Abstract
Unless a costimulatory signal is provided, TCR recognition of Ag bound to the MHC is insufficient to induce optimal T cell proliferation or the production of IL-2. Here we show that the stimulation of CD28, a T cell costimulatory receptor, by a specific Ab increases F-actin contents in T cells. The interaction between T cells and B7–2-transfected Chinese hamster ovary cells expressing the CD28 ligand leads to the rearrangement of the actin cytoskelton in the region of cell-cell contact. Within the Rho family of G proteins, Rac1, but not Rho, translocates to the sites of cell-cell contact where Tailin also accumulates. These results indicate that the interaction between B7–2 and CD28 establishes a focal adhesion-like cell contact between T cell and APCs. The results also suggest that CD28 signaling is primarily transduced by a cytoskeletal rearrangment/signaling pathway mediated by the Rho family G proteins.
CD28 is a 44-kDa homodimeric glycoprotein membrane-expressed by most mature T cells (1). The engagement of CD28 with the natural ligands B7–1 or B7–2 (2), or with a specific Ab, in conjunction with TCR-mediated stimulation causes a great increase in both proliferation and IL-2 production by T cells (3). In spite of a powerful influence on T cell activation, CD28 stimulation alone does not trigger detectable cell proliferation or IL-2 production. The existence of a phosphotyrosine-based motif pTyr-Met-Asn-Met (pYMNM) within the cytoplasmic tail suggests the recruitment of phosphatidylinositol 3-kinase. The pYMNM motif is also capable of binding to the GRB-2 SH2 domain, thus suggesting CD28 can complex with GRB-2-Sos complex and activates p21ras-mediated signaling (4). In fact, p21ras activation was observed in T cell lines stimulated with anti-CD28 Ab, although B7–1 did not trigger p21ras activation. Additionally, activation of 70 kDa S6 kinase (5) is reported and a T cell-specific protein tyrosine kinase, ITK, also binds to pYMNM motif and was activated in CD28-stimulated T cells (6, 7). We and others observed that CD28 stimulation triggers hydrolysis of sphingomyelin that generates ceramide in T cells (8, 9). Although the activation of various signaling molecules are reported in CD28-stimulated T cells, the CD28-coupled costimulatory signaling pathway is not fully understood.
The Rho family of G proteins, which consist of CDC42, Rac, and Rho are known as molecular signaling switches, which regulate various cellular functions. Currently, these G proteins are reported to be key in organizing two different but perhaps mutually interconnected functions. Their activity is essential for the organization of the actin cytoskelton (10, 11). Specifically, Rac regulates the formation of lamellipodia and membrane ruffles (12), Rho is required for the formation of focal adhesions and actin stress fibers (13), and CDC42 induces the formation of filopodia (14, 15). They also function to transduce signals from the cellular membrane to the nucleus by stimulating kinase cascades (16). A role of CDC42 in the polarization of T cells against APCs has been reported (17), although the functional role of these G proteins in lymphocytes remains largely elusive.
We show in this report that CD28, upon ligation with specific ligands, stimulates polymerization of actin. The stimulation with the natural ligand B7–2 results in the formation of cell-cell contact points in T cells, at the site of Rac and Talin accumulation. F-actin also localized closely to G proteins at the cell-cell contacts. The data suggest that the CD28 signal is primarily transduced by the cytoskeletal signaling that is mediated by Rho family G proteins. The potent costimulation induced by CD28 signaling may be in part due to a CD28-mediated enhancement of adhesion between T cell and APC.
Materials and Methods
Antibodies, proteins, and chemicals
Anti-RhoA, anti-Rac1, FITC-goat F(ab′)2 anti-rabbit IgG, FITC-goat F(ab′)2 anti-mouse IgG, and horseradish peroxidase conjugated anti-mouse IgG was purchased from SantaCruz Biotechnology; anti-Talin (8d4) and anti-Actin (AC-40) from Sigma. The Ab to mouse CD28 (37.51) were purified from culture supernatants. TRITC-labeled phalloidin was obtained from Sigma.
Mice
BALB/c ByJ (5–7 wk of age) were purchased from The Jackson Laboratory and were maintained in the conventional animal housing facility of the University of Western Ontario.
Cell culture and cell stimulation
EL4 cells were obtained from the American Type Culture Collection. Cells were grown in RPMI 1640 with 5% FCS (FCS) and 40 μg/ml gentamicin. EL4 cells were stimulated with anti-CD28 (final concentration 15 μg/ml) without cross-linking by secondary Ab, or treated with PMA in FCS-free RPMI 1640 as indicated.
Measurement of F-actin by immunoblotting
Triton X-100 insoluble F-actin was fractionated essentially as described (18, 19). Briefly, two million cells were sedimented and resuspended in Triton-PHEM buffer (0.75% Triton X-100, 60 mM PIPES, 25 mM HEPES, 10 mM EGTA, 2 mM MgCl2, 1 mM PMSF, 20 μg/ml leupeptin, and 80 μg/ml aprotinin) at 4°C, and allowed to incubate on ice for 20 min. The insoluble fraction, which contains Triton X-100-insoluble F-actin was sedimented by spinning at 12,000 g for 5 min. The pellet was dissolved in SDS sample buffer, boiled for 10 min, and analyzed by SDS-PAGE. After transferring the proteins to a PVDF membrane, the membrane was probed with mAb to actin and horseradish peroxidase-conjugated anti-mouse IgG, and visualized using the ECL system (Amersham).
Fluorescence and confocal microscopic analysis
Single-cell suspension of mesenteric lymph node T cells from BALB/c ByJ mice were prepared. T cells were enriched (over 95% cells were CD3ε positive) by a subtraction of anti-mouse Ig-reactive cells using Fe2O3 beads coated with anti-mouse IgG (DYNAL). Bound cells were removed using a magnetic stand (Advanced Magnetics). CHO and CHO-B7–2 were cultured to subconfluence in RPMI 1640 with 5% FCS (FCS) and 40 μg/ml gentamicin on 22-mm square microscope cover slips in 35-mm diameter petri dishes. The culture medium was removed and a suspension of lymph node T cells (2 × 106) in 100 μl RPMI 1640 was overlaid on the monolayer of cells. Cells were incubated for 30 min at 37°C and were stimulated, washed, permeabilized, and fixed using previously reported methods (13). Cells were then stained with TRITC-phalloidin, and incubated for 30 min at 20°C with various specific Abs at 1:20 dilution (overnight incubation at 4°C was performed for Talin staining), followed by staining for 1 h with FITC-goat F(ab′)2 anti-rabbit IgG or FITC-goat F(ab′)2 anti-mouse IgG at 1:20 dilution.
Samples were examined using a Zeiss Axioscope microscope equipped with epifluorescence filters or with a Zeiss LSM 410 confocal microscope equipped with a krypton/argon laser and the appropriate filters for distinguishing FITC and TRITC fluorochromes.
Results
CD28 ligation induces actin polymerization in T cells
An increasing number of reports have demonstrated that many surface receptors (e.g., the EGF receptor) interact with actin-based cytoskeltal components that, upon activation, induce actin polymerization (20). It has been demonstrated that numerous signal transduction molecules associate with cytoskeltal molecules (e.g., via SH3 domain) (21). It is therefore hypothesized that cytoskeletal networks have an important role in intracellular signal transduction. This fact prompted us to measure the levels of actin polymerization in T cells following CD28 stimulation. Western blotting using an actin-specific mAb (18) demonstrated a significant increase in the amount of F-actin in the CD28-stimulated mouse T cell line, EL4 (Fig. 1⇓). EL4 cells were stimulated, as a positive control, with phorbol 12-myristate 13-acetate (PMA), which also induces actin polymerization in T cells (22). The increase in F-actin levels peaked at 5 min following CD28 stimulation and remained at that level for the full 15 min of the assay.
Stimulation of CD28 causes actin polymerization in T cells. EL4 cells were stimulated for 15 min with various reagents as denoted above each lane. Triton X-100 insoluble fractions (lanes 1–4) and Triton X-100 soluble fractions (lanes 5–8) were analyzed by SDS-PAGE. Actin bands were visualized by Western blotting using a mAb to actin. In lane 9, EL4 cells were resuspended in loading buffer to extract whole actin and this lane serves as a positive control for a singular actin band. Extracts from nonstimulated cells are indicated as Control. Protein from identical numbers of cells was loaded onto each lane (lanes 1–8).
Ligation of CD28 with B7–2 induces the formation of cell-cell adhesions associated with an accumulation of Rac and Talin
Studies in fibroblasts have shown that actin polymerization is a basic component in the formation of filopodia, lamellipodia, and stress fibers. Moreover, these morphologic changes depend on Rho family G proteins and, since CD28 stimulation results in substantial F-actin formation, it is probable that CD28-costimulated T cells exhibit some of these same phenotypic alterations as previously observed in fibroblasts. To test this we utilized transfected Chinese hamster ovary (CHO) cells expressing mouse B7–2 (CHO-B7–2) (23). We plated freshly isolated mouse T cells onto both wildtype CHO and CHO-B7–2 transfectants and incubated the cells at 37°C. Following incubation, the cells were fixed and stained with TRITC-phalloidin to determine the distribution of F-actin in all three cell types. Remarkably, T cells incubated with CHO-B7–2 (Fig. 2⇓, c and d) demonstrated a more intense F-actin staining pattern than those incubated with wildtype CHO cells. (Fig. 2⇓a, b). Focal staining at the site of cell-cell adhesions was evident in more than 40% of T cells in contact with CHO-B7–2. Furthermore, we often found T cells which manifested ruffled membranes, and formed pseudopods as shown in Fig. 2⇓d. The pseudopod formation was always observed at the contact sites where F-actin accumulated. These molecular changes were inhibited when CHO-B7–2 was pretreated with anti-B7–2 Ab (data not shown). Although a similar number of T cells attached to wildtype CHO cells, the intensity of the staining and the number of cells with focal staining at the site of cell-cell adhesions was insignificant.
Distribution of F-actin, Rac, Rho, and Talin in fresh T cells cultured with CHO or CHO-B7–2 cells. Small round lymph node T cells were overlaid onto a monolayer culture of fibroblast-like CHO or CHO-B7–2 cells. CHO cells were used in a, b, e, f, j, and k. CHO-B7–2 cells were used in others. Cells were stained for F-actin with TRITC-phalloidin alone (a-d) and examined by fluorescence microscopy. Alternatively, cells were double labeled with TRITC-phalloidin and Abs specific for Rac1 (e-j), Rho (j-n), and Talin (o-q) using FITC-labeled secondary Abs and examined using the confocal microscope. F-actin staining is shown in red. Green fluorescence indicates: Rac in f, h, i, and Rho in k, m, n, and Talin in p, q. The combined red and green images are shown in i, n, and q to show colocalization. Size bars indicate 10 μm.
To induce membrane ruffling or to stimulate downstream signaling, Rho family G proteins translocate to the plasma membrane from the cytosol (24, 25) where they dissociate from GDP and a regulatory protein known as RhoGDI to complex with GTP (26). Accordingly, to determine the identity of Rho family G proteins involved in the formation of cell-cell adhesions we examined the translocation of Rac, Rho, and the cytoskeletal protein Talin, which has been shown to accumulate in the focal adhesions of adherent cells (27). As shown in Fig. 2⇑, e-q, the distribution of F-actin and Rac (Fig. 2⇑, e-i) or Talin (Fig. 2⇑, o-q) overlapped at the region of contact between CHO-B7–2 and T cells. The distribution of CDC42 and its co-localization with F-actin was similar to that of Rac (data not shown). Rho distribution was dissimilar to that of the other Rho family G proteins in that Rho was evenly distributed in CD28-stimulated T cells (Fig. 2⇑j-n). These molecular changes became evident 10 min after the initiation of T cell–CHO-B7–2 co-incubation. The data strongly indicate that CD28 stimulation induces edge ruffles and focal-like adhesions in which Rac1 and CDC42 but not Rho play an integral part.
Discussion
G proteins in a GTP-bound form function to transduce signals from the cellular membrane to the nucleus by stimulating kinase cascades (16), and several kinases have been reported as the immediate targets of these G proteins. Rac and CDC42 directly bind to and activate the p21-activated kinases (PAKs) (28). Upon activating PAKs, Rac appears to regulate the oxidative burst of phagocytic leukocytes (29). Rho associates with several kinases including protein kinase N (PKN), whose downstream pathway is not yet known (30, 31). CDC42 and Rac regulate transcription factors by activating both JNK/SAPK and p38 MAP kinase (32, 33, 34, 35) via PAK/MEKK3 activation. These kinase cascades also appear to control cell cycle progression (36, 37). These findings suggest that CD28 activates PAK signaling through a Rac1 or CDC42-coupled mechanism. Our results demonstrate that CD28 signal stimulates F-actin formation in T cells in which the Rac1 small G protein appeared to have an integral function. In fact, we have found that the stimulation with CD28-specific Ab of a 32P metabolically labeled human T cell line, Jurkat cells, induced turnover of GDP-Rac1 ↔ GTP-Rac1 resulting in a large increase of the amount of [32P]GDP-Rac1 (data not shown). Strikingly, in further experiments we detected an activation of PAK in CD28-stimulated T cells (Kaga et al., manuscript submitted.), thus this kinase is indeed a signaling element coupled to CD28 costimulation.
Our data indicate that the CD28 signal promotes adhesive interactions between T cells and APC, as well as suggesting the activation of Rac1 and CDC42-associated kinase pathway in T cells. It will be of interest to determine whether the F-actin formation and the kinase activation are in linkage or are regulated independently by CD28 signals. The present data appear to be in favor of the independent regulation of these consequences. First, CD28 signal has been known to trigger strong costimulation of PMA-activated T cells by markedly increasing cell proliferation and IL-2 mRNA levels, and inducing JNK activity (38, 39). PMA alone causes a strong actin polymerization in T cells (Fig. 1⇑), thus the pre-existing accumulation of F-actin does not alter CD28-dependent JNK activation. Second, a recent report demonstrated that different molecular domains of Rac1 are responsible for 1) F-actin formation and induction of membrane ruffles, and 2) PAK activation (40). These data indicate a segregation of the CD28-coupled signaling pathway that modulates cytoskeletal assembly from the pathway that enhances IL-2 production.
In summary, the data indicate that CD28 signaling stimulates cellular mechanisms that promote T cell-APC adhesion at localized contact sites where key cytoskeletal and G-protein molecules dynamically co-accumulate in activated T cells. Unlike integrin-mediated cell-cell contacts, this adhesion was independent of Rho but Rac1 and CDC42 translocated to contact sites. Correlated with the adhesion process is the CD28-dependent activation of Rac1 and CDC42-coupled signaling cascades that may contribute to the activation of JNK, and possibly play a modulatory role in cytoskeletal restructuring.
Acknowledgments
We thank J. Allison for the gift of B cell hybridoma lines for anti-CD28 and T. Watts for the gift of CHO-B7–2. We thank E. Negrou for her assistance in the confocal microscopic study. M. Cameron and H. Chou are also acknowledged for technical assistance. We also thank B. Gill for helpful discussions and input.
Footnotes
-
↵1 This work was supported by National Cancer Institute of Canada Grant 006349. A.O. is a Multiorgan Transplant Service Scholar. S.K. is a scholar from Showa University School of Medicine (Tokyo, Japan). S.J.R. is a Merk-Frost Scholar.
-
↵2 Address correspondence and reprint requests to Dr. Atsuo Ochi, John P. Robarts Research Institute, 1400 Western Road, London, Ontario N6G 2V4, Canada. E-mail: ochi{at}rri.uwo.ca
- Received August 22, 1997.
- Accepted October 23, 1997.
- Copyright © 1998 by The American Association of Immunologists
References
In this issue
