Journal of Virological Methods, 32 (1991) 255-263 0 1991 Elsevier Science Publishers B.V. / 0168-8510/91/$03.50

ADOh’IS016885109100128K

255

VIRMET 01159

Detection of total antibody against hepatitis A virus by an automated microparticle enzyme immunoassay

David J. Robbins, James Krater, Willie Kiang, Xiomara Alcalde, Shaun Helgesen, Joseph Carlos and Larry Mimms

Hepatitis R&D, Abbott Laboratories. Abbott Park, IL, U.S.A.

(Accepted 2 January 199 1)

Summary

A fully automated microparticle enzyme immunoassay, IMx HAVAB, was develo- ped for the detection of antibody against hepatitis A virus (anti-HAV). In the IMx HAVAB assay which is run on the IMx instrument, 24 tests are completed in less than 45 minutes. IMx HAVAB sensitivity was 18-25 World Health Organization U/l and was more sensitive than the commercial RIA or EIA, HAVAB and HAVAB EIA, respectively. Specimens from blood donors, diagnostic and hospital patients and indi- viduals with a variety of infectious and immune diseases were tested in parallel with IMx HAVAB and RIA or EIA. Overall agreement of 99.9% (2118/2121) was obtained.

Prevalence of anti-HAV tested by IMx ranged from 12.3% in volunteer blood donors in St. Louis to 64.3% for hospital patients in New York City. Discordant speci- mens were reactive by IMx HAVAB but borderline negative by EIA or RIA, due to the better sensitivity of the IMx assay. IMx HAVAB detected both IgM and IgG sub- classes of anti-HAV. Serial bleeds from six intravenous drug users with acute HAV infection were tested over 8 months for the presence of anti-HAV. At all time points, patients were strongly reactive for anti-HAV (titers > l/1000). Anti-HAV titers rose during the first 20 weeks after presentation of symptoms and then declined with time.

Hepatitis A virus; Enzyme immunoassay; Anti-HAV; HAV

Correspondence to: D.J. Robbins, Hepatitis R&D, Abbott Laboratories, Abbott Park, IL 60064, U.S.A.

256

Introduction

The World Health Organization has declared hepatitis A to be one of the four most important viral diseases. In most developing countries, hepatitis A has been primarily a childhood disease with prevalence of >90%. Improved living conditions and sani- tation in some countries has resulted in a greater number of adults being susceptible to and contracting acute hepatitis A (Gust, 1988). In the U.S.A., thirty to forty percent of the general population show serological evidence for previous exposure to HAV.

Antibody to hepatitis A virus (anti-HAV) is the first serological marker of HAV infection and is almost always detectable at the onset of symptoms. The earliest response is composed primarily of IgM antibody subclass which is transiently detected for 3-6 months after infection and therefore is a good marker for acute HAV infection (Decker et al., 1981). The IgM antibody response is followed by IgG anti- HAV which persists for years after recovery and confers permanent immunity against reinfection. Highly specific and sensitive radio- and enzyme-linked immunoassays (RIA and EIA) are commercially available for the detection of total antibody and IgM class specific antibody to HAV (Safford et al., 1980; Decker et al., 1981). Tests for measuring total anti-HAV have been used to assess immune status in naturally infected (Lemon and Binn, 1983; Recommendation of ACIP, 1990) and experimen- tally vaccinated individuals (Flehmig et al., 1988; Provost et al., 1988; Sjorgen et al., 1988) and for epidemiological studies (Tassioiykis et al., 1987; Yang et al., 1988).

In this paper, we report the development of a test for total anti-HAV using a micro- particle enzyme immunoassay (IMx HAVAB) which was fully automated on the Abbott IMx instrument. This assay is more sensitive and reproducible than previous procedures and allows results to be reported in less than 45 minutes. Specificity of the assay was demonstrated by testing 2 12 1 specimens in parallel with IMx HAVAB and current EIA or RIA.

Materials and Methods

Test kits Serum and plasma specimens were tested with commercial RIA and EIA reagents supplied by Abbott Laboratories: HAVAB and HAVAB EIA for total anti- HAV and HAVAB-M EIA or IMx HAVAB-M for IgM anti-HAV.

Samples Specimens for in-house and clinical testing were obtained from hospital diagnostic laboratories, reference laboratories and hospital, volunteer and commer- cial blood banks. Specimens from patients with diseases other than hepatitis A included infection with hepatitis B virus (HBV), Epstein-Barr virus (EBV), cytome- galovirus (CMV), hepatitis C virus (HCV), rubella and herpes simplex virus. Serial bleeds were collected from 6 intravenous drug abusers who contracted HAV infec- tion. The route of infection was unknown.

/MxHAVAB The IMx assay for anti-HAV is based on microparticle enzyme immu- noassay (MEIA) technology used in the IMx instrument system (Fiore et al., 1988).

257

Briefly, the MEIA technology uses microparticles approximately 0.2 microns in diameter as the solid phase. Separation of bound from unbound material is accomplished by capture of the microparticles on a glass fiber matrix. Microparticles allow increased kinetics of reaction (total time of assay is less than 4.5 min for 24 speci- mens) compared to other solid phase and may be easily manipulated by the instrument robotics for automation of the assay steps. No operator intervention is required other thanpipetting of the samples into the specimen wells, placing the reagent into the ana- lyzer and initiating the assay.

Carboxylated latex particles (0.1 to 0.3 ,um diameter) (Seradyn, Indianapolis, IN) were coupled to protein A purified mouse monoclonal anti-HAV (Abbott Laborato- ries; Dawson et al., 1984) using 1-ethyl-3,3-(dimethylaminopropyl) carbodiimide chemistry (Quash et al., 1978). After the unbound anti-HAV was washed away, the microparticles were overcoated with HAV produced by tissue culture (Dawson et al., 1984). DEAE purified human anti-HAV (Abbott Laboratories) was coupled to alka- line phosphatase using a hetero-bifunctional crosslinking reagent (N-hydroxy succinimide ester and maleimide) (Yoshitake et al., 1979).

In the first step of the IMx assay, microparticles coated with HAV are added to the specimen. This mixture is transferred to the glass fiber matrix and anti-HAV: alkaline phosphatase conjugate is added. Anti-HAV, if present in the specimen, blocks the binding of anti-HAV:alkaline phosphatase conjugate. Finally, the substrate, methyl- umbelliferyl phosphate, is added and a fluorescent signal is generated by dephospho- rylated product. This fluorescence rate increase is inversely related to the amount of anti-HAV captured by the microparticles from the specimen.

ZMs HAVAB assay cutoff Results are given as fluorescent rate values (counts/s/s) and as specimen rate divided by cutoff rate (S/CO). The assay cutoff is calculated as the Mode 1 Calibrator rate divided by 2. The Mode 1 Calibrator is a pool of recalcified human plasma testing negative for anti-HAV (NHP). A specimen is considered reac- tive for anti-HAV when S/CO is less than or equal to 1 .OOO.

Stutistical methods Intra-assay (within run), inter-assay (between run) and total variation were calculated as described by Krouwer and Rabinowitz (1984).

Sensitivity panel Recalcified pooled human plasma testing strongly reactive for anti-HAV was diluted into NHP to produce a sensitivity panel containing varying anti-HAV concentrations. Concentrations were determined by comparing each panel member to a World Health Organization (WHO) Standard provided by the Central Laboratory of the Netherlands Red Cross (Amsterdam). Sensitivity of an assay was determined as the anti-HAV concentration in WHO U/l at the assay cutoff read from a standard curve generated from the sensitivity panel.

2.58

Results

Sensitivity

Typical sensitivities for IMx HAVAB, HAVAB RIA, and HAVAB EIA deter- mined using a sensitivity panel calibrated against the World Health Organization Standard were 20,150,150 U/l, respectively.

The sensitivity of IMx HAVAB was further evaluated by testing twofold serial dilutions of 8 different anti-HAV specimens with widely varying titers from 50 to 3200. The IMx assay gives two to four fold better sensitivity for each of the eight specimens compared to EIA and RIA. When these data are plotted as S/CO vs. dilution factor (Fig. l), all eight specimens gave parallel dilution curves demonstrat- ing that end point titers can give a semi-quantitative measure of anti-HAV concentra- tion in a specimen.

Reproducibility

The IMx HAVAB assay reproducibility was determined using a panel of seven specimens containing various concentrations of anti-HAV. The panel was run at three sites in triplicate in five separate runs (Table 1). The within-run, between-run, and total variation ranged from 2.74.2%, 2.7-7.0%, and 3.8-8.2%, respectively. Current

2.2

2.0 -

1.8 -

1.6 -

1.4 - z

a? 1.2 - -1

:

ii 1.0 -

-2

z

E -3

cz 0.6 - -4

-5

-6

-7

_-t-6

10 100 1000 10000 100000

Dilution Factor

Fig. 1. Reactivity of eight anti-HAV positive specimens in IMx HAVAB as a function of specimen dilution factor. Two-fold serial dilutions ofeach positive specimen were made by dilution into human plasma testing

negative for anti-HAV.

259

TABLE 1

Reproducibility and sensitivity of IMx HAVAB

Sample Concentration n IMx HAVAB HAVAB (n=l8) U/l Mean Result %CV” Mean Result %CV

s/co Within- Between- Total s/co

run run

1 276.0 45 0.086 + 4.1 1.0 8.2 0.674 + 19.7 2 138.0 45 0.155 + 3.2 3.4 4.1 0.941 k 12.9 3 69.0 45 0.339 + 4.2 4.1 5.8 1.255 - 16.3 4 34.5 45 0.704 + 3.6 3.1 5.2 1.520 - 21.2 5 17.3 45 1.123 - 4.2 2.8 5.1 1.665 - 8.7 6 8.6 45 1.464 - 2.7 2.1 3.9 1.726 - 13.3 7 4.3 45 1.697 - 2.1 2.8 3.9 1.866 - 15.3

Negative control 13 2.009 - NA 3.8 3.8 1.931 - 11.3 Positive contra? 15 0.383 + NA 5.9 5.9 0.069 + 35.5

NA=not applicable. “Percent coefficient of variation. bPositive Controls of IMx HAVAB and HAVAB have different anti-HAV concentrations.

anti-HAV assays, HAVAB RIA and EIA, typically have variation in the range of 15%.

SpeciJi:city

Specimens from two hospitals and a plasmapheresis center in the Chicago area (n = 660) were tested by IMx HAVAB and compared to HAVAB. Agreement between IMx and RIA was 99.5% (657/660). 369/660 (55.9%) of specimens were reactive for anti- HAV by IMx.

Frequency histograms show good separation between anti-HAV negative and posi- tive populations for both assays (Fig. 2). However, a tighter distribution of negative specimens around the negative control was observed in the IMx assay (x = 2.14 + 0.15) compared to HAVAB (x = 2.46 f 0.36).

Two discrepant specimens were observed between IMx and RIA. Both were low level reactive by IMx with S/CO values of 0.973 and 0.818, but borderline negative by RIA with S/CO values of 1.07 1 and 1.248, respectively. These discrepants probably result because of the better sensitivity of the IMx assay.

Interfering substances

Sera from patients with a variety of infectious diseases or containing other serologi- cal components which are potential interferences and prescreened to be unreactive for anti-HAV by EIA were tested in the IMx HAVAB assay. The specimens (n=185) were drawn from dialysis patients (n=25), and patients with elevated alkaline phos- phatase (n=8), elevated bilirubin (n=6), elevated liver enzyme (n=9), HBsAg (n=54), other viral diseases (n=18), elevated IgG, IgM or IgE (n=28), and autoimmune dis-

260

Sample/Cutoff

Fig. 2. Frequency histogram of 660 plasma and sera from hospital patients and plasmapheresis donors

tested in parallel by IMx HAVAB and HAVAB RIA. In both assays, specimens giving S/CO less than or

equal to 1 .O are considered reactive.

eases (n=37). All specimens were unreactive by IMx HAVAB.

Clinical evaluation

Three clinical laboratories (St. Louis, MO: San Bemadino, CA, and Bronx, NY) tested 1377 specimens in parallel with IMx HAVAB and EIA or IMx HAVAB and RIA (Table 2). The following samples were included in the study: 559 volunteer blood donors, 300 diagnostic or hospital patients, 60 dialysis patients, 150 patients with liver diseases other than hepatitis B, 40 clinically diagnosed non-A, non-B patients, 60 HBsAg-positive specimens and 200 specimens which had previously tested positive for total or IgM anti-HAV. Results are shown in Table 2. Overall agree- ment between the IMx and current assays was 99.9% (1376/l 377).

The one discordant specimen was strongly positive by IMx (S/CO = 0.04) but bord- erline negative by EIA (S/CO = 1.061).

The overall prevalence of anti-HAV was 5 1.5% (709/l 377) but varied dramatically depending on specimen type and geographic origin. Anti-HAV prevalence rates ran- ged from 12.3% for volunteer blood donors from St. Louis to 64.3% for the hospital population in New York City. The blood donor population in San Bernadino and the hospital and plasmapheresis patients in Chicago had anti-HAV prevalence rates of 4 1.1 and 55.9%, respectively. Although the age distribution of the patients and donors

261

TABLE 2

Comparison of IMx HAVAB and HAVAB in testing clinical specimens

Specimen category Number Reactives Percent Number of tested IMx HAVAB HAVAB agreement discordants

Volunteer blood donor Hospital patients HAVAB reactive HAVAB-M reactive HBsAg positive Dialysis patients Non-A, non-B hepatitis Patients with diseases other than HBV Total

559 155 (27.7) I54 (27.6) 99.8 I 300 193 (64.3) 193 (64.3) 100.0 0 150 150(100) 150(100) 100.0 0 58 55 (94.8) 55 (94.8) 100.0 0 60 33 (55.0) 33 (55.0) 100.0 0 60 35 (58.3) 35 (58.3) 100.0 0 40 IO (25.0) IO (25.0) 100.0 0

372 90 (24.2) 90 (24.2) 100.0 0

1377 709 (51.5) 708 (5 I .4) 99.9 I

tested is unknown, these data suggesting a higher incidence of anti-HAV in Chicago and New York City may be attributed to the varied ethnic population and lower socio- economic status of the individuals served by the institution where specimens were drawn.

To determine the ability of IMx HAVAB to detect both IgG and IgM anti-HAV, a specimen testing reactive for IgM anti-HAV was subjected to rate zonal centrifuga- tion to separate IgM from IgG. Peak fractions of IgM and IgG anti-HAV activity were tested by EIA and IMx. IMx HAVAB was equivalent or more sensitive than HAVAB EIA in detection of both fractionated IgM and IgG anti-HAV.

Serial bleedsfrom acute hepatitis A patients

To compare the onset and duration of total and IgM subclass anti-HAV, serial bleeds from six intravenous drug abusers who became infected with HAV were tested by IMx HAVAB and IMx HAVAB-M. Six patients were tested for up to 8 months fol- lowing entry into the study. At each time, endpoint titers were determined to estimate total anti-HAV levels. All six patients were strongly reactive for total and IgM anti- HAV at the first bleed obtained after presentation of symptoms to the health clinic. Within 6 months after presentation of symptoms, all IgM anti-HAV levels had dimini- shed to nonreactive or gray zone reactive. In contrast, levels of total anti-HAV measu- red by IMx HAVAB continued to rise during the first 3 to 6 months but subsequently showed some decrease by 8 months after presentation of symptoms. All 6 patients remained strongly reactive for total anti-HAV (titers > l/1000) at the end of the study. A representative profile for one of these patients is shown in Fig. 3.

Discussion

We demonstrate in these studies that a fully automated microparticle EIA for the

262

--p_ IMx HAVAEM

--c IMxHAVAB

0.6 -

0 2 4 6 6 10 12 14 16 16 20 22 24 26 26 30 32 34 36 36 40

Weeks After Presentation of Symptoms

Fig. 3. Sequential bleeds of intravenous drug abuser who became infected with HAV. Each bleed was tested for total anti-HAV by IMx HAVAB, and for IgM anti-HAV by IMx HAVAB-M. Dashed lines indi- cate range for gray zone reactivity (index value between 1.2 and 0.8) for IgM anti-HAV. Index Values

greater than or equal to 1.2 are considered reactive for IgM anti-HAV.

detection of anti-HAV, IMx HAVAB, gives better sensitivity and equivalent speci- ficity to current methods. Sera and plasma specimens from patients with a variety of infectious diseases, random blood donors and hospital patients, and individuals hav- ing serological components which are potential interfering substances were tested by the IMx assay along with the commercial EIA and RIA. In-house studies demon- strated 99.7% (742/744) agreement between IMx and RIA or EIA. Overall agreement in clinical studies was 99.9% (1376/1377) between IMx and RIA or EIA. All discor- dants were reactive by IMx but borderline negative by current tests and result from the increased sensitivity of the IMx test.

The reproducibility of the IMx HAVAB assay was also significantly improved over EIA and RIA. The coefficient of variation for a cutoff specimen was less than 10% for IMx HAVAB compared to 15% observed for HAVAB (RIA or EIA). This good reproducibility in the IMx results from precise volume, temperature and assay timing control which is possible only in a fully automated assay.

In our studies of 6 intravenous drug abusers who contracted hepatitis A, we observed rising levels of total anti-HAV during the first 3-6 months after presentation of symptoms. This contrasted with IgM anti-HAV levels which dropped significantly from reactive to nonreactive levels during this time. Thereafter some decay in total anti-HAV titers was observed, however total anti-HAV levels remained high (titers > l/l 000) in all 6 patients after 8 months. These data were consistent with those reported

263

previously (Decker et al., 198 1). Several groups have reported the development of hepatitis A vaccines (Flehmig et

al., 1988; Provost et al., 1988; Sjorgen et al., 1988) and the IMx HAVAB should be useful for monitoring immune response generated to these vaccines.

Acknowledgements

We thank Laurie Metz for typing this manuscript and Adrian Spronk, Scott Detwiler, and Tracy Wilson for their expert technical assistance.

References

Dawson, G.J., Decker, R.H., Norton, D.K., Bryce, W.H., Whittington, R.O., Tribby, 1.1. and Mushahwar, I.K. (1984) Monoclonal antibodies to hepatitis A virus. J. Med. Virol. 14, l-8.

Decker, R.H., Kosakowski, SM., Vanderbilt, A.S., Ling, C.M., Chariez, R. and Overby, L.R. (1981) Diag- nosis of acute hepatitis A by HAVAB-M, a direct radioimmunoassay for IgM anti-HAV. Am. J. Clin. Pathol. 76,140-147.

Fiore, M., Mitchell, J., Doan, T., Nelson, R., Winter, G., Grandore, C., Zeng, K., Haraden, R., Smith, J., Harris, K., Leszynsk, J., Berry, D., Safford, S., Barnes, G., Scholnick, A. and Ludington, K. (1988) The Abbott IMx automated benchtop immunochemistry analyzer system. Clin. Chem. 34,1726-1732.

Flehmig, B., Haage, A., Heirricy, U. and Pfisterer, M. (1988) Studies with an inactivated hepatitis A vac- cine. In: A.J. Zuckerman (Ed), Viral Hepatitis and Liver Disease, pp. 100-105.

Gust, I. (1988) Prevention and control of hepatitis A. In: A.J. Zuckerman (Ed), Viral Hepatitis and Liver Diseases, pp. 77-80.

Krouwer, J,S. and Rabinowitz, R. (1984) How to improve estimates of imprecision. Clin. Chem. 30,290- 292.

Lemon, S.M. and Binn, L.N. (1983) Serum Neutralizing Antibody Response to Hepatitis A Virus. J. Infect. Dis. 148,1033-1039.

Provost, P.J., Emini, E.A., Lewis, J.A. andGerety, R.J. (1988) Progress toward developing the development of a hepatitis A vaccine. In: A.J. Zuckerman (Ed), Viral Hepatitis and Liver Disease, pp. 83-86.

Quash, G., Roth, A., Nivekau, A., Grange, J., Keolouanghahot, T. and Huppert, J. (1978) The preparation of latex particles with covalently bound polyamines, IgG, and measles, agglutinins and their use in viral agglutination tests. J. Immunol. Methods 22,165-174.

Recommendation of the Immunization Practice Advisory Committee (ACIP). Protection Against Viral Hepatitis. MMWR, 39/No. S-2, February 9,199O.

Safford, S.E., Needleman, S.B. and Decker, R.H. (1980) Radioimmunoassay for the detected of antibody to hepatitis A virus. Am. J. Clin. Pathol. 74,25-3 1.

Sjorgen,M.H.,Eckels,K.H.,Binn,L.N.,Dubois,D.R.,Hoke,C.H.,Burke,D.S.andBancroft, W.H. (1988) Safety and immunogenicity of an inactivated hepatitis A vaccine. In: A.J. Zuckerman (Ed), Viral Hepat- itis and Liver Disease, pp. 9496.

Tassioiykis, N.C., Roumeliotiu-Karayannis, A., Sakka, M., Ticehurst, J., Mihalik, K., Stephanou, T., Pur- cell, R.H. and Papaevangelou, G. (1987) An Epidemic of Hepatitis A in an Institution for Young Children. Am. J. Epidemiol. 125,302-307.

Yang, N.Y., Yu, P.H., Mao, Z.X., Chen, N.L., Chai, S.A. and Mao, J.S. (1988) Inapparent Infection of Hepa- titis A Virus. Am. J. Epidemiol. 127,599-604.

Yoshitake, S., Yamada, Y., Ishikawa, E. and Masseyeff, R. (1979) Conjugation of glucose oxidase from Aspergillis niger and rabbit antibodies using N-hydroxysuccinimide ester of N-(4-Carboxycyclohexyl- methyl) maleimide. Eur. J. Biochem. 101,395-399.