AIDS RESEARCH AND HUMAN RETROVIRUSES Volume 6, Number 11, 1990 Mary Ann Liebert, Inc., Publishers
Characterization of Serum Antibody Responses to Recombinant HIV-1 gpl60 Vaccine by Enzyme Immunoassay
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
RAPHAEL VISCIDI,1'2 EDWARD ELLERBECK,3 LESLIE GARRISON,3 KAREN MIDTHUN,2'3 MARY LOU CLEMENTS,2'3 BARBARA CLAYMAN,1 BRUCE FERME,4 GALE SMITH,5 and the NIAID AIDS VACCINE CLINICAL TRIALS NETWORK6
ABSTRACT An enzyme immunoassay (EIA) was developed to measure serum antibody responses of healthy adult volunteers vaccinated with 40 or 80 (xg of human immunodeficiency virus type 1 (HIV-1) recombinant gpl60 (rgpl60) vaccine at 0,1, 6, and 18 months. This assay, which used purified rgpl60 as antigen, was compared with the Biotech/Du Pont HIV-1 Western blot and the Abbott HIV-1 EIA. Of 33 volunteers who received three doses of rgpl60 vaccine, seroresponses were detected in 91% by rgpl60 EIA, 97% by Western blot, and 30% by HIV-1 EIA. The level of IgG rgpl60 EIA antibody (mainly IgG,) peaked after the third immunization; 64% of 33 vaccinées still had detectable antibody by 12 months. The fourth immunization induced anamnestic IgG EIA antibody in 23 of 24 vaccinées, with titers ranging from 1:200 to 1:25,600. Neutralizing antibody was not detected in postvaccination sera by microtiter syncytium formation inhibition assay. Additional testing of sera by EIA indicated that the immune response to the vaccine was directed toward epitopes on both gpl20 and gp41. Seroresponses to the immunodominant epitopes on gp41 were infrequent and none were detected to the neutralization epitope in the V3 region of gpl20. This highly sensitive EIA is useful for characterizing HIV-1-specifîc antibody responses induced by an HIV-1 gpl60 subunit vaccine.
INTRODUCTION
A nodeficiency vaccine, type (HIV-1) glycoprotein 160, rgpl60 (Vaxsyn® HIV-1,
of a human immucandidate recombinant MicroGeneSys, Inc., West Haven, CT), was conducted with safety and IMMUNOGENICITY trial
virus
1
healthy HIV-1 seronegative adult volunteers. ' Among 33 vol-
unteers immunized three times with the 40 or 80 p.g doses of rgp 160 vaccine, seroresponses were detected in 97% by Western blot (Biotech/Du Pont, Du Pont Co., Wilmington, DE), but in only 30% by an HIV-1 enzyme immunoassay (EIA) (Abbott Laboratories, North Chicago, IL).
The lack of sensitivity of the commercial EIA test and the need to quantitate the humoral immune responses induced by the rgp 160 vaccine prompted us to develop an EIA using highly purified rgp 160 as antigen. We were then able to fully characterize the class-specific and subclass-specific immunoglobulin responses and the magnitude and duration of IgG EIA antibody elicited by the vaccine. We also compared the titer of serum IgG rgp 160 antibody measured by EIA and the titer of neutralizing antibody in HIV-1 infected patients. Finally, to characterize the epitope specificity of the immune response to the vaccine, we measured IgG responses to synthetic and recombinant peptides from the HIV-1 envelope glycoprotein.
Departments of 'Pediatrics and 2Medicine, School of Medicine, and 3Center for Immunization Research, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD 21205. 4Division of Molecular Virology and Immunology, Georgetown University, Rockville, MD. 5MicroGeneSys, Inc., West Haven, CT. Principal collaborators from the NIAID Vaccine Evaluation Units include Stephen Greenberg, M.D., Baylor College of Medicine, Houston, TX; Robert Belshe, M.D., St. Louis University School of Medicine, St. Louis, MO; Carol Tacket, M.D., University of Maryland School of Medicine, Baltimore, MD; Raphael Dolin, M.D., University of Rochester School of Medicine and Dentistry, Rochester, NY; Barney Graham, M.D., Vanderbilt University School of Medicine, Nashville, TN; and Donald Stablein, Ph.D., the EMMES Corporation, Potomac, MD. 1251
1252
VISCIDI ET AL.
MATERIALS AND METHODS
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
Vaccine trial The Phase I safety and immunogenicity trial of the rgp 160 vaccine candidate (VaxSyn HIV-1 ) was carried out at six NIAID AIDS Vaccine Evaluation Units: Baylor College of Medicine, The Johns Hopkins Medical Institutions, Marshall University School of Medicine, University of Maryland School of Medicine, University of Rochester School of Medicine, and Vanderbilt Uni versity School of Medicine. Prior to initiation of the trial, study protocols were approved by the respective institutional review boards at each institution, and the study was conducted according to the human experimentation guidelines of the U.S. Department of Health and Human Services. Altogether, 72 healthy HIV-1 seronegative adult volunteers between the ages of 18 and 60 years, who gave written, informed consent were randomly assigned to one of four groups to receive an intramuscular injection in the deltoid muscle of 40 p,g rgpl60, 80 p,g rgpl60; or one of two control preparations: recombinant hepatitis B surface antigen vaccine (10 p-g Recombivax HB®, Merck Sharp & Dohme Laboratories, Rahway, NJ) or placebo (alum adjuvant alone) at 0,1, and 6 months. Approximately 18 months post vaccination, 24 rgpl60 vaccinées received a fourth dose of 80 pg (n 10) or 40 p,g (n = 14) rgp 160 vaccine. =
Sera Sera were collected from study participants before inoculation and at 0.5,1,1.5,2,4,6,6.5,7,9,12,18,18.5, and 19 months afterward. Coded serum specimens from 53 participants were tested by rgp 160 EIA at The Johns Hopkins University. Twenty individuals received hepatitis B virus vaccine or placebo. Of 33 subjects who received at least three doses of the rgp 160 vaccine, 24 received a fourth dose. Preinoculation sera from these 53 volunteers and serial serum specimens from the 20 hepatitis B vaccine or placebo recipients served as negative control sera. Single serum specimens from 20 HIV-1-infected patients with neutralizing antibody to HTLV-IIIB were also tested by rgp 160 EIA.
Commercial HIV-1 assays Serum samples from volunteers
were tested at one of six vaccine evaluation units and the Central Immunology Laboratory (Georgetown University) by HIV-1 EIA (Abbott Laboratories), and Biotech/Du Pont Western blot (Du Pont Co.) kits.2 Because of the high rate of false-positives with the Abbott HIV-1 EIA, a specimen was considered positive if the absorbance value detected by both laboratories was above the cutoff levels specified by the manufacturer.3 A serum specimen was considered to be positive for HIV-1 gp 160 antibody by Western blot if one or more bands of +/-, +, or ++ intensity corresponding to the molecular weights of gpl60, gpl20, or gp41 were detected by either laboratory.
Neutralization assay Twofold serial dilutions of postvaccination sera from rgp 160 vaccinées were tested for neutralizing antibody to HTLV-IIIB or
lymphadenopathy-associated virus (LAV) strain at the Central Immunology Laboratory by the microtiter syncytium formation inhibition assay with MOLT 3 human T-cell lines (Vujcic LK, Bethesda, MD) as described previously.4The endpoint titer was
determined by a 50% reduction in the number of syncytia relative to the mean number of syncytia per field in wells containing virus adsorbed with serum from an HIV-1 seronegative subject. Sera of 20 HIV-1-infected individuals were tested by the same assay at the Division of Virology, Food and Drug Administration.
rgpl60 enzyme immunoassay A purified preparation (2= 99% pure) of the rgp 160 used in the vaccine trials derived from the envelope of HIV-1 (LAV strain) was used as antigen in the EIA. The vaccine immunogen was expressed in lepidopteran cells by a recombinant baculovims vector containing the gene encoding for gpl60. The rgp 160 from cell culture and purified by protein was extracted chromatography. ' The rgp 160 (molecular mass —160 kD) contained less than 1% of baculovirus proteins. Wells of polyvinyl chloride microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with 50 p.1 of rgp 160 (1 u,g/ml) in 10 mM sodium phosphate, pH 7.2 and 150 mM NaCl (PBS) and incubated overnight at 4°C. The plates were washed six times with phosphate-buffered solution (PBS) containing 0.05% Tween-20 (PBST) before and after each incubation step. Volumes of 50 u.1 of serum samples and reagents were added and incubated in the following order: (1) serum samples diluted 1:100 in PBST containing 3% bovine serum albumin (BSA), 1 h at 37°C; (2) horseradish peroxidase-labeled goat antibody to 7-chain-specific human IgG (Zymed Laboratories Inc., South San Francisco, CA), diluted in PBST-3% BSA, 30 min at 37°C; (3) 2, 2'-azino-di-[3-ethylbenzthiazoline sulfonate (6)] (ABTS substrate) (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD), 30 min at room temperature. Serial specimens from each subject were tested in duplicate wells on the same plate. The absorbance of each well at 405 nm was read in a Titertek Multiskan MCC/340 microtiter plate colorimeter (Flow Laboratories, McLean, VA). A specimen was considered to be positive for rgp 160 antibody if the net absorbance value was greater than the mean absorbance value of the negative control samples plus three standard deviations. To measure other class-specific antibodies, the EIA was modified by using horseradish peroxidase-labeled goat antirabbit IgA or IgM conjugate. For measurement of subclassspecific antibody, isotype(IgG,_4)-specific monoclonal antibodies (Sigma Chemical Co., St. Louis, MO) and horseradish peroxidase-labeled antimouse IgG were used in place of the goat antihuman IgG conjugate. The HIV-1 specificity of rgp 160 EIA antibody responses was confirmed in a modified EIA in which the rgp 160 antigen was bound indirectly to the solid phase. Wells of a microtiter plate were coated overnight with a monoclonal antibody to the HIV-1 envelope glycoprotein (Monoclone 110-1, Genetics System Inc., Seattle, WA). Control wells were coated with normal mouse serum. Before use, the plate was washed six times with PBST. An aliquot of rgpl60 (1 p,g/ml) in PBST-3% BSA was added to the wells and the plate was incubated for 1 h at 37°C. After washing the plate six times with PBST, serum samples
EIA ANTIBODY RESPONSES TO HIV-1
gp!60 VACCINE
diluted 1:100 in PBST-3% BSA were added to duplicate test and control wells. The plate was incubated for 1 h at 37°C. The remaining steps of the assay were performed as described above. A net absorbance value was calculated by subtracting the mean absorbance value of wells coated with normal mouse serum from the mean absorbance value of wells coated with the monoclonal antibody. A specimen was considered to be positive for rgp 160 antibody if the net absorbance value was greater than the mean net absorbance value of the negative control samples, plus three standard deviations.
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
Peptide enzyme immunoassays To measure antibody responses to different regions of gpl60, wells of microtiter plates were coated with the following synthetic or recombinant peptides: (1) H2N-NNTRKSIRIQRGPGRAFVCTIGKIGC-OH (5 |xg/ml) in PBS (RP135);5 (2) H2NIWGCSGKLICTTAVPWNAS-OH (5 |xg/ml); and H2NAVERYLKDQQLLGIWGCSGKLI-OH (5 p,g/ml) in PBS (E32/E34);6 (3) recombinant gp41 (amino acids 542 to 633, numbering system of Ratner et al.)7 expressed in Escherichia coli (2 p.g/ml) in 1 M acetic acid (gift of Hardy Chan, Syntex, Inc.); (4) recombinant gpl20 expressed in yeast (1 u-g/ml) in PBS (gift of Vince Perna, SmithKline Biosystems Lab.). Peptides RP135, E32, and E34 were synthesized by an Applied Biosystems 430A peptide synthesizer using chemical and program cycles supplied by the manufacturer. Purity of the peptides was confirmed by high-performance liquid chromatography (HPLC) analysis. The recombinant gp41 peptide purified from E. coli extracts by HPLC contained less than 5% of E. coli protein. The EIAs were performed as described above except serum samples were diluted 1:10 in the assays using the synthetic peptides and recombinant gp41 peptide.
Comparison of liters of rgpl60 EIA antibody in vaccinées and HIV-1-infected patients Serial twofold dilutions of serum samples from 32 rgp 160 vaccinées who mounted seroresponses and from 20 HIV-1infected patients with neutralizing antibody (titers of 1:16 to 1:512) to HTLV-IIIB were tested by EIA to compare the level of IgG rgp 160 antibody induced by the HIV-1 vaccine with the level of IgG rgp 160 antibody detected in HIV-1-infected individuals with neutralizing antibody. Serum samples selected for testing were those that gave the highest absorbance readings by rgp 160 EIA after the third and fourth immunizations. The endpoint titer was defined as the final serum dilution which gave an absorbance value greater than the mean absorbance value of the negative control sera plus one standard deviation. The negative control sera was run at a 1:100 dilution.
Statistical
analysis
Data were entered into a DBase III software program (Ashtonat each of the six participating sites and transmitted electronically to the central data management and
Tate, Torrance, CA)
statistical analysis unit (EMMES Corporation, Potomac, MD). The data were analyzed by Mann Whitney U nonparametric,
1253
Spearman rank correlation, appropriate.
and Pearson correlation tests
as
RESULTS
Seroresponses to rgp 160
vaccination
Serial serum specimens from 53 volunteers receiving three doses of vaccine (33 rgp 160 vaccinées and 20 hepatitis B and placebo control recipients) were tested in a blinded fashion by rgpl60 EIA, Western blot, and Abbott EIA. Among the 33 rgp 160 vaccinées, seroresponses were detected in, after the third vaccination, 30 (91%) by rgpl60 EIA, 32 (97%) by Western blot, but in only 10 (30%) by Abbott EIA. Antibody to HIV-1 envelope proteins was detected in one postvaccination serum sample from the 20 controls by rgpl60 EIA, 2 by Abbott EIA, and 5 by Western blot. None of the positive Western blots or Abbott EIAs in control subjects were confirmed when tested in a second laboratory, but one control had antibody to gp41 on two separate occasions by Western blot. There was no clinical or laboratory evidence (including polymerase chain reaction) to suggest that any of the control subjects who had HIV-1 antibody detected by these tests were infected with HIV-1. The rgp 160 EIA and Western blot gave concordant results in measuring seroresponses of 44 (83%) of the 53 study participants, whereas rgp 160 and Abbott EIAs only agreed in 31 (58%) of the 53
subjects.
A false positive reaction by rgp 160 El A was detected in one of 33 rgp 160 vaccinées prior to vaccination. This individual mounted a fourfold increase in absorbance values after the fourth vaccination, suggesting that a significant rise in antibody to rgp 160 was detected even though the level of nonspecific background absorbance was high. To determine whether the rgp 160 EIA antibody response was directed against HIV-1 gpl60 (but not the baculovirus residue in the vaccine preparation), serum yielding a peak rgp 160 EIA antibody level after the third immunization was tested by EIA using rgp 160 antigen bound indirectly to the solid phase by a monoclonal antibody to the HIV-1 envelope glycoprotein. A positive response to rgp 160 was demonstrated in serum specimens from the 30 rgp 160 vaccine responders and in one serum specimen from a control vaccinée who had a positive rgp 160 EIA. No rgp 160 antibody could be detected by this assay in the serum sample that gave a false positive reaction by rgp 160 EIA prior to vaccination.
Kinetics and vaccine
magnitude of seroresponses to rgp 160
Since the frequency of IgG rgp 160 EIA antibody responses to 40 |xg or 80 |xg doses of vaccine was similar after each immunization, we pooled the results of the two dosage groups for analysis. An IgG rgp 160 EIA antibody response was detected in one of 33 volunteers after the first vaccination, in 24 (73%) of 33 after the second, in 30 (91%) of 33 after the third, and in 23 (97%) of 24 after the fourth vaccination. The level of IgG antibody rose slightly after the second vaccination, peaked after the third vaccination, and then declined rapidly between 7 and 12 months post vaccination. By 12 months, 21 (64%) of the 33 vaccinated volunteers had detectable IgG rgp 160 EIA anti-
VISCIDI ET AL.
1254
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
body in their sera. The fourth immunization given at 18 months elicited an anamnestic rgp 160 El A antibody response in all but one of 24 vaccinées. The pattern of antibody responses ofthe 24 volunteers immunized four times is shown in Figure 1. After each booster vaccination, mean antibody levels were higher in the 80 p,g dose recipients than in the 40 p.g dose recipients. However, this difference was not statistically significant. The predominant antibodies detected in all 30 rgp 160 vaccine responders were of the IgG class and IgG, subclass. The course of IgG 1 antibody response paralleled the IgG antibody response. We also detected antibody of the IgG3 subclass in sera of 3 of the 30 vaccine responders after the third vaccination, but we were unable to detect antibody of the IgG2 or IgG4 subclasses or IgM or Ig A classes in sera from any of the vaccinées.
Table 1. Antibody Responses to HIV-1 env Peptides 24 Volunteers After 4 Doses of rgp 160 Vaccine
in
Dose(n) 40 Mg (14) 80 Mg (10)
=
0.830 and0.793 for rgpl20 andrgp41, respectively,
.001, Spearman rank correlation test).
Correlation
of rgpl60 EIA antibody and neutralizing
antibody titers
The rgp 160 EIA detected rgp 160 antibody titers ranging from 1:100 to 1:6,400 in vaccinées after the third immunization and titers of 1:200 to 1:25,600 after the fourth immunization. Only 3 of the 33 vaccinées mounted a titer greater than 1:3,200. None of the vaccinées developed neutralizing antibody to HTLV-IIIB. In contrast, the EIA detected an IgG rgp 160 titer greater than 1:3,200 in sera from 19 of 20 HIV-1-infected patients who had HIV-1 neutralizing antibody (Fig. 2). There was a weak corre-
__. S
9
Q
1
§
» 90 ,ig I 40 pg
0.50
0 20
=
=
To determine the frequency, magnitude, and duration of antibody responses in serum of HIV-1 seronegative adult volunteers induced by a recombinant HIV-1 gpl60 vaccine preparation expressed in a baculovirus vector system, we developed a solid phase EIA using HIV-1 rgp 160 vaccine as antigen. Our
results indicate that the rgp 160 EIA, like the Biotech/Du Pont Western blot kit, was highly sensitive for detection of seroresponses to rgpl60 immunization. The concordance between the determination of seroconversion by Western blot and rgp 160 EIA for the rgp 160 vaccinées and HIV-1 seronegative controls was 83%. The small discrepancy between the results of these tests is due, in part, to the detection of Western blot bands corresponding to the molecular weights of the envelope glycoproteins (gp41, gp 120, or gp 160) in 4 of 20 HIV-1 seronegative volunteers. Studies carried out in our laboratories8 and in others9-10 indicate that commercial Western blot kits often detect indeterminate bands in sera of persons not infected with HIV-1. Moreover, lot to lot variability may account for the
rQp160
rgp160 E S
0.40
W
0.30
29 70
64 70
DISCUSSION
In an attempt to characterize the epitope specificity of the seroresponse following vaccination with rgp 160 vaccine, sera from volunteers who received four doses of vaccine were tested in EIAs using peptides from different regions of the HIV-1 envelope glycoprotein. The frequency of antibody responses to peptide-defined epitopes was higher in vaccinées receiving 80 pg doses of vaccine than 40 p,g doses (Table 1). Seroresponses to rgpl20, rgp41 and E32/E34 were detected in 67%, 46%, and 8% of the vaccinées, respectively. No vaccinées demonstrated a seroresponse to RP135. There was a strong correlation between levels of antibody to rgp 160 and levels of antibody to rgp 120 and p
Characterization of Serum Antibody Responses to Recombinant HIV-1 gpl60 Vaccine by Enzyme Immunoassay
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
RAPHAEL VISCIDI,1'2 EDWARD ELLERBECK,3 LESLIE GARRISON,3 KAREN MIDTHUN,2'3 MARY LOU CLEMENTS,2'3 BARBARA CLAYMAN,1 BRUCE FERME,4 GALE SMITH,5 and the NIAID AIDS VACCINE CLINICAL TRIALS NETWORK6
ABSTRACT An enzyme immunoassay (EIA) was developed to measure serum antibody responses of healthy adult volunteers vaccinated with 40 or 80 (xg of human immunodeficiency virus type 1 (HIV-1) recombinant gpl60 (rgpl60) vaccine at 0,1, 6, and 18 months. This assay, which used purified rgpl60 as antigen, was compared with the Biotech/Du Pont HIV-1 Western blot and the Abbott HIV-1 EIA. Of 33 volunteers who received three doses of rgpl60 vaccine, seroresponses were detected in 91% by rgpl60 EIA, 97% by Western blot, and 30% by HIV-1 EIA. The level of IgG rgpl60 EIA antibody (mainly IgG,) peaked after the third immunization; 64% of 33 vaccinées still had detectable antibody by 12 months. The fourth immunization induced anamnestic IgG EIA antibody in 23 of 24 vaccinées, with titers ranging from 1:200 to 1:25,600. Neutralizing antibody was not detected in postvaccination sera by microtiter syncytium formation inhibition assay. Additional testing of sera by EIA indicated that the immune response to the vaccine was directed toward epitopes on both gpl20 and gp41. Seroresponses to the immunodominant epitopes on gp41 were infrequent and none were detected to the neutralization epitope in the V3 region of gpl20. This highly sensitive EIA is useful for characterizing HIV-1-specifîc antibody responses induced by an HIV-1 gpl60 subunit vaccine.
INTRODUCTION
A nodeficiency vaccine, type (HIV-1) glycoprotein 160, rgpl60 (Vaxsyn® HIV-1,
of a human immucandidate recombinant MicroGeneSys, Inc., West Haven, CT), was conducted with safety and IMMUNOGENICITY trial
virus
1
healthy HIV-1 seronegative adult volunteers. ' Among 33 vol-
unteers immunized three times with the 40 or 80 p.g doses of rgp 160 vaccine, seroresponses were detected in 97% by Western blot (Biotech/Du Pont, Du Pont Co., Wilmington, DE), but in only 30% by an HIV-1 enzyme immunoassay (EIA) (Abbott Laboratories, North Chicago, IL).
The lack of sensitivity of the commercial EIA test and the need to quantitate the humoral immune responses induced by the rgp 160 vaccine prompted us to develop an EIA using highly purified rgp 160 as antigen. We were then able to fully characterize the class-specific and subclass-specific immunoglobulin responses and the magnitude and duration of IgG EIA antibody elicited by the vaccine. We also compared the titer of serum IgG rgp 160 antibody measured by EIA and the titer of neutralizing antibody in HIV-1 infected patients. Finally, to characterize the epitope specificity of the immune response to the vaccine, we measured IgG responses to synthetic and recombinant peptides from the HIV-1 envelope glycoprotein.
Departments of 'Pediatrics and 2Medicine, School of Medicine, and 3Center for Immunization Research, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, MD 21205. 4Division of Molecular Virology and Immunology, Georgetown University, Rockville, MD. 5MicroGeneSys, Inc., West Haven, CT. Principal collaborators from the NIAID Vaccine Evaluation Units include Stephen Greenberg, M.D., Baylor College of Medicine, Houston, TX; Robert Belshe, M.D., St. Louis University School of Medicine, St. Louis, MO; Carol Tacket, M.D., University of Maryland School of Medicine, Baltimore, MD; Raphael Dolin, M.D., University of Rochester School of Medicine and Dentistry, Rochester, NY; Barney Graham, M.D., Vanderbilt University School of Medicine, Nashville, TN; and Donald Stablein, Ph.D., the EMMES Corporation, Potomac, MD. 1251
1252
VISCIDI ET AL.
MATERIALS AND METHODS
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
Vaccine trial The Phase I safety and immunogenicity trial of the rgp 160 vaccine candidate (VaxSyn HIV-1 ) was carried out at six NIAID AIDS Vaccine Evaluation Units: Baylor College of Medicine, The Johns Hopkins Medical Institutions, Marshall University School of Medicine, University of Maryland School of Medicine, University of Rochester School of Medicine, and Vanderbilt Uni versity School of Medicine. Prior to initiation of the trial, study protocols were approved by the respective institutional review boards at each institution, and the study was conducted according to the human experimentation guidelines of the U.S. Department of Health and Human Services. Altogether, 72 healthy HIV-1 seronegative adult volunteers between the ages of 18 and 60 years, who gave written, informed consent were randomly assigned to one of four groups to receive an intramuscular injection in the deltoid muscle of 40 p,g rgpl60, 80 p,g rgpl60; or one of two control preparations: recombinant hepatitis B surface antigen vaccine (10 p-g Recombivax HB®, Merck Sharp & Dohme Laboratories, Rahway, NJ) or placebo (alum adjuvant alone) at 0,1, and 6 months. Approximately 18 months post vaccination, 24 rgpl60 vaccinées received a fourth dose of 80 pg (n 10) or 40 p,g (n = 14) rgp 160 vaccine. =
Sera Sera were collected from study participants before inoculation and at 0.5,1,1.5,2,4,6,6.5,7,9,12,18,18.5, and 19 months afterward. Coded serum specimens from 53 participants were tested by rgp 160 EIA at The Johns Hopkins University. Twenty individuals received hepatitis B virus vaccine or placebo. Of 33 subjects who received at least three doses of the rgp 160 vaccine, 24 received a fourth dose. Preinoculation sera from these 53 volunteers and serial serum specimens from the 20 hepatitis B vaccine or placebo recipients served as negative control sera. Single serum specimens from 20 HIV-1-infected patients with neutralizing antibody to HTLV-IIIB were also tested by rgp 160 EIA.
Commercial HIV-1 assays Serum samples from volunteers
were tested at one of six vaccine evaluation units and the Central Immunology Laboratory (Georgetown University) by HIV-1 EIA (Abbott Laboratories), and Biotech/Du Pont Western blot (Du Pont Co.) kits.2 Because of the high rate of false-positives with the Abbott HIV-1 EIA, a specimen was considered positive if the absorbance value detected by both laboratories was above the cutoff levels specified by the manufacturer.3 A serum specimen was considered to be positive for HIV-1 gp 160 antibody by Western blot if one or more bands of +/-, +, or ++ intensity corresponding to the molecular weights of gpl60, gpl20, or gp41 were detected by either laboratory.
Neutralization assay Twofold serial dilutions of postvaccination sera from rgp 160 vaccinées were tested for neutralizing antibody to HTLV-IIIB or
lymphadenopathy-associated virus (LAV) strain at the Central Immunology Laboratory by the microtiter syncytium formation inhibition assay with MOLT 3 human T-cell lines (Vujcic LK, Bethesda, MD) as described previously.4The endpoint titer was
determined by a 50% reduction in the number of syncytia relative to the mean number of syncytia per field in wells containing virus adsorbed with serum from an HIV-1 seronegative subject. Sera of 20 HIV-1-infected individuals were tested by the same assay at the Division of Virology, Food and Drug Administration.
rgpl60 enzyme immunoassay A purified preparation (2= 99% pure) of the rgp 160 used in the vaccine trials derived from the envelope of HIV-1 (LAV strain) was used as antigen in the EIA. The vaccine immunogen was expressed in lepidopteran cells by a recombinant baculovims vector containing the gene encoding for gpl60. The rgp 160 from cell culture and purified by protein was extracted chromatography. ' The rgp 160 (molecular mass —160 kD) contained less than 1% of baculovirus proteins. Wells of polyvinyl chloride microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with 50 p.1 of rgp 160 (1 u,g/ml) in 10 mM sodium phosphate, pH 7.2 and 150 mM NaCl (PBS) and incubated overnight at 4°C. The plates were washed six times with phosphate-buffered solution (PBS) containing 0.05% Tween-20 (PBST) before and after each incubation step. Volumes of 50 u.1 of serum samples and reagents were added and incubated in the following order: (1) serum samples diluted 1:100 in PBST containing 3% bovine serum albumin (BSA), 1 h at 37°C; (2) horseradish peroxidase-labeled goat antibody to 7-chain-specific human IgG (Zymed Laboratories Inc., South San Francisco, CA), diluted in PBST-3% BSA, 30 min at 37°C; (3) 2, 2'-azino-di-[3-ethylbenzthiazoline sulfonate (6)] (ABTS substrate) (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD), 30 min at room temperature. Serial specimens from each subject were tested in duplicate wells on the same plate. The absorbance of each well at 405 nm was read in a Titertek Multiskan MCC/340 microtiter plate colorimeter (Flow Laboratories, McLean, VA). A specimen was considered to be positive for rgp 160 antibody if the net absorbance value was greater than the mean absorbance value of the negative control samples plus three standard deviations. To measure other class-specific antibodies, the EIA was modified by using horseradish peroxidase-labeled goat antirabbit IgA or IgM conjugate. For measurement of subclassspecific antibody, isotype(IgG,_4)-specific monoclonal antibodies (Sigma Chemical Co., St. Louis, MO) and horseradish peroxidase-labeled antimouse IgG were used in place of the goat antihuman IgG conjugate. The HIV-1 specificity of rgp 160 EIA antibody responses was confirmed in a modified EIA in which the rgp 160 antigen was bound indirectly to the solid phase. Wells of a microtiter plate were coated overnight with a monoclonal antibody to the HIV-1 envelope glycoprotein (Monoclone 110-1, Genetics System Inc., Seattle, WA). Control wells were coated with normal mouse serum. Before use, the plate was washed six times with PBST. An aliquot of rgpl60 (1 p,g/ml) in PBST-3% BSA was added to the wells and the plate was incubated for 1 h at 37°C. After washing the plate six times with PBST, serum samples
EIA ANTIBODY RESPONSES TO HIV-1
gp!60 VACCINE
diluted 1:100 in PBST-3% BSA were added to duplicate test and control wells. The plate was incubated for 1 h at 37°C. The remaining steps of the assay were performed as described above. A net absorbance value was calculated by subtracting the mean absorbance value of wells coated with normal mouse serum from the mean absorbance value of wells coated with the monoclonal antibody. A specimen was considered to be positive for rgp 160 antibody if the net absorbance value was greater than the mean net absorbance value of the negative control samples, plus three standard deviations.
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
Peptide enzyme immunoassays To measure antibody responses to different regions of gpl60, wells of microtiter plates were coated with the following synthetic or recombinant peptides: (1) H2N-NNTRKSIRIQRGPGRAFVCTIGKIGC-OH (5 |xg/ml) in PBS (RP135);5 (2) H2NIWGCSGKLICTTAVPWNAS-OH (5 |xg/ml); and H2NAVERYLKDQQLLGIWGCSGKLI-OH (5 p,g/ml) in PBS (E32/E34);6 (3) recombinant gp41 (amino acids 542 to 633, numbering system of Ratner et al.)7 expressed in Escherichia coli (2 p.g/ml) in 1 M acetic acid (gift of Hardy Chan, Syntex, Inc.); (4) recombinant gpl20 expressed in yeast (1 u-g/ml) in PBS (gift of Vince Perna, SmithKline Biosystems Lab.). Peptides RP135, E32, and E34 were synthesized by an Applied Biosystems 430A peptide synthesizer using chemical and program cycles supplied by the manufacturer. Purity of the peptides was confirmed by high-performance liquid chromatography (HPLC) analysis. The recombinant gp41 peptide purified from E. coli extracts by HPLC contained less than 5% of E. coli protein. The EIAs were performed as described above except serum samples were diluted 1:10 in the assays using the synthetic peptides and recombinant gp41 peptide.
Comparison of liters of rgpl60 EIA antibody in vaccinées and HIV-1-infected patients Serial twofold dilutions of serum samples from 32 rgp 160 vaccinées who mounted seroresponses and from 20 HIV-1infected patients with neutralizing antibody (titers of 1:16 to 1:512) to HTLV-IIIB were tested by EIA to compare the level of IgG rgp 160 antibody induced by the HIV-1 vaccine with the level of IgG rgp 160 antibody detected in HIV-1-infected individuals with neutralizing antibody. Serum samples selected for testing were those that gave the highest absorbance readings by rgp 160 EIA after the third and fourth immunizations. The endpoint titer was defined as the final serum dilution which gave an absorbance value greater than the mean absorbance value of the negative control sera plus one standard deviation. The negative control sera was run at a 1:100 dilution.
Statistical
analysis
Data were entered into a DBase III software program (Ashtonat each of the six participating sites and transmitted electronically to the central data management and
Tate, Torrance, CA)
statistical analysis unit (EMMES Corporation, Potomac, MD). The data were analyzed by Mann Whitney U nonparametric,
1253
Spearman rank correlation, appropriate.
and Pearson correlation tests
as
RESULTS
Seroresponses to rgp 160
vaccination
Serial serum specimens from 53 volunteers receiving three doses of vaccine (33 rgp 160 vaccinées and 20 hepatitis B and placebo control recipients) were tested in a blinded fashion by rgpl60 EIA, Western blot, and Abbott EIA. Among the 33 rgp 160 vaccinées, seroresponses were detected in, after the third vaccination, 30 (91%) by rgpl60 EIA, 32 (97%) by Western blot, but in only 10 (30%) by Abbott EIA. Antibody to HIV-1 envelope proteins was detected in one postvaccination serum sample from the 20 controls by rgpl60 EIA, 2 by Abbott EIA, and 5 by Western blot. None of the positive Western blots or Abbott EIAs in control subjects were confirmed when tested in a second laboratory, but one control had antibody to gp41 on two separate occasions by Western blot. There was no clinical or laboratory evidence (including polymerase chain reaction) to suggest that any of the control subjects who had HIV-1 antibody detected by these tests were infected with HIV-1. The rgp 160 EIA and Western blot gave concordant results in measuring seroresponses of 44 (83%) of the 53 study participants, whereas rgp 160 and Abbott EIAs only agreed in 31 (58%) of the 53
subjects.
A false positive reaction by rgp 160 El A was detected in one of 33 rgp 160 vaccinées prior to vaccination. This individual mounted a fourfold increase in absorbance values after the fourth vaccination, suggesting that a significant rise in antibody to rgp 160 was detected even though the level of nonspecific background absorbance was high. To determine whether the rgp 160 EIA antibody response was directed against HIV-1 gpl60 (but not the baculovirus residue in the vaccine preparation), serum yielding a peak rgp 160 EIA antibody level after the third immunization was tested by EIA using rgp 160 antigen bound indirectly to the solid phase by a monoclonal antibody to the HIV-1 envelope glycoprotein. A positive response to rgp 160 was demonstrated in serum specimens from the 30 rgp 160 vaccine responders and in one serum specimen from a control vaccinée who had a positive rgp 160 EIA. No rgp 160 antibody could be detected by this assay in the serum sample that gave a false positive reaction by rgp 160 EIA prior to vaccination.
Kinetics and vaccine
magnitude of seroresponses to rgp 160
Since the frequency of IgG rgp 160 EIA antibody responses to 40 |xg or 80 |xg doses of vaccine was similar after each immunization, we pooled the results of the two dosage groups for analysis. An IgG rgp 160 EIA antibody response was detected in one of 33 volunteers after the first vaccination, in 24 (73%) of 33 after the second, in 30 (91%) of 33 after the third, and in 23 (97%) of 24 after the fourth vaccination. The level of IgG antibody rose slightly after the second vaccination, peaked after the third vaccination, and then declined rapidly between 7 and 12 months post vaccination. By 12 months, 21 (64%) of the 33 vaccinated volunteers had detectable IgG rgp 160 EIA anti-
VISCIDI ET AL.
1254
Downloaded by MIDWESTERN UNIVERSITY MULTISITE from www.liebertpub.com at 01/24/19. For personal use only.
body in their sera. The fourth immunization given at 18 months elicited an anamnestic rgp 160 El A antibody response in all but one of 24 vaccinées. The pattern of antibody responses ofthe 24 volunteers immunized four times is shown in Figure 1. After each booster vaccination, mean antibody levels were higher in the 80 p,g dose recipients than in the 40 p.g dose recipients. However, this difference was not statistically significant. The predominant antibodies detected in all 30 rgp 160 vaccine responders were of the IgG class and IgG, subclass. The course of IgG 1 antibody response paralleled the IgG antibody response. We also detected antibody of the IgG3 subclass in sera of 3 of the 30 vaccine responders after the third vaccination, but we were unable to detect antibody of the IgG2 or IgG4 subclasses or IgM or Ig A classes in sera from any of the vaccinées.
Table 1. Antibody Responses to HIV-1 env Peptides 24 Volunteers After 4 Doses of rgp 160 Vaccine
in
Dose(n) 40 Mg (14) 80 Mg (10)
=
0.830 and0.793 for rgpl20 andrgp41, respectively,
.001, Spearman rank correlation test).
Correlation
of rgpl60 EIA antibody and neutralizing
antibody titers
The rgp 160 EIA detected rgp 160 antibody titers ranging from 1:100 to 1:6,400 in vaccinées after the third immunization and titers of 1:200 to 1:25,600 after the fourth immunization. Only 3 of the 33 vaccinées mounted a titer greater than 1:3,200. None of the vaccinées developed neutralizing antibody to HTLV-IIIB. In contrast, the EIA detected an IgG rgp 160 titer greater than 1:3,200 in sera from 19 of 20 HIV-1-infected patients who had HIV-1 neutralizing antibody (Fig. 2). There was a weak corre-
__. S
9
Q
1
§
» 90 ,ig I 40 pg
0.50
0 20
=
=
To determine the frequency, magnitude, and duration of antibody responses in serum of HIV-1 seronegative adult volunteers induced by a recombinant HIV-1 gpl60 vaccine preparation expressed in a baculovirus vector system, we developed a solid phase EIA using HIV-1 rgp 160 vaccine as antigen. Our
results indicate that the rgp 160 EIA, like the Biotech/Du Pont Western blot kit, was highly sensitive for detection of seroresponses to rgpl60 immunization. The concordance between the determination of seroconversion by Western blot and rgp 160 EIA for the rgp 160 vaccinées and HIV-1 seronegative controls was 83%. The small discrepancy between the results of these tests is due, in part, to the detection of Western blot bands corresponding to the molecular weights of the envelope glycoproteins (gp41, gp 120, or gp 160) in 4 of 20 HIV-1 seronegative volunteers. Studies carried out in our laboratories8 and in others9-10 indicate that commercial Western blot kits often detect indeterminate bands in sera of persons not infected with HIV-1. Moreover, lot to lot variability may account for the
rQp160
rgp160 E S
0.40
W
0.30
29 70
64 70
DISCUSSION
In an attempt to characterize the epitope specificity of the seroresponse following vaccination with rgp 160 vaccine, sera from volunteers who received four doses of vaccine were tested in EIAs using peptides from different regions of the HIV-1 envelope glycoprotein. The frequency of antibody responses to peptide-defined epitopes was higher in vaccinées receiving 80 pg doses of vaccine than 40 p,g doses (Table 1). Seroresponses to rgpl20, rgp41 and E32/E34 were detected in 67%, 46%, and 8% of the vaccinées, respectively. No vaccinées demonstrated a seroresponse to RP135. There was a strong correlation between levels of antibody to rgp 160 and levels of antibody to rgp 120 and p
