Cutting Edge: TCR Signaling Induces Selective Exclusion of CD43 from the T Cell-Antigen-Presenting Cell Contact Site¹

The CD43 protein (leukosialin, sialophorin) was first described by Williams et al. (1) as a large sialoprotein expressed highly on rat thymocytes. CD43 is a member of a family of mucin-like proteins that includes CD34, CD68, CD45, CD96, CD162 (P-selectin glycoprotein ligand-1 (PSGL-1)), glycosylation-dependent cell adhesion molecule-1 (GlyCAM-1), and human platelet glycoprotein GPIbα (2). CD43 is highly expressed by T cells, monocytes, platelets, and neutrophils, and, to a lesser extent, by B cells, bone marrow precursors, and various other hemopoietic cells (3). CD43 has been found to extend out from the cell membrane 45 nm, making it one of the largest cell surface molecules on T cells (4). Due to its large size and abundance, CD43 has been estimated to cover 28% of the T cell surface.

The abundance and heavily glycosylated nature of CD43 has led to the hypothesis that CD43 provides a highly charged barrier to cellular interactions (1, 5, 6). This hypothesis has been supported by the studies of several research groups. It has been found that, when CD43-deficient T cells are compared with CD43 normal controls, they are more susceptible to cytotoxicity and spontaneous homotypic adhesion (6) and are hyperresponsive to activating reagents such as superantigen, Con A, and anti-CD3 stimulation (7). Conversely, overexpression of CD43 gives the opposite phenotype (8, 9). Transgenic mice that overexpress CD43 on all B cells and transgenic mice that express only the most highly glycosylated form of CD43 have been found to be hyporesponsive to stimulation (8, 9, 10). Taken together, these results suggest that overexpressing CD43 inhibits immune responses while underexpression of CD43 leads to increased immune responses.

There is also evidence, however, that CD43 plays a positive role in T cell activation (11, 12, 13). Cross-linking CD43 with Abs to either human or rat CD43 has been reported to enhance T cell responses induced by either mitogenic or allogenic stimulation (11, 14, 15). We have extended these results to establish that anti-CD43 alone, in the absence of any other costimulatory interactions, can provide a “second signal” necessary to induce T cell proliferation to suboptimal anti-CD3 stimulation (11). Indeed, direct cross-linking of human CD43 with certain CD43-specific mAbs can directly activate T cells in the absence of Ag (13, 16). If any of these effects are physiological, this would suggest that a natural ligand for CD43 exists. Several potential ligands for CD43 have been reported: ICAM-1 (17), Galectin-1 (18), MHC Class I (19), human serum albumin (20), and E-selectin (21). Despite these reports, no ligand has been shown to have a clear physiological effect on CD43. Thus, the question of the existence of a CD43 ligand remains unanswered.

Recently a topological view of T cell-APC interactions has been proposed (22, 23). In this model, molecules such as CD43 and CD45 are excluded from the T cell-APC interaction site due to their large size and highly glycosylated nature. This study represents the first direct test of this hypothesis. We demonstrate for the first time that CD43 is specifically excluded from the T cell-APC contact site. Optimal exclusion requires an Ag-dependent signal. Interestingly, CD3 signaling is sufficient to induce CD43 modulation in the absence of all other receptor-ligand interactions; thus, a putative CD43 ligand interaction is not required. In contrast, another large highly glycosylated molecule, CD45, is not excluded from the interaction site. Together, these data suggest that CD43 may represent a new type of cell surface regulatory protein that dampens T cell responses by its physical presence and is specifically removed upon T cell activation.

Conjugates were made with the OVA_323–339-specific Th1 T cell clone pGL2 (gift of Dr. Frank Fitch, Chicago, IL) and the B lymphoma A20. The A20 cells were pulsed for 2 h with 20 μg/ml OVA_323–339, washed, and combined with T cells at a 1:1 ratio. Conjugates were formed by centrifugation at 500 rpm for 5 min, and the pellets were incubated at room temperature for 10 min. The cells were gently resuspended and allowed to adhere to poly-l-lysine-coated glass slides for an additional 25 min at 37°C.

Polystyrene latex beads (Interfacial Dynamics, Portland, OR) were coated with either anti-CD3 (145-2C11; 24 or anti-H-2L^d (28.14.8 or 30.5.17; 25 at 100 μg/ml. To approximate the size of the T cells, 6-micron beads were used. After washing and blocking, 2 × 10⁵ beads were incubated with 2 × 10⁵ T cells in a volume of 100 μl for 5 min and allowed to adhere to poly-l-lysine-coated glass slides for an additional 10–25 min at 37°C. T cell-bead conjugates were scored when, during microscopic examination, adjusting the plane of focus revealed that the cell and the bead formed a clear interaction site. Those instances where it appeared that the bead was only resting on the T cell were not scored.

The slides were rinsed in PBS and fixed in 3% paraformaldehyde/PBS for 20 min. The paraformaldehyde was quenched with 50 mM NH₄Cl/PBS, the cells permeabilized for 1 min in 0.3% Triton-X100, and the cells were blocked with a solution of 0.01% saponin and 0.25% fish skin gelatin (Sigma, St. Louis, MO). Affinity-purified anti-mouse CD43 (S7; a gift of Dr. John Kemp, University of Iowa School of Medicine, Iowa City, IA), anti-CD45 (30-F11; PharMingen, San Diego, CA), and anti-talin (8d4; Sigma) were used to stain the conjugates. The primary Abs were detected by polyclonal FITC donkey anti-rat Ig and Texas Red donkey anti-mouse Ig (Jackson ImmunoResearch, West Grove, PA) respectively. The primary mAbs were incubated with the fixed cells for 30 min. After washing, the slides were incubated for 30 min with the fluorochrome-labeled secondary Ab, washed again, and mounted with coverslips in Mowiol (Hoechst Celanese, Charlotte, NC) containing 10% w/v 1,4-diazobicyclo-(2, 2, 2) octane as antifade. The immunofluorescence and corresponding DIC³ images of the cells were recorded by a cooled CCD camera (PXL, Photometrics, Tucson, AZ) that was mounted on a Zeiss Axioplan microscope, equipped with narrow-band optical filters (Chroma, Brattleboro, VT). Single images were processed using the Openlab deconvolution module (Improvision, Coventry, England) to remove out-of-focus haze.

It has been hypothesized that, because of their large size and strong negative charge, CD43 and CD45 may first have to move out of the interaction site between the T cell and APC before full activation occurs (22). To test this hypothesis, we used immunofluorescence microscopy to localize CD43 and CD45 during T cell-APC interactions. Ag-specific T cell-APC conjugates were formed by mixing the T cell clone pGL2 with the B lymphoma A20 that had been previously pulsed with the OVA_323–339 peptide Ag. Conjugates were double stained for the cytoskeletal protein talin and either CD43 or CD45. Talin translocation to the T cell-APC contact site is a hallmark of Ag-specific conjugate formation (26). While CD43 was equally distributed in a continuous ring on the cell surface of unbound T cells (data not shown), CD43 was absent from the interaction site of T cell-APC conjugates (Fig. 1,A, middle image) and seemed to be excluded specifically from the talin cluster regions (Fig. 1,A, right image). When CD43 exclusion was seen, we often noticed a concentration of CD43 in the areas directly adjacent to the interaction site (Fig. 1,A, arrows). This concentration may represent the movement of CD43 from the interaction site to the edges. In five separate experiments, CD43 removal from the interaction site was seen in 53.6 ± 3.1% of the conjugates scored. When the scoring was restricted to conjugates that demonstrated talin polarization, CD43 modulation was found in 68.6 ± 2.4% of the conjugates (Fig. 2).

In contrast to CD43, CD45 does not move away from the interaction site. In the majority of talin-polarized conjugates scored, a continuous fluorescence staining was seen all around the cell surface (Fig. 2 and data not shown). Interestingly, CD45 was often found to polarize toward the interaction site (Fig. 1,B, middle image; Fig. 2). Thus, while CD43 seems to be excluded from the interaction site, CD45 either remains uniformly distributed or is polarized into the interaction site. These data demonstrate for the first time that CD43 is selectively excluded from the T cell-APC interaction site.

To examine the signals involved in CD43 exclusion from the interaction site, we first tested the Ag dependence of CD43 modulation. In the absence of an Ag-specific signal (non-Ag-pulsed A20), most of the conjugates did not polarize talin (data not shown). When total conjugates, Ag-pulsed and non-Ag-pulsed, were compared, the Ag-specific conjugates were found to be more effective at excluding CD43 from the interaction site (Fig. 3). These results suggest that T cell activation is involved in effective exclusion of CD43 from the interaction site.

The experiments shown in Fig. 3 suggest that TCR signal transduction plays a role in CD43 modulation. However, these data do not address whether TCR signaling alone can induce CD43 modulation or whether TCR signaling is modulating CD43 through activation of other cell surface receptor-ligand pairs. To examine these questions, conjugates were made between T cells and anti-CD3-coated polystyrene beads. Efficacy of the beads was determined by two parameters. First, as previously shown (27, 28), the anti-CD3-coated beads, but not anti-Class I-coated beads, were capable of inducing polarization of the microtubule organizing center (MTOC) toward the bead (71.5% with anti-CD3-coated beads vs 39.3% with the anti-Class I-coated beads). Second, anti-CD3-coated beads were able to induce the proliferation of DO.11.10 TCR transgenic (Tg) cells. This proliferation was comparable to the proliferation induced by 0.3 μg/ml OVA_323–339 peptide presented by IA^d+ splenocytes (data not shown). As seen in Figs. 4 and 5, CD43 was effectively excluded from the T cell-bead contact site when anti-CD3-coated beads were used but not with anti-MHC class I-coated beads. Moreover, when the same conjugates were stained for CD45, only background levels of CD45 modulation from the interaction site was found (Fig. 5). Interestingly, unlike in the T cell-APC conjugates, the number of T cell-bead conjugates in which CD45 moved into the contact site was negligible (<5%). Together, these data suggest that TCR signaling is not sufficient to induce CD45 polarization into the contact zone but is sufficient to induce CD43 exclusion.

Unlike other receptors that are involved in T cell activation and that cluster at the T cell-APC contact site (29), we have found that CD43 normally moves out of the T cell-APC interaction site during T cell activation. It is likely that CD43 modulation occurs after the initial activation of the T cell since a significant number of conjugates had talin polarization, a sign of activation, but had not modulated CD43 (Fig. 2). We propose that CD43 negatively regulates T cell interaction by deterring cell-cell interactions. If, despite this obstacle, a T cell encounters an appropriate APC and undergoes the initial steps of activation, CD43 is actively removed from the contact site. By this mechanism, T cells that are not activated upon initial contact with the APC may more easily detach and move on to the next APC.

The mechanism by which CD43 is modulated from the interaction site is not clear at this time. However, we propose that TCR binding to Ag/MHC complexes may send a signal to the CD43 molecule that directs its movement away from the interaction site. Our data show that TCR signal transduction is sufficient to induce CD43 exclusion from the T cell-APC interaction site and argues against a role for a putative CD43 ligand. In keeping with our findings, Sanchez-Madrid et al. observed that CD3 ligation greatly increased anti-CD43-mediated movement into uropods (30). Together, these data suggest that signals through CD3 may play an important role in regulating CD43 movement.

Several groups have suggested that large, highly glycosylated molecules such as CD43 and CD45 would be naturally excluded from interaction sites (22, 23). This model was particularly attractive for CD45, which is a tyrosine phosphatase involved in regulating TCR signal transduction. One could imagine that CD45 exclusion from the interaction site might be important for allowing TCR signaling to occur unencumbered by the phosphatase activity of CD45. However, our results do not support this model. We have found that CD45 is not excluded from the interaction site, and frequently it is polarized into the site. Our data are consistent with the findings of Bottomly and colleagues that several isoforms of CD45 specifically associate with the TCR and CD4 in cocapping experiments (31). Since the pGL2 cells express only the smaller isoform of CD45 (data not shown), it is possible that the larger isoforms could be excluded from the interaction site. However, our data demonstrate that CD45 exclusion is not a necessary event for T cell activation. While the mechanism for CD45 polarization is not clear, our data suggest that CD3 signaling is not sufficient to induce CD45 movement into the contact zone. Since by contrast we find that TCR signaling is sufficient for CD43 exclusion, the signals involved in CD43 and CD45 movement are likely to be distinct.

In summary, we have demonstrated that CD43 moves out of the T cell-APC interaction site during T cell activation. This movement is not due to size and charge alone but probably involves specific intracellular signals. As a moveable barrier, CD43 may limit nonspecific T cell interactions while allowing Ag-specific interactions. Elucidating the mechanism by which CD43 movement is regulated will be important for understanding T cell interactions with other cell types.

We thank Catlin Sedwick, Alfred Chak, Richard J. DiPaolo, Amelie Collins, Dr. Jim McIlvain, Director of the Ben May Institute/Department of Pathology Digital Light Microscopy Facility, and Julie Auger, Director of the University of Chicago Cancer Research Center Flow Cytometry Facility, for their technical advice and assistance. We also thank Drs. Jeffrey A. Bluestone and Jim Miller for critical reading of the manuscript.