Targeting dendritic cells with CD44 monoclonal antibodies selectively inhibits the proliferation of naive CD4⁺ T-helper cells by induction of FAS-independent T-cell apoptosis

Abstract

CD44 is a multifunctional adhesion molecule that has been shown to be a costimulatory factor for T-cell activation in vitro and in vivo. The aim of the present study was to expand these findings by characterizing the role of CD44 during dendritic cell (DC) antigen presentation to naive, resting T cells. Certain monoclonal antibodies (mAbs) directed against all CD44 isoforms (pan CD44), or against the epitope encoded by the alternatively spliced exon v4 (CD44v4), dose-dependently inhibited the capacity of murine DC to induce proliferation of naive alloreactive T cells. Preincubation of the T cells or DC with these CD44 mAbs revealed that the effect was dependent upon mAb binding to DC, but not to T cells. DC treated with anti-pan CD44 and anti-CD44v4 mAbs induced CD4⁺ T-cell apoptosis, as shown by annexin V staining and TdT-mediated biotin–dUTP nick-end labelling (TUNEL) assays. However, CD4⁺ T-cell apoptosis was not dependent on the Fas/Fas ligand (Fas/FasL) system, as DC from FasL-deficient (Gld) mice and T cells from Fas-deficient (Lpr) mice were still susceptible to apoptosis induced by CD44-treated DC. To investigate whether CD44 treatment of DC affects early T-cell/DC interactions, time-lapse video microscopy was performed using peptide-specific T cells from T-cell receptor (TCR) transgenic mice. Interestingly, calcium signalling in CD4⁺ T cells was significantly diminished following interaction with CD44 mAb-treated DC, but this was not observed in CD8⁺ T cells. Taken together, we found that perturbation of distinct epitopes of CD44 on DC interfere with early Ca²⁺ signalling events during the activation of CD4⁺ T cells, resulting in T-cell apoptosis.

Introduction

Dendritic cells (DC) shape T-cell immune responses by expressing both activating and inhibiting factors. To facilitate this, DC express high amounts of stimulatory major histocompatibility (MHC) class I and MHC II molecules in combination with costimulatory molecules of the B7 family, enabling them to induce CD4⁺ and CD8⁺ T-cell responses.¹ DC are also able to terminate a T-cell response by the induction of activation-induced cell death (AICD) via the Fas/Fas ligand (Fas/FasL) system.² In addition to professional costimulatory receptors, such as members of the B7 family, other molecules that have an adhesive function during DC–T-cell interactions have been found to modulate T-cell responses. CD44 is one such molecule.³

The transmembrane protein CD44 is involved in a variety of immunological functions, cell–cell interactions, cell adhesion to the extracellular matrix, as well as in tumour metastasis and progression.^4–6 Differential utilization of 10 variably spliced exons, as well as variations in N- and O-glycosylation, generates multiple isoforms of this molecule, with molecular mass ranging from 85 to 230 kDa.⁶ CD44 is highly expressed in various lymphoid and non-lymphoid tissues, and contributes significantly to the recruitment of T cells and DC to sites of inflammation and to their migration into lymphatic tissues.^5–7 However, there is also substantial evidence to suggest that CD44 acts as a signalling receptor capable of directly modulating immune responses and cell survival.^8–10 For example, anti-CD44 monoclonal antibodies (mAbs) have been shown to inhibit superantigen-induced T-cell activation and proliferation by inducing apoptosis.¹¹ Similarly, mAb ligation of CD44 induces apoptotic cell death in fibroblasts.¹² On the other hand, cross-linking of CD44 by anti-CD44 mAbs provides costimulatory signals for anti-CD3 and anti-CD2 mAb-induced T-cell activation.^11,13

In vivo, injection of anti-CD44 mAb has a potent anti-inflammatory effect.^14–16 In particular, administration of the anti-CD44 mAb, IM7, was found to prevent cutaneous delayed-type hypersensitivity responses.^7,14 This same anti-CD44 mAb was found to prevent the progression of ongoing collagen-induced arthritis and blocked leucocyte infiltration and tissue swelling.^15,16 In view of this, at first sight it is surprising that CD44-deficient mice show no differences in T- and B-cell activation.^17,18 Moreover, T cells from these mice were recently described as being more resistant to activation-induced cell death (AICD), resulting in an enhanced development of concanavalin A (Con A)-induced hepatitis.¹⁹ Together, these data suggest that there are explicit differences between the total absence of CD44 and the blocking of distinct CD44 epitopes using mAbs.

To obtain more detailed information about the function of distinct CD44 epitopes during T-cell activation, it would be necessary to use cellular systems involving both T cells and antigen-presenting cells (APC). We therefore used a co-culture system of DC and T cells of either alloreactive or peptide-specific T-cell receptor (TCR) transgenic origins to differentiate between CD4⁺ and CD8⁺ T-cell responses. Here we show that the specific blockade of distinct CD44 epitopes by anti-CD44 mAbs dose-dependently inhibited the T-cell proliferation of CD4⁺ cells, but not CD8⁺ T-cell responses. This effect was only observed when the mAbs were bound to DC. Furthermore, we found that the inhibition of T-cell proliferation by the treatment of DC with CD44 mAb was the result of interference with early signalling events in CD4⁺ T cells.

Materials and methods

Antibodies and reagents

Hybridomas producing the rat anti-mouse pan CD44 mAbs KM201 (IgG1) and IM7 (IgG2b) were procured from the American Type Culture Collection (ATCC; Rockville, MD). Antibody was purified from conditioned medium of these hybridomas and from rat anti-mouse CD44v4 (IgG1) (Clone 10D1) and CD44v6 (IgG1) (9A4) hybridomas, as previously described.⁷ Fluorescein isothiocyanate (FITC)-conjugated affinity-purified F(ab′)₂ fragment goat anti-rat (H+L) mAbs were purchased from Dianova (Hamburg, Germany) and rat IgG isotype-control antibodies from Becton-Dickinson/Pharmingen (Heidelberg, Germany). [³H]Thymidine ([³H]TdR) was procured from Amersham (Braunschweig, Germany). Interleukin-4 (IL-4) and granulocyte–macrophage colony-stimulating factor (GM-CSF) were purchased from Promo-Cell (Heidelberg, Germany). FITC-labelled pan CD44 mAb, KM201, was purchased from SouthernBiotech (Birmingham, AL). RPMI-1640 (Gibco, Eggenstein, Germany) was supplemented with 5% heat-inactivated fetal calf serum (FCS), 2 mm l-glutamine, 25 mm HEPES buffer and 50 mg/ml penicillin–streptomycin (Gibco).

Generation of F(ab′)₂ fragments

IM7 mAb was digested into F(ab′)₂ fragments by using immobilized papain and purification on an immobilized protein A column, according to the instructions of the manufacturer (Pierce-Perbioscience, Bonn, Germany). The F(ab′)₂ fragments eluted from the column were adjusted to a concentration of 1 mg/ml for further use.

Experimental animals

Female mice (6–8 weeks of age), of the following strains, were used: C57BL/6; BALB/c (Charles River, Sulzfeld, Germany); C57BL/6-Gld (which are defective for Fas-L);²⁰ BALB/c-Lpr (which are defective for Fas);²¹ DO 11.10 mice on a BALB/c background, which carry a transgenic TCR for amino acids 323–339 of ovalbumin peptide (ISQAVHAAHAEINEAGR);²² and P14 TCR transgenic mice (line 318), which express a Vα2/Vβ8 TCR specific for amino acids 33–41 (GP33 peptide, KAVYNFATM) of the lymphocytic-choriomeningitis virus (LCMV) glycoprotein in association with the H-2D^b molecule.²³ Mice were housed and bred in the specific pathogen-free facility of the Max-Planck Institute for Immunobiology (Freiburg, Germany). The CD44-deficient mice¹⁷ were housed in the animal facility of the Institute for Genetics, Forschungszentrum Karlsruhe (Karlsruhe, Germany).

DC preparation

Bone marrow-derived DC were prepared from the femur and tibia of 6–8-week-old female C57BL/6 mice using the protocol of Inaba et al.,²⁴ with minor modifications. A total of 1 × 10⁶ cells/ml were resuspended in RPMI-1640 containing 40 ng/ml murine GM–CSF and 100 ng/ml IL-4. Cells were cultured in 1-ml aliquots in 24-well flat-bottomed culture plates and fed on days 3 and 5 of culture by replacing half of the medium per well with cytokine-supplemented RPMI-1640. On day 6, loosely adherent cells were harvested, washed in phosphate-buffered saline (PBS) and separated using a Percoll^® (Pharmacia, Freiburg, Germany) density gradient of 1·04–1·08 g/ml. After centrifuging the gradient for 20 min at 4° and 700 g, DC from the interphase were collected and washed extensively in PBS before further use.

T-cell isolation/T-cell proliferation assay

Spleens from 6–8-week-old female BALB/c mice (Charles River) were passed through a steel sieve and red blood cells lysed by the addition of 5 ml of ACK buffer (0·15 m NH₄Cl, 1 mm KHCO₃, 0·1 mm EDTA) for 3 min. Cells were washed with PBS several times before separation on a Percoll^® density gradient of 1·06–1·09 g/ml. After centrifuging the gradient for 20 min at 4° and 700 g, the lymphocytes were collected from the interphase and washed extensively in PBS before further use. The enriched lymphocyte population consisted of > 80% naive resting T cells and 20% resting B cells, as determined by fluorescence-activated cell sorter (FACS) analysis (data not shown). A total of 1 × 10⁵ cells/well were plated with 1 × 10⁴ DC in round-bottomed 96-well plates (Costar, Bodenheim, Germany) and incubated at 37° in an atmosphere of 5% CO₂. On day 4, 10 µCi/well of [³H]TdR (Amersham) was added. After further incubation for 18 hr the plates were harvested onto 96-well glass-fibre filter plates and the incorporated radioactivity determined by liquid scintillation counting using a TopCount β-counter (Canberra Packard, Dreieich, Germany).

Immunostaining and flow cytometry

Flow cytometry was performed as described previously.⁷ Briefly, DC were incubated with primary mAb for 30 min at 4°, washed and stained with the appropriate FITC-labelled secondary mAb. To determine cell viability and to exclude dead cells, propidium iodide (1 µg/ml; Sigma, Steinheim, Germany) was added. A total of 5 × 10⁴ cells were collected using a FACScan and analyzed with CellQuest research software (both Becton-Dickinson).

Annexin V staining

DC and T cells were co-cultured for 24 hr, then 3 × 10⁵ cells were labelled with annexin V–FITC (R&D Systems, Wiesbaden, Germany) and 7-amino-actinomycin D (7-AAD) for 20 min at 4°, followed by three-colour flow cytometry analysis using a FACScan flow cytometer and CellQuest analysis software (Becton-Dickinson), as described previously.²⁵

TdT-mediated biotin–dUTP nick-end labelling assays

TdT-mediated biotin–dUTP nick-end labelling (TUNEL) assays were performed to determine the percentage of T cells undergoing apoptosis following treatment with mAb. A total of 5 × 10⁵ cells were washed in PBS and incubated with phycoerythrin (PE)-labelled anti-CD3 mAb for 45 min on ice. Following incubation, the cells were washed twice in PBS containing 0·1% bovine serum albumin (BSA) and 0·1% NaN₃, then fixed with 3% paraformaldehyde for 30 min at room temperature. Cells were washed, then permeabilized for 3 min on ice with sodium citrate buffer containing 0·1% Triton-X-100. After washing, the cells were labelled with FITC-conjugated dUTP in the presence of terminal deoxynucleotidyl transferase enzyme solution and nucleotide mixture, for 1 hr at 37°, using the cell death detection kit (Boehringer Mannheim, Mannheim, Germany). To exclude cell debris, cells were stained with 7-AAD for 20 min at 4° followed by three-colour flow cytometry analysis using a FACScan flow cytometer and CellQuest analysis software (both Becton-Dickinson).

Time-lapse video microscopy

TCR transgenic T cells were labelled for 20 min at 37° in 2 ml of medium containing 1 µm FURA2/AM (1-[2-(5-Carboxyoxoazol-2-yl)-6-Aminobenzofuran-5-oxy]-2-(2′-Amino-5′-Methylphenoxy)-Ethan-N,N,N′,N′,-Tetraacetacetat, Pentaacetoxymethyl-ester) (Molecular Probes, Leiden, the Netherlands), and washed twice in PBS. Time-lapse video microscopy was performed using a T.I.L.L.-Photonics digital video imaging system consisting of a TILL-imago CCD camera, a Polychrome II monochromator (TILL Photonics, Munich, Germany) connected to a IMT-2 inverted microscope (Olympus, Hamburg, Germany). The cells were maintained in a heated (37°) incubation chamber under sterile conditions in RPMI-1640. Ca²⁺ influx was calculated using the TILLvision V3·03 software, deducting the absorption at 380 nm (free FURA-anions) from the absorption at 355 nm (Ca²⁺ FURA complexes), at an excitation wavelength of 480 nm.

Results

DC express pan CD44 epitopes and epitopes encoded by CD44 exons v4 and v6

CD44 expression has been investigated on human DC^7,26 and T cells,²⁷ but little is known about CD44 expression on murine DC. We have previously generated mAbs directed against murine CD44v4 and CD44v6 epitopes⁷ and we used these and other anti-CD44 antibodies to investigate CD44 expression on murine DC. FACS analysis of day-6 murine bone marrow-derived DC showed a high expression of pan CD44 epitopes, as well as clearly detectable amounts of CD44v4 and v6 (Fig. 1a). In contrast, naive splenic T cells were only weakly stained with mAb against CD44v4 and CD44v6, but showed high expression of pan CD44 epitopes (Fig. 1b).

Figure 1.

CD44 expression on murine bone marrow-derived dendritic cells (DC) and T cells. Analysis by flow cytometry of the indicated surface receptors on day-6 bone marrow-derived DC (a) or (b) naive splenic T-cells from C57BL/6 mice.

Anti-pan CD44 and CD44v4 mAbs dose-dependently inhibit the proliferation of naive resting T cells

The role of CD44 during T-cell stimulation is still unresolved. CD44 acts as a costimulatory and activating molecule on DC,^26,28,29 but there is also recent evidence that T-cell apoptosis is induced by CD44 cross-linking on T cells.⁸ Therefore, we wanted to characterize the role of CD44 during the interaction of DC and T cells. Allogeneic mixed leucocyte reactions (MLR) were performed using bone marrow-derived murine DC from C57BL/6 mice and naive, resting T lymphocytes from spleens of BALB/c mice. mAbs against CD44 were added to the MLR at a concentration of 10 µg/ml and were present for the whole 5-day incubation period. As shown in Fig. 2(a) a significant decrease in T-cell proliferation was observed after the addition of anti-pan CD44 mAb IM7 or CD44v4 mAb to the MLR, whereas a only a minor effect was observed with the addition of anti-CD44v6 mAb. Addition of the anti-pan CD44 mAb, KM201, had no effect. Titration of the IM7 and CD44v4 mAbs showed that this effect was dose dependent (Fig. 2b), with statistically significant inhibition at concentrations of ≥ 2 µg/ml.

Figure 2.

Pan CD44 and CD44v4 monoclonal antibodies (mAbs) dose-dependently inhibit the capacity of dendritic cells (DC) to stimulate the proliferation of naive resting T cells. Bone marrow-derived DC from C57BL/6 mice were co-incubated for 4 days with 1 × 10⁵ alloreactive T cells from BALB/c mice at a DC : T-cell ratio of 1 : 10 with the addition of 10 µg/ml of the indicated mAbs. T-cell proliferation was determined on day 5 by adding 1 µCi of [³H]thymidine ([³H]TdR) for the final 18 hr. Results are shown in counts per minute (c.p.m) + standard deviation (SD) of triplicate wells. Ctrl., control; PHA, phytohaemagglutinin; TC, T cells. (b) T-cell stimulation was performed as described for (a), but with the addition of the indicated concentrations of appropriate immunoglobulin G (IgG) control mAbs (• and ▴), anti-CD44v4 mAb clone 10D1 (▾) and anti-pan CD44 mAb clone IM7 (▪). *P > 0·01 compared with the appropriate IgG control. Results represent one of four independent experiments.

CD44 mAb interferes with T-cell activation by acting exclusively on DC

To discriminate between different effects of the CD44 mAbs on DC and T cells, each cell type was preincubated separately for 3 hr with 10 µg/ml of mAb. The cells were then washed and incubated with alloreactive T cells or DC, as appropriate, in DC–T-cell co-culture experiments. Interestingly, significant inhibition of T-cell proliferation was observed when DC were pretreated with anti-CD44 mAb (Fig. 3a), whereas preincubation of the T cells had no such effect (Fig. 3b). However, this effect was restricted to the anti-pan CD44 mAb (IM7) and the anti-CD44v4 mAb, whereas anti-pan CD44 mAb (KM201) and the anti-CD44v6 mAb had no effect. Taken together, our data suggest that mAb ligation of distinct CD44 epitopes on DC is able to interfere with the T-cell activation process.

Figure 3.

Selective preincubation of dendritic cells (DC) and T cells with anti-CD44 monoclonal antibody (mAb). DC (a) or T cells (b) were selectively preincubated for 3 hr at 37° with 10 µg/ml of the indicated mAb, washed extensively and co-incubated at a DC : T-cell ratio of 1 : 10, as described in the legend to Fig. 2. Results are shown in counts per minute (c.p.m.) + standard deviation (SD) of triplicate wells. *P > 0·01 compared with the appropriate immunoglobulin G (IgG) control. The results shown represent one of three independent experiments.

The effect of anti-CD44 mAb is not based on Fc interactions or CD44 expression on DC

We also wished to exclude that the effect of the anti-CD44 mAb was dependent on the interaction of the Fc parts of the mAb by cross-linking several CD44 on the DC surface. However, pretreatment of DC with F(ab′)₂ fragments of the anti-CD44 mAb, IM7, or IgG2b control mAb resulted in the same dose-dependent inhibition of T-cell proliferation as that of the complete mAb (Fig. 4a). As CD44 has been described as a costimulatory factor on DC,²⁸ we wished to compare our results generated by mAb-blocking experiments with a situation where the whole molecule is absent. DC and T cells were prepared from CD44-deficient or wild-type C57BL/6 mice¹⁷ and co-incubated in an allogenic MLR with cells from BALB/c mice. In accordance with data published by Schmitt et al.,¹⁷ we found no significant difference in the T-cell proliferation induced by CD44^−/− DC or wild-type DC incubated with CD44^−/− T cells compared with the wild-type control (Fig. 4b).

Figure 4.

The effect of anti-CD44 monoclonal antibody (mAb) is not attributable to Fc interactions or CD44 expression on dendritic cells (DC). (a) DC were pretreated for 3 hr at 37° with F(ab′)₂ fragments of the anti-CD44 mAb, IM7, and the appropriate IgG2b control mAb, at the concentrations indicated. DC were washed and co-incubated with T cells at a ratio of 1 : 10, as described in the legend to Figure 2. (b) DC and T cells were prepared from CD44-deficient or wild-type C57BL/6 mice¹⁷ and co-incubated with allogenic cells from BALB/c mice in ratios of 1 : 10 and 1 : 20, as indicated. Results are shown as counts per minute (c.p.m.) + standard deviation (SD) of triplicate wells. A representative of two independent experiments is shown.

CD44 mAb-treated DC induce T-cell apoptosis that is independent of Fas/FasL

To investigate changes induced by the anti-CD44 mAb in DC, FACS analysis of the cell-surface molecules MHC class II, B7-1/B7-2 and CD40 (relevant receptors for the induction of T-cell activation) was performed. However, no changes were detected in the expression of these receptors after treatment with CD44 mAb (data not shown). In contrast to human monocyte-derived DC,²⁶ there was also no difference in the activation status of murine DC after treatment with CD44 mAb, as determined by cell morphology, cell viability (as gauged by propidium iodide uptake) and quantification of the production of the proinflammatory cytokines, tumour necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) (data not shown). Furthermore, no evidence was found for intracellular signalling events, as electromobility shift assays (EMSA) assays performed with nuclear extracts from IM7-treated DC did not reveal any liberation of NF-κB to the nucleus (data not shown). Moreover, treatment with the anti-CD44 mAb, IM7, has been reported to induce shedding of CD44 from the surface of lymphocytes.¹⁴ To investigate this, DC were analysed for their CD44 surface expression 1, 3, 5, 7 and 9 hr after treatment with IM7. Again, no difference in CD44 expression was detected when the FITC-labelled anti-mAb, KM201 (which reacts with an epitope close to a hyaluronate-binding domain different from IM-7; data not shown), was used for FACS analysis. Therefore, a more direct effect of the antibodies appeared to be responsible for the inhibition of T-cell proliferation.

CD44 cross-linking by IM7 on T cells induces the apoptosis of T-helper cells in the presence of subthreshold levels of CD3 mAb.⁸ We therefore examined whether co-incubation of T cells with anti-CD44 mAb-treated DC resulted in the induction of apoptosis. To investigate early changes related to T-cell apoptosis, CD3⁺ T cells were stained for annexin V expression after 24 hr of co-incubation with pretreated DC (Fig. 5a). We found that up to 10% of the T cells were annexin V positive if co-incubated with IM7- or CD44v4-treated DC (Fig. 5a). In contrast, mAb pretreatment of the T cells had no effect, nor did preincubation with the non-blocking CD44 mAbs (Fig. 5a). To verify the induction of apoptosis, as indicated by the annexin V staining, TUNEL assays were performed on days 3 and 4 of the MLR to detect DNA fragmentation in apoptotic T cells. Again, IM7 and CD44v4 mAb treatment of DC induced a low, but significant, increase in the number of apoptotic T cells (Fig. 5b). To confirm that the decrease in T-cell proliferation is indeed related to an increased number of apoptotic cells, the number of propidium iodide-positive cells was measured in the assays on days 3 and 4 (Fig. 5b). Here, we found a background level (approximately 10%) of dying, propidium iodide-positive cells, which may result from unstimulated T cells that do not have the corresponding TCR on their surface. However, preincubation of the DC with IM7 and, to a lesser degree, also CD44v4 mAb, induced a significant increase of propidium iodide-positive cells, peaking at 37% on day 4 (Fig. 5b).

Figure 5.

Induction of annexin V staining and DNA fragmentation in T cells co-incubated with anti-CD44 monoclonal antibody (mAb)-treated dendritic cells (DC). DC or T cells were selectively preincubated for 3 hr at 37° with 10 µg/ml of the indicated mAb, washed and co-incubated with 10⁵ BALB/c T cells at a DC : T-cell ratio of 1 : 10, as described in the legend to Fig. 3. (a) After 24 hr the cells were labelled with phycoerythrin (PE)-conjugated CD3 mAb, stained with fluorescein isothiocyanate (FITC)–annexin V and 7-amino-actinomycin D (7-AAD), and analysed by flow cytometry. Annexin V-positive and 7-AAD-negative cells in the CD3-gated population were counted as apoptotic T cells. The percentage of apoptotic cells from one experiment, representative of three carried out, is shown as a histogram. (b) On day 3 or day 4, T cells were stained with PE–CD3 mAb followed by nicked DNA labelling with FITC-labelled dUTP using a TdT-mediated biotin–dUTP nick-end labelling (TUNEL) assay kit, as described in the Materials and Methods. The percentage of T cells undergoing apoptosis was determined by gating CD3-, dUTP–FITC- and 7-AAD-positive cells by flow cytometry.

These effects were only observed when DC, but not T cells, were preincubated with these antibodies, and suggests that blockade of distinct CD44 epitopes on DC can inhibit T-cell proliferation and promote T-cell apoptosis.

AICD is dependent on the interaction of Fas and its ligand FasL.³⁰ We therefore analysed the expression of FasL on the surface of DC. However, using antibody staining, no changes were detected in FasL surface expression on DC treated with IM7 (data not shown). To test definitively whether the Fas/FasL pathway is involved in the induction of T-cell apoptosis by CD44 mAb-treated DC, we co-cultured Fas-deficient T cells from Lpr mice together with wild-type DC (Fig. 6). Conversely, we co-cultured DC from FasL-deficient Gld mice together with wild-type T cells (Fig. 6). T cells co-incubated with IM7-pretreated, FasL-deficient DC from Gld mice again showed dose-dependent inhibition of proliferation and increased apoptosis (Fig. 6a; data not shown), comparable to the results obtained when Fas-deficient T cells from Lpr mice were co-cultured with wild-type DC (Fig. 6b). Therefore, the Fas/FasL system plays no role in the activation-dependent T-cell apoptosis induced by CD44 mAb-treated DC.

Figure 6.

Induction of T-cell apoptosis is independent of the expression of Fas/Fas ligand (Fas/FasL). Dendritic cells (DC) were preincubated for 3 hr at 37° with different concentrations of the anti-pan CD44 monoclonal antibody (mAb), IM7, washed and co-incubated at a DC : T-cell ratio of 1 : 10, as described in the legend to Fig. 3. To investigate the influence of Fas/FasL, either DC from FasL-deficient C57BL/6-Gld mice were co-incubated with allogenic T cells (TC) from BALB/c mice (a) or DC from C57BL/6 mice were co-incubated with allogenic T cells from Fas-deficient BALB/c-Lpr mice (b). Results are shown as counts per minute (c.p.m.) + standard deviation (SD) of triplicate wells. *P > 0·01 compared with the immunoglobulin G (IgG) control. The results represent one of two independent experiments carried out.

CD44 pretreated DC inhibit CD4⁺, but not CD8⁺, T-cell proliferation by interference with early Ca²⁺ signalling

Finally, we wished to investigate whether treatment of DC with CD44 mAbs affects equally CD8⁺ cytotoxic and CD4⁺ T-helper cells. Previous publications have described a function for CD44 on a T-helper cell line³¹ but nothing is known about CD8⁺-mediated T-cell responses. Therefore, P14 mice, expressing a Vα2/Vβ8 TCR specific for the MHC class I-restricted GP33 peptide of the LCMV²³ or the mouse strain DO 11.10 (carrying a TCR for amino acids 323–339 of the MHC class II-restricted ovalbumin peptide)²² were used. T cells from TCR-transgenic mice bear the advantage that they respond very uniformly to peptide-pulsed DC and are therefore an ideal tool for using to examine the early events that occur during DC–T-cell interactions at a single-cell level. Time-lapse video microscopy was established to measure the cytosolic Ca²⁺ influx of activated T cells as an early parameter occurring during the first seconds of DC–T-cell interactions depending on the f-actin bundling in DC.³² Furthermore, Ca²⁺ signalling has been described to be prerequisite for the formation of the immunological synapse leading to T-cell activation and proliferation.³³ DC were untreated or preincubated with IM7 mAb, as described above. Surprisingly, CD44-pretreated DC, in the presence of stimulatory concentrations of ovalbumin peptide, induced a significantly diminished Ca²⁺ influx in DO 11.10 CD4⁺ T cells, affecting both the initial peak as well as the lasting lower influx that followed (Fig. 7a, 7c). However, this effect was only observed in CD4⁺, not p14 CD8⁺, T cells (Fig. 7b, 7d). Furthermore, proliferation assays performed under the same conditions confirmed these results, as only DO 11.10 CD4⁺ T-helper cells showed a dose-dependent inhibition of proliferation in response to IM7-treated DC, whereas the proliferation of CD8⁺ cytotoxic T cells from p14 mice was significantly enhanced (Fig. 7e, 7f). These data provide the first evidence for a regulatory role of CD44 on DC for the activation of CD4⁺ T cells by interference with early cytosolic Ca²⁺ influx.

Figure 7.

CD44-pretreated dendritic cells (DC) inhibit CD4⁺, but not CD8⁺, T-cell proliferation by interference with early Ca²⁺ signalling. To investigate the influence of CD44-treated DC on T-cell responses, T cells (TC) from T-cell receptor (TCR) transgenic BALB/c DO 11.10 mice (CD4⁺) (panels a, c and e), or from TCR transgenic P14 mice (CD8⁺) (panels b, d and f), were incubated with stimulatory concentrations of ovalbumin 323–339 (CD4⁺) or lymphocytic-choriomeningitis virus (LCMV) 32–42 (CD8⁺) peptide, as described previously.⁴,³⁴ TCR transgenic T cells were labelled with FURA2/AM, and Ca²⁺ influx during cell–cell contacts with untreated (panels a and b) or 10 µg/ml IM7-treated (panels c and d) autologous murine bone marrow-derived DC was analysed using a time-lapse video microscopy system. The histograms represent the synchronized fluorescence changes of the indicated number of events from one film. (e) and (f) DC were preincubated for 3 hr at 37° with different concentrations of the anti-pan CD44 monoclonal antibody (mAb), IM7, washed and co-incubated at a DC : T-cell ratio of 1 : 10 with the same peptide concentrations described above. Results are shown as counts per minute (c.p.m.) + standard deviation (SD) of triplicate wells. *P > 0·01 compared with the IgG control. The results represent one of two independent experiments carried out.

Discussion

In this study we demonstrate that mAb ligation of CD44 epitopes on DC during antigen presentation dose-dependently interferes with T-cell activation. Obviously, it is very important which epitopes are bound by the antibodies, as not all anti-CD44 mAbs tested had an effect, and inhibitory activity was found only with the anti-pan CD44 mAb, IM7,^14,34 and with an anti-CD44v4 mAb.⁷ This is in accordance with in vivo studies demonstrating an inhibitory effect of the anti-CD44 mAb, IM7, on T-cell-mediated immune responses in contact hypersensitivity and collagen-induced murine arthritis.^14–16 However, anti-CD44 mAb-mediated effects seem to have a different immunological relevance than absence of the whole molecule in knockout animals. As an example of these differences, Con A induces an acute hepatitis which is more severe in CD44 knockout mice and leads to 100% mortality.¹⁹ Chen et al. correlated the disease severity with the resistance of CD44-deficient T cells to activation-induced apoptosis.¹⁹ Moreover, two different CD44 knockout mice have been generated, both without signs of severe immunodeficiency.^17,18 In a MLR using CD44-deficient T cells, similar proliferative responses against allogenic CD44^+/+or CD44^−/− stimulators were observed.¹⁷ Furthermore, CD44^−/− T cells responded in a similar manner as wild-type T cells when treated with soluble or cross-linked anti-CD3 antibodies, staphylococcal enterotoxin B or Con A.¹⁷ Our mAb experiments, however, show that CD44 plays an important role during the initiation of T-cell responses, but the outcome depends critically on the distinct CD44 epitopes engaged. These effects cannot be detected when T-cell responses are examined in the complete absence of CD44.^17–19

An important new finding arising from our study is that DC are the target of anti-CD44 mAb that block activation of naïve resting T cells. Naïve resting T cells do not express significant amounts of CD44v4 or CD44v6 isoforms, and pretreatment with either pan or variant anti-CD44 mAbs had no effect. This is in accordance with various in vivo studies, which show that only activated T cells are affected by treatment with anti-CD44 mAbs.^14,15,34 In this regard, it is interesting to note that CD44 variants are transiently upregulated during lymphocyte activation, and that this upregulation is necessary for an immune response to develop.^4,27 Taken together, our data support the model in which expression of CD44 variant-containing isoforms on immunocompetent cells acts as an activating switch.^26,27

Using TCR transgenic systems, we further demonstrated that the blocking effect of CD44 mAb-treated DC affected a CD4⁺-mediated, but not a CD8⁺-mediated, immune response. This is in accordance with data on collagen-induced arthritis, which is inhibited by CD44 mAb injection and clearly dependent on CD4⁺ T-helper cells.^15,34 Moreover, the development of a cutaneous, hapten-induced contact hypersensitivity is also strongly inhibited by the intravenous injection of anti-CD44 mAb, IM7.⁷

How does CD44 mAb binding to DC interfere with T-cell stimulation? No signalling activity or changes in surface receptor expression or cytokine production were detected in anti-CD44 mAb-treated DC. CD44 is closely linked to the cytoskeleton by the small GTPases Rho, Rac and Cdc42.^31,35 An essential step for T-cell activation is the polarized recruitment of receptors to the cell–cell interaction site, called the immunological synapse.³³ CD44 on T cells has been shown to be involved in the recruitment of signal transduction molecules to the site of the TCR–CD3 complex and lipid raft formation,^8,31 and it could be speculated that similar mechanisms might play a role on DC, because it has been shown recently that DC actively polarize their actin cytoskeleton during interaction with T cells.³² Our data provides the first evidence that CD44 is indeed involved in the early T-cell activation steps, as CD44 mAb-treated DC interfere with early Ca²⁺ signalling, leading to non-proliferation and the induction of T-cell apoptosis.

Furthermore, we have also shown that the induction of CD4⁺ T-cell apoptosis by CD44 mAb-treated DC is independent of the Fas/FasL system. The interaction of Fas with FasL is characteristic of AICD.³⁰ Various reports have shown that the induction of FasL on DC is the critical step leading to T-cell apoptosis and functions by engagement of the Fas that is constitutively expressed on T cells.^2,30 However, we show that the T-cell apoptosis induced by CD44 mAb-treated DC is independent of the Fas/FasL pathway, as the mAb effect was not influenced when Fas/FasL-deficient DC were used. This is in accordance with in vivo data from Chen et al. who found no differences in FasL regulation during the induction of T-cell apoptosis in an experimental hepatitis model.¹⁹ Fas-mediated AICD has been described to occur during the downregulation of an immune response.^2,30 In contrast, the induction of Fas-independent apoptosis via ligation of CD44 occurs during the early steps of T-cell stimulation and interferes with Ca²⁺ signalling.

In conclusion, we have shown that during antigen presentation to T cells by DC, engagement of distinct CD44 epitopes, namely those detected by mAbs IM7 and CD44v4, act exclusively on DC, thereby interfering with Ca²⁺ signalling in CD4⁺ T cells and inhibiting T-cell activation and proliferation.

Abbreviations

AICD

activation-induced cell death

APC

antigen-presenting cell

Con A

concanavalin A

DC

dendritic cells

FasL

Fas ligand

mAb

monoclonal antibody

FITC

fluorescein isothiocyanate

MLR

mixed leucocyte reaction

PBS

phosphate-buffered saline

PE

phycoerythrin

TCR

T-cell receptor