detection of igg antibody to epstein-barr virus viral capsid antigen in saliva by antibody capture...

9
ELSEVIER Journal of Virological Methods 63 (1997) 93-101 Journal of Virological Methods Detection of IgG antibody to Epstein-Barr virus viral capsid antigen. in saliva by antibody capture radioimmunoassay A.J. Vyse*, W.A. Knowles, B.J. Cohen, D.W.G. Brown Enteric and Respiratory Virus Laboratory, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK Accepted 11 August 1996 Abstract A ‘G’ antibody capture radioimmunoassay (GACRIA) to detect IgG to Epstein-Barr virus (EBV) viral capsid antigen (VCA) in saliva is described. The monoclonal antibody to EBV VCA used in the GACRIA bound non-specifically when testing saliva samples having a total IgG content of less than 2.0 mg/l, so giving false positive results. This problem was overcome by including 0.5% EBV-negative human serum in the monoclonal antibody diluent. The perfonmance of the assay was then evaluated by comparing the GACRIA test on serum and saliva samples to indirect immunofluorescence assay (IFA) results on sera using a panel of paired serum/saliva samples. Compared to the corresponding serum IFA the saliva GACRIA had a sensitivity and specificity of 93.5 and lOO%, respectively. Although less sensitive than IFA on serum samples, the saliva GACRIA is sufficiently sensitive to be used for epidemiological screening and will enable testing for anti-EBV VCA to be carried out easily and on a wide scale. Copyright 0 1997 Elsevier Science B.V. Keywords: Epstein-Barr virus; GACRIA; Non-specific binding; Saliva; Virus capsid antigen 1. Introduction Infection with Epstein-Barr virus (EBV), one of the eight known human herpesviruses, is * Corresponding amhor. Tel.: + 44 181 2004400; fax: + 44 181 2001569. widespread amongst humans. The virus is trans- mitted oro-orally and after primary infection persists in the host for life, in a latent state. Infection is associated with several disease states. Clinical manifestations of primary infec- tion range from asymptomatic in the majority of individuals to infectious mononucleosis, a condi- tion more frequently seen when primary infec- 0166-0934/97/$17.00 C’opyright0 1997 Elsevier Science B.V. All rights reserved PII SOl66-0934(96)02118-O

Upload: aj-vyse

Post on 16-Sep-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

ELSEVIER Journal of Virological Methods 63 (1997) 93-101

Journal of Virological Methods

Detection of IgG antibody to Epstein-Barr virus viral capsid antigen. in saliva by antibody capture radioimmunoassay

A.J. Vyse*, W.A. Knowles, B.J. Cohen, D.W.G. Brown

Enteric and Respiratory Virus Laboratory, Central Public Health Laboratory, 61 Colindale Avenue, London NW9 5HT, UK

Accepted 11 August 1996

Abstract

A ‘G’ antibody capture radioimmunoassay (GACRIA) to detect IgG to Epstein-Barr virus (EBV) viral capsid antigen (VCA) in saliva is described. The monoclonal antibody to EBV VCA used in the GACRIA bound non-specifically when testing saliva samples having a total IgG content of less than 2.0 mg/l, so giving false positive results. This problem was overcome by including 0.5% EBV-negative human serum in the monoclonal antibody diluent. The perfonmance of the assay was then evaluated by comparing the GACRIA test on serum and saliva samples to indirect immunofluorescence assay (IFA) results on sera using a panel of paired serum/saliva samples. Compared to the corresponding serum IFA the saliva GACRIA had a sensitivity and specificity of 93.5 and lOO%, respectively. Although less sensitive than IFA on serum samples, the saliva GACRIA is sufficiently sensitive to be used for epidemiological screening and will enable testing for anti-EBV VCA to be carried out easily and on a wide scale. Copyright 0 1997 Elsevier Science B.V.

Keywords: Epstein-Barr virus; GACRIA; Non-specific binding; Saliva; Virus capsid antigen

1. Introduction

Infection with Epstein-Barr virus (EBV), one of the eight known human herpesviruses, is

* Corresponding amhor. Tel.: + 44 181 2004400; fax: + 44 181 2001569.

widespread amongst humans. The virus is trans- mitted oro-orally and after primary infection persists in the host for life, in a latent state. Infection is associated with several disease states. Clinical manifestations of primary infec- tion range from asymptomatic in the majority of individuals to infectious mononucleosis, a condi- tion more frequently seen when primary infec-

0166-0934/97/$17.00 C’opyright 0 1997 Elsevier Science B.V. All rights reserved

PII SOl66-0934(96)02118-O

Page 2: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

94 A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101

tionis delayed until adolescence or adulthood. Further conditions associated with EBV infec- tion include undifferentiated nasopharyngeal carcinoma (NPC)-the most prevalent tumour in males from southeast China, Burkitt’s lymphoma, monoclonal and polyclonal lymphomas in immunocompromised patients and other lymphoproliferative conditions (Henle and Henle, 1980; Ambinder and Mann, 1994).

Primary EBV infection leads to the appear- ance of antibodies to various EBV-specific anti- gens. Typically, the antibody pattern of a primary EBV infection is characterized by IgM and IgG antibodies to virus capsid antigen (VCA), IgG antibodies to early antigens (EA) and the absence of antibodies to nuclear anti- gens (EBNA), which only appear several weeks after the onset of disease (Crawford and Ed- wards, 1990; Andersson et al., 1994). Once in- fected, IgG antibodies to VCA persist for life although other antibodies fluctuate depending upon the nature of infection and the course of disease, giving EBV-specific antibody patterns (Andersson et al., 1994). Therefore, as VCA-IgG is produced early in infection and persists, it makes a suitable marker to test for evidence of past EBV infection and for studying the epi- demiology of EBV.

Saliva is a body fluid that contains antibodies of diagnostic significance, the antibody content of salivary crevicular fluid reflecting that of plasma but at lower concentrations. However, by use of sensitive antibody capture rather than indirect assays it is possible to detect antibodies to a variety of viral antigens in saliva (Mortimer and Parry, 1991; Perry et al., 1993). Therefore, saliva is a viable alternative to blood for viral diagnosis. Furthermore, the collection of saliva specimens has several advantages over venepunc- ture: it is convenient and can be done by un- trained persons, e.g. parents; is non-invasive, painless and less hazardous, thus giving better access to large populations, hard-to-reach groups and children. We describe here the eval- uation of an assay to detect IgG antibody to EBV VCA in saliva which will be useful for epidemiological studies.

2. Materials and methods

2.1. Saliva collection

Saliva specimens, except where otherwise stated, were collected using a sterile foam swab (Brown et al., 1994). This was used to collect saliva from the gingival crevicular margin by brushing for approximately 1 min. Saliva was extracted from the swabs in 1 ml of phosphate buffered saline (PBS), pH 7.2, containing 10% fetal calf serum (FCS), 0.2% Tween 20 (T20)), 0.5% gentamicine and 0.2% fungizone, by vortex- ing and centrifugation (Mortimer and Parry, 1991).

2.2. Sera, paired serum /saliva panels and salivas

All samples were stored at - 20°C until re- quired for testing.

2.2.1. Sera (1) Thirty-seven sera (26 positive and 11 negative

for EBV VCA IgG by IFA) were provided by Prof. D. Crawford of the London School of Hygiene and Tropical Medicine.

2.2.2. Serum/saliva pairs

(1)

(2)

(3)

(4)

(5)

Twenty-one pairs from suspected NPC pa- tients, provided by Dr W.L. Lim, Queen Mary Hospital, Hong Kong. Eight pairs from a study of a measles out- break in Cumbria, provided by Alison Richards, Virus Reference Division (VRD) (Calvert et al., 1996). Two pairs of known EBV VCA IFA status from a study of oral hairy leucoplakia in human immunodeficiency virus (HIV) pa- tients, provided by Dr C.G. Teo, VRD. The saliva had been collected by dribbling into a clean plastic container. Forty-three pairs from healthy staff attending Occupational Health provided by Dr P. Rice, St Thomas’s Hospital, London. Four pairs provided by VRD staff.

Page 3: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101 95

2.2.3. Salivas (1) Nine salivas provided by VRD staff. (2) Salivas (486) from children aged l-5 years

enrolled in a Leukaemia Research Fund (LRF) study of childhood acute lymphoblas- tic leukaemia, provided by Dr F. Alexander (University of Edinburgh):

2.3. Total IgG assay

The total IgG content of saliva specimens measured using an indirect enzyme linked munosorbent assay (ELISA) as previously scribed (de Azevedo Neto et al., 1995).

2.4. Antigen preparation

was im- de-

EBV antigen was obtained from the EBV genome-positive F’,HRl lymphoblastoid cell line which was cultured in RPM1 1640 medium sup- plemented with 10)/o FCS. The cells were passaged every 7- 10 days b,y reseeding in fresh medium. To prepare viral antigen the cultures were centrifuged at 2000 x g for 10 min and the pellets resuspended in one tenth the original volume of supernatant fluid. The resulting suspension was sonicated for 60 s in a bath sonicator, aliquoted and stored at - 70°C.

2.5. Immunojluorescence assay

Cells were pelleted from 1 ml of P,HRl cell suspension, washed twice in PBS by centrifuging for 10 min at 400 x g, and resuspended in approx- imately 5 ml of P13S. Cell suspension (20 ~1) was applied to each well of a teflon coated slide and air dried. The slides were then fixed in acetone for 5 min at room temperature and stored at - 20°C until required. Doubling dilutions of each serum sample, from l/20 to l/2560, were made in PBS containing 10% FCS and 0.05% T20. Per well, 20 ~1 of each dilution was added, incubated at 37°C for 30 min and then washed twice with PBS followed by one wash with distilled water. Fluo- rescein conjugated rabbit immunoglobulin to hu- man IgG (Dakopatts code F202) was added at an optimal working dilution (l/40) with Evans Blue counterstain at l/100. The slides were incubated

for 30 min at 37°C washed twice with PBS and once with sterile distilled water before examining with a Zeiss D-7082 fluorescence microscope. Samples were considered negative if no fluores- cence was seen at a dilution of l/20 and positive if cells with brightly fluorescent cytoplasm (Henle and Henle, 1966) were observed at l/20 and higher dilutions.

2.6. EBV viral capsid antigen %’ antibody capture radioimmunoassay (GACRIA)

After a series of experiments to optimise assay conditions the following procedure was used.

Polystyrene beads (Northumbria Biologicals P201) were coated by immersion for 2 h at room temperature in rabbit antibody to human IgG (gamma chain specific, Dakopatts Code A423) diluted l/l000 in 0.05 M sodium carbonate/bicar- bonate buffer pH 9.6. The coated beads were stored at 4°C for 18 h or until used. After washing with PBS T20 (0.05% T20) the anti human IgG- coated beads were incubated successively in 200 ~1 of the following: a 1 in 3 dilution of saliva or 1 in 100 dilution of serum in PBS containing 10% FCS and 0.2% T20 for 2 h at 37’C; viral antigen, diluted 1 in 5 in PBS containing 10% FCS and 0.2% T20 at 37°C for 18 h; a l/l 000000 dilution of a mouse ascites containing anti-EBV viral capsid antigen gp 125 monoclonal antibody (MAb)(Chemicon MAB 817) diluted in PBS con- taining 10% FCS, 0.2% T20, 1% normal rabbit serum (NRS), 1% normal goat serum (NGS) and 0.5% human serum negative for anti-EBV VCA (NHS); 1251-labelled anti-mouse IgG (Amersham International, Aylesbury, UK, code no. IM131) diluted in PBS containing 10% FCS, 5% NHS, 5% NRS, 5% NGS and 0.2% T20 to give 100000 counts per min in 200 ml. Each of the last two stages was incubated for 2 h at 37°C. Between each stage of the assay, beads were washed with PBS T20 (0.05% T20) using a ‘Qwikwash’ appara- tus (Abbott Laboratories). The bound 125I was measured in a gamma counter (NE1600, Nuclear Enterprises Ltd) for 300 s.

In each assay two wells of NHS (diluted l/100) and two of a serum strongly positive for EBV

Page 4: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

96 A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101

VCA IgG (diluted l/ 100) were included as con- trols. Test results were calculated as the total bound radioactivity of each specimen divided by the mean radioactivity bound by the nega- tive serum controls and expressed as test: negative (T/N) ratios. A suitable cut-off value was determined using the results from 486 sali- vas from children aged l-5 years provided by Dr F. Alexander, using a mixture models method based on that described by Parker et al., 1990. It is not necessary to identify an unin- fected population using this technique which in- corporates methods to estimate the probability that an individual will be accurately classified by calculating a cut-off value that will give an equal number of false positive and false negative

results.

3.2. EfJ;?ct of total IgG concentration in saliva on

EBV VCA IgG detection

When testing salivas some non-specific reactiv- ity was observed with the assay, e.g. 45 of 98 LRF samples initially tested showed non-specific bind- ing when tested in parallel with a ‘no antigen’ control well. Results with ten salivas illustrating the range of reactions are shown in Table 1 (samples l-10). The source of the non-specific reactivity was investigated by the experiment illus-

A

26 -

Applying this mixture models method to the data gave a cut-off T/N ratio of 2.1. Samples were considered positive if they had a T/N > 2.1, equivocal at a T/N 2.1 - 1.8 and negative at a T/N < 1.8.

24 -

22 -

20 -

3. Results 18 -

3.1. Optimisation of assay conditions 16 -

A series of experiments was undertaken to determine optimal working dilutions of anti human IgG, test sample (serum and saliva),

antigen and monoclonal antibody (results not shown). This resulted in the chosen dilu- tions as described earlier in Materials and meth-

ods.

14 -

TIN

12 -

10 -

8- Initially the antigen incubation stage of the

assay was carried out at 4°C but in order to enhance assay sensitivity the effect of raising the temperature to 37°C was examined. For this, the 37 (of which 26 were VCA IgG positive by IFA) sera provided by Professor D. Crawford were tested by incubating with the antigen at 4°C or 37°C for 18 h. Increasing the incubation tempera- ture with the antigen to 37°C resulted in higher binding with positive samples and only a minimal increase in binding with negative samples (Fig. 1) thus giving higher T/N values and improved dis- crimination between positive and negative sera.

6-

4-

2-

4°C 37°C

Fig. 1. EBV VCA GACRIA T/N ratios obtained by testing a

panel of 37 sera of known EBV VCA IgG IFA status. Incuba-

tion with antigen was carried out for 18 h at 4°C or 37°C.

*

l

:

P Q

&

Page 5: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101 97

Table 1 EBV VCA GACRIA and total IgG results on saliva samples tested with and without antigen

Sample No. Dilution T/N (with Ag) T/N (without Ag) Total IgG mg/l

1 Neat 3.9 2.5 1.4

2 Neat 4.2 3.9 0.9

3 Neat 10.7 8.8 0.5 4 Neat 16.0 11.0 0.1

5 Neat 17.5 10.8 0.6 6 Neat 1.0 0.7 15.8 7 Neat 1.1 1.0 5.4 8 Neat 8.4 1.0 2.7

9 Neat 16.2 1.7 3.8 10 Neat 2.4 0.8 6.7

11 Neat 22.6 0.9 3.5 l/16 13.6 7.5 0.2

12 Neat 1.1 0.8 8.3

l/16 7.3 2.6 0.5 13 Neat 22.1 0.7 3.2

l/16 16.8 6.3 0.2

14 Neat 10.8 0.6 8.8 l/16 12.2 3.6 0.6

15 Neat 5.6 0.5 18.8 l/16 6.1 2.2 1.2

Samples l-5 with total IgG<2 mg/l show non-specific reactions.

Samples 6-10 with total IgG>2 mg/l show specific reactions.

Samples 1 l-15 were tested neat and diluted confirming the effect of low total 1gG concentration on test specificity.

trated in Fig. 2 in which the assay was carried out omitting reagents at different stages. Non-specific binding was observed when the specimen was

omitted and, to a lesser extent, when both the specimen and antigen were omitted implicating the monoclonal antibody stage. In addition it was observed that this non-specific binding occurred only with those salivas which had total IgG con- centrations of < 2 mg/l (Table 1, samples l-5), whereas salivas with total IgG concentrations > 2 mg/l gave specific reactions (Table 1, samples 6-10). It was confirmed by testing neat and di- luted salivas of known EBV VCA IgG status and total IgG concentrations > 2 mg/l that with the diluted salivas false positive results were liable to occur (Table 1, samples 1 l- 15). The critical total IgG concentration was determined as 2.0 mg/l by the experiment illustrated in Fig. 3 in which two salivas, one positive and one negative for EBV VCA IgG, were diluted to contain levels of total IgG from 1 to 8 mg/l. To overcome the problem of non-specific binding, 0.5% NHS was added to

the monoclonal antibody diluent. Under these conditions all of the 45 LRF samples initially showing non-specific binding gave satisfactory re-

sults. Overall, we tested 466 of the 486 saliva samples collected for the LRF study for total IgG. All 466 samples contained detectable IgG, range 0.1-56.2 mg/l, mean 9.3 mg/l, with 80/466 (17%) saliva samples containing less than 2 mg/l total IgG.

Further experiments were done to study the effect of the inclusion of 0.5% NHS in the mono- clonal antibody diluent on overall T/N values. A panel of salivas provided by VRD staff, contain- ing satisfactory total IgG concentrations (range 17.5-70.5 mg/l, mean 36.0 mg/l) was tested as described using the standard method with and without NHS. Without NHS T/N values ranged from 1.0 to 27.5 with a mean of 9.1, and with 0.5% NHS T/N values ranged from 0.6 to 30.1 with a mean of 8.4, thus confirming that the addition of 0.5% NHS did not significantly affect binding in adequate saliva samples.

Page 6: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

98 A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101

Solid phase Anti-Human Serum sample Antigen MAb 125, Mean counts

w (EBV VCA IgG Pas) per 5 min

0 ---d G 0 t-- h

J J J I % J 119

J J E J E J 183

J / E I( f J 132

J E I( I( J J 74

J E J I( I( J 78

J I( E J I( J 84

J - reagent included in test

I( - reagent omitted

Fig. 2. Identifying the source of non-specific binding in the EBV VCA GACRIA by omitting reagents at different stages of the assay. The test sample was a serum with a high titre of EBV VCA IgG from a NPC patient (provided by Dr Teo, VRD).

3.3. Evaluation of assay

The GACRIA for EBV VCA IgG was evalu- ated initially using serum samples that had been tested by IFA in other laboratories. A sensitivity and specificity of 88.5 and 81.8, respectively, was obtained for the serum GACRIA when compared with IFA. Evaluation of assay performance was then assessed using a panel of 78 serum/saliva

pairs (21 from Dr Lim, 43 from Dr Rice, 8 from the Cumbria study, 2 from Dr Teo and 4 from VRD staff), with sera tested by us using the IFA as described earlier. Of 78 serum/saliva pairs tested, 74 gave complete agreement between sali- vary and serum GACRIA and serum IFA results: 16 negative and 58 positive by all three tests. Four serum/saliva pairs gave discrepant results: two of the 43 sera provided by Dr Rice were salivary

Page 7: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101 99

GACRIA negative but serum GACRIA positive a problem when testing serum since it contains a and serum IFA positive; two (one of the 43 sera much higher immunoglobulin concentration than provided by Dr -Rice and one of the 21 sera saliva, thus completely saturating the solid phase provided by Dr Lim) were salivary GACRIA anti human IgG (Mortimer and Parry, 1991). The negative, serum GACRIA negative but serum phenomenon of non-specific binding by the mon- IFA positive. Overall, this assessment showed that oclonal antibody was overcome by including 0.5% the serum GACRIA had a sensitivity and specific- NHS in the monoclonal antibody diluent. Human ity of 96.8 and lOO%, respectively, compared to immunoglobulin in NHS added at this stage satu- corresponding serum IFA results (Fig. 4). With rates the solid phase anti human IgG, preventing the panel of 78 serum/saliva pairs the saliva the monoclonal antibody from binding non-spe- GACRIA had a sensitivity and specificity of 93.5 cifically and eliminating the possibility of false and 100% compared to serum IFA. positive results.

4. Discussion

During assay development false positive reac- tions were identified which were shown to occur only with salivas with a total IgG concentration of less than 2.0 mg/l. The false positive reactions occurred because the monoclonal antibody was binding non-specilically to unsaturated anti hu- man IgG on the solid phase (Fig. 2). This was not

Pm)-

The quantity of total IgG in a saliva sample depends on the collection device used. Parry (1993) found that the ‘Orasure’ device yielded higher total IgG (mean 40.5 mg/l) concentrations than the ‘Salivette’ (mean 20.6 mg/l). In that study it was concluded that a minimum of 0.1 mg/l of total IgG was required for HIV antibody testing. de Azevedo Neto et al. (1995) found that dribbled saliva contained higher total IgG (mean 18.5 mg/l) concentrations than saliva collected by the ‘Omnisal’ device (mean 1.25 mg/l) but they were unable to determine the minimum concen- tration required for rubella antibody studies. The levels of total IgG detected in saliva collected with the foam swab used in this study (mean 9.3 mg/l) were comparable to those obtained in other stud- ies using different commercially available devices.

- EB” VCA QG posmve wfm onngen

4cm- ...‘....’ EB” VCA l&s negattve vah onngen

.---- EB”“CAIgGrxxm”*timwt~~”

-.-.- EB” VCA IQG “*gotlve wHhou+ onfQ.3”

3ooo ,. ; k.

2 rm - ‘,, \

‘\ \

1 WI- \ ‘...

\. ‘\ ‘.. y,:“.. ,.

“. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ~~---______________--------______ ~.-._._._.-._._._._.~_.-.-.-._._._._._._._.

0, , + 10 20 30 40 5.0 60 70 80

Total IgG (mgll)

Fig. 3. The effect of total IgG concentration in saliva on the “‘1 count in the EBV VCA GACRIA. Results given by an EBV VCA IgG positke and by a negative saliva diluted to give various concentrations of total IgG are shown, tested with and without EBV VCA antigen.

Our evaluations showed the saliva GACRIA to be less sensitive than serum IFA. Some of the IFA results, however, were provided by other labora- tories, and the subjective nature of IFA and lack of standardisation between laboratories may ex- plain discrepancies between serum IFA and saliva GACRIA results. We are not advocating the use of the saliva EBV VCA GACRIA for diagnosis but our results suggests its performance is ade- quate for epidemiological studies. In preliminary tests (data not shown) we have modified the GACRIA to detect EBV EA IgA (AACRIA) by coating the solid phase with anti-human IgA and by using a monoclonal antibody specific for EA. This assay may be useful for the diagnosis/prog- nosis of NPC (Yao et al., 1991).

The assay for EBV VCA IgG described here is suitable for testing large numbers of samples and the results are read objectively and are amenable

Page 8: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

100 A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101

Serum IFA vs Serum GACRIA

IFA

+ SERUM GACRIA + 60 0

2 16

Sensitivity of serum GACRIA = 96.8% (95% C.I. = 88.8 - 99.6)

Specificity of swum GACRIA = 100% (95% C.I. = 79.4 - 100)

Serum IFA vs Saliva GACRIA

IFA

+ SALIVA GACRIA + 58 0

I 4 16

Sensitivity of saliva GACRlA = 93.5% (95% C.I. = 84.3 - 98.2)

Specificity of saliva GACRIA = 100% (95% C.I. = 79.4 - 100)

Fig. 4. Evaluation of serum and saliva EBV VCA GACRIA performance by comparison with serum IFA using 78 serum/saliva pairs.

to statistical analysis. Moreover, as it has the advantage of utilising saliva rather than serum it will enable easier access to large population

groups, including children. This will permit wider epidemiological studies of EBV infection than are currently possible because of limita- tions imposed by the need to collect blood sam- ples.

Acknowledgements

We thank Prof. D. Crawford, London School of Hygiene and Tropical Medicine for providing the sera of known EBV VCA IFA status; Dr W.L. Lim, Queen Mary Hospital, Hong Kong, Dr C.G. Teo, VRD, Alison Richards, VRD and

Dr P. Rice, St Thomas’s Hospital, London, for the paired serum/salivas; and Dr Freda Alexan- der, University of Edinburgh for saliva samples. We also thank Mr N. Gay, PHLS Communica- ble Disease Surveillance Centre, Colindale, for his help with the statistical analysing of EBV VCA antibody results and Mrs I. Seth for typ- ing the manuscript. This work was funded by the Leukaemia Research Fund.

References

Ambinder, R.F. and Mann, R.B. (1994) Detection and charac- terisation of Epstein-Barr Virus in clinical specimens. Am. J. Pathol. 145, 239-252.

Andersson, A., Vetter, V., Krevtzer, L. and Barer, G. (1994) Avidities of IgG directed against viral capsid antigen: use

Page 9: Detection of IgG antibody to Epstein-Barr virus viral capsid antigen in saliva by antibody capture radioimmunoassay

A.J. Vyse et al. /Journal of Virological Methods 63 (1997) 93-101 101

of markers for significant Epstein-Barr Virus serology. J. Med. Viral. 43, 23&244.

Brown, D.W.G., Ramsay, M.E.B., Richards, A.F. and Miller, E. (1994) Salivary diagnosis of measles: a study of notified cases in the United Kingdom, 1991-1993. Br. Med. J. 308, 1015--1017.

Calvert, N., Cutts, F.. Irving, R., Brown D., Marsh, J. and Miller, E., (1996) Measles immunity and responses to revaccination among secondary school children in Cum- bria. Epidem Infect, 116, 65-70.

Crawford, D.H. and Edwards, J.M.B. (1990) Epstein-Barr Virus. In A.J. Zuclterman, J.E. Banatvala and J.R. Patti- son (Eds), Principles and Practice of Clinical Virology, Wiley, Chichester, pp. 1 1 l- 133.

de Azevedo Neto, R.S., Richards, A., Nokes, D.J., et al. (1995) Salivary antibody detection in epidemiological sur- veys: a pilot study after a mass vaccination campaign against rubella in Sao Paula, Brazil. Trans. R. Sot. Trop. Med. Hyg. 89, 115-118.

Henle, G. and Henle, W. (1966) Immunofluorescence in cells derived from Burkitt’s Lymphoma. J. Bacterial. 91, 1248- 1256.

Henle, W. and Henle, G. (1980) Clinical Spectrum of Ep-

stein-Barr Virus Infection. In A.J. Nahmias, W.R. Dowdle and Schinazi, R.F. (Eds), The Human Her- pesviruses-An Interdisciplinary Perspective, Elsevier, New York, pp. 31-34.

Mortimer, P.P. and Parry, J.V. (1991) Non-invasive virologi- cal diagnosis: Are saliva and urine specimens adequate substitutes for blood? Rev. Med. Virol. 1, 73-78.

Parker, R.A., Erdman, D.D. and Anderson, L.J. (1990) Use of mixture models in determining laboratory criterion for identification of seropositlve individuals: application to parvovirus B19 serology. J. Virol. Methods. 27, 135-144.

Parry, J.V. (1993) Simple and reliable salivary tests for HIV and Hepatitis A and B virus diagnosis and surveillance. Ann. NY Acad. Sci. 694, 216-233.

Perry, K.R., Brown, D.W.G., Parry, J.V., Panday, S., Pipkin, C. and Richards, A.R. (1993) Detection of measles, mumps and rubella antibodies in saliva using antibody capture radiolmmunoassay. J. Med. Viral. 40, 235-240.

Yao, Q.Y., Rowe, M., Morgan, A.J., Sam, C.K., Prasad, V., Dang, H., Zeng, Y. and Rickinson, A.B. (1991) Salivary and serum IgA antibodies to the Epstein-Barr Virus gly- coprotein gp 340: Incidence and potential for virus neu- tralisation. Int. J. Cancer 48, 45-50.