was the result of

minor variations within the assay itself whereby samples with an actual titre of, say, 75

would have an observed titre varying from 50 to 100 in replicate assays.

It may be seen that, apart from minor variations, there was complete agreement

between the EIA and FAMA.

Plasma samples from Group C (134 of 135), which were presumed false VZ positive

by CF test, were further tested by FAMA using slides of HSV-1, prepared and treated

as for VZV. Fifty randomly selected plasma samples from Group A were also tested.

All 184 samples were coded before assay. Statistical analysis by the x2 test showed that

the two groups were highly significantly different (P < 0.00 1). In Fig. 1, the results are

presented as the percentage of plasmas showing the specified reactivity so that the two

groups can be directly compared.

Stability of EIA antigens

Plates and slides stored for 16 mth at 4°C and -20°C showed no significant drop in

titre by FAMA or EIA. Time required for full colour development in the EIA

appeared to be somewhat increased (from 20 to 30 min) though this may indicate a

small variation in the titre of the urease conjugate rather than a slight deterioration of

the VZ antigen.

DISCUSSION

The in situ EIA procedure described in this paper, in which antigen is grown and

fixed with glutaraldehyde in the well in which the subsequent EIA is performed, has a

major advantage in the detection of antibodies to VZV in that antigen expressed at the

host cell surface is the only viral antigen accessible to antibody. The in situ EIA can

therefore be expected to be specific for antibodies to VZV because cross-reacting

internal antigen is not expressed at the host cell membrane. This is especially impor-

tant if the assay is to be used to screen random plasma samples for inclusion in a pool

for the preparation of VZ immune globulin. The specificity of the EIA and FAMA for

VZV antibodies is illustrated in Fig. 1. Here, all 184 plasmas were negative to VZV by

both EIA and FAMA but were in two distinct groups according to their reactivity to

VZV in the CF test. It can be seen that the CF positive group (hatched bar) showed a

distinctly higher reactivity to HSV in the FAMA than did the CF negative group (P<

0.001). It is reasonable to conclude, therefore, that a number of the positive reactions

seen in the CF test are due to cross-reacting antibodies to HSV. Similarly the

143

50

40

30

20

10

/ / /

1 k

Open bar - Group A

Hatched bar - Group C

neg trace weak border- POS strong POS line pas POS

Extent of anti-HSV-1 activity as determined in

the HSV specific FAMA teat

Fig. 1. Specific anti-HSV-1 reactivity as determined by a HSV-specific FAMA test in human plasmas from

group A (VZV negative by all tests) and group C (VZV positive only by the CF test) (see Table 1).

remaining nonspecific reactions could be due to antibodies induced by other viruses of

the herpes group. An additional advantage of the EIA is that the use of glutaraldehyde

as fixative greatly reduces the risk of infection of laboratory staffwith VZV, a problem

of concern when most other diagnostic antigens are used in VZV serology. This was

demonstrated by the failure to recover viable virus from fixed culture unless sonica-

tion was used and even then only very low levels of virus could be detected.

There are two possible difficulties which may be anticipated to be associated with an

in situ EIA; an increased time required for the preparation of antigen plates and the

possibility of well-to-well variation. Experience has shown that neither difficulty

arises if the procedure, as outlined, is observed. Considering the growth of virus and

preparation of antigen plates, the number of manipulations is less when the virus is

144

grown and fixed in situ and the additional time required to perform these operations in

a sterile cabinet is more than compensated by not having to harvest or purify the virus.

Considering the second potential difficulty, it was expected that there would be a

greater well-to-well variation in the amount of antigen in in situ plates compared with

plates coated with soluble antigen because well-to-well variations in virus yield would

be expected to occur. While this may affect the maximum amount of antibody which

can be bound in a well, at a titration end-point, available antigen is in considerable

excess of antibody so that unless a well is grossly deficient in antigen, results would be

expected to be reproducible. This reproducibility was established by consistent read-

ings for the positive control plasma on a large number of plates. A gross deficiency in

antigen would generally be associated with poor cell growth which should show up as

an alkaline well during virus growth. If this is suspected subsequent to performance of

the EIA, the well can then easily be examined microscopically for viral cytopathic

effect on the cells.

In the determination of the optimum concentration of glutaraldehyde for in situ

antigen fixation, there were three considerations of importance:

(i) concentration of glutaraldehyde required to prevent intracellular penetration of

antibody;

(ii) concentration of glutaraldehyde which would preserve viral antigens expressed at

the cell surface;

(iii) concentration of glutaraldehyde required to inactivate free and intracellular virus.

The concentration chosen gaveseeminglycompletesealingofthe host cell membrane

without significant loss of antigenicity. This sealing of the host cell membrane gave the

in situ EIA procedure its two major advantages over other diagnostic procedures, viz.

non-involvement of nonspecific internal antigens, and effective operator safety from

residual infectious virus.

The basic EIA procedures described in this paper are essentially the same as

described elsewhere (Chandler et al., 1982) except for the increase in EDTA concen-

tration of the enzyme substrate solution from 1O-4 M to 2 X 1O-3 M. In the absence of

this comparatively high level of retardant buffer salt, nonspecific reactivity, due

perhaps to a neutral buffering effect of the fixed cultured cells, was found to be a

problem. The increase in EDTA effectively overcame this problem.

ACKNOWLEDGEMENTS

We thank Aishah Ibrahim for excellent technical assistance, David Pye for valued

guidance, and the Director of the Commonwealth Serum Laboratories for supporting

this investigation.

145

REFERENCES

Center for Disease Control, 1979, Varicella roster immune globulin. Morbidity and Mortality Weekly

Report 28, 589.

Chandler, H.M., J.C. Cox, K. Healey, A. MacGregor, R.R. Premierand J.G.R. Hurrell, 1982, J. Immunol.

Methods 53, 187-194.

Schiff, P., 1979, Med. J. Aust. 2, 604.

U.S. Department of Health, Education and Welfare, 1965, Standardized diagnostic complement fixation

method and adaptation to microtest, Public Health Monograph No. 74, Public Health Service Publication

No. 1228.

Zaia, J.A. and M.N. Oxman, 1977, J. Infect. Dis. 136, 519-530.