Increases in soluble CD8 antigen in plasma, and CD8+ and CD8+CD38+ cells in human immunodeficiency virus type-1 infection

CLINICAL

1MMUNOLObk

4NL)

IMML~~OPA’IHOLO~~~

Vol. 63. No. 2, May. pp. 12&134,

1992

Increases in Soluble CD8 Antigen in Plasma, and CD8 ’ and CD8+CD38+ Cells in Human lmmunodeficiency Virus Type-l Infection MARY JANE YAGI,

Fo-NIAN

CHU,

JIAN

DONG

JIANG,

JOYCE

WALLACE,

PATRICIA

MASUN,

YUAN LIU, JOYCE CARAFA,

AND J. GEORGE BEKESI Department

of Neoplastic

Diseases

and

the T. J. Martell Laboratory Medicine, 1 Gustave L. Levy

for Cancer, Place,

Increases in plasma levels of soluble CD8 @CDS) antigen and expansion of the CD8+CD38+ lymphocyte compartment were early immunologic alterations frequently observed prior to detection of antibodies against human immunodeficiency virus type 1 (HIV-l) and diminution of CD4+ cells in subjects at risk to develop AIDS. These increases identified in the 49 seronegative homosexual men were manifest in all 164 homosexual subjects and 45 intravenous drug users (IVDU) positive for HIV-l antibodies (HIV-l+), 19 patients with ARC, and 29 AIDS patients. Augmentation of plasma sCD8 antigen correlated with increases in both CD8+ and CD8+CD38+ cells in HIV-l- homosexual men (r = 0.35, P < 0.013; r = 0.48, P < 0.0005;respectively) and the 258HIV-l’ subjects (r = 0.25, P < 0.0003; r = 0.33, P < 0.0001, respectively). In vitro examination of unstimulated peripheral blood lymphocytes from HIV-l+ homosexuals and IVDU confirmed the fivefold higher constitutive levels of cellular release of sCD8 antigen in these subjects compared to heterosexual controls. Inclusion of radiolabeled amino acids during the S-day culture period in the presenceor absenceof phytohemagglutinin resulted in negligible levels of radioactivity associatedwith the sCD8 antigen indicative of a lack of de rwvo synthesis. Throughout clinical progression to AIDS, sCD8 antigen levels continued to escalate relative to the numbers of CD8+ cells bearing CD38+ antigen. The data confirm the interrelationship between sCD8+ antigen and CD8+ and CD8+CD38+ cells. o 1%~ Academic PRSS. 1~.

New

Leukemia York. New

and AIDS York 10029

Research,

The Mount

Sinaz

School

of

9), as a result of immune excitation ( lo), and infection with Epstein-Barr virus (11, 12) or human immunodeficiency virus type-l (HIV-l; 13-171. Elevation in plasma sCD8 antigen appears to occur when the pool size of CD8’ cells increases (11, 13, 16), suggesting that sCD8 may be useful for monitoring this cellular compartment and/or may be a prognostic indicator in pathologic disease states. Longitudinal studies performed in our laboratory on a large cohort of HIV-l-infected subjects consistently detected significant increases in the absolute numbers of CD8’ cells, particularly those bearing the CD38* antigen. These phenotypic changes in the T cell compartment were manifest prior to detection of HIV-l genomic sequences in peripheral blood lymphocytes using the polymerase chain reaction and before the appearance of HIV-l antibody or antigens (18). Initial studies carried out on this cohort of homosexual men further indicated that elevations in CD8+ cells were accompanied by increases in plasma levels of the sol., uble form of the CD8 antigen, sCD8. This investigation was undertaken to clarify the interrelationship(s) between sCD8 antigen, CD8+ and CD8+CD38cells and HIV-l infection and clinical progression of disease to AIDS in homosexual men and intravenous drug users. MATERIALS

AND METHODS

INTRODUCTION

Study Population Mature peripheral T lymphocytes with cytotoxic and suppressor functions selectively express CD8 cell surLongitudinal assessment of HIV-l infection and face antigen which is spontaneously released from both lymphoproliferative dysfunctions has been carried out activated and resting CD8’ T cells (1, 2). Increases in on a cohort of homosexual men, intravenous drug users plasma levels of soluble CD8 (sCD8) glycoprotein were (IVDU), and heterosexuals at low risk. The 360 subfirst described in acute lymphoblastic leukemia pa- jects in this study initiated in 1984 were subdivided tients (1, 3) and has since been observed in patients into seven groups according to risk factors resulting with non-Hodgkin’s lymphoma (3), hairy cell leukemia from acknowledged life styles, absence (HIV-1 -) or (4), Hodgkin’s disease (51, chronic liver diseases (61, presence (HIV-1 +) of HIV-l antibodies as determined autoimmune diseases such as rheumatoid arthritis (7- by enzyme-linked immunosorbent assays (ELISA) and 126 0090-1229/92 $1.50 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

ELEVATED

PLASMA

sCD8

ANTIGEN

Western blot (WB) analyses, and clinical diagnoses of ARC or AIDS. The first group constituting the low-risk, HIV-lpopulation contained 54 female and male heterosexuals with an average age of 37 + 11 years. The second group consisted of 49 HIV-l - homosexual males with an average age of 37 + 7 years who were at high risk of acquiring HIV infection due to involvement in highrisk behavior. The third group consisted of 164 HIV-l+ homosexuals with an average age of 39 f 9 years. The fourth group of 45 subjects with an average age of 34 + 5 years included 15 female and 30 male IVDUs diagnosed as HIV-l +. The fifth group of 19 ARC patients with an average age of 39 2 8 years exhibited weight loss, anogenital lesions and lymphadenopathy at multiple sites, and were at immediate risk of developing AIDS. The overall average age of the 29 AIDS patients was 40.6 + 9.6 years which was not significantly different between Group 6 consisting of 16 Kaposi’s sarcoma patients (AIDS-KS) and the 13 patients with opportunistic infections (AIDS-011 in Group 7. In order to investigate the possible function/role of sCD8 antigen in HIV-l infection and disease progression, a retrospective/prospective study quantifying sCD8 antigen levels was carried out on this cohort using both frozen samples maintained in our library and fresh samples obtained in the course of our ongoing longitudinal study. sCD8 antigen determinations were carried out only on samples which had been or were being examined in parallel by other immunologic and virologic assays. All samples were coded to ensure that technical personnel were blinded as to the clinical condition and HIV-l status of the donor. All participants were volunteers and signed Mount Sinai Medical Center consent forms. This study was carried out in adherence to guidelines for human experimentation as set forth by the Mount Sinai Medical Center, New York City and State, and the U.S. Department of Health and Human Services. Determination

of Total HIV-l

Antibodies

Total HIV-l antibodies were determined by ELISA (E.I. DuPont, Wilmington, DE) (19, 20). Each serum sample was also examined by WB analyses (21) (E.I. DuPont) to ascertain humoral responses to specific HIV-l antigens including the gag proteins ~17, ~24, and ~55, and the envelope glycoproteins gp41 and gp120/160. Detection

of HIV-l

p24 Antigen

p24 HIV-l core antigen was assayed by sandwich enzyme-linked immunosorbent assays (E.I. DuPont). Results are expressed in picograms per milliliter serum.

IN HIV-l

Phenotype

127

INFECTION

Determination

of Lymphocyte

Subsets

Absolute numbers of lymphocytes with specific surface phenotypes were determined using fluorescein isothiocyanate-conjugated monoclonal antibodies against human HLe-1 (CD45), Leu 4 (CD3), Leu 3 (CD4), Leu 2 (CD8), and Leu 12 (CD19), and phycoerythrin-conjugated Leu 2 (CD8), Leu 17 (CD38), HLADR (Ia), and Leu M3 (CD141 (Becton-Dickinson Immunocytometry Systems, San Jose, CA). Staining was carried out by the whole blood technique in which 150 ~1 of blood is incubated with 20 pl of specific pairs of monoclonal antibodies for 20 min at 26°C followed by the addition of 2 ml of FACS lysing solution (BectonDickinson Immunocytometry Systems) and further incubation for 10 min. Cells were washed twice with 0.01 it4 phosphate buffer in 0.85% saline (PBS) and fixed in 1% paraformaldehyde in PBS, pH 7.4. Flow cytometric analysis was performed on at least 1.5 x lo4 lymphocytes per monoclonal antibody set using a FACStar flow cytometer (Becton-Dickinson Immunocytometry System) calibrated with CaliBrite beads and glutaraldehyde-fixed chicken red blood cells, and was defined using Simulset software for exclusion of dead cells, platelets, granulocytes, and erythrocytes (24).

Determination

of Soluble

CD8 (sCD8) Antigen

The soluble, cell-free CD8 antigen was quantitated by ELISA (Cell-Free CD8 test kit; T Cell Sciences, Inc., Cambridge, MA) utilizing C9 and B12 anti-CD8 antibodies, with detection by horseradish peroxidase and ortho-phenylenediamine substrate (12). Serum samples stored at - 20°C were thawed slowly, clarified by centrifugation at 2000g for 10 min at 5°C and assayed undiluted or following a 1:lO dilution. Duplicate or triplicate lo-k1 aliquots of sera or samples from in vitro cultures were analyzed. CD8 antigen reference standards provided in the kit were assayed in parallel with every sample determination to generate a standard curve of units of sCD8 per milliliter. The minimum detectable concentration of sCD8 antigen was 50 units/ ml sample. Serologic levels of sCD8 antigen are also reported as the ratio of sCD8 antigen to the absolute number of CD8 + cells/mm3 quantitated by phenotypic analyses of the original whole blood specimen. In Vitro Culture of PBL for Production Extracellular sCD8 Antigen and Lymphocyte Transformation

of

The extracellular sCD8 antigen was determined using peripheral blood lymphocytes (PBL) separated on Ficoll-Hypaque gradients, washed in 0.85% NaCl, and suspended in RPM1 1640 (GIBCO, Grand Island, NY), pH 7.4, containing 15% heat-inactivated, fetal bovine serum, 100 U penicillin/ml, and 100 kg streptomycin/ ml. Lymphocytes at 1 x lo6 cells/ml were incubated in

128

YAGI E’l il.1

the presence of 25 &i [“Hlleucine (44 CiimM; ICN Radiochemicals, Irvine, CA) or 0.1 mCi [35S]methionine (>lOOO Ci/mM; Amersham Corp., Arlington Heights, IL) with or without 1.5 kg phytohemagglutinin (PHA)/ml for 72 hr at 37°C in a humidified atmosphere containing 5% CO, (25). Culture supernatants were harvested and clarified of cells and particulate material by centrifugation at 9OOg for 10 min. The extracellular sCD8 antigen was then isolated by immunomagnetic separation using Dynabeads M-450 (Dynal Inc., Great Neck, NY) coated with sheep antimouse IgG directly coupled with murine monoclonal anti-CD8 antibody (T8 MAb; Coulter, Hialeah, FL) by incubation at 4°C for 16 hr on a rocker platform (26). The optimal ratio of Dynabeads M-450 to monoclonal anti-CD8 antibody as determined by titration was 120 mg M-450 to 120 kg anti-CD8 antibody in 4 ml of PBS containing 0.1% BSA (PBS-BSA). The conjugated beads were harvested with a Dynal magnetic particle concentrator (Dynal MPC-1) and washed three times with PBS-BSA. sCD8 antigen was separated by directly mixing conjugated Dynabeads with culture supernatants at a ratio of 50 ml supernatant to 60 mg of conjugated Dynabeads for 16 hr at 4°C with gentle agitation. Conjugated beads were harvested and washed in PBS-BSA three times using the Dynal MPC-1 magnetic concentrator. Immunoadsorbed sCD8 antigen was eluted by incubation of Dynabeads with 4 ml of 0.1 M glycine containing 0.15 M NaCl, pH 2.7, at 4°C. The efficiency

of sCD8 antigen recovery was greater thari nA’( Gli; ates were assayed for sCD8 antigen by ELISA as dt, scribed above, and 35S and ‘H associated radioactiv: ties were determined in a Beckman scintillation spet trophotometer. Lymphocyte transformation induced by PIriA was de. termined in parallel using a radioisotopic assay as prrsviously described (22i. The results of quadruplicate samples are expressed as average counts per min I cpm 1 per 1 y lo6 lymphocytes. Stat&lea1

Analysis

Data analyses were carried out on an IBM 370 computer using the Statistical Analysis System (SAS: SAS Institute, Inc., Cary, NC) and the BioMedical Data Package (BMDP; BMDP Statistical Software. Inc., Los Angeles, CA). Values obtained from the heterosexual control subjects served as the normal standard. Significance between groups was evaluated using the Student’s t test with correction for multiple intergroup comparisons using the Bonferroni method. RESULTS Serologic and CD4’ the Study Cohort

Lymphocyte

Characteristics

of

Subjects identified as having low risk for HIV-l infection were negative for HIV-l p24 antigen and antibodies (Table 1). The mean percentage and absolute numbers of CD4’ cells and the ratio of CD4+ to CD8”

TABLE 1 Characteristics of Study Cohort: Detection of HIV-l Antibodies, p24 HIV-l Core Antigen, and CD4’ C;ells - .’

No. positivesubjects(9%) Total

Group 1. Heterosexualcontrolb High-risk,asymptomatic 2. Homosexual 3.

Homosexual

4.

Heterosexual-IVDUb

High-risk, symptomatic’ 5. ARC

Western

P24

~24

antibody

antigen

0 (0)

0 (0)

0 (0,

i 757

0 (0)

4 (8)

2 (4)

1977

2 790

164 (100)

115 (70)

33 (20)

45

1628

t 547

45 (100)

43 (96)

6 (13)

19

1563

i 630

19 (100)

7 (37)

5 (26)

lymphocytes

blot

54

21872 507

49

2341

164

No.

6.

AID%KS

16

1648

2 485

16 (100)

8 (50)

5 (31)

7.

AID%01

13

1126

? 373

13 (100)

4 (31)

5 (38)

aValuesof CD8+ cellsare given in Table2. bGroupcontainedmaleandfemalesubjects. c All symptomaticsubjectsin Groups5, 6, and 7 werehomosexuals. * Value is significantly different,P < 0.0001,comparedto Group1. t Value is significantly different,P < 0.0001,comparedto Groups1 and2. $Value is significantly different,P < 0.0005,comparedto Groups3 and4. 0Value is significantly different,P < 0.001,comparedto Groups3 and4.

% CD4

’ cells

(absoluteNo./mm”t SD) 48 2 9 I 1050 ? 326, 47 t965 20 1421 21 (355

i -+ k +t t

11 t

1190 13 (210 6 (69

‘+ t + t

12 4051 11’; 320)t 8t 2140 5t.t

119)f lot 153v 4:$ 60Wl

Ratio CD4+/CD8’ II 2.0 _i 0.6

1.4 z 0,7* 0.4 +~ 0.4.: 0.4 f 0.2:

0.2 1 0.1 f 0.2 = 0.21 0.1 _’ 0.1t

ELEVATED

PLASMA

sCD8

ANTIGEN

IN HIV-l

129

INFECTION

In comparison, the frequency of p24 antibody in subcells obtained for this heterosexual control group are characteristic of HIV-lcontrols (Table 1). jects with clinical symptoms of disease were lower than that observed in the asymptomatic HIV-l+ individuals Among the 306 high risk subjects, 49 homosexuals were classified as HIV-l - using the Center for Disease (Table 1). This reduction appeared to coincide with inControl criteria because they did not possess HIV-l creases in the detection of p24 core antigen in the serum of ARC and AIDS patients. The percentage and antibodies as determined by ELISA and WB. However, antibody against the p24 antigen of the HIV-l core was absolute numbers of CD4+ cells were severely dedetected in 4 subjects, 2 of whom also possessed p24 pressed in these subjects, resulting in lo- to 20-fold core protein. Although the percentage and absolute reductions in the ratio of CD4+ to CD8+ cells in the groups (Table 1). numbers of CD4+ cells in this group were similar to three high-risk, symptomatic the heterosexual controls, their ratio of CD4+ helper to CD8+ cells was significantly reduced due to elevation Alterations in CD8+ Lymphocyte Subsets of the percentage and absolute numbers of CDS+ cells Significant expansion of the CD8+ cell compartment (Tables 1 and 2). The 257 subjects who were all HIV-l+ by ELISA and was detected in all high-risk groups, regardless of their WB analyses represented 209 high-risk, asymptomatic HIV-l seropositivity or levels of CD4+ cells (Table 2). individuals of which 164 were homosexual males and Prior to any other indication of immunologic alterof CD8+ cells to the total lym45 were IVDU subjects, and 48 patients with ARC or ations, the contribution AIDS (Table 1). The asymptomatic, homosexual males phocyte population rose to 36% in the HIV-l - homoand heterosexual IVDU subjects displayed similar, sig- sexuals from the 26% observed in heterosexual control subjects. Increases in the numbers of CD8+ cells were nificant decreases in the numbers of CD4+ cells with concomitant identical depressions of their CD4+ to even greater in HIV-l + subjects; CD8 + cells comprised CD8+ cell ratios. Although many of these individuals 55 and 58% of the total lymphocytes in the asymptomwere positive for p24 antibodies, 43 of 45 (96%) IVDU atic HIV-l + homosexual and IVDU subjects, respectively (Table 2). Although the absolute numbers of subjects possessed reactivities against the HIV-l virion core protein compared to 115 of 164 (70%) of the CD8+ cells were not further elevated with the progresdisease, subsequent diminution of high-risk homosexual males. Comparison of these se- sion of clinical rologic and phenotypic parameters within the IVDU CD4+ cells resulted in CD8 + cells representing 68group revealed no differences between the 30 male and 69% of the total lymphocyte population of ARC and 15 female individuals demonstrating the unanimity of AIDS-KS subjects. Concomitant with the expansion of the CD8+ cell the IVDU subject population.

Alterations HIV-l status

Group” 1. Heterosexual

in CD8+, CD8+CD38+

controlb

High-risk, asymptomatic 2. Homosexual

-

-

3. Homosexual

+

4. Heterosexual-IVDUb

+

High-risk, 5. AFK!

% Cells

TABLE 2 Cells and Plasma Levels of sCD8 Antigen

(absolute

CD8+

No./mm3

2 SD)

CD8+CD38+

26 f 6 (556 f 152)

sCD8 levels [U/ml plasma]

Ratio

of sCD8 to CD8+ cells (X10m3)

12 k 6 (261 -c 135)

292 2 136

0.6 2 0.3

542 2 429t

0.8 ? 0.9

36 (793 55 (1094 58 (936

2 2 2 ? + 5

14t 324) 14$ 526)t lO$ 318Yt

13 (304 36 (714 37 (598

2 2 + 2 k 2

9 232) 16$ 448)$ 14$ 285)$

66 (1076 66 (1114 56 (638

* 2 -t 2 k +

12*3 51OYt 11*+ 424jt 12$ 267)

41 (673 42 (726 42 (497

2 2 k ” f 2

17$ 402)$ 12.t 437)$ 13$ 243)$

1372

k 757$

1.6 k 1.6$

1343

* 736$

1.6 2 0.9$

1642 k 434$

2.4 2 2.4$

1778 k 477$

1.7 2 0.7$

1251 k 471t.

2.3 -+ l.l$

symptomatic’ +

6. AIDSKS

+

7. AIDS-01

+

a Number of subjects in each group ’ Group contained male and female c All symptomatic subjects in Groups * Value is significantly different, P t Value is significantly different, P $ Value is significantly different, P

is presented in Table 1. subjects. 5, 6, and 7 were homosexuals. < 0.001, compared to Group 3. < 0.0001, compared to Group 1. < 0.0001, compared to Groups 1 and 2.

130

YAGI

pool, cells phenotypically expressing both CDS and the CD38 antigen also increased significantly in HIV-l ’ subjects (Fig. 1 and Table 2). The portion of CD8 + CD38 + cells to the total population of CD8 T cells increased from 47% for the heterosexual controls and 38% for the HIV-l ~~ homosexual group to 65% for the asymptomatic and symptomatic HIV-l ’ subjects (range of 63-69% for Groups 3-7). The correlation between the percentage of CD8 + to CD8 + CD38 + cells among total lymphocytes was statistically significant for the HIV-l homosexual males (r = 0.588, P = 0.0001) as well as for the entire group of asymptomatic and symptomatic HIV-l+ subjects (1. = 0.472, P = 0.0001) but not for the heterosexual controls (r = 0.22, P = 0.12). Detection

of sCD8 Antigen

in Plasma

Quantification of sCD8 antigen in the plasma of the study cohort revealed that all groups of high-risk individuals displayed significant elevation of this glycoprotein and that the proportion of individuals with sCD8 antigen levels greater than two standard deviations above the mean of the heterosexual controls increased from 40.8% for the HIV-l ~ group to greater than 90% of all asymptomatic and symptomatic HIV-l + individuals. The average level of sCD8 antigen per milliliter plasma in the HIV-l ~ homosexual population was 1.8fold higher than that observed in heterosexual controls. A further 2.5fold increase in sCD8 antigen was

4oooy

E 1 .\,

demonstrable m asymptomatic HIV-1 aomosexua~~ high-risk and IVDU subjects compared to the HIV-l homosexuals. An additional escalation in sCD8 antigen was observed with clinical progression to ARC and AIDS-KS, although these increases were not present ::I AIDS-01 patients. Examination of sCD8 antigen levels relative to oath percentage and absolute numbers of CD8 r)i CD8 i- CD38 - ceils confirmed the observed interreiationship between plasma sCD8 antigen and lympho. cytes in the CD8’ compartment. Comparison of the ratios of sCD8 antigen to CD8 * cells also revealed the progressive increase in the relative levels of sCD8 antigen in subjects with ARC and AIDS compared to HIV-1 , asymptomatic individuals (Table 2). The relationship between the percentage and numbers of CD8’ cells and plasma levels of sCD8 antigen was already demonstrable and significant in high-risk HIV-l individuals 0” == 0.35, P = 0.01) and in asymptomatic HIV-l+ subjects (I. = 0.25, P = 0.00031 compared to heterosexual controls (r == 0.043, P !!.76i An even greater correlation was observed between plasma sCD8 antigen and CD8’CD38 + cells in both high-risk HIV-l and HIV-l + subjects (Fig. 2) Determination of the ratio of sCD8 antigen t.o the absolute numbers of CD8’ cells also demonstrated that the eievation in sCD8 antigen in the plasma was due to expansion of this T cell compartment as well as to a 25 to 4.5fold increase in the amount of sCD8 antigen attributable to each CD8 * cell (Table 21.

4000 A

3oooi

% 3000

3

2000

ABSOLUTE

CD8+

CELLS

/ mm3

FIG. 1. Relationship between the absolute numbers of CD8+ and CD8+CD38+ cells. (A) Heterosexual controls (No. 54, r = 0.39, P = 0.0047); (B) high-risk, HIV-lhomosexuals (No. 49, r = 0.57, P = 0.0001); and (Cl HIV-l+ homosexuals and IVDU subjects, and ABC and AIDS patients (No. 257, r = 0.78, P = 0.0001). Each symbol represents values of one individual. Total number of subjects (No.), correlation coefficients (r), and significance levels (P) are given for each group.

ELEVATED

PLASMA

I

3000

sCD8 ANTIGEN

i

-&-; nm-

IN HIV-l

131

INFECTION

q Cl

00

ABSOLUTE CD8+CD38+ CELLS / mm3 FIG. 2. Relationship between CD8+CD38’ cells and plasma levels of soluble CD8 (sCD8) antigen. (A) Controls (No. 54, r = 0.08, P = 0.59); (B) high-risk, HIV-lhomosexuals (No. 49, P = 0.48, P = 0.0005); and (C) HIV-l+ homosexuals and IVDU subjects, and ARC and AIDS natients (No. 257. r = 0.33. P = 0.0001). Each symbol renresents values of one individual. Total number of subjects (No.), correlation coefficients (r.), and significance levels (P) are given for each group.

In Vitro Evaluation

of sCD8 Released

by PBL

Levels of extracellular sCD8 antigen were quantitated following in vitro culture of Ficoll-Hypaqueseparated PBL obtained from 7 heterosexual controls and 9 asymptomatic HIV-1 + subjects consisting of 3 IVDU and 6 homosexual males. Plasma levels of sCD8 antigen and pool sizes of CD4+, CD8+, and CD8+CD38+ cells for these PBL donors were similar to those obtained for subjects in their risk classification as exemplified in Tables 1 and 2. The average 507 units of sCD8 antigen spontaneously released in vitro during the 72-hr incubation period from lo6 PBL of HIV-l+ subjects was five times greater than the average 99 units obtained from PBL of the heterosexual controls (Fig. 3). The levels of spontaneous sCD8 antigen released correlated with the absolute numbers of CD8+CD38+ cells (T = 0.84, P = 0.005) for the HIVl+ subjects but not for the controls (r = 0.52, P = 0.4). Constant exposure of PBL to PHA during the 3-day culture period increased extracellular sCD8 in the HIV-lcontrol cultures from the 99 units released by unstimulated PBL to an average of 1041 units/lo6 for PHA-stimulated PBL. In contrast, lo6 PBL from HIV1 + subjects following PHA treatment only released 581 units of sCD8. This reduction in sCD8 release following exposure to PHA correlated inversely with the increased level of spontaneous antigen release (r = -0.94, P < 0.0002) and the absolute numbers of CDS+CD38+ cells (r = -0.74, P < 0.03). However,

these relative levels of sCD8 release following PHA induction did parallel the degree of PHA-induced blastogenic transformation of PBL from the control and HIV-1 + subjects (Fig. 3). The levels of 35S or 3H radioactivity associated with sCD8 antigen released from cultures of unstimulated PBL from either control or HIV-l + subjects were negligible (O-40 cpm/106 cells). Exposure of PBL to PHA did not alter the incorporation of radiolabeled precursors and only 8-40 cpm were associated with sCD8 antigen derived from lo6 PHA-stimulated PBL. DISCUSSION

Immunologic evaluation of this cohort of 306 highrisk, HIV-land HIV-l+ subjects and ARC and AIDS patients identified alterations in the cellular immune systems of the host as a result of exposure to and infection with HIV-l. Phenotypic analyses revealed significant diminution of CD4+ cells in seropositive subjects and established that lymphocytosis of CD8+ cells is demonstrable prior to the detection of HIV-l antibodies. This early increase in CD8+ cells has been observed by other investigators (13, 14, 16-18, 27, 281, and these cumulative findings indicate that alterations in the CD8+ compartment are symptomatic of early HIV-l infection and progressive clinical deterioration. Parallel quantification of sCD8 antigen in the plasma of these high-risk subjects identified an association between this cell-free antigen and the numbers of

132 sCD8 ~.

I-- -.~ i 1400

PLASMA

ANTIGEN

3/ 1 CEL : - i-UNC.iIOh ~-~

PBL

~~-

W-A

SPONTANEOUS

/

F’il/i

I ,~) PliA

lNDUCE0

~-.-.-.---

I

1 4,‘;;

:~ i

-800

-? r.

i

‘0

-600

- 400 P< 03

HIV-1 [-] CONTROLS

FIG. 3. Levels of sCD8 antigen in the plasma or released phocytes (PBL) and degree of PHA-induced transformation deviation.

200

- --,

m

I~~~

0

HIV-1 [+I SUBJECTS

from unstimulated or phytohemagglutinin-stimulated in control and HIV-l seropositive subjects.

CD8+ and/or CD8+CD38+ cells. The high correlation between sCD8 antigen and its cellular source(s) provides added evidence that this antigen can be used as an indicator for expansion of the CDS+ compartment (16,171. It is probable that the observed interrelationships between CD8+ cells, their activation, and the release of sCD8 antigen may not be unique to HIV-l infection. Similar alterations in the CD8’ compartment have been observed in other pathologic conditions such as infection with Epstein-Barr virus (121, rheumatic diseases (7,8), or allergic disorders (10) and these phenomena may be the result of nonspecific responses of CDS+ cells to infection or immunologic assault. The cause and mechanism of increased sCD8 antigen release from CD8 + and/or CD8 + CD38 + cells are not yet established. Quantification of plasma sCD8 antigen as a function of the percentage or absolute numbers of CD8+ and CDS+CD38+ cells revealed that the amounts of sCD8 antigen attributable to individual cells were significantly heightened in HIV-l + subjects. In vitro evaluation of spontaneous PBL release of sCD8 antigen confirmed these findings. A greater than fivefold elevation in extracellular sCD8 antigen was detected in PBL cultures from HIV-l + subjects compared to HIV-l - controls. This increase could not merely be accounted for by augmentation of the numbers of CD8+ or CD8+CD38+ cells since their proportion was only two-fold higher in the whole blood of HIV-l + vs control subjects. Induction of PBL from control subjects with PHA resulted in the expected blastogenic transformation and also in significant increases in the

E ::

5L : :

i r--l

r

Error

peripheral bars represent

blood lymstandard

amounts of sCD8 antigen (7, 10, 12). However, quantification of extracellular sCD8 antigen in PHAtreated PBL cultures from HIV-l + subjects revealed significantly lower levels of antigen release. Whether this reduction and the relative lack of PHA-induced blastogenic transformation observed in these cell populations are interrelated is unknown. Both alterations may occur if the cells are already maximally activated and are refractory to subsequent mitogenic induction (7-10) or if HIV-l infection resulted in selection of CD8’ cells with impaired clonogenic potential (29). Inclusion of radioisotopic precursors indicated that the sCD8 antigen released into the supernatant in either unstimulated or PHA-stimulated heterosexual control PBL cultures was not synthesized de nouo during the 72-hr period. A similar lack of radioisotopic incorporation was observed in PBL cultures from HIVl+ subjects, suggesting that infection did not alter the rate of sCD8 antigen synthesis and/or the mechanism of sCD8 antigen maturation or release. This inability to radiolabel sCD8 antigen could be due to the metabolic inertia of PBL compared to the in u&o-adapted cell culture lines or virally transformed T cells transfected with CD8-cDNA used in previous studies (1, 2, 12, 20, 31). It is also possible that the inability to effectively radiolabel the sCD8 antigen may be due to the limitation of the ELISA assay to only identify the (Yand not the 8 glycoprotein moiety of the CD8 antigen (12,32,33). Additionally, the (Yglycoprotein is present as either a membrane-bound molecule which can be released from the cell surface by proteolytic cleavage or a soluble molecule which is secreted from the T cells

ELEVATED

PLASMA

sCD8 ANTIGEN

(30, 31). Although the predominant species is membrane bound, the secreted form appears to be more actively synthesized in. vitro and has been identified in culture supernatants as the entity metabolically labeled following short-term incubation with radiolabeled precursors (30-33). The information from these biochemical studies suggests that the unlabeled extracellular sCD8 molecule detected in the PBL cultures was more likely to be structurally similar to the membrane form. Biochemical analyses of sCD8 antigen identified in the plasma and PBL cultures from HIVl-infected subjects are currently in progress.

13.

14.

15,

16.

17.

REFERENCES 1. Fujimoto, J., Levy, S., and Levy, R., Spontaneous release of the leu-2 (T8) molecule from human T cells. J. Exp. Med. 159, 752766,1983. 2. Fujimoto, J., Stewart, S. J., and Levy, R., Immunochemical analysis of the released Leu-2 (T8) molecule. J. Enp. Med. 160, 116 124, 1984. 3. Pui, C.-H., Ip, S. H., Dodge, R. K., Carrabis, S., Brown, M., Crist, W. M., Berard, C. W., Kung, P., Dahl, G. V., and Murphy, S. B., Serum levels of CD8 antigen in childhood lymphoid malignancies: A possible indicator of increased suppressor cell activity in poor-risk patients. Blood 72, 10X-1021, 1988. 4. Ho, A. D., Grossman, M., Trumper, L., Pezzutto, A., Pralle, H., Boedewadt-Radzun, S., Papendick, U., God, B., Zwinger, T., and Hunstein, W., Clinical implications of increased plasma levels of CD8 in patients with hairy cell leukemia. Blood 75, 1119-1124, 1990. 5. Pui, C.-H., Ip, S. H., Thompson, E., Dodge, R. K., Brown, M., Wilimas, J., Carrabis, S., Kung, P., Berard, C. W., and Crist, W. M., Increased serum CD8 antigen level in childhood Hodgkin’s disease relates to advanced stage and poor treatment outcome. Blood 73, 209-213, 1989. 6. Wagner, F., Assemi, C., Lersch, C., Hart, R., and Classen, M., Soluble interleukin-2 receptor and soluble CD8 in liver cirrhosis and obstructive jaundice. Clin. Exp. Zmmunol. 82, 344-349, 1990. 7. Zielinski, C. C., Pesau, B., and Muller, Ch., Soluble interleukin-2 receptor and soluble CD8 antigen in active rheumatoid arthritis. Clin. Zmmunol. Zmmunopathol. 57, 7p82, 1990. 8. Symons, J. A., Wood, N. C., Di Giovine, F. S., and Duff, G. W., Soluble CD8 in patients with rheumatic diseases. Clin. Exp. Zmmunol. 80, 354-359, 1990. 9. Carpenter, A. B., Eisenbeis, C. H., Jr., Carrabis, S., Brown, M. C., and Ip, S. H., Elevated soluble CD8 in the synovial fluid from patients with rhematoid arthritis. J. Clin. Lab. Anal. 4, 337-341,1990. 10. Hsieh, K.-H., Decreased production of CD8 (T8) antigen after immunotherapy. J. Clin. Zmmunol. 9, 111-118, 1989. 11. Tomkinson, B. E., Wagner, D. K., Nelson, D. L., and Sullivan, J. L., Activated lymphocytes during acute Epstein-Barr virus infection. J. Zmmunol. 139, 38023807, 1987.

INFECTION

133

12. Tomkinson, B. E., Brown, M. C., Ip, S. H., Carrabis, S., and Sul-

ACKNOWLEDGMENTS This work was supported by NIH-NIAID Grants NOl-AI-52572 and NOl-AI-72660 from the Vaccine Program, and by the T. J. Martell Foundation for Cancer, Leukemia and AIDS Research. The authors express their appreciation to Mrs. Sophie Kurdziel for her contribution to this study and to Tomi for art work.

IN HIV-l

18.

19.

20.

21.

22.

23.

24.

25.

26.

livan, J. L., Soluble CD8 during T cell activation. J. Zmmunol. 142, 2230-2236, 1989. Agostini, C., Semenzato, G., Vinante, F., Sinicco, A., Trentin, L., Zambello, R., Zuppini, B., Zanotti, R., Siviero, F., Veneri, D., Foa, R., and Pizzolo, G., Increased levels of soluble CD8 molecule in the serum of patients with acquired immunodeficiency syndrome (AIDS) and AIDS-related disorders. Clin. Zmmunol. Zmmunopathol. 50, 146-153, 1989. Reddy, M. M., Lange, M., and Grieco, M. H., Elevated soluble CD8 levels in sera of human immunodeficiency virus-infected populations. J. Clin. Microbial. 27, 257-260, 1989. Chu, F., Tsang, P. H., Roboz, J. P., and Bekesi, J. G., Increased spontaneous release of CD8 antigen from CD8’ cells reflects the clinical progression of HIV-l infected individuals. In “V Int. Congress AIDS,” 431, 1989. Yagi, M. J., Carrabis, S., and Bekesi, J. G., Determination of soluble CD8 T-cell antigen (sCD8) by enzyme-linked immunosorbent assay. In “Manual of Clinical Immunology” (N. R. Rose, E. C. DeMarco, J. L. Fahey, H. Friedman, and G. M. Penn, Eds.), pp. 261-268, ASM Publications, Washington, DC, 1992. Lin, R. Y., Nygren, E., Valinsky, J., and Franklin, M., Levels of soluble CD8 antigen and circulating immune complexes in intravenous drug abusers: Relationships to HIV antibody serology. AIDS Res. Hum. Retrovir. 5, 655-661, 1989. Yagi, M. J., Joesten, M. E., Wallace, J., Roboz, J. P., and Bekesi, J. G., Human immunodeficiency virus type 1 (HIV-l) genomic sequences and distinct changes in CD8+ lymphocytes precede detectable levels of HIV-l antibodies in high-risk homosexuals. J. Infect. Dis. 164, 183-188, 1991. Sarngadharan, M. G., Popovic, M., Bruch, L., Schupbach, J., and Gallo, R. C., Antibodies reactive with human T-lymphotropic retroviruses (HTLV-III) in serum of patients with AIDS. Science 224, 506-508, 1983. Brun-Vezinet, F., Rouzioux, C., Montagnier, L., Chamaret, S., Gruest, J., Barre-Sinoussi, F., Geroldi, D., Chermann, J. C., McCormick, J., Mitchell, S., Piot, P., Taelman, H., Mirlangu, K. B., Wobin, O., Mbendi, N., Mazebo, P., Kalambayi, K., Bridts, C., Desmyter, J., Feinsod, F. M., and Quinn, T. C., Prevalence of antibodies to lymphoadenopathy-associated retrovirus in African patients with AIDS. Science 226, 453456, 1984. Tsang, V. C., Peralta, J. M., and Simons, A. R., Enzyme-linked immunoelectrotransfer blot techniques (EITB) for studying the specificities of antigens and antibodies separated by gel electrophoresis. In “Methods in Enzymology” (J. Langone and S. Timasheff, Eds.), Vol. 92, pp. 377-391, Academic Press, New York, 1983. Bekesi, J. G., Tsang, P. H., Wallace, J. I., and Roboz, J. P., Immunorestorative properties of isoprinosine in the treatment of patients at high risk of developing ARC or AIDS. J. Clin. Lab. Zmmunol. 24, 155-161, 1987. Tsang, P. H., Zanjani, M. D., Warner, N., and Bekesi, J. G., Restoration of impaired B and T lymphocyte subsets and functions in vitro by isoprinosine in prodromal homosexuals and AIDS patients. J. Clin. Lab. Zmmunol. 20, 159-165, 1986. Tsang, P. H., Sei, Y., and Bekesi, J. G., Isoprinosine induced modulation of T-helper cell subsets and antigen presenting monocytes (Leu M3 + Ia + ) resulted in improvement of T and B lymphocyte function in vitro in ARC and AIDS patients. Clin. Zmmunol. Zmmunopathol. 45, 166-176, 1987. Terhorst, C., Agthoven, A. V., Reinherz, E., and Schlossman, S., Biochemical analysis of human T lymphocyte differentiation antigens T4 and T5. Science 209, 520-521, 1980. Warlock, A. J., Sidgwick, A., Horsburgh, T., and Bell, P. R. F.,

134

YAGl

El

A,

The use of paramagnetic beads for the detection of MHC class and II antigens. Biotechniques 10, 310-315, 1991. 27. Caruso, A., Gonsales, R., Stellini, R., Scalzini, A.. Peroni, L., and Turano, A., Interferon-y marks activated T lymphocytes in AIDS patients. AIDS Res. Hum. Retrouir. 6, 899-904, 1990. 28. Giorgi, J. V.. and Detels, R., T-cell subset alterations in HIVinfected homosexual men: NIAID Multicenter AIDS Cohort Study. Clin. Immunol. Immunopathol. 52, 10-18, 1989.

31.

29. Pantaleo, G., Koenig, S., Baseler, M., Lane. H. A. S.. Defective clonogenic potential of CD8 ’ T patients with AIDS: Expansion in vivo of a CD3 + CD8 + DR ’ CD25T cell population. J. 16961704,199o.

32. Johnson, P., A human homoiog of the mouse CD8 molecuie, i ,yt 3- Genomic sequence and expression. Immunogenetics 26. 1’74 177. 1987. 33. Lrttman, D. R The structure of the CD4 and CD8 genes. il ITII :d Rrr Immunol 5. 561-584. 1987

Received

June

6, 1991;

accepted

with

revision

January

C., and Fauci, lymphocytes in nonclonogenic Immunol. 144,

30, 1992

l’.. Ledbetter. .i. r%. dind Kavathas, P.. ri se:reteo r~)rnl i the human lymphocyte cell surface molecule CD8 arise-; fron alternative splicing. Pmc M&l Acad. Scl USA 86. 99%IOO? 1989 31. Norment, A. h F,onberg. X. ILacy. E, , and Litirnx. i; I; . A ternatively spliced mRNA encodes a secreted form ot’ humzli: CD8o: Characterization of the human CD8a gene. .I imrnunoi 142, 3312-3319. 1989 Glblln,