704
Maternal Antibodies to gp120 V3 Sequence Do Not Correlate with Protection against Vertical Transmission of Human Immunodeficiency Virus C. A. Robertson, J. Y. Q. Mok, K. S. Froebel, P. Simmonds, S. M. Burns, H. S. Marsden, and S. Graham
Medical Research Council Virology Unit. University of Glasgow; Regional Infectious Diseases Unit. City Hospital. and HIV Immunology Unit. University Department of Medicine. Royal Infirmary, and Department of Medical Microbiology. University of Edinburgh. Edinburgh. United Kingdom
A significant proportion ofchildren born to human immunodeficiency virus (HIV)-infected mothers are infected in utero. Transmission rates vary between 7%and 65% [1-5]. It has been reported that mothers who have high-affinity antibodies directed against the principal neutralizing determinant, contained within the third hypervariable domain (V3) of gp 120, are less likely to transmit HIV-I to their children [6-8]. In contrast, however, a more recent study [9] found no correlation between levels of maternal antibodies against the V3 loop and maternal-child transmission. The resolution of these apparently contrasting results is of importance for prognostic serodiagnosis and vaccine development. Such correlation would be helpful in predicting the clinical outcome of at-risk pregnancies and would lend support to vaccine strategies aimed at increasing antibody levels to the V3 region. The V3 region ofgpl20 comprises 34 amino acid residues that form a loop bounded by two invariant cysteines, amino acids 303-338 in the env protein of the BH 10 clone of the human T celllymphotropic virus type III isolate [10]. Many experiments indicate that antibodies to this region are immunologically important. Synthetic peptides from the V3 loop sequence elicit type-specific neutralizing antibodies [11-13].
Received 12 March 1992; revised 19 May 1992. Financial support: Medical Research Council (SPG 8710211 to P.L.Yap and J.Y.Q.M., SPG 8812779 to H.S.M., studentship to CA.R.) and Darwin Trust of Edinburgh (K.S.F.). Reprints or correspondence: Dr. Susan Graham. MRC Virology Unit. University of Glasgow. Church St., Glasgow Gil 5JR, Scotland, United Kingdom. The Journal of Infectious Diseases 1992;166:704-9 © 1992 by The University of Chicago. All rights reserved. 0022-1899/92/6604-0002$01.00
Within this region, a 9 amino acid sequence containing the highly conserved tetrameric GPGR motif, predicted to form a J3-turn structure, is the binding site ofthe type-specific neutralizing antibodies [10, 14, 15]. Monoclonal antibodies directed against this principal neutralizing determinant prevent HIV-I infection of MT-4 cells in vitro, while anti-V3 antibodies in polyclonal IgG inhibit infection of chimpanzees in vivo [16]. Most recently, a mouse-human chimeric monoclonal antibody specific for the V3 loop domain and having potent virus-neutralizing activity could protect chimpanzees against infection with HIV-I when administered either before or after challenge [ I7], Chimpanzees successfully protected from infection by HIV-I had antibodies to the principal neutralizing determinant [18, 19]. In addition, mutations in the V3 sequence can give rise to neutralization escape mutants [20, 21]. To further examine the relationship between maternal antibodies to the V3 sequence and the risk of vertical transmission, the reactivities of antibodies in sera from two groups of HIV-I-infected mothers, transmitters and nontransmitters, were analyzed. Our study differs from earlier studies in three aspects. First, antibody titers were determined as opposed to classifying maternal sera as positive or negative on the basis of whether a single dilution of maternal serum gave an ELISA titer above a cutoff value. Second, a long peptide representing the whole of the V3 sequence between the conserved cysteines was used, as it might be expected to contain conformational as well as sequential epitopes. Third, we investigated whether differences in antibody levels to the V3 sequence reflected differences in overall antibody levels or were peculiar for those against V3. To do this, the antibody levels to a peptide corresponding to an immunodominant and conserved domain of gp41 were also measured. In this
Downloaded from http://jid.oxfordjournals.org/ at University of Otago on July 13, 2015
A retrospective study of sera from mothers infected with human immunodeficiency virus (HIV -I) was undertaken to investigate whether the titers or affinities of antibodies against the third hypervariable region (V3 loop) of gpl20 correlated with transmission of the virus from mother to child. The cohort comprised 7 mothers who transmitted HIV-I to their children and 20 who did not. Sera were screened for reactivity against two synthetic peptides, one encompassing the entire V3100p of gpl20 (amino acids 297-330) and the other containing an immunodominant epitope from gp41 (amino acids 596-614). Doubling dilutions of sera were tested to obtain antibody titers against both peptides: Anti-gp41 titers were used to normalize the anti-V3 titers. Maternal sera were also screened for the presence of high-affinity antibodies against the V3 peptide. No differences were observed in either titers or affinities of maternal antibodies to the V3 sequence from transmitters and nontransmitters.
JID 1992; 166 (October)
V3 Antibodies and HIV-l Transmission
way, the maternal antibody levels against the variable V3 loop could be normalized to the antibody levels against the constant gp41 region.
Materials and Methods
brought to a final concentration of 0.2 JLg/mL with PBS. Microtiter plates (Immunlon 2; Dynatech, Billinghurst, UK) were coated with 50 JLL of branched peptide solutions (at 0.2 JLg/mL) to give lOng/well or various amounts of monomeric peptide as indicated in the text. The peptides were allowed to adsorb overnight at 37°C and then were washed five times with PBS-Tween 20. The plates were then blocked with a solution of 2%bovine serum albumin (BSA) in PBS for I h at 37°C. To determine antibody titers, human sera were diluted serially with 0.5%BSAPBS, and 50 JLL of the appropriate dilutions were added to the wells and incubated for I h at 37°C. For the "antigen-limited V3 ELISA" [8], the human sera were diluted I: 100 and incubated for I h at 37°C in the antigencoated wells. The plates were washed five times with PBSTween 20 before incubating for I hat 3rC with 50 JLL/well of horseradish peroxidase-conjugated sheep anti-human IgG (Scottish Antibody Production Unit, Carluke, UK) diluted I: 1000 in 0.5% BSA-PBS. Plates were then washed seven times in PBS-Tween 20 and reacted with 100 JLL of a 50 mg/mL solution of enzyme substrate 2,2'-azino-bis(3-ethylbenzthiazoline-c-sulfonic acid) in citrate phosphate buffer (pH 4.0) containing 0.01% hydrogen peroxide. After 15-30 min of color development, the plates were read on a Multiskan plate reader (Titertek: ICN Biomedicals, High Wycombe, UK) at 405 nm. The results are means of duplicate determinations. Statistical analysis. The groups were compared using the nonparametric Mann-Whitney test [32].
Results Maternal antibody ELISA titers. In an initial study, different amounts of branched V3 peptide and gp41 peptide were tested for reactivity by ELISA with a panel of sera from HIV-I-positive mothers and a panel of 14 seronegative controls. An aliquot of l O ng/well gave maximum signals with positive sera and low backgrounds with negative sera (00 = 0.040 ± 0.026 and 0.066 ± 0.030 for the V3 and gp41 peptide, respectively). Sera were coded and screened blind. The reactivities of maternal sera, serially twofold diluted, from 6 transmitters and 6 nontransmitters are shown in figure 1. High titers of V3 antibody were detected in both groups of mothers regardless of whether their children were infected (1 A) or not (1C). Sera from both transmitters (I B) and nontransmitters (1 D) also reacted with the gp41 peptide although the maximum optical densities were lower than with the V3 peptide. The data are summarized in table I; the titer ofeach serum sample is expressed as the reciprocal of the maximum dilution that would give an absorbance of 0.4 00 units. Also tabulated is the ratio of the titers against the V3 and gp41 peptides for each sample. The nonparametric Mann-Whitney test showed no significant difference between the median values for anti-VI titers (P = .82) or the anti-V3/antigp41 titers (P = .21). A similar analysis was also done for anti-V3 titers giving an absorbance of 1.0 00 unit. Again, no significant differences between the median values for anti-
Downloaded from http://jid.oxfordjournals.org/ at University of Otago on July 13, 2015
Synthesis of oligopeptides. gpl20 V3 loop sequence TRPNNNTRKRIHIGPGRAFYTTGQIIGDIRQAH is essentially that described by laRosa et al. [15]. This peptide represents residues 302-335 of strain MN gp 120 and was used because the sequence corresponds most closely to that of isolates from the Edinburgh area [22] (unpublished observations). The sequence GCSGKLICTTAVPWNAS corresponds to residues 596-614, the immunodominant domain, of HIV-I gp41 [2325]. These oligopeptides were made as branched (octameric) or multiple-antigenic peptides [26, 27], which have advantages over monomeric peptides in serodiagnosis: Lower amounts of peptides can be used and lower amounts of antibodies can be detected [28, 29]. The gp41 peptide was made with five additional glycine residues on each branch, (GCSGKLICTTAVPWNASG s)sK7A, as this can increase sensitivity in ELISAs [29]. The V3 peptide was also made in monomeric form for the antibody-affinity assays. The oligopeptides were synthesized by continuous flow Fmoc chemistry as previously described [30, 31]. After synthesis, the peptides were cleaved from the resin and side chain-protecting groups were removed by standard protocols. The amino acid compositions of the branched peptides were determined by amino acid analysis (Cambridge Research Biochemicals, Cambridge, UK) and were in agreement with expectations. The relative molecular mass of the monomeric V3 peptide was determined by fast atom bombardment mass spectrometry (M-Scan, Ascot, OK), which gave the expected value. Studv cohort. All infants born to Hl V-infected women in the Lothia~ region of Scotland were enrolled in a prospective study to examine the risk of mother-child transmission of HIV as well as to evaluate the natural history of vertically acquired HIV disease [2]. All women were clinically well during pregnancy, with the exception of 2 who had Pneumocystis carinii pneumonia at the time of delivery. After a median follow-up of 48 months (range, 4-73), infants were diagnosed as HIV-infected on the basis oftwo ofthe following criteria: persistence of HIV antibody beyond 18 months of age, positive HIV culture or HIV antigen tests on more than two occasions, and clinical evidence of HIV infection. Children were presumed uninfected when they were >2 years old, remained clinically healthy, had normal immune function tests, and tested negative for HIV antibody, antigen, and culture. Mothers were divided into two groups: Group I (n = 7) were those who had given birth to infected children (HIV transmitters); group 2 (n = 20) were those who delivered uninfected children (HIV nontransmitters). Maternal sera were collected during pregnancy or soon after delivery. Peptide ELISA. The V3 peptides were dissolved in PBSat an initial concentration of I mg/ml, and then diluted with PBS to give the required concentration for coating the wells (see Results). The gp41 peptide, which was insoluble in PBS, was dissolved in 100 JLL of 30% acetic acid at I mg/mL and then
705
Robertson et at.
706
2.0 . . . - - - - - - - - - - - - - - - - ,
JID 1992; 166 (October)
2.0 . . . . - - - - - - - - - - - - - - . ,
B
1.5
1.0
1.0
0.5
0.5
0.0
0.0
Ir)
It)
C'oI
0
"""
~ ~
~
~
U
s::::
0 0
0 0
""" """
"""
~
0 0 CD
""" p
0 0
~
CD
0 0 CD
"""
It)
C'oI
"""
0 0
""" """
p
~
0 0 CD
0 0 CD
0 0
~
~
""" """
It)
C'oI
p
"""
2.0
C
'€ 0
0 0
p
2.0
~
It)
C'oI
D
00
.,.0
Maternal Antibodies to gp120 V3 Sequence Do Not Correlate with Protection against Vertical Transmission of Human Immunodeficiency Virus C. A. Robertson, J. Y. Q. Mok, K. S. Froebel, P. Simmonds, S. M. Burns, H. S. Marsden, and S. Graham
Medical Research Council Virology Unit. University of Glasgow; Regional Infectious Diseases Unit. City Hospital. and HIV Immunology Unit. University Department of Medicine. Royal Infirmary, and Department of Medical Microbiology. University of Edinburgh. Edinburgh. United Kingdom
A significant proportion ofchildren born to human immunodeficiency virus (HIV)-infected mothers are infected in utero. Transmission rates vary between 7%and 65% [1-5]. It has been reported that mothers who have high-affinity antibodies directed against the principal neutralizing determinant, contained within the third hypervariable domain (V3) of gp 120, are less likely to transmit HIV-I to their children [6-8]. In contrast, however, a more recent study [9] found no correlation between levels of maternal antibodies against the V3 loop and maternal-child transmission. The resolution of these apparently contrasting results is of importance for prognostic serodiagnosis and vaccine development. Such correlation would be helpful in predicting the clinical outcome of at-risk pregnancies and would lend support to vaccine strategies aimed at increasing antibody levels to the V3 region. The V3 region ofgpl20 comprises 34 amino acid residues that form a loop bounded by two invariant cysteines, amino acids 303-338 in the env protein of the BH 10 clone of the human T celllymphotropic virus type III isolate [10]. Many experiments indicate that antibodies to this region are immunologically important. Synthetic peptides from the V3 loop sequence elicit type-specific neutralizing antibodies [11-13].
Received 12 March 1992; revised 19 May 1992. Financial support: Medical Research Council (SPG 8710211 to P.L.Yap and J.Y.Q.M., SPG 8812779 to H.S.M., studentship to CA.R.) and Darwin Trust of Edinburgh (K.S.F.). Reprints or correspondence: Dr. Susan Graham. MRC Virology Unit. University of Glasgow. Church St., Glasgow Gil 5JR, Scotland, United Kingdom. The Journal of Infectious Diseases 1992;166:704-9 © 1992 by The University of Chicago. All rights reserved. 0022-1899/92/6604-0002$01.00
Within this region, a 9 amino acid sequence containing the highly conserved tetrameric GPGR motif, predicted to form a J3-turn structure, is the binding site ofthe type-specific neutralizing antibodies [10, 14, 15]. Monoclonal antibodies directed against this principal neutralizing determinant prevent HIV-I infection of MT-4 cells in vitro, while anti-V3 antibodies in polyclonal IgG inhibit infection of chimpanzees in vivo [16]. Most recently, a mouse-human chimeric monoclonal antibody specific for the V3 loop domain and having potent virus-neutralizing activity could protect chimpanzees against infection with HIV-I when administered either before or after challenge [ I7], Chimpanzees successfully protected from infection by HIV-I had antibodies to the principal neutralizing determinant [18, 19]. In addition, mutations in the V3 sequence can give rise to neutralization escape mutants [20, 21]. To further examine the relationship between maternal antibodies to the V3 sequence and the risk of vertical transmission, the reactivities of antibodies in sera from two groups of HIV-I-infected mothers, transmitters and nontransmitters, were analyzed. Our study differs from earlier studies in three aspects. First, antibody titers were determined as opposed to classifying maternal sera as positive or negative on the basis of whether a single dilution of maternal serum gave an ELISA titer above a cutoff value. Second, a long peptide representing the whole of the V3 sequence between the conserved cysteines was used, as it might be expected to contain conformational as well as sequential epitopes. Third, we investigated whether differences in antibody levels to the V3 sequence reflected differences in overall antibody levels or were peculiar for those against V3. To do this, the antibody levels to a peptide corresponding to an immunodominant and conserved domain of gp41 were also measured. In this
Downloaded from http://jid.oxfordjournals.org/ at University of Otago on July 13, 2015
A retrospective study of sera from mothers infected with human immunodeficiency virus (HIV -I) was undertaken to investigate whether the titers or affinities of antibodies against the third hypervariable region (V3 loop) of gpl20 correlated with transmission of the virus from mother to child. The cohort comprised 7 mothers who transmitted HIV-I to their children and 20 who did not. Sera were screened for reactivity against two synthetic peptides, one encompassing the entire V3100p of gpl20 (amino acids 297-330) and the other containing an immunodominant epitope from gp41 (amino acids 596-614). Doubling dilutions of sera were tested to obtain antibody titers against both peptides: Anti-gp41 titers were used to normalize the anti-V3 titers. Maternal sera were also screened for the presence of high-affinity antibodies against the V3 peptide. No differences were observed in either titers or affinities of maternal antibodies to the V3 sequence from transmitters and nontransmitters.
JID 1992; 166 (October)
V3 Antibodies and HIV-l Transmission
way, the maternal antibody levels against the variable V3 loop could be normalized to the antibody levels against the constant gp41 region.
Materials and Methods
brought to a final concentration of 0.2 JLg/mL with PBS. Microtiter plates (Immunlon 2; Dynatech, Billinghurst, UK) were coated with 50 JLL of branched peptide solutions (at 0.2 JLg/mL) to give lOng/well or various amounts of monomeric peptide as indicated in the text. The peptides were allowed to adsorb overnight at 37°C and then were washed five times with PBS-Tween 20. The plates were then blocked with a solution of 2%bovine serum albumin (BSA) in PBS for I h at 37°C. To determine antibody titers, human sera were diluted serially with 0.5%BSAPBS, and 50 JLL of the appropriate dilutions were added to the wells and incubated for I h at 37°C. For the "antigen-limited V3 ELISA" [8], the human sera were diluted I: 100 and incubated for I h at 37°C in the antigencoated wells. The plates were washed five times with PBSTween 20 before incubating for I hat 3rC with 50 JLL/well of horseradish peroxidase-conjugated sheep anti-human IgG (Scottish Antibody Production Unit, Carluke, UK) diluted I: 1000 in 0.5% BSA-PBS. Plates were then washed seven times in PBS-Tween 20 and reacted with 100 JLL of a 50 mg/mL solution of enzyme substrate 2,2'-azino-bis(3-ethylbenzthiazoline-c-sulfonic acid) in citrate phosphate buffer (pH 4.0) containing 0.01% hydrogen peroxide. After 15-30 min of color development, the plates were read on a Multiskan plate reader (Titertek: ICN Biomedicals, High Wycombe, UK) at 405 nm. The results are means of duplicate determinations. Statistical analysis. The groups were compared using the nonparametric Mann-Whitney test [32].
Results Maternal antibody ELISA titers. In an initial study, different amounts of branched V3 peptide and gp41 peptide were tested for reactivity by ELISA with a panel of sera from HIV-I-positive mothers and a panel of 14 seronegative controls. An aliquot of l O ng/well gave maximum signals with positive sera and low backgrounds with negative sera (00 = 0.040 ± 0.026 and 0.066 ± 0.030 for the V3 and gp41 peptide, respectively). Sera were coded and screened blind. The reactivities of maternal sera, serially twofold diluted, from 6 transmitters and 6 nontransmitters are shown in figure 1. High titers of V3 antibody were detected in both groups of mothers regardless of whether their children were infected (1 A) or not (1C). Sera from both transmitters (I B) and nontransmitters (1 D) also reacted with the gp41 peptide although the maximum optical densities were lower than with the V3 peptide. The data are summarized in table I; the titer ofeach serum sample is expressed as the reciprocal of the maximum dilution that would give an absorbance of 0.4 00 units. Also tabulated is the ratio of the titers against the V3 and gp41 peptides for each sample. The nonparametric Mann-Whitney test showed no significant difference between the median values for anti-VI titers (P = .82) or the anti-V3/antigp41 titers (P = .21). A similar analysis was also done for anti-V3 titers giving an absorbance of 1.0 00 unit. Again, no significant differences between the median values for anti-
Downloaded from http://jid.oxfordjournals.org/ at University of Otago on July 13, 2015
Synthesis of oligopeptides. gpl20 V3 loop sequence TRPNNNTRKRIHIGPGRAFYTTGQIIGDIRQAH is essentially that described by laRosa et al. [15]. This peptide represents residues 302-335 of strain MN gp 120 and was used because the sequence corresponds most closely to that of isolates from the Edinburgh area [22] (unpublished observations). The sequence GCSGKLICTTAVPWNAS corresponds to residues 596-614, the immunodominant domain, of HIV-I gp41 [2325]. These oligopeptides were made as branched (octameric) or multiple-antigenic peptides [26, 27], which have advantages over monomeric peptides in serodiagnosis: Lower amounts of peptides can be used and lower amounts of antibodies can be detected [28, 29]. The gp41 peptide was made with five additional glycine residues on each branch, (GCSGKLICTTAVPWNASG s)sK7A, as this can increase sensitivity in ELISAs [29]. The V3 peptide was also made in monomeric form for the antibody-affinity assays. The oligopeptides were synthesized by continuous flow Fmoc chemistry as previously described [30, 31]. After synthesis, the peptides were cleaved from the resin and side chain-protecting groups were removed by standard protocols. The amino acid compositions of the branched peptides were determined by amino acid analysis (Cambridge Research Biochemicals, Cambridge, UK) and were in agreement with expectations. The relative molecular mass of the monomeric V3 peptide was determined by fast atom bombardment mass spectrometry (M-Scan, Ascot, OK), which gave the expected value. Studv cohort. All infants born to Hl V-infected women in the Lothia~ region of Scotland were enrolled in a prospective study to examine the risk of mother-child transmission of HIV as well as to evaluate the natural history of vertically acquired HIV disease [2]. All women were clinically well during pregnancy, with the exception of 2 who had Pneumocystis carinii pneumonia at the time of delivery. After a median follow-up of 48 months (range, 4-73), infants were diagnosed as HIV-infected on the basis oftwo ofthe following criteria: persistence of HIV antibody beyond 18 months of age, positive HIV culture or HIV antigen tests on more than two occasions, and clinical evidence of HIV infection. Children were presumed uninfected when they were >2 years old, remained clinically healthy, had normal immune function tests, and tested negative for HIV antibody, antigen, and culture. Mothers were divided into two groups: Group I (n = 7) were those who had given birth to infected children (HIV transmitters); group 2 (n = 20) were those who delivered uninfected children (HIV nontransmitters). Maternal sera were collected during pregnancy or soon after delivery. Peptide ELISA. The V3 peptides were dissolved in PBSat an initial concentration of I mg/ml, and then diluted with PBS to give the required concentration for coating the wells (see Results). The gp41 peptide, which was insoluble in PBS, was dissolved in 100 JLL of 30% acetic acid at I mg/mL and then
705
Robertson et at.
706
2.0 . . . - - - - - - - - - - - - - - - - ,
JID 1992; 166 (October)
2.0 . . . . - - - - - - - - - - - - - - . ,
B
1.5
1.0
1.0
0.5
0.5
0.0
0.0
Ir)
It)
C'oI
0
"""
~ ~
~
~
U
s::::
0 0
0 0
""" """
"""
~
0 0 CD
""" p
0 0
~
CD
0 0 CD
"""
It)
C'oI
"""
0 0
""" """
p
~
0 0 CD
0 0 CD
0 0
~
~
""" """
It)
C'oI
p
"""
2.0
C
'€ 0
0 0
p
2.0
~
It)
C'oI
D
00
.,.0
