Published online before print July 1, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Immunology Outpatient Clinic, Vienna, Austria
1Correspondence: Immunology Outpatient Clinic, Schwarzspanierstraße 15/1, A-1090 Vienna, Austria. E-mail: hermann.wolf{at}itk.at
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
with TNF-RII but not TNF-RI is directly costimulatory to TCR-mediated T cell activation, thereby augmenting T cell proliferation, expression of T cell activation markers (CD25, human leukocyte antigen-DR, TNF-RII), and secretion of cytokines such as interferon-
and TNF-. In contrast to the well-defined costimulatory molecule CD28, costimulation via TNF-RII showed significant differences in kinetics, requirement for cross-linking, redundancy of intracellular signaling pathways involved, and the capacity to induce interleukin (IL)-2, IL-10, and IL-13 secretion. In addition, cross-linking TNF-RII had the capacity to down-regulate TCR/CD28-induced Ca++ mobilization, IL-2 mRNA expression, and IL-2 and IL-10 secretion. Taken together, our findings demonstrate that TNF-RII plays a unique role among the T cell costimulatory molecules, as TNF-RII ligation can have positive and negative effects on TCR-dependent signaling. TNF-RII cross-linking has an inhibitory effect on early TCR signaling events proximal to induction of Ca++ flux, which ultimately leads to modulation of the T cell cytokine pattern expressed.
Key Words: T cell receptor • TNF- • IL-2 • Ca++ flux
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) is a protein of 17 kDa and acts as a homotrimer. It is mainly produced by monocytes and macrophages, but many other cells including lymphocytes have been found to release TNF- [1
]. It was originally described as a cell-derived factor causing hemorrhagic necrosis of tumors [2
], and its cytotoxic effects on transformed or normal cells have been investigated widely, aiming its use in anticancer treatment [3
]. In addition TNF- is a proinflammatory cytokine, as it activates monocytes and macrophages as well as granulocytes and endothelial cells, induces the secretion of various cytokines [interleukin (IL)-1, IL-6, IL-8], and thus, enhances and propagates local and systemic inflammatory responses [1
]. These proinflammatory properties are paralleled by antiapoptotic signals and are mediated by nuclear factor-
B (NF-B)-dependent pathways [4
], and cytotoxic effects of TNF- are not [5
].
Much less, however, is known about the effects of TNF- on B and T cells. It has been shown to enhance cytotoxicity of CD8+ T cell subsets and natural killer cells [6
, 7
], to stimulate proliferation of thymocytes and mature T cells in the presence of suboptimal doses of mitogenic lectins, e.g., phytohemagglutinin or concanavalin A (Con A) [8
9
10
], and to costimulate T cell proliferation together with other cytokines such as IL-2 [11
]. TNF- sustains proliferation in Staphylococcus aureus Cowan-activated B cells and costimulates with anti-µ the induction of B cell DNA synthesis [12
]. The membrane-bound form of TNF- expressed on activated CD4+ cells has been reported to costimulate IL-4-dependent secretion of immunoglobulin G (IgG)4 and IgE in human B cells [13
]. Moreover, TNF- seems to be essential in T lymphocyte development and differentiation within the thymus [14
], and the presence of endogenous TNF- has been described to be important for efficient T cell responses in vivo and in vitro. Neutralization of endogenous TNF- suppresses the development of contact sensitivity to the hapten trinitrophenyl in mice [15
], and TNF- knockout mice succumb to widespread dissemination of Mycobacterium tuberculosis as a result of deficiencies in granuloma formation [16
]. Neutralizing endogenous TNF- in CD3-activated T cells leads to unstable NF-B translocation to the nucleus [17
]. Chronic exposure to TNF-, conversely, seems to attenuate T cell responses in vitro and in vivo, and this has been linked to a reduction of T cell receptor (TCR)-
chain formation and down-regulation of TCR/CD3 complexes at the cell surface in mice [10
, 18
].
There are two receptors for TNF- [8
]: TNF-RI or p55, also classified as CD120a, and TNF-RII, also referred to as p75 or CD120b. Most of the TNF responses known (e.g., NF-B activation, cytotoxicity, IL-6 induction, fibroblast proliferation) occur by activation of p55, which is expressed on a wide range of cells [1
]. However, TNF effects on T cells seem to be TNF-RII-mediated, as anti-TNF-RII antisera or monoclonal antibodies (mAb) but not anti-TNF-RI antisera induce proliferation in the presence of a submitogenic dose of Con A [8
, 19
, 20
]. In the present study, we further investigated the role of TNF- and in particular, TNF-RII in T cell activation. We examined surface expression of TNFRs in resting and activated T cells and show that T cell activation leads to up-regulation of TNF-RII. Using monocyte-depleted peripheral blood mononuclear cells (PBMC) or highly purified CD4-positive T cells stimulated with anti-TCR mAb and mAb against the two different TNFRs, we confirm that TNF-RII is a costimulatory molecule to the TCR. In addition, the present study extends previous findings about the effect of TNF- on T cells by showing differences between TNF-RII costimulation and activation via the well-established costimulatory surface molecule CD28 [21
] with respect to kinetics, activation requirements, and the cytokine pattern induced. Furthermore, we show that concomitant triggering of TNF-RII not only has a costimulatory effect on TCR-induced T cell proliferation, cytokine secretion, and expression of T cell activation markers but is also able to inhibit distinct T cell functions such as IL-2 and IL-10 production. This inhibitory effect seems to be essentially different from the inhibitory effects of chronic TNF- exposure described so far [10
, 18
] or changes in receptor quality during different stages of cell activation as known for other receptors [22
], as it directly modulates TCR signaling proximal to the induction of Ca++ flux.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
CD4+ cells were isolated from nonadherent cells by positive immunoselection with the Dynal CD4-positive isolation kit® (Dynal, Oslo, Norway), according to the manufacturer’s protocol. In brief, PBMC were depleted twice from adherent cells by incubating in plastic tissue-culture plates and were incubated with anti-CD4-coated beads for 30 min on ice under gentle tilt rotation. Captured CD4+ cells were collected with a magnet (Dynal MPC®-M, Dynal) and detached from beads with DETACHaBEAD CD4/CD8® (Dynal). Purity was >99% CD4+ cells, as determined by flow cytometry.
For proliferation assays, cells were resuspended in RPMI-1640 medium supplemented with 2 mM/ml L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 10% heat-inactivated human AB serum at a concentration of 5 x 105/ml and were stimulated in triplicate in flat-bottomed, 96-well microtiter plates (200 µl/well; Sarstedt, Wiener Neudorf, Austria) for 5 days in a CO2 incubator (37°C, 5% CO2). Cells were pulsed with 3H-thymidine for the last 16 h and harvested onto microfilter plates (Wallac Oy, Turku, Finland), and 3H-thymidine incorporation was measured with a liquid scintillation counter (Wallac 1450 MicroBeta® Trilux). For cytokine measurements, cells were resuspended at a concentration of 1 x 106/ml, seeded into 24-well plates (400 µl/well; Sarstedt), and stimulated for 1, 3, or 5 days. Cytokines were measured with sandwich enzyme-linked immunosorbent assay (ELISA) kits [TNF-, interferon- (IFN-), IL-2, and IL-10 (FlexiaTM, Biosource Europe S.A., Fleurus, Belgium) and IL-13 (Pelikine CompactTM, CLB, Central Laboratory of the Netherlands Red Cross, Amsterdam)], according to the manufacturer’s instructions.
T cell stimulation
For T cell stimulation, the following reagents were used: phorbol myristate acetate (PMA; 10 ng/ml; Sigma-Aldrich Chemie, Vienna, Austria) plus ionomycin calcium (500 ng/ml; Sigma-Aldrich), anti-TCR mAb (clone BMA 031, Immunotech, Marseille, France; 750 ng/107 beads, 5x105 beads/ml), anti-CD28 mAb (clone CD28.2, Immunotech; soluble, 10 ng/ml; coated, 285 ng/107 beads, 5x105 beads/ml), recombinant human TNF- (20 ng/ml; R&D Systems, Minneapolis, MN), anti-TNF-RII mAb (clone 22221.311, R&D Systems; soluble, 1 µg/ml; coated, 7125 ng/107 beads, unless stated otherwise, 5x105 beads/ml), anti-TNF-RI mAb (1; clone 16803, R&D Systems; 1 µg/ml), anti-TNF-RI mAb (2; cat. no. AF225, R&D Systems; 1 µg/ml), IgG2a isotype-control antibody (clone 20102.1, R&D Systems; soluble, 1 µg/ml; coated, 7125 ng/107 beads, unless stated otherwise, 5x105 beads/ml), and anti-TNF- (clone 2C8, Upstate Biotechnology, Lake Placid, NY; 2 µg/ml). Anti-mouse IgG-loaded beads (Dynabeads® M-450, Dynal) were coated with specific or control mAb, according to the manufacturer’s instructions. Nonadherent monocyte-depleted PBMC or purified CD4+ cells were then stimulated for 5 days with anti-TCR mAb-coated beads in the presence or absence of anti-CD28 or TNF- as a costimulus. To show the requirement for cross-linking the costimulatory molecule, e.g., CD28, all binding sites for mouse IgG on the anti-TCR-coated beads, which were left unoccupied, were blocked by a further incubation step with mouse IgG (Sigma-Aldrich; 10 µg/ml, continuous rotation for 45 min, room temperature).
The following inhibitors of T cell signal-transduction pathways were used [inhibitory concentration (IC50), as indicated by the manufacturer]: BAY 11-7082 [inhibitor of B (IB), IC50=5–10 µM; Biomol Research Laboratories, Plymouth Meeting, PA], SB 202190 [inhibitor of stress-activated protein kinase (SAPK)/p38, IC50=280–350 nM], PD 98059 [inhibitor of mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK), IC50=1–20 µM], and LY 294002 [inhibitor of phosphatidylinositol-3 kinase (PI-3K), IC50=1.4 µM; all purchased from Upstate Biotechnology].
Flow cytometry
The following directly (fluorescein isothiocyanate, phycoerythrin, and peridinin chlorophyll protein) conjugated mAb were used in three-color immunofluorescence staining following a standard protocol: CD3, CD4, CD8, CD25, CD69, and human leukocyte antigen (HLA)-DR (all purchased from Becton Dickinson, Schwechat, Austria) and TNF-R1 and TNF-R2 (R&D Systems). Cells were analyzed with a FACScan (Becton Dickinson), and a minimum of 10,000 events within a lymphocyte light-scatter gate were aquired. Analysis was performed with the CellQuest® software (Becton Dickinson).
RNA extraction and reverse transcriptase-polymerase chain reaction (RT-PCR)
Whole cellular RNA was obtained by standard methods of homogenization with phenol-guanidin-isothiocyanate (Trizol-Reagent, Life Technologies) and phase separation with chloroform. The aqueous phase was collected, and RNA was precipitated with isopropyl alcohol. The RNA pellet was washed once with 75% ethanol, resuspended in RNase-free water, and stored at -80°C until RT-PCR was performed.
IL-2 mRNA was determined by one-step RT-PCR using the Titan One Tube RT-PCR kit (Roche-Boehringer Mannheim, Vienna, Austria). The reactions were performed with a 10-mM dNTP mix (Invitrogen, Lofer, Austria) and 5 U/µl RNAsin (Promega, Mannheim, Germany). For amplification of IL-2 mRNA, the following intron-spanning primers were used: AACCTCAACTCCTGCCACAATG (sense) and CAAGTTAGTGTTGAGATGATGC (antisense), coding for a product of 489 bp. For normalization of RNA quantity, the "housekeeping gene" glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was analyzed in parallel. The primers used were: GGTGAAGGTCGGAGTCAACGGA (sense) and GAGGGATCTCGCTCCTGGAAGA (antisense), coding for a product of 239 bp. Primers were purchased from VBC-Genomics Bioscience (Vienna, Austria). Amplification and synthesis of cDNA were performed with a Biometra Trio-Thermoblock (Biometra biomedizinische Analytik, Göttingen, Germany), programmed for 30 cycles (denaturation, 30 s at 94°C; annealing, 30 s at 60°C; elongation, 45 s at 68°C), preceded by a RT (45 min at 50°C) and denaturation step (2 min at 94°C). The RT-PCR products were then detected on ethidium bromide-stained, 1% agarose gels using a Fluor-S Multi-Imager system (Bio-Rad Laboratories, Vienna, Austria). For densitometric quantification, the electrophoretic gels were scanned and analyzed with Quantity-D One analysis software (Bio-Rad Laboratories).
Determination of intracellular-free Ca++
The increase in intracellular-free Ca++ following stimulation of T cells via the TCR was determined by flow cytometry [24
]. Nonadherent lymphocytes (2x106/ml) were loaded with Fluo-3 (Sigma-Aldrich; 1 µM) and Snarf-1 (Molecular Probes, Eugene, OR; 1 µM) in buffer A [Iscove’s modified Dulbecco’s medium (IMDM), Sigma-Aldrich] containing 10 mM Hepes (Sigma-Aldrich), pH 7.0, in a CO2 incubator at 37°C. Pluronic F-127 (Sigma-Aldrich; 0.2%) was used to increase the uptake of the two fluorochromes by the lymphocytes. After 30 min, an equal volume of buffer B [IMDM containing 10 mM Hepes and 10% fetal calf serum (FCS)] and the mAb (anti-TCR, 500 ng/ml; anti-CD28, 50 ng/ml, with or without anti-TNF-R1 or anti-TNF-R2; 5 µg/ml) were added. The cells were incubated for an additional 15 min at 37°C and thereafter, were washed three times and resuspended in buffer C (IMDM containing 10 mM Hepes, 10% FCS, and 10 µg/ml DNase I, Sigma-Aldrich). The flow cytometric analysis of intracellular Ca++ was performed using a FACScan (Becton Dickinson) by establishing a baseline calcium flux of at least 1 min in unstimulated cells. Biotinylated goat anti-mouse IgG (of the IgG class) and streptavidin (all purchased from Sigma-Aldrich; 4 µg/ml and 1 µg/ml, respectively) were then added to cross-link anti-TCR, anti-CD28, anti-TNF-RI, and anti-TNF-RII antibodies on the cell surface. Cells stimulated with ionomycin (Sigma-Aldrich; 2 µg/ml) were used as a positive control, and cells treated with the cross-linking reagents (biotinylated goat anti-mouse IgG and streptavidin) alone were used as a negative control. Analysis was performed by plotting FL-1 (Fluo-3) or FL-3 (Snarf-1) versus time and splitting the time axis in segments of 20 s, each comprising typically more than 2000 events. For each time segment, the percentage of responding cells was calculated with the cut-off between a responding and nonresponding cells set so that unstimulated cells contained between 1 and 5% responding cells.
Statistics
Results from repeated experiments with cells from different blood donors are given as mean of n ± SEM, and in n, each subject is represented only once. For calculation of statistically significant differences, the two-tailed unpaired Student’s t-test was applied. As a nonparametric test for paired samples, we used Friedman 
2 test for ranks. The statistical significance level was set at P < 0.05.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
costimulates TCR activation via TNF-RII but not TNF-RI (Fig. 1
and see Fig. 3
). The results depicted in Figure 1
show that TNF- acts as a costimulator and induces significant proliferation and secretion of IFN- if applied together with TCR stimulation (Fig. 1 a and 1b)
. Furthermore, TNF- acts as a costimulator with TCR by up-regulating TNF-RII, HLA-DR, and CD25 on the surface of CD4+ and CD8+ cells (Fig. 1 c
and 1d)
. For CD69 induction, TCR stimulation alone was sufficient in our system, as activation neither via TCR/CD28 (as has also been described earlier in a different experimental setup; see ref. [25
]) nor via TCR/TNF-RII increased CD69 expression further. The extent of TNF--mediated costimulation of T cell activation marker expression and IFN- release was comparable with that observed after CD28 costimulation (data not shown). TCR stimulation up-regulates TNF-RII expression, and newly up-regulated TNF-RII is equally effective in TCR costimulation, as TNF- increased TCR-dependent lymphoproliferative responses in cultures prestimulated via TCR for 24 h to an extent comparable with the costimulatory effect observed in freshly isolated lymphocytes (data not shown). TNF- does not in itself induce significant proliferation (Fig. 1a)
or secretion of IFN- or TNF- (data not shown), nor does it activate together with anti-CD28 mAb (Fig. 1a)
.
 View larger version (41K): [in a new window] |
Figure 1. TNF- costimulates TCR-induced but not CD28-stimulated proliferation, secretion of IFN-, and expression of T cell activation markers (TNF-RII, HLA-DR, CD25) via TNF-RII in peripheral blood lymphocytes. Nonadherent PBMC were stimulated with anti-CD28 or anti-TCR mAb coated onto beads in the presence or absence of TNF- (20 ng/ml), anti-TNF-RII mAb (1 µg/ml), two different anti-TNF-RI mAb (1 µg/ml), or isotype-matched control antibody (Isotype-Co) for 5 days (3H-thymidine uptake, IFN-, TNF-RII, HLA-DR, CD25) or 1 day (CD69). Liquid scintillation counter, ELISA, and three-color immunofluorescence flow cytometry, respectively, determined 3H-thymidine uptake (a, n=45 for anti TNF-RII), quantification of IFN- in culture supernatants (b, n=31 for anti-TNF-RII), and surface expression of T cell activation markers (c and d). Surface expression of activation markers is given as the percentage of lymphocytes positive for CD4 (c) or CD8 (d) and the respective activation marker. Bars represent the mean ± SEM. *, Statistically significant difference as compared with cells cultured in the absence of TNF- or anti-TNFR mAb (i.e., Med.-Co). n.t., Stimulation of lymphocyte-proliferative response with anti-CD28 plus anti-TNF-R1 has not been tested.
|
 View larger version (29K): [in a new window] |
Figure 3. Anti-TNF- inhibits TCR/CD28 but not PMA/ionomycin-induced activation. Lymphocytes were stimulated for 5 days with anti-TCR plus anti-CD28 mAb or with PMA and ionomycin as described in Materials and Methods in the presence or absence of TNF- (20 ng/ml), neutralizing anti-TNF- mAb (2 µg/ml), or isotype-matched control antibody (Isotype-Co), and 3H-thymidine uptake and IFN- secretion were measured as described in Materials and Methods. Results of seven (a) and three (b) independent experiments are shown. Bars represent the mean ± SEM. *, Statistically significant difference as compared with control cultures stimulated via TCR/CD28 alone (i.e., Med.-Co).
|
could be mimicked by an anti-TNF-RII antibody but not by two different anti-TNF-RI antibodies or an isotype-matched control antibody (Fig. 1 a
and 1b)
and was dose-dependent within a range of 0.02–20 ng/ml for TNF- and 0.01–1000 ng/ml for anti-TNF-RII mAb (costimulation of TCR-dependent lymphoproliferation; data not shown). Biologic activity of the two TNF-RI antibodies used was demonstrated by cytotoxicity assays in human CCL-75 fibroblasts (data not shown). We thus conclude that TNF-RI is not present on resting and activated T cells (always less than 1% TNF-RI-positive cells as determined by flow cytometry; data not shown), or low receptor expression, undetectable by flow cytometry, is functionally not sufficient. As TNF- was also costimulatory in highly purified CD4+ cells (Fig. 2
), we propose that costimulation through TNF- is a direct effect mediated by TNF-RII expressed on the T cell.
 View larger version (22K): [in a new window] |
Figure 2. Costimulation via TNF-RII is equally effective in purified CD4+ cells and in nonadherent lymphocytes. CD4+ cells were purified from PBMC by positive immunoselection with magnetic beads as described in Materials and Methods and were stimulated for 5 days with anti-TCR-coated beads in the presence or absence of TNF- or anti-TNF-RII mAb. Bars represent the mean ± SEM of eight independent experiments. *, Statistically significant costimulation as compared with cells stimulated via TCR alone.
|
for fully functional T cell activation by showing that endogenously secreted TNF- is required for optimal T cell activation. Neutralizing anti-TNF- antibodies could inhibit T cell proliferation and IFN- release induced by anti-TCR and anti-CD28 mAb, and stimulation with PMA plus ionomycin was unaffected (Fig. 3)
. These results indicate that the costimulatory effect of TNF- can be overcome by direct stimulation of protein kinase C and Ca++ influx bypassing TCR stimulation; however, a quantitative effect, e.g., through higher levels of TNF- induced following PMA/ionomycin stimulation, cannot be excluded. Addition of exogenous TNF- had no effect on TCR/CD28-induced proliferation or secretion of IFN- (Fig. 3)
, thus indicating that endogenous TNF- is sufficient for optimal costimulation following TCR/CD28 stimulation.
Costimulation via CD28 and costimulation via TNF-RII have different cellular effects and signaling requirements
As described previously by Lenschow et al. [21
], costimulation via CD28 required cross-linking, as CD28 mAb costimulated only if bound to the anti-TCR mAb-coated beads but not as a reagent kept in solution (data not shown). In contrast, soluble ligand was effective for TNF-RII costimulation, as soluble TNF- costimulated, and anti-TNF-RII mAb was fully effective if used as a soluble reagent (data not shown). Costimulation of cytokine release (e.g., INF- and TNF-) via CD28 and TNF-RII also followed different time kinetics (Fig. 4
). Peak secretion of IFN- and TNF- in lymphocytes costimulated via CD28 was approximately day 3 and showed a plateau until day 5, whereas peak secretion of IFN- and TNF- after TNF-RII costimulation only reached its peak on day 5 (Fig. 4)
. Furthermore, CD28 and TNF-RII coactivation pathways had different susceptibility to inhibitors of distinct signaling elements (Fig. 5
). Lymphocytes were stimulated for 5 days via TCR and CD28 or TNF-RII as a costimulus in the absence or presence of different concentrations of the following inhibitors: BAY 11-7082 (inhibitor of IB activation), SB 202190 (p38 inhibitor), PD 98059 (MKK1/2 inhibitor), and LY 294002 (PI-3K inhibitor). Inhibition of IB significantly reduced proliferation (Fig. 5a)
and cytokine secretion (TNF- and IFN-, Fig. 5b
) induced by TCR/CD28 as well as by TCR/TNF-RII. As compared with CD28, lower concentrations of p38, MKK1/2, and PI-3K inhibitors were sufficient to significantly inhibit costimulation of proliferative responses via TNF-RII (Fig. 5a)
. The effect of the different concentrations of inhibitors on costimulation of IFN- and TNF- secretion via CD28 versus TNF-RII (Fig. 5b
; data not shown) was comparable with their effect on the proliferative response (Fig. 5a)
. It is important that CD69 expression on CD4 and CD8 cells, which is known to be induced via TCR alone [25
], was not affected by the PI-3K inhibitor LY 294002, indicating that in this case, it was actually the costimulatory pathway and not the primary TCR signal that was blocked (data not shown).
 View larger version (18K): [in a new window] |
Figure 4. Costimulation via CD28 and via TNF-RII follows different kinetics. Lymphocytes were stimulated with anti-TCR coated onto beads and anti-CD28 (10 ng/ml) or TNF- (20 ng/ml) or anti-TNF-RII mAb (1 µg/ml) for the indicated time, and supernatants were harvested for measurement of cytokine release (IFN-, a; TNF-, b), as described in Materials and Methods. Data points represent the mean ± SEM of eight independent experiments. *, Statistically significant difference between costimulation via CD28 and TNF-RII.
|
 View larger version (34K): [in a new window] |
Figure 5. Costimulation via CD28 and TNF-RII differs in their sensitivity to inhibition of signal-transduction elements. Lymphocytes were stimulated for 5 days with the indicated stimulus in the presence of the indicated doses of the following inhibitors: BAY 11-7082, an inhibitor (Inh.) of IB; SB 202190, an inhibitor of SAPK/p38; PD 98059, an inhibitor of MAPK/ERK [MAPK kinase (MKK)1/2]; and LY 294002, a PI-3K inhibitor. Proliferation (a) and cytokine (b) release were measured as described in Materials and Methods. Data points and bars represent the mean ± SEM; n = 3–11. *, Statistically significant difference as compared with control cultures without inhibitors.
|
and TNF- to a comparable extent (Fig. 6a
). In contrast, levels of IL-13 were significantly lower if costimulated via TNF-RII as compared with CD28 costimulation (Fig. 6a)
, although with both costimuli, peak secretion was at day 5. Finally, in contrast to CD28, which is a strong costimulus for IL-2 secretion [21
, 26
], costimulation via TNF-RII did not induce secretion of IL-2 and was also unable to induce IL-10 release (Fig. 6a)
. This difference was not a result of a time-shift of IL-2 and IL-10 peak release following TNF-RII costimulation as compared with CD28 triggering, i.e., a delayed response to TNF-RII triggering (Fig. 6 b
and 6c)
.
 View larger version (25K): [in a new window] |
Figure 6. Costimulation via TNF-RII and via CD28 differentially regulates TCR-dependent T cell cytokine expression. (a) Lymphocytes were stimulated for 5 days (TNF-, IFN-, and IL-13), 3 days (IL-10), or 1 day (IL-2) with anti-TCR in the absence or presence of anti-CD28 or anti-TNF-RII mAb (10 ng and 1 µg/ml, respectively). Bars show the mean ± SEM. The number of experiments are: TNF-, n = 13; IFN-, n = 14; IL-13, n = 16; IL-10, n = 9; IL-2, n = 11. *, Statistically significant difference between costimulation via CD28 and TNF-RII; n.s., not significant. (b and c) Time kinetics of IL-10 (b) and IL-2 (c) release in response to TCR/CD28 versus TCR/TNF-RII from day 1 to 5 is shown. Data points represent the mean ± SEM of four (IL-10) and three (IL-2) independent experiments.
|
and IFN- (Fig. 7)
were not affected. Inhibition of TCR/CD28-dependent IL-2 secretion by anti-TNF-RII required extensive TNF-RII cross-linking; i.e., triggering with TNF-RII mAb bound to beads, as anti-TNF-RII mAb kept in solution or TNF- were unable to inhibit TCR/CD28-induced IL-2 release (data not shown).
 View larger version (19K): [in a new window] |
Figure 7. Concomitant triggering of TNF-RII inhibits TCR/CD28-induced IL-2 and IL-10 secretion but leaves secretion of TNF- and IFN- unaffected. Lymphocytes were stimulated for 5 days (TNF-, IFN-) or 1 day (IL-2, IL-10) with anti-TCR/anti-CD28-coated beads in the presence or absence of anti-TNF-RII mAb (1 µg/ml) or isotype-control antibody (Iso; 1 µg/ml) as indicated. Median values for TNF-, IFN-, IL-2, and IL-10 of unstimulated lymphocytes were <37 pg/ml. Bars represent the mean ± SEM. The number of experiments is given in the individual panel. *, Statistically significant difference between TCR/CD28-stimulated cultures with/without anti-TNF-RII.
|
 View larger version (39K): [in a new window] |
Figure 8. TCR/CD28-induced IL-2 expression is inhibited by TNF-RII cross-linking at the protein and mRNA level. Lymphocytes were stimulated for 24 h with beads coated with anti-TCR and anti-CD28 mAb alone or with beads coated with anti-TCR/anti-CD28 and anti-TNF-RII mAb or isotype-matched control antibody (Co- antibody) at the indicated doses. RNA was isolated, and IL-2 RT-PCR was performed as described in Materials and Methods. (A) Release of IL-2 into the cell supernatants, as determined by ELISA; (B) densitometric results; (C) fluoroimages of IL-2 RT-PCR products. (C) Lanes 1–4, Results from stimulated cells; lane 5, unstimulated cells; lane 6, negative PCR control without RNA. (B) Densitometric readings for GAPDH-specific RT-PCR products are given to show that expression of this housekeeping gene was comparable. The densitometric readings for IL-2-specific RT-PCR products are given after normalization for the respective GAPDH mRNA level, as calculated according to the following formula: IL-2GAPDH = IL-2/[GAPDH/average (GAPDHTCR+CD28, GAPDHTCR+CD28+aTNFRII 0.3 µg/ml, GAPDH TCR+CD28+aTNFRII 3.5 µg/ml, GAPDHTCR+CD28+isotype, GAPDHmedium control)].
|
 View larger version (51K): [in a new window] |
Figure 9. TNF-RII cross-linking inhibits TCR/CD28-induced intracellular Ca++ flux. (A and B) Lymphocytes were loaded with Fluo-3 and Snarf-1 and precoated with anti-TCR and anti-CD28 mAb with or without anti-TNF-RII, anti-TNF-RI, or isotype-control mAb as described in Materials and Methods and were analyzed by flow cytometry. After measurement of unstimulated cells for at least 1 min, mAb on the cell surface were cross-linked with biotinylated goat anti-mouse IgG (4 µg/ml) and streptavidin (1 µg/ml). Cells treated with the cross-linking reagents alone served as a negative control and cells treated with ionomycin (2 µg/ml), as a positive control. The addition of the respective stimulus (cross-linking reagents or ionomycin) is indicated by the break in the dot-plot recording over time (A) or an arrow (B). (A) Dot plots of one representative experiment out of three; (B) percentage of responding cells defined as described in Materials and Methods for Fluo-3 and Snarf-1 (mean±SEM of all three experiments). (C) Nonadherent lymphocytes (NA) were isolated from peripheral blood as described in Materials and Methods and immediately analyzed for surface expression of TNF-RII or cultured for 5 days with or without TCR/CD28 stimulation before flow cytometric analysis. The histograms show one representative experiment, and the percentages of TNF-RII-positive cells given in the panels represent the mean ± SEM of four independent experiments.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
[27
, 28
], but other studies showed a costimulatory effect of TNF- in highly purified peripheral human T cells [29
30
31
] or in human thymocytes [32
] activated with mitogenic anti-CD3 mAb or IL-2. Subsequently, it was found that TNF-RII (p75)-deficient mouse T cells have a decreased, proliferative response following stimulation via CD3, further supporting the concept that TNF-RII expressed on the T cell is responsible for the costimulatory activity of TNF- and that this cosignal is essential for sufficient T cell responses [33
].
Using mAb for T cell stimulation, we confirm that TNF- is directly costimulatory for TCR-mediated activation of resting as well as preactivated human T cells through activation via TNF-RII but not TNF-RI. Furthermore, we extend previous findings by showing that this costimulatory activity is distinct and independent from CD28-mediated activation signals. First, we noticed that maximum costimulatory activity for cytokine induction is later in TNF-RII costimulation as compared with CD28 costimulation. One possible explanation could be the requirement for up-regulation of TNF-RII on T cells, as only low but clearly detectable levels of TNF-RII are expressed on freshly isolated T cells. A further important difference between the costimulatory action of TNF-RII and CD28 is the requirement for receptor cross-linking. Although TNF-RII triggering by a soluble ligand such as TNF- was sufficient for costimulation, CD28-mediated activation strictly required cross-linking. These results reflect in vivo findings, where soluble TNF- is an effective cosignal for T cell activation in the mouse model [34
]. The fact that only particle-bound anti-CD28 costimulates TCR activation could reflect the requirement for multiple interactions between CD28 on the T cell with B7 molecules expressed on the antigen-presenting cell [21
].
In addition, costimulation via TNF-RII and CD28 differed in the pattern of T cell cytokines up-regulated, as TNF-RII costimulation led to TNF- and IFN- release without induction of IL-2 and IL-10, observed over an incubation period of 1–5 days, and peak IL-13 levels were significantly lower as compared with CD28 costimulation, which strongly up-regulated all five cytokines investigated, as has been previously shown [35
]. Differential regulation of IFN- and IL-2 has been described following stimulation via CD81 [36
] or IL-12 [37
], and cyclosporin A can inhibit anti-CD3/CD28-induced IL-2 expression on the mRNA and protein level, and IFN- production is largely unaffected or even increased [38
]. Our findings further support previous results showing that TNF- affects the T helper cell type 1 (TH1)/TH2 balance by promoting TH1-type regulation [39
], as two important TH2-type cytokines were significantly lower following costimulation via TNF-RII as compared with CD28 (i.e., IL-13) or were not costimulated at all (i.e., IL-10).
Although the mechanisms involved in induction of apoptosis by TNF-R family molecules are widely investigated [4
], much less is known about the costimulatory signaling requirements leading to TNF-RII-mediated costimulation of T cell activation. Vav1 has been identified as a crucial point of integration for TCR and CD28-mediated activation [40
], and transduction elements involved in NF-B activation seem to be prime candidates for the integration of signals derived from the TNF-RII and the TCR. The intracellular segment of TNF-RII can bind signaling molecules such as TNFR-associated factor 2 (TRAF2) or CD40bp [41
], which ultimately leads to NF-B activation known also to be indispensable for TCR-mediated T cell activation [42
]. Our findings show that CD28 and TNF-RII costimulation strictly depend on NF-B activation. However, TNF-RII-dependent signal-transduction requirements were more stringent and much more sensitive to inhibition of p38, MKK1/2, and PI-3K, as concentrations that left CD28 costimulation largely unaffected completely inhibited TNF-RII cosignaling. Multiple pathways are active in TCR/CD28-mediated activation [35
] and apparently are able to replace each other to a certain extent. This redundancy seems to be more limited or does not exist at all in the case of TCR/TNF-RII-mediated activation. Furthermore, our finding that IL-2 is unresponsive to TNF-RII triggering suggests that activation of NF of activated T cells is not involved in TNF-RII signaling, which is consistent with previous findings that TNF- leads to NF-B activation without stimulation of IL-2 gene expression [43
].
TRAF1, another member of the TRAF family that interacts with TNF-RII, has previously been implicated in the negative regulation of T cell activation, as TRAF1-deficient mice showed exaggerated T cell proliferation in response to stimulation with anti-CD3 [44
]. In view of their results, Tsitsikov et al. [44] suggested that TRAF1 normally inhibits TCR/CD3-mediated activation by interfering with signaling pathways different from CD28 or IL-2. The present results show for the first time that TNF-RII can have a negative effect on human T cell activation. Following stimulation via TCR plus CD28, extensive TNF-RII cross-linking specifically inhibited the expression of certain cytokines (IL-2, IL-10), and other activation events (T cell proliferation, IFN-, and TNF- secretion and up-regulation of T cell activation markers) were unaffected. A trivial explanation for this TNF-RII-mediated regulatory effect such as decreased cell viability seems unlikely, as inhibition of IL-2 secretion was specific for TCR-dependent T cell activation and left PMA plus ionomycin-induced IL-2 secretion unaffected. In addition, inhibition of IL-2 secretion was a rapid effect and thus completely different from the previously described attenuation of TCR stimulation by chronic exposure to TNF-, which required TNF exposure for several days [10
]. As IL-10 release is also inhibited, TNF-RII-mediated down-regulation of IL-2 expression cannot be explained by overinduction of this well-known inhibitor of IL-2 expression [45
].
As TNF-RII cross-linking inhibited IL-2 production at the mRNA level, a direct inhibitory effect of TNF-RII stimulation on signaling via TCR plus CD28 seemed likely. Our finding supported this assumption by showing that intracellular Ca++ flux following concomitant cross-linking of TNF-RII in TCR plus CD28-stimulated T cells is inhibited. This finding points toward a very proximal interaction point between TCR and TNF-RII signaling pathways and explains why TNF-RII-mediated down-regulation of TCR-dependent activation specifically involved a calcium influx-dependent cytokine such as IL-2. One possibility could be that extensive TNF-RII cross-linking, e.g., with mAb coated onto beads, involves distinct signaling elements. Differential effects depending on the extent of cross-linking have been described previously, e.g., in T cell hybridomas where soluble anti-CD3 antibody was able to stimulate IL-2 secretion, and plastic-bound anti-CD3 antibody stimulated TNF- production [46
]. Although further studies are required to clarify the molecular mechanism(s) involved, e.g., TRAF1 [44
], the present findings already underline the unique role of TNF-RII as an inhibitor of early TCR signaling events proximal to intracellular calcium mobilization, which ultimately leads to modulation of the T cell cytokine pattern expressed.
Received March 19, 2003; revised May 27, 2003; accepted May 28, 2003.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. E. Munroe and G. A. Bishop A Costimulatory Function for T Cell CD40 J. Immunol., January 15, 2007; 178(2): 671 - 682. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. B. Thornley, M. A. Brehm, T. G. Markees, L. D. Shultz, J. P. Mordes, R. M. Welsh, A. A. Rossini, and D. L. Greiner TLR Agonists Abrogate Costimulation Blockade-Induced Prolongation of Skin Allografts J. Immunol., February 1, 2006; 176(3): 1561 - 1570. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Y. Kim, J. J. Priatel, S.-J. Teh, and H.-S. Teh TNF Receptor Type 2 (p75) Functions as a Costimulator for Antigen-Driven T Cell Responses In Vivo J. Immunol., January 15, 2006; 176(2): 1026 - 1035. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Depuydt, G. Van Loo, P. Vandenabeele, and W. Declercq Induction of apoptosis by TNF receptor 2 in a T-cell hybridoma is FADD dependent and blocked by caspase-8 inhibitors J. Cell Sci., February 1, 2005; 118(3): 497 - 504. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
