The biological function of 2B4, a CD48-binding molecule expressed on T cells with an activation/memory phenotype, is not clear. In this report, we demonstrate that proliferation of CD8+ T cells is regulated by 2B4. Proliferative responses of CD8+ T cells were significantly reduced by anti-2B4 Ab. The effects were not potentiated by anti-CD48 Ab, suggesting that the observed responses were driven by 2B4/CD48 interactions. Surprisingly, the 2B4/CD48-dependent proliferative responses were also observed in the absence of APCs. This suggests that 2B4/CD48 interactions can occur directly between T cells. Furthermore, when activated 2B4+CD8+ T cells were mixed with 2B4−CD8+ TCR-transgenic T cells and specific peptide-loaded APC, the proliferation of the latter T cells was inhibited by anti-2B4 Ab. Taken together, this suggests that 2B4 on activated/memory T cells serves as a ligand for CD48, and by its ability to interact with CD48 provides costimulatory-like function for neighboring T cells.
The 2B4 molecule (CD244) is expressed on NK cells, monocytes, and basophils, and on subsets of TCRγδ+ T cells and CD8+ T cells (1, 2, 3, 4). Most of the information available on 2B4 comes from studies of NK cells. Cross-linking of 2B4 on IL-2-activated NK cells leads to stimulation of lytic activity (1, 4, 5), IFN-γ secretion (1), and granule exocytosis (6). The biological function of 2B4 on CD8+ T cells remains largely unclear. 2B4+CD8+ T cells have been associated with non-MHC-restricted cytotoxicity (6), but cross-linking of 2B4 on CD8+ T cells does not trigger redirected lysis of FcR-expressing targets, cytokine production, or proliferation (4).
To date, the only known 2B4-binding molecule is CD48. Binding studies have shown that CD48 has a 5–10 times stronger affinity for 2B4 than for CD2 (7, 8). Anti-CD48 Abs have been found to inhibit activation of CD4+ and CD8+ T cells (9, 10). Whereas CD2−/− mice are phenotypically similar to wild-type mice (11), T cells from CD48−/− mice show a severe reduction in proliferation and IL-2 production in response to lectins, anti-CD3 Ab, and alloantigen (8, 12). The relatively normal phenotype of the CD2−/− mice suggests that CD48-binding ligands other than CD2 may compensate for the absence of CD2.
In this report, we demonstrate that 2B4 is associated with an activation/memory phenotype of CD8+ T cells. Furthermore, the results suggest that 2B4 on activated/memory T cells serves as a ligand for CD48, and by its ability to interact with CD48 provides costimulatory-like function for neighboring T cells.
Materials and Methods
Adult C57BL/6 (B6) and TCR-transgenic (Tg)4 (F5) mice specific for an influenza nucleoprotein epitope, ASNENMDAM, presented on H-2Db (13) were bred at the Microbiology and Tumor Biology Center, Karolinska Institutet (Stockholm, Sweden). Animal care was in accordance with national and institutional guidelines.
All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified. The influenza virus nucleoprotein-derived peptide (nucleoprotein (NP)366–374), ASNENMDAM, was purchased from Research Genetics (Huntsville, AL). All Ab used were purchased from BD PharMingen (San Diego, CA) except for anti-CD2 (Southern Biotechnology Associates, Birmingham, AL).
Preparation of CD8+ T cells
CD8α+ cells were purified from spleens of mice using the MACS separation system (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s guidelines and resuspended in complete medium (αMEM, 10 mM HEPES, 2 × 10−5 M 2-ME, 10% FCS, 100 U/ml penicillin, 100 U/ml streptomycin; Life Technologies, Paisley, U.K.) plus recombinant human IL-2 (1000 U/ml; PeproTech, Rocky Hill, NJ) for 6 days. FACS analysis of these cells demonstrated that only CD8+TCR+ cells were present in these cultures. Purification of CD8+ T cells from influenza-infected mice has been described previously (14).
2B4 transcripts were detected using RNA from naive or activated CD8+ T cells by RT-PCR using the following primers: 2B4 short transcripts were detected using 5′-ACTGTTTTTGTCCTGCTTGGTG and 3′-AGGAAACTGGTGGAGAAGAAAA (Cybergene, Stockholm, Sweden) encompassing nt 573-1093 of murine 2B4 (m2B4) short. m2B4 long transcripts were detected using 5′-TGGAAGAAGACAGAACAGGG and 3′-TGGAATCAGAAGGCTTGCAC encompassing nt 283-1136 of m2B4 long. cDNA quality was confirmed by amplification of a β-actin gene fragment.
CD8+ T cell proliferation assays
Bone marrow-derived dendritic cells (DC) were pulsed with ASNENMDAM peptide overnight at 37°C and seeded at 2 × 104 cells/well in U-bottom 96-well plates. A total of 3 × 104 TCR-Tg CD8+ T cells (naive or IL-2-activated) or TCR-Tg lymphokine-activated T (LAT) cells (FACS sorted into NK1.1+ or NK1.1−) were added to each well with the indicated azide-free Ab. When B6 LAT cells were mixed with naive TCR-Tg CD8+ T cells, 4 × 104 B6 LAT cells (sorted into NK1.1+ or NK1.1−) were added with 3 × 104 naive TCR-Tg CD8+ T cells to each well. In APC-free experiments, plates were coated overnight at 4°C with purified Db-NP366 monomers (14, 15) at various concentrations with anti-CD28 Ab (10 μg/ml) in PBS. TCR-Tg LAT cells (3 × 104) were added to each well. In cytokine-induced proliferation assays, 3 × 104 B6 CD8+ T cells (naive or IL-2-activated) and TCR-Tg LAT cells (sorted into NK1.1+ or NK1.1−) were stimulated with IL-2 (1000 U/ml). All proliferation assays were for 72 h except the naive B6 CD8+ T cell cytokine-induced proliferation assay, which was for 120 h. [3H]Thymidine (1 μCi/ml) was added for the last 18 h of all assays. The plates were harvested on glass fiber filters (Wallac, Turku, Finland) and analyzed in a beta-scintillation counter (Wallac).
Results and Discussion
CD8+ T cells acquire expression of 2B4 upon activation in vitro and in vivo
One to 2% of all splenic CD8+ T cells from normal uninfected mice express 2B4. In the present study, we show that activation by IL-2, IL-4, and IL-15 induced expression of 2B4 on CD8+ T cells (Fig. 1,A and data not shown). CD8+2B4+ T cells were predominantly CD44high and CD62Llow (data not shown), in line with recent findings showing that expression of NK cell receptors on CD8+ T cells is associated with a memory phenotype (16, 17). Alternative splicing of the 2B4 mRNA has been shown to give rise to two different polypeptides, 2B4 long and 2B4 short, differing in their cytoplasmic tails while having identical extracellular domains (18, 19). As shown in Fig. 1,B, both the short and long forms of 2B4 were equally up-regulated in CD8+ T cells after stimulation with IL-2. Taken together, the data indicated that expression of 2B4 was increased in IL-2-activated CD8+ T cells both at the cell surface and at the transcriptional level. 2B4 expression was also observed on CD8+ T cells on day 10 post-influenza infection (peak of infection). Five to 10% of all CD8+ T cells in the lungs of infected mice expressed 2B4 compared with 1–2% in untreated mice (Fig. 1,C). The expression of 2B4 on virus-specific T cells was confirmed using tetramers of MHC class I molecules refolded with an influenza nucleoprotein epitope (Db-NP366) (Fig. 1 C). Although 2B4 was observed also on tetramer− T cells, this could be related to the other CD8+2B4+ T cells recognizing other flu epitopes, or bystander activation, or both.
2B4/CD48 interactions are involved in Ag-specific and cytokine-induced proliferation of preactivated CD8+ T cells
The effect of anti-2B4 Ab on the proliferation of CD8+ T cells was examined using CD8+ T cells from mice transgenic for a TCR specific for an influenza nucleoprotein epitope. The Ag-specific response of purified naive TCR-Tg CD8+ T cells was compared with that of TCR-Tg CD8+ T cells which were first cultured in IL-2 for 6 days (hereafter denoted as CD8 LAT cells). In response to peptide-pulsed DCs, anti-2B4 Ab had no effect on the proliferation of naive CD8+ T cells (Fig. 2,A). However, using TCR-Tg LAT cells, the proliferative response to peptide-pulsed DCs was significantly diminished by anti-2B4 Ab compared with isotype control Ab (Fig. 2 B).
The effect of anti-2B4 Ab was also compared with that of anti-CD48 and anti-CD2 Ab. Addition of anti-CD48 or anti-CD2 Ab resulted in ∼45 and ∼20% reduction in the Ag-specific proliferation of naive TCR-Tg CD8+ T cells at peptide concentrations of 0.01 μM and 0.1 μM, respectively (Fig. 2,A). In contrast, the Ag-specific proliferation of TCR-Tg CD8 LAT cells was suppressed by anti-CD48 but not by anti-CD2 Ab (Fig. 2,B). However, both anti-CD48 and anti-2B4 Ab caused a similar degree of inhibition (∼45%) in CD8 LAT cell proliferation at all peptide concentrations. Interestingly, the proliferation of CD8 LAT cells could not be further inhibited by anti-2B4 Ab in the presence of anti-CD48 Ab, suggesting that the effects of anti-2B4 Ab were mediated through blocking of the 2B4/CD48 interaction (Fig. 2 C). CD48 has been shown to have a 5–10 times stronger affinity for 2B4 than for CD2 (7, 8). Thus, as 2B4 expression was increased on activated CD8+ T cells, it was likely that the CD2/CD48 interaction was replaced by 2B4/CD48 interactions. This notion is supported by the fact that anti-2B4 Ab blocks the Ag-specific proliferation of CD8 LAT cells but not of naive CD8+ T cells and that the opposite is true for anti-CD2 Ab.
It has recently been suggested that the proliferation of memory T cells is MHC independent and probably regulated by cytokines (20). Because CD8+2B4+ T cells in normal mice were primarily of memory phenotype, we examined whether 2B4 also played a role in the cytokine-induced proliferation of CD8+ T cells. In line with the above observations, a 40–50% reduction in proliferation of naive B6 CD8+ T cells was observed with anti-CD2 or anti-CD48 Ab, whereas no inhibition in proliferation was seen with anti-2B4 Ab in response to IL-2 (Fig. 2,D). However, a 20% decrease in the IL-2-stimulated proliferation of B6 CD8 LAT cells was observed in the presence of anti-2B4 or anti-CD48 Ab but not anti-CD2 Ab (Fig. 2,D). Furthermore, taking advantage of the fact that NK1.1 is coexpressed with 2B4 (data not shown), B6 CD8 LAT cells were sorted into NK1.1+ and NK1.1− subsets. The addition of anti-2B4 or anti-CD48 Abs to NK1.1+ CD8 LAT cells resulted in an ∼40% attenuation of the IL-2-induced proliferation while no effect was observed by the addition of the same Ab to NK1.1− CD8 LAT cells (Fig. 2 D). Taken together, these results suggested that non-MHC-restricted routes of T cell proliferation can also be regulated by 2B4/CD48 interactions and that such interactions may be involved in the propagation of memory T cells.
The 2B4/CD48 interaction takes place between CD8+ T cells
Initially, we believed that the 2B4/CD48 interaction took place between the APC and T cell, because others have reported that 2B4 cross-linking does not directly activate CD8+ T cells and 2B4 was presumed to function as an adhesion molecule for T cells to APCs (4, 21). However, as observed from the inhibition of IL-2-induced proliferation of CD8 LAT cells by anti-2B4 Ab, it was apparent that the 2B4/CD48 interaction did not require APCs. Moreover, we found that both anti-2B4 and anti-CD48 Ab inhibited the Ag-specific proliferation of CD8 LAT cells in the absence of APCs where proliferation was induced by cross-linking of the TCR using purified MHC class I-peptide complexes (Fig. 3 A). These data suggested that the 2B4/CD48 interaction does not necessarily take place at the T cell/APC interface, but can also occur between neighboring CD8+ T cells.
These results led to the question of whether 2B4 was directly triggering proliferation of 2B4+ CD8 LAT cells or was affecting the proliferation of T cells by virtue of its ability to function as a ligand for CD48. To answer this question, we first looked at which population of cells, 2B4− or 2B4+, was proliferating in the CD8 LAT cell/DC cocultures. To do this, TCR-Tg CD8 LAT cells were sorted into NK1.1+ and NK1.1− subsets and cocultured with peptide-pulsed DCs. Of the two subsets, only the NK1.1− (2B4−) subset proliferated in response to specific Ag (Fig. 3,B). In contrast, both subsets proliferated equally well to IL-2, indicating that the NK1.1+ cells were not exhausted but were just unresponsive to peptide (data no shown). This was also confirmed by FACS analysis of CFSE-labeled CD8 LAT cells, which showed that only the 2B4− cells proliferated after a coculture with peptide-pulsed DCs (data not shown). Thus, from these results it was evident that the anti-2B4 Ab was not inhibiting the proliferation of 2B4+CD8+ T cells, because these cells did not proliferate in response to Ag. More compelling evidence came from experiments where B6 CD8 LAT cells were sorted into NK1.1+ (2B4+) and NK1.1− (2B4−) subsets and mixed with naive TCR-Tg CD8+ T cells. The addition of anti-2B4 Ab to B6 NK1.1+ CD8 LAT/TCR-Tg cocultures resulted in significant inhibition of Ag-specific proliferation, whereas no effect was observed by the addition of the same Ab to NK1.1− CD8 LAT cells (Fig. 3, C and D). The addition of anti-CD48 Ab to either B6 NK1.1+ CD8 LAT/TCR-Tg or NK1.1− CD8 LAT/TCR-Tg cocultures inhibited Ag-specific proliferation (Fig. 3 D). This inhibition was related to the interaction of CD2 expression on the naive TCR-Tg, as anti-CD2 Abs could also inhibit in both cocultures (data not shown).
This suggested that the 2B4 expression on nonspecific CD8+ T cells (B6 CD8 LAT cells in this case) affects the proliferation of neighboring naive 2B4− Ag-specific CD8+ T cells (TCR-Tg CD8+ T cells), because the B6 CD8 LAT cells did not proliferate in response to specific Ag (Fig. 3, C and D). Furthermore, anti-2B4 Ab had no effect on the proliferation of T cell/DC cultures when the B6 CD8 LAT cells were separated from the naive TCR-Tg CD8+ T cells by a cell-impermeable porous membrane, indicating that cell-cell contact was required (data not shown). Taken together, these results strongly suggest that 2B4 functions as a ligand for an activating receptor on CD8+ T cells, most likely CD48, rather than being an activating receptor itself. Although CD48 is a GPI-anchored receptor and lacks intracytoplasmic domains, it has been shown to be physically associated to G proteins (22) or to the SRC family member of tyrosine kinases in glycolipid-enriched microdomains of the cell membrane (23). Thus, it is not surprising that a signal is transduced by CD48 upon ligation with 2B4 and can affect CD8+ T cell proliferation. However, exactly what signals lead to the enhancement of proliferation by 2B4/CD48 interactions is an issue that remains to be established.
In summary, we have shown that 2B4 plays a significant role in the Ag-specific and cytokine-induced proliferation of CD8+ T cells. Although only a small proportion of CD8+ T cells normally express 2B4 in the mouse, this molecule may play a significant role in the propagation of T cells through interaction with CD48. We speculate that 2B4 could provide costimulatory signals necessary for the survival of memory T cells in response to low levels of Ag or cytokines.
We thank Dr. T. N. M. Schumacher for providing us with the H-2Db H chain construct, Dr. P. J. Travers for the murine β2-microglobulin construct and Dr. D. Kioussis for the TCR-Tg (F5 transgenic) mice. We also thank members of the H.-G. Ljunggren laboratory for fruitful discussions.
This work was supported by the Swedish Foundation of Strategic Research, the Karolinska Institutet, the Swedish Medical Research Council, the Swedish Cancer Society, the Tobias Foundation, the Åke Wiberg Foundation, the Alex and Eva Wallström Foundation, and the Lars Hiertas Foundation.
Abbreviations used in this paper: Tg, transgenic; DC, dendritic cell; LAT cell, lymphokine-activated T cell; NP, nucleoprotein; m2B4, murine 2B4.