765
Detection of Infectious Immune Complexes in Human Immunodeficiency Virus Type 1 (DIV-l) Infections: Correlation with Plasma Viremia and CD4 Cell Counts Susan A. Fiscus, Ewa B. Wallmark,* J. D. Folds, J. Fryer, and C. M. van der Horst
Departments of Microbiology, Medicine, and Biostatistics, University of North Carolina at Chapel Hill
Various surrogate laboratory markers have been used to predict the likelihood of progression to AIDS and to monitor the effectiveness of different drug therapies in people infected with human immunodeficiency virus (HIV). The assay most frequently used is the CD4 cell count [l , 2]. However, decreases in these values often occur late in the pathogenesis of HI V infection. A virologic marker that could indicate events before T cell death might be better than changes in CD4 cell coun ts. One marker that has been proposed is p24 antigen. However, only about one-third of HIV-infected people display HIV p24 antigenemia [l , 3]. The frequency of p24 antigenemia in general increases with disease progression, but many patients with frank AIDS are HIV p24 antigen-negative, and some asymptomatic patients have demonstrable p24 antigen. Investigators have attempted to improve the sensitivity of the p24 antigen test by treating serum with acids to hydrolyze immune complexes, followed immediately by neutralization just before testing in the p24 antigen assay [4-6]. More recently, cell-free viremia has been suggested as a
Received 22 February 1991; revised 28 May 1991. Presented in part: 10th AIDS Clinical Trial Groups meeting, Washington, DC, November 1990 (abstract 6). Informed consent was obtained from participants. The study protocol was approved by the Committee on the Protection of the Rights of Human Subjects, the Institutional Review Board of the University ofNorth Carolina at Chapel Hill. Financial support: National Institutes of Health (cooperative agreement AI-25868). Reprints or correspondence: Dr. Susan A. Fiscus. Department of'Microbiology, Rm. 436 Taylor Hall, CB#7l40. University of North Carolina at Chapel Hill. Chapel Hill, NC 27599. *Present address: Infektionskliniken, Malmo all manna sjukhus, Sweden. The Journal of Infectious Diseases 1991;164:765-9 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6404-0021 $0 1.00
better surrogate marker for progression to AIDS [3, 7-9]. However, the plasma culture assay is expensive, time consuming, and appropriately done only in special laboratories. Although one study found that 100% of HIV-seropositive patients not on antiretroviral drug therapy were positive for plasma viremia [7], other studies have shown only 30%-75% of patients to be plasma viremic [3, 8, 9]. Circulating immune complexes have been demonstrated many times in patients infected with HIV [l0-13]. A few investigators have shown that at least some of these complexes contain HIV antigen [11-13], while others have failed to detect any HIV antigen [10]. To date, there has been only a single report of infectious HIV in the circulating immune complexes of AIDS patients [13]. Since polyethylene glycol (PEG) precipitation is a simple and efficient method of isolating immune complexes from serum and plasma samples, we examined 2% PEG precipitates of HIV-seropositive plasmas.
Materials and Methods Patients. A total of 63 HIV type 1 (HI V-1)-seropositive patients provided heparinized blood samples. All patients were evaluated by one of the authors. Some of the patients were being screened for enrollment in various AIDS Clinical Trials Group (ACTG) protocols, while others were already on antiviral drug therapy while this study was being conducted. Twenty milliliters of heparinized blood was obtained from each patient. Blood was also obtained for CD4 cell counts. Twenty HIV-I-seronegative donors (all healthy hospital personnel) were also tested. Normal donor cells for coculture. Buffy coat cells used as feeder cells in the culture experiments were purchased from the American Red Cross, isolated, washed, and cultured in phytohemagglutinin (PHA; Sigma, St. Louis)-containing medium using previously described methods [3].
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
The detection of infectious immune complexes in plasma after human immunodeficiency virus (HIV) infection may be useful as a surrogate marker of progression of disease and may help in understanding the pathogenesis of AIDS. Polyethylene glycol (PEG) precipitates of plasma were tested for the presence of HIV p24 antigen and infectious virus. Results were compared with data from cell and plasma cultures, plasma p24 antigen, CD4 cell counts, and stage of disease. PEG precipitation increased the detection rate of the p24 antigen assay from 38.3% to 58.7%. There was a significant correlation between precipitable p24 antigen and plasma viremia, changes in CD4 cell counts, and progression of disease. The sensitivity of the PEG-precipitable p24 antigen assay versus traditional p24 antigen testing was 59.0% and the specificity 91.7%. The assay was reproducible and may be a useful determinant of viral load, clinical progression, and antiretroviral efficacy.
766
Concise Communications
with commercially available directly conjugated monoclonal antibodies. All CD4 determinations were done within 2 weeks of the culture and p24 antigen assays. Most were done on the same day. The Clinical Immunology Laboratory was a certified participant in the flow cytometry quality assurance program of the National Institutes of Health ACTG during the course of the study. Statistics. We used standard x 2 tests for 2 X 2 tables. Where cell numbers were very small we applied Yates's [14] correction. At the same time we calculated a Q statistic, which measures association and its standard error, originally proposed by Yule (cited in [14]).
Results There were 15 asymptomatic subjects, 32 with AIDS-related complex (ARC), 16 with AIDS, and 20 seronegative healthy workers in the study population. Samples from the seronegative donors were tested in all assays except CD4 cell counts and found to be negative in all cases. CD4 cell counts of the patients ranged from 840 to 2 with a median of 100/ mrrr'. All but one sample was positive for HIV -1 in the cell culture assay (61/62, 98.4%), but only 50.8%(30/59) of the plasma cultures were positive. The patient who was negative in cell culture had a CD4 count of 840/mm 3 and was asymptomatic. HIV p24 antigen was found in only 38.1 % (24/63) of the plasma samples. In contrast, 58.7% (37/63) of the PEG-precipitated p24 antigen assays were positive and 52.7% (29/55) of the cultured PEG precipitates were positive. The PEG precipitate culture and PEG precipitate p24 antigen tests were significantly associated with each other (Q = .97, P < .001) and with the incidence of plasma viremia (Q = .90, P < .001). The standard plasma p24 antigen assay had a lower level of association with the PEG precipitate p24 antigen assay (Q = .81, P < .001) and virtually no association with plasma viremia results (Q = .45, P < .12). The sensitivity of the PEG precipitate p24 antigen was 59% compared with the traditional p24 antigen assay, with a specificity of 91. 7%. All but one of the HIV-seropositive patients were cell culture-positive. The incidence of p24 antigen was somewhat higher in patients with plasma viremia (15/30, 50%) than in those without (8/29,27.6%) (figure lA). However, there was a significant difference in PEG-precipitable HIV between patients with and without plasma viremia. Only 31.0% (9/29) of those lacking plasma viremia had precipitable p24 antigen, and 29.6% (18/27) had precipitable infectious HIV. In contrast, 90.0% (27/30) of patients with plasma viremia had precipitable HIV antigen and 82.6% (19/23) had precipitable infectious virus. There was also a clear correlation in CD4 cell counts and the ability to detect infectious immune complexes (figure 1B). Again, no difference was observed in the cell culture results. Plasma p24 antigen, plasma viremia, and precipitable HIV showed marked differences, however. In patients
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
HIV-l isolation by quantitative plasma culture. Heparinized blood was centrifuged at 600 g to obtain plasma and buffy coat cells. The plasma was filtered through 0.45-JLm filters, and 10fold serial dilutions ranging from 10- 1 to 10- 3 were made in culture medium (RPMI 1640, 20% heat-inactivated fetal calf serum, 5% human interleukin-2 [Cellular Products, Buffalo, NY], 2 JLg/ml polybrene [Sigma], penicillin, and streptomycin). Aliquots (1 ml) of undiluted or diluted plasma were added to 1 ml of 2 X 106 3- to 4-day-old PHA-stimulated donor peripheral blood mononuclear cells (PBMC), and the samples were incubated overnight at 37°C. In the morning the cells were washed three times at 400 g to remove any free p24 antigen. They were then resuspended in 2 ml of fresh culture medium and placed in 24-well plates. Each week half of the medium of each well was removed and replaced with fresh PHA-stimulated donor PBMC in culture medium. Culture supernatants were tested at 3 weeks for p24 antigen by ELISA. HIV isolation by cell cocultivation. The buffy coat cells obtained from the first centrifugation were resuspended in Hanks' balanced salt solution (HBSS) and layered on lymphocyte separation medium (Organon Teknika, Durham, NC). After centrifugation at 400 g for 30 min, the isolated cells were washed twice in HBSS, counted, and resuspended in culture medium. Dilutions of the patients' cells ranging from I X 106 to I X 102 were placed in 24-well plates with 2 X 106 PH A-stimulated donor PBMC. At weekly intervals half of the supernatant medium was replaced with fresh PHA-stimulated donor PBMC. The culture supernatants were tested for HIV p24 antigen during weeks 2, 3, and 4. The University of North Carolina at Chapel Hill Retrovirology Laboratory was a certified participant in the virologic quality assurance program of the National Institutes of Health ACTG during the course of this study. PEG precipitation. One milliliter of0.45-JLm-filtered plasma was mixed with 0.2 ml of 12%PEG-8000 (Sigma), vortexed, and refrigerated overnight at 4°C. In the morning the samples were centrifuged at 2000 g for 20 min at 4°C. The supernatants were saved and the pellets were washed twice in 2% PEG with 0.02 M EDTA in HBSS and resuspended in 0.6 rnl of culture medium. Aliquots of the PEG precipitates and supernatants (0.2 ml) were tested directly in the p24 antigen assay. The rest of the samples were incubated with 2 X 106 PHA-stimulated donor PBMC overnight at 37°C. Cells were then washed three times in HBSS, resuspended in 1.5 ml of culture medium, placed in 24-well plates, and incubated 3 weeks at 37°C. Cell culture supernatants were tested at 3 weeks for the presence of p24 antigen. Acid hydrolysis. Test plasma or resuspended PEG precipitates (100 JLl) were mixed with an equal volume of 1.5 M glycine, pH 2.2, and incubated at 37°C for 1 h. Then the samples were neutralized with 100 JLl of 1.5 MTRIS, pH 8.6. These were tested along with nonhydrolyzed plasma and PEG precipitates in the p24 antigen ELISA. HIV p24 antigen ELISA. p24 antigen was assayed by p24 antigen ELISA (Coulter Diagnostics, Hialeah, FL). Manufacturer's instructions were followed, and the standard provided in the kit was used to quantitate p24 antigen. Results were considered positive if the amount of antigen was ;?o8 pg/rnl. CD4 cell counts. CD4 cells were counted by the Clinical Immunology Laboratory at North Carolina Memorial Hospital using flow cytometry (FACScan; Becton Dickinson, San Jose, CA)
lID 1991; 164 (October)
Concise Communications
JID 1991 :1 64 (October)
A
Negat ive
10 01r-:
75
-
0
•-
a
P l a sm .
•
"V
"v
P EG Pp l
Ppl C U l ture
w
>
•
l-
iiiiii
UI
r-e-
iiiiii
P P I "9 PEG PP I Cul tun
a
jm~
II
I
••
C 04 < 1 70 / m m
3
-
• ?;
@ I
~ ~
3
Cell Culture
~
75
I
C 04 !: 1701 m m
0
100
• • !!
P l a sm a
"9
Pla sma
Cultur .
PEG
a
~
a
iii ii• ]i@ iii [il iii !!• iii iii !!• un!1~li ra it ~ •• li ra • a
~~
P < .001
::::::::::
50
:::::»~
5
I
~
•
ii
ff
~ ~
ii
:~:::::~
t ill
= rr
Figure 1. A, Correlation of plasma viremia with positivehuman immunodeficiency virus (HIV) cell culture, plasma p24 antigen (Plasma Ag). and polyethylene glycol- precipitable p24 antigen (PEG Ppt Ag) and infectious HIV (PEG Ppt Culture). B, Correlation ofC D4 cell counts ;;.170/ mmJ or < 170/mm J with positivecell cultures, plasma p24 antigen (Plasma Ag). plasma viremia (Plasma Culture), and PEG-precipitable p24 antigen (PEG Ppt Ag) and infectious HIV (PEG Ppt Cultur e).
with CD4 counts > I 70 , only 19% (4/21) had plasma p24 a ntigen, 0 of 20 demonstrated plasma viremia, 23 .8%(5/21 ) had precipitable p24 antigen . and 5.6%(1/1 8) had infectiou s immun e complexes. When C D4 cell co unts dropped below 170, 48 .7% ( 19/ 39) dem on strated plasma p24 antigen . 80 .6% (29/ 36) had pla sma viremi a , 74.4% (29/39) had precipitable p24 antigen. an d 77 .1 W·(2 7/35) had precipitable infectiou s HI V. As disease progressed from asymptoma tic to AR C to AIDS. there was a decre ase in C D4 cell counts and corr espo nding increases in th e incidence of plasma viremia a nd PEG-precipit abl e p24 a n tige n and in fectiou s HI V (figure 2). The mean C D4 cell count for asy m pto ma tic patients was 36 7/ m m J • Patients with ARC had a n average of 96/ mm J and those with AIDS on ly 52/mm J . T he re was a n insignificant increase in the incidence o f p24 a n tigen: 26. 7'70 (4/15) in asymptomatic patients , 40 .6% (1 3/32) in those with AR C. and 43.8%(7/16) in patients with AIDS . Plasma viremia. on the other hand, showed a sign ifica nt inc rease as disease pro-
gresse d . None of the asymptomat ic patients had plasm a viremia, but 60 %(1 8/30) of th ose with ARC and 80%( 12/1 5) o f th ose with AIDS did . Th is trend was mimicked in the immune complex assays. Precipitable HIV p24 antigen wa s found in 26 .7% (4/1 5) of asy mp to ma tic, 62.5 % (20/ 32) of ARC. a nd 81.3% (13/16) of AIDS pati ents. Precipit able in fectious HIV was found in 18.2% (2/ I I ), 56.3 %( 18/ 32 ), a n d 75 % (9/12), respectively. We tried to improve both the sta nda rd pla sma p24 and PEG-precipitated p24 assays by ac id treatment. Eighteen plasma samples from Hlv-serop ositive patients whose CD4 cell counts ranged from 500 to 2/mm J were tested . Six patients with CD4 cell counts> 170/mm J were negative for plasma viremia. and only one had detectable p24 antigen using the standard technique . One additional patient was detected when the pla sma was treated with acid. Four of the six patients had p24 antigen in th e PEG precipitates, but on ly one of these was still po sitive after ac id treatment. T en o f the 12 patients with CD4 cell co unts < 170/mm J had plasma viremi a. Seven of th ese had detectable p24 a n tige n with out acid hydrolysis and on ly o ne othe r was un covered by ac id trea tment. The same 10 pati ents who had plasm a viremia also had PEG-precipitabl e p24 a n tigen . Two o f th ese rea cti vities were lost after acid treatment. Norm al human plasma tre ated in thi s way was negati ve in each instance.
Discussion Better sur roga te mark ers of th e clinical stage of HI V infe ction and viral replic at ion ar e needed to evaluate the efficacy of antiretroviral dru g therapi es. A good surrogate m arker sho uld have the following cha racteristics: be prevalen t amo ng infected indi vidual s, correlate clinically, change with effective therapy or with clinical improvement, be reproducib le, and make biologic sense . In addition, it would be adva n tag eous to hav e a n assay that is easy to perform , rapid, relati vely inexpensive, a nd safe. Several such markers hav e been proposed, including C D 4 cell counts. pla sma viremia. immune co mplexes. a n d p24 a n tigene m ia. C D4 cell counts a re clearl y the best clin ical su rrogate mark ers found to dat e but are ex pensive a nd ca n be don e on ly in specially eq uipped la bora tori es [I , 2]. Plasm a viremi a has also been prop osed as a goo d surroga te m arker but is tim e co nsuming (up to 4 wee ks), lab oriou s, ex pe ns ive, a nd a ppro pria tely don e in specia lized laborato ries [3, 7-9] . One gro up has reported tha t th eir plasma culture assay overa ll had a sensi tivity of 4 5% a nd a spec ificity of 82.7% co m pa red with the tradition al p24 a ntige n assay [3]. The res u lts from our regular plasm a cultures show ed a sensitivity o f 48.4 % a nd a specificity of 6 7.9%. T he PEG-precipitate c u lture had a sensitivity of 56.0% a nd a specificity of 80% co mpar ed with p24 antigen testing; how ever. the PEG-precipitate p24 antigen test had a sensitivity of 59%an d a specificity
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
B
Go
it t is It
ii ii
~ ~ ~
25
a
ii ii
!!
p < .0 0 1
50
o
V i r emi a
Ce II Cu lt ur •
iili PEG a
P os i ti ve
V i rem i 8
767
Concise Communications
768
0
10 0
Cell Cu Iture Plasma
Ag
-
r-
~ PI asma Culture
7S
-
Ppl Ag
CD
>
PEG Ppt Culture
u I
o'" a.. so ~
t
P
Detection of Infectious Immune Complexes in Human Immunodeficiency Virus Type 1 (DIV-l) Infections: Correlation with Plasma Viremia and CD4 Cell Counts Susan A. Fiscus, Ewa B. Wallmark,* J. D. Folds, J. Fryer, and C. M. van der Horst
Departments of Microbiology, Medicine, and Biostatistics, University of North Carolina at Chapel Hill
Various surrogate laboratory markers have been used to predict the likelihood of progression to AIDS and to monitor the effectiveness of different drug therapies in people infected with human immunodeficiency virus (HIV). The assay most frequently used is the CD4 cell count [l , 2]. However, decreases in these values often occur late in the pathogenesis of HI V infection. A virologic marker that could indicate events before T cell death might be better than changes in CD4 cell coun ts. One marker that has been proposed is p24 antigen. However, only about one-third of HIV-infected people display HIV p24 antigenemia [l , 3]. The frequency of p24 antigenemia in general increases with disease progression, but many patients with frank AIDS are HIV p24 antigen-negative, and some asymptomatic patients have demonstrable p24 antigen. Investigators have attempted to improve the sensitivity of the p24 antigen test by treating serum with acids to hydrolyze immune complexes, followed immediately by neutralization just before testing in the p24 antigen assay [4-6]. More recently, cell-free viremia has been suggested as a
Received 22 February 1991; revised 28 May 1991. Presented in part: 10th AIDS Clinical Trial Groups meeting, Washington, DC, November 1990 (abstract 6). Informed consent was obtained from participants. The study protocol was approved by the Committee on the Protection of the Rights of Human Subjects, the Institutional Review Board of the University ofNorth Carolina at Chapel Hill. Financial support: National Institutes of Health (cooperative agreement AI-25868). Reprints or correspondence: Dr. Susan A. Fiscus. Department of'Microbiology, Rm. 436 Taylor Hall, CB#7l40. University of North Carolina at Chapel Hill. Chapel Hill, NC 27599. *Present address: Infektionskliniken, Malmo all manna sjukhus, Sweden. The Journal of Infectious Diseases 1991;164:765-9 © 1991 by The University of Chicago. All rights reserved. 0022-1899/91/6404-0021 $0 1.00
better surrogate marker for progression to AIDS [3, 7-9]. However, the plasma culture assay is expensive, time consuming, and appropriately done only in special laboratories. Although one study found that 100% of HIV-seropositive patients not on antiretroviral drug therapy were positive for plasma viremia [7], other studies have shown only 30%-75% of patients to be plasma viremic [3, 8, 9]. Circulating immune complexes have been demonstrated many times in patients infected with HIV [l0-13]. A few investigators have shown that at least some of these complexes contain HIV antigen [11-13], while others have failed to detect any HIV antigen [10]. To date, there has been only a single report of infectious HIV in the circulating immune complexes of AIDS patients [13]. Since polyethylene glycol (PEG) precipitation is a simple and efficient method of isolating immune complexes from serum and plasma samples, we examined 2% PEG precipitates of HIV-seropositive plasmas.
Materials and Methods Patients. A total of 63 HIV type 1 (HI V-1)-seropositive patients provided heparinized blood samples. All patients were evaluated by one of the authors. Some of the patients were being screened for enrollment in various AIDS Clinical Trials Group (ACTG) protocols, while others were already on antiviral drug therapy while this study was being conducted. Twenty milliliters of heparinized blood was obtained from each patient. Blood was also obtained for CD4 cell counts. Twenty HIV-I-seronegative donors (all healthy hospital personnel) were also tested. Normal donor cells for coculture. Buffy coat cells used as feeder cells in the culture experiments were purchased from the American Red Cross, isolated, washed, and cultured in phytohemagglutinin (PHA; Sigma, St. Louis)-containing medium using previously described methods [3].
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
The detection of infectious immune complexes in plasma after human immunodeficiency virus (HIV) infection may be useful as a surrogate marker of progression of disease and may help in understanding the pathogenesis of AIDS. Polyethylene glycol (PEG) precipitates of plasma were tested for the presence of HIV p24 antigen and infectious virus. Results were compared with data from cell and plasma cultures, plasma p24 antigen, CD4 cell counts, and stage of disease. PEG precipitation increased the detection rate of the p24 antigen assay from 38.3% to 58.7%. There was a significant correlation between precipitable p24 antigen and plasma viremia, changes in CD4 cell counts, and progression of disease. The sensitivity of the PEG-precipitable p24 antigen assay versus traditional p24 antigen testing was 59.0% and the specificity 91.7%. The assay was reproducible and may be a useful determinant of viral load, clinical progression, and antiretroviral efficacy.
766
Concise Communications
with commercially available directly conjugated monoclonal antibodies. All CD4 determinations were done within 2 weeks of the culture and p24 antigen assays. Most were done on the same day. The Clinical Immunology Laboratory was a certified participant in the flow cytometry quality assurance program of the National Institutes of Health ACTG during the course of the study. Statistics. We used standard x 2 tests for 2 X 2 tables. Where cell numbers were very small we applied Yates's [14] correction. At the same time we calculated a Q statistic, which measures association and its standard error, originally proposed by Yule (cited in [14]).
Results There were 15 asymptomatic subjects, 32 with AIDS-related complex (ARC), 16 with AIDS, and 20 seronegative healthy workers in the study population. Samples from the seronegative donors were tested in all assays except CD4 cell counts and found to be negative in all cases. CD4 cell counts of the patients ranged from 840 to 2 with a median of 100/ mrrr'. All but one sample was positive for HIV -1 in the cell culture assay (61/62, 98.4%), but only 50.8%(30/59) of the plasma cultures were positive. The patient who was negative in cell culture had a CD4 count of 840/mm 3 and was asymptomatic. HIV p24 antigen was found in only 38.1 % (24/63) of the plasma samples. In contrast, 58.7% (37/63) of the PEG-precipitated p24 antigen assays were positive and 52.7% (29/55) of the cultured PEG precipitates were positive. The PEG precipitate culture and PEG precipitate p24 antigen tests were significantly associated with each other (Q = .97, P < .001) and with the incidence of plasma viremia (Q = .90, P < .001). The standard plasma p24 antigen assay had a lower level of association with the PEG precipitate p24 antigen assay (Q = .81, P < .001) and virtually no association with plasma viremia results (Q = .45, P < .12). The sensitivity of the PEG precipitate p24 antigen was 59% compared with the traditional p24 antigen assay, with a specificity of 91. 7%. All but one of the HIV-seropositive patients were cell culture-positive. The incidence of p24 antigen was somewhat higher in patients with plasma viremia (15/30, 50%) than in those without (8/29,27.6%) (figure lA). However, there was a significant difference in PEG-precipitable HIV between patients with and without plasma viremia. Only 31.0% (9/29) of those lacking plasma viremia had precipitable p24 antigen, and 29.6% (18/27) had precipitable infectious HIV. In contrast, 90.0% (27/30) of patients with plasma viremia had precipitable HIV antigen and 82.6% (19/23) had precipitable infectious virus. There was also a clear correlation in CD4 cell counts and the ability to detect infectious immune complexes (figure 1B). Again, no difference was observed in the cell culture results. Plasma p24 antigen, plasma viremia, and precipitable HIV showed marked differences, however. In patients
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
HIV-l isolation by quantitative plasma culture. Heparinized blood was centrifuged at 600 g to obtain plasma and buffy coat cells. The plasma was filtered through 0.45-JLm filters, and 10fold serial dilutions ranging from 10- 1 to 10- 3 were made in culture medium (RPMI 1640, 20% heat-inactivated fetal calf serum, 5% human interleukin-2 [Cellular Products, Buffalo, NY], 2 JLg/ml polybrene [Sigma], penicillin, and streptomycin). Aliquots (1 ml) of undiluted or diluted plasma were added to 1 ml of 2 X 106 3- to 4-day-old PHA-stimulated donor peripheral blood mononuclear cells (PBMC), and the samples were incubated overnight at 37°C. In the morning the cells were washed three times at 400 g to remove any free p24 antigen. They were then resuspended in 2 ml of fresh culture medium and placed in 24-well plates. Each week half of the medium of each well was removed and replaced with fresh PHA-stimulated donor PBMC in culture medium. Culture supernatants were tested at 3 weeks for p24 antigen by ELISA. HIV isolation by cell cocultivation. The buffy coat cells obtained from the first centrifugation were resuspended in Hanks' balanced salt solution (HBSS) and layered on lymphocyte separation medium (Organon Teknika, Durham, NC). After centrifugation at 400 g for 30 min, the isolated cells were washed twice in HBSS, counted, and resuspended in culture medium. Dilutions of the patients' cells ranging from I X 106 to I X 102 were placed in 24-well plates with 2 X 106 PH A-stimulated donor PBMC. At weekly intervals half of the supernatant medium was replaced with fresh PHA-stimulated donor PBMC. The culture supernatants were tested for HIV p24 antigen during weeks 2, 3, and 4. The University of North Carolina at Chapel Hill Retrovirology Laboratory was a certified participant in the virologic quality assurance program of the National Institutes of Health ACTG during the course of this study. PEG precipitation. One milliliter of0.45-JLm-filtered plasma was mixed with 0.2 ml of 12%PEG-8000 (Sigma), vortexed, and refrigerated overnight at 4°C. In the morning the samples were centrifuged at 2000 g for 20 min at 4°C. The supernatants were saved and the pellets were washed twice in 2% PEG with 0.02 M EDTA in HBSS and resuspended in 0.6 rnl of culture medium. Aliquots of the PEG precipitates and supernatants (0.2 ml) were tested directly in the p24 antigen assay. The rest of the samples were incubated with 2 X 106 PHA-stimulated donor PBMC overnight at 37°C. Cells were then washed three times in HBSS, resuspended in 1.5 ml of culture medium, placed in 24-well plates, and incubated 3 weeks at 37°C. Cell culture supernatants were tested at 3 weeks for the presence of p24 antigen. Acid hydrolysis. Test plasma or resuspended PEG precipitates (100 JLl) were mixed with an equal volume of 1.5 M glycine, pH 2.2, and incubated at 37°C for 1 h. Then the samples were neutralized with 100 JLl of 1.5 MTRIS, pH 8.6. These were tested along with nonhydrolyzed plasma and PEG precipitates in the p24 antigen ELISA. HIV p24 antigen ELISA. p24 antigen was assayed by p24 antigen ELISA (Coulter Diagnostics, Hialeah, FL). Manufacturer's instructions were followed, and the standard provided in the kit was used to quantitate p24 antigen. Results were considered positive if the amount of antigen was ;?o8 pg/rnl. CD4 cell counts. CD4 cells were counted by the Clinical Immunology Laboratory at North Carolina Memorial Hospital using flow cytometry (FACScan; Becton Dickinson, San Jose, CA)
lID 1991; 164 (October)
Concise Communications
JID 1991 :1 64 (October)
A
Negat ive
10 01r-:
75
-
0
•-
a
P l a sm .
•
"V
"v
P EG Pp l
Ppl C U l ture
w
>
•
l-
iiiiii
UI
r-e-
iiiiii
P P I "9 PEG PP I Cul tun
a
jm~
II
I
••
C 04 < 1 70 / m m
3
-
• ?;
@ I
~ ~
3
Cell Culture
~
75
I
C 04 !: 1701 m m
0
100
• • !!
P l a sm a
"9
Pla sma
Cultur .
PEG
a
~
a
iii ii• ]i@ iii [il iii !!• iii iii !!• un!1~li ra it ~ •• li ra • a
~~
P < .001
::::::::::
50
:::::»~
5
I
~
•
ii
ff
~ ~
ii
:~:::::~
t ill
= rr
Figure 1. A, Correlation of plasma viremia with positivehuman immunodeficiency virus (HIV) cell culture, plasma p24 antigen (Plasma Ag). and polyethylene glycol- precipitable p24 antigen (PEG Ppt Ag) and infectious HIV (PEG Ppt Culture). B, Correlation ofC D4 cell counts ;;.170/ mmJ or < 170/mm J with positivecell cultures, plasma p24 antigen (Plasma Ag). plasma viremia (Plasma Culture), and PEG-precipitable p24 antigen (PEG Ppt Ag) and infectious HIV (PEG Ppt Cultur e).
with CD4 counts > I 70 , only 19% (4/21) had plasma p24 a ntigen, 0 of 20 demonstrated plasma viremia, 23 .8%(5/21 ) had precipitable p24 antigen . and 5.6%(1/1 8) had infectiou s immun e complexes. When C D4 cell co unts dropped below 170, 48 .7% ( 19/ 39) dem on strated plasma p24 antigen . 80 .6% (29/ 36) had pla sma viremi a , 74.4% (29/39) had precipitable p24 antigen. an d 77 .1 W·(2 7/35) had precipitable infectiou s HI V. As disease progressed from asymptoma tic to AR C to AIDS. there was a decre ase in C D4 cell counts and corr espo nding increases in th e incidence of plasma viremia a nd PEG-precipit abl e p24 a n tige n and in fectiou s HI V (figure 2). The mean C D4 cell count for asy m pto ma tic patients was 36 7/ m m J • Patients with ARC had a n average of 96/ mm J and those with AIDS on ly 52/mm J . T he re was a n insignificant increase in the incidence o f p24 a n tigen: 26. 7'70 (4/15) in asymptomatic patients , 40 .6% (1 3/32) in those with AR C. and 43.8%(7/16) in patients with AIDS . Plasma viremia. on the other hand, showed a sign ifica nt inc rease as disease pro-
gresse d . None of the asymptomat ic patients had plasm a viremia, but 60 %(1 8/30) of th ose with ARC and 80%( 12/1 5) o f th ose with AIDS did . Th is trend was mimicked in the immune complex assays. Precipitable HIV p24 antigen wa s found in 26 .7% (4/1 5) of asy mp to ma tic, 62.5 % (20/ 32) of ARC. a nd 81.3% (13/16) of AIDS pati ents. Precipit able in fectious HIV was found in 18.2% (2/ I I ), 56.3 %( 18/ 32 ), a n d 75 % (9/12), respectively. We tried to improve both the sta nda rd pla sma p24 and PEG-precipitated p24 assays by ac id treatment. Eighteen plasma samples from Hlv-serop ositive patients whose CD4 cell counts ranged from 500 to 2/mm J were tested . Six patients with CD4 cell counts> 170/mm J were negative for plasma viremia. and only one had detectable p24 antigen using the standard technique . One additional patient was detected when the pla sma was treated with acid. Four of the six patients had p24 antigen in th e PEG precipitates, but on ly one of these was still po sitive after ac id treatment. T en o f the 12 patients with CD4 cell co unts < 170/mm J had plasma viremi a. Seven of th ese had detectable p24 a n tige n with out acid hydrolysis and on ly o ne othe r was un covered by ac id trea tment. The same 10 pati ents who had plasm a viremia also had PEG-precipitabl e p24 a n tigen . Two o f th ese rea cti vities were lost after acid treatment. Norm al human plasma tre ated in thi s way was negati ve in each instance.
Discussion Better sur roga te mark ers of th e clinical stage of HI V infe ction and viral replic at ion ar e needed to evaluate the efficacy of antiretroviral dru g therapi es. A good surrogate m arker sho uld have the following cha racteristics: be prevalen t amo ng infected indi vidual s, correlate clinically, change with effective therapy or with clinical improvement, be reproducib le, and make biologic sense . In addition, it would be adva n tag eous to hav e a n assay that is easy to perform , rapid, relati vely inexpensive, a nd safe. Several such markers hav e been proposed, including C D 4 cell counts. pla sma viremia. immune co mplexes. a n d p24 a n tigene m ia. C D4 cell counts a re clearl y the best clin ical su rrogate mark ers found to dat e but are ex pensive a nd ca n be don e on ly in specially eq uipped la bora tori es [I , 2]. Plasm a viremi a has also been prop osed as a goo d surroga te m arker but is tim e co nsuming (up to 4 wee ks), lab oriou s, ex pe ns ive, a nd a ppro pria tely don e in specia lized laborato ries [3, 7-9] . One gro up has reported tha t th eir plasma culture assay overa ll had a sensi tivity of 4 5% a nd a spec ificity of 82.7% co m pa red with the tradition al p24 a ntige n assay [3]. The res u lts from our regular plasm a cultures show ed a sensitivity o f 48.4 % a nd a specificity of 6 7.9%. T he PEG-precipitate c u lture had a sensitivity of 56.0% a nd a specificity of 80% co mpar ed with p24 antigen testing; how ever. the PEG-precipitate p24 antigen test had a sensitivity of 59%an d a specificity
Downloaded from http://jid.oxfordjournals.org/ at UB der TU Muenchen on July 12, 2015
B
Go
it t is It
ii ii
~ ~ ~
25
a
ii ii
!!
p < .0 0 1
50
o
V i r emi a
Ce II Cu lt ur •
iili PEG a
P os i ti ve
V i rem i 8
767
Concise Communications
768
0
10 0
Cell Cu Iture Plasma
Ag
-
r-
~ PI asma Culture
7S
-
Ppl Ag
CD
>
PEG Ppt Culture
u I
o'" a.. so ~
t
P
