Secondary Logo

Journal Logo

Clinical Science

Elevated CD38 Antigen Expression on CD8+ T Cells Is a Stronger Marker for the Risk of Chronic HIV Disease Progression to AIDS and Death in the Multicenter AIDS Cohort Study Than CD4+ Cell Count, Soluble Immune Activation Markers, or Combinations of HLA-DR and CD38 Expression

Liu, Zhiyuan*; Cumberland, William G.; Hultin, Lance E.; Prince, Harry E.§; Detels, Roger*; Giorgi, Janis V.

Author Information
Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology: October 1, 1997 - Volume 16 - Issue 2 - p 83-92
  • Free


Activation of the immune system in HIV infection includes elevated expression of HLA-DR and CD38 antigens on CD8+ cells(1-3). These cells include anti-HIV-directed cytotoxic T lymphocytes, which seem to contribute to control of viral replication in the HIV-infected host even though they fail to eradicate HIV infection(4-6). Ultimately, viral persistence induces a state of chronic immune stimulation and increased lymphocyte turnover(7-10). The persistent immune activation that characterizes chronic HIV infection aggravates the immunodeficiency caused by HIV infection and contributes to HIV-induced pathogenesis(11-17).

We and others have previously reported that an elevated level of CD38 antigen expression on CD8+ cells is a predictor that HIV disease will progress to AIDS(18-21). Other studies have reported that several serologic markers of immune activation including neopterin,β2-microglobulin, soluble interleukin-2 (IL-2) receptor(sIL-2R), soluble CD8 (sCD8), and soluble tumor necrosis factor-α receptor type II (sTNFRII) are also elevated during HIV-infection and predict disease progression(22-26). No single study has simultaneously compared the predictive value of soluble markers of immune activation and increased expression of CD38 and HLA-DR antigens on CD8+ cells.

The purpose of this study was to compare available immune activation markers for their prognostic value in predicting the development of clinical AIDS or death among HIV-infected individuals. In this analysis, the prognostic power of elevated CD38 antigen expression on CD8+ cells was greater than that of any other activation marker and greater than that of CD4+ cell number and percent. These results suggest that CD38 measurements may be useful in the clinical management of HIV-infected individuals.


Study Population

This study was conducted at the Los Angeles center of the Multicenter AIDS Cohort Study (MACS)(27,28). Participants were homosexual or bisexual men who had been evaluated for lymphocyte subset levels and interviewed every 6 months since 1984. Immunologic measurements used for this analysis were made on blood samples obtained from January through June 1992. Flow cytometric measurements were made on fresh peripheral blood within 6 hours after collection in EDTA from 369 HIV-seropositive and 218 HIV-seronegative homosexual men. Of the former, 316 were AIDS-free and 53 had already been diagnosed with clinical AIDS when the measurements were taken. Two hundred and fifty-eight of the 316 AIDS-free men were seropositive at the beginning of the MACS in 1984 and had thus been HIV infected for at least 8 years when the measurements were made. The remaining 58 of the 316 AIDS-free men had seroconverted between 1985 and 1992 and had a median seropositive time of 4.2 years when the measurements were made. The data from all 316 patients were combined in the final data analyses because a preliminary analysis indicated that the prognostic values of the immune markers for AIDS development were similar between the 58 seroconverters and the 258 AIDS-free seroprevalent men.

The analysis of the prognostic value of the markers was limited to the 316 men who were AIDS-free at the time of the measurements. Except where noted, the evaluations of the prognostic value of the markers were done on 289 of the 316 men who had stored plasma collected at the same time that the flow cytometry measurements were done. Neopterin, β2-microglobulin, sIL-2R, and sCD8 levels were measured in these 289 HIV-seropositive AIDS-free men, 46 of the AIDS patients, and 31 of the HIV-seronegative controls; sTNFRII was measured in plasma of 243 of the 316 HIV-seropositive, AIDS-free men and 31 of the HIV-seronegative controls. The plasma tested for the seropositive AIDS-free and AIDS patients for each marker was based on the availability of samples stored in the repository from the time that the flow cytometric measurements were done; seronegative controls were chosen at random to obtain comparison values.

Study Outcome and Covariates

Among the 316 initially AIDS-free, HIV-seropositive men, 87 clinical AIDS cases and 45 deaths were observed during the 3 years of follow-up through March 1995. AIDS was defined by clinical presentations as described by the Centers for Disease Control and Prevention revised case definition(29). AIDS development and death were used in evaluating the prognostic value of the measured immunologic markers. The median follow-up times for men who remained AIDS-free and those who developed AIDS were 2.8 and 1.3 years, respectively. Diagnoses were 28, Kaposi's sarcoma; 20, Pneumocystis carinii pneumonia; 13, wasting syndrome as a result of HIV; 7, non-Hodgkin's lymphoma; 5, cytomegalovirus disease or retinitis; 5, candidiasis; 4,Mycobacterium avium infections; 4, other opportunistic infections; and 1, HIV-related encephalopathy. Data on antiviral treatments (i.e., zidovudine, dideoxyinosine, and dideoxycytosine) were collected semiannually. Other drugs were not Food and Drug Administration-approved at the time the measurements were made and were not considered in the analysis. Three groups of antiviral treatments, chosen because of the reported short-term effects of these antiviral drugs on immunologic markers(30), were classified: no use ever of any of the treatments (n= 154), use of one or more of the drugs for no more than 6 months(n= 37), or use of one or more of the drugs for >6 months (n= 125). Data on sexually transmitted diseases (STD) were also collected semiannually. Three groups of STD (gonorrhea and syphilis) occurrence were classified: no history of STD (n = 62), STD within the last 5 years (n = 25), or STD before the last 5 years (n= 229).

Flow Cytometric Measurements

Analyses were done on a FACScan flow cytometer (Becton Dickinson Immunocytometry Systems, San Jose, CA, U.S.A.) equipped with the standard filter setup. Relative fluorescence intensity (RFI) values were standardized over the course of the study, and flow cytometric measurements of lymphocyte subsets were made as described previously(4,31-33). All monoclonal antibodies were purchased from Becton Dickinson Immunocytometry Systems. Expression of CD38 and HLA-DR antigens on CD8+ T cells was measured on whole blood as RFI using anti-HLA-DR fluorescein isothiocyanate (FITC) and CD38 phycoerythrin (PE) in combination with gating on CD8 PerCP to select CD8bright cells(4,31). The CD38 RFI measurements were converted into the median number of molecules of CD38 per cell by multiplying by 41, the approximate number of PE molecules detected per RFI channel on our flow cytometer(34,35). This conversion was not done for the number of HLA-DR molecules per CD8+ cell because the number of molecules detected per RFI channel for FITC was not determined.

Other flow cytometric measurements made on the anti-HLA-DR FITC, CD38 PE, and CD8 PerCP-stained blood included the percentage and number of CD8+ T cells that were CD38+ using an arbitrarily assigned threshold of CD38 RFI >100 to score cells as CD38+(36). Measurements were also made to obtain the percentages and numbers of CD8+ T cells that were CD38-HLA-DR-, CD38+HLA-DR+, CD38-HLA-DR+, and CD38+HLA-DR- using arbitrarily assigned thresholds of 100 for CD38 RFI and 15 for HLA-DR.

Measurements of Serum Markers

Neopterin was measured by radioimmunoassay (RIA; Henning Berlin GMBH, Postfach, Berlin, F.R.G.). β2-Microglobulin was measured by microparticle enzyme immunoassay (Abbott Laboratories, Abbott Park, IL, U.S.A.). Enzyme-linked immunoabsorbent assays (ELISA) were used to measure sIL-2 R (T Cell Diagnostics, Cambridge, MA, U.S.A.), sCD8 (T Cell Diagnostics) and sTNFRII (R&D Systems, Minneapolis, MN, U.S.A.).

Statistical Analysis

The Cox proportional hazards model(37) was used to evaluate the prognostic value of single measurements of the immunologic markers. The χ2 statistics were calculated as twice the change in the log partial likelihoods, and all had one degree of freedom. Continuous independent variables were used either as continuous variables or were categorized into four levels using group sizes dictated by commonly used CD4+ cell categories (>500, 350-500, 200-349, and <200). The follow-up times for AIDS as the outcome were calculated from the time of the measurements of the study variables to the time of AIDS diagnosis or last visit for AIDS-free men. The follow-up times for death as the outcome were calculated from the time of the measurements to the time of death or last contact. Cox proportional hazards models were performed using the BMDP statistical package(38). Differences of the measured immunologic markers between the groups (Table 1) were tested using Wilcoxon rank sums.

Baseline values for immunologic measurements


Immunologic Measures in the Study Group

Table 1 shows the lymphocyte subset levels and serologic marker levels in the three groups of MACS subjects: HIV-seronegative, AIDS-free, and AIDS. The median levels for all the immune activation markers were significantly higher (p < 0.0001) in the AIDS-free men at baseline compared with the HIV-seronegative controls. CD38 on CD8, as well as neopterin, β2-microglobulin, and sTNFRII, were significantly higher (p < 0.0001) in the subjects who had clinical AIDS compared with those who were AIDS-free; sIL-2R (p = 0.067), sCD8 (p = 0.331), and HLA-DR expression on CD8+ cells (p = 0.147) did not differ significantly between these two groups.

Alternate ways of measuring CD8+ cell activation were also evaluated, and five of these measurements are included in Table 1. These measurements use cursor placements on the flow cytometer and report percentages of "positive" or "negative" cells. The values for the percentage of CD38+CD8+ T cells, CD38+HLA-DR-CD8+ T cells, and CD38+HLA-DR+CD8+ T cells followed virtually the same pattern across the groups as the CD38 on CD8 measurements (i.e., progressively higher levels in the AIDS-free seropositive and AIDS groups). The median percentage of CD38-HLA-DR-CD8+ T cells, which are considered resting cells, followed an inverse trend, with progressively lower values across the groups. Values for CD38-HLA-DR+CD8+ T cells were higher in AIDS-free HIV-infected subjects than in seronegative controls or AIDS patients, in keeping with our prior observation that these cells are favorable predictors of future stable CD4+ cell levels in HIV disease(36).

As illustrated in Figure 1, for representative donors from each of the three study groups, CD38 expression showed a continuous pattern of fluorescence brightness on each donor. For each sample, the median relative brightness of CD38 staining on the CD8+ cells was measured in RFI units, then converted to the median number of molecules of CD38 expressed per CD8+ cells, as described in the Methods section. In this article, CD38 on CD8 refers to the median number of CD38 molecules expressed per CD8+ cell. For the representative donors shown in Figure 1, the representative AIDS-free, HIV-seropositive subject had a fivefold elevation in the CD38 on CD8 measurement compared with the HIV-seronegative control; the person with AIDS had a CD38 on CD8 value that was a further fivefold higher. HLA-DR expression on CD8+ cells, which was measured with a FITC-conjugated monoclonal antibody (mAb), is expressed as RFI in this article. As shown in Table 1, levels of HLA-DR were elevated in the seropositive AIDS-free subjects compared with HIV-seronegative controls but were not further elevated in AIDS subjects.

FIG. 1
FIG. 1:
Examples of expression of CD38 antigen on CD8+ T cells detected with CD38 phycoerythrin (PE) monoclonal antibody in three study subjects. The median relative fluorescence intensity (RFI) of CD38 antigen expression on the CD8+ T cells was determined by gating on CD8bright cells. The median number of molecules of CD38 per CD8+ T cells (CD38 on CD8) was then estimated by multiplying by a factor for the number of PE molecules detected per channel. On the flow cytometer used for this study, that factor was 41. The CD4+ cell values of the subjects shown were 572, 281, and 279/mm3 for the seronegative control (left), AIDS-free seropositive (middle), and AIDS subject (right), respectively.

Prognostic Value of Immunologic Variables Analyzed as Continuous Variables

The Cox proportional hazards model was used to calculate the significance of the markers with AIDS development as the outcome. The results are shown in Table 2. The prognostic value of each marker was evaluated by the χ2 improvement achieved by adding each marker into the Cox model. The larger the χ2 statistics, the greater the significance of the variable with respect to prognostic value; χ2 values of 6.63, 10.83, 15.13, and > 15.13 were significant at p = 0.01, 0.001, 0.0001, and <0.0001 levels, respectively. Among all the markers, CD38 on CD8(χ2 = 133) provided the best fit to the data, with a much higher χ2 value than neopterin (χ2 = 54),β2-microglobulin(χ2 = 42), sIL-2R (χ2 = 33), or sCD8(χ2= 3.2). In a separate analysis because of smaller sample size for sTNFRII, CD38 on CD8 (χ2 = 117) provided a better fit to the data than did sTNFRII (χ2 = 70) or neopterin (χ2 = 60).

A number of other interesting comparisons were made in the analysis of the 289 men shown in Table 2. CD38 on CD8 had a higher χ2 value than CD4+ cell number(χ2= 77) or CD4+ percent (χ2 = 86). Theχ2 value for HLA-DR RFI (χ2 = 8.5) was much smaller than that of the other activation markers and became nonsignificant(p > 0.05) after adjustment for CD4+ cell level. Mean CD38 on CD8 (χ2 = 128) had a slightly lower χ2 value than the median CD38 on CD8. Finally, the placement of cursors to report results as a percent of CD8+ cells that were CD38+2 = 121) or that reacted with either one or both CD38 and HLA-DR mAbs (e.g., CD38+HLA-DR+2 = 85) resulted in χ2 values that were lower than for the median CD38 on CD8 measurements.

Chi-square statistics for improvement of the fit of the data to AIDS development in the initially AIDS-free HIV-infected subjects

Absolute numbers of each CD8+ T-cell subset gave χ2 values that were smaller than the χ2 values obtained with the corresponding percentages (data not shown). If the CD38 on CD8 measurement was in the model, no other immune marker contributed much to predicting AIDS as an outcome; thus, after adjustment for CD38 on CD8, theχ2 values of neopterin, β2-microglobulin, and sIL-2R as well as that for both CD4+ cell absolute number and CD4+ cell percent were reduced, with none of these χ2 values >6.63 (p = 0.01). In contrast, the χ2 values for the fit to the model of CD38 on CD8 remained highly significant after adjustment for CD4+ absolute number or CD4+ cell percent (χ2 = 58 and 53, respectively).

Relative hazards for clinical AIDS development within 3 years of the marker measurement in the initially AIDS-free subjects

Prognostic Value for AIDS Development of Immunological Markers Categorized into Four Groups

CD4+ cell numbers, CD38 on CD8, and neopterin (as the most predictive of the soluble markers of immune activation) were selected for further analysis in the 289 HIV-seropositive initially AIDS-free men for whom all three markers were measured. Groups were classified by progressive abnormalities in immunologic markers(Table 3), with group sizes determined by the number of men whose CD4+ cell numbers at the time of the measurement stratified them into four commonly used categories for CD4+ cell number, i.e., >500, 350 to 500, 200 to 349, and <200/mm3. These group sizes were also used for classification of the study participants according to CD38 on CD8 and neopterin. Use of same-size groups allowed the relative risks of the markers to be compared. Furthermore, the use of group sizes dictated by commonly used categories for CD4+ cell numbers provides a familiar reference point for comparing the predictive value of the markers with each other.

Seventy-nine of the 289 (27%) initially AIDS-free men developed AIDS during the ~3 years (March 1992-March 1995) of follow-up. Relative hazards and 95% confidence intervals for development of AIDS during follow-up are shown in Table 3. A dose-response relation between AIDS risk and progressively more abnormal levels of markers of CD4+ cell number, neopterin, and CD38 on CD8 was observed in univariate analysis. As expected, more progression to AIDS was observed in subjects in the two lower categories of CD4+ cells, with men in the most abnormal strata (CD4 count <200/mm3)~14 times more likely to develop clinical AIDS than the men in the reference group (>500/mm3). Elevated neopterin also increased the risk of progression to AIDS, with men in the most abnormal strata (≥ 19.6 nmol/L) about nine times more likely to develop clinical AIDS than the men in the reference group (<19.6 nmol/L). β2-Microglobulin, sTNFRII, and sIL-2R followed the same trend as for neopterin (data not shown). For CD38 on CD8, a continuous and marked dose-response relationship was observed across the groups. Risks were significantly different from those of the reference group (CD38 on CD8 < 2470 molecules) for all three categories of elevated CD38 on CD8 measurements; relative hazards(95% confidence interval [CI]) were 5.0 (range, 1.7-15), 12.3 (range, 4.3-36), and 41.4 (range, 15-117) for progressively more elevated CD38 on CD8 measurements of 2470 to 3899, 3900 to 7249, and > 7250 molecules per CD8+ cell (Table 3). If percent of CD8 + cells that were CD38+ was used in the model, relative hazards for the groups were 5.7 (range, 1.9-17), 11.9 (range 4.1-34), and 39.3 (range, 14-111) for measurements of 29.7% to 42.5%, 42.6% to 63.4%, and ≥63.5%, respectively.

Adjustment for CD4+ cell number reduced the relative hazards for predicting AIDS development associated with the CD38 on CD8 and neopterin measurements, but they remained significant (95% CI excluded 1.0) at all three strata of elevated CD38 on CD8 (hazards were 4.0, 7.6, and 18.4, respectively) and at the two highest strata of neopterin measurements (hazards were 2.8 and 4.3). After adjustment for CD38 on CD8, the relative hazards for other immunologic measurements were no longer significant. This was true for all the immune activation markers, including neopterin (Table 3), β2-microglobulin, sIL-2R, and sTNFRII (not shown), as well as CD4+ cell number (Table 3) and percent (not shown).

Relative hazards for death with an AIDS diagnosis within 3 years of the marker measurement in the initially AIDS-free subjects

Prognostic Value of Immunologic Markers for AIDS-related Death

Among the initially AIDS-free HIV-seropositive men in whom soluble markers were measured (n = 289), the relative hazards for predicting AIDS-related death (Table 4) were similar to the hazards for these markers in predicting AIDS development (Table 3). Again, CD38 on CD8 was more predictive for AIDS-related death in the three highest strata (hazards were 8.6, 15.8, and 69.3, respectively) compared with decreased CD4+ cell levels where only the two lowest strata excluded 1.0 (hazards were 3.5 and 18.2). After adjustment for CD4+ cell levels, CD38 on CD8 remained highly predictive of death.

Contribution of CD38 on CD8 Measurements to Predicting Outcome at High, Intermediate, and Low CD4+ Cell Levels

A further analysis of measurements from all 316 men on whom CD38 on CD8 and CD4+ cell measurements were made indicated that the contribution of CD38 on CD8 to outcome was especially marked in men with CD4+ cells >500 mm3. Among this group, men with CD38 on CD8 ≥ 3900 (the highest two strata of CD38 on CD8 in Table 4) were 10.6-fold more likely to develop AIDS than those with CD38 on CD8 < 3900. In contrast, at CD4+ cell counts between 200 and 500/mm3, the relative risk of developing AIDS for men with CD38 levels ≥ 3900 versus those with CD38 levels < 3900 was 7.9 versus 2.2. A small but not significant effect was also apparent in patients with CD4+ cell levels <200/mm3(25.8 vs. 16.7). Elevated CD38 on CD8 measurements also contributed more at high than at low levels of CD4+ cell numbers as a predictor of progression to death (data not shown).

Adjustment for Potential Confounding Effects of Antiviral Treatments and Sexually Transmitted Diseases on the Association Between CD38 on CD8 Levels and AIDS Development

Antiviral drug treatments could lead to a decrease in CD38 antigen expression, whereas STD could activate the immune system, leading to increased CD38 antigen expression. Either effect could confound the association between CD38 on CD8 expression and HIV disease progression. To examine the effect of antiviral drug treatments, these associations were considered in three groups as described in the Methods section; two indicator variables were generated for men who used one or more of the drugs for only 6 months and men who used them for >6 months. After adjustment for antiviral treatments, the relative hazards for AIDS associated with CD38 on CD8 levels were 4.8, 11.8, and 37.4, respectively, very similar to the unadjusted relative hazards(5.0, 12.3, and 41.4, respectively) in Table 3. An analogous analysis adjusting for STD was also done. The adjusted relative hazards (5.0, 12.0, and 40.4, respectively) were almost identical to the unadjusted relative hazards. Thus, even though transient changes in CD38 levels may have occurred in our subjects as a result of the effects of antiretroviral drug therapy or STDs, the similar hazards in the unadjusted and adjusted analyses indicate that antiviral drug treatments and STD were not likely to have confounded the association between elevated CD38 on CD8 expression and HIV disease progression observed in this study.


In the current study, which compared measures of immune system activation as predictors of progression of chronic HIV infection to AIDS and death, flow cytometric measurement of CD38 on CD8 was a stronger predictor of outcome than levels of soluble markers of immune activation, including neopterin, β2-microglobulin, sTNFRII, sIL-2R, and sCD8, measured by ELISA or RIA. Once CD38 on CD8 was known, none of the other markers provided predictive value for AIDS development or death. CD38 on CD8 was also a stronger predictor of outcome than CD4+ cell number or percent, which are markers of immunodeficiency. CD38 on CD8 was also compared in this study with additional flow cytometric measures of CD8+ T-cell activation, including several made using CD38 and anti-HLA-DR antibodies, and was found to be a stronger predictor of outcome than any of these.

The CD38 on CD8 measurement and its close corollary, the CD38+CD8+ percentage, are the only two CD8+ cell subset markers that, to our knowledge, have consistently shown predictive value for HIV disease progression(18-21). For example, in our previous studies on alterations in CD8+ cell subsets in HIV-infected MACS participants, we reported that elevated levels of CD57+CD8+ and HLA-DR+CD8+ cells did not have predictive value at all for AIDS development, and after adjusting for CD4+ cell level, the predictive value of elevated proportions of CD62L-CD8+ cells also became nonsignificant(19). We and others have used complex flow cytometry panels to provide insight into the pathogenesis of HIV disease (39; reviewed in 40). Our study extends a recent cross-sectional study that reported changes in antigen density, measured as RFI, on numerous leukocyte subsets in HIV disease, although that study did not assess whether these alterations were prognostic for the development of AIDS(41).

In a clinical setting, the most useful markers are those that have prognostic value that supplements that provided by commonly used markers such as CD4+ cell number. CD38 on CD8 is a candidate for such a marker. Results presented here suggest that it may be the most informative single immune activation marker. Table 5, which is based on data reported in this article, provides a suggested range of low (<2500 molecules/CD8+ cell), intermediate (2500-3999), high (4000-7000), and very high (>7000) CD38 values. This is a tentative classification that can be modified as values from other cohorts become available.

Suggested categories of CD38 on CD8 measurements for clinical applications

Although the CD38+CD8+ percentage and CD38 on CD8 measurements had similar χ2 values (χ2 = 121 and 133, respectively) and almost identical relative risks in our model, we prefer the latter measurement because it reflects the fact that the distribution of CD38 expression on CD8 cells is continuous. As shown in Figure 1, CD38 expression is heterogeneous and does not show a clear negative or positive population in either HIV-infected or uninfected control donors. The method of fluorescence quantitation that we applied in this study, which uses 1:1 conjugates of PE:CD4 and PE:CD38 and takes advantage of the tight conservation of CD4 antigen at ~50,000 molecules per CD4+ lymphocyte, allows standardization of CD38 on CD8 measurements across instruments and laboratories(34,35). Although such methods are theoretically feasible for CD38+CD8+ percent measurements, we have been unable to develop an adequate approach for such standardization (unpublished results).

The calculation of number of molecules per cell in this study was made by first determining that the number of PE molecules detected per channel on our flow cytometer was 41, as described elsewhere(35). This value was used to convert the median CD38 RFI on each sample to number of molecules of CD38 per CD8 cell. If the samples analyzed in this study had been run on a different flow cytometer, the RFI measurements might have been different but the number of molecules reported on each sample should have been similar. This is because the value determined on each flow cytometer for the number of PE molecules detected per RFI channel corrects for differences between instruments.

This study was done in MACS participants, most of whom had been HIV infected for 8 years when the measurements were taken and whose risk of acquiring HIV infection was through homosexual transmission. Studies of the prognostic value of CD38 on CD8 in other cohorts-including women and those who have acquired HIV by intravenous drug use-are needed to verify the prognostic power of the marker in these groups. There was a high rate of progression to AIDS and death in our cohort, as would be expected since most participants had been infected for 8 or more years; therefore, additional studies on subjects measured shortly after infection would be of value. Clinical follow-up in our study for development of AIDS or death was for 3 years after the markers were measured; follow-up for a longer period would also be of value. CD38 on CD8 also needs to be evaluated as a prognostic marker in infants and children. Although newborns express high levels of CD38 on their T cells, levels decrease in uninfected children within a few months after birth, so that it is possible this marker could also be useful in pediatric HIV infection after the first year of life(42,43).

In adults, intercurrent infections cause transient increases in CD38 expression in some herpes-type viral infections, but these increases resolve within a few weeks or months (J. V. Giorgi, unpublished observations). In this study, we could not address whether transient elevations in CD38 on CD8 occurred in our cohort due to STDs because only a single measurement was taken and no STDs concurrent with the measurements occurred. However, adjusting for past STDs in the current analysis did not affect the relative hazards for AIDS development of the CD38 on CD8 measurements.

There was a high rate of Kaposi's sarcoma diagnosed in the MACS, raising concern that our conclusions might not be generalizable to other cohorts in which Kaposi's sarcoma was not a frequent outcome. However, in a separate analysis of the data (not shown), we found that for those whose initial AIDS diagnosis was Kaposi's sarcoma and for those whose initial AIDS diagnosis was an opportunistic infection, CD38 on CD8 was a stronger predictive marker for progression to AIDS than CD4+ cell number or any of the other markers. In addition, for both Kaposi's sarcoma outcomes and Pneumocystis carinii pneumonia outcomes, median CD4+ cell number at the time of the measurement was 284/mm3 and median time from the measurement to an AIDS diagnosis was 1.4 years. The similarity of the two groups in these parameters suggests that the prognosis of the two groups after an AIDS diagnosis might also be similar. Thus, our conclusions are most likely also applicable to other HIV-infected cohorts in which Kaposi's sarcoma is not as frequent an outcome.

The reason that elevated CD38 expression on CD8+ cells is predictive of progression of HIV disease to AIDS and death is unknown. CD38 is an ectoenzyme that hydrolyzes nicotinamide adenine dinucleotide (NAD+) to nicotinamide and ADP-ribose(42). An increase in the level of CD38 antigen expression is a marker of cellular activation(42,44-46). In fact, CD38 is the first in a cascade of three enzymes whose expression is simultaneously induced when T cells are activated(47). These enzymes provide an extracellular salvage pathway for purines. The purine degradation products, including adenine, are then presumably transported into the cell, possibly to assist the cell in its rapid metabolism(47). Whether this is related to the defective de novo ribonucleotide synthesis by T cells that has been reported in HIV-infected people is unknown(48).

The cause of CD8+ cell activation in HIV infection probably at least in part reflects the persistent but unsuccessful attempt of the host's immune response to clear HIV-infected cells(9). We have hypothesized that the chronic immune stimulation in HIV-infected individuals leads to excessive CD8+ cell division and ultimately to immunologic exhaustion(10). CD8+ cells with high-level expression of CD38 antigen are prone to undergo apoptosis in vitro, and this could also contribute to the association with poor outcome(49). Alternatively, high levels of expression of CD38 on CD8+ cells could reflect an immune response to HIV which contributes to destruction of HIV-infected CD4+ cells, because anti-HIV-directed cytotoxic T cells express CD38(4). Other functions have been associated with the CD38 molecule, but how these activities may play a role on the CD8+ cells in HIV infection is not obvious(42,45,50,51).

The use of laboratory markers as predictors of disease progression has entered a new era as a result of the availability of effective antiretroviral drugs. CD38 has already been shown to respond to effective suppression of viral replication mediated by antiretroviral drugs; decreases were observed with zidovudine, and levels of CD38 expression on CD8+ cells fell over the first 24 weeks of protease inhibitor therapy(30,52,53). The prognostic value for AIDS development and death demonstrated in our study for the CD38 on CD8 marker suggests that it should be useful for evaluating the extent of immune dysfunction before the initiation of therapy and could be valuable for determining drug regimens and intensity of treatment that are optimal for a given patient. Additionally studies are being conducted to examine the predictive power of combinations of virologic and immunologic measurements taken at different times during the follow-up period for this cohort.

Acknowledgments: We thank Dr. John L. Fahey and Najib Aziz for the neopterin and β2-microglobulin measurements; Ms. Patricia Hultin, Ms. Mary Ann Hausner, and Mr. Jose Matud for assistance with the conduct of the study and their contribution to data analysis; and Drs. Donna Mildvan and Alan Landay for critical review of the article. This study was supported by National Institutes of Health awards Al-30540, Al-37613, Al-28697 (UCLA Center for AIDS Research, CFAR), and CA-16042 (UCLA Jonsson Comprehensive Cancer Center).


1. Salazar-Gonzáles JF, Moody DJ, Giorgi JV, Martinez-Maza O, Mitsuyasu RT, Fahey JL. Reduced ecto-5'-nucleotidase activity and enhanced OKT10 and HLA-DR expression on CD8 (T suppressor/cytotoxic) lymphocytes in the acquired immune deficiency syndrome: evidence of CD8 cell immaturity. J Immunol 1985;135:1778-85.
2. Giorgi JV, Detels R. T-cell subset alterations in HIV-infected homosexual men: NIAID Multicenter AIDS Cohort Study. Clin Immunol Immunopathol 1989;52:10-8.
3. Prince HE, Jensen ER. Three-color cytofluorometric analysis of CD8 cell subsets in HIV-1 infection. JAIDS 1991;4:1227-32.
4. Ho H-N, Hultin LE, Mitsuyasu RT, et al. Circulating HIV-specific CD8+ cytotoxic T cells express CD38 and HLA-DR antigens. J Immunol 1993;150:3070-9.
5. Borrow P, Lewicki H, Hahn BH, Shaw GM, Oldstone MBA. Virus-specific CD8+ cytotoxic T-lymphocyte activity associated with control of viremia in primary human immunodeficiency virus type 1 infection. J Virol 1994;68:6103-10.
6. Koup RA, Safrit JT, Cao Y, et al. Temporal association of cellular immune responses with the initial control of viremia in primary human immunodeficiency virus type 1 syndrome. J Virol 1994;68:4650-5.
7. Pantaleo G, Graziosi C, Demarest JF, et al. HIV infection is active and progressive in lymphoid tissue during the clinically latent stage of disease. Nature 1993;362:355-8.
8. Embretson J, Zupancic M, Ribas JL, et al. Massive covert infection of helper T lymphocytes and macrophages by HIV during the incubation period of AIDS. Nature 1993;362:359-62.
9. Ferbas J, Kaplan AH, Hausner MA, et al. Virus burden in long-term survivors of human immunodeficiency virus (HIV) infection is a determinant of anti-HIV CD8+ lymphocyte activity. J Infect Dis 1995;172:329-39.
10. Effros RB, Allsopp R, Chiu C-P, et al. Shortened telomeres in the expanded CD28-CD8+ cell subset in HIV disease implicate replicative senescence in HIV pathogenesis. AIDS 1996;10:F17-F22.
11. Fahey JL, Giorgi J, Martinez-Maza O, Detels R, Mitsuyasu R, Taylor J. Immune pathogenesis of AIDS and related syndromes.Ann Inst Pasteur Immunol 1987;138:245-52.
12. Gowda SD, Stein BS, Mohagheghpour N, Benike CJ, Engleman EG. Evidence that T cell activation is required for HIV-1 entry in CD4+ lymphocytes. J Immunol 1989;142:773-80.
13. Zack JA, Arrigo SJ, Weitsman SR, Go AS, Haislip A, Chen ISY. HIV-1 entry into quiescent primary lymphocytes: molecular analysis reveals a labile, latent viral structure. Cell 1990;61:213-22.
14. O'Brien WA, Grovit-Ferbas K, Namazi A, et al. Human immunodeficiency virus-type 1 replication can be increased in peripheral blood of seropositive patients after influenza vaccination. Blood 1995;86:1082-9.
15. Staprans SI, Hamilton BL, Follansbee SE, et al. Activation of virus replication after vaccination of HIV-1-infected individuals. J Exp Med 1995;182:1727-37.
16. Pantaleo G, Fauci AS. New concepts in the immunopathogenesis of HIV infection. Annu Rev Immunol 1995;13:487-512.
17. Stanley SK, Ostrowski MA, Justement JS, et al. Effect of immunization with a common recall antigen on viral expression in patients infected with human immunodeficiency virus type 1. N Engl J Med 1996;334:1222-30.
18. Levacher M, Hulstaert F, Tallet S, Ullery S, Pocidalo JJ, Bach BA. The significance of activation markers on CD8 lymphocytes in human immunodeficiency syndrome: staging and prognostic value.Clin Exp Immunol 1992;90:376-82.
19. Giorgi JV, Liu Z, Hultin LE, Cumberland WG, Hennessey K, Detels R. Elevated levels of CD38+CD8+ T cells in HIV infection add to the prognostic value of low CD4+ T cell levels: results of 6 years of follow-up. JAIDS 1993;6:904-12.
20. Liu Z, Hultin LE, Cumberland WG, et al. Elevated relative fluorescence intensity of CD38 antigen expression on CD8+ T cells is a marker of poor prognosis in HIV infection: results of 6 years of follow-up. Commun Clin Cytometry 1996;26:1-7.
21. Bofill M, Mocroft A, Lipman M, et al. Increased numbers of primed activated CD8+CD38+CD45RO+ T cells predict the decline of CD4+ T cells in HIV-1-infected patients. AIDS 1996;10:827-34.
22. Fuchs D, Jäger H, Popescu M, et al. Immune activation markers to predict AIDS and survival in HIV-1 seropositives. Immunol Lett 1990;26:75-80.
23. Moss AR, Bacchetti P, Osmond D, et al. Seropositivity for HIV and the development of AIDS or AIDS related condition: three year follow up of the San Francisco General Hospital cohort. Br Med J 1988;296:745-50.
24. Fahey JL, Taylor JMG, Detels R, et al. The prognostic value of cellular and serologic markers in infection with human immunodeficiency virus type 1. N Engl J Med 1990;322:166-72.
25. Giorgi JV. CD4 counts in relation to markers of immune activation. In: Janossy G, Autran B, Miedema F, eds. Immunodeficiency in HIV infection and AIDS. Basel: Karger, 1992:1-17.
26. Godfried MH, van der Poll T, Weverling GJ, et al. Soluble receptors for tumor necrosis factor as predictors of progression to AIDS in asymptomatic human immunodeficiency virus type 1 infection.J Infect Dis 1994;169:739-45.
27. Kaslow RA, Ostrow DG, Detels R, Phair JP, Polk BF, Rinaldo CR Jr. The Multicenter AIDS Cohort Study: rationale, organization, and selected characteristics of the participants. Am J Epidemiol 1987;126:310-8.
28. Detels R, Phair JP, Saah AJ, et al. Recent scientific contributions to understanding HIV/AIDS from the Multicenter AIDS Cohort Study. J Epidemiol 1992;2:S11-9.
29. Centers for Disease Control and Prevention. 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR 1992;41(No. RR-17):1-19.
30. Bass HZ, Hardy WD, Mitsuyasu RT, Wang Y-X, Cumberland W. Fahey JL. Eleven lymphoid phenotypic markers in HIV infection: selective changes induced by zidovudine treatment. JAIDS 1992;5:890-7.
31. Schmid I, Schmid P, Giorgi JV. Conversion of logarithmic channel numbers into relative linear fluorescence intensity. Cytometry 1988;9:533-8.
32. Schenker EL, Hultin LE, Bauer KD, Ferbas J, Margolick JB, Giorgi JV. Evaluation of a dual-color flow cytometry immunophenotyping panel in a multicenter quality assurance program. Cytometry 1993;14:307-17.
33. Calvelli T, Denny TN, Paxton H, Gelman R, Kagan J. Guideline for flow cytometric immunophenotyping: a report from the National Institute of Allergy and Infectious Diseases, Division of AIDS. Cytometry 1993;14:702-15.
34. Poncelet P, George F, Lavabre-Bertrand T. Immunological detection of membrane-bound antigens and receptors. In: Masseyeff RF, Albert WH, Staines NA, eds. Methods of immunological analysis. Weinheim, Germany: VCH Verlagsgesellschaft mbH, 1993:388-417.
35. Hultin L, Matud J, Giorgi JV. Tight regulation of CD4 antigen expression on lymphocytes of healthy donors used directly as a fluorescence quantitation standard. Cytometry 1996;(Suppl 8):90.
36. Giorgi JV, Ho H-N, Hirji K, et al. CD8+ lymphocyte activation at human immunodeficiency virus type 1 seroconversion: development of HLA-DR+CD38-CD8+ cells is associated with subsequent stable CD4+ cell levels. J Infect Dis 1994;170:775-81.
37. Cox DR. Regression models and life-tables. J R Stat Soc 1972;34:187-220.
38. Dixon WJ. BMDP statistical software manual. Los Angeles: University of California Press, 1988.
39. Giorgi JV, Boumsell L, Autran B. Reactivity of workshop T-cell section mAb with circulating CD4+ and CD8+ T cells in HIV disease and following in vitro activation. In: Schlossman SF, Boumsell L, Gilks W, et al. eds. Leucocyte typing V: White cell differentiation antigens, 1st ed. Oxford: Oxford University Press, 1995:446-61.
    40. Giorgi JV. Phenotype and function of T cells in HIV disease. In: Gupta S, ed. Immunology of HIV infection. New York: Plenum Press, 1996:181-99.
      41. Roederer M, Herzenberg LA, Herzenberg LA. Changes in antigen densities on leukocyte subsets correlate with progression of HIV disease. Int Immunol 1996;8:1-11.
      42. Malavasi F, Funaro A, Roggero S, Horenstein A, Calosso L, Mehta K. Human CD38: a glycoprotein in search of a function. Immunol Today 1994;15:95-7.
      43. Plaeger-Marshall S, Hultin P, Bertolli J, et al. Activation and differentiation antigens on T cells of healthy, at-risk, and HIV-infected children. JAIDS 1993;6:984-93.
      44. Hercend T, Ritz J, Schlossman SF, Reinherz EL. Comparative expression of T9, T10, and la antigens on activated human T cell subsets. Hum Immunol 1981;3:247-59.
      45. Funaro A, Spagnoli GC, Ausiello CM, et al. Involvement of the multilineage CD38 molecule in a unique pathway of cell activation and proliferation. J Immunol 1990;145:2390-6.
      46. Prince HE, Kleinman S, Czaplicki C, John J, Williams AE. Interrelationships between serologic markers of immune activation and T lymphocyte subsets in HIV infection. JAIDS 1990;3:525-30.
      47. Deterre P, Gelman L, Gary-Gouy H, et al. Coordinated regulation in human T cells of nucleotide-hydrolyzing ecto-enzymatic activities, including CD38 and PC-1: possible role in the recycling of nicotinamide adenine dinucleotide metabolites. J Immunol 1996;157:1381-8.
      48. Bofill M, Fairbanks LD, Ruckemann K, Lipman M, Simmonds HA. T-lymphocytes from AIDS patients are unable to synthesize ribonucleotides de novo in response to mitogenic stimulation. J Biol Chem 1995;270:29690-7.
      49. Prince HE, Jensen ER. HIV-related alterations in CD8 cell subsets defined by in vitro survival characteristics. Cell Immunol 1991;134:276-86.
      50. Howard M, Grimaldi JC, Bazan JF, et al. Formation and hydrolysis of cyclic ADP-ribose catalyzed by lymphocyte antigen CD38. Science 1993;262:1056-9.
      51. Dianzani U, Malavasi F. Lymphocyte adhesion to endothelium.Crit Rev Immunol 1995;15:167-200.
      52. Kelleher AD, Zaunders ACJ, Cooper DA. Alterations in the immune response of human immunodeficiency virus (HIV)-infected subjects treated with an HIV-specific protease inhibitor, ritonavir.J Infect Dis 1996;173:321-9.
      53. Autran B, Carcelain G, Li TS, et al. Positive effects of combined antiretroviral therapy on CD4+ T cell homeostasis and function in advanced HIV disease. Science 1997;278:112-116.

      CD38 antigen; Immune activation; CD8+ cells; Prognostic marker; Flow cytometry; Quantitative fluorescence; Neopterin; β2-Microglobulin; Soluble IL-2 receptor; Soluble CD8; Soluble TNF-α-receptor type II.

      © Lippincott-Raven Publishers.