Antigen-induced regulation of T-cell motility, interaction with antigen-presenting cells and activation through endogenous thrombospondin-1 and its receptors

Abstract

Antigen recognition reduces T-cell motility, and induces prolonged contact with antigen-presenting cells and activation through mechanisms that remain unclear. Here we show that the T-cell receptor (TCR) and CD28 regulate T-cell motility, contact with antigen-presenting cells and activation through endogenous thrombospondin-1 (TSP-1) and its receptors low-density lipoprotein receptor-related protein 1 (LRP1), calreticulin and CD47. Antigen stimulation induced a prominent up-regulation of TSP-1 expression, and transiently increased and subsequently decreased LRP1 expression whereas calreticulin was unaffected. This antigen-induced TSP-1/LRP1 response down-regulated a motogenic mechanism directed by LRP1-mediated processing of TSP-1 in cis within the same plasma membrane while promoting contact with antigen-presenting cells and activation through cis interaction of the C-terminal domain of TSP-1 with CD47 in response to N-terminal TSP-1 triggering by calreticulin. The antigen-induced TSP-1/LRP1 response maintained a reduced but significant motility level in activated cells. Blocking CD28 co-stimulation abrogated LRP1 and TSP-1 expression and motility. TCR/CD3 ligation alone enhanced TSP-1 expression whereas CD28 ligation alone enhanced LRP1 expression. Silencing of TSP-1 inhibited T-cell conjugation to antigen-presenting cells and T helper type 1 (Th1) and Th2 cytokine responses. The Th1 response enhanced motility and increased TSP-1 expression through interleukin-2, whereas the Th2 response weakened motility and reduced LRP1 expression through interleukin-4. Ligation of the TCR and CD28 therefore elicits a TSP-1/LRP1 response that stimulates prolonged contact with antigen-presenting cells and, although down-regulating motility, maintains a significant motility level to allow serial contacts and activation. Th1 and Th2 cytokine responses differentially regulate T-cell expression of TSP-1 and LRP1 and motility.

Introduction

T-cell recognition of antigen plays a central role for cellular and humoral immunity. The T-cell antigen receptor (TCR) recognizes processed antigen peptides in complex with MHC molecules but T cells also require a second co-stimulatory signal to become fully activated. One of the best characterized co-stimulatory molecules expressed by T cells is CD28, which interacts with B7 on the surface of antigen-presenting cells. CD28 signals regulate T-cell trafficking, stabilize the immunological synapse, induce long-term survival, prevent anergy and generate T helper type 1 (Th1) and Th2 polarized cells.^1–4 CD28 co-stimulation seems to be particularly important for the development of Th2 cells through interleukin-2 (IL-2) -mediated priming for formation of the Th2 cytokine IL-4.^4–6 The CD28 family members cytotoxic T-lymphocyte antigen 4 (CTLA-4) and programmed death-1 are expressed following T-cell activation and inhibit excessive expansion of already activated cells, promote peripheral tolerance and protect against autoimmunity^7–9 and immune-mediated tissue damage.^10–12

Recognition of antigen by TCR induces formation of lymphocyte function associated antigen-1 (LFA-1)-mediated synapses with antigen-presenting cells, which play a fundamental role for the induction of adaptive immune responses.^13,14 However, the regulation of adhesive interactions between T cells and antigen-presenting cells and how TCR and CD28 signals are integrated by the T-cell still remains poorly understood.^15–18 It is interesting in this context that induction of protective immune responses, tolerance and prevention of autoimmunity collectively seem to depend on regulatory effects on T-cell motility through the TCR or co-stimulatory pathways.^19–24 The T-cell response to antigen challenge in vivo is characterized by a reduction of motility over several hours associated with brief serial contacts with antigen-presenting cells accompanied by prolonged contact.^19–24 The T cell therefore seems to integrate antigen signals from multiple antigen-presenting cells to be able to reduce motility and establish prolonged contacts. In contrast, antigen-specific tolerance is associated with transient T-cell contacts with antigen-presenting cells and the cells remain motile. There is also evidence that the development of specific T-cell immune responses correlate with differences in motility. Accordingly, Th1 and Th2 effector cells exhibit differences in tissue localization and chemokine receptor expression^25–27 and the Th1 cytokine IL-2 stimulates T-cell motility through endogenous T-cell thrombospondin-1 (TSP-1) whereas the Th2 cytokine IL-4 antagonizes this effect.²⁸

TSP-1 is a trimolecular calcium-binding protein with binding sites for integrins, integrin-associated protein (CD47), CD36, low-density lipoprotein receptor-related protein 1 (LRP1) and calreticulin, which mediates cell-to-cell and cell-to-matrix interactions and inhibits angiogenesis.^29–31 LRP1 is an endocytic and intracellular signalling protein with a broad repertoire of ligand interactions.^32,33 Calreticulin is a calcium-binding chaperone protein that associates with LRP1 on the cell surface and acts as a co-receptor for TSP-1.^34,35 Interaction of endogenous TSP-1 with its receptors CD47, LRP1 and calreticulin in cis within the same T-lymphocyte plasma membrane has been shown to regulate the development of polarized shape and translocation (migration) as well as adhesion to intercellular adhesion molecule-1 (ICAM-1) and fibronectin.^36–38 This integrated regulation of motility and adhesion makes adhesive stimuli from integrin ligands or CXCL12 prioritize motile responses before adhesion through LRP1-dependent proteolytic processing of TSP-1 and Janus kinase/signal transducer and activator of transcription signalling.^28,36–38 Formation of a 130 000 molecular weight fragment therefore seems to promote motility,^28,36–38 whereas intact TSP-1 mediates transient adhesion to ICAM-1 and fibronectin through the C-terminal domain via CD47 upon N-terminal triggering by calreticulin. In support of a role of TSP-1 for the function of the immune system, TSP-1-deficient mice show inflammatory infiltrates in multiple organs, which was attributed to poor TSP-1-dependent activation of transforming growth factor-β1.^39,40 TSP-1-deficient mice also exhibit prolonged inflammation in response to oxazolone treatment, which was attributed to defective elimination of activated T cells.⁴¹ Over-expression of TSP-1 further has a beneficial effect in contact hypersensitivity and TSP-1-derived peptides ameliorate inflammation.^42,43 Multiple independent evidence therefore indicates that TSP-1 has an anti-inflammatory function. However, deficiency of TSP-1 has also been shown to attenuate experimental autoimmune encephalomyelitis,⁴⁴ which suggests a more complex role with participation also in induction of immune responses. Here we show that antigen-induced reduction in T-cell motility, enhancement of contacts with antigen-presenting cells and activation depend on TCR and CD28-mediated regulatory effects on TSP-1 and LRP1. Furthermore, TSP-1 and LRP1 are shown to participate in the induction of immune responses and in the control of motility in Th1 and Th2 polarized cells.

Materials and methods

Chemicals and antibodies

Rat tendon collagen type I and fibronectin were purified and prepared as described elsewhere.³⁷ Poly-l-lysine (molecular mass 5300) was purchased from Miles-Yeda Ltd (Rehovoth, Israel). IL-2, IL-4, anti-IL-2 (clone 5334) and anti-IL-4 (clone 34019) were from Genzyme Diagnostics (Cambridge, MA). Dynabeads were from Dynal Biotech (Oslo, Norway). Anti-fibronectin (clone IST1, IgG1) was obtained from Sera Lab (Loughborough, UK). Anti-CD3 (clone SK7, IgG1) and anti-CD4 (clone SK3, IgG1) were obtained from Becton Dickinson (Mountain View, CA). Dynabeads were from Dynal Biotech. Mouse IgG, goat anti-mouse IgG, anti-CD45RA (clone 4KB5) and anti-CD45RO (clone UCHL1) were obtained from Dako (Glostrup, Denmark).). Anti-TSP-1 clone A6.1 (also called TSP-Ab-4, IgG1), clone C 6.7 (also called TSP-Ab-3, IgG1) and clone MBC 200.1 (also called TSP–Ab-9, IgG1) were from NEO-MARKERS (Fremont, CA). Anti-CD91 (clone A2MRα2, IgG1) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti anti-CD28 (clone CD28.2) was from BD Biosciences (San Jose, CA). Anti-calreticulin (clone FMC75) was from Biosite (Täby, Sweden). Biotinylated peroxidase and avidin were from Vector Laboratories (Burlingame, CA). The birch antigen Bet v G75 was obtained from ALK (Hoersholm, Denmark). Receptor associated protein (RAP) was obtained from Oxford Biomedical Research (Oxford, MI). ELT GAA RKG SGR RLV KGP D (hep1) was synthesized by the Biomolecular Resource Facility (University of Lund, Sweden). RSK AGT LGE RDL KPG ARV G (scrambled hep1 peptide), KRFYVVMWKK (4N1K) and KVFRWKYVMK (scrambled 4N1K) were synthesized by Tri pep (Novum Research Park, Huddinge, Sweden). RWI ESKHKS DFGKFVLSS (the TSP-1 binding site in calreticulin) and a scrambled control peptide (RSVWIKELGSKDSFHSF) were synthesized by the Biomolecular Resource Facility (University of Lund, Lund, Sweden).

Cells

Blood lymphocytes were purified using Lymphoprep and depleted of monocytes by treatment with carbonyl iron and magnetic removal of phagocytic cells. The cell preparations obtained consisted of 82–93% CD3-positive cells. Further enrichment of T cells was accomplished by depleting CD56-, CD19-, CD14-, CD45RA- and CD45RO-positive cells using beads coated with the corresponding antibodies. The experiments were performed with cells that had been cultured overnight and with the birch (Bet v) -specific CD4-positive T-cell clone AF 24. AF 24 was stimulated with anti-CD3 or specific antigen (Bet v G75) presented by autologous B cells. In the cell culture and during mixed lymphocyte culture (MLC) activation the cells were cultured in RPMI-1640 (Gibco Ltd, Paisley, UK) supplemented with 2 mm l-glutamine, 0·16% sodium bicarbonate, 10 000 U/ml benzylpenicillin, 10 000 μg/ml streptomycin and 10% fetal calf serum. In all experiments the cells were cultured in serum-free AIM-V medium (Gibco).

Small interfering RNA-mediated gene silencing

The expression of TSP-1 was suppressed using the human T-cell Nucleofector kit (Lonza, Köln, Germany) and a Nucleofector device (Amaxa biosystems, Köln, Germany) as previously described.⁴⁵ Briefly, 5 × 10⁶ cells were resuspended in 100 μl of nucleofector solution and transfected with 500 nm final concentration of small interfering RNA (siRNA) using protocol U14. The siRNA consisted of TSP-1 siRNA (human) (Alternative 1) (A: Sense: CCACGAUGAUGACAACGAUtt. Antisense: AUCGUUGUCAUCAUCGUGGtt. B: Sense: CGAGACGAUUGUAUGAAGAtt. Antisense: UCUUCAUACAAUCGUCUCGtt. C: Sense: GAAGAAGCGUAAAGACUAUtt. Antisense: AUAGUCUUUACGCUUCUUCtt) and control siRNA (sc-37007) from Santa Cruz Biotechnology and TSP-1 siRNASuppl (human): (Sense: GCAUGACCCUCGUCACAUAtt. Antisense: UAUGUGACGAGGGUCAUGCca.) from Applied Biosystems (Stockholm, Sweden). The degree of gene silencing and the influence of silencing on conjugate formation were determined 40 hr after introducing siRNA. Cytokine production was determined as specified in the figure legends.

Quantitative immunocytochemistry

The expression of various antigens was analysed in cells fixed in 2% paraformaldehyde at 4° for 20 min attached to glass slides coated with poly-l-lysine (10 μg/ml) at 4° overnight. Antigen expression was detected with monoclonal antibodies and a complex of biotinylated peroxidase and avidin (Vector Laboratories). For detection of intracellular antigens, cells were fixed in 2% paraformaldehyde and permeabilized by 0·1% saponin. The cells were examined in a Nikon Eclipse E1000M microscope (Nikon Corporation, Tokyo, Japan). The intensity of the immunocytochemical staining was quantified using the image processing and analysis program imagej.

Real-time PCR

Primer and probe sequences for the transcripts were: TSP1-F: 5′-TGC ACT GAG TGT CAC TGT CAG AA-3′, TSP1-R: CAT TGG AGC AGG GCA TGA T, TSP1-probe: 6-FAM-TA CCA TCT GCA AAA AGG TGT CCT GCC C-TAMRA, CRT-F: CAC GGA GAC TCA GAA TAC AAC ATC AT, CRT-R: TCA TCC TTG CAA CGG ATG TC, CRT-probe: 6-FAM-CA AGG GCA AGA ACG TGC TGA TCA ACA A-TAMRA, LRP-F: TGA CGA GGC CCC TGA GAT T, LRP-R: CAG GCA GTT ATG CTC GTT TGG, LRP-probe: 6-FAM-CA CAG AGT AAG GCC CAG CGA TGC C-TAMRA. ABL was used as an internal control gene. ABL-F: 5′-CGA AGG GAG GGT GTA CCA TTA C-3′; ABL-R:5′-CGT TGA ATG ATG ATG AAC CAA CTC-3′; ABLProbe: 5′-6-FAM-TTC TGA TGG CAA GCT CTA CGT CTC CTC C-TAMRA-3′. Quantification was performed with the ΔCt method using the formula 2^{(Ct ABL−Ct GENE)}. The PCR was performed and analysed on the ABI 7500 Sequence Detection System.

Biosynthetic labelling, immunoprecipitation and gel electrophoresis

Biosynthetic labelling of polypeptides synthesized during a 3-hr culture period was carried out with [³⁵S]methionine at 50 μCi/ml (Amersham Biosciences, Piscataway, NJ: 54·6 TBq/mmol, 533 MBq/ml) in methionine-free medium supplemented with 10% complete medium. Material from conditioned medium was dialysed and lyophilized and either applied directly in gel analysis or after immunoprecipitation with protein G agarose beads as described elsewhere (Roche, Basel, Switzerland). The dissolved material was mixed with 1 μg antibody at 4° overnight followed by centrifugation at 12 000 g at 4° for 20 seconds. Subsequently, the supernatants were discarded and the beads were resuspended in 1 ml washing buffer, and centrifuged again at 12 000 g at 4° for 20 seconds, the same procedure was repeated twice. After washing, 20 μl reducing buffer (2× , containing 0·15 g DTT in 5 ml JB buffer) was mixed with the beads and heated at 95° for 4 min and subsequently centrifuged at 7000 g for 1 min to spin down the beads. The proteins were separated on SDS–PAGE) gels. Proteins were transferred to the Hybond ECL membrane (Amersham) and detected using the BMC chemiluminescence blotting kit (Roche).

Western blotting

The samples were separated on SDS–PAGE gels and blotted onto a nitrocellulose membrane (Amersham), blocked overnight with PBS, 4% BSA and 0·5% Tween. Filters were washed with PBS with 1·5% BSA and incubated with antibodies. ECL Western blotting detection reagents were used for detection with Hyperfilm TM (Amersham).

Cell motility

Collagen type 1 was diluted in serum-free RPMI-1640 and H₂O (8/1/1), applied in plastic Petri dishes 1 ml/dish (30 mm; BD Biosciences) and allowed to polymerize at room temperature. A total of 1·0 × 10⁶ cells in AIM-V medium was added to each well and allowed to migrate for different times. Cytochalasin B, 10 μg/ml prevented migration into the collagen showing that it is an active cellular process. The cells were fixed in 2·5% glutaraldehyde or for immunocytochemistry in 2% paraformaldehyde and washed twice with PBS. Cell morphology and cell migration were evaluated in nine fixed positions in each well and at 50-μm intervals throughout the gel by the use of an inverted microscope (Nikon Eclipse TE300) and a digital depth meter (Heidenheim ND221). The results are given as per cent polarized cells, mean number of infiltrating cells/field (× 20 objective) per infiltration depth (50 μm for the first two layers immediately beneath the gel surface and 100 μm for other layers further down), as total number of infiltrating cells throughout the gel (× 20 objective) or as maximal infiltration depth. The infiltrating cells were identified in situ in the collagen gels using immunocytochemistry after fixation in paraformaldehyde. The transwell assay was performed using 48-well Boyden chambers. The lower wells were filled with RPMI-1640 containing 1 mg/ml BSA whereupon 8-μm nucleopore filters were placed in the chambers. The upper chamber was filled with 50 μl of 2 × 10⁶ cells/ml in AIMV. Following incubation for 1 hr the number of cells in the lower chamber was counted in triplicate.

Cell adhesion

To study cell adhesion, plastic Petri dishes (90 mm; Heger A/S, Rjukan, Norway) were coated with fibronectin (10 μg/ml), ICAM-1 (2 μg/ml) or BSA (10 μg/ml), and extensively washed before use. Cells (10 000/position) in AIM-V medium were incubated on the substrates for different times, fixed in 2·4% cold glutaraldehyde for 10 min or in 2% paraformaldehyde for 20 min for immunocytochemistry and unbound cells were removed by gentle aspiration. The number of adherent cells per microscope field (20 × objective) was counted. Cell adhesion was routinely, unless otherwise stated, evaluated in six fixed positions in triplicate.

Coupling assay

AF 24 T cells were mixed with Bet v G75-pulsed B cells, allowed to settle in a Petri dish and incubated at 37° for 45 min. After incubation the cells were fixed in 2% paraformaldehyde and analysed using immunocytochemistry. T cells forming conjugates were quantified as a percentage of all T cells in the sample.

Cytokine production

The production of interferon-γ, IL-2 and IL-4 at the single cell level was determined using Elispot kits from Mabtech (Stockholm, Sweden) according to the manufacturer's instructions.

Statistical analysis

The Mann–Whitney U-test was used to evaluate differences between groups except in the transwell assay where Student's t-test was used. Values of P < 0·05 were considered statistically significant.

Results

T-cell activation regulates the expression of TSP-1 and LRP1

To investigate the influence of antigen stimulation on TSP-1 and LRP1 expression we reasoned that TCR recognition of MHC–peptide complexes, which induces an exceptionally vigorous T-cell response, might provide useful information. Gel analysis of material from conditioned media of biosynthetically labelled cells showed that an allogeneic MLC released newly synthesized TSP-1 into the medium whereas the constitutive production by control cells was negligible (Fig.1a). CD3-depleted activated cells showed negligible TSP-1 synthesis, indicating that T cells were responsible for the production. Immunostaining of T cells from an allogeneic MLC after mixing cells from three donors to enhance stimulation demonstrated increased cell surface expression of LRP1 and TSP-1 after 4 hr (Fig.1b). After 24 hr the cells revealed a marked increase of TSP-1 on the cell surface and also increased LRP1 expression. After 72 hr the cells exhibited increased TSP-1 expression on the surface compared with control cells whereas the LRP1 expression was decreased. Control cells from the separate individuals used for MLC activation cultured at the same density did not exhibit any changes in the cell surface expression of TSP-1 and LRP1. The MHC–peptide complex-induced changes in the cell surface expression of TSP-1 and LRP1 were seen in both RA and RO-positive cells although RO-positive cells showed a stronger response (see Supporting information, Fig. S1). Quantitative RT-PCR further showed that cells from an allogeneic MLC exhibited a more than 60-fold up-regulation of the TSP-1 mRNA level (threefold changes are significant) and a concomitant down-regulation of LRP1 mRNA (10-fold) already after 4 hr in comparison with control cells that persisted after 24 hr (Fig.1c). As a control for the TSP-1 and LRP1 mRNA levels we examined calreticulin mRNA. In contrast to the changes in the expression of TSP-1 and LRP1 mRNA, the expression of calreticulin was unaffected after 4 hr. After 24 hr calreticulin mRNA increased threefold and then persisted at a slightly higher level than before stimulation. Control cells from the separate individuals used for MLC activation did not exhibit any changes in LRP1, calreticulin or TSP-1 mRNA expression (not shown).

Figure 1.

An antigen-induced thrombospondin-1 (TSP-1)/ low-density lipoprotein receptor-related protein 1 (LRP1) response. (a) Gel analysis (6% SDS–PAGE) showing mixed lymphocyte reaction-induced release of newly synthesized TSP-1 into the culture medium after 24 hr. The gels show unprecipitated (left) and immunoprecipitated (right) material from biosynthetically labelled (³⁵S methionine; 3-hr pulse) conditioned medium of co-cultured mononuclear cells from two individuals and a Western blot of unprecipitated material. Biosynthetic labelling of polypeptides synthesized during a 3-hr culture period was carried out with [³⁵S]methionine at 50 μCi/ml (Amersham: 54·6 TBq/mmol, 533 MBq/ml) in methionine-free medium supplemented with 10% complete medium. (b) Cell surface expression of TSP-1 and LRP1 in CD3 enriched lymphocytes after co-culture of mononuclear cells from three individuals for different times as revealed by quantitative immunocytochemistry. (c) Messenger RNA expression of TSP-1, LRP1 and calreticulin after co-culture of cells from three individuals as determined by quantitative RT-PCR. Results were normalized against sample ‘MLC 0 hr’. (d) Cell surface expression of TSP-1 and LRP1 in CD3 enriched lymphocytes, as revealed by quantitative immunocytochemistry, after co-culture of mononuclear cells from three individuals for 4 and 24 hr with and without abatacept (100 μg/ml) or infliximab (100 μg/ml) for 4 and 24 hr. (a) Representative experiments of three to nine independent experiments; (b) and (d) show mean values of three and (c) mean values of two independent experiments. *P < 0·05 versus the result at 0 hr; ***P < 0·001 versus the result at 0 hr; ns, not significant.

The influence of co-stimulation on the MHC–peptide complex-induced TSP-1/LRP1 response was investigated using abatacept, a CTLA4 molecule that effectively inhibits T-cell activation in different model systems including MHC-induced responses.⁴⁶ Abatacept contains a high-affinity binding site for B7, and works by binding to the B7 protein on antigen-presenting cells. This prevents them from delivering a CD28 co-stimulatory signal, so preventing the full activation of T cells. As a control molecule we applied infliximab, a chimeric monoclonal antibody (30% mouse) against tumour necrosis factor, the beneficial action of which has been proposed to reflect reduction of production of other pro-inflammatory cytokines, decreased production of acute phase reactants and increased apoptosis of cells in the synovium. Abatacept virtually abolished the increased cell surface expression of TSP-1 in MLC-activated T cells as determined after 4 and 24 hr (Fig.1d). Abatacept also inhibited the increased cell surface expression of LRP1 after 4 and 24 hr. In contrast, abatacept did not inhibit the baseline TSP-1 and LRP1 expression in controls. Infliximab did not affect the MHC-induced TSP-1/LRP1 response or the baseline TSP-1 and LRP1 expression in controls. The results in Fig.1(d) indicated that the MHC–peptide complexes-induced TSP-1/LRP1 response requires CD28 co-stimulation.

Ligation of CD28 induces cell surface expression of LRP1 and enhances anti-CD3-induced TSP-1 expression

We next examined the influence of ligation of the TCR/CD3 complex and CD28 using antibodies on the T-cell expression of TSP-1 and LRP1. Ligation of CD3 with insoluble antibody under serum-free conditions to exclude any interference of exogenous proteins and peptides increased the cell surface expression of TSP-1 in comparison with control cells but did not affect LRP1 expression (Fig.2a). This TSP-1 increase was seen already after 15 min (not shown). Ligation of CD28 alone increased the cell surface expression of LRP1 significantly but did not increase TSP-1 expression (Fig.2a). Co-ligation of CD3 and CD28 enhanced the cell surface expression of TSP-1 markedly in comparison with anti-CD3 alone showing that co-stimulation of CD28 enhances TCR/CD3-induced TSP-1 expression (Fig.2a).

Figure 2.

Ligation of CD28 and T-cell receptor (TCR)/CD3 have distinct effects on the cell surface expression of low-density lipoprotein receptor-related protein 1 (LRP1) and thrombospondin-1 (TSP-1). (a) Quantitative immunocytochemistry showing the expression of LRP1 and TSP-1 after incubation of non-activated cells on an uncoated plastic surface or surfaces coated with anti-CD28, anti-CD3 or anti-CD3 plus anti-CD28, respectively for 4 and 24 hr. Goat anti-mouse IgG was coated at a concentration of 2 μg/ml, anti-CD3 and anti-CD28 at a concentration of 50 ng/ml. Control IgG generally yielded values < 5 arbitrary units in intensity measurements. (b) Influence of incubation of cells on anti-CD3, anti-CD28 and anti-CD3 plus anti-CD28 on mRNA expression of TSP-1, LRP1 and calreticulin as determined by RT-PCR. Results were normalized against the ‘Control’ samples. Threefold changes are significant. The results in (a) and (b) show mean values of three and two independent experiments, respectively. *P < 0·05 versus control; **P < 0·01 versus control; ***P < 0·001 versus control; ns, not significant.

Quantitative PCR showed that anti-CD3 stimulation was permissive for basal TSP-1 and LRP1 mRNA levels after 4 hr but suppressed TSP-1 and LRP1 mRNA levels after 24 hr (Fig.2b). The specificity of this anti-CD3 effect is supported by the fact that anti-CD28 did not affect TSP-1 and LRP1 mRNA levels (Fig.2b), whereas IL-2 up-regulates TSP-1 mRNA.²⁸ The finding that anti-CD3 reduces the TSP-1 mRNA level confirms our previous finding using a T-cell clone.²⁸ Importantly, the results demonstrate that the TCR exerts positive as well as negative control of TSP-1 and LRP1 expression and that CD28 co-stimulation is necessary for a positive effect of TCR/CD3 stimulation.

Antigen stimulation induces a characteristic TSP-1/LRP1 response

In the light of the evidence that TCR/CD3 ligation caused a similar but weaker TSP-1/LRP1 response than MHC–peptide complexes (Figs2) it was important to compare the influence of antigen and anti-CD3 in another test system. Therefore, we examined the influence on TSP-1 and LRP1 expression of stimulation of a CD4-positive birch-allergen-specific T-cell clone (AF 24) with anti-CD3 and antigen presented by autologous B7-positive B cells over several days (Fig.3). Presence of IL-2, which supports TSP-1 synthesis,²⁸ was necessary to promote the growth of AF 24 but would not be expected to affect cells stimulated by antigen different from cells stimulated with anti-CD3 used as a comparison. The antigen-presenting cells were removed after activation before the analysis of TSP-1 and LRP1 expression and so did not contribute to the results. Birch antigen stimulation induced a marked increase in TSP-1 expression in comparison with anti-CD3. Stimulation with birch antigen or anti-CD3 decreased LRP1 expression on day 1 and day 3 accompanied by increased expression on day 6 (suggestive) and day 9 (significant). Abatacept also abrogated the increased TSP-1 expression induced by antigen (see legend to Fig.3) The marked stimulatory effect of specific antigen on TSP-1 expression together with the decreased LRP1 expression seems to confirm the findings in Fig.1 that collaboration of the TCR and CD28 induces a characteristic TSP-1/LRP1 response.

Figure 3.

Influence of specific antigen and anti-CD3 on the cell surface expression of thrombospondin-1 (TSP-1) and low-density lipoprotein receptor-related protein 1 (LRP1) in the birch allergen-specific T-cell clone AF 24. The expression of TSP-1 and LRP1 in AF 24 cells stimulated with specific antigen or anti-CD3 for different times was determined by quantitative immunocytochemistry. The results show one representative experiment of multiple independent experiments. The intensity of TSP-1 staining in the absence of abatacept in a separate representative experiment was 56 ± 36 and in the presence of abatacept from the beginning of the stimulation 16 ± 12 (P < 0·01). **P < 0·01 versus unstimulated cells; ***P < 0·001 versus unstimulated cells.

The TCR regulates T-cell motility in collaboration with CD28

T-cell motility was examined under serum-free conditions to exclude any interference of exogenous proteins and peptides through determination of polarized shape and migration in three-dimensional (3D) type 1 collagen matrices. This is a well established model for analysis of lymphocyte motility and infiltration of tissues where the cells move independent of adhesive interactions with the substrate.^47–50 Collagen substrata maintain a significantly higher motility level in T cells than plastic and the cells within 3D collagen develop a more persistent bipolar locomotor morphology.⁴⁹ In contrast, the motile behaviour of adherent cells is characterized by cytoplasmic spreading and formation of lamellipodia.⁵¹

Lymphocytes activated by allogeneic MHC–peptide complexes for 15 hr and identified as CD3-positive showed reduced migration into a collagen matrix in comparison with non-activated control cells (Fig.4a,b). However, MHC–peptide complex-activated cells incubated with abatacept exhibited an even lower level of motility, as demonstrated by the fact that virtually all cells were non-polarized and spherical without pseudopodia or other active cell edges and had lost the capacity to migrate (Fig.4a–c). In contrast, abatacept did not inhibit development of a polarized shape and migration into collagen in non-stimulated control cells. The conclusion that motility was down-regulated but not arrested in MHC–peptide complex-activated cells was supported by the finding that receptor-associated protein (RAP), an endoplasmic reticulum-resident protein that prevents ligand binding to LRP1⁵² and inhibits LRP1- and TSP-1-dependent T-cell motility,²⁸ abrogated development of a polarized locomotor morphology in MHC–peptide complex-activated cells (Fig.4d). Additional support for the significance of the motility in MHC–peptide complex-activated cells was provided by the finding that hep1, a peptide mimetic of the N-terminal calreticulin binding site in TSP-1, abrogated development of a polarized locomotor morphology (Fig.4d). Non-activated control cells from the individual donors exhibited variable morphology with motile forms in different stages of polarization and migrated in the collagen. These results indicated that CD28 co-stimulation is necessary to maintain a basic level of T-cell motility during antigen stimulation and that TCR stimulation alone suppresses motility. We next examined whether transport of LRP1 to the cell surface is critical for motility in MHC–peptide complex-activated T cells. The dynamin inhibitor dynasore, which inhibits transport of LRP1 to the cell surface (see legend to Fig.4e) but does not inhibit the expression of TSP-1, was found to abrogate formation of active cell edges and development of a polarized locomotor morphology (Fig.4e) indicating that CD28-mediated transport of LRP1 to the cell surface maintains motility.

Figure 4.

CD28 co-stimulation is critical for motility in activated T cells. (a–c) Migration of lymphocytes into a collagen gel (a, b) and development of locomotor morphology (c) were examined after culture with allogeneic cells in the absence and presence of abatacept and infliximab for 15 hr. The cells were identified as CD3 positive in situ within the gel using immunocytochemistry. The maximal depth of migration (a), the total number of cells at different levels throughout the gel (b), and % cells with a polarized morphology (c) were determined. (d) Influence of RAP 50 (μg/ml) and hep1 (50 μM), inhibitors of low-density lipoprotein receptor-related protein 1 (LRP1) -dependent motility, on lymphocyte morphology. (e) Influence of inhibition of cell surface expression of LRP1 using dynasore (80 μm) on lymphocyte morphology. Colchicine (10–6 m) was used as a control. Dynasore reduced the cell surface expression of LRP1 from 45 to 5 arbitrary units as determined using immunocytochemistry. The figures show mean values of three (a, b) and two (c–e) representative experiments of three to four independent experiments. ***P < 0·001 versus control cells.

In the light of the requirement of co-stimulation for a characteristic TSP-1/LRP1 response (Figs1 and 3) and the more potent effect of birch antigen than anti-CD3 (Fig.3) we compared the influence of anti-CD3 and antigen stimulation on T-cell motility using AF 24. On day 1 and 3 after the beginning of stimulation with birch antigen or anti-CD3 migration of AF 24 in a transwell assay and in collagen decreased (Fig.5a,b). However, two to four times as many cells migrated on day 3 after the beginning of antigen stimulation than on day 1 although the level was relatively low and cannot be seen in the figure. Migration subsequently increased markedly during a 2-week period whereas the increase with anti-CD3 was less pronounced, supporting the finding in Fig.4 that co-stimulation is important for motility. Abatacept also abrogated the increased motility during the activation (see legend to Fig.5). It is important to notice that migration in the Transwell assay was pronounced on uncoated filters (Fig.5a,b). Furthermore, coating of the lower filter surface simply delayed migration into the lower well (not shown). A role of TSP-1 and LRP1 was supported by the finding that RAP and hep1, which inhibit LRP1- and TSP-1-dependent T-cell motility,²⁸ inhibited migration (Fig.5c,d). A comparison of adhesion to fibronectin and ICAM-1 (not shown here but used in adhesion experiments with AF 24 in Fig.6) at different times after the beginning of stimulation showed that adhesion was markedly higher on day 1 than on day 7 (Fig.5e). This difference is consistent with the finding that highly motile T cells are non-adhesive.^28,53

Figure 5.

Influence of specific antigen and anti-CD3 on motility and adhesion of the birch allergen-specific T-cell clone AF 24. (a, b) AF 24 cells were stimulated with specific antigen or anti-CD3 for different times before analysis of migration using a transwell assay as determined after 1 hr (a) or migration into a collagen matrix (b) after a 30-min period. (b) Mean values of total number of cells throughout the gel in a high-power field. The number of transmigrated cells in a separate experiment using cells stimulated with antigen for 7 days in the absence of abatacept was 5680 ± 1068 and in the presence of abatacept 1973 ± 388. (c, d) Inhibition of migration of AF 24 into a collagen gel by RAP (50 μg/ml) and hep1 (50 μM), inhibitors of low-density lipoprotein receptor-related protein 1 (LRP1) -dependent T-cell motility as previously demonstrated.²⁸ (e) Adhesion of AF 24 to fibronectin (10 μg/ml) as determined at different times in parallel with the experiment in (a, b). The figures show mean values of three representative experiments of three to four independent experiments. **P < 0·01 versus day 0; ***P < 0·001 versus control cells.

Figure 6.

Thrombospondin-1 (TSP-1) and the TSP-1 binding site in calreticulin enhance T-cell conjugate formation through CD47. (a) Silencing of TSP-1 inhibits the formation of conjugates between AF 24 and antigen-presenting cells. The figure shows formation of conjugates of AF 24 T cells with antigen-presenting HLA-identical B cells after transfection with control small interfering RNA (siRNA) and TSP-1 siRNA (alternative 1 see Material and methods). (b) Expression of TSP-1 in permeabilized cells after transfection with scrambled control siRNA or TSP-1 siRNA. The cells were incubated with brefeldin A for 30 min before detection using quantitative immunocytochemistry. (c) The CD47 (4N1K) and calreticulin (hepI) binding sites of TSP-1 at a concentration of 50 μM inhibit conjugate formation of AF 24 T cells with antigen-presenting HLA-identical B cells, whereas a peptide mimetic of the TSP-1 binding site in calreticulin (CRT19–36) (50 μM) stimulates conjugate formation. It is also evident that scrambled control peptides of 4N1K, hep1 and CRT19–36 at a concentration of 50 μM did not affect conjugate formation. (d) The stimulation of conjugates of AF 24 with antigen-presenting HLA-identical B cells by CRT19–36 depends on LFA1, as indicated by the reduced number of conjugates by the presence of an antibody to CD18. (e) Adhesion of AF 24 to intercellular adhesion molecule 1 (ICAM-1; 2 μg/ml) is inhibited in the presence of hep1 (50 μM) and stimulated by CRT19–36 (50 μM) as determined after 30 min. The figures show representative experiments of two to five independent experiments. **P < 0·01 versus control cells.

TSP-1 regulates T-cell adhesion to antigen-presenting cells

The finding that newly activated T cells exhibited increased adhesion (Fig.5e) associated with up-regulation of TSP-1 expression (Figs3) pointed to the possibility that TSP-1 affected T-cell conjugation to antigen-presenting cells. To clarify this we examined whether knockdown of TSP-1 affected the number of conjugates formed during activation of AF 24. Knockdown of TSP-1 reduced conjugate formation between AF 24 and antigen-presenting B cells significantly after 1, 4, 12 and 24 hr (Fig.6a,b). In contrast, conjugate formation in control cells was markedly enhanced after 12 and 24 hr, which correlates with the up-regulation of TSP-1 expression. The validity of the siRNA silencing experiments was further examined using a separate TSP-1 siRNA (TSP-1 siRNASuppl in Materials and methods; not shown). This yielded essentially the same results as the TSP-1 siRNA used in the experiments shown in Fig.6(a), which argues against the possibility that off-target effects were responsible.

Given the evidence that TSP-1 stimulated conjugate formation (Fig.6a,b) it was important to examine the mechanism of this stimulatory effect. We reasoned that the stimulatory effect of TSP-1 on T-cell adhesion to antigen-presenting cells may be accounted for by enhancement of integrin-mediated adhesion similar to the previous finding that TSP-1 enhances T-cell adhesion to fibronectin and collagen type IV through interaction of its C-terminal domain with CD47 in response to a stimulus by the LRP1-associated protein calreticulin at the N-terminal domain.³⁶ To examine this possibility we added peptides mimetic of the TSP-1 binding sites for CD47 and calreticulin. The CD47 binding site in TSP-1 (4N1K) was found to inhibit conjugate formation whereas a scrambled control peptide did not affect conjugate formation (Fig.6c). The calreticulin-binding site in TSP-1, hep1, also inhibited conjugate formation, whereas a peptide mimetic of the TSP-1 binding site in calreticulin, CRT19-36, had a marked stimulatory effect on conjugate formation that was inhibitable by an anti-integrin β₂ antibody (Fig.6d). In addition, hep1 inhibited adhesion to ICAM-1 whereas CRT19-36 stimulated adhesion (Fig.6e). The results in Fig.6(c–e) indicated that the calreticulin–LRP1 complex delivers a triggering signal for conjugate formation through the N-terminal domain of TSP-1 and via the C-terminal domain of TSP-1 to CD47.

TSP-1, T-cell activation and motility

To investigate the influence of endogenous TSP-1 on T-cell activation its expression in blood T cells was silenced using siRNA and the cells were activated using inactivated allogeneic stimulator cells as previously described.⁵⁴ Transfection with TSP-1 siRNA caused a significant reduction of interferon-γ and IL-2 production compared with cells transfected with scrambled control siRNA (Fig.7a) (see Supporting information, Fig. S2). Transfection with TSP-1 siRNA also inhibited anti-CD3-induced interferon-γ production in human blood T cells whereas control siRNA was not inhibitory (Fig.7b) (Fig. S2). To further investigate the influence of TSP-1 on T-cell activation Th2 polarized AF 24 T cells⁵⁵ were transfected with TSP-1 siRNA and the cells were activated using antigen-pulsed B cells. Transfection with TSP-1 siRNA was found to inhibit IL-4 production, whereas scrambled control siRNA did not inhibit IL-4 production (Fig.7c) (Fig. S2). These results indicated that Th1 as well as Th2 responses depend on TSP-1 and the antigen-induced TSP-1/LRP1 response (Fig.1).

Figure 7.

Silencing of thrombospondin-1 (TSP-1), T helper type 1 (Th1) and Th2 responses and motility. (a) Silencing of TSP-1 inhibits interferon-γ (IFN-γ) and interleukin-2 (IL-2) production induced by co-culture of blood T cells with inactivated allogeneic stimulator cells as determined by an Elispot assay after 144 hr. (b) Silencing of TSP-1 inhibits anti-CD3-induced IFN-γ production as determined by an Elispot assay after 40 hr. (c) Silencing of TSP-1 inhibits IL-4 production in Th2 polarized AF 24 as determined by an Elispot assay after 40 hr. (d) The influence of antibodies to IL-2 and IL-4 at a concentration of 2 μg/ml on motility in blood T cells co-cultured with inactivated allogeneic stimulator cells (Th1) for 144 hr and in Th2 polarized AF24 cultured for 120 hr. For determination of motility the cells were pre-incubated with antibodies for 60 min and subsequently allowed to migrate into a collagen gel for 30 min. (e–f) The influence of antibodies to IL-2 and IL-4 at a concentration of 2 μg/ml on the cell surface expression of TSP-1 and low-density lipoprotein receptor-related protein 1 (LRP1) in blood T cells co-cultured with inactivated allogeneic stimulator cells (Th1) for 144 hr (e) and in Th2 polarized AF 24 for 120 hr (f) as revealed by quantitative immunocytochemistry. Before determination of TSP-1 and LRP1 expression the cells were pre-incubated with antibodies for 90 min. The figures show mean values of three independent experiments.

Given our earlier demonstration that IL-2 stimulated TSP-1 expression and T-cell motility in non-activated cells whereas IL-4 inhibited motility and enhanced LRP1 expression²⁸ we next examined the possible influence of ongoing Th1 and Th2 responses on T-cell motility. Figure7(d,e) illustrates that an anti-IL-2 antibody inhibited motility of MHC–peptide complexes-activated Th1 cells as shown by migration into collagen matrices and also reduced TSP-1 expression. In contrast, an anti-IL-4 antibody did not affect motility and TSP-1 expression of MLC-activated cells. Figure7(d, f) further shows that Th2 polarized AF 24 cells exhibited increased motility and decreased LRP1 expression in the presence of an anti-IL-4 antibody, whereas an anti-IL-2 antibody did not change the motility level. Nor did addition of IL-2 increase the motility of AF 24 (see Supporting information, Fig. S3). The results in Fig.7 indicate that the IL-2 response of Th1 cells delivers a motogenic stimulus associated with an effect on TSP-1 whereas the IL-4 response of Th2 cells delivers IL-2-insensitive inhibition of motility associated with an effect on LRP1.

Discussion

The present findings indicate that the TCR exerts co-stimulation-dependent control of T-cell motility, interaction with antigen-presenting cells and activation by directing the synthesis and cell surface expression of TSP-1 and LRP1 as depicted in Fig.8. Antigen stimulation up-regulated the synthesis and cell surface expression of TSP-1 whereas the cell surface expression of LRP1 transiently increased and subsequently decreased. The effects of blocking co-stimulation indicated that TCR ligation alone arrests motility through suppression of LRP1 and TSP-1 expression whereas co-stimulation maintains a reduced but significant motility while enhancing the cell surface expression of LRP1. CD28 co-stimulation is probably instrumental in T-cell activation through this effect on LRP1 and LRP1-dependent up-regulation of TSP-1 synthesis. Suppression of LRP1 and TSP-1 expression may explain the antigen-specific hyporesponsiveness induced by TCR ligation alone.⁵⁶

Figure 8.

T-cell receptor (TCR) -induced co-stimulation-dependent regulation of T-cell motility, adhesive contacts and activation through thrombospondin-1 (TSP-1) and its receptors low-density lipoprotein receptor-related protein 1 (LRP1), calreticulin and CD47. (a) Co-stimulation of the TCR and CD28 regulates T-cell motility and capacity to form conjugates with antigen-presenting cells through differential effects on the synthesis and cell surface expression of TSP-1 and LRP1, so eliciting a powerful TSP-1 response while down-regulating LRP1. Co-stimulation-dependent up-regulation of TSP-1 synthesis induces T-cell conjugation to antigen-presenting cells through its C-terminal domain via CD47 in response to N-terminal triggering by the LRP1 co-receptor calreticulin. TSP-1 induces CD47-mediated stabilization of the high-affinity state of the integrin-ligand binding site. (b) Ligation of CD28 per se enhances cell surface expression of LRP1 and development of a polarized cell shape, whereas ligation of the TCR alone inhibits motility through inhibition of the expression of TSP-1 and LRP1. CD28 ligation therefore counterbalances the inhibitory effect of TCR ligation alone on T-cell motility.

TSP-1 seems to be important for induction of Th1 and Th2 cytokines (Fig.7), which is consistent with the evidence that TSP-1 and LRP1 regulate T-cell activation by controlling motility and contact with antigen-presenting cells. The antigen-induced TSP-1/LRP1 response peaks within 24 hr (Fig.1) whereas IL-2 and IL-4 production peaks after 144 and 48 hr, respectively.^54,57 Unlike activation through the TCR and CD28, IL-2 does not induce a prominent TSP-1 peak but causes a persistent less pronounced up-regulation of TSP-1 synthesis.²⁸ The TCR- and CD28-induced effects on TSP-1 and LRP1 expression are also distinguishable from the effects of CXCL12/SDF-1 and integrin ligands, which induce rapid up-regulation of the cell surface expression of both LRP1 and TSP-1.³⁷

Ligation of the TCR and CD28 elicits a TSP-1/LRP1 response that down-regulates but does not arrest motility. This modulated motility may serve to allow multiple serial contacts with antigen-presenting cells while stimulating prolonged contact and activation, which is consistent with in vivo imaging data of T-cell behaviour.^19–24 This behaviour may be explained by the integrated regulation of motility and adhesion through TSP-1 and its receptors that make T cells prioritize movement before permanent adhesion.^28,38 Antigen stimulation probably changes the balance within this regulatory cell surface cascade from prioritization of a normally dominating motogenic pathway, driven by LRP1-dependent processing of TSP-1, to a normally down-regulated pro-adhesive pro-proliferative LRP1–calreticulin–TSP-1–CD47–integrin pathway driven by intact TSP-1.^28,38 Intact TSP-1 probably promotes T-cell adhesion to ICAM-1 and conjugate formation with antigen-presenting cells through interaction of its C-terminal domain with CD47 as a consequence of calreticulin interaction with the N-terminal domain (Fig.6). We propose that the association of CD47 to LFA-1 and VLA-4⁵⁸ promotes adhesion through TSP-1-induced CD47-mediated stabilization of the high-affinity state of the integrin-ligand binding site.

The present findings suggest that CD28, like the other members of the same family, CTLA-4 and programmed death-1, also promotes motility although these receptors are operative at different stages of T-cell activation and in different settings.^10–12,19 In support of a vital role of CD28 for T-cell motility, CD28 knockout mice lack germinal centres, indicating that T cells fail to migrate to sites of interaction with B cells,⁵⁹ and CD28 regulates in vivo migration of memory T cells.¹ Failure to mount a CD28-dependent TSP-1/LRP1 response together with the fact that TSP-1 can induce regulatory T cells may also explain why mice deficient in CD28 develop autoimmunity and have reduced numbers of regulatory T cells.^60,61

T-cell motility, adhesion and activation therefore appear to be tightly coupled and inter-regulated through cell surface interactions in cis between TSP-1, LRP1, calreticulin, CD47 and integrins. The basal interaction between the C-terminal domain of TSP-1 and CD47 is probably motogenic through stimulation of cell surface expression of LRP1 and LRP1-dependent processing of TSP-1 because an antibody to CD47 mimics this motogenic effect.^28,36 In contrast, adhesion signalling by intact TSP-1 is conditional and dependent on N-terminal calreticulin triggering.^28,36 The demonstration that anti-CD47 antibodies block allogeneic mixed lymphocyte reactions⁶² may reflect inhibition of the pro-adhesive pro-proliferative LRP1–calreticulin–TSP-1–CD47–integrin pathway and is consistent with the motogenic effect of such antibodies.^28,36 The present and earlier evidence for an integrated regulation of adhesion, motility and activation through LRP1, calreticulin, TSP-1 and CD47^28,38 may explain why TSP-1 on one hand can inhibit TCR-induced T-cell activation through CD47^63,64 but on the other hand can stimulate proliferative responses through CD47.^65,66

Antibody blocking experiments indicated that endogenous IL-2 and IL-4 differentially regulate the motility of polarized Th1 and Th2 cells (Fig.7). Interleukin-2 therefore enhances T-cell motility during a Th1 response and increases TSP-1 expression, whereas IL-4 weakens motility during a Th2 response and enhances LRP1 expression. This supports the previous conclusions with exogenous IL-2 and IL-4.²⁸ However, although IL-2 and IL-4 counterbalanced each other in non-stimulated cells²⁸ the effect of IL-4 seemed to dominate during an ongoing Th2 response (Fig.7). In the light of the evidence that the Th2 response is more dependent on CD28 co-stimulation than Th1^67–69 it is interesting that CD28 ligation as well as the Th2 cytokine IL-4 stimulated LRP1 expression. Hence, besides the established Th2 stimuli, inhibition of the Th1 inducing cytokine IL-12, and high-affinity peptide stimulation,^70,71 sustained LRP1 expression may also favour a Th2 response.

The association of the antigen-induced pro-adhesive pro-proliferative LRP1–calreticulin–TSP-1–CD47–integrin pathway and the antagonistic motogenic LRP1–TSP-1 pathway²⁸ endows the process of T-cell activation with a counterbalancing mechanism that may prevent unwanted activation. It is therefore conceivable that activation by low amounts or transient expression of MHC-complexed antigen will be prevented due to dominance of the motogenic TSP-1–LRP1 pathway whereas higher amounts of antigen or persistent antigen expression will overcome the motogenic pathway and activate the pro-adhesive pro-proliferative LRP1–calreticulin–TSP-1–CD4–integrin pathway. The motogenic pathway may therefore represent a tolerance mechanism.

The TCR- and CD28-dependent regulation of T-cell motility and contact with antigen-presenting cells through LRP1 and TSP-1 (Fig.8) may account for the fact that induction of protective immune responses, tolerance and prevention of autoimmunity collectively seem to depend on T-cell motility and adhesion.^19–24 The conclusion that TCR- and CD28-dependent effects on LRP1 and TSP-1 expression regulate immune responses is further underscored by the fact that TSP-1 regulates inflammatory responses.^39–44 This suggests that the T-cell response to antigen is designed to prevent inflammatory adverse side effects through up-regulation of TSP-1 synthesis. The present findings provide a conceptual background for better understanding of normal immune regulation and dysregulated T-cell functions in autoimmune and allergic diseases as well as for development of better T-cell therapy for treatment of immunological diseases.

Disclosures

None.

Supporting Information

Additional Supporting Information may be found in the online version of this article: