WIP deficiency reveals a differential role for WIP and the actin cytoskeleton in T and B cell activation

Immunity, Vol. 16, 193–204, February, 2002, Copyright 2002 by Cell Press WIP Deficiency Reveals a Differential Role for WIP and the Actin Cytoskeleton in T and B Cell Activation as Wiskott Aldrich Syndrome protein (WASP) and Ena/ VASP family members (Grakoui et al., 1999; Krause et al., 2000; Monks et al., 1998). The accumulation of F-actin at the T cell-APC interface is thought to stabilize a continuous contact between T cells and APCs, which is required Ine ´s M. Anto ´n, 1,5,6 Miguel A. de la Fuente, 1,5 Tasha N. Sims, 2 Sheryl Freeman, 1 Narayanaswamy Ramesh, 1 John H. Hartwig, 3 Michael L. Dustin, 2 and Raif S. Geha 1,4 1 Division of Immunology for optimal T cell activation. Inhibition of actin polymerChildren’s Hospital and Department of Pediatrics ization by cytochalasin blocks formation of the immunoHarvard Medical School logical synapse and T cell activation (Holsinger et al., Boston, Massachusetts 02115 1998; Wulfing et al., 1998). 2 Skirball Institute of Biomolecular Medicine Wiskott Aldrich Syndrome (WAS) is a severe X-linked New York University School of Medicine immunodeficiency caused by mutations in the WASP gene New York, New York 10016 (Derry et al., 1994). WASP is the first identified member of 3 Division of Experimental Medicine an expanding family of proteins involved in signaling and Brigham and Women’s Hospital and cytoskeletal organization that include N-WASP and Scar/ Harvard Medical School WAVE (Machesky and Insall, 1998; Miki et al., 1996, Boston, Massachusetts 02115 1998a). TCR signaling results in the activation of Vav, which is an exchange factor for the small GTPase Cdc42. GTP-loaded Cdc42 binds to WASP (Aspenstrom et al., Summary 1996; Kolluri et al., 1996; Symons et al., 1996) and its homolog N-WASP (Miki et al., 1998b), causing a conforWIP stabilizes actin filaments and is important for filomational change that allows them to interact with the podium formation. To define the role of WIP in immuArp2/3 complex and initiate actin polymerization (Kim nity, we generated WIP-deficient mice. WIP⫺ / ⫺mice et al., 2000; Miki et al., 1996; Rohatgi et al., 2000). WASP have normal lymphocyte development, but their T cells plays a critical role in T cell activation and in the reorganifail to proliferate, secrete IL-2, increase their F-actin zation of the actin cytoskeleton that follow TCR engagecontent, polarize and extend protrusions following ment. This is evidenced by the markedly impaired prolifT cell receptor ligation, and are deficient in conjugate eration of T cells from WAS patients and WASP-deficient formation with superantigen-presenting B cells and mice to anti-CD3 mAb and by their decreased ability to anti-CD3 bilayers. In contrast, WIP-deficient B lymphoform caps and to increase their F-actin content following cytes have enhanced proliferation and CD69 expresTCR/CD3 ligation (Gallego et al., 1997; Molina et al., sion following B cell receptor ligation and mount nor1993; Snapper et al., 1998; Zhang et al., 1999). mal antibody responses to T-independent antigens. The WASP-interacting protein (WIP) is a proline-rich 503 Both WIP-deficient T and B cells show a profound aa long protein that shows homology in its N-terminal defect in their subcortical actin filament networks. end to the yeast polarity development protein verprolin These results suggest that WIP is important for immu- (Ramesh et al., 1997). Several lines of evidence point to nologic synapse formation and T cell activation. the importance of WIP in regulating the actin cytoskeleton. First, WIP interacts with G-actin, F-actin, and with Introduction the actin polymerization regulating protein profilin (Martinez-Quiles et al., 2001; Ramesh et al., 1997). Second, The actin cytoskeleton is emerging as an integral comintroduction of human WIP into verprolin-deficient yeast ponent in T lymphocyte activation following the interaccorrects their cell growth, cytoskeletal organization, ention between the T cell receptor (TCR) and peptidedocytosis, and cell polarity defects (Vaduva et al., 1999). loaded major histocompatibility complex (MHC) moleThird, WIP overexpression in human B cell lines causes cules on the surface of antigen-presenting cells (APCs) an increase in cellular F-actin content and induces the (Dustin and Cooper, 2000; Penninger and Crabtree, formation of subcortical patches of actin (Ramesh et 1999). Interaction between T lymphocytes and APCs al., 1997). Fourth, WIP regulates N-WASP-induced actin induces the formation of molecular clusters at the connucleation (Martinez-Quiles et al., 2001). Finally, WIP is tact site that are enriched in polymerized filamentous important for filopodium formation (Martinez-Quiles et actin (F-actin) (Grakoui et al., 1999; Monks et al., 1998). al., 2001) and for actin tail generation by vaccinia virus These supramolecular activation clusters (SMACs), also (Moreau et al., 2000). To investigate the role of WIP in TCR-mediated T cell activation and cytoskeletal organinamed immunological synapses, contain several signalzation, we generated WIP null mice by gene targeting. ing components. They include TCR/CD3, src and Syk family kinases, and adaptor proteins such as SLP-76, Results Fyb, and Nck that are linked directly or indirectly to proteins that are involved in actin polymerization, such Generation of WIP-Deficient Mice A targeting construct in which coding exons two to five 4 Correspondence: [email protected] of the murine wip gene were replaced with a neomycin5 These authors contributed equally to this work. resistant gene was introduced in ES cells (Figure 1A). 6 Present address: Dipartimento di Scienze Cliniche e Biologiche, Ospedale San Luigi Gonzaga, 10043 Orbassano, Italy. ES clones with targeted disruption of one wip allele were Immunity 194 Figure 1. Generation of WIP-Deficient Mice (A) Genomic structure of the wip gene and predicted structure of the targeted allele after homologous recombination. Exons are represented by black boxes. Neo, neomycin resistance gene. The 650 bp SacI/EcoRI hybridization probe is shown as a patterned box. B, BamHI; E, EcoRI; S, SacI; and X, XbaI (not all restriction sites are shown). (B) Southern blot analysis of tail DNA. Genomic DNA was digested by EcoRI and probed with the 650 bp fragment immediately 3⬘to the targeted locus, shown in (A). The WT allele is represented by the 4 kb band. The knockout allele is represented by the 5.8 kb band. (C) Western blot analysis of WIP from thymocytes using rabbit antibody against an ESRSGSNRRERGAP WIP peptide. identified in Southern blots by the presence of a novel Increased Proliferation and IL-2 Receptor Expression in B Cells from WIP⫺ / ⫺Mice5.8 kb fragment derived from the targeted allele in addition to the 4 kb fragment derived from the WT allele. Of 97 in Response to Stimulation Pure B cells (⬎80% B220⫹) from WIP⫺ / ⫺mice showedES clones analyzed, one was found to have a disrupted allele and was used to generate WIP⫺ / ⫺mice, which markedly increased proliferation in response to antiIgM, lipopolysaccharide (LPS), and anti-CD40 mAb withwere identified by Southern blot analysis (Figure 1B). Western blot analysis of lysates from thymocytes, lymph or without IL-4 (Figure 2A). Activated B cells express increased levels of CD69 (Risso et al., 1989). There wasnodes (LN), and splenocytes confirmed the absence of WIP expression in WIP⫺ / ⫺mice (Figure 1C and data not increased expression of CD69 on B cells from WIP⫺ / ⫺ mice following stimulation with LPS, anti-CD40, LPS⫹IL-4,shown). WIP⫺ / ⫺mice did not display apparent differences from WT littermates in growth, weight, or health. and anti-CD40⫹IL-4 compared to B cells from WT littermates (Figure 2B and data not shown). The enhanced response of WIP⫺ / ⫺B cells to BCR ligation was notNormal Lymphoid Development in WIP-Deficient Mice Thymus cellularity in WIP⫺ / ⫺mice 6–10 weeks of age simply due to failure of internalization of the BCR, which terminates signaling, because they internalized theirwas significantly reduced compared to WT littermates (106 ⫾66 ⫻10 6 cells in KO versus 220 ⫾94 ⫻10 6 cells BCR normally following anti-IgM crosslinking, as assessed by FACS (data not shown).in WT, n ⫽6, p ⫽0.03). Detailed FACS analysis of thymocytes at 6–10 weeks of age revealed no obvious BCR ligation results in tyrosine phosphorylation of several proteins (reviewed in Benschop and Cambier,differences in the percentages of CD4⫹and CD8⫹cells, or of CD3, TCR␣␤, TCR␥␦, and CD2 positive cells (data 1999 and Kurosaki et al., 2000). Following anti-IgM stimulation, tyrosine phosphorylation of proteins ⬍50 kDanot shown). There were no obvious differences between spleen and LN from 6to 8-week-old WIP⫺ / ⫺and WT in mol wt was enhanced in WIP⫺ / ⫺B cells (Figure 2C). littermates as to numbers or percentages of CD4 ⴙ , CD8 ⴙ , B220 ⴙ , IgM ⴙ , IgD⫹, and Thy1⫹cells (data not shown). Serum Immunoglobulins and Antibody Responses in WIP⫺ / ⫺MiceBone marrow cells from the same WIP⫺ / ⫺mice had a normal profile of staining for B220, IgM, and CD43, and Serum IgM and IgE levels were significantly elevated in 6to 8-week-old WIP⫺ / ⫺mice compared to WT lit-the numbers of B220⫹, CD5⫹B1 cells in the peritoneum of WIP⫺ / ⫺mice were normal (data not shown). These termates, while serum IgG subclasses and serum IgA were not significantly different (Figure 3A).results suggest that WIP is not essential for the development of B and T lymphocytes. To determine the role of WIP in antibody immune Defective Activation in WIP-Deficient T Cells 195 Figure 2. B Cell Activation in WIP ⫺ / ⫺ Mice (A) Proliferation: Purified spleen B cells from 6to 9-week-old mice were cultured in medium alone or in the presence of stimuli, and proliferation was measured by [ 3 H]-thymidine incorporation. Error bars represent SDs of triplicate values. Similar results were obtained in four experiments. (B) Activation markers: Splenic B cells from 6to 9-week-old mice were incubated for 20 hr with the indicated stimuli and then analyzed by FACS for expression of CD69 and B220. Similar results were obtained in two other experiments. (C) Protein tyrosine phosphorylation: Splenic B cells from 6to 9-week-old mice were incubated at 37⬚C for the indicated times (min) with anti-IgM, lysed, run on SDS-PAGE, transferred to nitrocellulose, and developed with anti-phosphotyrosine antibody. The membrane was stripped and reprobed with anti-PLC␥2 as a loading control. Identical results were obtained in two other experiments. responses, we immunized WT, WIP⫺ / ⫺mice and WIP⫹ / ⫺able from those of WT controls (data not shown). WIP⫺ / ⫺ mice had normal or slightly increased IgM antibody re-littermates with the T-dependent (TD) antigen TNP-KLH and T-independent (TI) antigens. WIP⫹ / ⫺mice had serum sponses but virtually undetectable IgG responses to both KLH (Figure 3B) and TNP (data not shown). Bothimmunoglobulins and antibody responses indistinguish- Immunity 196 Figure 3. Antibody Production in WIP ⫺ / ⫺ Mice (A) Serum immunoglobulin levels from nonimmunized 6to 8-week-old WIP ⫺ / ⫺ (open circles) mice and WT littermates (filled circles) as determined by isotype-specific ELISA. (B) IgM and IgG antigen-specific antibody responses of 12to 15-week-old mice to KLH following immunization with the TD antigen TNP-KLH as determined by ELISA. (C) IgG antigen-specific antibody responses of 12to 15-week-old mice to TNP following immunization with the TI type I (TNP-LPS) or TI type II (TNP-Ficoll) antigens as determined by ELISA. IgG and IgM antibody responses of WIP⫺ / ⫺mice to the even upon costimulation with anti-CD28 as determined by bioassay (Figure 4D) and ELISA (data not shown).type I TI antigen TNP-LPS and to the type II TI antigen TNP-Ficoll were normal (Figure 3C and data not shown). TCR/CD3 ligation causes rapid activation of a number of tyrosine kinases, resulting in tyrosine phosphorylation of several proteins that play important roles in TCR sig-WIP Is Essential for T Cell Activation via the TCR/CD3 Complex nal transduction (reviewed in Acuto and Cantrell, 2000). Following anti-CD3 stimulation, protein tyrosine phos-Proliferation in response to plate-bound anti-CD3 was abolished in WIP⫺ / ⫺T lymphocytes at all coating conphorylation was grossly intact in WIP⫺ / ⫺T cells (data not shown), and calcium fluxes were only slightly decreasedcentrations tested (Figure 4A). In contrast, proliferation to PMA ⫹ionomycin, which bypasses receptor signal- (ⵑ15% decrease, n ⫽4 experiments). Erk phosphorylation was also slightly decreased, whereas JNK phos-ing, was normal. Interactions between IL-2 and IL-2 R play an important phorylation was normal (Figure 4E). Furthermore, there was no detectable impairment in the nuclear transloca-role in T cell proliferation, and optimal production of IL-2 by T cells requires costimulation via CD28 (Rudd, 1996). tion of NF␬B and NFATc as assessed by immunofluorescence (data not shown).The failure of WIP⫺ / ⫺T cells to proliferate to anti-CD3 was corrected by addition of IL-2 but not by costimulation with anti-CD28 mAb (Figure 4B). WIP⫺ / ⫺T cells poorly WIP⫺ / ⫺T Cells Do Not Increase F-Actin Content or Form Protrusions and Pseudopodiaupregulated IL-2R␣chain expression following anti-CD3 stimulation (Figure 4C). This was only modestly enhanced following TCR/CD3 Ligation TCR/CD3 ligation causes an increase in the cellular F-actinby costimulation with anti-CD28 and almost completely corrected by addition of IL-2. Similar results were obof normal T cells (Phatak and Packman, 1994). The effect of TCR/CD3 ligation on cellular F-actin content was exam-tained for CD69 expression (data not shown). WIP⫺ / ⫺ T cells failed to secrete detectable IL-2 after CD3 ligation ined by staining permeabilized cells with TRITC-labeled Defective Activation in WIP-Deficient T Cells 197 phalloidin. In four experiments, the amount of F-actin in hypothesis using a planar bilayer substrate in which fluorescently labeled anti-CD3 mAb is laterally mobile unstimulated WT and WIP⫺ / ⫺T cells was similar (data and ICAM-1 provides adhesion. Equivalent clustering of not shown). However, in contrast to T cells from WT anti-CD3 in contacts of WT and WIP⫺ / ⫺T cells was littermates, T cells from WIP⫺ / ⫺mice failed to increase observed by 4 min (data not shown). By 20 min, the their F-actin content following stimulation with anti-CD3 WT cells displayed extensive contact areas with the (Figure 5A). planar bilayer and colocalization of F-actin and GM1 in Engagement of the TCR/CD3 complex triggers a disthe contact area (Figure 6A). Total GM1 in WT and tinct pattern of reorganization of cellular F-actin (Bunnell WIP⫺ / ⫺cells was similar in level, but the amount of GM1 et al., 2001; Parsey and Lewis, 1993). Cytoskeletal reorin the contact areas was greater in WT cells. The mobility ganization was examined in T cells activated using covof the WT cells on the surface resulted in dispersion of erslip-bound anti-CD3 antibody. This system closely the clustered anti-CD3, but the total amount of clustered mimics T cell activation by MHC class II-peptide comanti-CD3 in WT contacts remained the same when inteplexes arrayed on the surface of APCs by creating a grated over the large contact areas. After 20 min, WIP⫺ / ⫺ polarized stimulus that induces T cells to undergo the T cells formed smaller contact areas with lower levels morphological changes necessary to maximize the conof F-actin and GM1 than WT cells but still contained tact between their antigen receptors and the activating clustered anti-CD3 mAb (Figure 6A). The difference in surface (Bunnell et al., 2001). Unstimulated T cells from contact area was statistically significant (Figure 6B), WT and WIP⫺ / ⫺mice exhibited a similar staining pattern while the trends in F-actin and GM-1 localization were with prominent actin rings, as previously described consistently observed but were not statistically signifi- (Parsey and Lewis, 1993). After incubation for 20 min cant upon quantification (data not shown). over anti-CD3 coated glass coverslips, T cells from WT The above results suggested that WIP⫺ / ⫺T cells might mice acquired a polarized shape, had less prominent have defects in forming an immunological synapse upon actin rings, and spread by making microspikes and blunt interaction with APCs. Since WIP⫺ / ⫺mice have not yet pseudopodia enriched in F-actin (Figure 5B and video been bred on a TCR transgenic background, we emframes in Figure 5C). The surface projections were visuployed I-E k positive APC and the superantigen Staphyloalized by fluorescence microscopy, although they were coccal Enterotoxin A (SEA) to trigger initial stages of difficult to detect by differential interference contrast synapse formation (Wulfing et al., 2000). An early actinused in the time lapse videomicroscopy. In contrast, dependent stage in immunological synapse formation T cells from WIP⫺ / ⫺mice maintained a round shape with has been identified as an extensive interface between persistence of a strong actin ring. More importantly, they the T cell and APC (Negulescu et al., 1996; Wulfing et al., spread very poorly on anti-CD3 coated surface and had 1998). Formation of this broad interface with pseudopod markedly reduced ability to make cortical protrusions extension is most prominent at 2 min and is actin depen- (Figure 5B and video frames in Figure 5C). However, dent. WT cells formed a classical extended interface in actin cap formation following stimulation with immobiwhich clustering of the integrin LFA-1 was observed lized anti-CD3 for 20 min was similar in WIP⫺ / ⫺T cells (Figure 6Ca and 6Cc). The extension of pseudopodia and controls (52% versus 54% capped cells). resulted in a contact angle greater than 90⬚with the The failure of T cells from WIP⫺ / ⫺mice to spread APC in all WT cells conjugates observed (Figure 6Ca). At and to develop projections following TCR/CD3 ligation this time, TCR redistribution was not detectable (Figure prompted us to examine their dynamics when exposed 6Cb), although SEA was required to observe conjugates to an anti-CD3 coated surface. Purified T cells were providing evidence of TCR-SEA-MHC interactions. plated on anti-CD3 coated or anti-B220 coated glass WIP⫺ / ⫺T cells displayed 2-fold fewer conjugates (data coverslips and observed with videomicroscopy for 20 not shown) and a significantly smaller interface both at min. T cells from WT and WIP⫺ / ⫺mice attached to a the 2 and 5 min time points (Figure 6D). The contact similar extent to the anti-CD3 coated surface but did angle was also always smaller than 90⬚indicating that not attach to the anti-B220-coated control surface (data the WIP⫺ / ⫺T cells remained round, a situation that was not shown). The change in shape and polarized appearnever observed with WT cells at that time. Despite the ance of T cells from WT mice after incubation with bound overall defect in contact formation, LFA-1 was still accuanti-CD3 for the indicated times from 0–14 min are illusmulated to a similar extent as in the WT T cell conjugates. trated in Figure 5C. During the course of the video, 90% The cortical actin network is disrupted in WIP-defiof tracked T cells from WT mice cells extended protrucient lymphocytes. WIP stabilizes actin filaments sions that later retracted, only to appear again in other (Ramesh et al., 1997). This prompted us to examine the areas of the cell. In contrast, T cells from WIP⫺ / ⫺mice architecture of the actin network in WIP⫺ / ⫺lymphocytes. did not acquire a polarized shape; only 9% of them Purified T and B cells were adhered to anti-CD3 and exhibited protrusions at some point, and the number anti-IgM coated coverslips, respectively, and their actin and the movement of these protrusions were drastically network was directly visualized by electron microscopy reduced. by unroofing pseudopods to reveal the basal cell membrane. The cytoplasmic surface of the plasma membrane of adherent unstimulated WT T cells was decorContact Formation with Anti-CD3 Bilayers ated with a loose network of F-actin (Figure 7A, upper and Conjugate Formation with APCs left panel). Activation over anti-CD3 coated coverslips Are Impaired in WIP⫺ / ⫺T Cells for 15 min led to recruitment of additional F-actin (Figure The poor response of WIP⫺ / ⫺T cells to anti-CD3 could 7A, bottom left panel) and formation of prominent protrusions (Figure 7A, inset in bottom left panel). In markedbe related to defects in TCR clustering. We tested this Immunity 198 Figure 4. Antigen Receptor-Induced Proliferation and Activation in Purified T WIP ⫺ / ⫺ T Cells (A) Purified LN T cells from 6to 10-week-old WT mice and WIP ⫺ / ⫺ littermates were stimulated in wells coated with increasing concentrations of anti-CD3⑀or in the presence of PMA and ionomycin. T cells were cultured for 48 hr, pulsed with 1 ␮Ci [ 3 H]-thymidine for an additional 16 hr, then Defective Activation in WIP-Deficient T Cells 199 contrast, although the total basal F-actin content was ing in vivo following immunization with the TD antigen TNP-KLH (Figure 3B) reflects deficient T cell help rathersimilar in WT and WIP ⫺ / ⫺T cells (Figure 5A), the plasma membrane of adherent WIP⫺ / ⫺T cells was sparsely than an intrinsic abnormality in the B cells since IgG isotype switching in response to type I and type II TIcoated with F-actin, and activation did little to alter this coating (Figure 7A, upper and bottom right panels), reantigens was normal (Figure 3C). WIP was not essential for the early biochemical eventssulting in less prominent protrusions than in WIP⫹ / ⫺cells (Figure 7A, inset in bottom right panel). These findings that follow TCR engagement since calcium mobilization and total protein tyrosine phosphorylation, JNK phos-suggest that WIP is essential for the integrity of the actin cytoskeleton in T cells and for its reorganization phorylation, and nuclear translocation of NF␬B and NFAT were all essentially normal in WIP⫺ / ⫺T cells. How-following TCR engagement. A similar defect in the actin network was also found ever, T cells from WIP⫺ / ⫺mice failed to increase their F-actin content following TCR/CD3 ligation (Figure 5A).in WIP-deficient B cells. The cytoplasmic surface of the plasma membrane of B cells from WIP⫹ / ⫺mice was Inhibition of actin polymerization with cytochalasin D blocks both calcium flux and IL-2 secretion induced bydecorated with a loose network of F-actin (Figure 7B, upper left panel), and there was no increase in the dencrosslinking of the TCR/CD3 complex (Valitutti et al., 1995). Furthermore, TCR ligation induces association ofsity of the actin network following anti-IgM stimulation (Figure 7B, bottom left panel). This was consistent with tyrosine-phosphorylated CD3␨with the actin cytoskeleton. Disruption of this association results in reducedthe observation that anti-IgM stimulation caused a marginal increase in F-actin content of B cells (data not IL-2 production but preserved tyrosine phosphorylation (Rozdzial et al., 1995). The failure of T cells from WIP⫺ / ⫺ shown). The plasma membrane of adherent WIP⫺ / ⫺ B cells is sparsely coated with F-actin (Figure 7B, upper mice to secrete IL-2 is consistent with their failure to increase F-actin content. The well-preserved calciumright panel) and, like the others, was not changed following anti-IgM stimulation (Figure 7B, bottom right panel). flux in these cells could be explained by a less stringent requirement for actin polymerization and by residualThese findings suggest that WIP is also important for the integrity of the actin cytoskeleton in B cells. actin polymerization activity in WIP⫺ / ⫺T cells, compared to its complete inhibition caused by cytochalasin D. Activation of Jurkat T cells with immobilized anti-CD3Discussion antibodies provokes a distinctive pattern of F-actin reorganization and characteristic changes in cell shapeOur results indicate that WIP is critical for the integrity of the actin cytoskeleton in both T and B lymphocytes (Bunnell et al., 2001). This includes progressive dissolution of actin rings followed by cellular spreading andand is essential for T cell, but not B cell, activation. Although WIP is widely expressed, WIP⫺ / ⫺mice apformation of pseudopods (Parsey and Lewis, 1993). We found similar results when we stimulated WT murinepear normal and show no gross abnormalities. Moreover, except for a modest reduction in thymocyte numT cells with coverslip bound anti-CD3 antibody: the actin ring attenuates, actin-rich microspikes appear on thebers in WIP⫺ / ⫺mice, WIP is dispensable for T and B lymphocyte development but is essential for T cell acticell surface, and pseudopodia form and attach to the stimulating surface. In contrast, T cells from WIP⫺ / ⫺micevation and proliferation in response to ligation of the TCR/CD3 complex. T cells from WIP⫺ / ⫺mice completely conserve their actin rings and fail to form protrusions and pseudopodia following anti-CD3 stimulation (Figurefailed to proliferate and secrete IL-2 in response to antiCD3 (Figure 4). Correction of the proliferation defect and 5B). This suggests an essential role of WIP in cytoskeletal reorganization after TCR engagement.upregulation of CD25 expression by anti-CD3 suggest residual CD3 signaling in WIP⫺ / ⫺T cells. The failure of Upon initial TCR triggering by engagement with peptide-MHC in the APC surface, T cells polarize towardanti-CD28 costimulation to correct the proliferative defect of WIP⫺ / ⫺T cells may reflect a requirement for a the APC, crawl around them, and spread lamellipodia. These changes, which require a functional actin cy-higher threshold of signaling via CD3 and/or a requirement for WIP in CD28 signaling. CD28 may be linked to toskeleton (Valitutti et al., 1995), allow the sustained association of multiple TCR/peptide-MHC complexesWIP by Vav (Klasen et al., 1998), which synergizes with WIP in inducing NFAT activity and IL-2 production (Sarequired for triggering optimal T cell activation. WIP⫺ / ⫺ T cells fail to acquire a polarized shape and exhibit very fewvoy et al., 2000). Failure of WIP⫺ / ⫺mice to undergo IgG isotype switchmotile protrusions when exposed to an anti-CD3 coated collected and scintillation counted. Error bars represent SDs of triplicate values. Displayed are the results of one experiment representative of the four performed with similar results. (B) Purified LN T cells from 6to 10-week-old WT mice and WIP ⫺ / ⫺ littermates were stimulated in wells coated with 10 ␮g/ml anti-CD3 in the presence of plate-bound anti-CD28 (10 ␮g/ml) or soluble IL-2 (40 ng/ml). Experimental conditions are as described in (A). Error bars represent SDs of triplicate values. Displayed are the results of one experiment representative of the three performed with similar results. (C) FACS analysis of CD25 surface expression on purified CD4 ⫹ /CD8 ⫹ T cells from 6to 10-week-old mice before and after stimulation in wells coated with 10 ␮g/ml anti-CD3 with or without the addition of plate-bound anti-CD28 (10 ␮g/ml) or IL-2 (40 ng/ml) for 20 hr. (D) Purified LN T cells from 6to 10-week-old WT mice and WIP ⫺ / ⫺ littermates were stimulated in wells coated with 10 ␮g/ml anti-CD3 with or without addition of anti-CD28 (10 ␮g/ml) for 96 hr. Supernatants were then collected and assayed for IL-2 by examining their capacity to induce the proliferation of the IL-2-dependent cell line CTLL2. (E) Purified LN T cells from 12to 14-week-old mice were stimulated on anti-CD3⫹anti-CD28 coated plates for the indicated times. Cell lysates were blotted with anti-phosphoERK and anti-ERK as loading control and with anti-phosphoJNK and anti-JNK as loading control. Immunity 200 surface (Figure 5C), suggesting that WIP⫺ / ⫺T cells may be unable to establish the sustained contact with the APC necessary for antigen-driven T cell activation. Experiments with anti-CD3 on planar bilayers demonstrated that that CD3 is engaged and clustered in WIP⫺ / ⫺ T cells similarly to WT T cells. Nevertheless, there was a profound defect in expansion of the contact area, which is a hallmark of early immunological synapse formation. Similar results were obtained with T cell-APC interactions triggered by SEA (Figure 6). LFA-1 clustering in the nascent immunological synapse was observed in response to antigen in WIP-deficient T cells, but the interface was not normal in that it never underwent the characteristic expansion observed for WT cells within 2 min of cell-cell contact. The finding that WIP⫺ / ⫺T cells can cluster LFA-1 in the interface suggests that LFA-1 activation may not require formation of new actin filaments and demonstrates that some actin-dependent functions, which also include actin cap formation, are intact in WIP⫺ / ⫺mice. Electron microscopy revealed that the amount of F-actin associated with the cytoplasmic side of the adherent membrane was reduced in WIP ⫺ / ⫺T cells and did not increase after CD3 ligation. Furthermore, actin filaments associated with adherent membranes were sparse, disrupted, and disorganized (Figure 7A). Disruption of the actin network is likely to explain the T cell defect in WIP⫺ / ⫺mice. Crosslinking of the BCR has been reported to cause its translocation to lipid rafts and its association with the cytoskeleton in a detergent insoluble fraction (Braun et al., 1982; Cheng et al., 1999; Jugloff and JongstraBilen, 1997). In spite of a reduced association of the actin cytoskeleton with the plasma membrane (Figure 7B), B cells from WIP⫺ / ⫺mice exhibited increased proliferation and CD69 expression in response to ligation of the BCR, anti-CD40, and LPS. This suggests that WIP and an intact actin cytoskeleton may be negative regulators of B cell activation. The opposite effect of WIP deficiency on the activation of T and B cells suggests that WIP and the actin network play fundamentally different roles in signaling by the TCR versus the BCR. While the TCR aggregates after stimulation, the BCR exists as an oligomer that is thought to dissociate following receptor ligation (Schamel and Reth, 2000). This dissociation may be inhibited by actin filaments. Since WIP stabilizes actin filaments, it may normally inhibit BCR signaling. Furthermore, while inhibition of actin polymerization by cytochalasin blocks raft assembly following TCR ligation, translocation of the BCR into rafts followFigure 5. F-Actin Content and Distribution and Time Lapse Videomiing crosslinking is not affected (Cheng et al., 2001), and croscopy of T Cells after Stimulation with Anti-CD3 B cell proliferation in response to BCR ligation is en- (A) Purified LN T cells from 9to 14-week-old WT or WIP ⫺ / ⫺ mice hanced (Rothstein, 1985). were stimulated with anti-CD3 followed by crosslinking with a secSome of the functions of WASP and its homolog ondary anti-rat antibody. After 20 min, the cells were fixed, permeN-WASP may be mediated by WIP. The ability of WASP abilized, stained with phalloidin-TRITC, and analyzed by FACS. Simito enhance TCR induction of NFAT activity and IL-2 lar data were obtained in three other experiments. (B) Purified T lymphocytes were stimulated on anti-CD3 coated glass expression depends on the WIP binding WH1 region of coverslips. After 20 min, the cells were fixed, stained with phalloidinWASP (Silvin et al., 2001). It has been suggested that TRITC, and examined by fluorescent microscopy. WASP/N-WASP and WIP function as a unit (Martinez- (C) Purified T cells from WT and WIP ⫺ / ⫺ mice were added to antiQuiles et al., 2001). This may explain the similarities CD3 coated glass coverslips and allowed to sediment for 30 min at between the phenotype of WIP⫺ / ⫺mice and that of WAS 4⬚C. After washing off unattached cells, warm (37⬚C) medium was patients and WASP-deficient mice. WASP-deficient T cells added, and the cells were videotaped for 20 min. The indicated time proliferate poorly and fail to increase their F-actin content frames were processed using Adobe Photoshop software. Arrowheads point to protrusions. Bar is 10 ␮m. to anti-CD3. However, there are important differences be- Defective Activation in WIP-Deficient T Cells 201 Figure 6. WIP ⫺ / ⫺ T Cells Cluster the TCR/ CD3 Complex but Display a Significant Defect in Contact Area Expansion and Fail to Form an Expanded Interface with APC (A) At 20 min, WT LN T cells interact with ICAM-1and anti-CD3 FITC-containing bilayers (green). The cells were fixed and stained using phalloidin-rhodamine (for F-actin, red) and CtxB-cy5 (for GM1, blue). Therefore, purple indicates overlap of actin and cholera toxin. The CD3 antibody is laterally mobile in the bilayer, so accumulations of green indicate CD3-Ab interaction. The motility of WT cells results in some shedding of CD3 clusters that appear around the contact areas. Images were taken at the bilayer as determined by FITC brightness. Control bilayers lacking anti-CD3 showed no contact formation. At 20 min, WIP ⫺ / ⫺ lymph node T cells also interacted with the same type of bilayer and were fixed and stained as WT cells. The white outline indicates the contact area for each cell. Scale bar, 7 ␮m. (B) Quantification of contact area with anti-CD3 bilayer based on two separate experiments. Statistical analysis used Student’s t test with an n ⫽94. (C) WIP ⫺ / ⫺ T cells polarize LFA-1 but fail to form an expanded interface with the APC. (A and D) Transmitted laser light image taken at 2 min of a lymph node T cell interacting with A20 B cell loaded with SEA (1 ␮M). (B and E) H57 Fab fragments labeled with Cy5 to mark the TCR. (C and F) H155 Fab fragments labeled with Cy3, marking LFA-1 fluorescence. The intersecting lines in (A) and (D) indicate the contact angle, which is greater than 90⬚ for WT and less than 90⬚for the WIP ⫺ / ⫺ . The arrows in (B) and (E) indicate the length of the contact interface. Scale bar, 4.5 ␮m. (D) Analysis of length of contact interface at 2 and 5 min. T cell-A20 conjugates were measured at their widest interface, and averages were taken. Similar data were obtained in three separate experiments. Statistical analysis used Student’s t test with an n ⫽31.