Published online before print November 21, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Centro Nacional de Microbiología, Instituto de Salud Carlos III, Majadahonda, Madrid, Spain;
Robarts Research Institute and Departments of Microbiology and Immunology and Medicine, The University of Western Ontario, London, Canada; and
Centro de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas, Madrid, Spain
1 Correspondence: Centro Nacional de Microbiología, Instituto de Salud Carlos III, Ctra. Majadahonda-Pozuelo km 2, Majadahonda, 28220-Madrid, Spain. E-mail: pportols{at}isciii.es
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-associated protein-70 (ZAP-70), Vav-1, Akt, and extracellular signal-regulated kinase (ERK) phosphorylation but also costimulation-dependent mitogen-activated protein kinases (MAPK), such as the stress-activated c-Jun N-terminal kinase (JNK). It is intriguing that Crry costimulus enhanced p38 MAPK activation in T helper cell type 1 (Th1) but not in Th2 cells. A fraction of Crry is found consistently in the detergent-insoluble membrane fraction of Th1 or Th2 cells or CD4+ lymphoblasts. Crry costimulation induced clustering of lipid rafts, increasing their content in Crry, CD3
, and p59-60 forms of p56lck, and caused actin polymerization close to the site of activation in Th2 cells. Such events were inhibited by wortmannin, suggesting a role for phosphatidylinositol-3 kinase in these effects. The Crry cytoplasmic domain was required for JNK activation and interleukin-4 secretion but not for the presence of Crry in rafts or activation of p56lck, ZAP-70, Akt, Vav-1, or ERK. This suggests that Crry costimulation involves two different but not mutually exclusive signal transduction modules. The dual function of Crry as a complement regulatory protein and as a T cell costimulator illustrates the importance of complement regulatory proteins as links between innate and adaptive immunity.
Key Words: costimulation • cell surface molecule • signal transduction
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-associated protein-70 (ZAP-70) and the scaffolding protein linker for activation of T cells (LAT), which then recruits different adaptor proteins and enzymes such as Grb-2, phospholipase C 
1 (PLC1), and phosphatidylinositol-3 kinase (PI-3K). This cascade eventually leads to activation of Ras and the mitogen-activated protein kinase (MAPK) pathways, including the extracellular signal-regulated kinase (ERK) pathway (for reviews, see refs. [1
, 2
]) and the "stress"-activated c-Jun N-terminal kinase (JNK) pathway. Efficient activation of JNK requires not only TCR/CD3 engagement but also second signals provided by costimulatory membrane molecules [3
4
5
]. ERK and JNK phosphorylation promotes activation/differentiation of the T cell and may modify the outcome of the immune response by shifting the differentiation pattern of T cells [6
7
8
9
]. Thus, ligation of costimulatory molecules is important to achieve efficient and adequate effector immune responses. CD28 is the main costimulatory molecule for naïve T lymphocytes. Its ligation amplifies interleukin (IL)-2 secretion, which is the primordial growth factor for T cells. However, CD28 is dispensable, and other molecules, such as members of the immunoglobulin (Ig) superfamily (CD2, CD4, CD8), the tumor necrosis factor family (CD154), or the integrin family (lymphocyte function-associated antigen-1), may play a costimulatory role in T cell activation [7 ].
The family of regulators of complement activation (RCA) includes a group of proteins expressed on the surface of cells, including hemopoietic cells, which protect them against the deleterious effects of autologous complement activation. Some members of this family are transmembrane proteins, such as CD35 [complement receptor 1 (CR1)], CD21 (CR2), CD46, and murine CR1-related protein Crry/p65 (Crry), and others are glycosylphosphatidylinositol (GPI)-anchored proteins, such as CD55 [decay accelerating factor (DAF)] or CD59. Besides protection, other functions have been described for some RCA members, as amplification of B cell responses [10 , 11 ] or binding of pathogens (see ref. [12 ] for a review). In addition, some RCA molecules expressed by T cells (i.e., Crry, CD46, CD55, CD59) can play a costimulatory role [9 , 13 14 15 16 17 18 19 ] and may thus be critical to link innate with adaptive immunity. This linkage function is extremely important, as signals promoted by the immediate innate immunity mediators (as CRs and complement regulators) may later shape and fine-tune the adaptive response to pathogens or self-antigens.
Expression of membrane complement regulatory proteins is different in human and mouse T lymphocytes. Unlike their human counterparts, murine T cells do not express CR1, CR2, or CD46 [20 , 21 ]. Thus, Crry [9 , 22 ], and perhaps DAF [23 ] provide the main protection from autologous complement in mouse T lymphocytes. We have described that monoclonal antibodies (mAb), recognizing different epitopes on Crry, promote costimulation of CD4+ T lymphocytes. One of them (P3D2) interferes with Crry complement-regulatory function and also triggers a costimulatory signal, which promotes differentiation of CD4+ naïve T cells toward a T helper cell type 2 (Th2) phenotype [9 ]. This fact suggests the possibility that ligands, related or not to complement, can regulate protection from complement attack and T cell costimulation, controlling immune responses.
Here, we report that Crry enhances not only TCR/CD3 signaling by increasing the activation of proximal molecules such as p56lck, ZAP-70, Vav, Akt, and ERK but also costimulation-dependent MAPK such as JNK in Th cell lines. We have used a Th2 cell line to show that deletion of the intracellular domain of Crry differentially affects ERK and JNK activation, indicating the existence of two different modules of signal transduction associated to Crry. In addition, we observe that a fraction of Crry is within lipid rafts in Th1 or Th2 cell lines and CD4+ lymphoblasts. Such pool of Crry increases upon cross-linking with a specific mAb. Coligation of Crry and TCR/CD3 promotes raft clustering and actin polymerization close to the site of activation. All these signals strongly potentiate IL-4 and IL-5 secretion in Th2 cells. These data highlight the importance of CRs as multifunctional molecules influencing the outcome of the immune response and reveal some of the signaling mechanisms used by Crry to costimulate CD4+ T cells.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, rat IgG2b) [24
]; P3D2 (anti-mouse Crry, rat IgG2a) [9
]; GK1.5 (anti-mouse CD4, rat IgG2b) [25
]; 53-6.72 (TIB-105, anti-mouse CD8, rat IgG2a) [26
]; M 1/70 (anti-CD11b, rat IgG2b) [27
]. These antibodies were purified by affinity chromatography on protein G- or A-sepharose columns (Amersham Biosciences, Little Chalfont, UK). Rabbit anti-Crry, anti-Lck, and anti-CD3 antisera were obtained by immunization with fusion proteins of glutathione S-transferase with Crry (residues 123–253), human Lck (residues 22–51), and mouse CD3 (residues 1–56), respectively. The antibodies were purified by affinity chromatography over columns of the immunizing fusion protein coupled to Sepharose. The rabbit antiserum against ZAP-70 has been described previously [28
]. The PY-20 mouse antiphosphotyrosine mAb was obtained from Transduction Laboratories (BD Bioscience, Erembodegen, Belgium). Affinity-purified rabbit polyclonal antibodies specific for active ERK and active JNK were purchased from Promega Corp. (Madison, WI). Rabbit antiphospho-p38 antibodies were obtained from Cell Signal Technologies (Beverly, MA). Anti-ERK2, -JNK1, -Vav, or -p38 antibodies were from Santa Cruz Biotechnology (CA). Horseradish peroxidase (HRP)-coupled goat anti-rabbit or anti-mouse Ig antibodies (Sigma Chemical Co., St. Louis, MO) were used as secondary antibodies for immunoblot. Culture supernatants of the cell lines X63Ag8-653 BMG-Neo murine IL-2 (mIL-2) and X63Ag8-653 BMG-Neo mIL-4 were used as a source of mouse IL-2 and IL-4, respectively [29
]. Mouse recombinant (mr)IL-1
, cytochalasin D, and wortmannin were purchased from Calbiochem (La Jolla, CA). IL concentration in culture supernatants was determined by capture enzyme-linked immunosorbent assay (ELISA) using 11B11 (capture antibody) and biotinylated BVD6-24G2 (antibody for detection, BD PharMingen, San Diego, CA) for IL-4 detection and OptEIA mouse set for mIL-5 detection (BD PharMingen).
cDNA cloning and generation of cytoplasmic domain-deleted forms of Crry (
Cyt-Crry) cDNA constructs
Total RNA was isolated from SR.D10 cells using the SV total RNA isolation system (Promega Corp.). cDNA coding for Crry was obtained by reverse transcriptase-polymerase chain reaction (RT-PCR) with the Access RT-PCR system (Promega Corp.). The 5'-primer oligonucleotide (sense, 5' CTACACCATTTGCCGTAAAACGTTGTTTGAGAACGGTG 3') encompassed nucleotides –48 to –8 of the 5'-untranslated region of Crry, and the 3'-primer (antisense, 5' GCTATTTAGGAGACTTCTTGAGTGAGTGAATTCCGTGC 3') included nucleotides 1772–1809, which includes the stop codon (1804). Both sequences were obtained from GenBank (accession number L19874). The resulting PCR fragment was cloned into the pNeoSR vector using the TOPO-TA cloning system (Invitrogen BV, The Netherlands). This plasmid was termed pSR-Crry full-length (FL). Crry, lacking the cytoplasmic domain, was constructed by adding a stop codon at nucleotide 1699, corresponding to amino acid residue 387 (first residue in the cytoplasmic domain) using the QuickChange site-directed mutagenesis kit (Stratagene Cloning Systems, La Jolla, CA). This construct was subcloned into pNeoSR and was termed pSR-CrryCyt. The mutation was confirmed by DNA sequencing.
Cell lines and transfection
SR.D10 [30
] is a clone obtained from the murine CD4+ Th2 cell line D10.G4.1 specific for Conalbumin fragment 134–146, bound to I-Ak class II major histocompatibility complex molecules [31
]. It was maintained in Click’s medium supplemented with 10% heat-inactivated fetal calf serum (FCSi; culture medium) containing 5 U/ml mrIL-2, 10 U/ml mrIL-4, and 25 pg/ml mrIL-1 (IL medium).
AE-103 is an I-Ak-specific Th1 CD4+ mouse T cell line [32 ] and was grown in culture medium supplemented with mouse IL-2.
To obtain Crry-deficient cells, SR.D10 cells (107 in 10 ml culture medium) were irradiated (500 Rad, 10 min) and grown for 10 days in IL medium. The cells were then negatively selected by two cycles of "panning" over P3D2 antibody-coated plates, and unbound cells were grown for 4 days in IL medium. Then, the cells (70% Crry–, 100% TCR+, as determined by flow cytometry) were cloned in IL medium by limiting dilution. The resulting clones were checked for absence of Crry expression by flow cytometry and RT-PCR. Functional studies to check normal pattern of tyrosine phosphorylation were performed on selected clones activated by pervanadate. One Crry-deficient representative (SR.Crry– G3) clone was selected for further use.
To obtain Crry transfectants, plasmids pNeoSR, pSR-FL-Crry, and pSR-Cyt-Crry were linearized with ScaI and transfected (10 µg DNA/ml) into SR.Crry–-G3 [5x106 cells in 0.8 ml phosphate-buffered saline (PBS)] by electroporation at 960 µF and 260 V using a Gene Pulser (Bio-Rad, Hercules, CA). Electroporated cells were distributed in flat-bottom 96-well culture plates and cultured overnight in IL medium; Geneticin (Sigma Chemical Co., 800 µg/ml) was added to the surviving cells. Neomycin-resistant cells were cloned, and clones expressing adequate levels of Crry, TCR/CD3, and CD4 were selected.
Cell stimulation and lysis
Cells were stimulated using polystyrene microspheres (Polybeads, 4.5 µm diameter, Polysciences Europe GmbH, Eppelheim, Germany). The microspheres were coated by incubation (16 h at 4°C) with combinations of anti-CD3 mAb (YCD3-1) or isotype control (M1/70; 2 µg/ml) plus costimulatory mAb (P3D2, GK1.5, or isotype control TIB105; 10 µg/ml) in PBS unless stated otherwise. Beads were washed, mixed with T cells at a 1:1 cell:bead ratio, 108 cells/ml, and incubated at 37°C for the times indicated. The reaction was stopped by adding ice-cold PBS, 0.5 mM EDTA, and 1 mM Na3VO4. Cells were spun down and lysed in 50 mM Tris, 300 mM NaCl, pH 7.6, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, 1 µg/ml leupeptin, 1 mM Na3VO4, 1 mM EGTA, 50 mM NaF, and 1 mM MgCl2 (lysis buffer) at 4 x 107 cells/ml. Postnuclear supernatants were used for immunoprecipitation and/or immunoblot.
For lipid rafts analysis, T cells were activated by antibodies cross-linked to sheep red blood cells (SRBC) instead of using polystyrene beads. Briefly, SRBC (Oxoid, Madrid, Spain) were washed twice in freshly prepared saline, and the antibodies were coupled, as described previously [33 ], with minor modifications as follows: 108 SRBC in 1 ml saline were mixed with antibodies (4 µg/ml anti-CD3 or control antibodies plus 20 µg/ml costimulator or control antibodies), resuspended in saline in the presence of 0.1 ml 0.1% CrCl3, and allowed to react for 5 min at room temperature while shaking. The reaction was stopped by addition of PBS, and eventually, antibody-coated SRBC were suspended in PBS containing 10% FCSi and kept at 4°C until use. T cells and the antibody-coated SRBC were incubated for 30 min at 37°C as described for polystyrene beads. The reaction was stopped by addition of ice-cold PBS, 0.5 mM EDTA, and 1 mM Na3VO4 and the resulting lysates used for lipid raft isolation by sucrose gradient centrifugation (see below).
Immunoprecipitation and immunoblotting
Immunoprecipitation and immunoblotting were carried out as described [19
]; lysis buffer containing Triton X-100, as indicated above, was used. Immunoprecipitates from 107 cells were performed unless indicated otherwise. Immunoblots of cell lysates were made with lysates from 3 x 105 cells.
Densitometric analysis was performed with Fuji Bas equipment (Tokyo, Japan) and the PC Bas 2.09 computer program or the National Institutes of Health Image 1.61 image analysis program.
Immunofluorescence confocal microscopy
For visualization of lipid raft clustering or actin polymerization, T cells were stimulated for 20 min at 37°C with polystyrene beads coated with antibodies as described above. The stimulated cells were seeded on poly-L-lysine-coated coverglasses, fixed with 4% paraformaldehyde in PBS for 5 min at 37°C, and washed. Surface ganglioside monsialic acid (GM1) was then stained with cholera toxin subunit B-fluorescein isothiocyanate (CTB-FITC) at 5 µg/ml for 30 min at 4°C, washed, and mounted. For detection of polymerized actin, the fixed cells were permeabilized 5 min with 0.1% saponin in PBS/0.1% bovine serum albumin (PBS saponin), stained (10 min at 20°C) with 20 nM FITC-coupled phalloidin (Sigma Chemical Co.) in PBS saponin, washed with PBS saponin, and coverglasses mounted as above.
Confocal microscopy analysis was performed in an Axiomat135 Zeiss microscope equipped with a MRC1024 Bio-Rad confocal sytem (Bio-Rad) or a Leica TCS-SP2-AOBS-UV ultraspectral confocal microscope. Immunofluorescence analysis was done using a 63 x 1.4 objective lens at 0.5 µm intervals. Signals from different fluorescent probes were taken in parallel. One hundred cells were analyzed for each labeling condition, and representative results are shown. Image processing was performed with Adobe Photoshop (Adobe Systems, Mountain View, CA).
Lipid raft isolation
Resting cells or cells activated with antibodies coupled to SRBCs were spun and used for isolation of lipid rafts by centrifugation in a step gradient of sucrose as described [34
]. Twelve 1 ml fractions were collected from the top. Fractions 5–6 contained a cloudy band of rafts, and fractions 11–12 were considered as the soluble ones. Unless otherwise indicated, fractions were precipitated with acetone and resuspended in 1x sodium dodecyl sulfate (SDS) Laemmli sample buffer before electrophoresis.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (69K): [in a new window] |
Figure 1. Effect of Crry costimulation on early signaling in CD4+ cells. (A) SR.D10 Th2 cells were incubated for 7 min with polystyrene beads coated with anti-CD3, anti-Crry, or control antibodies, as indicated in the figure. ZAP-70 (left panel) and Vav immunoprecipitates (IP; middle panel) or postnuclear lysates of SR.D10 cells (right panel) were immunoblotted for phosphotyrosine (pTyr; in ZAP-70 and Vav) or phosphorylated Akt (pAkt). The blots were stripped and reprobed with specific antibodies as controls for loading. (B) Kinetic analysis of ERK and JNK phosphorylation in lysates of SR.D10 stimulated at different times with combinations of antibodies as above and immunoblotted with anti-pERK, -ERK2, -pJNK, or -JNK antibodies. (C) Analysis of p38 MAPK phosphorylation in lysates of SR.D10. Cells were activated and costimulated as in A and immunoblotted for activated, pP38, or total p38. (D) Analysis of signaling intermediaries in AE-103 Th1 cells activated as in A. ERK, JNK, p38, and Akt were analyzed by immunoblot with specific antibodies as in A–C.
|
1 and PI-3K. Generation of phosphatidylinositol 3-phosphate by PI-3K leads to the recruitment of the serine/threonine kinase Akt/protein kinase B to the cell membrane, where it is activated. Our results show that Crry also fostered early activation of Akt (Fig. 1A
, right panel), indirectly indicating the participation of PI-3K in the costimulatory process mediated by Crry signaling.
Crry-mediated costimulation enhances activation of ERK and JNK MAPK
We have previously demonstrated an enhanced activation of the ERK MAPK mediated by Crry costimulatory signals in unprimed CD4+ lymphocytes [9
]. Then, in Th2 cells, we asked whether Crry costimulatory signal increased TCR-dependent MAPK activation, such as ERK, or if it can also activate JNK, which is strongly potentiated by costimulatory molecules such as CD28 [3
, 35
, 36
] or inducible costimulator (ICOS) [28
], but it is not efficiently activated by TCR signals alone. Thus, we tested the effects of Crry costimulation on the activation of the three MAPK pathways described in T cells (ERK, JNK, and p38; Fig. 1B
and 1C
). We observed that CD3 ligation alone induced a low-level of ERK activation at all the times assayed (Fig. 1B
, upper panels). Anti-CD3 activation induced low-level pJNK after 15 min, and longer exposure of the blots did not reveal JNK activation at shorter times (Fig. 1B
, lower panels), indicating different kinetics of ERK and JNK activation. In contrast, activation by anti-CD3 plus anti-Crry strongly enhanced the phosphorylation of ERK (
40-fold) and JNK (100-fold), modifying the activation kinetics, as pERK and pJNK appeared earlier and lasted longer when compared with anti-CD3 stimulus alone. It is interesting that we found that activation of the p38 MAPK was not enhanced by Crry signaling (Fig. 1C)
, illustrating the selective potentiation of the ERK and the JNK MAPK pathways by Crry in Th2 cells.
To ensure the extent of these findings, the effect of Crry costimulus on MAPK and Akt activation was also analyzed in AE-103 Th1 cells. Crry costimulation markedly enhanced ERK, JNK, and p38 phosphorylation (Fig. 1D) . Akt phosphorylation was also fostered by the Crry costimulatory signal, indicating the participation of PI-3K in this process in Th1 cells (Fig. 1D) .
Role of Crry cytoplasmic domain in costimulation
As the complement regulatory activity of Crry does not need its cytoplasmic domain [37
38
39
], we asked whether this part of Crry was required for its costimulatory effects. To address this question, we first generated Crry-deficient Th2 cell lines. Crry deficiency was confirmed by flow cytometry and RT-PCR (data not shown). The Crry-deficient T cells had comparable TCR, CD3, CD4, and CD45 expression to the wild-type counterpart and also showed similar tyrosine phosphorylation patterns upon pervanadate activation or anti-CD3 stimulation (not shown). These cell lines were transfected with FL-Crry, with Cyt-Crry, or with empty vector as control. Transfectants were selected for normal expression of cell-surface activation molecules and proliferative and IL-4 responses to polyclonal stimuli. Surface expression of Crry in these reconstituted transfectants was similar in FL- or Cyt-Crry transfectants (not shown), although their surface expression levels were lower than those observed in the parental line SR.D10. Furthermore, stimulation with anti-CD3 alone yielded similar tyrosine phosphorylation patterns in the selected transfectants (not shown).
The costimulatory activity of Crry for ERK or JNK activation was determined in these transfectants. We found that coligation of Crry with CD3 efficiently enhanced ERK activation, independent of the presence of the Crry cytoplasmic tail (Fig. 2A
, FL-H5, Cyt-F6). This result was confirmed in a set of independent transfectants upon normalization with the ERK activation induced by anti-CD3 plus anti-CD4 antibodies (Fig. 2B)
, finding that the mean ERK activation was higher in cells expressing FL-Crry than in those expressing Cyt-Crry (70±7.3 vs. 50±5.2%; Fig. 2B ).
 View larger version (51K): [in a new window] |
Figure 2. Role of the cytoplasmic domain in Crry-dependent costimulation of ERK and JNK activation. (A) Parental SR.D10 cells, a Crry-defective control transfectant (Neo-A8), a FL-Crry transfectant (FL-H5), and a cytoplasmic domain-defective Crry transfectant (Cyt-F6) were activated for 7 min with polystyrene beads coated with control filling antibodies (lane 1), anti-CD3 plus filling antibodies (lane 2), or anti-CD3 plus anti-Crry (lane 3). Activation with anti-CD3 plus anti-CD4 (lane 4) was used as a positive control for activation. Postnuclear lysates were electrophoresed and blotted for activated or total ERK and JNK, as indicated. (B) Analysis of ERK and JNK phosphorylation in a panel of seven FL-Crry (FL) or 11 -Cyt-Crry transfectants (-Cyt). Cells were stimulated using combinations of antibodies as in A, and lysates were electrophoresed and blotted for active ERK and active JNK. Densitometric results in lanes activated with anti-CD3 plus anti-Crry were normalized against their respective load controls (total ERK or JNK1) and then referred to the maximal response (100%) obtained by activation with anti-CD3 plus anti-CD4 in each case. Bars represent the mean ± SE of the data from individual transfectants.
|
Differential effects of deleting the cytoplasmic domain of Crry on IL-4 and IL-5 production
Crry signals enhance IL-4 secretion and favor Th2 differentiation of normal CD4+ T lymphocytes, and CD28 promotes interferon- (IFN-) secretion in the same system [9
]. In SR.D10 Th2 cells, cocross-linking of CD3 and Crry also enhanced IL-4 secretion (Fig. 3A
). To analyze the effect of deleting the cytoplasmic domain of Crry on late steps of activation, we determined the production of IL-4 and IL-5 by FL-Crry or tail-less Crry transfectants in response to anti-CD3, plus or minus anti-Crry, using activation by anti-CD3 plus anti-CD4 as a positive control (Fig. 3)
. CD3-Crry cocross-linking had a clear, costimulatory effect on IL-4 secretion in FL-Crry-reconstituted transfectants (Fig. 3A
, middle row). In contrast, costimulation by anti-Crry did not significantly increase IL-4 secretion in Cyt-Crry transfectants (Fig. 3A
, bottom row), indicating that Crry-dependent costimulation on IL-4 production requires its cytoplasmic domain.
 View larger version (26K): [in a new window] |
Figure 3. Deletion of the cytoplasmic domain of Crry differently affects Crry-dependent costimulation of IL-4 or IL-5 response. A panel of Th2 cells expressing native Crry (SR.D10), cells defective for Crry expression (Neo), FL-Crry transfectants (FL), and transfectants of Crry lacking the cytoplasmic domain (Ct) were activated for 6 h with polystyrene beads coated with control antibodies (Control Ab), anti-CD3 plus filling antibodies (CD3 + Control Ab), or anti-CD3 plus anti-Crry (CD3 + Crry). Activation with anti-CD3 plus anti-CD4 (CD3 + CD4) was used as a positive control for activation. IL-4 (A) or IL-5 (B) secreted to supernatants was measured by capture ELISA. Mean of triplicate determinations and standard errors are shown. Results have been normalized, considering as 1 the value obtained by anti-CD3 activation. Asterisks indicate statistically significant differences (P<0.05) from the cultures activated by anti-CD3 alone.
|
Next, we analyzed IL-5 secretion in these cells (Fig. 3B)
. We found that Crry costimulation was vigorous in FL-Crry transfectants (Fig. 3B
, middle row). Albeit low, Cyt-Crry transfectants still had detectable and statistically significant Crry costimulation on IL-5 production, indicating that the cytoplasmic domain of Crry is not as important for IL-5 secretion as for IL-4 secretion.
Crry partitions within lipid rafts
As tail-less Crry was still able to provide some costimulatory signals, we hypothesized that this might be a result of partition of Crry into lipid rafts. This would allow for interactions between Crry and molecules participating in T cell activation in a way similar to what has been reported for GPI-anchored receptors.
In preliminary experiments, we observed by confocal microscopy that Crry could form patches, which colocalized with the GM1 ganglioside, a molecule used as a marker of lipid rafts [40 ] (data not shown). To confirm that there is a pool of Crry within lipid rafts, we performed sucrose gradient centrifugation and immunoblotting of lysates from resting cells. Figure 4A shows that most Crry is in the detergent-soluble fractions (fractions 10–12), yet 6–10% of Crry (as calculated by densitometry) is consistently present in the rafts (fractions 5–6) of Th2 resting cells (Fig. 4A) . p56lck kinase was used as a protein marker mainly localizing in raft fractions and ERK2 (Fig. 4A) , a cytosolic protein, as a marker of detergent soluble fractions, demonstrating that no cross-contamination between soluble and raft fractions occurred during the experiments. The integrity of raft fractions was further assessed by the presence of the ganglioside GM1 as a marker of rafts (Fig. 4A) .
 View larger version (33K): [in a new window] |
Figure 4. A pool of Crry partitions within lipid rafts. Analysis of sucrose gradient fractions of Triton-X 100 lysates from resting Th2 SR.D10 (A) or Th1 AE-103 cells (B). All fractions were processed as indicated in Materials and Methods, blotted, and developed with anti-Crry antiserum (Crry), anti-p56lck antiserum (Lck), and anti-ERK2 (ERK), or dot-blotted and developed with CTB-HRP to show the presence of GM1 in the detergent-insoluble membrane fractions. (C) The partition of Crry within lipid rafts in CD4+ T lymphoblasts is shown. CD4+ T cells were activated with concanavalin A and expanded in IL-2 containing medium as described [28
]. Four days later, cells were collected and incubated with anti-Crry P3D2 antibody (+) or purified normal rat Ig coupled to SRBC as control (–) and processed to be fractionated on a sucrose gradient as indicated in Materials and Methods. Fractions 5 and 6 containing cloudy bands (rafts) were pooled and precipitated with acetone before SDS-polyacrylamide gel electrophoresis (PAGE), yielding a 20x concentration. Soluble fractions were not acetone-precipitated before SDS-PAGE. Raft and soluble fractions were blotted and developed as in A.
|
 View larger version (38K): [in a new window] |
Figure 6. Crry partitioning into lipid rafts and role of its cytoplasmic domain. (A) Lysates from SR.D10 and FL- or Cyt-Crry cells activated with various stimuli were fractionated under sucrose gradient, electrophoresed, and immunoblotted with anti-Crry antiserum (Crry) or anti-Lck (Lck). Antibodies coupled to SRBC were used as stimuli in the combinations indicated in the figure. After each gradient, fractions 5 and 6 containing cloudy bands (rafts) were pooled and precipitated with acetone before SDS-PAGE, as indicated in Materials and Methods, yielding a 20x concentration. Analysis of raft fractions is shown in lanes 1–4. The soluble fractions (lanes 5–8) were not acetone-precipitated before electrophoresis. Arrows indicate low mobility and normal Lck forms. Presence of GM1 in rafts or soluble fractions was detected by dot-blot developed with CTB-HRP in the same samples used above. (B) Western blot and detection of CD3 chains and Lck in detergent-insoluble fractions of FL-Crry and Cyt-Crry (DCYT-Crry) activated and processed as in A, except that fractions were not precipitated with acetone before electrophoresis. Arrows indicate low mobility and normal Lck forms.
|
 View larger version (64K): [in a new window] |
Figure 5. CD3/Crry coligation induces lipid raft aggregation and reorganization of actin cytoskeleton at the activation interface. (A) Confocal microphotographs of SR.D10 cells activated with polystyrene beads coated with anti-CD3 (CD3), anti-Crry (Crry), or both antibodies (CD3+Crry) during 20 min. Phase contrast (PC) shows the contact between the cells and the activating bead. Staining by CTB-FITC indicates the presence of GM1 (lipid rafts). Lipid raft clusters are indicated with arrows. For CD3 + Crry, two different fields are shown. (B) Confocal microphotographs of a FL-Crry transfectant and a tail-less Crry transfectant (CYT-Crry) activated with beads coated with anti-CD3 plus anti-Crry and stained with CTB-FITC to show the presence of rafts as in A. GM1 patches are indicated with arrows. (C) Cocross-linking of CD3 and Crry induces actin polymerization in the bead–T cell interface. SR.D10 cells were activated for 20 min with microbeads coated with anti-CD3 (CD3), anti-Crry (Crry), or both antibodies (CD3+Crry) in the absence or presence of cytochalasin D (CytD; 10 µM) or wortmannin (Wort; 50 nM). Cells were then fixed, and polymerized actin was detected with phalloidin-FITC (arrows). Cells were analyzed by confocal microscopy for differential interference contrast (DIC) or green fluorescence (Actin). Control beads (Nil) were coated with poly-L-lysine. Original white bar represents 10 µm.
|
CYT-Crry). Actin polymerization has been linked to the formation of supramolecular activation complexes (SMACs) and caps during antigen or antibody T cell activation. Therefore, we analyzed if Crry costimulation induced actin polymerization close to the bead–T cell interaction site. Figure 5C shows that separated ligation of Crry or CD3 induced low actin polymerization in the bead–T cell contact zone. In contrast, beads coated with anti-CD3 plus anti-Crry antibody induced strong actin polymerization in the site of interaction. Such actin filament formation was blocked by cytochalasin D, an inhibitor of actin polymerization, and by wortmannin, an inhibitor of PI-3K.
Cross-linked Crry is translocated into lipid rafts
We then analyzed the presence of Crry in rafts after stimulation/costimulation of the cell by using separate or joint cross-linking of CD3 and Crry. In these experiments, we immobilized antibodies to SRBC to activate the cells so that the cross-linking base would not interfere with the isolation of lipid rafts. Figure 6A
(top) shows that in SR.D10 cells, cross-linking of Crry increases the amount of this molecule in lipid rafts (lanes 1–4) when this molecule is coligated with CD3 (lane 3) or alone (lane 4). This behavior was also observed in FL-Crry transfectants (Fig. 6A
, middle), and it is not affected by Crry cytoplasmic domain deletion (Fig. 6A
, bottom). In lymphoblasts from CD4+ T cells, the presence of Crry in rafts was also increased by cross-linking with specific anti-Crry antibodies (see Fig. 4C
).
Lck was used as a load control and rafts fraction marker. Figure 6A also shows that only when CD3 and Crry are cocross-linked, p56lck modifies its apparent molecular weight, yielding 59–60 kDa forms, regardless of the deletion of Crry-cytoplasmic domain. Seveal groups have previously described these slower mobility forms of p56lck after TCR/CD3 activation (i.e., refs. [42 43 44 45 ]) and are mainly a result of serine phosphorylation. The presence of GM1 in insoluble (Fig. 6A , lanes 1–4) but not in the soluble fractions (Fig. 6A , lanes 5–8) indicates the integrity of the sucrose fractions obtained in each case.
It is interesting that ligation of FL- or Cty-Crry in the presence of anti-CD3 also promoted the partition of CD3 chains into the glycosphingolipid-enriched membrane domain fraction (Fig. 6B)
. Together, these results indicate that the cytoplasmic domain of Crry is not necessary for Crry recruitment into rafts and that the Crry costimulus facilitates the inclusion of signaling molecules such as CD3 chains into lipid rafts.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Crry potentiates TCR/CD3-dependent signals in CD4+ Th2 as well as Th1 cells. These signals include phosphorylation of ZAP-70, Vav, and Akt and changes in electrophoretic mobility of p56lck.
The effect of the inhibitor wortmannin indicates that some of the effects of Crry signaling, such as cytoskeletal reorganization or ERK activation, involve PI-3K signaling (Fig. 5 and data not shown). The effect of Crry on Akt phosphorylation in Th2 as well as in Th1 cells is consistent with this claim, as this serine/threonine kinase is regulated by PI-3K [1 , 46 ], which can also regulate recruitment and activation of Vav-1. In T lymphocytes, Vav-1 activation is induced by tyrosine kinases of the src family and by ZAP-70 (for a review, see [47 ]). Thus, the enhanced activation of Vav is in agreement with the effect of Crry on Lck, ZAP-70, and PI-3K.
As for CD28 signaling, Vav activation is probably critical for Crry costimulation. The Vav guanine nucleotide exchange factor activity induces Rac-dependent cortical actin changes, which favor membrane reorganization, leading to the formation of SMACs and clustering of lipid rafts (see refs. [1 , 47 , 48 ]). In this way, the aggregation of rafts induced by TCR signaling is amplified further [49 ]. As shown in Figure 5 , Crry-dependent costimulation induces raft aggregation at the site of activation and detectable incorporation of CD3 chains into rafts (Fig. 6B) , two phenomena linked with TCR activation and also with costimulation [41 , 50 , 51 ].
Partition of Crry into rafts and role of the cytoplasmic domain
To our knowledge, this is the first time that the partition into rafts of a membrane-spanning RCA molecule has been described in T cells. We have observed this finding in Th1, Th2, as well as in CD4+ lymphoblasts, demonstrating that the presence of a fraction of Crry in lipid rafts is a general finding for CD4+ T cells; other cell types possibly show a similar pattern. Transmembrane proteins belonging to other families, which are a resident of or induced to enter rafts, have been described. These include CD2 [52
], CD4, CD8, multichain receptors such as TCR and BCR (for a review, see ref. [40
]), or the TM-4-member CD81 [53
]. The fraction of Crry in lipid rafts found by us is similar or even higher to that found for CD2 and other transmembrane proteins in T cells [52
, 54
]. It is intriguing that ligation of Crry enhanced the fraction of this molecule included in rafts in Th2 (Fig. 6A)
as well as in CD4+ lymphoblasts (Fig. 4C)
, a phenomenon also observed in human CD2 [54
]. Our finding opens the possibility for ligands of RCA molecules, including pathogens, promoting the inclusion of these proteins into rafts, facilitating their interaction with signal transduction molecules, and even for the entry of pathogens, as recently described in humans [55
].
One remarkable finding of our studies is that the cytoplasmic domain of Crry is not required for its partitioning into lipid rafts (Fig. 6) or for some of its costimulatory effects. Surface proteins lacking a cytoplasmic domain such as CD48 [56 ] or the complement regulatory proteins CD55 (DAF) [57 ] and CD59 [15 ] are able to transduce activation signals in T lymphocytes by means of their presence into membrane lipid rafts. However, in the case of Crry, enhanced partitioning in the lipid raft fraction is not the sole cause for its costimulatory function (Fig. 6) . Moreover, Crry costimulus requires coligation of Crry with the CD3 in cis, suggesting that segregation into separate lipid rafts and multimerization of either molecule are not sufficient for productive costimulation (data not shown). Such a behavior has been reported for CD59 [15 ]. As the presence of Crry into rafts was not dependent on its cytoplasmic domain, we concluded that other parts of the molecule are likely implicated in its inducible raft association. In a different system, human cytotoxic T lymphocyte antigen 4 (CTLA-4) coclusters with the TCR and the ganglioside GM1 in the immune synapse in a CTLA-4 tail-dependent manner [34 ]. However, as happens to Crry costimulation, raft localization is not sufficient for CTLA-4-mediated, negative signaling. Although dispensable for Crry inclusion into rafts, its cytoplasmic domain is needed to achieve optimal raft coalescence to the activation zone (Fig. 5B) and a better costimulation of lymphokine secretion (Fig. 3) .
Crry has two signaling modules for cytokine production
Another question addressed here is that of the mechanisms used by Crry to transduce the costimulatory signal to the intracellular milieu of the T cell. Comparison of FL- and tail-less Crry allowed us to distinguish two signal transduction modules in this molecule. The first one is linked to early TCR-dependent signals (i.e., phosphorylation of ZAP-70, Vav-1, Akt, or p56lck or CD3 partition into rafts), which are not markedly affected by Crry cytoplasmic domain deletion. The other module is linked to JNK activation and requires the cytoplasmic domain of Crry. The low but detectable level of JNK activation induced by Crry in cells expressing tail-less Crry forms may be explained by the activation of Vav and Akt in these cells, as Vav activation induces JNK phosphorylation [58
]. Recently, Michel et al. [59
] observed that truncated forms of CD28 could efficiently costimulate some early signals including ZAP-70, LAT, or Vav phosphorylation but not others such as SLP-76, Itk, or PLC1 phosphorylation. The former signals were dependent on the strength of the interaction between the cell and its stimuli, whereas the latter would depend on specific interactions of the costimulatory molecule with enzymes or adaptor molecules.
The role of the cytoplasmic tail of Crry in anchoring/stabilizing the interaction of enzymes or adaptor molecules participating in TCR signaling needs to be ascertained further, as some of the tyrosine residues of its cytoplasmic sequence might interact with Src homology 2 domains upon phosphorylation. Candidates include the recently described molecule, Lck-interacting membrane protein [60 , 61 ], as a novel transmembrane raft-associated adaptor protein, associating with p56lck and favoring T cell activation of ERK1/2 and JNK but not p38. Another possibility is that the cytoplasmic domain of Crry recruits PI-3K, which by interaction and activation of Rac, might allow the costimulatory signal to activate JNK, independently on Vav or Akt activation [62 ]. Although we have not detected a direct Crry–PI-3K interaction so far (not shown), a clear role for PI-3K in Crry costimulation has been observed (Fig. 5 and data not shown).
Costimulation mediated by Crry in SR.D10 cells produces a strong and stable enhancement in TCR/CD3-induced ERK and JNK activation but not of p38. ERK phosphorylation was detectable at early times and lasted longer upon Crry costimulation. MAPK kinase inhibitors diminished IL-4 secretion (results not shown), confirming the linkage of ERK activation with Crry costimulation of effector function in Th2 cells. In mouse Th1 AE-103 cells, Crry enhances ERK, JNK, and p38 phosphorylation, promoting an increased IFN- production (unpublished results) and suggesting that p38 is a key element in the production of different ILs by T cells. It is interesting that ligands of CD46, a human complement regulatory protein homologous to Crry, also synergized with CD3 activation to enhance ERK phosphorylation [18
, 19
]. However, the eventual effect of ERK activation by CD46 is different, as it leads to inhibited IL-5 and enhanced IFN- and IL-2 secretion, favoring Th1 differentiation [19
]. In contrast, our data indicate that ERK contributes to costimulation of IL-4 and IL-5 secretion by Crry in mouse Th2 cells (not shown). These data show how the same activation pathways modified by costimulatory molecules such as Crry can differentially shape the immune responses in different systems.
It has been proposed that Jun kinase activation is a key point for integration of TCR and CD28 signals [3 ] and contributes to IL-2 mRNA stabilization [63 , 64 ]. In Th2 cells, which do not secrete IL-2, JNK is implicated in IL-4 production [28 , 65 ]. In this context, we show here that Crry costimulation enhances IL-4 secretion in SR.D10 Th2 cells in a JNK-dependent manner. As JNK inhibitors clearly inhibit IL-4 secretion by SR.D10 cells (ref. [28 ] and our unpublished data), deficient JNK activation in tail-less Crry transfectants might be linked to loss of IL-4 costimulation in these cells (Fig. 3) .
Crry also enhanced the secretion of IL-5, another lymphokine characteristic of Th2 cells. Like IL-4, IL-5 production was also diminished in tail-less Crry T cell transfectants, although to a lower extent than IL-4. Such a difference might be a result of different requirements of IL-4 or IL-5 production for transcription factors downstream of JNK [66 , 67 ]. However, we cannot rule out that the absence of the cytoplasmic domain of Crry might also affect other signal pathways activated by Crry (i.e., nuclear factor of activated T cells, our unpublished data).
In summary, our data suggest that Crry, by transducing signals to T cells, may provide a link between innate and adaptive immune responses. Such a possibility is shared with other RCA molecules such as CD46 [16 17 18 19 ]. Taking into account that many RCA proteins, including Crry, can mediate pathogen binding (for a review, see ref. [68 ] and for Crry, ref. [69 ]), it is tempting to speculate that complement proteins or pathogen-derived molecules can contribute to modification of host immune responses by binding to complement regulatory proteins.
Received November 3, 2004; revised July 26, 2005; accepted August 29, 2005.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
define distinct epitopes, one of which may interact with CD4 during T cell activation J. Immunol. 142,4169-4175[Abstract]This article has been cited by other articles:
 |
|
 |
|
  R. Bello, M. J. Feito, G. Ojeda, P. Portoles, and J. M. Rojo Loss of N-terminal Charged Residues of Mouse CD3{epsilon} Chains Generates Isoforms Modulating Antigen T Cell Receptor-mediated Signals and T Cell Receptor-CD3 Interactions J. Biol. Chem., August 3, 2007; 282(31): 22324 - 22334. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
