Cytotoxic T lymphocyte-associated antigen-4 inhibits integrin-mediated stimulation

Abstract

The negative role exerted by cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) in the regulation of T-cell activity, as induced by T-cell receptor (TCR)/CD3 and CD28 costimulation, has been widely described. In the present work we investigated the role of CTLA-4 in the control of cell activation, as induced by costimulation of the adhesion molecule lymphocyte function-associated antigen-1 (LFA-1) in murine CD4⁺ T cells. Results show that CTLA-4 engagement inhibits interleukin-2 (IL-2) production, not only when induced by CD3/CD28 costimulation, but also when CD4⁺ T cells are costimulated by anti-CD3 and anti-LFA-1 monoclonal antibodies (mAbs). LFA-1 has been described to induce Ca²⁺ mobilization also in the absence of TCR engagement. Moreover, we found that CTLA-4 engagement negatively affects Ca²⁺ mobilization and NF-AT activation, as induced by LFA-1 engagement alone. PLCγ1 phosphorylation was also dampened within minutes after CTLA-4 engagement. Altogether these data indicate that through the control of signals induced by different receptors, CTLA-4 could be a global attenuator of T-cell activation.

Introduction

T-cell receptor (TCR) engagement by the antigen–major histocompatibility complex II (MHC II) is the primary event in CD4⁺ T-cell activation, but TCR signals must be amplified through engagement of costimulatory receptors for successful T-cell functions. CD28, the main costimulatory receptor, plays a relevant role in T-cell activation by up-regulating cytokine production and CD25 expression, leading to T-cell protection from anergy and apoptosis. Integrins, formerly defined as adhesion molecules involved only in cell–cell and cell–matrix interactions, have also been more recently found to act as costimulatory receptors delivering functional signals.^1–4 The β₂ integrin, lymphocyte function-associated antigen-1 (LFA-1), is alone unable to induce cell activation but does induce Ca²⁺ flux, as recently described.⁵ LFA-1 has also been described to synergize with TCR signals sustaining inositol-1,4,5-triphosphate (IP3) generation.^3,6 Ca²⁺ plays a key role in T-cell function as it cross-talks with signalling pathways regulated by mitogen-activated protein kinase (MAPK) and inositol-1,4,5-triphosphate kinase (IP3K), and controls cytokine production and cell proliferation.^7–9 The increase in Ca²⁺ concentration is induced by IP3 which, as generated by phospholipase C gamma 1 (PLCγ1)-catalysed phosphatidylinositol-4,5-biphosphate (PIP₂) hydrolysis, opens the Ca²⁺ channels. The resulting rise in Ca²⁺ concentration activates the Ser/Thr phosphatase, calcineurin, to dephosphorylate NF-AT, which then enters the nucleus and activates gene transcription.¹⁰

CD28 and LFA-1 promote T-cell activation, whereas cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) exerts a negative role.¹¹ CTLA-4-deficient mice show an elevated frequency of T cells expressing activation markers, autoimmune diseases and massive lymphoproliferative disorders leading to death at 3–4 weeks of age.^12,13 CTLA-4 engagement by its ligands CD80 and CD86 inhibits CD3/CD28 costimulation-induced cytokine production, CD25 expression and cell proliferation.¹¹ In cells stimulated through TCR and CD28 engagement, CTLA-4 exerts its function by inhibiting the activation of nuclear factors such as NF-κB,¹⁴ NF-AT¹⁵ and AP1,¹⁶ and by competing with CD28 with higher affinity for CD80 and CD86 ligands.¹¹

There is no evidence, however, as to whether CTLA-4 inhibits other signals delivered by integrin molecules such as LFA-1. In the present work we investigated the effect of CTLA-4 engagement on PLCγ1 and calcium pathway activation, as well as on NF-AT activation, in purified mouse CD4⁺ T cells following LFA-1 stimulation.

Materials and methods

Cell purification and culture

Splenic CD4⁺ T cells from C57Bl/6 mice were purified by immunomagnetic cell sorting, as previously described.¹⁷ Collected cells were found to be almost exclusively (> 98%) TCR-αβ⁺ CD4⁺, by flow cytometry analysis. To induce cytokine production, CD4⁺ T cells were cultured in plates precoated with anti-CD3ε monoclonal antibody (mAb) (clone 145-2C11; 10 µg/ml) or anti-LFA-1 mAb (clone M17/4; 10 µg/ml). Anti-CD28 mAb (clone 37·51) was used in soluble form (1 µg/ml). Anti-CTLA-4 mAb (clone UC10-4F10-11) was bound to plates at a concentration of 1 or 10 µg/ml. For transcription factor and calcium flux analysis, CD4⁺ cells were cultured with anti-CD3ε and anti-CD28 mAbs for 40 hr to induce optimal CTLA-4 expression on the cell membrane. After 4 hr of starving, cells were restimulated with different antibody combinations and/or ionomycin, as described below. For Western blotting, cells were collected after 1, 3 and 10 min of stimulation, while for the electrophoretic mobility shift assay (EMSA) cells were collected after 60 min. When required, cyclosporin A (CsA) was added to cell culture 90 min prior to stimulation. All the antibodies employed in culture were sodium azide- and endotoxin-free, as certified by the producer (PharMingen, San Diego, CA).

Cytokine titration

Interleukin (IL)-2 was titrated in culture supernatants by sandwich enzyme-linked immunosorbent assay (ELISA), as previously described.¹⁷ JES6-1A12 (1 µg/ml) purified mAb and JES6-5H4 (1 µg/ml) biotin-conjugated mAb were used to capture and detect IL-2, respectively. Serially diluted cytokine standard [recombinant murine (rm)IL-2, 1·6 × 10⁶ U/mg; PharMingen] or culture supernatants were added to the wells. The reference straight line obtained by plotting the absorbance versus the standard cytokine concentrations was used to calculate cytokine concentrations in the supernatants.

Flow cytometry analysis

Cells (5 × 10⁵) were preincubated with Fc Block [rat immunoglobulin G (IgG)2b anti-CD16/32, clone 2·4G2, PharMingen] to prevent cytophilic binding of labelled Abs and then stained with phycoerythrin (PE)-conjugated anti-CTLA-4 (clone UC10-4F10-11). PE-conjugated isotype-matched Ab was used as a control (clone G235-2356). The optimal concentrations of the Abs were assessed in preliminary experiments. Samples of 20 × 10⁴ cells were analysed, and fluorescence signals were collected in log mode using a fluorescence-activated cell sorter (FACScan; Becton-Dickinson, Mountain View, CA).

Preparation of cell extracts

Cells (5 × 10⁶) were collected, washed twice and pelleted by centrifugation at 200 g for 10 min. The pellet was resuspended in 75 µl of cold lysis buffer [MgCl₂ 1 mm, NaCl 350 mm, HEPES 20 mm, EDTA 0·5 mm, EGTA 0·1 mm, dithiothreitol (DTT) 1 mm, Na₄P₂O₇ 1 mm, phenylmethylsulphonyl fluoride (PMSF) 1 mm, aprotinin 1·5 mm, leupeptin 1·5 mm, 1% phosphatase inhibitor cocktail II (P5726; Sigma, St Louis, MO), glycerol 20%, Nonidet P-40 (NP-40) 1%], vigorously vortexed for 10 seconds and, after incubation for 10 min on ice, centrifuged at 20 000 g for 15 min at 4°. The supernatant was used as cell extract.

Western blot analysis

Cell extracts at equal protein concentration were immunoprecipitated from cleared extract using anti-PLCγ1 polyclonal antibody and subjected to sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) using a 7·5% polyacrylamide gel. Proteins were then electrotransferred onto polyvinylidene difluoride (PVDF) membranes (Amersham Biosciences, Little Chalfont, UK), as previously described.¹⁴

Membranes were incubated with anti-phospho PLCγ1 (2821; Cell Signaling Tech., Beverly, MA) and, after stripping, probed with anti-PLCγ1 (sc-426; Santa Cruz Biotech, Inc., Santa Cruz, CA). Immunoblots were developed by chemifluorescence and acquired by the phosphor/fluorescence imager STORM 840 (Molecular Dynamics, Sunnyvale, CA). The intensity of the bands was directly quantified by Image QuaNT software (Molecular Dynamics), which gives rise to a volume report by integrating the area of the band and its density.

EMSA

The binding reaction [containing 12 µg of total protein extract, 10 mm Tris–HCl (pH 7·6), 50 mm NaCl, 1 mm DTT, 1 mm EDTA, 5% glycerol and 200 µg/ml of poly(dI-dC) as non-specific competitor] was incubated for 30 min with 5000 counts per minute (c.p.m.) of ³²P end-labelled double-stranded oligonucleotide (corresponding to the NF-ATp-binding site) in a total volume of 20 µl. The resulting DNA–protein complexes were resolved by electrophoresis on a 6% non-denaturating polyacrylamide gel.¹⁸ After drying, gels were exposed on a phosphor screen and subsequently analysed by phosphor/fluorescence imager STORM 840 and quantified as described above.

Intracellular free-calcium measurements

Intracellular calcium concentration was evaluated by measuring the fluorescence of the calcium-indicating dye, Fura-2-AM.¹⁹ Cells (2·5 × 10⁶/ml) were washed twice, suspended in 50 mm Ringer-HEPES, pH 7·4, and loaded with 3 µm of Fura-2-AM, as previously described.²⁰ After incubation at 4° for 30 min with different combinations of anti-CD3ε, anti-CTLA-4 and anti-LFA-1 mAbs, cells were washed to remove excess antibody. Stimulation started with the addition of appropriate cross-linking antibody and soluble CD28, as indicated. Changes in intracellular Ca²⁺ concentration were determined by monitoring the fluorescence emission of the Ca²⁺-bound Fura-2-AM in a Perkin-Elmer (Norwalk, CT) LS3B fluorometer (excitation 340 nm, emission 510 nm), to minimize dye leakage from the cells. Calibration was performed according to the protocol previously described.²¹ A value of 184 nm was used as the K_d for the calcium–Fura-2-AM complex at 30°. The recorded changes in fluorescence are directly proportional to stimulus-dependent alterations in the intracellular free Ca²⁺ concentration.

Statistical analysis

Each culture was set up in triplicate for the evaluation of cytokine production. The two-tailed Student's t-test was performed on the arithmetic mean of each experimental point. For EMSA and Western blot, arithmetic means were calculated on volume reports (as generated by Image QuaNT software) from three independent experiments, and statistical evaluation was performed using the two-tailed Student's t-test.

Results

CTLA-4 expression

Several signals contributing to T-cell activation also induce later events leading to its inhibition. CTLA-4 expression, indeed, can be induced by several activating stimuli such as TCR/CD3 engagement, CD28 costimulation, LFA-1 triggering and IL-2.^22,23 Purified CD4⁺ cells were stimulated with different combinations of anti-CD3, anti-CD28 and anti-LFA-1 mAbs. At the end of the stimulation period, cells were permeabilized and stained with PE-conjugated anti-CTLA-4 mAb or PE-conjugated anti-trinitrophenyl (TNP) (isotype control) mAb. Figure 1 shows that CTLA-4 expression is barely detectable in unstimulated CD4⁺ cells (Fig. 1a) but that it increases after 48 hr of stimulation. CD3 engagement alone induces low CTLA-4 expression, the percentage of positive cells being 13% (Fig. 1b). At variance, CD3/CD28 costimulation increased the percentage of positive cells to 44% (Fig. 1c). LFA-1 alone did not induce CTLA-4 expression (data not shown). Interestingly, when both anti-CD3 and anti-LFA-1 mAbs were present, the percentage of CD4⁺ cells expressing CTLA-4 was 32% (Fig. 1d). The columns in (Fig. 1e) show mean fluorescence intensities.

Figure 1.

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) expression was evaluated in the intracellular compartment of CD4⁺ cells unstimulated (a) or stimulated for 48 hr with anti-CD3 alone (b), anti-CD3 and anti-CD28 (c), or anti-CD3 and anti-lymphocyte function-associated antigen 1 (LFA-1) (d) monoclonal antibodies (mAbs). Phycoerythrin (PE)-conjugated anti-trinitrophenyl (TNP) A19-3 mAb was used as isotype control for the staining (filled curve). Numbers indicate the percentage of positive cells. Panel (e) shows the CTLA-4 mean fluorescence intensity (MFI). Each column represents the difference between the MFI of the sample stained with the PE-conjugated anti-CTLA-4 mAb and the corresponding sample stained with PE-conjugated anti-TNP A19-3 mAb. The results are representative of four independent experiments.

CD28 or LFA-1 costimulation-induced IL-2 production is inhibited by CTLA-4 engagement

CD4⁺ T cells were stimulated with different combinations of anti-CD3 and anti-CD28 or anti-LFA-1 mAbs in the presence of either anti-CTLA-4 or anti-TNP (isotype control) immobilized mAb. Optimal concentrations of the Abs were assessed in preliminary experiments. Figure 2 shows the kinetics and dose–response of IL-2 production. As expected, CD3 engagement alone induced a negligible IL-2 production (Fig. 2a). Conversely, upon CD3/LFA-1 costimulation, CD4⁺ cells produced a large amount of IL-2, reaching 17–18 ng/ml after 48–72 hr of stimulation. In the presence of the anti-CTLA-4 mAb, IL-2 production induced by CD3/LFA-1 costimulation was 32–34% lower after 48–72 hr of stimulation than in the absence of anti-CTLA-4 mAb (Fig. 2a). A 10-fold lower concentration of the anti-CTLA-4 mAb was weakly or not effective on IL-2 production induced by anti-CD3 and anti-CD28 mAbs (Fig. 2b). Similar results were obtained when anti-CD3 and anti-LFA-1 mAbs were used (Fig. 2c). Thus, CTLA-4 engagement inhibits both the CD3/CD28- and CD3/LFA-1-induced IL-2 production (Fig. 2b, 2c).

Figure 2.

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) engagement inhibits interleukin-2 (IL-2) production, as induced by CD3–CD28 or CD3–anti-lymphocyte function-associated antigen 1 (LFA-1) costimulation. (a) Culture supernatants from CD4⁺ T cells stimulated for 24–72 hr with anti-CD3, anti-CD3 and anti-CD28 monoclonal antibodies (mAbs) or anti-CD3 and anti-LFA-1 mAbs, in the presence of either isotype-control hamster mAb at 10 µg/ml (Iso-10) or anti-CTLA-4 mAb at 10 µg/ml (A4-10), were analysed using enzyme-linked immunosorbent assay (ELISA) (anti-CD3/CD28/CTLA4 mAbs versus anti-CD3/CD28/Iso mAbs, P < 0·01 at 48 and 72 hr; anti-CD3/LFA1/CTLA4 mAbs versus anti-CD3/LFA1/Iso mAbs, P < 0·05 at 48 and 72 hr). (b) CD4⁺ T cells were stimulated for 48 hr with anti-CD3 mAb and different concentrations of anti-CD28 mAb in the presence of either isotype-control mAb at 10 µg/ml (Iso-10) or anti-CTLA-4 mAb at 1 µg/ml (A4-1) and 10 µg/ml (A4-10) (anti-CD3/A4-1 mAbs versus anti-CD3/Iso-10 mAbs, P < 0·05 at 0·5 µg of anti-CD28 mAb; anti-CD3/A4-10 mAbs versus anti-CD3/Iso-10 mAbs, P < 0·05 at 0·5–2 µg of anti-CD28 mAb). (c) CD4⁺ T cells were stimulated for 48 hr with anti-CD3 mAb and different concentrations of anti-LFA-1 mAb in the presence of either isotype-control mAb at 10 µg/ml (Iso-10) or anti-CTLA-4 mAb at 1 µg/ml (A4-1) and 10 µg/ml (A4-10) (anti-CD3/A4-10 mAbs versus anti-CD3/Iso-10 mAbs, P < 0·05 at 3·5–14 µg of anti-LFA1 mAb). Data shown are from one representative out of three independent experiments. Values in each panel represent mean ± standard error.

CTLA-4 inhibits NF-AT activation induced by CD3 and/or LFA-1 engagement

In primary cultures (such as the one used to study cytokine production), as soon as cell activation proceeds, CD4⁺ cells begin to express CTLA-4 which, upon engagement, can exert its negative effects. In this type of culture CTLA-4 expression reaches a maximum after 24–48 hr, even if negative effects can be observed earlier. To investigate the effects of CTLA-4 engagement on signal events required for NF-AT activation, relevant for IL-2 production,¹⁰ CD4⁺ cells were prestimulated for 40 hr (see the Materials and methods) to induce optimal CTLA-4 expression, as previously described.^24,25 Then, cells were washed and suspended in medium alone for 4 hr to reduce the activation of transcription factors and signalling cascades. These cells, which still express CTLA-4 (as verified by flow cytometry), were washed and used to study the effects of CTLA-4 engagement on signalling events.

Prestimulated cells were restimulated for 30 min with different combinations of antibodies, as shown in Fig. 3. Cell extracts were analysed by EMSA using a radiolabelled consensus-binding sequence for NF-AT (Fig. 3a). Densitometric analyses from three independent experiments are shown as mean ± standard error in Fig. 3b. Results show that even if CD3 stimulation alone does not induce IL-2 production, its engagement induces NF-AT activation (Fig. 3a, lane 2 versus lane 1). CTLA-4 engagement reduces (Fig. 3a, lane 3 versus lane 2, P < 0·05) this activation to the level of the unstimulated group (Fig. 3a, lane 1), as previously described.¹⁵ For comparison, an immunosuppressive concentration of the calcineurin inhibitor, CsA, reduced CD3-induced NF-AT-binding by 75% (Fig. 3a, lane 4 versus lane 2, P < 0·05). Also, LFA-1 engagement, which per se does not induce IL-2 production, activated NF-AT (Fig. 3a, lane 5). Interestingly, CTLA-4 engagement inhibited NF-AT activation when cells were stimulated with anti-LFA1 mAb (Fig. 3a, lane 6 versus lane 5, P < 0·05), unmasking a negative effect of CTLA-4 on signals derived from an integrin molecule. CTLA-4 inhibited NF-AT activation also when CD4⁺ cells were stimulated with both anti-CD3 and anti-LFA-1 mAb (Fig. 3a, lane 10 versus lane 9, P < 0·05). As previously described, CTLA-4 engagement also inhibited CD3/CD28 costimulation-induced NF-AT activation.

Figure 3.

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) engagement inhibits NF-AT activation. (a) Preactivated CD4⁺ cells (see the Materials and methods), after starving, were restimulated with different combinations of monoclonal antibodies (mAb) for 30 min. (Iso ctrl, isotype control for the α-CTLA-4 mAb; CsA, cyclosporin A; Iono, ionomycin.) Cellular extracts were incubated with NF-AT-radiolabelled consensus sequence and resolved by electrophoresis. (b) Supershift analysis. Before analysis using the electrophoretic mobility shift assay (EMSA), extracts were incubated with α-NF-AT or the isotype control (ctrl) mAb to assess the specificity of the band. (c) Densitometric analysis was performed using the Image QuaNT software. Values, expressed in arbitrary units, represent the means of three independent experiments ± standard error. (d) Cellular extracts from murine EL4 thymoma cells, stimulated with α-CD3 and α-CD28 mAb in the presence of either isotype-control mAb (Iso ctrl) or α-CTLA-4 mAb, were analysed using EMSA for NF-AT activation. The results are representative of three independent experiments.

In cells stimulated with anti-CD3 mAb and ionomycin, CTLA-4 engagement did not inhibit NF-AT binding (Fig. 3a, lane 12 versus lane 11), suggesting that CTLA-4 affects signal events upstream of Ca²⁺ mobilization. This hypothesis is further sustained by the finding that CTLA-4 engagement did not inhibit NF-AT activation (Fig. 3c, 3d) and IL-2 production (data not shown) in EL4 cells which (express CTLA-4 and) carry a gene mutated in the sequence coding for the autoregulatory site of the calcineurin phosphatase.²⁶

CTLA-4 inhibits Ca²⁺ flux

CD4⁺ T cells were analysed for Ca²⁺ mobilization using the fluorescence indicator, Fura-2-AM, in a fluorometer. CD4⁺ cells prestimulated as described in Fig. 3 were stimulated with different combinations of antibodies in the presence of either anti-CTLA-4 mAb or the isotype control. In preliminary experiments we observed that anti-CTLA-4 mAb does not induce Ca²⁺ mobilization. Stimulation of CD4⁺ cells started when cross-linking mAb was added (indicated by an arrow in the figure) and resulted in an increase of [Ca²⁺]_i, followed by a sustained plateau (Fig. 4). [Ca²⁺]_i increased upon CD3 engagement but was reduced by CTLA-4 engagement (Fig. 4a). CD28 showed a drastic synergic effect on CD3-induced [Ca²⁺]_i increase, which was reduced by CTLA-4 engagement (Fig. 4b). Interestingly, Ca²⁺ flux induced by LFA-1 stimulation alone (Fig. 4c) was completely abolished by CTLA-4 engagement. Stimulation of both CD3 and LFA-1 induced a lower [Ca²⁺]_i increase as compared to CD3/CD28 costimulation (Fig. 4d versus 4b). Yet, CTLA-4 engagement inhibited CD3/LFA-1-induced [Ca²⁺]_i increase less effectively than upon CD3/CD28 costimulation (Fig. 4d versus 4b). These data reveal that CTLA-4 can interfere not only with signals generated by CD3 but also with events induced by LFA-1 engagement and leading to Ca²⁺ flux (independently from TCR engagement).

Figure 4.

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) inhibits Ca²⁺ flux. Fura-2-AM-loaded CD4⁺ T cells were stimulated with different combinations of monoclonal antibodies (mAb) in the presence of either α-CTLA-4 mAb (closed squares) or isotype-control mAb (open squares). The cross-linker mAb was added after 1 min (arrow). In (a) the horizontal line represents basal [Ca²⁺] (88 nm). The results shown are representative of three independent experiments.

CTLA-4 engagement blocks LFA-1-induced PLCγ1 phosphorylation

PLCγ1 activation, which requires tyrosine phosphorylation, induces Ca²⁺ flux in T cells.²⁷ Prestimulated CD4⁺ cells were used to study the effects of CTLA-4 engagement on PLCγ1 phosphorylation. Results (Fig. 5) show that anti-CD3 mAb-induced PLCγ1 phosphorylation is reduced and delayed by anti-CTLA-4 mAb. CD3/CD28 costimulation causes higher and more sustained PLCγ1 phosphorylation. CTLA-4 engagement lowers PLCγ1 phosphorylation. LFA-1 stimulation alone induces a lower PLCγ1 phosphorylation as compared to the other stimuli and CTLA-4 engagement reduces PLCγ1 phosphorylation to the level of the unstimulated group. CD3/LFA-1 costimulation induces a less sustained PLCγ1 phosphorylation as compared to CD3/CD28 costimulation. The inhibitory effect of CTLA-4 engagement on CD3/LFA-1-induced PLCγ1 phosphorylation is faster and higher as compared to its effect upon CD3/CD28 costimulation.

Figure 5.

Cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) inhibits phospholipase C gamma 1 (PLCγ1) phosphorylation. CD4⁺ T cells were stimulated for 1, 3 or 10 min with α-CD3 monoclonal antibody (mAb) and α-CD3 and α-CD28 mAb (a), or α-lymphocyte function-associated antigen 1 (LFA-1) mAb and α-CD3 and α-anti-LFA-1 mAb (b) in the presence of either α-CTLA-4 mAb (+) or isotype control mAb (−). Cell extracts were immunoprecipitated with α-PLCγ1 mAb and resolved by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) (7·5% gel). After electrotransfer, filters were immunostained with anti-phosphoPLCγ1, stripped and reprobed with α-PLCγ1. Densitometric analysis was performed using Image QuaNT software. Values, expressed in arbitrary units, are means of three independent experiments ± standard error. White column, unstimulated control; black columns, stimulated groups + isotype control; grey columns, stimulated groups +α-CTLA-4. Similar results were obtained in two other independent experiments.

Discussion

The outcome of T-cell activation, namely induction of cytokine production or stimulation of effector functions, is dependent on the integration of several signals delivered by membrane receptors. TCR stimulation, indeed, which ensures antigen specificity of the response, in the absence of other signals can lead T cells to a state of anergy. At variance, costimulation of additional receptors, such as CD28, induces T-cell activation. CTLA-4 has been described to control T-cell activity directly by interfering with CD3/CD28-induced signals and indirectly by competing for B7 ligands with higher affinity, as compared to CD28.²⁸ The majority of resting T cells express CD28 on the cell surface, whereas CTLA-4 is up-regulated after T-cell activation, with peak expression at 24–48 hr in both CD4⁺ and CD8⁺ cells. Recent studies show that signals from TCR and CD28 act synergistically to induce the expression of CTLA-4 by enhancing the rate of transcription and increasing mRNA stability.²⁹ However, even at maximal expression levels, remarkably little CTLA-4 is found on the cell surface. This low expression results from the interaction of the cytoplasmic domain of CTLA-4 with the clathrin pit adaptor complex (AP-50), which causes its rapid internalization. Upon activation, the CTLA-4 cytoplasmic domain is Tyr-phosphorylated resulting in disengagement from the AP-50 internalization system and therefore stabilization on the cell surface.³⁰ Noteworthy, other signals contributing to T-cell activation up-regulate CTLA-4 expression. As shown in the present report and as previously reported,²² LFA-1 together with CD3 engagement leads to CTLA-4 expression. Soluble mediators, such as IL-2, can also increase CTLA-4 expression²³ and thus contribute to the preparation of the signalling machine for shutting down T-cell activation, after the effector function has been carried out.

It has been demonstrated that CTLA-4 inhibits the activation of the linker for activation of T cell (LAT)²⁵ which leads to recruitment of PLCγ1 to the plasma membrane.³¹ Upon phosphorylation, PLCγ1 generates IP3 which, in turn, induces Ca²⁺ flux.³² In our work we found, indeed, that PLCγ1 phosphorylation and Ca²⁺ flux induction, as well as NF-AT activation induced by CD3 engagement, are inhibited by CTLA-4 engagement. Interestingly, we also found that CD28 costimulation induces a drastic increase of CD3-induced PLCγ1 phosphorylation and Ca²⁺ flux. The CD28 CsA-sensitive pathway is mainly limited to T cells of the CD4⁺ subset³³ and dominant in activated T cells^34,35 in which the [Ca²⁺]_i increase is preferentially induced by CD80-mediated stimulation of CD28. Even if molecular mechanisms underlying costimulation are largely unknown, new findings show that CD28 does not lead to increased Zap-70/LAT phosphorylation but amplifies PLCγ1 phosphorylation (and thus Ca²⁺ flux) through Itk activation.³⁶ This mechanism might account for the drastic increase in Ca²⁺ flux induced by CD28, as observed in our experiments. Our data also show that CD3/CD28 costimulation-induced Ca²⁺ flux is inhibited by CTLA-4 engagement, suggesting that other mechanisms could be involved in the control exerted by CTLA-4.

LFA-1 has been described to contribute to activating signals for IL-2 production in TCR-stimulated T cells.³⁷ LFA-1 per se is, indeed, unable to induce cell activation but it contributes to TCR-induced signalling by enhancing free intracellular Ca²⁺ concentration.^3,6 An efficient sensor for [Ca²⁺]_i is calmodulin which, upon Ca²⁺ binding, activates the protein phosphatase calcineurin to dephosphorylate NF-AT, which then enters the nucleus. As nuclear kinases can rapidly phosphorylate NF-AT, sustained Ca²⁺ signalling is required to maintain NF-AT in its active form, able to induce cytokine production and cell proliferation. If the inhibitory effect of CTLA-4 on Ca²⁺ flux induced by CD3 can be explained by the effects of CTLA-4 on LAT activation, it remains unclear how CTLA-4 can affect [Ca²⁺]_i induced by LFA-1. The mechanisms whereby LFA-1 contributes to Ca²⁺ flux are also not completely understood. It is difficult, indeed, to dissect out the contribution of LFA-1 to the TCR signalling induced by favouring T cell–antigen-presenting cell (APC) interactions from its own signals. Under our experimental conditions we observed that, in preactivated CD4⁺ cells, LFA-1 engagement alone induces [Ca²⁺]_i increase, PLCγ1 phosphorylation and NF-AT activation. We provide the first evidence that CTLA-4 engagement inhibits these LFA-1-induced effects. Considering the ability of Ca²⁺ to recruit control elements of other signalling pathways, such as the MAP kinase cascade, these findings show a critical role of CTLA-4 in the control of lymphocyte activity. This notion is further sustained by the finding that CTLA-4 engagement inhibits NF-κB activation,¹⁴ which can also be affected by changes of [Ca²⁺]_i.⁷

Thus, CTLA-4 can control signals generated by CD3, LFA-1 and the costimulatory effects of CD28 and, as shown in a recent publication,³⁸ of inducible costimulatory molecule (ICOS), indicating that it could play a general role of gatekeeper to T-cell activation. Although CTLA-4 engagement inhibits PLCγ1 phosphorylation very early and efficiently in all groups, it only partially inhibits the [Ca²⁺]_i increase and cytokine production, as induced by CD3 and LFA-1 costimulation. These findings suggest that LFA-1 and CD28 might act differently in sustaining calcium signals generated from TCR engagement. How CTLA-4 interferes with signals generated by different receptors and possibly with their integration, as can occur in the supramolecular activation cluster (SMAC),³⁹ requires further investigation. A recent publication shows that CTLA-4 can affect the release of rafts to the surface of T cells,⁴⁰ providing a mechanism through which CTLA-4 could interfere with the signalling machine of different receptors.

Abbreviations

CTLA-4

cytotoxic T lymphocyte-associated antigen-4

IL

interleukin

LFA-1

lymphocyte function-associated antigen 1

mAb

monoclonal antibody

PLCγ1

phospholipase C gamma 1

TCR

T-cell receptor