After an overnight incubation at 37°C in 5% C02, the cultures ... and supernatants were stored at -20°C. PBLs were main- ..... S. T. & Wilson, C. (1991) J. Clin.
Proc. Natl. Acad. Sci. USA Vol. 90, pp. 2340-2344, March 1993 Immunology
Ontogeny of anti-human immunodeficiency virus (HIV) antibody production in HIV-1-infected infants (perinatal infection/AIDS/immune response)
HENRY POLLACK, MING XIA ZHAN, TIINA ILMET-MOORE, KEN AJUANG-SIMBIRI, KEITH KRASINSKI, AND WILLIAM BORKOWSKY Division of Infectious Disease and Immunology, Department of Pediatrics, New York University Medical Center and Bellevue Hospital Center, New York, NY 10016
Communicated by Saul Krugman, September 22, 1992
ABSTRACT The early serologic response of infants to infection with human immunodeficiency virus type 1 (HIV-1) is normally obscured by the presence of transplacentaHly acquired maternal HIV antibody. By measuring HIV antibody produced in vitro by lymphocytes isolated from peripheral blood of infants and children of HIV-1-infected mothers, we have been able to study the natural acquisition of humoral immunity to perinatal HIV-1 infection. One hundred ninety-seven infants of HIV-1-infected women were studied prospectively and longitudinally from birth. In the neonatal period, infected infants produced only small amounts of HIV-specific IgG antibodies to a restricted number of antigens. The amount of immunoglobulin to HIV-1 and the number of HIV-1 antigens recognized increased with age. After 6 months of life 85% of infected infants made detectable antibody to two or more viral proteins. Antibody to gpl60 appeared first and was the most frequently found at all ages, followed by antibody to the envelope proteins gpl20 and gp4l. The amount of HIV antibody produced correlated positively with the percentage of CD4+ T lymphocytes in peripheral blood. This assay provides a method of studying the immunogenicity of vaccines against HIV-1 in HIV-1-infected infants and of assessing the effect of early therapeutic interventions on the humoral response to HIV-1.
children (6-11). Although most of these studies have limited themselves mainly to older children and adults, several studies have employed this technique in the detection of HIV-1 infection in infants (7-9, 11). In the present study the technique of measuring antibody production in vitro is used to study the development of the antibody repertoire to HIV-1 in a large cohort of perinatally HIV-1-infected infants followed prospectively and longitudinally during the first 2 years of life.
MATERIALS AND METHODS Human Subjects. Subjects were enrolled from the Obstetrical and the Pediatric Infectious Disease Services at Bellevue Hospital Center and New York University Medical Center, between March 1989 and November 1991, as part of an ongoing prospective study of perinatal HIV-1 infection. One hundred ninety-seven HIV-1-seropositive infants and children, 46 HIV-1-infected adults, and 115 HIV-1seronegative infant and adult controls were studied. Infants and children were classified as HIV-1-infected ifthey met any of the following criteria: Centers for Disease Control defined AIDS; HIV antibody positivity beyond 18 months of age; two positive PCR or HIV-1 culture results; or one positive PCR or HIV-1 culture with p24 antigenemia. Seroreversion was defined as the loss of HIV antibody in the absence of criteria for HIV-1 infection or hypogammaglobulinemia. With these criteria, 109 infants were identified as HIV-1-infected and 88
As interest in carrying out vaccine studies in children infected with human immunodeficiency virus type 1 (HIV-1) develops, a fuller understanding of the infants' antibody repertoire to HIV-1 will be necessary. While some aspects of infants' humoral response to perinatal HIV-1 infection are already known, the ontogeny and full scope of this response are incompletely described. This is, to a large extent, a result of the presence of confounding transplacentally acquired maternal antibody. To circumvent this problem several investigators have measured antibodies that are not transferred across the placenta. However, this strategy has yielded limited results: the measurement of HIV IgM responses has been troubled with problems of specificity (1, 2); IgG subclass responses have been measured in only a handful of infants (3); and IgA responses are infrequently present in the first few months of life (4, 5). More detailed information including the development of antibodies to specific HIV-1 antigens has been lacking. A different approach has been to measure HIV-specific antibody production of B cells from infected infants. The B cells are washed free of serum and maintained in culture. Antibody that is detected in the culture supernatant is, in theory, produced de novo by the infant's B cells. This technique of measuring in vitro antibody production (IVAP) has been used to diagnose HIV-1 infection in infants and
as uninfected. In Vitro Antibody Assay. Blood from subjects was drawn by venipuncture into collection tubes containing either sodium heparin or EDTA and processed within 2 hr. Peripheral blood lymphocytes (PBLs) were isolated by centrifugation over Ficoll/Hypaque, washed twice in Hanks' balanced salt solution, and resuspended in culture medium containing RPMI 1640 (GIBCO), 10%o fetal bovine serum, gentamicin (50 ,ug/ml), vancomycin (50 ,ug/ml), and amphotericin B (2.5 pg/ml). Approximately 2 x 106 PBLs were used in each assay. The percentage of CD4+ and CD8+ T lymphocytes was determined by immunofluorescence with fluorescein- or phycoerythrin-conjugated anti-Leu-3a and anti-Leu-2 monoclonal antibodies (Becton Dickinson) as observed with a Zeiss epifluorescence phase-contrast microscope. Fifty microliters of supernatant from a marmoset B95-8 cell-line culture containing infectious Epstein-Barr virus (EBV) was added to each culture. In a limited number (n = 32) of cases, two sets of PBL cultures were set up in parallel for each patient, one into which EBV was added and the other without EBV. After an overnight incubation at 37°C in 5% C02, the cultures were centrifuged, the supernatant was removed, and
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Abbreviations: HIV-1, human immunodeficiency virus type 1; IVAP, in vitro antibody production; PBL, peripheral blood lymphocyte; EBV, Epstein-Barr virus; SI, secretion index.
2340
Immunology: Poflack et al. the cells were washed twice and resuspended in culture medium without B95-8 supernatant. An aliquot of supernatant was removed at this time and stored as the initial time-point (day 0) supernatant. Medium was changed weekly and supernatants were stored at -20°C. PBLs were maintained in culture for 4 weeks. HIV-specific IgG was measured in culture supernatants (diluted 1:4) by solid-phase ELISA (HIV EIA; Abbott) and Western blot (Epitope, Beaverton, OR). Western blots were considered to be positive when at least one band corresponding to an envelope protein or two bands corresponding to other HIV-1 structural proteins were present. Of 100 Western blot-positive supernatants, 99 had ELISA values >0.025 OD unit and a 2-fold increase in the ELISA reading above that of the day 0 supernatant of the same culture. These results were used as guidelines for selection of subsequent samples for Western blot analysis. The amount of antibody produced in culture was ascertained by measuring the increase in the HIV-1 ELISA OD after 1 week of culture over the initial day 0 ELISA reading. To compare different individuals or the same individual at different times, measurements were standardized in the following manner: the OD value obtained from the culture supernatant on day 7 was divided by the value obtained on day 0. This ratio was called the secretion index (SI). Data Analysis. Statistical analysis was performed using PRODAS software (Conceptional Software, Houston). Crosssectional and longitudinal data were analyzed by ANOVA and logistic regression models. Both linear and quadratic models were considered.
RESULTS Sensitivity and Specificity of Assay. HIV-specific IVAP was detected for 45 of 46 (97%) HIV-1-infected adults [66 of 72 (91.7%) assays] and 57 of 59 (96.7%) HIV-1-infected infants >2 years of age. Only 1 of 115 (0.85%) [1 of 158 (0.63%) assays] seronegative infants, children, and adults had culture supernatants in which Western blot-confimed HIV-specific IVAP was detected. Eight additional seronegative individuals had cultures that produced a >2-fold increase in baseline ELISAs but were negative by Western blot. HIV-specific IVAP was also detected on at least one occasion for 12 of 88 (13.6%) uninfected infants and children born to HIV-1infected mothers. Seven of these 12 infants had positive assays in the first 2 months of life (6 in the first month); 3 between 2 and 6 months of life; and 2 when older (one child at 30 and 38 months, and another at 30 months). Only 3 of these infants had more than one positive IVAP assay; all eventually lost IVAP. Kinetics of In Vitro Antibody Production. In >95% of positive IVAP cultures, HIV antibody was detected in culture supernatants in the first few days of incubation and peaked by days 7-14. In several cases antibody was not detected until after 2 weeks of culture, and in two cases only after 3 weeks of culture. Analysis of the 32 paired cultures set up with and without the addition of EBV showed no effect of EBV on the mean amount and time course of HIV-specific IVAP. Evolution of HIV IVAP in Infants of HIV-i-Infected Mothers. The percentage of infants for whom HIV IVAP was detected at various times during the first 2 years of life is shown in Fig. 1. The proportion of infected infants with detectable IVAP increased from 58.8% in the first 2 months of life to 85.7% between 6 and 12 months. Between 12 and 24 months 83.3% had IVAP. In the group of uninfected seroreverter infants the proportion with IVAP decreased from -28% in the first 2 months of life to 24
12-24
Age, months
FIG. 1. Proportion of infants of HIV-infected mothers with positive HIV IVAP as a function of age. The proportion of HIVinfected infants (hatched bars) producing HIV IgG in vitro increases steadily over the first 12 months, reaching adult values after 2 years of life. The percentage of uninfected HIV-seroreverters (solid bars) for whom IVAP is detected declines rapidly after the first 2 months of life. The number above each bar refers to the number of infants assayed for each time period.
for 7 of 8 (87.5%) HIV-1-infected infants compared with 6 of 17 (35.2%) seroreverters (P = 0.02). The amount of antibody produced, as represented by the SI values, from birth to age 12 months of each of the 105 prospectively followed infants of HIV-1-infected women is shown in Fig. 2. The SI for the 115 seronegative children and adult controls was 1.22 + 0.39 (mean ± SD). The SI was considered positive only if it was 2 SDs (SI > 2) above this level. Values for HIV-1-infected infants (n = 47) and uninfected seroreverter infants (n = 58) are plotted separately (Fig. 2 Left and Right, respectively). The SI values of the two HIV-infected
Seroreverter
-00 CQ
0
4
8
12
0
4
8
12
Age, months FIG. 2. HIV-specific IVAP during the first year of life. The antibody SI values of HIV-infected (n = 47) (Left) and HIVuninfected seroreverter (n = 58) (Right) infants of HIV-infected mothers, followed during the first 12 months of life, are plotted. The amount of antibody produced in vitro is proportional to the SI. In infected infants, SI increases steadily over the first year of life. The regression line of this rise in SI is drawn. The SI of uninfected seroreverters, by contrast, decreases from birth to approximate that of 112 seronegative controls (SI = 1.22 + 0.39). Significant differences in IVAP between infected and seroreverter infants are measured after the age of 2 months.
2342
Immunology: Pollack et al.
Proc. Natl. Acad. Sci. USA 90 (1993)
groups were comparable in the newborn period (2.8 ± 1.7 and 2.4 ± 3.2, respectively). In the group of infected infants the SI increased during the first year of life, reflecting an increase in the amount of HIV-specific antibody produced. The amount of increase for HIV-1-infected infants is described by the regression equation SI = -1.72 + (4.42)-(age in months). The strong upward slope of this equation is significantly different (P = 0.001) from the slope of SI, for the same period of time, of HIV-1-uninfected seroreverters, described by the regression equation SI = 2.47 - (0.11) (age in months). A difference in the mean SI between infected and uninfected infants of HIV-1-infected mothers was first apparent between 2 and 6 months of age, when the mean SI was 20.7 ± 35.3 for infected infants compared with 1.9 ± 3.6 for uninfected seroreverter infants (P = 0.001). The difference between these groups remained statistically significant for each subsequent interval. IVAP continued to rise in the second half of the first year of life of infected infants (SI = 36.4 + 34.9). In the second year of life the SI was 23.8 ± 27.9. For infected children .2 years of age the SI was 14.2 ± 25.3, similar to the SI of infected adults, 17.7 ± 20.7. IVAP to Specific mIV-1 Antigens. The ability to make antibody to each of the specific structural HIV-1 antigens varied with the age of the infant (Fig. 3). Antibody to more than one Western blot band was present in 90% (180/200) of cases. The ability to recognize multiple HIV-1 antigens developed sequentially. The total number of different HIV-1 antigen-specific antibodies increased with the age of the infant from 2.0 ± 1.15 in the first month of life to 3.86 ± 1.58 by 12 months of age. The predominant antigens recognized were the envelope proteins gpl60, gpl20, and gp4l. Antibodies to gpl60 were present earlier and in a higher percentage of infected infants than antibodies to gpl20 or gp4l, whereas antibodies to the pol and gag proteins appeared later and were present in only a small percentage of infants. By contrast, antibodies to a much larger number and array of HIV-1 antigens (5.14 ± 1.34) were detected in the first 2 months of life for infants who produced IVAP but later seroreverted. The distribution of these antibodies was different from that of infected infants; a much higher proportion 100-
GP160 GP120
80 --
~0
C,)
P65
CO
a)
._
-5-
60
///-
-
P65/51
--
._
P41 40
P31
ol
.
-al
100 7
0
500
c
cn
0~
.0
a
0
o~~~~
10-
o
5-
0
0
/
0
I 0
.
. 10
.
20
30
TI 40
50
% CD4+ PBLS
FIG. 4. Variation of antibody SI as a function of percentage of CD4+ T lymphocytes in peripheral blood in HIV-infected infants. The SI is plotted as a function of % CD4+ T-lymphocytes present in peripheral blood and IVAP culture of the HIV-infected infants at the time of IVAP assay. Infants with a higher percentage of CD4+ lymphocytes produced more HIV-specific antibody than infants with a lower percentage of CD4+ cells. This relationship is described by the regression equation log SI = 0.65 + 0.026(% CD4+).
of these infants had antibody directed to gp120, p65, gp4l, p24 (85%), to p55/51 (71%), and to p31 antigen (14%). Characteristics of HIV-1-Infected Infants Who were IVAPNegative. HIV IVAP was not detected for 16 HIV-1-infected infants and children. The inability to detect IVAP by cells from these children was closely associated with severe immunodeficiency. Fifteen of these children had 24
Age, months FIG. 3. Development of the ability to make antibody to specific HIV antigens in infancy and early childhood. The Western blot patterns of HIV-infected infants and children producing antibody in vitro to HIV-1 were examined. The proportion of infants producing antibodies to each specific antigen, among all the HIV-infected infants studied for each age group, is plotted over time. Antibody to p18 was not detected in any of the culture supernatants (data not shown on graph).
Although the IVAP assay is less sensitive and specific than viral culture or PCR in detecting HIV-1 infection in infants (12, 13), this technique provides an effective method of studying the repertoire of HIV-specific antibody response normally masked by maternal antibody. Because B cells from these infants are releasing HIV-specific antibody spontaneously, as demonstrated by the similar antibody production of EBV-stimulated versus unstimulated cultures, it provides a measure ex vivo, at a defined point in time, of the rate of HIV-antibody production in vivo. This spontaneous production of HIV antibodies is characteristic of the polyclonal
Immunology: Poflack et al. B-cell activation reported in HIV-1-infected individuals (14, 15). Both the proportion of HIV-1-infected infants producing antibody and the amount of antibody produced increased gradually with age. In the first 2 months of life, 60%o of infected infants produce HIV-specific IgG, but only in very small amounts. By 6 months of life, >85% of infected infants produced antibody in amounts that approached those of adults. Whether the increase in IVAP is due to increased antigenic exposure or merely reflects the maturational (16) and numerical* changes in infant B cells that allow increased antibody production is unresolved. Viral activity, as measured by p24 antigen, increases dramatically in the first few weeks of life (12, 13). This may provide the stimulus that drives the postnatal increase in HIV-antibody production. That IVAP accurately reflects physiologic in vivo processes is suggested by the fact that the increase in HIV IVAP during the first year of life closely resembles the time course of total immunoglobulin production seen in normal infants born to hypogammaglobulinemic mothers (17, 18), and the reported time course of IgG3 production to HIV-1 (3). Moreover, the time course of HIV-antibody production corresponds to the development of hypergammaglobulinemia, one of the early markers of HIV-1 infection in infants.t The acquisition of HIV-1 antigen-specific IgG synthesis appears to be sequential and selective. In the first 2 months of life, infants make antibodies to only one or two HIV-1specific structural antigens. By 6 months of life, infants are producing antibodies to three or four antigens. A similar pattern has been noted in the development of HIV-specific IgA in infants (4, 5). Overall, antibodies to envelope proteins predominate at all ages, especially antibodies to gp160, which appear very early and increase steadily during infancy. Antibodies to gpl20 and gp4l appear later but increase more rapidly:
