ADAP is dispensable for NK cell development and function.

NK cells are key mediators of the innate immune response and anti-tumor surveillance. Adhesion and degranulation-promoting adapter protein (ADAP, formerly known as SLAP-130 or Fyb) is a hematopoietic-specific adapter that is required for efficient TCR signaling and T cell activation. Herein, we examine a potential role for ADAP in NK development and function. ADAP is expressed in primary NK cells and in IL-2 stimulated lymphokine-activated killers. However, ADAP-deficient mice show no defects in NK development. Further, ADAP is dispensable for key NK functions, including cytotoxicity in response to engagement of activating receptors, cytokine production, conjugate formation and tumor suppression in vivo. These results indicate that, unlike events stimulated by TCR engagement, signaling events engaged by immunoreceptor tyrosine-based activation motif-associated and cytokine receptors on NK cells can occur independently of ADAP.


Introduction
Adhesion and degranulation-promoting adapter protein (ADAP, formerly known as SLAP-130 or Fyb) is a hematopoieticspecific adapter molecule that was identified on the basis of TCR-inducible association with signaling proteins fyn (1) and SH2-containing leukocyte protein of 76 kDa (SLP-76) (2). In addition to the T lineage, ADAP is expressed in myeloid cells and in a human NK cell line [(3) and N. Fang and G. Koretzky, unpublished results]. ADAP contains a number of domains that mediate association with proteins implicated in cellular adhesion and migration, cytoskeletal rearrangement TCR signal transduction (4).
ADAP regulates TCR-mediated cellular responses (5). Early over-expression studies implicated ADAP in TCR-induced IL-2 production (1,2). Later, analysis of ADAP-deficient T cells established a positive regulatory role for ADAP in TCRmediated cellular activation (6,7). Upon TCR ligation, ADAPdeficient peripheral T cells show intact proximal signaling, but they inefficiently up-regulate early-activation antigens CD25 and CD69, fail to produce IL-2 and proliferate poorly in response to TCR stimulation. ADAP-deficient T cells also demonstrate impaired activation of integrins, with decreased clustering of leukocyte function antigen 1 (LFA-1) and impaired adhesion to inter-cellular adhesion molecule-1 upon TCR ligation. More recently, ADAP has been shown to be required for thymocyte positive and negative selection, processes that are critically dependent upon TCR engagement (8). Although data supporting roles for ADAP in mast cell and osteoclast function or differentiation have been reported (9,10), little work has been done investigating the role of ADAP in the biology of non-T lymphoid cells.
NK cells regulate innate immune responses, and perform vital tumor and virus-infected cell surveillance functions. Like T cells, NK cells derive from a common lymphoid progenitor and develop the capacity for effector functions like cytokine secretion and cytotoxicity (11,12). Analysis of the molecular basis for these shared functions reveals usage of identical or similar signaling proteins and cascades in T and NK cells. For example, both NK and T cells can be activated by ligation of receptors associated with immunoreceptor tyrosine activation motifs (ITAMs). The TCR relays signals through ITAM-bearing members of the CD3 complex. FccRIIIA, the IgGR that mediates antibody-dependent cellular cytotoxicity (ADCC) on NK cells, also associates with the ITAM-bearing CD3f or the FceRIc chains (13). Similar to what is observed after TCR engagement, activation of src and syk family protein tyrosine kinases (PTKs) is the earliest detectable biochemical event following FccRIIIA ligation, and inhibition of PTK activity blocks ADCC by NK cells (14). Activating Ly49 family members form another group of murine NK and TCRs that associate with the ITAM-bearing transmembrane protein DAP-12/KARAP, through a transmembrane charged residue (13). Additionally, selected members of the signaling lymphocytic activation molecule (SLAM) family of proteins, including 2B4, have recently been shown to regulate T cell co-stimulation and NK lytic capacity (15). Finally, integrins mediate adhesion and signaling in both NK and T cells (16,17). Activity of the integrin LFA-1, which is impaired in ADAP-null T cells (6,7), is also required for optimal NK cell cytotoxicity (17).
In view of the shared biologic features of T and NK cells, and in light of the requirement for ADAP in efficient T cell development, adhesion and function (6, 7), we hypothesized that ADAP would also play a role in NK cell development and function. In this study, we describe ADAP expression in primary NK cells, and evaluate NK cell development and function in ADAP-deficient mice. We find no defects in NK cell development and distribution in ADAP-null animals. ADAPdeficient, lymphokine-activated killer (LAK) cells conjugate with and kill standard murine NK targets and are controlled by inhibitory signals comparable to wild-type cells. Furthermore, ADAP-deficient splenocytes respond to cytokine stimulation, as evidenced by enhanced IL-2-induced cytotoxicity and by IL-12-dependent secretion of IFN-c. Surprisingly, ADAP is dispensable for ADCC and for 2B4-mediated killing and cytokine secretion. ADAP-null mice show normal NK-dependent in vivo tumor rejection function. These results suggest that signaling through cytokine and activating receptors in NK cells can occur independently of ADAP.

Mice and cell culture
The generation of ADAP-deficient mice has been described elsewhere (6). All experiments were performed with animals backcrossed to the C57Bl/6 background at least six times. Mice were housed in the Specific Pathogen Free facility at the UMN; animals were 6-16 weeks of age at time of study. Spleens were removed and disrupted through 70-lm nylon filters to generate single-cell suspensions. RBCs were depleted by splenocyte incubation in ammonium chloride (1%) solution for 1 min. Liver lymphocyte populations were enriched by passing hepatic tissue through a tissue disaggregator, RBC depletion and collection of interface-localized cells after density gradient centrifugation (Histopaque, Sigma, St Louis, MO, USA). Cell culture and cytotoxicity assays were performed in RPMI with 10% FCS, supplemented with penicillin, streptomycin, L-glutamine and 2-mercaptoethanol (2-ME). LAKs were generated by culture of RBC-depleted splenocytes at 3 3 10 6 cells ml ÿ1 in the presence of rhIL-2 [Tecin, National Institutes of Health (NIH)] 1000 U ml ÿ1 for 5-8 days. Non-adherent cells were removed by gentle washing at day 3, and fresh IL-2-containing medium was added to the remaining cells. Ly49A+ LAKs were generated as previously described (19). Briefly, after 6 days in culture, LAKs were stained with anti-Ly49A mAb A1 (18). After panning on antimouse Ig-coated culture flasks, adherent cells were cultured for 3 additional days before use in cytotoxicity assays.

Conjugate assay
The protocol was adapted from that described in (22). YAC-1 target cells were labeled for 10 min in 5 lM 5,6-carboxyfluorescein diacetate succinimidyl ester (Invitrogen, Carlsbad, CA, USA) and mixed 1:1 with allophycocyanin-NK1.1-labeled LAKs. For blocking conditions, LAKs were pre-treated with 20 lg ml ÿ1 anti-LFA-1 or IgG2a control for 20 min on ice. Cells were gently spun down in round-bottom FACS tubes at 300 r.p.m. 4°C for 2 min and then incubated at 37°C for various times. One volume ice-cold buffer (2% FCS in PBS) was added to terminate stimulation and samples were immediately analyzed on a FACSCalibur.

Cytotoxicity assays
Standard 4-h 51 Cr-release assays were performed as previously described (23,24). Briefly, splenocytes or LAKs were used to assess natural cytotoxicity. Targets (5 3 10 3 or 10 4 per well) were labeled with 51 Cr for 90 min, washed extensively and incubated with effector cells in triplicate at indicated ratios in V-bottom 96-well cell tissue culture plates (Becton Dickinson Labware, Franklin Lakes, NJ, USA). For the ADCC assay, NKresistant 51 Cr-labeled EL4 (murine T lymphoma cell line) targets were washed and incubated with anti-Thy1 mAb (5 ll per 10 6 cells) or with isotype control antibody (IgG2a; 1 lg per 10 6 cells; MP Biomedicals) for 10 min at room temperature; targets were then placed in 96-well culture plates (10 4 per well). For the redirected lysis assay, P815 mastocytoma cells (6 3 10 3 ) were incubated with anti-2B4 antibodies for 10 min before addition of LAK effectors (6 3 10 4 per well). After 4 h at 37°C, samples were centrifuged, and supernatants were assessed for released 51

Western analysis
Cells were lysed using 1% NP-40 and protease inhibitors; insoluble material was removed by centrifugation at 14 000 r.p.m. for 10 min in a tabletop microfuge. Supernatants were mixed with 43 sample buffer [Tris-HCl (pH 6.8), 0.5 M; SDS, 277 mM; glycerol, 40%; 2-ME, 20%; bromophenol blue, 1%] and boiled for 5 min. Samples were resolved by SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Millipore). Membranes were blocked with 5% BSA in PBS for 30 min and then probed overnight with the appropriate antibody diluted in Tris-buffered saline T/5% BSA. Antibody binding was detected by Alexa-Fluor 680-conjugated secondary antibody (Molecular Probes, Eugene OR, USA) and the Odyssey imaging system (Licor).

In vivo tumor suppression assay
This assay was performed as previously described (26). Briefly, 0.5 3 10 5 B16 melanoma cells were injected by tail vein. Twenty-one days later, mice were killed, lungs were removed into PBS and black lesions on the lung surface were enumerated by visual inspection using a 32.5 objective on a dissecting microscope. In order to deplete NK cells, some mice received tail vein injection of anti-NK1.1 mAb (200 lg per mouse) 3 days prior to inoculation with B16 cells.

ADAP is expressed in murine NK cells
ADAP protein expression has been reported in T lymphocytes, platelets and in cells of the myeloid lineages (1,2,27). Western analysis of lysates from NK3.3, a human NK-like cell line, showed that these cells also express ADAP (N. Fang, E. Peterson and G. Koretzky, unpublished results). We therefore addressed the question of ADAP expression in primary murine NK cells.
NK cells (defined by a surface phenotype of NK1.1+ or DX-5+, CD3ÿ) comprise 2-5% murine splenocytes. To obviate the need for extensive cell purification prior to western blotting for ADAP expression in NK cells, we employed intracellular staining with polyclonal anti-ADAP antiserum. Figure 1(A) shows the sensitivity and specificity of anti-ADAP antiserum using this approach. We incubated fixed and permeabilized splenocytes with either pre-immune or anti-ADAP serum. Anti-ADAP staining produced no signal in ADAP-deficient CD3+ splenocytes. Staining with anti-ADAP resulted in a >10-fold, specific shift in mean fluorescence intensity in wild-type CD3+ splenic T cells (Fig. 1A). Anti-ADAP staining produced fluorescent signal in <2% of wild-type CD19+ splenocytes, consistent with previous studies indicating lack of ADAP expression in B cells (2). The few ADAP+CD19+ cells are likely follicular dendritic cells (28). Like CD3+ cells, CD3-negative splenocytes co-staining with NK surface markers DX-5 or NK1.1 express ADAP (Fig. 1B).
In vitro IL-2 stimulation of splenocytes generates LAK cells often utilized for ex vivo assessment of cytotoxic function. We asked whether ADAP is expressed in LAKs after 6 days in culture. By western analysis, we found that ADAP is expressed in LAKs (purity = 97% NK1.1+, CD3ÿ; data not shown) at a level comparable to that seen in thymocytes (Fig. 1C). Together, these data indicate that ADAP is expressed in primary murine DX-5-and NK1.1-positive splenocytes and in LAKs.

ADAP-deficient mice contain normal numbers of NK and NKT cells
ADAP is required for normal thymocyte development and for population of the periphery with T lymphocytes (6)(7)(8). In mice lacking the ADAP-binding partner fyn, deficiency in lymphocytes bearing both T and NK markers (NKT cells) have been reported (29). Although we have previously identified DX-5+ splenocytes in ADAP-deficient mice, we wished to further explore a potential role for ADAP in the development and distribution of cells bearing NK surface markers. Antibodies against TCRb or CD3 and DX-5 or NK1.1 were used to distinguish NK, NKT and T cell populations in thymus, bone marrow, spleen and liver by flow cytometric analysis. We found that both NK and NKT cells were present in these tissues from ADAP-deficient mice at levels (both percentages and absolute numbers) comparable to those found in control animal tissues ( Fig. 2A and B and data not shown). Furthermore, we found that equivalent percentages of control and ADAP-deficient spleen, bone marrow and blood-derived NK1.1+ cells express CD11b, a marker indicating more mature NK cells (30) (Fig. 2C and data not shown). As ADAP-null mice <4 months of age have normal splenic cellularity (6), absolute numbers of NK cells are also comparable in ADAP-deficient and control spleens. These data indicate that ADAP is not required for establishment of mature, peripheral NK or NKT cell populations.

ADAP is dispensable for 'natural' and LAK cytotoxicity
The NK activating receptor NKG2D can mediate cytotoxicity against murine NK-sensitive YAC-1 targets (31). NKG2D, like the TCR, can engage src family PTK activity and induce phosphorylation of the ADAP-binding partner SLP-76 (13). As ADAP is tyrosine phosphorylated after TCR engagement in T lymphocytes, we tested the role of ADAP in the development of NK lytic function against YAC-1. Figure 3(A) shows that, in a standard 4-h 51 Cr-release cytotoxicity assay, freshly isolated ADAP-deficient splenocytes are comparable to wild-type and heterozygous mice in their ability to kill YAC-1 cells. We also observed normal enhancement of lytic activity against YAC-1 in ADAP-deficient LAKs (Fig. 3B). In another test of NK activating receptor function, we found no defect in the ability of ADAP-deficient LAKs to lyse class I-deficient RMA-S cells (data not shown). These data indicate that signals leading to 'natural cytotoxicity' and to IL-2-activated enhancement of cytotoxic capacity (LAK development) do not require ADAP.
ADCC and target lysis induced by NK activating receptors do not require ADAP NK cells develop ADCC against some otherwise resistant target cells through cross-linking of FccRIIIA (CD16) receptors on effector cells by target cell-bound antibody (12). Given the structural and functional similarities of signal transduction through the TCR and FccRIIIA, and defects in TCR-dependent events in ADAP-deficient mice, we speculated that loss of ADAP might result in defective FccRIIIA-dependent signal transduction and impaired ADCC. We incubated killingresistant, Thy1 + , 51 Cr-labeled EL4 T cell tumor cell targets with either isotype-matched control or with anti-Thy1 antibody prior to mixing target cells and LAK effectors. As expected, we found that EL4 cells incubated with control antibodies resisted killing by either control or ADAP-deficient LAKs. Surprisingly, ADAP-null LAKs also killed anti-Thy1.1-bound EL4 cells with efficiency comparable to control LAKs (Fig. 4A).
The NKRs Ly49D and 2B4 surface antigens can deliver activating signals through distinct biochemical cascades, leading to cytotoxic responses. Ly49D associates with DAP12; when ligated by target MHC class I molecules on CHO cells, Ly49D is responsible for induction of cytotoxicity (21,32). 2B4 is a member of the SLAM family of surface receptors that can mediate NK cytotoxicity following signaling that requires the presence of the signaling lymphocytic activation moleculeassociated protein (SAP) adaptor molecule (25). SAP association with the ADAP-binding partner fyn is critical for signaling through SLAM in T cells (33). We assessed integrity of Ly49Dmediated signaling in ADAP-deficient LAKs in a cytotoxicity assay with CHO cell targets. As shown in Fig. 4(B), ADAPdeficient LAKs lysed CHO target cells with efficiency similar to control LAKs. Furthermore, we found no difference between ADAP-deficient and control LAKs in capacity to kill resistant FcR+ P815 targets bound to anti-Ly49D antibodies in a reverse reverse antibody-dependent cellular cytotoxicity (rADCC) assay (data not shown). Signaling through 2B4Rs was assayed in rADCC by LAK against P815 target cells. We observed comparable, anti-2B4 dependent, enhanced lysis of NK-resistant P815 cells by either control or ADAP-deficient LAKs (Fig. 4C). Together, these data indicate that cytotoxicity signaling downstream of key Ly49 and SLAM family NK cell activating receptors on LAKs does not require ADAP.

Ly49A-mediated NK inhibition does not require ADAP
Interactions between MHC class I molecules and cognate inhibitory receptors on NK cells constitute a principal mechanism whereby NK cells are prevented from damaging host tissues (13). Ly49A is a prototypic, inhibitory receptor that mediates inhibition of NK cytotoxicity when ligated by H-2D d expressed on candidate target cells (34). Inhibitory signaling through Ly49 family members involves diminished phosphorylation of adaptors (such as ADAP-binding partner SLP-76) that, like ADAP, are implicated in TCR signaling (13,35). Thus, we examined ADAP-deficient cells for their capacity to respond to inhibitory signals through Ly49A. Control or ADAP-deficient Ly49A+ LAKs were purified by panning with anti-Ly49A antibody. The Ly49A+ cells (>90% pure) were then assayed for their ability to kill the Tcell lymphoma C1498 stably transfected with H-2D d (C1498.D d ) (19). As shown in Fig. 5, both ADAP-deficient and control Ly49A+ LAKs showed poor cytolytic activity toward C1498.D d . ADAP-null and wild-type LAKs also showed equivalent robust cytotoxic activity against C1498 parental targets (data not shown), indicating the specificity of target-expressed H-2D d in mediating inhibition of lysis. While isotype control mAb had no effect (data not shown), addition of anti-Ly49A-blocking antibody resulted in enhanced C1498.D d lysis by both control and ADAP-null effectors (Fig. 5). These results confirm that Ly49A engagement by H-2D d mediates inhibition in the assay. Further, ADAP is not required for signaling function of Ly49A.

ADAP is dispensable for IFN secretion by NK cells
Production of IFN-c in response to inflammatory signals such as IL-12 constitutes a principal function of NK cells (36).
Recently, SLAM family member receptors such as 2B4 have also been shown to mediate IFN-c production by NK cells (25). To examine the role of ADAP in NK responses to these stimuli, we investigated the role of ADAP in IFN-c secretion. We cultured control or ADAP-deficient LAKs for 24 h with medium alone, with anti-2B4 antibody-either cell bound or solubleor with recombinant murine IL-12. IFN-c secretion was measured in cell supernatants. 2B4 antibody stimulated equivalent levels of IFN-c by both control and ADAP-deficient LAKs (Fig. 6A). Likewise, Fig. 6(B) shows that ADAP-deficient LAKs produced IFN-c at wild-type levels when stimulated with IL-12.  Together, these results indicate that ADAP-deficient NK cells can secrete IFN-c normally in response to IL-12 and to 2B4 stimulation.

LFA-1-mediated NK target adhesion does not require ADAP
ADAP-deficient T cells display defective adhesion to integrin ligands (6,7). Integrins such as LFA-1 have been shown to be critical for NK adhesion and cytolytic function (37). We therefore studied the ability of ADAP-deficient LAKs to form LFA-1-dependent conjugates with target cells. We first established that LFA-1 expression on LAKs does not depend upon ADAP (Fig. 7A). Next, fluorescently labeled LAKs and targets were mixed, and the percentage of LAKs bound to target cells according to FACS analysis was observed over time. We found a time-dependent increase in the percentage of LAKs forming conjugates with YAC-1 targets (Fig. 7B). Pretreatment of LAKs with anti-LFA-1 reduced the percent conjugate formation by~50%. We found no difference in conjugate-forming behavior between control and ADAPdeficient LAKs. These data reveal that ADAP is dispensable for LFA-1-dependent effector-target adhesion in LAKs.

NK cell-mediated tumor rejection proceeds independently of ADAP
NK cells participate in host defense against malignancy. A critical role for NK cells in host defense against progression of the metastatic tumor B16 has been well established (26). To test the role of ADAP in NK-mediated tumor rejection, we studied the capacity for ADAP-null mice to control B16 pulmonary metastases. We injected control or ADAP-deficient mice with B16 cells by tail vein. To confirm that NK1.1+ cell populations were mediating suppression of B16 metastases, we also injected some animals with depleting anti-NK1.1 mAb 3 days before inoculation with B16. After 21 days, lungs were removed, and numbers of pulmonary metastases were quantified. In both ADAP-sufficient and -deficient mice, anti-NK1.1 treatment resulted in marked increase in the size and number of pulmonary tumor nodules evident at day 21 compared with control IgG-treated animals (Fig. 8A). We found that numbers of visible pulmonary metastases were comparable in the lungs of control and ADAP-deficient mice (Fig. 8B). These data indicate that ADAP deficiency does not impair in vivo tumor suppressive capacity attributable to NK cells.

Discussion
This study presents evidence that the hematopoietic adaptor ADAP is dispensable for development and major functions of NK cells. Splenocytes bearing NK surface numbers express ADAP; however, lymphoid tissues from ADAP-deficient mice contain normal populations of DX-5 + CD3 ÿ (and DX-5 + CD3 + ) lymphocytes. ADAPÿ/ÿ splenocytes kill standard NK targets. Furthermore, similar to cells from control mice, ADAP-deficient LAKs show intact cytolytic function induced through a number of activating and inhibitory receptors, including ITAM-associating FcRcR and Ly49DR and immunoreceptor tyrosine-based inhibition motif (ITIM)-containing Ly49A.  Cytokine-and activating receptor-stimulated IFN-c production occur in the absence of ADAP, and ADAP-null mice suppress tumor metastasis in an NK cell-dependent manner. Together, these data suggest that signals required for NK development and function can proceed independently of ADAP.
ADAP is required for efficient thymocyte development and T cell population of the periphery (6)(7)(8). Although T cells and NK cells share a common lymphoid progenitor in the bone marrow, their molecular developmental requirements differ. T cell maturation requires pre-TCR-and TCR-mediated signal transduction in addition to signals mediated by select cytokines [e.g. IL-7 (38)]. In contrast, cytokine or growth factor signaling components, but not antigen receptor signaling proteins [excepting MHC class I ligands for NKRs (39)], are required for NK development (40). Normal NK development and distribution in ADAP-deficient mice and normal capacity for LAK generation from ADAP-null splenocytes suggest that ADAP does not regulate signaling by cytokines that utilize the common gamma chain [IL-2 and IL-15 (41)]. These results are perhaps not surprising, since most functional data reported to date place ADAP in a signaling niche downstream of ITAM-associated molecules [e.g. TCR (5) and the FceRI (9)] rather than of cytokine receptors.
FccRIIIA, the low-affinity IgGR on NK cells, associates with ITAM-bearing adaptor proteins CD3f and/or FceRIc. FccRIIIA can engage activation of src and syk family kinases, calcium mobilization and ERK activation, which in turn culminate in cytokine secretion, degranulation and killing function (12). All TCR-dependent proximal signaling events studied to date are intact in ADAP-deficient T cells, despite markedly impaired TCR-stimulated IL-2 production and proliferation (6,7). In contrast to TCR-mediated functions, however, FccRIIIAdependent cytotoxicity (ADCC) proceeds normally in the absence of ADAP (Fig. 4).
Previous work showed that ADAP is also dispensable for signaling through FceRI, another FcR expressed on myeloid (mast) cells (42). Interestingly, the ADAP-associated adaptor SLP-76, while dispensable for NK function (24), is required for FceRI function in mast cells (43). Furthermore, the C-terminal portion of SLP-76 that mediates ADAP association is required  for FceRI signaling (42). Together, these results suggest that a SLP-76 SH2 domain binder other than ADAP plays a role in FceRI function.
The unexpected differential in requirements for ADAP between TCR and FcR signaling suggests that ADAPregulated events entrained by the TCR differ from FcRmediated events leading to NK cytotoxicity or to mast cell function. Nuclear factor-jB (NF-jB) signaling is a candidate cascade in this regard. Recent genetic studies of molecules that regulate NF-jB suggest that NF-jB activity is required for certain antigen receptor-induced functions like T cell proliferation (44), yet is dispensable for FcR-mediated NK cytotoxicity (45). Although ADAP is dispensable for TCRmediated PKCh phosphorylation [upstream NF-jB regulator in T cells (5)], further examination of a potential role for ADAP in NF-jB activation is needed.
Ly49D is the prototype member of a family of activating NK cell receptors that signal through the ITAM-containing adaptor DAP12 (13). Similar to the TCR and FcR, Ly49D engagement triggers activation of syk family kinases that are implicated in ADAP phosphorylation. Ly49D cross-linking also induces mobilization of calcium and ERK activation, and mediates rADCC. Given these similarities between Ly49D and antigen receptor signaling, the apparent lack of a role for ADAP in regulating Ly49D-mediated cytotoxicity is unexpected. One explanation may lie in the observation that there is considerable redundancy at the level of the proximal mediators (adaptors and enzymes) that transduce signals from ITAMassociated receptors like Ly49D (13). The ADAP-binding partner SLP-76 and the transmembrane adaptor linker for activated T cells (LATs), both critical for TCR signaling, are phosphorylated after activating receptor engagement (46). However, neither SLP-76 nor LAT is alone required for NK activating receptor function (24,47).
Functional redundancy also appears to characterize the signaling machinery engaged by inhibitory receptors like Ly49A. Ly49A shares sequence homology with activating Ly49 family members, but contains a cytoplasmic domain ITIM that recruits SHP-1 en route to blocking signaling through NK activating receptors. Both SLP-76 and LAT can be dephosphorylated by SHP-1 during inhibitory receptor function (46). These data suggest a potential intersection between activating and inhibitory pathways that could be redundantly regulated by hematopoietic adaptors like SLP-76, LAT and ADAP. While the status of Ly49-dependent inhibitory signaling has not been reported for NK cells deficient for SLP-76 or LAT, the present study showing dispensability of ADAP for inhibitory signaling is consistent with the notion that overlapping pathways may regulate inhibition.
The absence of a role for ADAP in NKT development and 2B4-mediated NK functions is interesting in light of ADAP physical association with the src family kinase fyn (1). Fyn is required for the development and function of NKT cells (29), and is known to associate with the adaptor SAP. SAP is also required for NKT development and for signaling through members of the SLAM receptor family that promote T h 2 cytokine production (48). Additionally, both fyn and SAP have recently been shown to be critical for NK cytotoxic and cytokine secretion functions mediated by SLAM family member 2B4 (25). However, our results indicate that neither NKT development (Fig. 1) nor 2B4 signaling in LAKs (Figs 4 and 6) require ADAP, arguing against a role for ADAP in fynmediated or 2B4-induced signaling in NK cells.
TCR-stimulated integrin-mediated adhesion is defective in ADAP-deficient T cells (6,7). Recent studies show that the b2 integrin LFA-1 likely functions to promote both adhesion and cytotoxic function of NK cells (37,49). Indeed, LFA-1 is required for efficient LAK cytotoxicity against a range of tumor targets, including the YAC-1 and RMA cells employed in the current study (50). Interestingly, ADAP is tyrosine phosphorylated after b1 integrin engagement on T cells (51), a stimulus that induces tyrosine kinase activation in NK cells (52). However, our observations of normal cytotoxic function against YAC-1 and RMA targets ( Fig. 3 and data not shown), and of normal conjugate formation with YAC-1 (Fig. 7) by ADAP-deficient LAKs, suggest that LFA-1-dependent functions in NK cells do not require ADAP.