Comparison of enzyme immunoassay and reverse transcriptase assay for detection of HIV in culture supernates

Journal of Virological Methods, 17 (1987) 237-245 Elsevier

237

JVM 00614

Comparison transcriptase

of enzyme immunoassay and reverse assay for detection of HIV in culture supernates

D.S. Healey, J.B.M. Jowett, F. Beaton, W.J. Maskill and I.D. Gust National HIV Reference

Laboratory,

Fairfield Hospital for Communicable Victoria, Australia

(Accepted

29 April

Diseases, Fairfield,

1987)

Summary An enzyme immunoassay (EIA) was developed for detection of human immunodeficiency virus antigen (HIV Ag) in tissue culture supernatants. The assay was found to be specific for HIV and cheaper, easier to perform and more sensitive than the generally used reverse transcriptase (RT) assay. Cultures of peripheral blood leucocytes (PBL) from 106 patients with acquired immune deficiency syndrome (AIDS), AIDS related complex (ARC), healthy antiHIV positive subjects and healthy anti-HIV negative subjects were held for 35 days and the supernatant fluid tested at regular intervals by EIA and RT. Of these 106 cultures, the presence of HIV was detected by EIA in 27 and by RT in 21. While six cultures were positive by EIA alone, none were positive by RT alone; the specificity of the results in the six EIA positive RT negative cultures was confirmed by subculture. In the 21 cultures in which HIV was detected by both techniques, the EIA became positive first on 10 occasions; in the remaining cultures both tests became positive at the same time. The HIV Ag assay reduces the time taken to process specimens and thus increases the efficiency and reduces the cost of isolation procedures. HIV;

Human

Correspondence 3078. 0166-0934/87/$03.50

retrovirus;

AIDS;

IO: Dr. Ian Gust,

0

Antigen

Fairfield

1987 Elsevier

Science

capture

Hospital,

Publishers

Yarra

EIA;

RT

Bend Road,

B.V. (Biomedical

Fairfield,

Division)

Victoria,

Australia

238

Introduction

Isolation of human immunodeficiency virus (HIV) (Coffin et al., 1986) from blood and tissues obtained from infected patients has become an important tool in studying the natural history of HIV infections. While the virus can be propagated in some human and lymphoblastoid cell lines (Levy et al., 1984; Popovic et al., 1984) most workers prefer to attempt isolation in peripheral blood lymphocyte cultures (Neate et al., 1987). Because of the increasing demand for HIV isolation, the time taken to obtain results and the high cost of maintaining lymphocytes in culture, there is an urgent need to develop more sensitive means of detecting virus infection. At present, the most widely used method for detecting replication of HIV in lymphocyte culture is the reverse transcriptase (RT) assay, a test which is labour intensive and expensive, especially when used to monitor a large number of cultures. In addition. because the test detects the activity of any magnesium dependent RT, it is not specific for HIV. Recently enzyme immunoassays (EIAs) have been used to detect HIV antigens in both culture supernates and body fluids (McDougal et al., 1985; Paul and Falk, 1986; Goudsmit et al., 1986). In this paper we evaluate an EIA for detecting specific HIV antigens in culture of peripheral blood leucocytes.

Materials

and Methods

Origin of speciinens

Peripheral blood leucocytes were separated from 20 ml aliquots of heparinized venous blood from 12 patients with acquired immune deficiency syndrome (AIDS), 26 patients with AIDS-related complex (ARC), 27 healthy anti-HIV positive subjects and 26 healthy anti-HIV negative controls. The criteria used to define clinical categories were those recommended by the Australian AIDS Task Force (1985). Multiple specimens were collected from most patients with AIDS (25 specimens from 12 patients) whilst in the remaining groups usually only a single specimen was tested (28 specimens from 26 patients with ARC, 27 specimens from 27 anti-HIV positive subjects and 26 specimens from the antibody negative controls).

Virus isolation and reverse transcriptase (RT) assays were performed as previously described (Neate et al., 1987). Briefly fresh PBL from patients were purified in Ficoll Paque (Pharmacia) and resuspended at approximately 2 x 10” cells/ml in 10 ml culture medium base (RPM1 1640 with 10% heat inactivated foetal calf serum, 15 mM Hepes, 0.1% sodium bicarbonate with penicillin 100 IUiml and streptomycin 100 pgiml containing 10 pgiml phytohaemagglutinin A (PHA) (Sigma). After 3 days incubation at 37°C the medium was removed and replaced

239

with 10 ml complete culture medium (i.e. culture medium base plus 5% v/v interleukin-2 (Boehringer), 1:lOOO anti-interferon (Miles) and 2 pg/ml DEAE dextran or Polybrene (Sigma)) supplemented with approximately 2 x 10’ PHA stimulated PBL from a healthy seronegative subject. The stationary cultures were maintained at a cell concentration of l-2 x 10”iml for 35 days. Every 3 days half the supernatant fluid was harvested and replaced with fresh medium. RT assays were performed by the method of Popovic et al. (1984) with minor modifications (Neate et al., 1987). Cultures which were positive by EIA and negative by RT were passaged once and observed for a further 35 days. The results of these tests were analysed by chi square and McNeman’s tests. Detection

of HIV untigen

The method used to detect HIV antigen was a solid phase EIA based on the method of McDougal et al. (1985), performed in microtitre trays with the following modifications. Capture

antibody

Human antibodies to HIV (anti-HIV) were obtained from the serum of an adult male with lymphadenopathy syndrome (LAS). The serum produced a titre of 1:20000 by an ‘in house’ anti-HIV EIA and showed activity to all major viral proteins by Western blot. The immunglobulin G (IgG) component was separated by adsorption to a Sepharose Gl-4B protein A column (Pharmacia, Uppsala, Sweden), eluted with 0.1 M glycine-HCl buffer (pH 2.8), neutralized immediately with 1 M TRIS (pH 7.8) and stored at 4°C in 0.01% thiomerasal. Anti-HIV

conjugate

Human anti-HIV IgG was prepared as described above from a pool of sera obtained from six adult males all of whom had LAS and were reactive by anti-HIV EIA and to all major viral proteins by Western blot. Horseradish peroxidase (HRP, type VI, Sigma, St. Louis, MO, USA) was conjugated to the IgG by the method of Wilson and Nakane (1978). The conjugate was stored at -20°C in PBS (pH 7.4) containing 50% glycerol and 0.01% thiomersal. HIV antigen

capture

EIA

The optimal concentrations of capture antibody and conjugate were determined by checkerboard titration and found to be 7.0 kg/ml and 5.4 pgiml respectively when using 10 kg/ml of HIV antigen. Flat-bottom 96-well microtitre plates (Nunc, Immulon II) were coated with 200 lJJwel1 of capture antibody diluted in PBS (pH 7.4) and incubated at 37°C for 2 h. After aspiration, free protein adsorption sites were blocked by addition of 200 Fl/well of 1% casein diluted in PBS (pH 7.4) and incubated for 1 h at 37°C. Wells

240

were then washed four times with distilled water containing 0.05% Tween-20 (DWT). Supernates were prepared by addition of 1 volume 10% Triton X-100 in PBS (pH 7.4) to 9 volumes of supernate. Each supernate preparation (100 kliwell) was added to duplicate wells and incubated at 4°C overnight. After washing four times in DWT, 100 lJl/well of anti-HIV conjugate diluted in PBS (pH 7.4) containing 1% casein and 0.05% Tween-20 (CT20) was added and the plates incubated for 2 h at 37°C. After a further four washes in DWT, 100 ~1 of substrate (2 0.3% v/v H,O, in 0.05 M citrate mgiml o-phenylenediamine (OPD)) containing buffer (pH 5.6) was added to each well, and incubated at room temperature in the dark. The colour reaction was stopped after 15 min by adding 50 kl/well of 2N H$O,. Absorbance was read at 492 nm with a reference wavelength of 620 mm in a Flow Titertek Multiskan ELISA plate reader. Determination of cutoff The cutoff was defined as three standard deviations above the mean of eight negative controls (supernatant fluids from uninfected PBLs). In addition, a positive control with expected absorbance range of 1.0-1.4 was included in quadruplicate in each plate. The test was considered invalid if the mean of these positives fell outside the expected range. Specificity of the assay To determine the specificity of the HIV Ag assay supernatant fluids from cell cultures infected with a number of common viruses and other agents were tested. These included herpes simplex types 1 and 2 (HSVl, HSV2) and cytomegalovirus (CMV) grown in HEL cells, Epstein Barr virus (EBV) infected WIL lymphocytes and Vero cultures infected with Toxoplasma gondii (T. gondii). Supernates were from the following cell lines: MT-2 (Harada et al., 1978), MOLT-4 (Popovic et al., lung fibroblasts 1984), CCRF CEM (McCarthy et al., 1965)) human embryonic (HEL) (Lewis and Kennett, 1976), African green monkey epithelial cells (VERO) (MacFarlane and Sommerville, 1969), and baby hamster kidney cells (BHK21) (Stoker and Macpherson, 1964) were also tested to provide negative controls. PBL supernates from normal uninfected human peripheral blood lymphocyte cultures were assayed in every plate. In addition all specimens which were positive for HIV Ag by EIA were tested for specificity in a blocking test in which a 1:400 dilution of an unconjugated antiHIV positive serum (from a local haemophiliac with ARC) was used. This serum was selected because it was reactive for anti-HIV by EIA, containing antibodies to all major viral proteins by Western blot but did not contain antibodies to cytomegalovirus or herpes simplex virus types 1 and 2 when tested by commercial EIAs. Prior to addition of conjugated anti-HIV, 100 ~1 of unconjugated blocking serum (diluted 1:400 in CT20) was added and the mixture incubated for 1 h at 37°C. After four washes the conjugate was added and the procedure completed as previously described. All tests were performed in duplicate.

241

A 25% or greater reduction in the mean absorbance of the duplicate blocked samples (when compared with the unblocked samples) was considered evidence of the presence of HIV Ag. This figure was five times the mean reduction in absorbance obtained when the anti-HIV positive blocking serum was replaced by a panel of 40 anti-HIV negative sera obtained from healthy blood donors. Positive supernatant fluids (which produced approximately 25% and 50% reduction in absorbance) and negative control fluids (supernatants collected from uninfected PBL cultures) were included in each blocking assay.

Sensitivity of the assay The sensitivity of the HIV Ag EIA was determined by testing serial dilutions of a purified standard antigen (HIV strain HTLV III B grown in H9 cells, containing 7.12 Fgiml of viral protein). This antigen was kindly provided by Dr. Larry Arthur of the National Cancer Institute Research Facility, Frederick, Maryland, USA. A plot of absorbance versus dilution was made and the sensitivity of the assay defined as the minimum dilution of antigen which could be detected. The cut off point chosen was three standard deviations above the mean of 10 negative control wells in which 10% casein was substituted for the capture antibody.

Comparison of the RT and EIA techniques To determine the relative sensitivity of the two assays serial two-fold dilutions of HIV positive supernatant fluids (obtained from H9 cells infected with HTLV III B and CEM cells infected with LAV 1) were tested by both techniques. Each sample was tested in quadruplicate and on two separate occasions. The LAV 1 isolate propagated in CEM cells, originally from the Pasteur Institute, was kindly provided by the Commonwealth Serum Laboratories, Parkville, Melbourne, Australia.

Results In a preliminary study several samples of supernatant fluid were collected from a number of specimens which had been submitted to the laboratory for HIV isolation. Each specimen was tested in parallel and under code for RT activity and the presence of HIV Ag. Of a total of 199 specimens tested, 180 were negative and six positive by both techniques. A further 13 samples were positive for HIV Ag alone. When the code was broken it was found that the 13 samples had been obtained from six patients, all of whom were known to be infected with HIV. In each instance when these specimens were subpassaged subsequent supernatant fluids became positive by both tests. Following these encouraging results a larger study was undertaken to define the sensitivity and specificity of the antigen assay and to compare its value with RT in monitoring specimens inoculated into cell cultures.

242 TABLE

1

Antigen

capture

Clinical

category

EIA and RT assay results

No. specimens tested

AIDS (12)* ARC (26) Healthy anti-HIV Healthy anti-HIV

positive negative

(27) (26)

Total * No. subjects

for the 106 monitored HIV detected

cultures. by

Total

EIA and RT

EIA alone

25 28 27 26

17 1 3 0

1 2 3 0

18 3 6 0

106

21

6

27

in each group.

Specificity Supernatant fluids from cell cultures infected with T. gondii HSV2, Epstein Barr virus and from seven different types of uninfected cell culture, were uniformly negative in the HIV Ag assay. By contrast supernatant fluids from cell cultures infected with CMV or HSVl showed low levels of reactivity (1.9-2.6 times the cut off). These specimens were readily distinguished from true positives as they remained positive in the blocking assay. Sensitivity The minimum quantity of purified HIV Ag (HTLV IIIB) which could be detected in the EIA was 10 ng/ml. In addition when serial dilutions of RT positive supernatant fluids from H9 cell cultures infected with HTLV BIB and CEM cells infected with LAV 1 were tested in parallel by RT and EIA, the mean EIA titres were 40-fold higher (1:800 compared to 1:20). The HIV Ag assay has also been found to be of value in testing stored specimens. We have noted that the RT activity of specimens held for 6 mth or more at -70°C may decline dramatically so that many initially positive specimens are found to be RT negative on retesting. On each occasion HIV Ag remained detectable by EIA (data not shown). Comparison of the two assay systems Both assays appear to be highly specific - none of the cultures from the 26 antibody negative subjects being positive by either RT or EIA. Table 1 demonstrates the results of the two assays in specimens obtained from patients in each of the clinical categories. Of the 21 specimens in which HIV was detected by EIA and RT the EIA test became positive earlier in 10 and on the same day as the RT assay in 11. The period for which HIV Ag could be detected by EIA alone ranged from 3-22 days with a mean of 6 days (see Fig. 1). All specimens positive by EIA showed greater than 25.0% reduction in activity in the blocking EIA.

243

+RT

+

.

EIA:

”B

TIME IN CULTURE

TIME IN CULTURE

TIME IN CULTURE

(WEEKS)

(WEEKS)

(WEEKS)

.

_....____

._._...............__._..........

.___.._...

Fig. 1. Comparison of antigen capture EIA (--X-) and RT (-o-) assays for three cultures established from specimens taken at weekly intervals from an anti-HIV positive individual. The cultures were monitored by both assays at intervals of 3-4 days for 3-5 wk. Results are presented as ratios of vaiues obtained with each sample/mean of the negative control (sample/cutoff ratio).

In cultures in which the presence of HIV was detected simultaneously by EIA and RT, activity was detected earlier in the culture period (mean 11.4 days) than in those cultures in which antigen detection preceded detectable RT activity (mean 14.3 days) suggesting that these effects may be dose related. Once HIV Ag was detected in the supernatant of a culture, subsequent samples from the same culture were usually positive. By contrast RT activity tended to be transient, rising to a peak then disappearing within 3-7 days (mean 5.3 days). Although six cultures from infected individuals were initially positive by EIA alone, in each instance passage of the specimen resulted in activity detectable by both assays. Fig. 1 demonstrates the results of seriaf RT and HIV Ag assays in three cultures which were established from specimens taken at weekly intervals from an anti-HIV positive individual. For purposes of comparison results are presented as ratios of values obtained with each sample/cutoff (sample/cutoff ratio). In cultures A and B, HIV Ag was detected a few days and a few weeks prior to RT activity. In culture C both assays became detectable at the same time.

Discussion The antigen capture assay which we have established appears to be specific for HIV and more sensitive than the widely used RT assay. Specificity was demon-

244

strated by the failure of supernatant fluids from a variety of uninfected and infected cell cultures to react in the test and the absence of reactivity with PBL obtained from a panel of 26 healthy anti-HIV negative subjects. Furthermore the reactivity obtained when specimens from infected individuals were tested, could always be blocked with unlabelled sera containing anti-HIV. While these results are encouraging, large scale use of the test will require a constant supply of capture and detector antibodies which have a known titre and defined reactivities. As human antisera are likely to contain antibodies to a number of other common viruses, we intend to replace them with animal sera raised against purified virus or pools of monoclonal antibodies. The HIV antigen capture assay is clearly more sensitive than the RT assay as currently performed and detected as little as 10 ng viral protein/ml. The assay not only consistently detected HIV Ag in higher dilutions (up to 40-fold lower) than RT, but was able to identify the presence of HIV in cell culture earlier in the course of infection. The HIV Ag assay not only detected HIV in every specimen which was positive by RT but in several others in which it would only have been detected on blind passage. In cultures in which HIV was detectable by both techniques, the antigen assay reduced the time needed to issue a report by several days. In those cultures in which RT activity was not detected in the first 5 wk but only became apparent in subculture (6 of 27) the use of the antigen assay reduced the time needed to produce a report by several weeks. The HIV Ag assay has several advantages over the RT assay, particularly in ease of performance, cost and sensitivity. In addition the assay is readily automated which makes it more suitable as a screening test. Isolation of HIV from clinical specimens is important for increasing our understanding of the pathogenesis and mode of transmission of the disease as well as for resolving the diagnosis in subjects with equivocal anti-HIV serology and for evaluating the efficacy and mode of action of antiviral chemotherapeutic agents. The HIV Ag assay is likely to prove a most useful test in clinical and research laboratories, especially those handling large numbers of specimens. Acknowledgements

This study was performed by staff of the National Reference Laboratory with funds generously provided by the Commonwealth Department of Health. The authors wish to acknowledge the assistance of Barbara Gray in preparation of the manuscript.

References Australian AIRS Task Force (February 1985) In: Infection Control Guidelines, p. 2. Coffin, J., Haase, A., Levy, J.A., Montagnier, L., Oroszlan, S., Teich. N., Temin, H., Toyoshima, K.. Varmus, H., Vogt, P. and Weiss, R. (1986) Science 232, 697.

245 Goudsmit, J., Paul, D.A., Lange, J.M.A., Speelman, H., Van Der Noordaa, J., Van Der Helm, H.J., De Wolf, F., Epstein, L.G., Krone, W.J.A., Wolters, E.C., Oleske, M. and Coutinho. R.A. (1986) Lancet ii, 177-180. Harada, S., Koyanagi: Y. and Yamamoto, N. (1985) Science 229. 563-566. Levy, J.A., Hoffman, A.D., Kramer, SM., Landis, J.A. and Shimabukuro, J.M. (1984) Science 225, 840-842. Lewis, F.A. and Kennett, M.L. (1976) J. Clin. Microbial. 3. 528-532. Macfarlane, D.E. and Sommerville, R.G. (1969) Arch. Gesamte Virusforsch. 27, 379-385. McCarthy, R.E., Junius, V., Farber, S., Lazarus, H. and Foley. GE. (1965) Exp. Cell Res. 40, 197-200. McDougal. J.S., Cart, S.P., Kennedy, MS., Cabridilla, C.D., Feorino, P.M., Francis, D.P., Hicks, D., Kalyanaraman, V.S. and Martin, L.S. (1985) J. Immunol. Methods. 76. 171-183. Neate, E.V., Pringle, R.C., Jowett, J.B.M.. Healey, D.S. and Gust, I.D. (1987) In press. Aust. N.Z. J. Med. Paul, D.A. and Falk, L.A. (1986). J. Cell Biochem. Supplement lOA, 224. Popovic, M., Sarngadharan, M.G., Read, E. and Gallo. R.C. (1984) Science 224, 497-500. Stoker, M., and Macpherson, I. (1964) Nature 4952, 1355-1357. Wilson, M.B. and Nakane, P.K. (1978) In: Immunofluorescence and Related Staining Techniques (Knapp, W., Holubar, K. and Wick, G., eds.), p. 215. Elsevier, Amsterdam.