CTLA4CD80/CD86 interactions on primary mouse CD4+ T cells integrate signal-strength information to modulate activation with Concanavalin A

CTLA4-CD80/CD86 interactions on primary mouse CD4� Tcells integrate signal-strength information to modulateactivation with Concanavalin A

Sambuddho Mukherjee, Asma Ahmed, and Dipankar Nandi1

Department of Biochemistry, Indian Institute of Science (IISc), Bangalore, India

Abstract: The mechanisms by which concanava-lin A (Con A), a lectin, activates T cells are poorlystudied. A low dose of Con A is stimulatory for Tcells, whereas a high dose of Con A results insuppression of proliferation and enhanced T celldeath. The expression and functional roles of co-stimulatory receptors, CD28 and cytotoxic T-lym-phocyte antigen 4 (CTLA4), and their ligands,CD80 and CD86, on primary mouse CD4� T cellsafter activation with different doses of Con A werestudied. CTLA4-CD80/CD86 interactions in thisT:T cell activation model demonstrate distinct out-comes depending on the dose of Con A. CTLA4-CD80/CD86 interactions inhibit CD4� T cell cy-cling and survival after activation with a suppres-sive dose of Con A by increasing oxidative stressand decreasing levels of BclXL. The enhancedCD4� T cell death with a suppressive dose of ConA is dependent on excess H2O2 and nitric oxide butis independent of Fas and caspase activity. It issurprising that the increased proliferation of CD4�

T cells with a suppressive dose of Con A on block-ing CTLA4-CD80/CD86 interactions is largely in-terleukin (IL)-2-independent but is cyclosporineA-sensitive. On activation with a stimulatory doseof Con A, CTLA4-CD80/CD86 interactions en-hance T cell activation and survival by reducing theproduction of reactive oxygen species, increasingIL-2 and BclXL levels. Here IL-10 but not trans-forming growth factor-� plays a functional role. Insummary, CTLA4-CD80/CD86 interactions on Tcells integrate signal strength, based on the dose ofCon A, to enhance or inhibit primary mouse CD4�

T cell cycling and survival. J. Leukoc. Biol. 78:000–000; 2005.

Key Words: costimulation � T cell cycling and survival � IL-2-independent � oxidative stress � IL-10 � TGF-�

INTRODUCTION

Optimal activation of CD4� T cells requires two distinct sig-nals: Signal 1 is T cell receptor (TCR)-CD3-mediated, arisingfrom interaction between the TCR and the cognate majorhistocompatibility complex (MHC)- peptide complex, and sig-

nal 2 is antigen-independent and involves the binding of thecostimulatory receptor CD28 on T cells to its ligands CD80/CD86 on antigen-presenting cells (APCs). CD28 has a longhalf-life and is constitutively expressed on the surface of Tcells. CD28 binding to CD80/CD86 together with TCR signal-ing (signal 1) lead to the production of high levels of interleu-kin (IL)-2 and other cytokines followed by increased T cellcycle progression. In addition, CD28 signaling enhances sur-vival by increasing levels of antiapoptotic proteins, includingBclXL. Cytotoxic T-lymphocyte antigen 4 (CTLA4; CD152),which is closely related to CD28, is another important T cellcostimulatory receptor. CTLA4 has a short half-life, and themajority of molecules is present intracellularly with low levelsof cell-surface expression. CTLA4 binding to CD80/CD86 re-sults in decreased IL-2 production and T cell cycle progres-sion. The opposing roles of CD28 and CTLA4 during T cellactivation are reinforced by the phenotype of mice lackingthese costimulatory receptors. cd28�/� mice are able to initiatebut are unable to sustain T cell immune responses. A moredramatic phenotype is displayed by ctla4�/� mice, which diewithin 3–4 weeks of age as a result of hyperproliferation ofCD4� T cells (reviewed in refs. [1–3]).

One mechanism by which CTLA4 dampens T cell activationis by binding to CD80/CD86 with at least tenfold higher affinityand sequestering these ligands from binding to CD28. Inaddition, the cytoplasmic tail of CTLA4 is required for optimalinhibition of T cell activation [4, 5]. CTLA4 ligation alsoreduces extracellular signal-regulated kinase activation [6],lowers IL-2 production, and reduces T cell cycling [7, 8]. In theinitial study, CTLA4 ligation in primary CD4� T cells, togetherwith anti-CD3 activation, reduced T cell cycling without af-fecting survival [9]. However, a subsequent study using atransgenic TCR mouse demonstrated that CTLA4 expressionreduced primary CD4� T cell cycling and survival [10]. Therelationship between CTLA4 ligation and survival is contro-versial, and some reports demonstrate a role in enhancing[11–14] or inhibiting [15–17] T cell survival. CTLA4 ligationresults in T cell anergy in some systems [18–20] but not others[21, 22]. Also, CTLA4 ligation results in increased levels oftransforming growth factor-� (TGF-�) [23, 24], but the func-

1 Correspondence: #126, Department of Biochemistry, Indian Institute ofScience, Bangalore, 560012 India. E-mail: [email protected]

Received November 5, 2004; revised December 22, 2004; accepted Febru-ary 22, 2005; doi: 10.1189/jlb.1104644.

0741-5400/05/0078-0001 © Society for Leukocyte Biology Journal of Leukocyte Biology Volume 78, July 2005 1

Uncorrected Version. Published on March 23, 2005 as DOI:10.1189/jlb.1104644

Copyright 2005 by The Society for Leukocyte Biology.

tional role of this TGF-� is controversial [25]. AlthoughCTLA4 clearly plays important roles in the T cell immuneresponse, the mechanisms by which it acts are not fully un-derstood.

T cells are known to express costimulatory receptors, andtheir ligands, although the functional consequences of theseinteractions, are not well studied. B7 molecules on mouse Tcells are hypoglycosylated and bind to CTLA4 but not CD28[26, 27]. In addition, differences in the roles of CD80 on T cellsand APCs are known [28]. Finally, these interactions may beclinically important, as shown in a disease model [29]. Ingeneral, CTLA4 has been shown to inhibit T cell responses,although there are some studies that demonstrate a role ofCTLA4 in enhancing T cell activation [13, 30–34]. In fact, arecent report demonstrated that a single-chain Fv ligand toCTLA4 enhances T cell activation [34]. However, it is unclearwhether the inhibiting or enhancing roles of CTLA4 in theliterature are a result of the use of different model systems, andone is unable to predict when CTLA4 would act as an enhanceror an inhibitor of CD4� T cell responses. To study the func-tional roles of the B7 family of costimulatory receptors andligands on T cells, we developed a primary CD4� T cellactivation model and demonstrated that CTLA4-CD80/CD86interactions inhibit or enhance primary CD4� T cell activationdepending on the stimulatory conditions used: ActivatingCD4� T cells with plate-bound anti-CD3 and blocking CTLA4-CD80/CD86 interactions increase T cell proliferation; i.e.,CTLA4-CD80/CD86 interactions inhibit T cell activation. Con-versely, activating CD4� T cells with phorbol 12-myristate13-acetate (PMA; P) and ionomycin (I) and blocking CTLA4-CD80/CD86 interactions greatly inhibit T cell proliferation;i.e., CTLA4-CD80/CD86 interactions enhance T cell activation[13].

Concanavalin A (Con A) is a lectin that binds to cell-surfaceglycoproteins, including the TCR, and has been used exten-sively to study T cell activation [35–40]. In this report, westudied the functional consequences of blocking CTLA4-CD80/CD86 interactions after activating CD4� T cells withdifferent amounts of Con A. We show that with a stimulatorydose of Con A, CTLA4-CD80/CD86 interactions enhance,whereas with a suppressive dose of Con A, the same interac-tions inhibit CD4� T cell cycling and survival. This studyclearly demonstrates that CTLA4-CD80/CD86 interactions in-tegrate signal strength, based on the dose of Con A, to modulateprimary mouse CD4� T cell cycling and survival.

MATERIALS AND METHODS

Mice

CD4� T cells were obtained from C57BL/6 mice, usually 6- to 10-weeks old.Mice were obtained from the Central Animal Facility [Indian Institute ofScience (IISc), Bangalore, India] or the National Institute of Nutrition (Hyder-abad, India) and housed in our departmental facility, as per institutionalguidelines.

Media, antibodies, and cell lines

Primary CD4� T cells were cultured in RPMI 1640, supplemented with 25 mMHEPES (Sigma Chemical Co., St. Louis, MO), 2 mM L-glutamine (Life Tech-

nologies, Gaithersburg, MD), 5 �M �-mercaptoethanol (Merck, Rahway, NJ),100 �g/ml penicillin, 250 �g/ml streptomycin, 50 �g/ml gentamycin (HiMediaLabs, Mumbai, India), and 5% heat-inactivated fetal bovine serum (FBS;Sigma Chemical Co.). Anti-CD3 (145-2C11), anti-CD28 (37.51), and hamstercontrol antibody were sourced from eBioScience (San Diego, CA). Ascitescontaining anti-CTLA4 and murine CTLA4 human immunoglobulin G1(mCTLA4hIgG1) were used for all blocking studies, as described previously[13]. Anti-CD8 (3.155) and heat-stable antigen (J11D) culture supernatantswere used to purify lymph node CD4� T cells. All other antibodies (e.g.,anti-IL-2, anti-IL-4, and others) were obtained from eBioScience. Anti-TGF-�1 was kindly provided by Dr. Paturu Kondaiah (IISc). For flow cytom-etry, anti-BclXL was obtained from eBioScience, and secondary antibodieswere from Jackson ImmunoResearch Laboratories (West Grove, PA).

Isolation of CD4� T cells and activation

CD4� T cells were purified by complement-mediated lysis of J11D� andCD8� cells, as described previously [13]. Live cells were obtained by densitygradient centrifugation with Histopaque 1083 (Sigma Chemical Co.) andsubjected to panning over a T25 flask coated with 100 �g/ml anti-mouse Ig(Jackson ImmunoResearch Laboratories). CD4� T cell preparations weretypically �95% pure, as measured by flow cytometric analysis for key markers.Purified T cells were plated at 6–7 � 104 cells/well in 96-well U-bottom plates(Costar, Corning Inc., NY) in a final volume of 100 �l/well. To minimizenonspecific adhesion of monoclonal antibodies (mAb) to the plate, all wellswere precoated with RPMI 1640 containing 5% FBS. In most assays, T cellswere activated with different doses of Con A (Sigma Chemical Co.), asmentioned in the figure legends. Anti-CD28 was used at a concentration of0.3–0.5 �g/ml, and anti-CTLA4 and mCTLA4hIgG ascites were used at a finalconcentration of 1:100. Fetuin, glutathione (GSH), catalase, N-methyl-L-argi-nine (L-NMA; all from Sigma Chemical Co.), IL-2, IL-4 (PeproTech, Israel),and cyclosporine A (CsA; Sigma Chemical Co.) were titred and used at theindicated concentrations. Unless otherwise mentioned, CD4� T cell cultureswere pulsed 36 h after activation with �0.4 �Ci/well [3H]-thymidine (BRIT,Mumbai, India) and harvested 12 h later. Incorporated radioactivity wasmeasured using a liquid scintillation counter (Beckman LS6500) to assesslevels of proliferation. The data are presented as mean � SD of replicates in onerepresentative of multiple individual experiments.

Cytokine assays

Supernatants from T cell assays were collected at different time-points or 36 hafter activation, and cytokine-specific enzyme-linked immunosorbent assay(ELISA; eBioscience) or bioassays were performed for IL-2 and TGF-� [usingthe cell line CC chemokine ligand (CCL)-64], respectively. The amount ofcytokine in the supernatants was determined using an equation derived fromvalues obtained from known amounts of standard cytokines, and specific T cellsecretion of cytokines was determined by deducting appropriate controls.ELISA was performed with standard amounts of recombinant IL-2 (rIL-2) andvarious dilutions of culture supernatants. Typically, the linear detection rangeof the IL-2 assay was 30–900 pg/ml. Active TGF-� was measured as an indexof growth inhibition of CCL-64 cells, which were cultured (�5000 cells/well)with supernatants or with known amounts of rTGF-�1 (PeproTech), as de-scribed previously [13]. The linear detection range of the TGF-� bioassay was80–1250 pg/ml.

Flow cytometric analysis

For surface staining, �2 � 105 cells were washed in cold Hanks’ balancedsaline solution (Sigma Chemical Co.), containing 0.5% FBS, stained withpretitred amounts of culture supernatants or direct conjugates, washed, andincubated with the appropriate fluorescein isothiocyanate (FITC)-conjugated,preadsorbed secondary antibodies. For intracellular staining of BclXL, cellswere fixed with 4% paraformaldehyde (E Merck, San Diego, CA) and perme-abilized with 0.2% saponin (Sigma Chemical Co.) prior to staining. Flowcytometry was performed on FACScan (Becton Dickinson, San Jose, CA) usingCellQuest (Becton Dickinson) software for acquisition and WinList (Verity,Topsham, ME) software for analysis. Debris and cellular fragments wereexcluded from the analysis by electronic logical gates based on forward- andside-scatter profiles. Cell-cycle analysis was performed using propidium iodide(PI; Sigma Chemical Co.), as reported previously [13]. Production of peroxides,

2 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

peroxinitrites, and other reactive oxygen species (ROS) was assessed using theoxidation-sensitive fluorescent probe 2�,7�-dichlorofluorescein diacetate(DCFDA). Cells treated under the different conditions were incubated with 2.5�M DCFDA for 20 min and acquired on a FACScan. The total membranepotential was measured on a FACScan by incubating cells with the membranepotential-responsive carbocyanine dye dihexa-oxacarbocyanine iodide (DiOC6;Sigma Chemical Co.), as reported previously [13].

RESULTS

Distinct roles of CTLA4-CD80/CD86 interactionsduring activation of mouse CD4� T lymphocytesare dependent on the dose of Con A

Purified CD4� T cells were obtained from mouse lymph nodesand activated with different amounts of Con A [38] in thepresence of control antibody, anti-CD28, anti-CTLA4, ormCTLA4 (Fig. 1A). At extremely suboptimal doses (0.2 �g/ml), little proliferation was observed. In keeping with theknown role of CD28 in reducing the threshold of primary signalrequired for activation, T cells activated with a suboptimaldose of Con A and anti-CD28 showed increased proliferation.Con A, at stimulatory doses (e.g., 1 �g/ml), increased T cellproliferation, which was further enhanced with anti-CD28.However, at suppressive doses of Con A (3–4.25 �g/ml),proliferation of CD4� T cells together with control antibodywas suppressed greatly, and triggering with anti-CD28 was notable to rescue proliferation. We used this system to study theroles of CTLA4-CD80/CD86 interactions, using soluble anti-CTLA4, which blocks the interaction of CTLA4 with CD80/CD86, and the monovalent reagent mCTLA4hIgG1, whichbinds CD80 and CD86 and blocks their interactions with CD28and CTLA4 [41]. Activation of T cells with a stimulatory doseof Con A and soluble anti-CTLA4 or mCTLA4IgG1 decreasedproliferation compared with control antibody. However, as ConA concentrations were increased (2.5–4.25 �g/ml), this inhi-bition of proliferation disappeared gradually, and enhancedproliferation was observed. At suppressive doses of Con A,which are known to enhance T cell death [38], blockade ofCTLA4-CD80/CD86 interactions enhanced T cell proliferationover and above that in cells treated with control antibody oranti-CD28. Similar results were obtained with a combination ofblocking antibodies to CD80 and CD86, which mimicked theeffects of mCTLA4hIgG1 (Fig. 1B). However, this effect wasnot observed with antibodies against the adhesion moleculeCD62L (Fig. 1B), demonstrating that the effects were specificto CTLA4-CD80/CD86 interactions. These results clearly dem-onstrated that CTLA4-CD80/CD86 interactions on CD4� Tcells modulated T cell activation depending on the dose of ConA used for T cell activation.

CTLA4-CD80/CD86 interactions are requiredearly during CD4� T cell activation formanifestation of Con A responses

Next, we studied the activation kinetics of CD4� T cellsstimulated with a stimulatory or a suppressive dose Con A (Fig.1C). The optimal proliferation of CD4� T cells activated witha stimulatory dose of Con A was observed at 36 h afteractivation. In keeping with the known ability of CD28 to

enhance and sustain T cell-proliferative responses, CD4� Tcells activated with a stimulatory dose of Con A and anti-CD28displayed high, proliferative responses, optimally at 48 h afteractivation. Similar to Figure 1, A and B, blockade of CTLA4-CD80/CD86 interactions reduced proliferation in T cells acti-vated with a stimulatory dose of Con A. Also, T cells activatedwith a suppressive dose of Con A, together with control anti-body or anti-CD28, displayed little proliferation. However,blockade of CTLA4-CD80/CD86 interactions greatly enhancedproliferative responses in T cells activated with a suppressivedose of Con A, optimally after 36 h post-activation. To under-stand the roles of these interactions during CD4� T cellactivation, anti-CD28, anti-CTLA4, or mCTLA4hIgG1 wasadded at 0, 3, 6, and 12 h of T cell activation with a stimulatoryor a suppressive dose of Con A (Fig. 1D). The modulation of Tcell proliferation as a result of CTLA4-CD80/CD86 blockade,with a stimulatory or a suppressive dose of Con A, was mostevident when blockade was applied early, i.e., within 6 h ofactivation. It is possible that a suppressive dose of Con Agenerated an overly strong, primary signal that reduced T cellproliferation. However, blockade of CTLA4-CD80/CD86 inter-actions enhanced T cell proliferation in the presence of thisstrong signal, suggesting that these interactions played aninhibitory role. In keeping with the same logic, the reducedproliferation with blockade of CTLA4-CD80/CD86 interactionswith a stimulatory dose of Con A suggests that these interac-tions enhanced T cell activation under this condition.

Costimulatory receptors and their ligands areexpressed on CD4� T cell activation with Con A

We studied the expression of cell-surface molecules on lymphnode primary mouse CD4� T cells before and after activationwith different doses of Con A (Fig. 2). Flow cytometric anal-ysis revealed that the vast majority of cells was CD4� andCD3�. CD28 was constitutively expressed on T cells, andincreased levels were detected on activation with a stimulatoryor a suppressive dose of Con A. CTLA4 was not detected onunstimulated cells but was induced by 12 h of activation. Asimilar profile was observed with CD80, which was present atlow levels on unactivated T cells. CD86 was also stronglyup-regulated upon activation with a stimulatory and a suppres-sive dose of Con A. Thus, in this T cell:T cell interactionmodel, costimulatory receptors and ligands are present onCD4� T cells after activation with Con A.

CTLA4-CD80/CD86 interactions modulate levelsof ROS, membrane potential, cell-cycleprogression, and death, depending on the doseof Con A used for activation

We measured different phenotypic manifestations of activationof primary CD4� T cells activated with a stimulatory or asuppressive dose of Con A together with different antibodies.Intracellular ROS was detected using the oxidation-sensitivefluorescent probe DCFDA, which results in increased fluores-cence upon oxidation. As seen in Figure 3A, T cells activatedwith a suppressive dose of Con A displayed over 2.5-fold moreintracellular ROS compared with T cells activated with astimulatory dose of Con A and control antibody. Treatment with

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 3

anti-CD28 did not significantly change ROS levels with astimulatory or a suppressive dose of Con A. With a stimulatorydose of Con A, levels of ROS increased on blockade of CTLA4-CD80/CD86 interactions. This profile was reversed with a

suppressive dose of Con A, where the blockade reduced ROSto levels observed in T cells activated with a stimulatory doseof Con A and control antibody. The ROS profile observed,under different conditions, is similar at 24 h and 36 h of

Fig. 1. Plasticity of CTLA4-CD80/CD86 interactions on primary CD4� T cells depends on the dose of Con A used for activation. (A) Primary lymph node mouseCD4� T cells were isolated and activated with different concentrations of Con A together with control antibody (5 �g/ml), anti-CD28 (aCD28; 0.3 �g/ml),anti-CTLA4 (1:100), or mCTLA4hIgG1 (1:100) for 36 h and pulsed for 12 h with 3H-thymidine. (B) Different amounts of purified antibodies against CD80 andCD86 were added to culture at 0 h and compared with anti-CTLA4 and mCTLA4 (1:100), control antibody (Ctrl Ab), or anti-CD62 ligand (CD62L). (C) The kineticsof T cell activation was studied with a stimulatory (1 �g/ml) or a suppressive (4 �g/ml) dose of Con A together with antibodies for the indicated duration, includingthe final 12-h pulse with 3H-thymidine. (D) CD4� T cells were activated with a stimulatory or a suppressive dose of Con A, and control antibody, anti-CD28,anti-CTLA4, or mCTLA4hIgG1 was added at the indicated time-points. Cells were harvested after 48 h of activation. cpm, Counts per minute.

4 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

activation (Fig. 3B). Next, we used DiOC6 to measure mem-brane potential changes in these cells; this dye accumulatesprimarily in mitochondria that maintain mitochondrial mem-brane potential, thereby showing high fluorescence. T cellstreated with a stimulatory dose of Con A and control antibodydisplayed �1.5-fold higher fluorescence compared with cells

activated with a suppressive dose of Con A (Fig. 3C). Activa-tion of T cells with a suppressive dose of Con A and anti-CD28increased membrane potential, but only a slight increase wasobserved with a stimulatory dose of Con A. CTLA4-CD80/CD86 blockade of T cells slightly reduced membrane potentialin cells activated with a stimulatory dose of Con A at 24 h ofactivation but resulted in over 60% enhancement when acti-vation was carried out with a suppressive dose of Con A. Asobserved in Figure 3D, significant reduction in membranepotential was observed in T cells activated for 36 h with astimulatory dose of Con A together with CTAL4-CD80/CD86blockade. The kinetics of ROS production (Fig. 3B) and mem-brane potential (Fig. 3D) demonstrates that excess ROS isproduced before reduction in membrane potential in T cellsactivated with a stimulatory dose of Con A together withCTLA4-CD80/CD86 blockade.

Next, PI staining, under different conditions of activationcells in the S/G2M phase of the cell cycle and hypodipoid cells,was detected by CD4� T lymphocytes activated with a sup-pressive dose of Con A, and control antibody showed a reducedproportion of actively cycling cells and an enhanced hypodip-loid population compared with cells activated with a stimula-tory dose of Con A after 48 h of activation (Fig. 3, E and F),which is consistent with the results obtained with kinetics ofproliferation (Fig. 1C). No major differences were observed onCD28 triggering in conjunction with a suppressive dose of ConA. However, T cells activated with a stimulatory dose of Con Aand anti-CD28 demonstrated the highest proportion of activelycycling cells and the lowest proportion of hypodiploid cells(Fig. 3, E and F). Conversely, T cells activated with a stimu-latory dose of Con A and CTLA4-CD80/CD86 blockadeshowed a decreased proportion of cycling cells, with a con-comitant, �1.4-fold increase in hypodiploid cells over timecompared with the control population. T cells activated with asuppressive dose of Con A and CTLA4-CD80/CD86 blockadedemonstrated an �3.5-fold increase in the actively cyclingpopulation, coupled with the reduced hypodiploid population.Thus, an inverse correlation was observed across multiplephenotypic effects on blockade of CTLA4-CD80/CD86 inter-actions in primary CD4� T cells activated with stimulatory orsuppressive doses of Con A.

Differential expression of BclXL and functionalROS-mediated effects during Con A-mediatedactivation of primary CD4� T cells

The decreased proliferation and increased T cell death inCD4� T cells activated with a suppressive dose of Con A ledus to address the roles of Fas, caspase, oxidative stress, andBclXL levels in this culture system [42, 43]. We studied theroles of Fas, using lpr�/� mice, which contain a point mutationin Fas, rendering it nonfunctional. In other studies, the pan-caspase inhibitor, BDfmk, was used to study the roles ofcaspases. CD4� T cells were activated for 36 h with P � I,washed, and rested for 36 h as a positive control. Reactivationof these T cells with plate-bound anti-CD3 resulted in 41% ofhypodiploid cells in lpr�/�, whereas death was reduced to 23%in lpr�/� T cells (data not shown), clearly demonstrating therole for Fas during activation-induced T cell death (AICD). Inthe same model of AICD, T cell death in cultures treated with

Fig. 2. Primary mouse CD4� T cells express surface costimulatory receptorsand their ligands upon Con A stimulation. Lymph node CD4� T cells (0 h)were activated with a stimulatory (1 �g/ml) or a suppressive (4 �g/ml) dose ofCon A for 12 h or 42 h and stained with specific mAb to different cell-surfacemarkers followed by flow cytometric analysis. The light-gray, dotted linesindicate control antibody; solid gray lines with gray arrows indicate unacti-vated cells; thin black lines with black arrows indicate 12 h-activated cells;and solid black lines with large arrowheads indicate cells activated for 42 h.Mean fluorescence intensities (MFIs) are indicated in three rows representing0, 12, and 42 h of activation. This pattern of expression is representative of fourindependent experiments.

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 5

BDfmk was 37% compared with 62% hypodiplody in culturestreated with the control peptide, zFAfmk, demonstrating thatcaspase activity is required for AICD (data not shown). How-ever, the use of lpr�/� mice or the use of BDfmk failed tomodulate proliferation or cell death (data not shown), demon-

strating that Fas-Fas ligand (FasL) binding and caspase activ-ity were not involved during Con A-mediated T cell activation(Fig. 4, A and B).

Next, we studied the expression of the survival factor BclXL

in CD4� T cells activated under different conditions. There

6 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

was over threefold reduced expression of BclXL in cells treatedwith a suppressive dose of Con A and control antibody, ascompared with those treated with a stimulatory dose of Con A.CD28 promotes T cell survival by enhancing BclXL levels [44].Although T cells activated with a suppressive dose of Con Aand anti-CD28 increased intracellular expression of BclXL

(Fig. 4C), these levels were probably insufficient to enhance Tcell cycling and survival on activation with this dose of Con A(Fig. 3E). CTLA4-CD80/CD86 blockade with a suppressivedose of Con A resulted in an approximate fourfold increase inBclXL levels over the control population. However, with astimulatory dose of Con A, an �2.5-fold decrease in BclXL

levels was observed on CTLA4-CD80/CD86 blockade. Thus,CTLA4-CD80/CD86 interactions play an important role inmodulating the expression of BclXL.

To directly establish a functional link between ROS and Tcell cycle progression and survival [42], T cells were activatedwith a stimulatory or a suppressive dose of Con A in thepresence of different antioxidants (Fig. 4D). GSH reducesincreased oxidation in cells, exogenous catalase specificallycleaves excess H2O2 that diffuses out of the cell, and L-NMAis an inhibitor of nitric oxide (NO) synthase. In T cells acti-vated with a stimulatory dose of Con A, these agents did notcause any significant effects on the proportion of cycling cellsor T cell death across multiple experiments (Fig. 4E). However,in T cells activated with a suppressive dose of Con A, theseagents significantly increased T cell cycling and reduced theproportion of hypodiploid T cells (Fig. 4, D and E). Theseresults clearly implicate H2O2 and NO in enhanced primaryCD4� T cell death on activation with a suppressive dose of ConA. Together, these results demonstrate that T cell death with asuppressive dose of Con A is dependent on ROS but is inde-pendent of Fas and caspase activity.

Functional roles of inhibitory cytokines, TGF-�and IL-10, during Con A-mediated activation ofCD4� T cells

TGF-� is known to decrease cell-cycle progression and en-hance death in T cells, and there are several studies about theroles of CTLA4 ligation in TGF-� production [23–25]. T cellsactivated with a suppressive or a stimulatory dose of Con Aproduced similar amounts of TGF-�, and triggering with anti-CD28 did not modulate TGF-� levels on activation with asuppressive dose of Con A (Table 1). However, on activationof T cells with a stimulatory dose of Con A and anti-CD28,TGF-� levels were not detectable. This effect is similar to theeffect of CD28 in greatly reducing the production of activeTGF-� on activation of T cells with P � I [13]. Increased levels

of TGF-� were produced T cells activated with suppressive andstimulatory doses of Con A-mediated activation and CTLA4-CD80/CD86 blockade. In fact, CTLA4 blockade with a stim-ulatory dose of Con A resulted in over a 4.5-fold increase inactive TGF-� levels and appeared to be consistent with resultsobtained with respect to T cell cycling and survival (Figs. 1 and3). However, the �2.7-fold increase in active TGF-� observedin T cells activated with a suppressive dose of Con A andCTLA4-CD80/CD86 blockade was inconsistent with the resultson T cell cycling and survival (Figs. 1 and 3) and led us tocheck for functional effects of TGF-� in this system. Additionof a neutralizing antibody to TGF-�1, the dominant isoform ofTGF-� in T cells, or fetuin, which binds to TGF-� and seques-ters it from TGF-� receptors [45], did not display any signif-icant effects on T cell activation with a stimulatory or asuppressive dose of Con A (Fig. 5A). Although high levels ofactive TGF-� are produced on Con A activation and CTLA4-CD80/C86 blockade, no functional role for TGF-� was de-tected in this system. This is unlike the situation in P �I-mediated activation of T cells, where the addition of fetuin oranti-TGF-�1 results in �30% rescue of proliferation observedon CTLA4 blockade [13]. These results led us to study thefunctional role of another major inhibitory cytokine, IL-10 (Fig.5B). Dose-dependent rescue of inhibition of proliferation onCTLA4-CD80/CD86 blockade with a stimulatory dose of Con Awas observed on neutralization using anti-IL-10, demonstratingthat IL-10 played an important role. In T cells activated with asuppressive dose of Con A, a slight increase in the level ofproliferation was observed with anti-IL10, but there was nofundamental change in the activation profile as that observedwith a stimulatory dose of Con A and CTLA4-CD80/CD86blockade. In summary, CTLA4-CD80/CD86 interactions, insuppressive and stimulatory-dose activation of T cells by ConA, inhibited TGF-� production; however, TGF-� does not playa functional role here. In fact, IL-10 is functionally moreimportant during activation of T cells with a stimulatory dose ofCon A and CTLA4-CD80/CD86 blockade.

Blockade of CTLA4-CD80/CD86 interactions onCD4� T cells and activation with a suppressivedose of Con A results in largely IL-2-independentbut CsA-sensitive proliferation

IL-2 is a key cytokine responsible for T cell proliferation.The amount of IL-2 secreted in culture supernatants wasdetected using a specific IL-2 ELISA. As seen in Table 1, Tcells activated with a stimulatory or a suppressive dose ofCon A produced similar amounts of IL-2, and these wereenhanced with anti-CD28. This was quite unlike the prolif-

Fig. 3. Distinct differences in levels of ROS, membrane potential, cell-cycle progression, and hypodiploid population are observed on activation of primary CD4�

T cells with a stimulatory (1 �g/ml) or a suppressive (4 �g/ml) dose of Con A together with blockade of CTLA4-CD80/CD86 interactions. (A) Total levels of ROSwere detected using DCFDA and flow cytometric analysis profiles after 24 h of activation are depicted, and numbers in the plot indicate MFI. (B) The relative levelsof ROS under different conditions after 24 h or 36 h of activation are represented as mean � SE of three independent experiments. (C) Membrane potential wasstudied using DiOC6 fluorescence after 24 h of activation. (D) The relative changes in membrane potential under different conditions at 24 h and 36 h of activationare represented as mean � SE of three independent experiments. (E) Cell-cycle progression and hypodiploidy were studied using PI staining of CD4� T cells,activated with a stimulatory or a suppressive dose of Con A together with the indicated antibodies. (F) A kinetic profile of cell cycling and hypodiploidy with mean �SE of four independent experiments is also depicted for T cells activated with a stimulatory and a suppressive dose of Con A. The results are normalized to cellstreated with control antibody alone. The numbers in the plots represent the mean percentage of the cell population in G0/G1, S/G2M, and hypodiploid phases,respectively.

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 7

Fig. 4. ROS-mediated cell death and BclXL expression depend on the dose of Con A used for CD4� T cell activation. (A) Lymph node CD4� T cells fromIpr�/� or Ipr�/� mice were activated for 36 h in the presence of a stimulatory (1 �g/ml) or a suppressive (4 �g/ml) dose of Con A, and an index ofproliferation was obtained. (B) Cells were treated at 0 time with the pan-specific caspase inhibitor BDfmk or the control peptide zFAfmk. 3H-Thymidinecounts are depicted after activation for 36 h with a 12-h pulse. (C) Intracellular staining of the survival factor BclXL was followed by flow cytometric analysisat �36 h of activation. The numbers in the plots indicate the MFI of the population. (D) T cells were activated with a stimulatory or a suppressive doseof Con A in the absence or presence of different antioxidants for 48 h, and cell-cycle analysis was performed. The antioxidants were used at the followingconcentrations: GSH, 50 �M; catalase (Cat), 10 �g/ml; L-NMA, 250 �M. The numbers in the plots represent the mean percentage of the cell populationin G0/G1, S/G2M, and hypodiploid phases, respectively. (E) The effect of antioxidants in modulating T cell cycling and hypodiploidy under differentconditions is represented as mean � SE of three independent experiments.

8 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

eration profile, where T cell activation with a stimulatorydose of Con A and control antibody or anti-CD28 displayedmuch higher proliferation compared with corresponding cul-tures treated with a suppressive dose of Con A (Fig. 1).Conversely, anti-CTLA4 or mCTLA4hIgG1 treatment ofCD4� T cells decreased IL-2 production on activation witha stimulatory and a suppressive dose of Con A (Table 1).Although the decreased IL-2 production is consistent withthe inhibition of proliferation observed on CTLA4-CD80/CD86 blockade in T cells activated with a stimulatory doseof Con A, it is inconsistent with the enhanced proliferationobserved on CTLA4-CD80/CD86 blockade and activation

with a suppressive dose of Con A. These observations led usto study the functional role of IL-2 during activation ofCD4� T cells with different doses of Con A. The addition ofexogenous IL-2 completely rescued the inhibition of prolif-eration observed with a stimulatory dose of Con A togetherwith CTLA4-CD80/CD86 blockade (Fig. 6A). This rescuein proliferation was also mediated with exogenous IL-4,although higher amounts were required compared with IL-2.It is most important that the addition of IL-2 or IL-4 did notresult in any significant increase in proliferation of T cellsactivated with a suppressive dose of Con A. To furtherconfirm the roles of IL-2 in this system, we used neutralizingantibodies against IL-2 (Fig. 6B). Increased amounts ofanti-IL-2 but not anti-IL-4 greatly reduced the proliferationof T cells activated with a stimulatory dose of Con A,demonstrating the key role of IL-2. However, only a mildreduction in proliferation was observed in T cells activatedwith a suppressive dose of Con A and blockade of CTLA4-CD80/CD86. Together, these results demonstrate that pro-liferation of T cells activated with a suppressive dose of ConA with CTLA4-B7 blockade was largely IL-2-independent,whereas T cell proliferation with a stimulatory dose of ConA was largely dependent on IL-2. Finally, we used CsA todetermine the importance of the calcineurin pathway in thissystem (Fig. 6C). The inhibition observed with CsA additionin Con A-mediated activation was comparable with that seenwith plate-bound anti-CD3 (data not shown). Thus, T cellsactivated with Con A, in the absence or presence of CTLA4blockade, used the CsA-sensitive calcineurin TCR signaltransduction pathway involving nuclear factor activated Tcell signaling to proliferate.

Fig. 5. IL-10 but not TGF-� plays a major, functional role in CD4� T cell activation with a stimulatory (1 �g/ml) dose of Con A and CTLA4-CD80/CD86 blockade.Cells were treated with anti-TGF-�1 or fetuin (A) or with anti-IL-10 (B) in the presence or absence of blockade of CTLA4-CD80/CD86 interactions. Cells wereactivated for 36 h, followed by a 12-h pulse with 3H-thymidine.

TABLE 1. Differences in Production of IL-2 and TGF-� onBlockade of CTLA4-CD80/CD86 Interactions by CD4� T CellsActivated with a Stimulatory or a Suppressive Dose of Con A

IL-2 (pg/ml) Con A (1 �g/ml) Con A (4 �g/ml)

Control Ab 1013 � 41 1115 � 31�CD28 2005 � 48 2000 � 24�CTLA4 668 � 28 638 � 42�mCTLA4 670 � 10 648 � 35

TGF-� (pg/ml)

Control Ab 265 � 21 230 � 18�CD28 nd 209 � 4�CTLA4 1258 � 15 653 � 59�mCTLA4 1280 � 17 639 � 65

Culture supernatants were collected at 36 h post-activation. The data arerepresentative of three independent experiments with mean � SE. nd, Notdetected.

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 9

DISCUSSION

Con A is a known T cell mitogen, but the mechanisms involvedin this process are poorly studied. APCs are thought to berequired for Con A-mediated T cell activation [35, 37]. How-ever, another study demonstrated that purified T cells areactivated by Con A without the requirement for APCs [36]. Wefind that primary mouse CD4� T cells expressed costimulatoryreceptors CD28 and CTLA4 and their ligands CD80 and CD86after activation with Con A (Fig. 2). These cells responded toCon A depending on the concentration used [38]: A low dose ofCon A was stimulatory for CD4� T cells, whereas a high dose

of Con A suppressed proliferation followed by increased T celldeath (Figs. 1 and 3). To study the roles of costimulatoryreceptors CD28 and CTLA4 and their ligands CD80 and CD86on CD4� T cells activated with a stimulatory or a suppressivedose of Con A, three independent reagents were used. Anti-CD28 binds and signals via CD28, together with a stimulatorydose of Con A, to enhance T cell cytokine production andproliferation. It is interesting that anti-CD28 triggering, onactivation with a suppressive dose of Con A, was not able toenhance cell survival, perhaps as a result of the dominanteffects of the CTLA4-CD80/CD86 interactions in this system(Fig. 1). It is also possible that CD28 signaling in the presence

Fig. 6. Blockade of CTLA4-CD80/CD86 interactions on CD4� T cells results in mainly IL-2-independent proliferation on activation with a suppressive (4 �g/ml)dose of Con A. Exogenous IL-2 or IL-4 (A) or neutralizing antibodies to IL-2 or IL-4 (B) in the indicated concentrations were added at 0 h to cells stimulated witha stimulatory or a suppressive dose of Con A in the presence of different antibodies. (C) CsA was added at the indicated concentrations at 0 h, and the inhibitionof proliferation after activation for 36 h, followed by a 12-h pulse with 3H-thymidine, is depicted.

10 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

of a strong signal curtails T cell responses by inducing apo-ptosis [46]. CD28-CD80/CD86 interactions do not appear to beplaying a major role, perhaps as a result of the fact that it doesnot bind effectively with hypoglycosylated B7 moleculespresent on mouse T cells [26, 27]. As all our results areconsistent with soluble anti-CTLA4 and mCTLA4hIgG1, mostlikely the major interactions observed are a result of CTLA4binding to CD80/CD86. It is unlikely that CD80 and CD86 aresignaling in this T cell:T cell system, as no major effects wereobserved on cross-linking with anti-CD80 and anti-CD86 (datanot shown). It was shown previously that mCTLA4IgG1 doesnot reduce Con A binding to cell-surface molecules [37].Consistent with this study, Con A binding to cell-surfacemolecules was not affected with anti-CTLA4 or mCTLA4IgG1,as preincubating, activated cells with Con A did not decreasestaining with FITC-conjugated anti-CTLA4, anti-CD80, oranti-CD86 (data not shown).

The roles of CTLA4-CD80/CD86 interactions post-activa-tion with a stimulatory or a suppressive dose of Con A aresummarized in Figure 7. Binding of CTLA4 to CD80/CD86 onCD4� T cells activated with a stimulatory dose of Con A-en-hanced T cell cycling and survival, as blocking these interac-tions led to increased ROS and TGF-� levels but decreasedIL-2. This is similar to the roles of CTLA4-CD80/CD86 inter-actions on T cells activated with P � I [13]. It is important topoint out that CD4� T cell-proliferative responses to a stimu-latory dose of Con A were reduced but were not abrogatedcompletely (Fig. 1) on blocking CTLA4-CD80/CD86 interac-tions, suggesting roles for other cell-surface proteins in thisprocess. Also, in T cells activated with a stimulatory dose ofCon A and CTLA4-CD80/CD86 blockade, excess ROS wasproduced by 24 h (Fig. 3, A and B), followed by reduction inmembrane potential (Fig. 3, C and D) and reduced proliferationand cell cyling (Figs. 1C and 3F) by 36–48 h. A more dramaticphenotype was observed with T cells activated with a suppres-sive dose of Con A, which resulted in increased ROS produc-tion and greatly inhibited CD4� T cell activation. BlockingCTLA4-CD80/CD86 responses under this condition led togreatly increased T cell cycling and survival, which was largelyIL-2 independent. This pathway was sensitive to CsA, demon-

strating dependence on the phosphatase calcineurin. IL-2-deficient T cells are activated with anti-CD3 and B7-trans-fected cells, demonstrating that IL-2 is not essential for B7-induced T cell proliferation [47]. Also, CD28-mediatedproliferation of T cells involves IL-2-dependent and indepen-dent pathways [48]. It is most interesting that CD4� T cellhyperproliferation in ctla4�/� mice is IL-2-independent [49].Our results demonstrate that the roles of CTLA4-CD80/CD86interactions differ based on the strength of activation of T cells.Thus, on activation with a suppressive dose of Con A, CTLA4-CD80/CD86 responses inhibit T cell responses, consistent withits role as a negative regulator during T cell activation [1–3].These studies clearly demonstrate that CTLA4 is an “intelli-gent”, costimulatory receptor that integrates signal strengthinformation to modulate primary mouse CD4� T cell activa-tion.

The role of CTLA4 in survival of T cells is controversial.CTLA4 ligation of activated T cells results in death [15, 16]and enhances -irradiation-induced apoptosis [17]. However,CTLA4 expression on differentiated T cells increases T cellsurvival in the AICD model [12, 14]. In primary CD4� T cells,activation by anti-CD3 and ligation of CTLA4 reduced T cellcycling without affecting survival [6]. However, in CD4� Tcells expressing a transgenic TCR, CTLA4 was found to reduceT cell expansion and survival [10]. Also, CD4� T cells acti-vated with P � I, together with blocking CTLA4-CD80/CD86interactions, reduced T cell cycling and survival [13]. The roleof CTLA4 in survival of primary T cells is unclear, and thesediscrepancies may be a result of the use of different popula-tions of T cells (primary vs. differentiated) or T cells expressingTCRs with different affinities (anti-CD3 vs. a high-affinity,transgenic TCR) or a result of differences in activation condi-tions (P�I or anti-CD3). In this study, using primary CD4� Tcells, we clearly show that CTLA4-CD80/CD86 interactionsmodulate T cell cycling and survival depending on the dose ofCon A. On activation with stimulatory doses of Con A, CTLA4-CD80/CD86 interactions enhance T cell survival (Fig. 3, E andF). Although the effects are much reduced, it is similar to theroles of these interactions observed in the P � I system [13].However, the role of CTLA4-CD80/CD86 interactions in re-ducing T cell survival was uncovered on blocking CTLA4-CD80/CD86 interactions and activation of T cells with a sup-pressive dose of Con A. Notably, enhanced T cell death withsuppressive doses of Con A is Fas- and caspase-independent(Fig. 4, A and B). Previously, Con A was shown to inducesignaling ROS in mouse thymocytes, implicating flavonoidreduced nicotinamide adenine dinucleotide phosphate oxi-dase(s) but not NO synthase [39]. Here, we show that primarymouse CD4� T cells activated with a suppressive dose of ConA induced excess ROS (Fig. 3, A and B), resulting in oxidativestress. Functionally, H2O2 and NO are involved in this processas an exogenous catalase, and L-NMA, independently or incombination, rescued primary CD4� T cell cycling and sur-vival on activation with a suppressive dose of Con A (Fig. 4, Eand F). Together with the greatly modulated levels of BclXL

and the functional roles of oxidative stress, it appears thatenhanced primary mouse CD4� T cell death with a suppressivedose of Con A is not activation-induced cell death but mostlikely autonomous T cell death [42].

Fig. 7. The functional effects of CTLA4-CD80/CD86 interactions in a CD4�

T cell:T cell activation model depend on the dose of Con A used for primaryactivation. These interactions are beneficial at stimulatory but are inhibitory atsuppressive doses of Con A.

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 11

Although many studies have demonstrated that CTLA4-CD80/CD86 interactions result in increased TGF-� produc-tion, the functional roles of CTLA4-modulated TGF-� levelsare controversial [23–25]. Unlike other reports where CTLA4-mediated production of T cell inhibitory cytokines was limitedto specific subsets of regulatory T (Treg) cells, our observationsare likely to be a property of the global CD4� T cell popula-tion, as depletion of CD25�CD4� Treg cells prior to culturedid not have any effects on the activation profile on blockade(data not shown). Blockade of CTLA4-CD80/CD86 interactionsenhanced TGF-� production by CD4� T cells activated with P� I. The TGF-� produced in this system did play a functionalrole by enhancing T cell death [13]. In the current study,anti-CD28 triggering greatly reduced the amount of TGF-�produced by T cells stimulated with Con A. However, blockadeof CTLA4-CD80/CD86 interactions also enhanced activeTGF-� production in T cells activated with a stimulatory anda suppressive dose of Con A; however, no significant rescuewas observed with fetuin or specific anti-TGF-�1 antibodies.The reasons for the lack of a functional effect of TGF-� areunclear and led us to enquire about the role of another immu-nosuppressive cytokine, IL-10. A recent study demonstratedthat CTLA4 ligation reduces interferon- secretion by T cellsin an IL-10-dependent manner. Also, CTLA4-induced IL-10may play an important role in anti-tumor T cell responses [50].Although no major effect on neutralizing of IL-10 was observedwith a suppressive dose of Con A, significant rescue in prolif-eration was observed in T cells activated with a stimulatorydose of Con A, together with blockade of CTLA4-CD80/CD86interactions (Fig. 5B). Thus, IL-10 is the key suppressivecytokine and is likely to be a major player in the cell-cyclearrest observed in this system. This is unlike the P � I-medi-ated activation system [13], where IL-10 does not play a role(data not shown). As CD4� T cells are known to expressdifferent cytokines depending on varying doses of antigen [51],it is possible that expression of these immunosuppressivecytokines may differ depending on the primary activation con-ditions. Further studies are required to address the expression,cross-talk, and functional roles of TGF-� and IL-10 on T cellsactivated under different conditions.

There are at least three studies that demonstrate a role forCTLA4 in modulating immune responses depending on thestrength of signal. First, higher levels of CTLA4 accumulationat the immunological synapse are found with increased signalstrength. Thus, a strong signal results in increased CTLA4surface expression, which inhibits T cell activation [52]. Sec-ond, stimulation of T cells with high concentrations of antigen,together with CTLA4 blockade, favors a T helper cell type 2response [32]. Finally, a role for CTLA4 in enhancing T cellresponses was demonstrated in a study using an autoimmuneencephalitis model. Here, immunization with a disease antag-onistic peptide, but not a disease agonistic peptide, togetherwith CTLA4 blockade inhibit generation of cross-reactive Tcell clones. It is possible that antagonistic peptides generatesub-optimal primary signals, and CTLA4 interactions enhanceT cell responses under this condition [33]. It is most importantthat CTLA4 blockade inhibits or enhances the generation of Tcells expressing distinct TCRs of identical specificities [32].Our results about the roles of CTLA4-CD80/CD86 interactions

in modulating T cell activation, based on the dose of Con Aused for activation, are consistent with these studies. However,validation of this model in other T cell activation systems isrequired. This point is particularly relevant for studies on Tcell activation, as Con A binds to additional glycoproteinsurface receptors in addition to the TCR [53]; hence, thefunctional consequences of activating T cells with Con A oranti-CD3 may be different [41]. We find that increasing iono-mycin concentration, but not PMA, on activation of CD4� Tcells with P � I [13] abrogates T cell cycle arrest and death onblocking CTLA4-CD80/CD86 interactions by inducing highlevels of IL-2 (data not shown). Preliminary experiments (datanot shown) suggest that CTLA4-CD80/CD86 interactions en-hance CD4� T cell activation after activation with solubleanti-CD3 cross-linking, whereas these act in an inhibitorymanner in the presence of plate-bound anti-CD3, which sendsa stronger signal [54].

Two mutually nonexclusive models have been proposed toexplain the role of CTLA4 during immune responses, thresh-old, and attenuator [2]. The former predicts that CTLA4 sets astimulatory threshold for optimal T cell activation. Thus, in theabsence of CTLA4 or with CTLA4 blockade, T cells proliferatein response to weak activation signals (e.g., TCR-MHC signalsduring peripheral survival of CD4� T cells). Conversely, in thepresence of a strong signal (e.g., inflammation by pathogens),CTLA4 regulates the extent of T cell division after initialactivation. In this case, in the presence of CTLA4 blockade orctla4�/�, cells divide with greater frequency. Both these mod-els, with some modifications, appear to be at play in this study.It is possible that with a stimulatory dose of Con A, thethreshold model plays an important role, as CTLA4 integratesthe signal strength to enhance T cell activation. Conversely, itis possible that on T cell activation with suppressive doses ofCon A, the attenuator model comes into play. Therefore, block-ing CTLA4-CD80/CD86 interactions reduces signal strength toenhance T cell cycling and survival. Based on these results, wesuggest that CTLA4-CD80/CD86 interactions on CD4� T cellsenhance T cell activation in the presence of a suboptimal orstimulatory signal, whereas CTLA4-CD80/CD86 inhibits T cellactivation in the presence of a strong or overly strong signal. Itis possible that CTLA4-CD80/CD86 interactions inhibit thegeneration of dominant TCRs, which recognize antigens with ahigh affinity, whereas the same interactions may enhance theproliferation of T cells with lower affinities to enlarge the T cellimmune response. These results are consistent with studiesthat demonstrate CTLA4-CD80/CD86 interactions regulate thediversity and the extent of the primary CD4� T cell immuneresponse [2, 32, 33, 52]. In summary, this study clearly dem-onstrates that CTLA4-CD80/CD86 interactions on primarymouse CD4� T cells modulate T cell activation depending onthe dose of Con A. This data implicate the strength of primarysignal in conjunction with CTLA4-CD80/CD86 interactions tomodulate primary T cell responses. Further studies are re-quired to understand the mechanisms by which CTLA4, asingle receptor, can switch from an enhancer to a negativemodulator of CD4� T cell cycling and survival, depending onthe signal strength used for activation.

12 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org

ACKNOWLEDGMENTS

This study was supported by a grant from the Department ofBiotechnology, Government of India. S. M. was awarded aresearch fellowship from CSIR. We are grateful to Dr. S. Rathfor insightful comments about this study. Dr. P. Kondaiahprovided the CCL64 cell line for the TGF-� bioassay andneutralizing antibodies to TGF-�. We thank Prof. T. Ra-masarma for suggestions on oxidative stress, Dr. A. Sarin forcaspase-inhibitors, and Drs. V. Bal and S. Rath for access tolpr�/� mice. The assistance of Dr. O. Joy and H. Krishnan,DBT-FACS facility, is highly appreciated.

REFERENCES

1. Alegre, M-L., Frauwirth, K. A., Thompson, C. B. (2001) T cell regulationby CD28 and CTLA4. Nat. Rev. Immunol. 1, 220–228.

2. Egen, J. G., Kuhns, M. S., Allison, J. P. (2002) CTLA-4: new insights intoits biological function and use in tumor immunotherapy. Nat. Immunol. 3,611–618.

3. Rudd, C. E., Schneider, H. (2003) Unifying concepts in CD28, ICOS andCTLA4 co-receptor signaling. Nat. Rev. Immunol. 3, 544–556.

4. Masteller, E. L., Chuang, E., Mullen, A. C., Reiner, S. L., Thompson, C. B.(2000) Structural analysis of CTLA4 function in vivo. J. Immunol. 164,5319–5327.

5. Carreno, B. M., Bennett, F., Chau, T. A., Ling, V., Luxenberg, D., Jussif,J., Baroja, M. L., Madrenas, J. (2000) CTLA-4 (CD152) can inhibit T cellactivation by two different mechanisms depending on its level of cellsurface expression. J. Immunol. 165, 1352–1356.

6. Krummel, M. F., Allison, J. P. (1996) CTLA-4 engagement inhibits IL-2accumulation and cell cycle progression upon activation of resting T cells.J. Exp. Med. 183, 2533–2540.

7. Calvo, C. R., Amsen, D., Kruisbeek, A. M. (1997) Cytotoxic T lymphocyteantigen 4 (CTLA-4) interferes with extracellular signal-regulated kinase(ERK) and Jun NH2-terminal kinase (JNK) activation, but does not affectphosphorylation of T cell receptor � and ZAP70. J. Exp. Med. 186,1645–1653.

8. Brunner, M. C., Chambers, C. A., Chan, F. K-M., Hanke, J., Winoto, A.,Allison, J. P. (1999) CTLA4-mediated inhibition of early events of T cellproliferation. J. Immunol. 162, 5813–5820.

9. Krummel, M. F., Allison, J. P. (1995) CD28 and CTLA-4 have opposingeffects on the response of T cells to stimulation. J. Exp. Med. 182,459–465.

10. Doyle, A. M., Mullen, A. C., Villarino, A. V., Hutchins, A. S., High, F. A.,Lee, H. W., Thompson, C. B., Reiner, S. (2001) Induction of CTLA4restricts clonal expansion of helper T cells. J. Exp. Med. 194, 893–902.

11. Blair, P. J., Riley, J. L., Levine, B. L., Lee, K. P., Craighead, N.,Francomano, T., Perfetto, S. J., Gray, G. S., Carreno, B. M., June, C. H.(1998) CTLA-4 ligation delivers a unique signal to resting human CD4 Tcells that inhibits interleukin-2 secretion but allows BclXL induction.J. Immunol. 160, 12–15.

12. de Rocha Dias, S., Rudd, C. E. (2001) CTLA-4 blockade of antigen-induced cell death. Blood 97, 1134–1137.

13. Mukherjee, S., Maiti, P. K., Nandi, D. (2002) Role of CD80, CD86 andCTLA4 on mouse CD4� T lymphocytes in enhancing cell cycle progres-sion and survival after activation with PMA and ionomycin. J. Leukoc.Biol. 72, 921–931.

14. Pandiyan, P., Gartner, D., Soezeri, O., Radbruch, A., Schulze-Osthoff, K.,Brunner-Weinzierl, M. C. (2004) CD152 (CTLA4) determines the unequalresistance of Th1 and Th2 cells against activation induced cell death bya mechanism requiring PI3 kinase function. J. Exp. Med. 199, 831–842.

15. Gribben, J. G., Freeman, G. J., Boussiotis, V. A., Rennert, P., Jellis, C. L.,Greenfield, E., Barber, M., Restivo Jr., V. A., Ke, X., Gray, G. S., et al.(1995) CTLA4 mediates antigen-specific apoptosis of human T cells. Proc.Natl. Acad. Sci. USA 92, 811–815.

16. Scheipers, P., Reiser, H. (1998) Fas-independent death of activated CD4�

T lymphocytes induced by CTLA-4 crosslinking. Proc. Natl. Acad. Sci.USA 95, 10083–10088.

17. Bergman, M. L., Cilio, C. M., Penha-Goncalves, C., Lamhamedi-Cherradi,S. E., Lofgran, A., Colucci, F., Lejon, K., Garchon, H. J., Holmerg, D.(2001) CTLA4�/� mice display T cell-apoptosis resistance resembling

that ascribed to autoimmune-prone non-obese diabetic (NOD) mice. J.Autoimmun. 16, 105–113.

18. Perez, V. L., Van Parijs, L., Biuckians, A., Zheng, X. X., Strom, T. B.,Abbas, A. K. (1997) Induction of peripheral T cell tolerance in vivorequires CTLA-4 engagement. Immunity 6, 411–417.

19. Walunas, T. L., Bluestone, J. A. (1998) CTLA-4 regulates toleranceinduction and T cell differentiation in vivo. J. Immunol. 160, 3855–3860.

20. Greenwald, R. J., Boussiotis, V. A., Lorsbach, R. B., Abbas, A. K., Sharpe,A. H. (2001) CTLA-4 regulates induction of anergy in vivo. Immunity 14,145–155.

21. Frauwirth, K. A., Alegre, M-L., Thompson, C. B. (2000) Induction of T cellanergy in the absence of CTLA-4/B7 interaction. J. Immunol. 164,2987–2993.

22. Frauwirth, K. A., Alegre, M-L., Thompson, C. B. (2001) CTLA4 is notrequired for induction of CD8� T cell anergy in vivo. J. Immunol. 167,4936–4941.

23. Chen, W., Jin, W., Wahl, S. M. (1998) Engagement of cytotoxic Tlymphocyte-associated antigen-4 (CTLA-4) induces transforming growthfactor � (TGF�) production by murine CD4� T cells. J. Exp. Med. 188,1849–1857.

24. Kato, T., Nariuchi, H. (2000) Polarization of naıve CD4� T cells towardthe Th1 subset by CTLA-4 costimulation. J. Immunol. 164, 3554–3562.

25. Sullivan, T. J., Letterio, J. J., van Elsas, A., Mamura, M., van Amelsfort,J., Sharpe, S., Metzler, B., Chambers, C. A., Allison, J. P. (2001) Lack ofa role for transforming growth factor-� in cytotoxic T lymphocyte antigen-4-mediated inhibition of T cell activation. Proc. Natl. Acad. Sci. USA 98,2587–2592.

26. Hollsberg, P., Scholz, C., Anderson, D. E., Greenfield, E. A., Kuchroo,V. K., Freeman, G. J., Hafler, D. A. (1997) Expression of a hypoglycosy-lated form of CD86 (B7.2) on human T cells with altered binding prop-erties to CD28 and CTLA4. J. Immunol. 159, 4799–4805.

27. Greenfield, E. A., Howard, E., Paradis, T., Nguyen, K., Benazzo, F.,McLean, P., Hollsberg, P., Davis, G., Hafler, D. A., Sharpe, A. H.,Freeman, G. J., Kuchroo, V. K. (1997) B7.2 expressed by T cells does notinduce CD28-mediated costimulatory activity but retains CTLA4 binding.J. Immunol. 158, 2025–2034.

28. Schweitzer, A. N., Sharpe, A. H. (1999) Mutual regulation between B7–1(CD80) expressed on T cells and IL-4. J. Immunol. 163, 4819–4825.

29. Taylor, P. A., Lees, C. J., Fournier, S., Allison, J. P., Sharpe, A. H., Blazar,B. R. (2004) B7 expression on T cells down-regulates immune responsesthrough CTLA-4 ligation via T-T interactions. J. Immunol. 172, 34–39.

30. Wu, Y., Guo, Y., Huang, A., Zheng, P., Liu, Y. (1997) CTLA-4-B7interaction is sufficient to costimulate T cell clonal expansion. J. Exp.Med. 185, 1327–1335.

31. Zheng, P., Yan, W., Yong, G., Lee, C., Liu, Y. (1998) B7-CTLA4 inter-action enhances both production of anti-tumor cytotoxic T lymphocytesand resistance to tumor challenge. Proc. Natl. Acad. Sci. USA 95, 6284–6289.

32. Anderson, D. E., Bieganowska, K. D., Bar-Or, A., Oliveira, E. M. L.,Carreno, B., Collins, M., Hafler, D. A. (2000) Paradoxical inhibition ofT-cell function in response to CTLA-4 blockade; heterogeneity within thehuman T-cell population. Nat. Med. 6, 211–214.

33. Kuhns, M. S., Epshteyn, V., Sobel, R. A., Allison, J. P. (2000) CytotoxicT lymphocyte antigen-4 (CTLA-4) regulates the size, reactivity, and func-tion of a primed pool of CD4� T cells. Proc. Natl. Acad. Sci. USA 97,12711–12716.

34. Madrenas, J., Chau, L. A., Teft, W. A., Wu, P. W., Jussif, J., Kasaian, M.,Carreno, B. M., Ling, V. (2004) Conversion of CTLA4 from inhibitor toactivator of T cells with a bispecific tandem single chain Fv ligand.J. Immunol. 172, 5948–5956.

35. Ahmann, G. B., Sachs, D. H., Hodes, R. J. (1978) Requirement for anIa-bearing accessory cell in Con A-induced T cell proliferation. J. Immu-nol. 121, 1981–1989.

36. Quintans, J., Yokoyama, A., Evavold, B., Hirsch, R., Mayforth, R. D.(1989) Direct activation of murine resting T cells by con A or anti-CD3 Ig.J. Mol. Cell. Immunol. 4, 225–235.

37. Perrin, P. J., Davis, T. A., Smoot, D. S., Abe, R., June, C. H., Lee, K. P.(1997) Mitogenic stimulation of T cells reveals differing contributions forB7–1 and B7–2 costimulation. Immunology 90, 534–542.

38. Nagase, F., Abo, T., Hiramatsu, K., Suzuki, S., Du, J., Nakashima, I.(1998) Induction of apoptosis and tyrosine phosphorylation of cellularproteins in T cells and non-T cells by stimulation with concanavalin A.Microbiol. Immunol. 42, 567–574.

39. Pani, G., Colavitti, R., Borrello, S., Galeotti, T. (2000) Endogenous oxygenradicals modulate protein tyrosine phosphorylation and JNK-1 activationin lectin-stimulated thymocytes. Biochem. J. 347, 173–181.

40. Pongracz, J., Parnell, S., Anderson, G., Jaffrezou, J-P., Jenkinson, E.(2003) Con A activates an Akt/PKB dependent survival mechanism to

Mukherjee et al. Roles of CTLA4 and B7 during Con A-mediated T cell activation 13

modulate TCR induced cell death in double positive thymocytes. Mol.Immunol. 39, 1013–1023.

41. Lane, P., Gerhard, W., Hubele, S., Lanzavecchia, A., McConnell, F.(1993) Expression and functional properties of mouse B7/BB1 using afusion protein between mouse CTLA4 and human 1. Immunology 80,56–61.

42. Hildeman, D. A., Mitchell, T., Kappler, J., Marrack, P. (2003) T cellapoptosis and reactive oxygen species. J. Clin. Invest. 111, 575–581.

43. Jaattela, M., Tschopp, J. (2003) Caspase-independent cell death in Tlymphocytes. Nat. Immunol. 4, 416–423.

44. Boise, L. H., Minn, A. J., Noel, P. J., June, C. H., Accavitti, M. A.,Lindsten, T., Thompson, C. B. (1995) CD28 costimulation can promote Tcell survival by enhancing the expression of BclXL. Immunity 3, 87–98.

45. Demetriou, M., Binkert, C., Sukhu, B., Tenenbaum, H. C., Dennis, J. W.(1996) Fetuin/2-HS glycoprotein is a transforming growth factor-� type IIreceptor mimic and cytokine antagonist. J. Biol. Chem. 271, 12755–12761.

46. Yu, X-Z., Martin, P. J., Anasetti, C. (2003) CD28 signal enhances apo-ptosis of CD8 T cells after strong TCR ligation. J. Immunol. 170,3002–3006.

47. Razi-Wolf, Z., Hollander, G. A., Reiser, H. (1996) Activation of CD4� Tlymphocytes from interleukin 2-deficient mice by costimulatory B7 mol-ecules. Proc. Natl. Acad. Sci. USA 93, 2903–2908.

48. Appleman, L. J., Berezovskaya, A., Grass, I., Boussiotis, V. A. (2000)CD28 costimulation mediates T cell expansion via IL-2-independent and

IL-2-dependent regulation of cell cycle progression. J. Immunol. 164,144–151.

49. Tivol, E. A., Borriello, F., Schweitzer, A. N., Lynch, W. P., Bluestone,J. A., Sharpe, A. H. (1995) Loss of CTLA-4 leads to massive lymphopro-liferation and fatal multiorgan tissue destruction, revealing a criticalnegative regulatory role of CTLA-4. Immunity 3, 541–547.

50. Jovasevic, V. M., Gorelik, L., Bluestone, J. A. Mokyr, M. B. (2004)Importance of IL-10 for CTLA4-mediated inhibition of tumor-eradicatingimmunity. J. Immunol. 172, 1449–1454.

51. Ise, W., Totsuka, M., Sogawa, Y., Ametani, A., Hachimura, S., Sato, T.,Kumagai, Y., Habu, S., Kaminogawa, S. (2002) Naıve CD4� T cellsexhibit distinct expression patterns of cytokines and cell surface mole-cules on their primary responses to varying doses of antigen. J. Immunol.168, 3242–3250.

52. Egen, J. G., Allison, J. P. (2002) Cytotoxic T lymphocyte antigen-4accumulation in the immunological synapse is regulated by TCR signalstrength. Immunity 16, 23–35.

53. Watanabe, Y., Morita, M., Akaike, T. (1996) Concanavalin A inducesperforin-mediated but not Fas-mediated hepatic injury. Hepatology 24,702–710.

54. Brogdon, J. L., Leitenberg, D., Bottomly, K. (2002) The potency of TCRsignaling differentially regulates NFATc/p activity and early IL-4 tran-scription in naive CD4� T cells. J. Immunol. 168, 3825–3832.

14 Journal of Leukocyte Biology Volume 78, July 2005 http://www.jleukbio.org