Costimulation via CD55 on Human CD4⁺ T Cells Mediated by CD97

Decay accelerating factor (CD55) is a 70-kDa single chain complement regulatory protein, that consists of four extracellular short consensus repeat (SCR)³ domains, linked to the membrane by a GPI anchor (1). It is widely expressed on human cells, including PBLs, and protects them from complement-mediated lysis. CD55 induces the decay of the C3 and C5 convertases, preventing C3b deposition and downstream assembly of the MAC complex (1, 2). CD55 is also used by many viral and bacterial pathogens as an adhesion or invasion receptor (3, 4, 5, 6, 7). Moreover, CD55 has been shown to interact with an epidermal growth factor-seven-span transmembrane (EGF-TM7) receptor, CD97 (8), an Ag that becomes immediately up-regulated on most leukocytes during activation (9) and is abundantly expressed by macrophages, dendritic cells, and some epithelial tumors (10, 11). CD97 consists of a membrane-spanning region, homologous to the secretin receptor superfamily, and an extended extracellular region with several epidermal growth factor-like (EGF) domains at the N terminus (12). Alternative RNA splicing results in expression of different isoforms, containing three (EGF 1,2,5), four (EGF 1,2,3,5), or five (EGF 1,2,3,4,5) EGF-like domains (13). CD55 binds with the highest affinity to the smallest isoform (EGF 1,2,5), and only these three EGF-like domains have been shown to be necessary for this interaction (14).

Early studies suggested a role for CD55 in costimulating submitogenic doses of phorbol esters resulting in T cell proliferation (15), implying that direct stimulation of CD55 could have an effect on T cell activation. Furthermore, cross-linking of CD55 with Abs was shown to result in signal transduction in lymphoid and nonlymphoid cells (16, 17, 18). More recently, two studies on CD55-deficient mice have shown enhanced T cell responses, suggesting that CD55 can directly suppress T cell function (19, 20). However, the T cell hyperresponsiveness was shown to be related to enhanced complement activation, due to lack of CD55 complement regulation, because normal responses were restored in CD55^−/−C3^−/− or CD55^−/−Factor D^−/− double-knockout mice.

In the present study, we investigated whether CD55 engagement directly costimulates human CD4⁺ T cell activation or if such an effect can only be mediated by a reduction in complement inhibition. We show that direct CD55 engagement using either cross-linking Abs or the ligand CD97 can costimulate CD3 engagement, resulting in enhanced T cell proliferation, cytokine production, and expression of activation markers, whereas the effect of CD55 on complement inhibition did not have a direct role in this system.

All reagents used were from Sigma-Aldrich unless otherwise stated. The mAb BRIC 216 (SCR3) and BRIC 220 (SCR1) were purchased from International Blood Group Reference Laboratory. The mAb 791T/36 (21) is an IgG2b that recognizes SCR1 and SCR2 of CD55 and was produced in-house. The anti-CD3 mAb, clone OKT-3, was purified from the B cell hybridoma (American Type Culture Collection). The anti-CD28 mAb and the anti-CD25 Ab, PE conjugated, were purchased from BD Biosciences. The anti-CD69-PCy5 and the anti-IL-2-PE Abs were purchased from IO TEST (Beckman Coulter). The mouse IgG1 and IgG2b isotype controls were purchased from Serotec and BD Biosciences, respectively. The rabbit anti-C3c FITC-conjugated Ab and the anti-CD3 FITC-conjugated Ab were purchased from DakoCytomation. The fixative reagent Cellfix and the reagent containing brefeldin A, GolgiPlug, were purchased from BD Biosciences. CFSE was purchased from Invitrogen Life Technologies.

The EGF1,2,5-Fc fusion constructs of EMR2 and CD97 were made as previously described (22). HEK293T cells were maintained in DMEM, supplemented with 10% heat-inactivated FCS and 2 mM l-glutamine. Cells at 60% confluence were transfected with 10 μg of DNA/78-cm² culture dishes (Corning) using calcium phosphate precipitation. Six hours posttransfection, medium was replaced with 15 ml of DMEM, 1% FCS, and incubated for a further 72 h. Conditioned culture supernatant was collected, buffered to pH 7.6, and passed repeatedly at 4°C over protein G-Sepharose column, using a dialysis pump. The column was washed with 100 mM citrate buffer at pH 6.0 and Ab was eluted with 100 mM glycine buffer at pH 2.5. Eluate was collected into 1 M Tris (pH 7.0) and dialyzed against PBS overnight at 4°C.

PBMCs were purified from blood of anonymous human healthy donors by Histopaque-1077 separation. CD4⁺ T cells were isolated from PBMCs by negative selection, using the CD4⁺ T cell enrichment kit from R&D Systems. Purity of the preparation was between 92 and 97%, as routinely checked by flow cytometric analysis, with the main contaminants being CD8⁺ T cells and NK cells (data not shown). Naive CD4⁺ T cells were also isolated from PBMCs by negative selection, using a naive CD4⁺ T cell (CD3⁺CD4⁺CD45RO⁻) enrichment kit, purchased from R&D Systems. CD3⁺CD4⁺ cells were 96–97% of the preparation and 99.5% CD45RA⁺ (data not shown).

PBMCs were isolated by Histopaque-1077 separation and CD4⁺ T cells were negatively selected with a CD4⁺ T cell enrichment kit (R&D Systems), before staining with anti-CD3-FITC and anti-CD4-PCy5 Abs for 20 min at room temperature. After two washes with T cell medium, cells were sorted on a Coulter Epics Altra (Beckman Coulter) with Expo32 software. Cells were cultured for 4 days at a concentration of 2 × 10⁵/well, in the absence or presence of 100 U/ml IL-2, with plate-bound mAbs as described below.

CD4⁺ T cells (1 × 10⁵/well) and naive CD4⁺ T cells (1.5 × 10⁵/well) were stimulated with combinations of plate-bound anti-CD3 mAb (1 or 5 μg/ml), plate-bound anti-CD55 mAbs (BRIC 216, 791T/36, BRIC 220), or irrelevant IgG1 and IgG2b mAbs (all at 10 μg/ml, unless otherwise stated) in flat-bottom 96-well plates for 72 h at various concentrations, in the absence or presence of IL-2 (100 U/ml). The anti-CD28 mAb was also used at 10 μg/ml, either fluid-phase or plate-bound, because there were no significant differences between the two. CD97-Fc and EMR2-Fc proteins were also used at 10 μg/ml and plate-bound to the wells. Cultures were pulsed for the final 8 h with 0.5 μCi of [³H]thymidine (Amersham Biosciences), harvested on a 96-well plate Harvester (Filtermate 196; Packard; PerkinElmer), and incorporated activity was measured using a TopCount (Packard; PerkinElmer).

Purified CD4⁺ T cells were stimulated as described above. After 72 h, they were fixed with 70% ethanol and frozen at −20°C. Before propidium iodide staining, they were washed with PBS, treated with DNase-free RNase (Roche) for 30 min, and then stained with 10 μg/ml propidium iodide solution for 15 min. Samples were then analyzed on a FACScan (BD Biosciences) with CellQuest software. Cell cycle analysis was conducted by diploid peak analysis using the ModFitLT Mac 3.1 SP2 software.

After 3-day activation, cells (1 × 10⁵) were stained for 1 h on ice with saturating concentrations of anti-CD25 PE-labeled and anti-CD69 PCy5-labeled Abs. After washing with PBS, they were fixed with 1× Cellfix and analyzed on a FACScan with CellQuest software.

Purified CD4⁺ T cells (1 × 10⁵/well in 200 μl) were cultured in 96-well plates in the absence or presence of IL-2 (100 U/ml), and supernatants were collected after 64 h. Simultaneous analysis of 10 different cytokines was performed with the Bio-Plex fluorometric bead-based assay (Bio-Rad). The cytokines investigated were as follows: IL-2, IL-10, IL-4, IL-5, IFN-γ, TNF-α, GM-CSF, IL-12 p70, IL-13, and IL-17, and the analysis was conducted according to the manufacturer’s protocol on 50 μl of culture supernatants.

Purified CD4⁺ T cells were cultured as described above for 12 h, in the presence of brefeldin A. After washing with PBS, cells were fixed, permeabilized with 0.1% saponin, and stained for 2 h with IL-2-PE Abs or relevant isotype control. After two further washes with PBS, cells were fixed again and analyzed on a FACScan with CellQuest software.

A peptide-specific CD4⁺ T cell clone was generated as follows. PBMCs were obtained from heparinized blood by centrifugation over Histopaque-1077, washed, and incubated at 2 × 10⁶ cells/well in 24-well plates for 7 days with 20 μg/ml Tie-2 peptide (GGITIGRDFEALMNQHQDPLEV). The responding T cells were cloned by limiting dilution in 20 μl Terasaki wells using 1 × 10⁶/ml irradiated (8.54 Gy) autologous PBMCs as APCs, 50 U/ml rIL-2, and 20 μg/ml specific peptide. The culture medium (T cell medium) was RPMI 1640 supplemented with 5% human serum, 1% Na-pyruvate, 1% l-glutamine, 2% HEPES, 1% penicillin-streptomycin, 1% nonessential amino acids, and 50 U/ml rIL-2 (PeproTech). Clones were expanded at 21-day intervals by restimulation with 5 × 10⁵/ml irradiated allogeneic PBMCs, IL-2 (50 U/ml), and 2 μg/ml PHA. Ag specificity of clones was examined at least 10 days after restimulation with PHA. A selected clone was characterized and shown to have a Th1 phenotype, secreting IFN-γ in response to peptide presented on autologous APCs (data not shown).

The effect of fluid-phase and plate-bound CD55 mAbs 791T36, BRIC 220, or CD97-Fc on C3b deposition was assessed as described by Morgan et al. (23) with minor modifications. Cells were incubated for 1 h on ice with 10 μg/ml fluid-phase 791T/36, BRIC 220, IgG1 isotype control, IgG2b isotype control, CD97-Fc, and EMR2-Fc. After addition of 10% fresh human serum, plates were transferred to a 37°C, 5% CO₂ incubator for 2 h. Cells were then washed twice with PBS and stained with saturating concentrations of a rabbit anti-human C3c-FITC Ab for 1 h on ice. This Ab also recognizes C3 and C3b. After washing with PBS, samples were analyzed on a FACScan with CellQuest software.

Five millimolar CFSE stock solution was prepared by dissolving the CFSE in DMSO. Before addition to the cells, an aliquot was dissolved in 1.5 ml of T cell medium to give a final concentration of 3 μM. Freshly isolated CD4⁺ T cells were resuspended in 2.5 ml of T cell medium and kept under gentle agitation while CFSE was added. After 5 min, cells were washed twice with 5 vol of medium and plated out on mAbs, CD97-Fc- and EMR2-Fc-precoated 96-well plates. After 5 days in culture, cells were fixed with 1× Cellfix and analyzed on a FACScan with CellQuest software.

Specificity of CD97-CD55 interaction was assessed in a proliferation assay, masking CD55 with an anti-CD55 mAb, BRIC 220. CD4⁺ purified T cells (1 × 10⁵/quadruplicate) were preincubated with 50 μg/ml BRIC 220 or IgG1 isotype control in T cell medium, before being plated out in 96-well plates precoated with 5 μg/ml anti-CD3 mAb, 10 μg/ml CD97-Fc or negative control EMR2-Fc, keeping the cells in the presence of saturating concentrations of BRIC 220 or matched isotype control. Cells were cultured for 3 days and pulsed with 0.5 μCi of [³H]thymidine for the last 8 h, and then harvested on a 96-well plate harvester, and incorporated activity was measured using a TopCount.

Statistical analysis on proliferation, cytokine secretion, and C3b deposition data was conducted using Student’s t test. A value of p < 0.05 was considered statistically significant. Values shown are the mean ± SD of not less than three replicates.

To establish whether anti-CD55 mAbs (anti-CD55) could costimulate purified human polyclonal CD4⁺ T cells, cells were first stimulated with a range of concentrations (0.01–10 μg/ml) of plate-bound anti-CD3 mAb OKT-3 (anti-CD3). This resulted in increased proliferation at concentrations of 1 μg/ml and above but required the presence of IL-2 (100 U/ml) to produce a significant increase (Fig. 1,a). Plate-bound anti-CD55 did not induce any effect in the absence of CD3 stimulation (data not shown). However, in the presence of CD3 stimulation, plate-bound 791T/36 mAb, which binds to the SCR1 and SCR2 regions of CD55, induced a 12-fold increase in proliferation in the absence of IL-2 (Fig. 1,b, ▪) and a 3-fold increase in its presence (Fig. 1,b, □). In contrast, an isotype-matched control mAb failed to costimulate CD3 engagement. Another anti-CD55 mAb, BRIC 216, directed to SCR3, also enhanced CD3 stimulation (Fig. 1 c), whereas the IgG1 isotype control did not.

Stimulation of CD4⁺ T cells with anti-CD55 (791T/36, unless otherwise specified) and anti-CD3 mAbs (CD55/CD3) was highly reproducible, because all 13 donors screened showed enhanced proliferation (Table I). However, there was a high variability between donors. Generally, those who were less responsive to CD3 engagement were more responsive to CD55 costimulation.

The effect of CD55 costimulation was also compared with CD28 costimulation, in the presence of 1 μg/ml anti-CD3 and 10 μg/ml plate-bound or fluid-phase anti-CD28. Fig. 1 d shows the increase in proliferation measured by [³H]thymidine incorporation was very similar for plate-bound and fluid-phase anti-CD28. Conversely, anti-CD55 mAbs result in costimulation only when used plate-bound (data not shown).

Coengagement of CD28 and CD3 on T cells results in increased cell cycle progression compared with CD3 engagement alone, resulting in a higher number of cells leaving G₀ phase (24). Similar results were obtained with CD55/CD3 stimulation. Fig. 1, e–g, shows that percentages of cells in G₂ and S phases increased 1.6- to 2.1-fold upon activation with CD55/CD3, compared with CD3 stimulation alone. Comparable results were obtained in the presence of IL-2 (data not shown).

T cell activation results in increased levels of CD69 and CD25 at the cell surface. Expression of these molecules was investigated in response to activation with anti-CD55 and anti-CD3. Fig. 2,a shows that, in the absence of any stimulus, the expression of both CD25 and CD69 remained low. Although CD3 stimulation alone induced high expression of CD69 on 70% of the cell population (Fig. 2,b), only a modest increase in the expression of CD25 was observed. As expected, CD28 costimulation further enhanced CD69 expression and significantly increased the number of CD4⁺ T cells expressing high levels of CD25 (Fig. 2,c). Interestingly, upon CD55/CD3 engagement, the majority of CD4⁺ T cells also expressed high levels of both CD69 and CD25 (Fig. 2 d).

Concomitant with cellular activation is the secretion of cytokines by CD4⁺ T cells. To investigate what effect CD55 costimulation could have on cytokine production, culture supernatants were analyzed after 3 days of CD55/CD3 stimulation, in the presence or absence of exogenous IL-2. Fig. 2 e shows that there was a considerable increase in all measurable cytokines. In the absence of IL-2, the most relevant effect was an increase in secretion of IL-10 (7-fold) and GM-CSF (5-fold). Interestingly, in the presence of exogenous IL-2, there was a significant change in IL-10 secretion, where a 12-fold increase was observed. Smaller increases were measured for GM-CSF. Although IFN-γ, TNF-α, and IL-13 were affected by CD55 costimulation, this was to a lesser extent, whereas IL-4, IL-5, IL-12 p70, and IL-17 were undetectable (data not shown).

A comparison of CD28 and CD55 costimulation on cytokine production revealed significant differences in both IL-2 and IL-10 secretion (Fig. 2, f and g). IL-2 production, investigated by intracellular staining, was significantly increased by CD28 (27% positive cells) (Fig. 2,f) but not CD55 costimulation. Similar results were obtained in the presence of IL-2 (data not shown). However, in the presence of exogenous IL-2, CD55 costimulation showed a 15-fold higher increase in IL-10 compared with CD28, which had a marginal effect on IL-10 secretion (Fig. 2 g). There were no significant differences in the production of GM-CSF, TNF-α, IFN-γ, and IL-13 in these assays (data not shown).

Negatively selected CD4⁺ T cells were still contaminated with a small proportion of CD3⁻ and CD4⁻ cells. To exclude any enhancing effect by contaminating cells, APCs in particular, CD3⁺CD4⁺ cells were positively selected after the first round of negative selection. The highly pure (99.9%; data not shown), sorted cell population was stimulated with plate-bound anti-CD3 (1 μg/ml) in combination with anti-CD55 or anti-CD28 (10 μg/ml) for 4 days, and then proliferation and cytokine production were measured (Fig. 3). The results confirmed previous data observed with negatively selected cells. CD55 engagement significantly increased CD3-induced proliferation (5-fold) (Fig. 3,a) and IL-10 production in the presence of exogenous IL-2, where a 40-fold increase was observed (Fig. 3 b).

Freshly isolated CD4⁺ T cells are a mixed population of cells at different developmental stages and states of activation. To investigate the result of CD55/CD3 engagement on a homogeneous population that has already acquired an effector/memory phenotype, a CD4⁺ T cell clone, generated against a specific peptide, was used. The clone showed a Th1 phenotype, secreting IFN-γ in response to peptide presented on autologous APCs (data not shown).

Upon CD55/CD3 stimulation, the T cell clone also responded with increased proliferation and cytokine secretion (Fig. 4). Lower concentrations of plate-bound anti-CD3 were required to induce proliferation compared with purified CD4⁺ T cells, and optimal CD3 stimulation was achieved at a concentration of 0.01 μg/ml (Fig. 4,a). The bar in Fig. 4,a indicates the range of suboptimal doses of anti-CD3, which effect can be enhanced by a second stimulus, such as IL-2. Addition of increasing concentrations of anti-CD55 to 0.0005 μg/ml (▵), 0.001 μg/ml (▪), and 0.002 μg/ml (•) of anti-CD3 significantly potentiated T cell proliferation (5- to 50-fold) (Fig. 4,b). Furthermore, a reduced but similar effect was observed with the same suboptimal concentrations of anti-CD3, in the presence of IL-2 (Fig. 4,c). When the effect of CD55/CD3 engagement on cytokine production was assessed, no change in the cytokine profile was observed. IFN-γ and GM-CSF were up-regulated 2.6- and 2.2-fold, respectively, whereas TNF-α secretion was increased to a lesser extent (1.6-fold) (Fig. 3 d). No increase or induction of IL-4, IL-5, or IL-10 was detected, demonstrating that CD55-mediated costimulation does not alter the established phenotype of a committed T cell population.

CD55 costimulation is not able to modify the cytokine profile of committed T cells; therefore, we investigated whether it could modulate the priming of a naive CD4⁺ T cell population.

Naive CD3⁺CD4⁺ (99.5% CD45RA⁺; data not shown) showed enhanced proliferation when stimulated with 5 μg/ml anti-CD3 and 10 μg/ml anti-CD55, compared with CD3 stimulation alone, as assessed by [³H]thymidine incorporation (Fig. 5 a).

Cytokine secretion required the presence of exogenous IL-2, because other cytokines investigated were below detection level in its absence (data not shown). As previously observed with purified polyclonal CD4⁺ T cells, the major effect was observed on IL-10 production (15-fold increase) (Fig. 5 b), whereas other cytokines investigated were not significantly affected (data not shown). These results demonstrate that CD55 engagement is a sufficient costimulatory signal to prime CD4⁺ naive cells and it also able to induce a preferential production of IL-10 when compared with CD28.

Besides direct stimulation, binding of mAbs to CD55 might inhibit CD55 mediated complement regulation, which has been implicated in enhanced T cell responses (19, 20). To investigate the effect of anti-CD55 mAbs on complement regulation in our setting, we conducted a series of experiments to assess whether a correlation existed between costimulation with anti-CD55 mAbs and their effect on CD55 complement inhibition.

This was analyzed, as previously described (23, 25, 26), by measurement of C3b deposition on the cells, in the presence of a CD55 neutralizing mAb, 791T/36, and a known nonneutralizing mAb, BRIC 220 (23), which binds to SCR1 on CD55, a domain not involved in complement regulation (27). Purified CD4⁺ T cells were incubated with the mAbs and isotype-matched controls (IgG2b and IgG1, respectively) for 1 h and then exposed to 10% of fresh human serum for 2 h. Staining for C3b showed no deposition in the absence of fresh serum (data not shown), whereas there was a basal deposition on cells alone (Fig. 6 a). This level of C3b was not increased by the nonneutralizing Ab (BRIC 220). In contrast, the CD55 neutralizing mAb (791T/36) resulted in a 1.8-fold increase in complement deposition on cells.

The same mAbs were then used in a CD4⁺ T cell proliferation assay (Fig. 6,b). Cells were stimulated for 3 days with 10 μg/ml plate-bound 791T/36 and BRIC 220, in combination with 1 μg/ml anti-CD3. As Fig. 6 b shows, the two anti-CD55 mAbs had very similar costimulatory effects, proving that CD55 engagement and increased complement deposition are separate events. Therefore, costimulation, in this study, is a result of direct stimulation of CD55.

In addition to its role in complement regulation, CD55 has a cellular ligand, CD97, an EGF-TM7 receptor. To evaluate what effect CD97-CD55 interaction has on T cell activation, freshly isolated CD4⁺ T cells were stimulated with plate-bound anti-CD3 and plate-bound CD97-Fc (Fig. 7). The isoform of human CD97 that binds CD55 with the highest affinity (EGF 1,2,5) was generated as a fusion protein linked to mouse Fc (22). CD97-Fc alone did not induce any T cell proliferation (Fig. 7,a), whereas in combination with anti-CD3 an increase in proliferation occurred, both in the absence and in the presence of IL-2 (Fig. 7, a and b). The protein EMR2, expressed as EMR2 domains 1,2,5 linked to mouse Fc (EMR2-Fc) (22), was used as a negative control. EMR2 also belongs to the protein family EGF-TM7 and differs from CD97 only in six residues in the EGF-like domain region. Despite the high similarity with CD97, EMR2 does not effectively bind CD55, its affinity for CD55 being one order of magnitude lower than CD97. Fig. 7, a and b, shows there was no enhancing effect of plate-bound EMR2-Fc on proliferation of purified CD4⁺ T cells. To study the effect of CD97-mediated costimulation of CD55 on cell division, cells were stained with CFSE and then maintained in culture with plate-bound anti-CD3 and CD97-Fc or EMR2-Fc for 5 days (Fig. 7 c). CD97-Fc/anti-CD3 stimulation activated the majority of the cells (>95%), which progressed into five further divisions, compared with background proliferation seen with EMR2-Fc/anti-CD3. CD55 costimulation with anti-CD55 gave similar results to CD97-Fc (data not shown).

To verify that CD97-Fc was mediating its costimulatory effect through CD55, cells were precoated with the anti-CD55 mAb BRIC 220 before stimulation with plate-bound anti-CD3 and CD97-Fc. BRIC 220 binds to SCR1 on CD55, one of the domains involved in the binding to CD97, together with SCR2 and SCR3 (8). As Fig. 7 d shows, there was a 75% reduction in CD97-Fc mediated costimulation when cells were activated in the presence of BRIC 220, whereas an IgG1 isotype control did not have any effect.

The effect of CD97-Fc mediated costimulation on cytokine secretion (Fig. 7, e and f) was similar to anti-CD55, whereas the negative control EMR2 had no effect. There was a 10-fold increase in IL-10 and a 2.5-fold increase in IFN-γ. Intracellular staining of IL-2 also showed a similar profile for CD97-Fc and anti-CD55, with no significant IL-2 production, both in the presence and the absence of exogenous IL-2 (data not shown). Anti-CD55 mAbs were shown to increase C3b deposition on cells, upon addition of fresh human serum (Fig. 7). To study whether CD97-CD55 interaction interfered with CD55-mediated complement inhibition, a C3b deposition assay was conducted on CD4⁺ T cells in the presence of CD97-Fc and 791T/36 mAb. Fig. 7 f shows there was no increase in C3b deposition with CD97-Fc, further indicating that CD97 induces a direct stimulation of CD55, resulting in costimulation of CD4⁺ T cells.

CD55 is a complement regulatory protein, which binds to both the classical and alternative C3 and C5 convertases, accelerating their decay and preventing C3b deposition and downstream assembly of the MAC complex (1, 2). CD55 also binds the early leukocyte activation Ag, CD97 (8). In this study, we have shown that CD55 engagement, in combination with CD3 stimulation, induces increased activation of human peripheral blood CD4⁺ T cells.

Our findings, using three plate-bound anti-CD55 mAbs, demonstrate that CD55 costimulation results in enhanced T cell proliferation, cytokine secretion, and up-regulation of the high-affinity IL-2 receptor α subunit CD25. Although stimulation of CD55 alone does not influence T cell responses, in combination with TCR engagement it potentiates this response. This is also supported by studies on a Th1 cell clone, whose activation with CD55/CD3 coengagement resulted in enhanced secretion of IFN-γ but did not alter the cytokine profile. Furthermore, earlier studies showed that cross-linking of CD55 with Ab complexes stimulated T cell proliferation in the presence of submitogenic doses of phorbol esters (15). Similarly, Shenoy-Scaria et al. (16) and Tosello et al. (18) showed that cross-linking CD55 Abs could result in T cell activation and IL-2 secretion.

The stimulation of cells with anti-CD55 Abs may be a result of direct signaling through CD55 (16, 17, 18) or it may be related to their ability to block CD55-mediated complement inhibition (19, 20). Recent studies have shown an enhanced T cell responsiveness in CD55^−/− mice. This effect was abrogated by concomitant deletion of C3 or factor D, indicating that complement itself enhances T cell responses and that the role of CD55 is to prevent complement activation (19, 20) (reviewed in Ref. 28). In particular, Heeger et al. (20) questioned the results observed with cross-linking anti-CD55 Abs (15), suggesting that they were not mediated by direct stimulation of CD55 but rather its neutralization.

Therefore, in this setting, binding of mAbs to CD55 could either block CD55-mediated complement inhibition or have a direct effect on stimulating CD55, or both. Indeed, 791T/36 and BRIC 216 anti-CD55 used in this study do increase complement deposition on cells. However, the nonneutralizing anti-CD55 mAb BRIC 220 does not increase C3b deposition (Ref. 23 and our unpublished observations) but is also able to costimulate CD4⁺ T cell activation. This demonstrates that interference with CD55-mediated complement inhibition is not responsible for the CD55-mediated costimulation observed in this study.

In the present work, we also provide evidence, for the first time, that the ligand for CD55, CD97, is able to costimulate CD4⁺ T cells, elucidating a further role for CD97-CD55 interaction. CD97, an EGF-TM7 family member, is abundantly expressed on most leukocytes upon activation and is constitutively expressed by granulocytes and monocytes (29). Little is known about the function of CD97-CD55 interaction, which has been shown to play an adhesive role in neutrophil migration (30). CD97 also binds chondroitin sulfate and the integrins α₅β₁ and α_vβ₃ (22, 31), but for those interactions the EGF-like domain 4 and the RGD motif in the stalk region, respectively, are essential. To specifically investigate CD97-CD55 interaction, we stimulated CD4⁺ T cells with a CD97-Fc construct that comprises the three EGF-like domains (EGF 1,2,5) necessary for the binding to CD55 only. Plate-bound CD97-Fc showed strong costimulation of T cells and produced a similar cytokine profile, as the cross-linking mAbs. Furthermore, CD97-Fc did not interfere with CD55-mediated complement inhibition, indicating a direct effect on CD55. The specificity of their interaction was confirmed with a blocking experiment, where CD55 was masked with a fluid-phase CD55 mAb, before stimulation with plate-bound anti-CD3 mAb and CD97-Fc.

A fluid-phase anti-CD55 mAb could be used to block CD55, rather than stimulate it, because stimulation may require a number of molecules to be brought together, as with plate-bound Abs or cross-linking with a secondary Ab. This is not unusual, because a similar effect has been observed for other GPI-anchored molecules, like murine Thy-1 (32), and non-GPI-anchored complement regulatory proteins, like the murine Crry (25). For CD55 in particular, cross-linking of CD55 Abs was previously shown to be an essential prerequisite for exerting costimulation (15).

It could be argued that the engagement of CD55 might tighten the immunological synapse and increase signal strength. If this was the case, CD55 costimulation should give the same phenotype as maximal TCR stimulation. In our system, maximal signal 1, obtained with optimal concentrations of anti-CD3 mAb and the presence of exogenous IL-2, did not result in a polarized increase in IL-10 production (data not shown). This implies that the IL-10 generated by CD55 costimulation was the result of additional signaling.

The cytokine profile also supports the hypothesis of direct stimulation of CD55, with up-regulation of IL-10 and no significant increase in IFN-γ. In contrast, in the hyperresponsive CD55^−/− T cells, IL-10 was found to be down-regulated whereas IFN-γ was up-regulated (19). Elimination of C3, as well as CD55, in the double-knockout mice restored normal levels of IFN-γ but not of IL-10, indicating that IL-10 down-regulation is not related to an increase in complement activation but might be related to CD55 downstream signaling. Among T lymphocytes, IL-10 is produced by the T regulatory cells, both the naturally occurring thymic CD4⁺CD25⁺ or those generated in the periphery, the adaptive regulatory T cells such as T-reg 1 and Th3 (33, 34). In CD55-deficient mice, IL-10 reduction did not correlate with differences in thymic regulatory cells (19). However, adaptive regulatory T cells were not investigated.

Increased IL-10 production has also been linked to another human complement regulatory protein with costimulatory properties, CD46. Coengagement of CD46 and CD3 was shown to induce a T-reg 1 phenotype in naive CD4⁺ T cells (35). The IL-10 production observed with naive CD4⁺ T cells further indicates that CD55 costimulation might have a similar role to CD46 in generating adaptive regulatory cells, which would also explain why many different viruses and bacteria use CD46 and CD55 as adhesion and invasion receptors (3, 4, 5, 6, 7, 36, 37, 38, 39, 40). Nonetheless, a possible involvement of CD55 in generating regulatory T cells in the periphery warrants further investigation.

In summary, our study shows that CD55 engagement with its natural ligand CD97 can act as a potent costimulator of human CD4⁺ T cells, resulting in cellular activation and promoting enhanced proliferation and cytokine secretion. This finding elucidates a novel role for CD55, expanding its function beyond complement regulation.

The authors have no financial conflict of interest.