All‐ Trans Retinoic Acid Inhibition of Anti‐CD3–Induced T Cell Apoptosis in Human...

1288

All-Trans Retinoic Acid Inhibition of Anti-CD3–Induced T Cell Apoptosis inHuman Immunodeficiency Virus Infection Mostly Concerns CD4 TLymphocytes and Is Mediated via Regulation of CD95 Ligand Expression

Zsuzsa Szondy, Herve Lecoeur, Laszlo Fesus, and Marie-Lise Gougeon

Department of Biochemistry, University Medical School of Debrecen,Debrecen, Hungary; Department of AIDS and Retroviruses, Institut

Pasteur, Paris, France

This study analyzes the influence of all-trans retinoid acid (tRA) on apoptosis of peripherallymphocytes from human immunodeficiency virus (HIV)–positive patients. tRA inhibits theex vivo apoptosis in T cells; a more potent effect was observed on activation-induced apoptosis.Phenotypic characterization of T cell subsets prevented from anti-CD3–induced apoptosis bytRA revealed a more potent effect on CD4 T cells. A central regulatory system for apoptosisis the CD95 system, and inappropriate induction of this pathway is thought to contribute toAIDS pathogenesis. In investigation of CD95-based apoptosis, tRA had no effect on acti-vation-dependent induction of CD95 on T lymphocytes, but it inhibited the induction ofCD95 ligand expression on anti-CD3–activated T cells. The previously reported in vivo effectof tRA inhibiting HIV-associated apoptosis and the present observations suggest that tRAcould be considered to down-regulate apoptosis associated with AIDS pathogenesis.

Physiologic cell death is primarily mediated through apop-tosis. One of the central regulatory systems for apoptosis is theCD95 receptor/ligand system [1, 2]. The important role playedby CD95-mediated death in T cell homeostasis was supportedby the finding that CD95 and the CD95 ligand (CD95L) aremutated in mouse strains with severe autoimmune diseases andlymphoproliferation [3, 4]. Thus, mutations of the CD95 mol-ecule in lpr mice and mutations of the CD95L in gld miceconstitute the first genetically defined syndromes of defectiveapoptosis. Recently, human counterparts of the lpr mutationin mice have been identified [5–7].

The CD95 molecule is a cell surface receptor, of the tumornecrosis factor receptor superfamily, that is readily expressedin peripheral T cells after activation. To induce apoptosis, CD95receptors on the cell surface have to oligomerize [8, 9]. CD95-mediated death depends on an apoptosis-sensitive phenotypethat is acquired during prolonged activation of T cells. Sensi-tivity or resistance toward CD95-induced apoptosis may de-pend on the ability to transmit the death signal or may be

Received 12 February 1998; revised 16 June 1998.Study subjects gave informed consent.Grant support: Hungarian Ministry of Welfare (T-01 409/1994), National

Research Fund (OTKA T022705), Agence Nationale de Recherche sur leSIDA to M.-L.G., Fondation pour la Recherche Medicale (Sidaction) toM.-L.G., Ministere des Affaires Etrangeres (Balaton program), Pasteur In-stitute, Centre National de la Recherche Scientifique and European Com-munity (Biomed-2 program, contract BMH4-CT 97-2055), and Inco-Copernicus Program (IC15-CT0901, coordinator M.-L.G.).

Reprints or correspondence: Dr. Marie-Lise Gougeon, Institut Pasteur,Unite d’Oncologie Virale, 28 rue du Dr. Roux, 75724 Paris Cedex 15, France([email protected]).

The Journal of Infectious Diseases 1998;178:1288–98q 1998 by the Infectious Diseases Society of America. All rights reserved.0022-1899/98/7805-0008$02.00

modulated by differential expression of anti-apoptotic proteinsof the Bcl-2 family [10, 11]. The CD95L is a type II transmem-brane protein expressed by activated T cells, and it is also pro-duced in soluble form by proteolytic cleavage. T cell receptor(TCR) triggering in activated peripheral T cells may induceapoptosis that involves autocrine suicide or paracrine deathmediated via CD95 receptor/ligand interaction [12–14]. Just asa defect of the CD95 system is intimately linked to an impairedremoval of autoreactive lymphocytes, inappropriate inductionof apoptosis may lead to various pathologic conditions [15].This is exemplified in chronic virologic diseases such as AIDS[16, 17].

Indeed, T cells from human immunodeficiency virus(HIV)–infected persons are highly prone to in vitro spontaneousand activation-induced apoptosis [18–22]. Although apoptosisis not detected in freshly isolated peripheral blood lymphocytes,a high level of apoptosis may be induced in both CD4 andCD8 T cells from patients after overnight culture in the presenceof mitogens or superantigens or after triggering of the TCR[18, 19, 21]. Recent characterization of apoptotic cells (frompatients) indicated that the majority were in an activated state[22, 23], and the intensity of activation-induced apoptosis in Tcells correlated with the degree of immune activation in thepatients [22, 23] and with disease progression [23].

The in vivo involvement of the CD95 pathway in T cellapoptosis during HIV infection is supported by reports showingan increased expression of the CD95 receptor [24, 25] and cellsurface CD95L [26] on CD4 and CD8 T lymphocytes of HIV-infected patients. This is associated with an increased suscep-tibility of patients’ T cells to apoptosis following ligation ofCD95 by agonist antibodies or soluble human CD95L and thesusceptibility of activation-induced apoptosis to inhibition by

at University of D

ebrecen, Faculty of Medicine, C

entral Library on January 11, 2012

http://jid.oxfordjournals.org/D

ownloaded from

http://jid.oxfordjournals.org/

JID 1998;178 (November) tRA-Induced Inhibition of Apoptosis in AIDS 1289

antagonist anti-CD95 antibodies [25–29]. Moreover, the pro-portion of compliant CD95-expressing cells dramatically in-creases with disease progression [23, 25, 29]. Of interest, ex-acerbated CD95 expression and CD95 sensitivity seem to beconfined to primed/memory (CD45R0) T cells, supporting theidea that virus-driven T cell activation is responsible for theincreased apoptosis and contributes to the loss of memory cellsthroughout HIV infection [30].

Some have speculated that inappropriate activation-inducedapoptosis of T cells from HIV-infected patients contributes tothe decline of CD4 T cell number in HIV infection [16, 17].The finding that a massive cell death is detected in situ in lymphnodes from patients, affecting all lymphocyte subsets [22], in-cluding CD4 T cells and dendritic cells [31], and concerningmostly noninfected cells [32], argues for this hypothesis. Themechanisms contributing to the priming for apoptosis of non-infected patients’ lymphocytes are actually misunderstood, butthey might involve some viral gene products, such as gp120and Tat, which accelerate activation-induced apoptosis in nor-mal T cells via triggering of the CD95 pathway [33].

A recent study evaluating the influence of ingestion by 6 HIV-infected persons of all-trans retinoic acid (tRA) on ex vivo Tcell apoptosis showed that peripheral blood mononuclear cells(PBMC) from these patients exhibited a reduced spontaneousand anti-CD3–induced apoptosis [34]. This finding was com-pleted by in vitro studies demonstrating that tRA preventedactivation-induced apoptosis of patients’ PBMC, whereas noeffect was observed on lymphocytes from control donors [34].These observations prompted us to determine whether the in-hibitory effect of tRA on apoptosis was similar on CD4 andCD8 T cells from HIV-infected persons at different stages ofthe disease and to characterize the molecular mechanism in-volved in this inhibition.

Materials and Methods

Blood samples. Peripheral blood was obtained from HIV-in-fected persons at various stages of the disease in the Service forInfectious Diseases (R. Roue), Begin Military Hospital, SaintMande, France. Patients were classified according to their ex vivoCD4 cell percentage (129%, 14%–29%, !14%), following the revisedclassification of HIV infection (CDC, Atlanta). Control blood sam-ples were drawn from HIV-seronegative healthy donors (CentreNational de Transfusion Sanguine, Paris).

Monoclonal antibodies (MAbs). The mouse MAbs specific forhuman surface antigens used in this study included anti-CD4(IgG1k, clone SK3) and anti-CD8 (IgG1k, clone SK1) conjugatedwith fluorescein isothiocyanate (FITC; Becton Dickinson, San Jose,CA). Anti-CD95 (IgG1k, clone UB2) MAb conjugated with FITCwas purchased from Immunotech (Marseille, France). Two anti-CD95L MAbs were used, clone 4A5 (Immunotech) to block ac-tivation-induced apoptosis and clone 33 (IgG1k; TransductionLaboratories, Pantin, France), in Western blot analysis. Controlantibodies for cell surface labeling included mouse FITC or phy-coerythrin-conjugated IgG1k (Becton Dickinson).

Lymphocyte isolation and stimulation. Human PBMC were iso-lated from heparinized blood by centrifugation on a Ficoll-Paque(Pharmacia, Uppsala, Sweden) density gradient. Cells were washedand resuspended in RPMI 1640 (BioWhittaker, Verviers, Belgium)supplemented with 10% heat-inactivated fetal calf serum (InstitutJacques Boy, Reims, France), 10 IU/mL penicillin, 10 mg/mL strep-tomycin, 20 mM HEPES, and 2 mM L-glutamine (completemedium).

In the apoptosis assay, PBMC ( /mL) were cultured for55 3 1024 h in complete medium alone or in the presence of an optimalconcentration of soluble (0.1 mg/mL) or coated (10 mg/mL, 1-h 30-min incubation in complete medium at 47C) anti-CD3 MAb (IoT-3; Immunotech) as previously described [23]. tRA (Sigma, St.Louis) diluted in dimethyl sulfoxide (DMSO) was added in somecultures at 1 mL of tRA/mL of culture, and control cultures received1 mL of DMSO. In some experiments, anti-CD95L MAbs (clone4A5) were added at the initiation of the culture at 1 mg/mL. After24 h, cells were harvested and studied for apoptosis quantification.

Phenotypic analysis of apoptotic cells. In recent comparativeanalyses of flow cytometric methods for apoptosis quantification[35], we found that the most reliable way to phenotype lymphoidapoptotic cells was a combination of cell surface staining and 7-amino actinomycin D (7-AAD) incorporation. 7-AAD discrimi-nates between early and late apoptotic cells and indicates the in-crease of fluorescence of apoptotic cells due to alteration in theirmembrane permeability [35].

Phenotypic analysis of apoptotic cells was performed as follows:cultured lymphocytes were washed in PBS containing 1%55 3 10

bovine serum albumin and 0.1% sodium azide (PBS-BSA-NaN3)and incubated 20 min at 47C with MAbs conjugated to FITCspecific for surface antigens CD4 or CD8. After being washed,lymphocytes were incubated with 20 mg/mL 7-AAD (Sigma) in PBSfor 20 min at 47C in the dark, as described [35]. Stained cells werefurther fixed with 1% paraformaldehyde in PBS in the presence of20 mg/mL nonfluorescent actinomycin D (Sigma) to block 7-AADstaining within apoptotic cells and to avoid nonspecific labeling ofliving cells [35]. Finally, the double-stained lymphocytes were in-cubated overnight at 47C in the dark.

For each sample, 20,000 lymphocytes were acquired for flowcytometry (FACScan; Becton Dickinson). The spectral propertiesof 7-AAD allow staining of apoptotic cells by fluorescence emissionin the red channel, FL-3 (wavelength, 650–850 nm), and the easyand simultaneous labeling of cell surface antigens (FL-1 and FL-2). In some experiments done to test the influence of tRA on apop-tosis, data are given as the percentage of inhibition of apoptosiscalculated as follows: % inhibition of spontaneous apoptosis 5

apoptosis in apoptosis in{[% medium 2 (% apoptosis 1 tRA)]/%; for anti-CD3–induced apoptosis, {[% apoptosismedium} 3 100

with apoptosis with apop-anti-CD3 2 (% anti-CD3 1 tRA)]/[%tosis with apoptosis in .anti-CD3 2 % medium]} 3 100

CD95 detection on the cell surface of CD4 and CD8 T lympho-cytes. Lymphocytes stimulated for 24 h by soluble anti-CD3MAbs (1 mg/mL) in the presence or absence of tRA (10 mM) werewashed twice and resuspended in ice-cold PBS-BSA-NaN3 and co-stained with phycoerythrin-conjugated anti-CD8, phycoerythrin-conjugated anti-CD4, and FITC-labeled anti-CD95, or controlMAbs. The cells were incubated 30 min at 47C, further fixed with1% paraformaldehyde, and acquired for flow cytometry (FAC-

at University of D




ownloaded from


1290 Szondy et al. JID 1998;178 (November)

Figure 1. Effect of tRA on ex vivo spontaneous and anti-CD3–induced apoptosis of peripheral blood mononuclear cells (PBMC) from healthycontrols or HIV-infected donors. PBMC ( /mL) from controls ( ) or HIV-infected persons at different stages of disease (CD4 cells55 3 10 n 5 7129%, ; 14%–29%, ; !14%, ) were cultured overnight in medium or activated by 0.1 mg/mL anti-CD3 monoclonal antibodies, inn 5 7 n 5 7 n 5 7presence or absence of various concentrations of tRA (0.01–10 mM). Quantification of apoptosis and calculations to obtain % of inhibition ofapoptosis were performed as described in Materials and Methods. Means of % of apoptosis or of % of inhibition of apoptosis in total lymphocytesare shown; bars represent SE.

Scan). For each sample, 5000–10,000 double-stained viable lym-phocytes were gated following size (forward scatter) and granu-larity (side scatter) criteria and analyzed (Lysis II software, BectonDickinson).

CD95L bioassay. CD95L on activated lymphocytes was de-tected using, as a target, radiolabeled U937 cells (incubated for 12h with 5 mCi/mL [3H]thymidine; Amersham, Amersham, UK),which had been preincubated for 12 h with IFN-g (200 U/mL; giftof Ara Hovanessian, Institut Pasteur, Paris). Interferon-g (IFN-g)treatment of U937 cells induced up-regulation of the CD95 mol-ecule and rendered the cells sensitive to CD95L-induced apoptosis[36]. IFN-g–treated radiolabeled U937 cells were then washed anddivided into 2 groups: The first was incubated at 38.17C for 1 hwith neutralizing anti-CD95 MAb (ZB4; 70 ng/mL) to cover theCD95 receptor; the other group was kept at 38.17C for 1 h andused as a control. The human lymphocyte population (PBMC fromcontrol or HIV-positive donors) to be tested for CD95L expressionwas preactivated for 3 h with anti-CD3 MAbs (1 mg/mL), in thepresence or absence of tRA (10 mM), and then incubated for 15 hwith either anti-CD95–treated or control U937 cells in the presenceof 1 mM cold thymidine. Apoptosis of U937 cells was measuredfollowing cell lysis, and the released radioactivity associated withfragmented DNA and that of the intact chromatin were measuredin counts per minute (cpm) as previously described [36]. The per-centage of fragmented DNA was expressed as (cpm released/total

. Data are expressed as the percentage of fragmentedcpm) 3 100DNA released above the background level detected in coculturesof U937 cells with nonstimulated PBMC.

CD95L detection by Western blot analysis. PBMC were stim-ulated for 24–48 h with coated anti-CD3 MAbs in the presence ofthe metalloprotease inhibitor, BB3103 (1 mg/mL; gift of Alan H.Drummond, British Biotech Pharmaceuticals, Oxford, UK). Celllysates (100 mg of total proteins/sample) were separated by SDS-PAGE and transferred onto a nitrocellulose membrane, which wasthen blocked with blocking buffer (5% nonfat milk in 10 mM Tris[pH 7.5], 100 mM NaCl, and 0.1% Tween 20) for 1 night at 47C

with agitation. The membrane was then incubated with primaryanti-CD95L MAbs (Transduction Laboratories, Pantin, France;1/1000 in blocking buffer) for 60 min at room temperature. Sixwashes were performed in washing buffer (10 mM Tris [pH 7.5],100 mM NaCl, and 0.1% Tween 20), and protein was detectedusing the ECL protein detection system (Amersham, Courtaboeuf,France). Briefly, peroxidase-conjugated secondary anti–mouse IgGantibodies were added and incubated for 30 min at room temper-ature. Again six washes were performed; substrate was added andleft for 1 min. The chemiluminescent blots were exposed to KodakXAR-5 films for recording; exposure times ranged from 5 to 15min. Positive control for CD95L expression consisted of lysatesfrom PBMC cultured at 377C for 3 days with 5 mg/mL concanavalinA and 10 IU/mL interleukin (IL)-2 in the presence of BB-3103,then stimulated with 10 ng/mL PMA and 500 ng/mL ionomycinfor 1 day, according to the report of Tanaka et al. [37].

Results

tRA inhibits both the spontaneous and activation-induced apop-tosis of PBMC of HIV-infected persons. We and others pre-viously reported that a short-term incubation in medium aloneof PBMC from HIV-infected persons induced a rapid spon-taneous apoptosis that could be detected by several features,including morphologic criteria, alteration in membrane per-meability, condensed chromatin, or the presence of fragmentedDNA on agarose gels [18–21]. Cytofluorometric analysis per-mitted the quantification and characterization of apoptotic cells[23], and, as is shown in figure 1A, the level of spontaneousapoptosis in lymphocytes from HIV-infected persons, detectedwith 7-AAD staining, was significantly higher than that of lym-phocytes from HIV-seronegative persons. A significant increaseof apoptosis could be induced in lymphocytes from HIV-in-fected persons after 24 h of stimulation with anti-CD3 MAbs

at University of D




ownloaded from



Figure 2. Influence of tRA on anti-CD3–induced apoptosis of CD4 and CD8 T cells from HIV-infected donor. Peripheral blood mononuclearcells ( /mL) from HIV-infected donor were stimulated overnight with 0.1 mg/mL anti-CD3 monoclonal antibodies (MAbs), in presence or55 3 10absence of 10 mM tRA, as indicated. Quantification of apoptotic cells in each subset was performed following dual staining with anti-CD4 oranti-CD8 MAbs and with 7-amino actinomycin D (7AAD) for detection of apoptotic lymphocytes. Percentage of apoptotic lymphocytes amongeach subset is indicated on dot blots. FSC, forward scatter.

(figure 1) [19, 21, 23], whereas no effect was detected on lym-phocytes from control donors. Moreover, as previously re-ported [23], the rate of spontaneous and anti-CD3–inducedapoptosis correlated with disease evolution, as evaluated by theex vivo percentage of CD4 T cells (figure 1A).

When various concentrations of tRA (0.01–10 mM) wereadded to the short-term cultures of control PBMC, no effectwas observed on spontaneous or on anti-CD3–induced apop-tosis (data not shown). Conversely, tRA had an inhibitory effecton spontaneous apoptosis of PBMC from HIV-infected personsthat was more pronounced on anti-CD3–induced apoptosis(figure 1B). Indeed, when 10 mM of tRA was added, a meanof 38% inhibition of apoptosis was detected in nonstimulatedcultures, whereas 60% inhibition was observed in anti-CD3–stimulated cultures. The failure of tRA to totally inhibitspontaneous and anti-CD3–induced apoptosis of patients’PBMC is not due to the lack of conversion of tRA to 9-cisretinoic acid within the time frame of the experiment. Thus,the same concentration (10 mM) of tRA was completely suc-cessful in inhibiting activation-induced death of mouse thy-mocytes or T cell hybridomas within the same time period (data

not shown). Comparative analysis of PBMC from individualHIV-infected donors revealed a certain heterogeneity in theirsusceptibility to tRA inhibitory effect. As discussed below, theintensity of the tRA-dependent inhibition of apoptosis of pa-tients’ lymphocytes is dependent on the stage of the disease.These observations are in agreement with the recently publishedstudy from Yang et al. [34].

Differential effect of tRA on CD4 and CD8 T lymphocytesfrom HIV-infected persons. Phenotypic analysis of apoptoticcells was performed with the 7-AAD dye, which allows thecombined staining of apoptotic cells and cell surface antigenssuch as CD4 or CD8 [35] (figure 2). As is shown in figures 2and 3A, both CD4 and CD8 T cells from patients are involvedin the process of apoptosis, whether spontaneous or inducedby anti-CD3 MAbs [19, 23]. A correlation was found betweenthe intensity of apoptosis in both CD4 and CD8 T cells anddisease evolution, and this was true for both spontaneous andanti-CD3–induced apoptosis (figure 3A) [23]. Addition of tRA(10 mM) to nonstimulated short-term cultures had marginal orno effect on apoptosis of CD4 and CD8 T cells from patientswith a high in vivo percentage of CD4 cells (figure 3B). In

at University of D




ownloaded from



Figure 3. Preferential inhibitory effect of tRA on activation-induced apoptosis of CD4 T lymphocytes from HIV-infected persons. A, Peripheralblood mononuclear cells ( /mL) from control donors ( ) or HIV-infected persons at different stages of disease (CD4 cells 129%,55 3 10 n 5 7 n 5

; 14%–29%, ; !14%, ) were cultured overnight in medium or activated by 0.1 mg/mL of anti-CD3 monoclonal antibodies. Apoptosis7 n 5 7 n 5 7was quantified in CD4 and CD8 T cell subsets. Data are mean 5 SE of % of apoptosis in each subset. B, Cultures were performed as indicatedabove, in presence of 10 mM tRA. Percentage of inhibition of apoptosis in each subset was calculated as described in Materials and Methods.Data are of % of inhibition of apoptosis in each subset.mean 5 SE

patients with more advanced disease (CD4 cells !14%), theeffect of tRA on spontaneous apoptosis was selectively ob-served on the CD4 T cell subset (figure 3B). When anti-CD3–induced apoptosis was studied, a strong inhibitory effectof tRA on apoptosis of CD4 T cells was also observed, evenin the early stages of HIV-infection (CD4 cells 129%) (figure3B, 3C). Strikingly, tRA had no effect on anti-CD3–inducedapoptosis of CD8 T cells in most of the patients, whatever thestage of the disease (figure 3B, 3C). Moreover, it must be noticedthat not only CD4 and CD8 T cells are involved in the process

of apoptosis in HIV-infected persons, but also B and NK cells[23]. Of interest, we observed that tRA also inhibited the spon-taneous apoptosis of B cells (data not shown). Together, thesedata reveal that tRA has a selective inhibitory effect on apop-tosis of CD4 T lymphocytes from HIV-infected persons, andthis inhibition is particularly strong on activation-inducedapoptosis.

tRA does not inhibit activation-dependent CD95 expression onCD4 and CD8 T cells. The CD95 receptor has been reportedto be a potent inducer of apoptosis [1, 2], to be mainly present

at University of D




ownloaded from



Figure 3. Continued. C, Influence of 10 mM tRA on spontaneous and anti-CD3–induced apoptosis of CD4 and CD8 T cells is shown forindividual patients, separated according to their ex vivo % of CD4 T cells.

on activated human T cells [38], and to be highly expressed onperipheral T cells from HIV-infected children [24]. We and oth-ers recently reported that CD95 expression was increased in Tcells from adult HIV-infected persons, and the cells were highlysensitive to apoptosis triggered through ligation of CD95[25–30]. One of the possible mechanisms of tRA-dependentinhibition of activation-induced apoptosis is interference in theinduction of CD95 expression on T cells following anti-CD3stimulation. Therefore, CD95 expression was analyzed on CD4and CD8 T cells from HIV-infected persons and controls after24 h of activation with anti-CD3 MAbs in the presence orabsence of tRA (10 mM). Anti-CD3 activation of T cells frompatients induced an increase in the mean fluorescence intensityof CD95, which was unchanged when tRA was added to thestimulated culture (figure 4B). In addition, comparison of thepercentages of CD41CD951 or CD81CD951 cells among pa-tients’ or controls’ total CD4 or CD8 T cells in unstimulated

or stimulated cultures indicated that anti-CD3 activation in-duced an increase in the percentage of CD951 cells that wasnot affected by tRA (figure 4A). The absence of effect of tRAon anti-CD3–induced CD95 expression was observed on T cellsfrom control donors and from HIV-infected persons, whateverthe stage of the disease (figure 4A). To address whether tRAaffected CD95 signaling, PBMC from patients were incubatedin the presence of the agonistic anti-CD95 MAbs (CH-11, 1mg/mL), and tRA (10 mM) was added simultaneously or 6 hlater. The rate of apoptosis was similar in the presence or ab-sence of tRA, suggesting that tRA had no influence on theCD95 signaling pathway (data not shown).

tRA prevents anti-CD3–induced apoptosis in patients’ T lym-phocytes through inhibition of CD95L expression. Since tRAdid not affect CD95 expression, we tested whether its inhibitoryeffect on anti-CD3–induced apoptosis of patients’ T cells wasrelated to interference of CD95L expression. As reported by

at University of D




ownloaded from



Figure 4. Absence of effect of tRA on activation-induced CD95 expression on CD4 and CD8 T lymphocytes from HIV-infected donors. A,Peripheral blood mononuclear cells ( /mL) from control donors ( ) or HIV-infected persons at different stages of disease (CD4 cells55 3 10 n 5 4129%, ; 14%–29%, ; !14%, ) were cultured overnight in medium or activated by 0.1 mg/mL anti-CD3 monoclonal antibodies inn 5 9 n 5 9 n 5 9presence or absence of 10 mM tRA. Percentage of CD95 expressing cells among CD4 and CD8 T lymphocytes was determined by cytofluorometryat end of culture. Data represent of % of CD951 cells in each subset. B, Comparative flow cytometry profiles of CD95 expressionmean 5 SEon CD4 T cells from representative HIV-infected donor, following overnight stimulation under conditions described above.

Tanaka et al. [37], sequential activation with concanavalin Aplus IL-2 and subsequently with PMA plus ionomycin inducesCD95L expression in PBMC from healthy donors (figure 5A).Similarly, CD95L can be detected in PBMC from HIV-infectedpersons that are activated under the same conditions (figure5A). With this protocol of activation, CD95L induction wasobserved for all patients tested. Anti-CD3 activation could alsoinduce CD95L expression on PBMC from healthy and infecteddonors, but a significant increase in CD95L in stimulated cul-tures compared with nonstimulated ones was optimally de-tected after 48 h of stimulation (figure 5B, lanes 4 and 5).

The influence of tRA on CD95L expression is shown in figure

5B. tRA was added at the initiation of the culture, and a stronginhibition of CD95L expression, triggered by anti-CD3 acti-vation, was observed in patients’ T cells (figure 5B, lane 6 vs.5). This specific inhibitory effect of tRA was confirmed in afunctional assay. PBMC from control donors or HIV-infectedpersons were stimulated for 3 h with anti-CD3 MAbs to induceCD95L expression and were cocultured for 15 h with radio-labeled U937 cells, which had been pretreated with IFN-g (200U/mL) to induce CD95 expression. This coculture induced celldeath by apoptosis of CD951 U937 cells through a CD95-basedmechanism, provided that CD95L was expressed by anti-CD3–activated lymphocytes.

at University of D




ownloaded from



Figure 5. Inhibition by tRA of CD95L expression by anti-CD3–stimulated lymphocytes. A, Peripheral blood mononuclear cells(PBMC) from control donor (lane C) and 3 HIV-infected patients(P1–P3) were sequentially stimulated with concanavalin A plus inter-leukin-2 and PMA plus ionomycin to induce CD95L expression [47].CD95L was detected by Western blot analysis. B, PBMC from HIV-infected patient were cultured in medium (lanes 1 and 4), stimulatedwith coated anti-CD3 monoclonal antibodies in absence (lanes 2 and5) or presence (lanes 1 and 6) of 10 mM tRA. Western blot analysis ofCD95L was performed after 24 h (lanes 1–3) or 48 h (lanes 4–6) ofstimulation. Representative experiment is shown.

Quantification of apoptosis in U937 cells was performed bycalculating the percentage of fragmented DNA released, as de-scribed in Materials and Methods, and the results were ex-pressed as the increase in fragmented DNA released above thebackground level of apoptosis (10%–15%) in cocultures of U937cells and unstimulated PBMC. Cocultures of IFN-g–pretreatedU937 cells with anti-CD3–stimulated PBMC of control donorswere associated with a slight increase of apoptosis in the U937cells (figure 6). This was more pronounced when anti-CD3–stimulated lymphocytes came from HIV-infected patients(figure 6). This induction of apoptosis in U937 cells was CD95-mediated, since it was almost completely prevented by pre-incubation of IFN-g–pretreated U937 cells with anti-CD95neutralizing antibodies (figure 6). Patients’ PBMC that werestimulated with anti-CD3 in the presence of tRA became unableto induce apoptosis in IFN-g–pretreated U937 cells, in a man-ner similar to that occurring when U937 were protected by anti-Fas neutralizing antibodies (figure 6). For a control, we checkedthat neither tRA nor anti-CD95 neutralizing antibodies hadeffect when added to cocultures of IFN-g–pretreated U937 cellsand unstimulated PBMC (figure 6). We also observed that tRAhad no blocking effect on CD95-induced death of U937 cells,whether tRA was added at the same time as agonistic anti-CD95 MAbs (CH-11, 1 mg/mL) or 6 h later (anti-CD95 MAbsinduced of U937 apoptosis vs.24.5% 5 2.2% 26.5% 5 1.8%and for anti-CD95 plus 10 mM tRA added at24.2% 5 3.1%time 0 or 6 h later, respectively). These data suggest that tRAdoes not influence CD95 signaling.

Taken together, our data show that tRA inhibited the acti-

vation-dependent induction of CD95L expression on T cellsand support the hypothesis that the prevention of anti-CD3–induced apoptosis of patients’ T cells under the influenceof tRA is the consequence of decreased CD95–CD95Linteraction.

Discussion

In the present study, we investigated the influence of tRA onex vivo spontaneous and activation-induced apoptosis of CD4and CD8 T lymphocytes from HIV-infected persons at differentstages of the disease. Retinoids are a group of natural andsynthetic vitamin A derivatives that modulate the growth anddifferentiation of many different cells in vivo and in vitro. tRAis the high-affinity ligand for retinoic acid receptors, and it wasfound to prevent specifically TCR-induced apoptosis of murineT cell hybridomas and thymocytes, while not affecting TCR-mediated cellular activation itself [39]. tRA was also foundrecently to inhibit ex vivo activation-induced apoptosis ofPBMC from HIV-infected persons [34]. However, the cell sub-sets involved were not studied, and the mechanism of inhibitionnot addressed.

In the present report, we confirm the ability of tRA to preventex vivo activation-induced apoptosis of PBMC from HIV-in-fected persons, and we show that the inhibitory effect of tRAconcerns essentially the CD4 T cell subset. Moreover, the pre-vention of apoptosis in anti-CD3–stimulated T cells from pa-tients was found to be associated with the inhibition of CD95Linduction following TCR triggering. Therefore tRA preventsex vivo activation-induced apoptosis in patients’ PBMC by in-hibiting the CD95 pathway.

The reason why not only T cells but also B cells, NK cells[22, 23], and accessory cells [31] from HIV-infected patientsundergo apoptosis is not clear, but several mechanisms havebeen proposed. The in vivo priming of CD4 T cells by gp120and subsequent deletion following T cell triggering [33, 40]would represent a CD4-specific mechanism. The possible de-fective antigen-presenting cell function [41] or alteration in thepattern of cytokines produced [42, 43] in patients might con-tribute to apoptosis of both CD4 and CD8 T cells. However,recent studies, including analyses of patients’ lymph-nodes, sug-gested that apoptosis detected in B and T cell subsets was re-lated and caused by the continuous stimulation of the immunesystem observed throughout HIV infection [22, 23, 31]. Thishyperactivation causes the dysregulation of the expression ofsurvival (Bcl-2) and death (CD95, CD95L) molecules, inducinga preapoptotic state in patients’ circulating lymphocytes[22–31].

We recently reported that tissue transglutaminase (tTG) ishighly expressed in lymph nodes of HIV-positive patients, onlymphocytes as well as dendritic cells, while undetectable inlymph nodes of control donors; it is also selectively expressedin blood CD4 T cells from patients, thus defining a preapoptotic

at University of D




ownloaded from



Figure 6. Inhibition by tRA of CD95L activity of anti-CD3–stimulated lymphocytes from control donors or HIV-infected patients. Radiolabeledinterferon (IFN)-g–pretreated U937 cells were cocultured with peripheral blood mononuclear cells (PBMC) from control donors ( ) or HIV-n 5 3infected patients ( ), previously activated for 3 h by 0.1 mg/mL anti-CD3 monoclonal antibodies (MAbs), in presence or absence of 10 mMn 5 6tRA. In some experiments, IFN-g–pretreated U937 cells were preincubated with neutralizing anti-Fas MAb (70 ng/mL), as indicated. Data are

% of fragmented DNA released above background level, as indicated in Materials and Methods.mean 5 SE

marker in this subset [31]. The in vitro protective effect of tRA([34] and the present report) on apoptosis of patients’ lympho-cytes is probably related to the activation state of these lym-phocytes. Indeed, tRA was also shown to prevent activation-induced death of PBMC from healthy donors, provided thecells were preactivated [34]. The relation between tTG expres-sion on patients’ PBMC and the inhibitory effect of tRA on Tcell apoptosis was not addressed in this study. However, a pro-apoptotic effect of tRA was reported in different cell systems,including tumor cells, tracheobronchial cells, and murine Tcells, in which tTG was up-regulated [44]. In the present study,an anti-apoptotic effect of tRA was observed on patients’ CD4T cells, which express high levels of tTG, suggesting that theinfluence of tRA on the apoptosis pathway is related not justto the expression of tTG. The expression of other molecules,such as Bcl-2 or its relatives, might also influence thesepathways.

The present report suggests that tRA prevents anti-CD3–induced apoptosis of patients’ lymphocytes by interferingin the CD95 pathway. CD95 is highly expressed on PBMC fromHIV-infected persons, and at the AIDS stage, around

80%–100% of these cells are CD95-positive [24–30]. Increasedexpression of CD95 is detected on both CD4 and CD8 T cellsfrom patients, and those cells are susceptible to apoptosis in-duced by CD95 triggering [25, 28–30]. CD95L expression canbe induced following T cell activation [14], and we show herethat the increased apoptosis induced by anti-CD3 stimulationin patients’ PBMC is associated with an induction of CD95Lexpression. Moreover anti-CD95L MAbs could inhibit anti-CD3–induced apoptosis (data not shown), in agreement withprevious observations of Baumler et al. [45], who reported sim-ilar effects of anti-CD95 MAbs. Therefore, activation-inducedapoptosis in patients’ PBMC is probably mediated via the en-gagement of the CD95 receptor by activation of up-regulatedCD95L.

The CD95 pathway might also be involved in spontaneousapoptosis. Thus, neutralizing anti-CD95 MAbs inhibited par-tially spontaneous apoptosis of T cells from HIV-positive pa-tients in the late stage of disease, whereas no effect was observedon patients with earlier disease stages (data not shown). Thiswould suggest that, in AIDS, spontaneous apoptosis is partiallymediated by CD95L, which is released by the patient’s hyper-

at University of D




ownloaded from



activated lymphocytes and promotes apoptosis on CD95-pos-itive susceptible lymphocytes in this patient.

tRA inhibited activation-induced apoptosis by blocking theactivation-dependent induction of CD95L expression on Tcells. These findings are consistent with several recent reportsdemonstrating that tRA inhibits TCR-mediated apoptosis in Tcell hybridomas by blocking the expression of CD95L [46–48].The selective effect of tRA on the CD4 T cell subset frompatients is intriguing. It would suggest that apoptosis of pa-tients’ CD4 T cells preferentially involves the CD95 death path-way. This is indeed compatible with observations of Katsikiset al. [25] and our own, showing that, following ligation of theCD95 receptors with agonistic antibodies, patients’ CD4 T lym-phocytes are more susceptible than CD8 T lymphocytes toCD95-based apoptosis. This is also in agreement with a recentstudy of the normal immune system, which showed that theCD95 pathway mediates TCR-induced apoptosis of most CD4T cells, while tumor necrosis factor mediates the death of CD8T cells [49].

tRA is currently used to treat acute promyelocytic leukemiaand has recently been investigated as a possible treatment forAIDS-related Kaposi’s sarcoma [50]. Yang et al. [34] recentlyadministered tRA orally to 6 HIV-infected patients and ob-served a reduction in ex vivo activation-induced apoptosis ofPBMC from treated patients compared with nontreated ones.No overall elevation of circulating CD4 T cells was noted inpatients undergoing tRA therapy [34] but, as discussed by theauthors, such an effect would not be expected from this smalland time-limited study.

The in vitro influence of tRA on activation-induced CD95Lexpression makes it attractive as a means of affecting pathologicapoptosis. The in vivo persistence of HIV associated with theunceasing expression of retroviral antigens is probably the pri-mary mechanism for the chronic stimulation of the immunesystem. This mechanism is responsible for the dysregulation inthe expression of molecules involved in the balance betweencell survival and cell death, such as Bcl-2 and CD95, respec-tively down-regulated and up-regulated in patients’ lympho-cytes [23, 24–30]. It also triggers the expression of CD95L [26,45], thus creating appropriate conditions for excessive CD95-based peripheral T cell deletion throughout HIV infection.

Recently, combination antiretroviral therapies, includinganti-HIV proteases, have been shown to be quite efficient inreducing the plasmatic virus load in HIV-infected persons [51,52] and to limit the development of AIDS. A decrease in lym-phocyte activation and in the ex vivo expression of CD95L [53]was reported in T cells of patients under combination anti-retroviral therapy, and we found that the rate of spontaneousand activation-induced apoptosis was significantly reduced intreated patients, concomitant with decreased CD95 expression(Gougeon ML, et al., unpublished data). However, these com-bined therapies are not always successful from an immunologicpoint of view, and integrated treatments that will help interrupt

the viral triggering of apoptotic cell death should be considered.The coadministration of tRA, involved in the regulation of Tcell apoptosis [44], with antiretroviral therapy might representa strategy to sustain T cell numbers and recover immune func-tions in AIDS.

Acknowledgment

We thank Luc Montagnier for constant support in the course of thisstudy.

References

1. Nagata S, Golstein P. The Fas death factor. Science 1995;267:1449–56.2. Krammer PH, Dhein J, Walczak H, et al. The role of APO-1 mediated

apoptosis in the immune system. Immunol Rev 1994;142:175–91.3. Watanabe-Fukunaga R, Brannan CI, Copeland NG, Jenkins NA, Nagata

S. Lymphoproliferation disorder in mice explained by defects in Fas an-tigen that mediates apoptosis. Nature 1992;356:314–7.

4. Takahashi T, Tanaka M, Brannan CI, et al. Generalized lymphoproliferativedisease in mice caused by a point mutation in the Fas ligand. Cell 1994;76:969–76.

5. Rieux-Laucat F, Le Deist F, Hivroz C, et al. Mutations in Fas associatedwith human lymphoproliferative syndrome and autoimmunity. Science1995;268:1347–9.

6. Fisher GH, Rosenberg FJ, Straus SE, et al. Dominant interfering Fas genemutations impair apoptosis in a human autoimmune lymphoproliferativesyndrome. Cell 1995;81:935–46.

7. Drappa J, Vaishnaw AK, Sullivan KE, Chu JL, Elkon KB. Fas gene mu-tations in the Canale-Smith syndrome, an inherited lymphoproliferativedisorder associated with autoimmunity. N Engl J Med 1996;335:1643–9.

8. Peter ME, Kischkel FC, Hellbardt S, Chinnaiyan AM, Krammer PH, DixitVM. CD95 (APO-1/Fas)-associating signalling proteins. Cell Death Differ1996;3:161–5.

9. Nagata S. Apoptosis by death factor. Cell 1997;88:355–65.10. Peter ME, Kischkel FC, Scheuerpflug CG, Medema JP, Debatin KM, Kram-

mer PH. Resistance of cultured peripheral T cells towards activation-induced cell death (AICD) involves a lack of recruitment of FLICE tothe death-inducing signaling complex (DISC). Eur J Immunol 1997;27:1207–12.

11. Reed JC. Double identity for proteins of the Bcl-2 family. Nature 1997;387:773–5.

12. Dhein J, Walczak H, Baumler C, Debatin KM, Krammer PH. Autocrine T-cell suicide mediated by APO-1/(Fas/CD95). Nature 1995;373:438–41.

13. Brunner T, Mogil RJ, LaFace D, et al. Cell-autonomous Fas (CD95)/Fasligand interaction mediates activation induced apoptosis in T cell hybrid-omas. Nature 1995;373:441–4.

14. Alderson MR, Tough TW, Davis-Smith T, et al. Fas ligand mediates acti-vation induced cell death in human T lymphocytes. J Exp Med 1995;181:71–7.

15. Debatin KM. Disturbances of the CD95 (APO-1/Fas) system in disorders oflymphohaematopoietic cells. Cell Death Differ 1996;3:185–9.

16. Gougeon ML. Does apoptosis contribute to CD4 T cell depletion in HIVinfection. Cell Death Differ 1995;2:1–7.

17. Ameisen JC, Estaquier J, Idziorek J, De Bels F. Programmed cell death andAIDS: significance, perspectives and unanswered questions. Cell DeathDiffer 1995;2:9–13.

18. Gougeon ML, Olivier R, Garcia S, et al. Demonstration of an engagementprocess towards cell death by apoptosis in lymphocytes of HIV infectedpatients. C R Acad Sci 1991;312:529–37.

at University of D




ownloaded from



19. Meyaard L, Otto SA, Jonker RR, Mijnster MJ, Keet RP, Miedema F. Pro-grammed death of T cells in HIV-1 infection. Science 1992;257:217–9.

20. Groux H, Torpier G, Monte D, Mouton Y, Capron A, Ameisen JC. Acti-vation-induced death by apoptosis from human immunodeficiency vi-rus–infected asymptomatic individuals. J Exp Med 1992;175:331–40.

21. Gougeon ML, Garcia S, Heeney J, et al. Programmed cell death in AIDS-related HIV and SIV infections. AIDS Res Hum Retroviruses 1993;9:553–63.

22. Muro-Cacho CA, Pantaleo G, Fauci A. Analysis of apoptosis in lymph nodesof HIV-infected persons. Intensity of apoptosis correlates with the generalstate of activation of the lymphoid tissue and not with the stage of diseaseor viral burden. J Immunol 1995;154:5555–66.

23. Gougeon ML, Lecoeur H, Dulioust A, et al. PCD in peripheral lymphocytesfrom HIV-infected persons. The increased susceptibility to apoptosis ofCD4 and CD8 T cells correlates with lymphocyte activation and withdisease progression. J Immunol 1996;157:3502–19.

24. Debatin KM, Fahrig-Faissner A, Enenkel-Stoodt S, Kreuz W, Benner A,Krammer PH. High expression of APO-1 (CD95) on T lymphocytes fromhuman immunodeficiency virus-1–infected children. Blood 1994;83:3101–3.

25. Katsikis PD, Wunderlich ES, Craig CA, Herzenberg LA, Herzenberg LA.Fas antigen stimulation induced marked apoptosis of T lymphocytes inHIV-virus infected individuals. J Exp Med 1995;181:2029–36.

26. Sloand EM, Young NS, Kumar P, Weichold FF, Sato T, Maciejewski JP.Role of Fas ligand and receptor in the mechanism of T-cell depletion inAIDS: effect on CD41 lymphocyte depletion and HIV replication. Blood1997;89:1357–63.

27. Boudet F, Lecoeur H, Gougeon ML. Apoptosis associated with ex vivo down-regulation of bcl-2 and up-regulation of Fas in potential cytotoxic CD81

T lymphocytes during HIV infection. J Immunol 1996;156:2282–93.28. Yang YL, Liu ZH, Ware CF, Ashwell JD. A cysteine protease inhibitor

prevents activation-induced T cell apoptosis and death of peripheral bloodcells from human immunodeficiency virus–infected individuals by inhib-iting upregulation of Fas ligand. Blood 1997;89:550–7.

29. Gougeon ML, Lecoeur H, Boudet F, et al. Lack of chronic immune activationin HIV-infected chimpanzees correlates with the resistance of T cells toFas/Apo-1 (CD95)–induced apoptosis and preservation of a Th1 phe-notype. J Immunol 1997;158:2964–76.

30. Bohler T, Nedel S, Debatin KM. CD95-induced apoptosis contributes to lossof primed/memory but not resting/naive T cells in children infected withhuman immunodeficiency virus. Pediatr Res 1997;41:878–85.

31. Amendola A, Gougeon ML, Poccia F, Bondurand A, Fesus L, PiacentiniM. Induction of tissue transglutaminase in HIV pathogenesis: evidencefor high rate of apoptosis on CD41 T lymphocytes and accessory cells inlymphoid tissues. Proc Natl Acad Sci USA 1996;93:11057–62.

32. Finkel TH, Tudor-Williams G, Banda NK, et al. Apoptosis occurs predom-inantly in bystander cells and not in productively infected cells of HIV-and SIV-infected lymph nodes. Nature Med 1995;1:129–34.

33. Westendorp MO, Frank R, Oschsenbauer C, et al. Sensitization of T cellsto CD95-mediated apoptosis by HIV-1 Tat and gp120. Nature 1995;375:497–500.

34. Yang Y, Bailey J, Vacchio MS, Yarchoan R, Ashwell JD. Retinoic acid in-hibition of ex vivo human immunodeficiency virus–associated apoptosisin peripheral blood cells. Proc Natl Acad Sci USA 1995;92:3051–5.

35. Lecoeur H, Gougeon ML. Comparative analysis of flow cytometric methodsfor apoptosis quantitation in thymocytes and human peripheral blood

lymphocytes of controls and HIV1 persons. Evidence for interferences ofgranulocytes and erythrocytes. J Immunol Methods 1996;198:87–99.

36. Sumimoto S, Ishigami T, Horiguchi Y, et al. Anti-Fas antibody induces dif-ferent types of cell death in the human B cell line B104: the role of single-strand DNA breaks and poly (ADP-ribosylation) in cell death. Cell Im-munol 1994;153:184–93.

37. Tanaka M, Suda T, Haze K, et al. Fas ligand in human serum. Nature Med1996;2:317–22.

38. Miyawaki T, Uehara T, Nibu R, et al. Differential expression of apoptosis-related Fas antigen on lymphocyte subpopulations in human peripheralblood. J Immunol 1992;149:3753–8.

39. Yang Y, Vacchio MS, Ashwell JD. 9-cis retinoic acid inhibits activation-drivenapoptosis: implications for retinoid X receptor involvement in thymocytedevelopment. Proc Natl Acad Sci USA 1993;90:6170–4.

40. Banda NM, Bernier J, Kurahara DK, et al. Cross-linking by HIV gp120primes T cells for activation-induced apoptosis. J Exp Med 1992;176:1099–106.

41. Meyaard L, Schuitemaker H, Miedema F. T cell dysfunction in HIV infection:anergy due to defective antigen-presenting cell function? Immunol Today1993;14:161–4.

42. Clerici M, Sarin A, Coffman RL, et al. Type 1/type 2 cytokine modulationof T-cell programmed cell death as a model for human immunodeficiencyvirus pathogenesis. Proc Natl Acad Sci USA 1994;91:1181–5.

43. Ledru E, Lecoeur H, Garcia S, Roue R, Gougeon ML. Differential suscep-tibility to activation-induced apoptosis among peripheral Th1 subsets.Correlation with Bcl-2 expression and consequences for AIDS pathogen-esis. J Immunol 1998;160:3194–206.

44. Szondy Z, Reichert U, Fesus L. Retinoid acids regulate apoptosis of T lym-phocytes through an interplay between RAR and RXR receptors. CellDeath Differ 1998;5:4–10.

45. Baumler CB, Bohler T, Herr I, Benner A, Krammer PH, Debatin KM. Ac-tivation of the CD95 (APO-1/Fas) system in T cells from human im-munodeficiency virus type-1–infected children. Blood 1996;88:1741–6.

46. Bissonnette RP, Brunner T, Lazarcchik SB, et al. 9-cis retinoic acid inhibitionof activation-induced apoptosis is mediated via regulation of fas ligandand requires retinoic acid receptor and retinoid X receptor activation. MolCell Biol 1995;15:5576–85.

47. Yang Y, Minucci S, Ozato K, Heyman RA, Ashwell JD. Efficient inhibitionof activation-induced fas ligand up-regulation and T cell apoptosis byretinoids requires occupancy of both retinoid X receptors and retinoidacid receptors. J Biol Chem 1995;270:18672–7.

48. Yang Y, Mercep M, Ashwell JD. Fas and activation-induced Fas ligandmediate apoptosis of T cell hybridomas: inhibition of Fas ligand expres-sion by retinoic acid and glucocorticoids. J Exp Med 1995;181:1673–82.

49. Zheng L, Fisher G, Miller RE, Peschon J, Lynch D, Lenardo MJ. Inductionof apoptosis in mature T cells by TNF. Nature 1995;377:348–51.

50. Smith MA, Parkinson DR, Cheson BD, Friedman MA. Retinoids in cancertherapy. J Clin Oncol 1992;10:839–64.

51. Ho DD, Neumann AU, Perelson AS, Chen W, Leonard JM, Markowitz M.Rapid turnover of plasma virions and CD4 lymphocytes in HIV-1 infec-tion. Nature 1995;373:123–6.

52. Wei X, Ghosh SL, Taylor ME, et al. Viral dynamics in human immunode-ficiency virus type 1 infection. Nature 1995;373:117–22.

53. Bohler T, Herr I, Geiss M, Haas J, Debatin KM. Downregulation of increasedCD95 (APO-1/Fas) ligand expression in T cells following antiretroviraltherapy in HIV-1 infected children. Blood 1997;90:886–8.

at University of D




ownloaded from


All‐ Trans Retinoic Acid Inhibition of Anti‐CD3–Induced T Cell Apoptosis in Human...

Documents

Transcript of All‐ Trans Retinoic Acid Inhibition of Anti‐CD3–Induced T Cell Apoptosis in Human...