The Notch Regulator Numb Links the Notch and TCR Signaling Pathways
http://www.100md.com
免疫学杂志 2005年第2期
Abstract
Both the Notch and TCR signaling pathways play an important role in T cell development, but the links between these signaling pathways are largely unexplored. The adapter protein Numb is a well-characterized inhibitor of Notch and also contains a phosphotyrosine binding domain, suggesting that Numb could provide a link between these pathways. We explored this possibility by investigating the physical interactions among Notch, Numb, and the TCR signaling apparatus and by examining the consequences of a Numb mutation on T cell development. We found that Notch and Numb cocluster with the TCR at the APC contact during Ag-driven T cell-APC interactions in both immature and mature T cells. Furthermore, Numb coimmunoprecipitates with components of the TCR signaling apparatus. Despite this association, T cell development and T cell activation occur normally in the absence of Numb, perhaps due to the expression of the related protein, Numblike. Together our data suggest that Notch and TCR signals may be integrated at the cell membrane, and that Numb may be an important adapter in this process.
Introduction
The importance of signals through the pre-TCR and TCR in cell fate decisions during T cell development is well established (reviewed in Refs. 1 and 2). More recently, signaling through the Notch receptor has also been implicated in both early (3, 4, 5) and late (6, 7, 8) cell fate decisions during T cell development (reviewed in Refs. 9, 10, 11). Notch signals appear to be critical for T cell development because T cells fail to develop from Notch1-deficient bone marrow precursors (12, 13). In addition, deletion of Notch1 at the DN2/3 stage of T cell development results in a block in T cell development (3). This block is associated with an impairment of V to DJ rearrangement at the TCR locus in DN3 stage thymocytes. Given the importance of Notch in cell fate decisions during T cell development, it is important to address how Notch activity may be regulated during this process.
One of the proteins that has been implicated as a regulator of Notch is Numb (14, 15). Numb is an adapter protein that contains an N-terminal phosphotyrosine binding (PTB)4 domain and a C-terminal proline-rich region containing several putative Src homology 3 binding domains. Numb was first identified as a gene controlling cell fate specification in development of the Drosophila peripheral nervous system. Further studies of Drosophila neural development showed that Numb acts by antagonizing Notch signals. Recent reports suggest that Numb may regulate Notch activity by promoting the down-regulation and/or degradation of Notch receptor (16, 17). In mammals, two homologues of Numb, Numb and Numblike, have been described (18, 19, 20). Because Numb is expressed in the thymus (21), it is a good candidate for regulation of Notch activity during T cell development.
Notch receptors and ligands are also expressed on cells of the peripheral immune system (22, 23, 24) and appear to have a role in T cell responses. Some studies suggest that Notch may have a role in inducing regulatory T cells (22, 25, 26, 27). Other studies suggest a role for Notch in modulating T cell responses (28, 29, 30) and affecting Th1/Th2 differentiation (31, 32). Thus, it appears that Notch also plays a role in peripheral T cell responses, although the exact mechanism and effects of Notch are still unclear.
In this study we provide evidence for the physical association of TCR, Notch, and Numb in both immature and mature T cells. We also explored the role of Numb in T cell development by generating mice in which Numb is deleted specifically in T cells, and we found that there is no appreciable effect of Numb deficiency on T cell development, perhaps due to redundancy with the homologous protein Numblike. Our results provide a basis for future examination of the roles of Notch and Numb in modulating TCR signaling.
Materials and Methods
Mice
Lck-Cre and CD4-Cre transgenic mice (33) were provided by C. Wilson (University of Washington). Numbflox/flox and Numb+/– mice have been previously described (34). AND TCR transgenic (H-2b) mice were purchased from The Jackson Laboratory. AND TCR transgenic RAG–/– (H-2d) mice were purchased from Taconic Farms. All animals were housed in accordance with the guidelines established by the animal care and use committee at University of California, Berkeley.
Colocalization assay
DCEK.ICAM cells (I-EK) (35) were cultured on four-well, poly-L-lysine-coated slides at 105 cells/well in complete RPMI 1640 and incubated overnight at 37°C. Twelve to 16 h before the addition of T cells, DCEK.ICAM cells were either pulsed with 10 μM pigeon cytochrome c (PCC) peptide or left untreated. Thymocytes (106) from AND TCR transgenic (H-2d) mice or 106 lymph node cells (LNC) from AND TCR transgenic (H-2b) mice were then added with or without freshly added peptide and cultured for 20 min. After incubation, slides were spun down at 700 rpm for 5 min. Cells were then fixed with 4% paraformaldehyde for 15 min at 4°C, permeabilized, and stained with Abs against Numb, Notch, and CD3. The Abs used are affinity-purified rabbit antisera generated against the C terminus of murine Numb (aa 489–524) as previously described (19), goat anti-rabbit Alexa-546, goat anti-rabbit Alexa-488, goat anti-hamster Alexa-546 (Molecular Probes), anti-CD3-allophycocyanin, anti-CD4 (BD Biosciences; which we conjugated in-house to Alexa-546 according to the manufacturer’s protocol (Molecular Probes)), and anti-Notch (clone 8G10; see below). T cell/APC conjugates were visualized on a Nikon E800 fluorescence microscope retrofit for three-dimensional capture. Images were deconvolved using SlideBook software (Intelligent Imaging Innovations).
Isolation of mAb reactive with mouse Notch1
A hamster mAb against mouse Notch1 (clone 8G10) was generated by immunizing a Syrian hamster with a bacterial fusion protein consisting of the region from epidermal growth factor repeat 33 through LIN-12, Notch, GLP-1 repeat 1 (nucleotides 3898–4476) of mouse Notch1 fused to GST. Spleen cells were fused with a hamster hybridoma partner, and the clones were screened by ELISA for reactivity with the fusion protein. Positive clones were then recloned and screened for reactivity against full-length mouse Notch1 by immunoblotting. Anti-Notch1 Ab was generated from hybridoma culture supernatant and was either purified on a protein G column or used as a concentrated culture supernatant. The 8G10 Ab reacts with mouse and rat Notch1, but not with other Notch family members (data not shown). The Ab has been licensed to Upstate Biotechnology.
Immunoprecipitation and immunoblotting
For immunoprecipitations, 108 thymocytes from either C57BL/6 or AND-TCR transgenic (H-2d) mice were lysed in lysis buffer (20 mM Tris-HCl (pH 8.0), 138 mM NaCl, 9.5% glycerol, 1 mM sodium orthovanadate, 2 mM EDTA, 10 mM NaF containing either 1% Brij 58 detergent (Pierce), or 1% Nonidet P-40 (Sigma-Aldrich)) supplemented with mini complete EDTA-free protease inhibitor mixture (Roche). Cell lysates were precleared with protein G agarose beads (Sigma-Aldrich) bound with rabbit anti-GST for 2 h at 4°C. Samples were then briefly centrifuged, and the supernatant was added to protein G agarose beads bound with Numb antisera and incubated overnight at 4°C. Samples were separated on a 10% SDS-PAGE gel, transferred onto nitrocellulose (Hybond ECL; Amersham Pharmacia), and immunoblotted. For kinetic analysis of Numb associations after TCR stimulation, DCEK.ICAM cells were plated at 105 cells/ml in complete RPMI 1640 and incubated overnight at 37°C. Four hours before the addition of T cells, DCEK.ICAM cells were either pulsed with 10 μM PCC peptide or left untreated. Thymocytes (1.5 x 108) were then added with or without freshly added peptide and cultured for 5, 15, or 20 min at 37°C. Thymocytes were then washed off and immediately lysed as described above. Cells (106) were removed from the samples for the lysate control. The remaining cells were immunoprecipitated with Numb as described above. For the analysis of Numb deletion, cell lysates were prepared by resuspending 106 cells in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris (pH 7.4), and 0.02% NaN3) supplemented with mini complete protease inhibitor mixture (Roche). Cell lysates were separated as described above, transferred onto nitrocellulose, and immunoblotted using Numb antisera. Where indicated, CD3+ T cells were purified using T cell enrichment columns (R&D Systems). Abs used for immunoblotting are as follows: affinity-purified rabbit antisera generated against the C terminus of murine Numb (aa 489–524), anti-phosphotyrosine (4G10), anti-Lck (clone 3A2), anti-Vav, anti-GST (Z-5), anti-c-Cbl (A-9; Santa Cruz Biotechnology), and anti-CD4 (L3T4; BD Biosciences). Bound Abs were detected with goat anti-rabbit HRP-conjugated secondary Ab (Southern Biotechnology Associates) and goat anti-mouse IgG-HRP (Jackson ImmunoResearch Laboratories) and visualized using ECL (Amersham Pharmacia).
Flow cytometry
Single-cell suspensions from thymus, lymph node, and spleen were prepared and stained with the following Abs: anti- TCR, anti-CD3 labeled with FITC (BD Biosciences/eBiosciences), anti-CD19, labeled with PE (BD Biosciences/eBiosciences), anti-CD4 labeled with PE-Texas Red (Caltag Laboratories), and anti-CD8 labeled with PE-Cy5 (eBiosciences). All data was collected on a Coulter EPICS-XL cytometer (Beckman Coulter) and analyzed with FlowJo software (TreeStar).
Assays for T lymphocyte function
Activation with anti-CD3. CD3+ T cells were purified from lymph nodes using CD3 enrichment columns (R&D Systems) CD3+ T cells (1 x 104–5 x 104/well) were then cultured in triplicate in flat-bottom 96-well plates that had been coated with anti-CD3 (BD Biosciences). Cells were cultured 48 h and pulsed with 1 μCi/well [3H]thymidine for the last 16 h. Cells were then harvested, and [3H]thymidine incorporation was determined in a rack beta scintillation counter (Beckman Coulter).
LNC proliferation. Mice were immunized s.c. at five sites with 100 μg of OVA (Sigma-Aldrich) emulsified in CFA (Difco). Draining lymph nodes were harvested 10 days later. LNCs (4 x 105/well) were cultured in triplicate in 96-well, round-bottom plates in the presence of various concentrations of Ag for 48 h and pulsed with 1 μCi/well [3H]thymidine for the last 16 h. [3H]thymidine incorporation was determined as described above.
Homeostatic proliferation. CD3+ lymph node T cells were purified from Numb–/– mice and littermate controls (Ly5.2) using T cell enrichment columns (R&D Systems). Cells were then labeled with 2 μM CFSE (Molecular Probes) and transferred i.v. (106/mouse) to B65.1 (Ly5.1) that had been irradiated with 400 rad 24 h earlier. On day 6 post-transfer, spleen cells were harvested and stained with Abs to CD4, CD8, and Ly5.1. Data were collected on a Coulter EPICS XL cytometer (Beckman Coulter) and analyzed as described above.
RT-PCR
Total RNA was extracted from either thymocytes or purified CD3+ LNC using TRIzol (Invitrogen Life Technologies). RNA was treated with DNase I to eliminate contaminating genomic DNA before cDNA synthesis using SuperScript (Invitrogen Life Technologies). First-strand cDNA was then used in PCR. The following primers were used: -actin: forward, 5'-tggaatctgtggcatccatgaaa c-3'; reverse, 5'-taaaacgcagctcagtaacagtccg-3'; and Numblike: forward, 5'-ctgaaaccttcaggacggag-3'; reverse, 5'-cacaggacagacttcacgga-3'. PCR products were visualized by agarose gel electrophoresis.
Results
Physical association of Notch and Numb with the T cell signaling apparatus
Given the likely involvement of the Notch and TCR signaling pathways in both immature and mature T cell responses, we hypothesized that these pathways may be physically linked in T cells. To test this hypothesis, we examined the localization of Notch and CD3 in lymph node T cells from AND TCR transgenic (H-2b) mice after exposure to peptide-pulsed APCs by fluorescence microscopy. We found that Notch is recruited to the T cell/APC interface and coclusters with CD3 after Ag exposure (Fig. 1A). Because Numb has been reported to associate with the intracellular region of Notch (15, 19), we examined whether Numb also coclusters with CD3 after Ag exposure. As shown in Fig. 1B, Numb is also recruited to the T cell/APC interface and coclusters with CD3 after Ag exposure.
FIGURE 1. Coclustering of Notch and Numb with TCR at the T cell/APC interface in response to Ag stimulation. DCEK.ICAM cells were cultured on poly-L-lysine-coated slides overnight at 37°C. Twelve to 16 h before the addition of T cells, DCEK.ICAM were pulsed with 10 μM PCC peptide or were left untreated. LNC or thymocytes (106) were then added and cultured for 20 min. After incubation, slides were spun down, and cells were fixed, permeabilized, and stained with Abs against Numb, Notch, and CD3. T cell/APC conjugates were then visualized by fluorescence microscopy. The rightmost panels show the brightfield image with the fluorescent image superimposed. Colocalization of red and green signals appears as yellow. A and B, Data from LNC from AND TCR transgenic (H-2b) mice. C–E, Data from thymocytes from AND TCR transgenic RAG–/– (H-2d) mice.
To determine whether Notch, Numb, and CD3 also cocluster at the T cell/APC interface in thymocytes in an Ag-dependent manner, we used thymocytes from AND TCR transgenic (H-2d) mice. Because the AND TCR is not selected in mice carrying the H-2d haplotype, thymocytes in these mice are arrested at the double-positive stage of T cell development, thereby providing a source of naive thymic precursors that have not yet received an Ag receptor signal. We exposed thymocytes from these mice to peptide-pulsed APCs and visualized the localization of Notch, Numb, and CD3. Because the levels of Notch and Numb are lower on thymocytes relative to the levels on APCs, coclustering is more difficult to discern than in mature T cells. Nevertheless, we found that Notch and Numb were also recruited to the T cell/APC interface and coclustered with CD3 in thymocytes after Ag exposure (Fig. 1, C and D). To ensure that the coclustering observed was specific, we examined the localization of CD8. Because the AND TCR is an MHC class II-restricted receptor, CD8 would not be part of the TCR complex in these cells. As expected, CD3 redistributed and clustered at the T cell/APC interface in an Ag-dependent manner, whereas CD8 remained evenly distributed over the cell surface (Figs. 1E and 2B).
FIGURE 2. Quantitation of immunofluorescence data. A, T cell/APC conjugates were identified using brightfield imaging. After images were recorded, clustering of the indicated proteins at the T cell/APC interface in the presence or the absence of Ag was scored. Values are expressed as the percentage of total conjugates scored. The left panel shows data from nine (–Ag) and 11 (+Ag) conjugates. The right panel shows data from 13 (–Ag) and 14 (+Ag) conjugates. The slightly increased colocalization of CD3 observed in the experiment shown in the right panel may reflect slight differences in the rate of T/APC conjugate formation. B, The pixel intensity of each indicated protein over the surface of the whole thymocyte and at the thymocyte/APC contact in the presence or the absence of Ag was determined using SlideBook software. Values indicate the percentage of signal at the thymocyte/APC contact relative to the signal over the whole cell. Data shown for CD3 are the average of 18 (+Ag) and 12 (–Ag) conjugates. The data shown for CD8 are the average of nine (+Ag) and four (–Ag) conjugates. The data shown for Numb are the average of nine (+Ag) and eight (–Ag) conjugates.
We used two approaches to quantify the coclustering of Notch and Numb with CD3. First, we identified mature T cell/APC conjugates using brightfield, recorded fluorescent images and then scored them for clustering of CD3, Notch, Numb, CD3 plus Notch, or CD3 plus Numb at the T cell/APC interface in the presence or the absence of Ag. For CD3 and Notch, 30% of conjugates had CD3 plus Notch clustered at the interface. No clustering of CD3, Notch, or CD3 plus Notch was observed in the absence of Ag (Fig. 2A). For the experiments examining the clustering of CD3 and Numb, some clustering of CD3 or Numb alone to the interface was observed in the presence and the absence of Ag. However, coclustering of CD3 plus Numb was only observed in the presence of Ag (Fig. 2A). Second, we chose representative examples of thymocyte/APC conjugates and determined the pixel intensity of the CD3, CD8, and Numb signals at the T cell/APC interface relative to the pixel intensity over the whole cell. The percentage of CD3 and Numb signals that localized at the T cell/APC interface in thymocytes increased in the presence of Ag, whereas the amount of CD8 signal at the interface remained the same in the presence or the absence of Ag (Fig. 2B). Thus, it appears that both Notch and Numb are recruited to the T cell/APC interface and cocluster with the TCR complex upon Ag exposure in immature and mature T cells, suggesting a physical association between the Notch and TCR signaling pathways in these cells.
Given that Numb has the structural features of an adapter protein, we hypothesized that it may provide a link between the TCR and Notch signaling pathways. To investigate this, we examined what proteins interact with Numb in thymocytes. Because Numb contains a PTB domain, we first looked for associated proteins containing phosphorylated tyrosines by immunoprecipitation with Numb antisera, followed by immunoblotting with an anti-phosphotyrosine Ab. We observed several prominent tyrosine-phosphorylated proteins in the Numb immunoprecipitates that were not present in control immunoprecipitations using species-matched antisera of irrelevant specificity (preclear lane, Fig. 3A). Upon reprobing the blot with specific Abs, we found that Lck (Fig. 3B), c-Cbl (Fig. 3C), and Vav (Fig. 3D), coprecipitated with Numb in thymocytes. Similar results were obtained in peripheral T cells (data not shown). The comigration of these bands with prominent phosphotyrosine-containing proteins together with previous reports that these proteins are tyrosine phosphorylated in T cells suggest that the indicated bands in the anti-phosphotyrosine blot correspond to phospho-c Cbl, phospho-Lck, and phospho-Vav (Fig. 3A). We also probed the blot with anti-CD4 and found CD4 in the immunoprecipitate (Fig. 3E) as expected, because Lck is known to associate strongly with CD4 (36). Thus, Numb appears to exist in a preformed complex with TCR signaling proteins, and this complex gets recruited to the T cell/APC interface in an Ag-dependent manner.
FIGURE 3. Numb interacts with components of the T cell signaling complex in thymocytes. B6 thymocytes (108) were lysed, immunoprecipitated with antisera generated against the C terminus of Numb, and analyzed by immunoblotting using Abs against the indicated proteins. The preclear Ab is an irrelevant Ab species-matched to the IP Ab as a control for nonspecific binding. Abs for immunoblotting were specific for anti-phosphotyrosine (A), Lck (B), c-Cbl (C), Vav (D), and CD4 (E).
To provide evidence for the functional relevance of these interactions, we determined whether the interaction of Numb with any of these proteins changes upon TCR stimulation. To do this, AND TCR transgenic (H-2d) thymocytes were exposed to APCs with or without Ag for 5, 15, or 20 min. Cells were then lysed, lysates were immunoprecipitated with Numb antisera, and immunoprecipitates were analyzed by immunoblotting. As shown in Fig. 4, the interaction of Numb with phospho-Lck and phospho-c-Cbl increased at 5 min, followed by a progressive decrease at 15 and 20 min (Fig. 4). The change in association with Numb was not due to a change in the total amounts of Lck and c-Cbl in the cells, because the amounts of these proteins in the lysate lanes did not change over time (Fig. 4). Taken together, our data suggest that TCR, Notch, and Numb are physically linked in T cells.
FIGURE 4. Interaction of Numb with Lck and c-Cbl changes upon TCR stimulation. DCEK.ICAM cells were cultured overnight at 37°C. Four hours before addition of T cells, DCEK.ICAM were pulsed with 10 μM PCC peptide or were left untreated. Thymocytes from AND TCR transgenic RAG–/– (H-2d) mice were then added for 5, 15, or 20 min. Thymocytes were harvested, lysed, and immunoprecipitated with Numb antisera. A, Anti-phosphotyrosine blot and c-Cbl blot. B, Quantification by densitometry of the indicated proteins present in the lysate (left panel) and the Numb immunoprecipitates (right panel). This experiment was repeated three times with similar results. Representative data are shown.
Conditional deletion of the Numb gene in thymocytes
Given our data implicating Numb as a link between the TCR and Notch signaling pathways, we chose to address the role of Numb at different stages of T cell development. Because germline deletion of Numb leads to lethality on embryonic day 11.5 (34, 37), we used Cre/lox technology to target deletion of the Numb gene specifically to developing T cells. To achieve this we crossed Numbflox/floxmice (34) with mice expressing the Cre recombinase under the control of the Lck proximal and CD4 promoters. These promoters have been shown to effect deletion at the DN2/3 and DN3/4 stages of thymic development, respectively (3, 33, 38). We generated mice in which one allele of Numb was already deleted by first crossing Numb+/– (34) mice to Cre transgenic mice and subsequently crossing progeny from this cross to Numbflox/flox mice. To assess the extent of deletion of Numb, we extracted total genomic DNA from unfractionated thymocytes and examined the Numb locus by PCR using primers that distinguish wild-type (WT), floxed, and deleted Numb alleles (34). Fig. 5A shows a representative example of the deletion of Numb. In thymocytes from Lck Cre– Numbfl/+ mice, both the WT and floxed alleles are detected; however, in the Lck-Cre+ Numbfl/– (Numb-deficient) littermate, the deleted allele is the predominant product with a very low level of floxed allele detected. This residual Numb could be due to the presence of DN cells that have not yet deleted Numb, or it could be the result of incomplete deletion of Numb. We next examined Numb protein expression in both unfractionated thymocytes and purified peripheral T cells by immunoblotting (Fig. 5B). We estimate that the deletion of Numb protein in thymocytes and mature T cells from Numb-deficient mice is >95%, with deletion occurring to a similar extent in Numb-deficient mice generated under the Lck-proximal and CD4 promoters. We also observed the presence of a lower m.w. band in lysates from Numb+/– and Numb-deficient mice that was not present in lysates from Numb+/+ or floxed mice (Fig. 5B and data not shown). Sequence analysis of RT-PCR products from these mice indicates that this minor band corresponds to a splice around the targeted exons of Numb (data not shown). This mutant Numb protein is unlikely to be functional given that it lacks a significant portion of the PTB domain and an independently generated Numb-deficient mouse displays a similar embryonic phenotype (34, 37).
FIGURE 5. Deletion of Numb. A, Analysis of genomic DNA by PCR in unfractionated thymocytes from control Lck-Cre–Numbfl/+ and Lck-Cre+Numbfl/– (Numb-deficient) mice. Arrows indicate the presence of WT, floxed, and targeted Numb alleles. B, Analysis of Numb protein in unfractionated thymocytes and CD3+ peripheral LNC from Lck-Cre- Numbfl/+ and Lck-Cre+Numbfl/– (Numb-deficient) mice. Lysates (106 cell equivalents/lane) were separated on a 10% SDS-PAGE gel and blotted with polyclonal rabbit antisera generated against the C terminus of Numb (aa 489–524).
T cell development in the absence of Numb
We examined T cell development in mice in which Numb deletion was driven by CRE expressed under the control of the Lck proximal and CD4 promoters. In Lck-Cre+ Numbfl/ –mice, we observed a modest block between DN3 and DN4 that could be attributed to the presence of the Lck Cre transgene because it was also observed in Lck-Cre+ Numb+/+ controls (data not shown). Similarly, a reduction in the number of peripheral T cells and a skew in the ratio of CD4/CD8 T cells in Lck-Cre+ Numbfl/ –mice could also be attributed to the presence of the Lck Cre transgene alone (Fig. 6, A and B). In addition, we did not observe any defects in thymic development or in the periphery of Numb-deficient mice generated under the CD4 promoter (Fig. 6, C and D, and data not shown). Taken together these data suggest that T cell development can proceed normally in the absence of Numb.
FIGURE 6. Peripheral lymphocyte subsets in Numb-deficient mice. B/T (A and C) and CD4/8 (B and D) ratios in Lck–Cre+Numb-deficient mice (Lck–Cre+Numbfl/–) and controls (Lck–Cre–Numbfl/–, Lck–Cre–Numbfl/+, Lck–Cre+Numb+/+) and in CD4–Cre+Numb-deficient mice (CD4–Cre+Numbfl/–) and controls (CD4+Cre–Numbfl/+). The B/T ratio is based on the number of spleen cells staining positively for CD19 and CD3 as determined by flow cytometry. The CD4/CD8 ratio is based on the number of spleen cells staining positively for CD4 and CD8 as determined by flow cytometry. Points represent values from individual mice, and bars represent mean values. All data are from mice <30 wk of age.
Peripheral T cell function in the absence of Numb
Given our data implicating Numb in TCR signaling, we examined next the ability of Numb-deficient T cells to respond to signals through the TCR. As shown in Fig. 7A, Numb-deficient T cells proliferate equally well in response to plate-bound anti-CD3 as do Numb WT T cells. To examine the ability of Numb-deficient T cells to respond to a more physiologic stimulus, we assessed the in vitro recall response to OVA after immunization with OVA in CFA. As shown in Fig. 7B, there is no significant difference in the ability of Numb-deficient T cells to proliferate in response to Ag stimulation relative to Numb WT littermate controls. We also found no significant differences in the production of IFN- by Numb-deficient T cells relative to littermate controls (data not shown). Finally, we examined the ability of Numb-deficient T cells to respond to homeostatic signals that regulate T cell survival and drive proliferation in response to lymphopenia (reviewed in Refs. 39 and 40). We transferred CFSE-labeled peripheral T cells isolated from Numb WT or Numb-deficient mice to irradiated recipients and analyzed the ability of T cells to undergo homeostatic proliferation by comparing CFSE dilution profiles. We found that Numb-deficient T cells and Numb WT T cells undergo homeostatic division to a similar degree (Fig. 8). We also examined the phosphorylation of TCR-, which is constitutively phosphorylated in normal T cells in response to environmental survival signals (41), and found that its phosphorylation is similar in Numb-deficient T cells compared with T cells from littermate controls (data not shown). Taken together, these data suggest that Numb deficiency does not affect peripheral T cell function and homeostasis.
FIGURE 7. Numb-deficient T cells respond normally to stimulation. A, Proliferative response of Numb-deficient T cells to plate-bound anti-CD3. Purified CD3+ T cells were cultured in triplicate with the indicated amounts of anti-CD3. [3H]thymidine was added at 48 h, and cells were harvested 16 h later. Data are shown as the mean cpm of triplicate wells, where cpm is the mean cpm in test wells – the mean cpm in wells with medium only. An independent experiment gave similar results. B, Recall response to OVA. LNCs were harvested from immunized mice and tested in triplicate over a dose response of 0.1–100 μg/ml OVA. [3H]thymidine was added at 48 h, and cells were harvested 16 h later. Data are shown as the mean cpm of triplicate wells, where cpm is the mean cpm in test wells – the mean cpm in wells with medium only.
FIGURE 8. Homeostatic proliferation of Numb-deficient T cells. CD3+ lymph node T cells (106) from Numb-deficient mice and littermate controls (Ly5.2) were labeled with 2 μM CFSE and transferred i.v. to irradiated B65.1 (Ly5.1) mice. Histograms show the CFSE profile of either CD4+Ly5.1– or CD8+Ly5.1– cells. Each panel represents an individual B65.1 recipient.
Expression of Numblike in T cells
The lack of an appreciable effect of Numb deficiency on thymocyte development and peripheral T cell responses despite evidence for functional associations between Numb and the TCR signaling apparatus raises the possibility of functional redundancy by a related protein. Numblike is highly homologous to Numb (20), and low levels of Numblike have been shown to functionally compensate for Numb deficiency (42). We examined the expression of Numblike in both thymocytes and peripheral T cells and found that Numblike is expressed in both thymocytes and peripheral T cells (Fig. 9 and data not shown), providing a possible explanation for the normal T cell development and functional responses we observe in Numb-deficient mice.
FIGURE 9. Numblike expression in T cells. Numblike message was amplified by RT-PCR from CD3+ lymph node T cells. Amplification of -actin message from the same sample is shown as a control.
Discussion
The importance of the TCR and Notch signaling pathways in T cell development together with evidence for functional interplay between these two pathways prompted us to investigate how these two signaling pathways are linked. In this study we provide evidence that both Notch and Numb are recruited to the T cell/APC interface and cocluster with T cell signaling components in both immature and mature T cells upon exposure to Ag. We also provide evidence from coimmunoprecipitation experiments that Numb interacts with multiple components of the TCR signaling apparatus. Taken together, our data suggest that Notch and Numb may be part of the TCR signalosome in thymocytes and mature T cells. A recent report that Notch interacts with Lck in T cells also fits with this view (43).
A number of recent studies point to a role for Notch in modulating TCR signaling. In some studies, Notch activity was found to enhance TCR-induced proliferation by potentiating IL-2R expression (28) and NF-B activity (29), whereas another study found an inhibitory effect of Notch on TCR signaling (30). Notch has also been implicated in regulating the balance of Th1/Th2 cell development (31, 32). For thymocytes, we have found that Notch synergizes with TCR signals to regulate gene expression in thymocytes (44), whereas others have reported that Notch activity dampens TCR signals (45). Together, these studies point to a complex functional interplay between the TCR and Notch signaling pathways. Our data showing physical associations among TCR, Notch, and Numb provide a framework for further examination of how Notch and TCR signals could be influencing one another.
In addition to its role in regulating Notch, Numb has been implicated in the cellular processes of endocytosis and ubiquitination (16, 17, 46). Two recent reports suggest that these functions of Numb may be related to its ability to regulate Notch. First, Numb was found to associate with the ubiquitin ligase Itch, thereby promoting the ubiquitination and degradation of membrane-associated Notch receptor (16). Second, Numb was found to associate with -adaptin, a protein involved in receptor-mediated endocytosis, and this interaction was found to be important in down-regulating Notch expression (17). Given these considerations and our data suggesting Numb as a link between the TCR and Notch signaling pathways, several nonmutually exclusive models of how these pathways operate in T cells emerge. In the first model, Numb regulates Notch activity in T cells, and Notch acts directly to modulate T cell fate. Although this model fits well with classic studies of Numb and Notch in other systems, it does not take into account our data showing the interaction of Numb with components of the T cell signaling apparatus or the recent studies suggesting regulation of TCR signaling by Notch (discussed above). In a second model, Numb functions as an adapter protein in T cells and regulates T cell signaling as a part of the TCR signalosome. This model fits with the idea of Numb as an adapter protein and our interaction data, but does not incorporate the well-established role of Numb in regulating Notch. In a third model, Numb regulates Notch signaling, and Notch then regulates signals through the TCR. In considering these last two models, Numb could regulate TCR, Notch, or both by targeting them for ubiquitination and degradation (16, 17). Our observation that the ubiquitin ligase c-Cbl is part of the complex containing Numb fits with this idea. Although we did not observe any difference in the levels of cell surface TCR or Notch expression in Numb-deficient T cells (data not shown), it is still possible that Numb targets a subset of activated TCR or Notch receptors for endocytosis and degradation, in line with previous suggestions (16, 17). The extent to which any of these models may operate in T cells will be the subject of future experiments.
Despite these indications for a connection between Notch signaling and TCR signaling, we did not see an appreciable effect of Numb deficiency on peripheral T cell responses. One possible explanation for this is that Numblike compensates for lack of Numb. Numblike is highly homologous to Numb, and low level expression of Numblike has been shown to be functionally redundant with Numb during neural development (42) (Y.-N. Jan and W. Zhong, unpublished observations). Because, Numblike is expressed in thymocytes (data not shown) and peripheral T cells, it is likely that Numblike compensates for the absence of Numb in Numb-deficient T cells. This may explain the lack of a thymic defect in Numb-deficient mice. Another possibility is that the T cells that are seeded to the periphery in Numb-deficient mice represent those that have best adapted to function normally in the absence of Numb. One way to resolve these issues is to examine development and function in T cells deficient for both Numb and Numblike.
In summary, we have shown that the TCR and Notch signaling pathways may be connected in T cells via the adapter protein Numb. Given the indications that Notch signaling affects multiple steps in T cell development and the possibility of complex interplay between the Notch and TCR signaling pathways, it is not surprising that many different effects have been attributed to Notch in T cells. Clarification of the role of Notch in T cell development and peripheral T cell function may come from a better understanding of which effects of Notch are direct consequences of Notch target gene expression vs modulation of TCR signals. Our data provide a framework for future work in deciphering how Notch mediates its effects in both immature and mature T cells.
Acknowledgments
We acknowledge Mimi Mong and Xiao Jun-Kang for production of the Notch monoclonal 8G10, Alice Liao and Rakesh Ahuja for assistance with genotyping mice, Hector Nolla for assistance with cell sorting, and B. J. Fowlkes, Astar Winoto, and members of the Robey laboratory for comments on this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Heath Grant AI32985 (to E.R.). A.C.A. is supported by National Institutes of Heath Postdoctoral Fellowship F132AI50415-01.
2 A.C.A. and E.A.K. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Ana C. Anderson, Center for Neurologic Diseases, Brigham and Women’s Hospital and Harvard Medical School, HIM, 77 Avenue Louis Pasteur, Room 784, Boston, MA 02115. E-mail address: aanderson@rics.bwh.harvard.edu
4 Abbreviations used in this paper: PTB, phosphotyrosine binding; DN, double negative; LNC, lymph node cell; PCC, pigeon cytochrome c; WT, wild type.
Received for publication May 11, 2004. Accepted for publication November 3, 2004.
References
von Boehmer, H., H. J. Fehling. 1997. Structure and function of the pre-T cell receptor. Annu. Rev. Immunol. 15:433.
Marrack, P., J. Kappler. 1997. Positive selection of thymocytes bearing - T cell receptors. Curr. Opin. Immunol. 9:250.
Wolfer, A., A. Wilson, M. Nemir, H. R. MacDonald, F. Radtke. 2002. Inactivation of Notch1 impairs VDJ rearrangement and allows pre-TCR-independent survival of early lineage thymocytes. Immunity 16:869.
Washburn, T., E. Schweighoffer, T. Gridley, D. Chang, B. Fowlkes, D. Cado, P. Salmon, E. Robey. 1997. Notch activity influences the vs. T cell lineage decision. Cell 88:833.
Schmitt, T. M., J. C. Zuniga-Pflucker. 2002. Induction of T cell development from hematopoietic progenitor cells by -like-1 in vitro. Immunity 17:749.
Robey, E., D. Chang, A. Itano, D. Cado, H. Alexander, D. Lans, G. Weinmaster, P. Salmon. 1996. An activated form of Notch influences the choice between CD4 and CD8 T cell lineages. Cell 87:483.
Yasutomo, K., C. Doyle, L. Miele, R. N. Germain. 2000. The duration of antigen receptor signaling determines CD4+ versus CD8+ T-cell lineage fate. Nature 404:506
Robey, E. A., B. J. Fowlkes. 2002. A reassessment of the effect of activated Notch1 on CD4 and CD8 T cell development. J. Immunol. 169:1817.
MacDonald, H. R., A. Wilson, F. Radtke. 2001. Notch1 and T-cell development: insights from conditional knockout mice. Trends. Immunol. 22:155.
Anderson, A. C., E. A. Robey, Y. H. Huang. 2001. Notch signaling in lymphocyte(Ana C. Anderson, Elizabet)
Both the Notch and TCR signaling pathways play an important role in T cell development, but the links between these signaling pathways are largely unexplored. The adapter protein Numb is a well-characterized inhibitor of Notch and also contains a phosphotyrosine binding domain, suggesting that Numb could provide a link between these pathways. We explored this possibility by investigating the physical interactions among Notch, Numb, and the TCR signaling apparatus and by examining the consequences of a Numb mutation on T cell development. We found that Notch and Numb cocluster with the TCR at the APC contact during Ag-driven T cell-APC interactions in both immature and mature T cells. Furthermore, Numb coimmunoprecipitates with components of the TCR signaling apparatus. Despite this association, T cell development and T cell activation occur normally in the absence of Numb, perhaps due to the expression of the related protein, Numblike. Together our data suggest that Notch and TCR signals may be integrated at the cell membrane, and that Numb may be an important adapter in this process.
Introduction
The importance of signals through the pre-TCR and TCR in cell fate decisions during T cell development is well established (reviewed in Refs. 1 and 2). More recently, signaling through the Notch receptor has also been implicated in both early (3, 4, 5) and late (6, 7, 8) cell fate decisions during T cell development (reviewed in Refs. 9, 10, 11). Notch signals appear to be critical for T cell development because T cells fail to develop from Notch1-deficient bone marrow precursors (12, 13). In addition, deletion of Notch1 at the DN2/3 stage of T cell development results in a block in T cell development (3). This block is associated with an impairment of V to DJ rearrangement at the TCR locus in DN3 stage thymocytes. Given the importance of Notch in cell fate decisions during T cell development, it is important to address how Notch activity may be regulated during this process.
One of the proteins that has been implicated as a regulator of Notch is Numb (14, 15). Numb is an adapter protein that contains an N-terminal phosphotyrosine binding (PTB)4 domain and a C-terminal proline-rich region containing several putative Src homology 3 binding domains. Numb was first identified as a gene controlling cell fate specification in development of the Drosophila peripheral nervous system. Further studies of Drosophila neural development showed that Numb acts by antagonizing Notch signals. Recent reports suggest that Numb may regulate Notch activity by promoting the down-regulation and/or degradation of Notch receptor (16, 17). In mammals, two homologues of Numb, Numb and Numblike, have been described (18, 19, 20). Because Numb is expressed in the thymus (21), it is a good candidate for regulation of Notch activity during T cell development.
Notch receptors and ligands are also expressed on cells of the peripheral immune system (22, 23, 24) and appear to have a role in T cell responses. Some studies suggest that Notch may have a role in inducing regulatory T cells (22, 25, 26, 27). Other studies suggest a role for Notch in modulating T cell responses (28, 29, 30) and affecting Th1/Th2 differentiation (31, 32). Thus, it appears that Notch also plays a role in peripheral T cell responses, although the exact mechanism and effects of Notch are still unclear.
In this study we provide evidence for the physical association of TCR, Notch, and Numb in both immature and mature T cells. We also explored the role of Numb in T cell development by generating mice in which Numb is deleted specifically in T cells, and we found that there is no appreciable effect of Numb deficiency on T cell development, perhaps due to redundancy with the homologous protein Numblike. Our results provide a basis for future examination of the roles of Notch and Numb in modulating TCR signaling.
Materials and Methods
Mice
Lck-Cre and CD4-Cre transgenic mice (33) were provided by C. Wilson (University of Washington). Numbflox/flox and Numb+/– mice have been previously described (34). AND TCR transgenic (H-2b) mice were purchased from The Jackson Laboratory. AND TCR transgenic RAG–/– (H-2d) mice were purchased from Taconic Farms. All animals were housed in accordance with the guidelines established by the animal care and use committee at University of California, Berkeley.
Colocalization assay
DCEK.ICAM cells (I-EK) (35) were cultured on four-well, poly-L-lysine-coated slides at 105 cells/well in complete RPMI 1640 and incubated overnight at 37°C. Twelve to 16 h before the addition of T cells, DCEK.ICAM cells were either pulsed with 10 μM pigeon cytochrome c (PCC) peptide or left untreated. Thymocytes (106) from AND TCR transgenic (H-2d) mice or 106 lymph node cells (LNC) from AND TCR transgenic (H-2b) mice were then added with or without freshly added peptide and cultured for 20 min. After incubation, slides were spun down at 700 rpm for 5 min. Cells were then fixed with 4% paraformaldehyde for 15 min at 4°C, permeabilized, and stained with Abs against Numb, Notch, and CD3. The Abs used are affinity-purified rabbit antisera generated against the C terminus of murine Numb (aa 489–524) as previously described (19), goat anti-rabbit Alexa-546, goat anti-rabbit Alexa-488, goat anti-hamster Alexa-546 (Molecular Probes), anti-CD3-allophycocyanin, anti-CD4 (BD Biosciences; which we conjugated in-house to Alexa-546 according to the manufacturer’s protocol (Molecular Probes)), and anti-Notch (clone 8G10; see below). T cell/APC conjugates were visualized on a Nikon E800 fluorescence microscope retrofit for three-dimensional capture. Images were deconvolved using SlideBook software (Intelligent Imaging Innovations).
Isolation of mAb reactive with mouse Notch1
A hamster mAb against mouse Notch1 (clone 8G10) was generated by immunizing a Syrian hamster with a bacterial fusion protein consisting of the region from epidermal growth factor repeat 33 through LIN-12, Notch, GLP-1 repeat 1 (nucleotides 3898–4476) of mouse Notch1 fused to GST. Spleen cells were fused with a hamster hybridoma partner, and the clones were screened by ELISA for reactivity with the fusion protein. Positive clones were then recloned and screened for reactivity against full-length mouse Notch1 by immunoblotting. Anti-Notch1 Ab was generated from hybridoma culture supernatant and was either purified on a protein G column or used as a concentrated culture supernatant. The 8G10 Ab reacts with mouse and rat Notch1, but not with other Notch family members (data not shown). The Ab has been licensed to Upstate Biotechnology.
Immunoprecipitation and immunoblotting
For immunoprecipitations, 108 thymocytes from either C57BL/6 or AND-TCR transgenic (H-2d) mice were lysed in lysis buffer (20 mM Tris-HCl (pH 8.0), 138 mM NaCl, 9.5% glycerol, 1 mM sodium orthovanadate, 2 mM EDTA, 10 mM NaF containing either 1% Brij 58 detergent (Pierce), or 1% Nonidet P-40 (Sigma-Aldrich)) supplemented with mini complete EDTA-free protease inhibitor mixture (Roche). Cell lysates were precleared with protein G agarose beads (Sigma-Aldrich) bound with rabbit anti-GST for 2 h at 4°C. Samples were then briefly centrifuged, and the supernatant was added to protein G agarose beads bound with Numb antisera and incubated overnight at 4°C. Samples were separated on a 10% SDS-PAGE gel, transferred onto nitrocellulose (Hybond ECL; Amersham Pharmacia), and immunoblotted. For kinetic analysis of Numb associations after TCR stimulation, DCEK.ICAM cells were plated at 105 cells/ml in complete RPMI 1640 and incubated overnight at 37°C. Four hours before the addition of T cells, DCEK.ICAM cells were either pulsed with 10 μM PCC peptide or left untreated. Thymocytes (1.5 x 108) were then added with or without freshly added peptide and cultured for 5, 15, or 20 min at 37°C. Thymocytes were then washed off and immediately lysed as described above. Cells (106) were removed from the samples for the lysate control. The remaining cells were immunoprecipitated with Numb as described above. For the analysis of Numb deletion, cell lysates were prepared by resuspending 106 cells in lysis buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris (pH 7.4), and 0.02% NaN3) supplemented with mini complete protease inhibitor mixture (Roche). Cell lysates were separated as described above, transferred onto nitrocellulose, and immunoblotted using Numb antisera. Where indicated, CD3+ T cells were purified using T cell enrichment columns (R&D Systems). Abs used for immunoblotting are as follows: affinity-purified rabbit antisera generated against the C terminus of murine Numb (aa 489–524), anti-phosphotyrosine (4G10), anti-Lck (clone 3A2), anti-Vav, anti-GST (Z-5), anti-c-Cbl (A-9; Santa Cruz Biotechnology), and anti-CD4 (L3T4; BD Biosciences). Bound Abs were detected with goat anti-rabbit HRP-conjugated secondary Ab (Southern Biotechnology Associates) and goat anti-mouse IgG-HRP (Jackson ImmunoResearch Laboratories) and visualized using ECL (Amersham Pharmacia).
Flow cytometry
Single-cell suspensions from thymus, lymph node, and spleen were prepared and stained with the following Abs: anti- TCR, anti-CD3 labeled with FITC (BD Biosciences/eBiosciences), anti-CD19, labeled with PE (BD Biosciences/eBiosciences), anti-CD4 labeled with PE-Texas Red (Caltag Laboratories), and anti-CD8 labeled with PE-Cy5 (eBiosciences). All data was collected on a Coulter EPICS-XL cytometer (Beckman Coulter) and analyzed with FlowJo software (TreeStar).
Assays for T lymphocyte function
Activation with anti-CD3. CD3+ T cells were purified from lymph nodes using CD3 enrichment columns (R&D Systems) CD3+ T cells (1 x 104–5 x 104/well) were then cultured in triplicate in flat-bottom 96-well plates that had been coated with anti-CD3 (BD Biosciences). Cells were cultured 48 h and pulsed with 1 μCi/well [3H]thymidine for the last 16 h. Cells were then harvested, and [3H]thymidine incorporation was determined in a rack beta scintillation counter (Beckman Coulter).
LNC proliferation. Mice were immunized s.c. at five sites with 100 μg of OVA (Sigma-Aldrich) emulsified in CFA (Difco). Draining lymph nodes were harvested 10 days later. LNCs (4 x 105/well) were cultured in triplicate in 96-well, round-bottom plates in the presence of various concentrations of Ag for 48 h and pulsed with 1 μCi/well [3H]thymidine for the last 16 h. [3H]thymidine incorporation was determined as described above.
Homeostatic proliferation. CD3+ lymph node T cells were purified from Numb–/– mice and littermate controls (Ly5.2) using T cell enrichment columns (R&D Systems). Cells were then labeled with 2 μM CFSE (Molecular Probes) and transferred i.v. (106/mouse) to B65.1 (Ly5.1) that had been irradiated with 400 rad 24 h earlier. On day 6 post-transfer, spleen cells were harvested and stained with Abs to CD4, CD8, and Ly5.1. Data were collected on a Coulter EPICS XL cytometer (Beckman Coulter) and analyzed as described above.
RT-PCR
Total RNA was extracted from either thymocytes or purified CD3+ LNC using TRIzol (Invitrogen Life Technologies). RNA was treated with DNase I to eliminate contaminating genomic DNA before cDNA synthesis using SuperScript (Invitrogen Life Technologies). First-strand cDNA was then used in PCR. The following primers were used: -actin: forward, 5'-tggaatctgtggcatccatgaaa c-3'; reverse, 5'-taaaacgcagctcagtaacagtccg-3'; and Numblike: forward, 5'-ctgaaaccttcaggacggag-3'; reverse, 5'-cacaggacagacttcacgga-3'. PCR products were visualized by agarose gel electrophoresis.
Results
Physical association of Notch and Numb with the T cell signaling apparatus
Given the likely involvement of the Notch and TCR signaling pathways in both immature and mature T cell responses, we hypothesized that these pathways may be physically linked in T cells. To test this hypothesis, we examined the localization of Notch and CD3 in lymph node T cells from AND TCR transgenic (H-2b) mice after exposure to peptide-pulsed APCs by fluorescence microscopy. We found that Notch is recruited to the T cell/APC interface and coclusters with CD3 after Ag exposure (Fig. 1A). Because Numb has been reported to associate with the intracellular region of Notch (15, 19), we examined whether Numb also coclusters with CD3 after Ag exposure. As shown in Fig. 1B, Numb is also recruited to the T cell/APC interface and coclusters with CD3 after Ag exposure.
FIGURE 1. Coclustering of Notch and Numb with TCR at the T cell/APC interface in response to Ag stimulation. DCEK.ICAM cells were cultured on poly-L-lysine-coated slides overnight at 37°C. Twelve to 16 h before the addition of T cells, DCEK.ICAM were pulsed with 10 μM PCC peptide or were left untreated. LNC or thymocytes (106) were then added and cultured for 20 min. After incubation, slides were spun down, and cells were fixed, permeabilized, and stained with Abs against Numb, Notch, and CD3. T cell/APC conjugates were then visualized by fluorescence microscopy. The rightmost panels show the brightfield image with the fluorescent image superimposed. Colocalization of red and green signals appears as yellow. A and B, Data from LNC from AND TCR transgenic (H-2b) mice. C–E, Data from thymocytes from AND TCR transgenic RAG–/– (H-2d) mice.
To determine whether Notch, Numb, and CD3 also cocluster at the T cell/APC interface in thymocytes in an Ag-dependent manner, we used thymocytes from AND TCR transgenic (H-2d) mice. Because the AND TCR is not selected in mice carrying the H-2d haplotype, thymocytes in these mice are arrested at the double-positive stage of T cell development, thereby providing a source of naive thymic precursors that have not yet received an Ag receptor signal. We exposed thymocytes from these mice to peptide-pulsed APCs and visualized the localization of Notch, Numb, and CD3. Because the levels of Notch and Numb are lower on thymocytes relative to the levels on APCs, coclustering is more difficult to discern than in mature T cells. Nevertheless, we found that Notch and Numb were also recruited to the T cell/APC interface and coclustered with CD3 in thymocytes after Ag exposure (Fig. 1, C and D). To ensure that the coclustering observed was specific, we examined the localization of CD8. Because the AND TCR is an MHC class II-restricted receptor, CD8 would not be part of the TCR complex in these cells. As expected, CD3 redistributed and clustered at the T cell/APC interface in an Ag-dependent manner, whereas CD8 remained evenly distributed over the cell surface (Figs. 1E and 2B).
FIGURE 2. Quantitation of immunofluorescence data. A, T cell/APC conjugates were identified using brightfield imaging. After images were recorded, clustering of the indicated proteins at the T cell/APC interface in the presence or the absence of Ag was scored. Values are expressed as the percentage of total conjugates scored. The left panel shows data from nine (–Ag) and 11 (+Ag) conjugates. The right panel shows data from 13 (–Ag) and 14 (+Ag) conjugates. The slightly increased colocalization of CD3 observed in the experiment shown in the right panel may reflect slight differences in the rate of T/APC conjugate formation. B, The pixel intensity of each indicated protein over the surface of the whole thymocyte and at the thymocyte/APC contact in the presence or the absence of Ag was determined using SlideBook software. Values indicate the percentage of signal at the thymocyte/APC contact relative to the signal over the whole cell. Data shown for CD3 are the average of 18 (+Ag) and 12 (–Ag) conjugates. The data shown for CD8 are the average of nine (+Ag) and four (–Ag) conjugates. The data shown for Numb are the average of nine (+Ag) and eight (–Ag) conjugates.
We used two approaches to quantify the coclustering of Notch and Numb with CD3. First, we identified mature T cell/APC conjugates using brightfield, recorded fluorescent images and then scored them for clustering of CD3, Notch, Numb, CD3 plus Notch, or CD3 plus Numb at the T cell/APC interface in the presence or the absence of Ag. For CD3 and Notch, 30% of conjugates had CD3 plus Notch clustered at the interface. No clustering of CD3, Notch, or CD3 plus Notch was observed in the absence of Ag (Fig. 2A). For the experiments examining the clustering of CD3 and Numb, some clustering of CD3 or Numb alone to the interface was observed in the presence and the absence of Ag. However, coclustering of CD3 plus Numb was only observed in the presence of Ag (Fig. 2A). Second, we chose representative examples of thymocyte/APC conjugates and determined the pixel intensity of the CD3, CD8, and Numb signals at the T cell/APC interface relative to the pixel intensity over the whole cell. The percentage of CD3 and Numb signals that localized at the T cell/APC interface in thymocytes increased in the presence of Ag, whereas the amount of CD8 signal at the interface remained the same in the presence or the absence of Ag (Fig. 2B). Thus, it appears that both Notch and Numb are recruited to the T cell/APC interface and cocluster with the TCR complex upon Ag exposure in immature and mature T cells, suggesting a physical association between the Notch and TCR signaling pathways in these cells.
Given that Numb has the structural features of an adapter protein, we hypothesized that it may provide a link between the TCR and Notch signaling pathways. To investigate this, we examined what proteins interact with Numb in thymocytes. Because Numb contains a PTB domain, we first looked for associated proteins containing phosphorylated tyrosines by immunoprecipitation with Numb antisera, followed by immunoblotting with an anti-phosphotyrosine Ab. We observed several prominent tyrosine-phosphorylated proteins in the Numb immunoprecipitates that were not present in control immunoprecipitations using species-matched antisera of irrelevant specificity (preclear lane, Fig. 3A). Upon reprobing the blot with specific Abs, we found that Lck (Fig. 3B), c-Cbl (Fig. 3C), and Vav (Fig. 3D), coprecipitated with Numb in thymocytes. Similar results were obtained in peripheral T cells (data not shown). The comigration of these bands with prominent phosphotyrosine-containing proteins together with previous reports that these proteins are tyrosine phosphorylated in T cells suggest that the indicated bands in the anti-phosphotyrosine blot correspond to phospho-c Cbl, phospho-Lck, and phospho-Vav (Fig. 3A). We also probed the blot with anti-CD4 and found CD4 in the immunoprecipitate (Fig. 3E) as expected, because Lck is known to associate strongly with CD4 (36). Thus, Numb appears to exist in a preformed complex with TCR signaling proteins, and this complex gets recruited to the T cell/APC interface in an Ag-dependent manner.
FIGURE 3. Numb interacts with components of the T cell signaling complex in thymocytes. B6 thymocytes (108) were lysed, immunoprecipitated with antisera generated against the C terminus of Numb, and analyzed by immunoblotting using Abs against the indicated proteins. The preclear Ab is an irrelevant Ab species-matched to the IP Ab as a control for nonspecific binding. Abs for immunoblotting were specific for anti-phosphotyrosine (A), Lck (B), c-Cbl (C), Vav (D), and CD4 (E).
To provide evidence for the functional relevance of these interactions, we determined whether the interaction of Numb with any of these proteins changes upon TCR stimulation. To do this, AND TCR transgenic (H-2d) thymocytes were exposed to APCs with or without Ag for 5, 15, or 20 min. Cells were then lysed, lysates were immunoprecipitated with Numb antisera, and immunoprecipitates were analyzed by immunoblotting. As shown in Fig. 4, the interaction of Numb with phospho-Lck and phospho-c-Cbl increased at 5 min, followed by a progressive decrease at 15 and 20 min (Fig. 4). The change in association with Numb was not due to a change in the total amounts of Lck and c-Cbl in the cells, because the amounts of these proteins in the lysate lanes did not change over time (Fig. 4). Taken together, our data suggest that TCR, Notch, and Numb are physically linked in T cells.
FIGURE 4. Interaction of Numb with Lck and c-Cbl changes upon TCR stimulation. DCEK.ICAM cells were cultured overnight at 37°C. Four hours before addition of T cells, DCEK.ICAM were pulsed with 10 μM PCC peptide or were left untreated. Thymocytes from AND TCR transgenic RAG–/– (H-2d) mice were then added for 5, 15, or 20 min. Thymocytes were harvested, lysed, and immunoprecipitated with Numb antisera. A, Anti-phosphotyrosine blot and c-Cbl blot. B, Quantification by densitometry of the indicated proteins present in the lysate (left panel) and the Numb immunoprecipitates (right panel). This experiment was repeated three times with similar results. Representative data are shown.
Conditional deletion of the Numb gene in thymocytes
Given our data implicating Numb as a link between the TCR and Notch signaling pathways, we chose to address the role of Numb at different stages of T cell development. Because germline deletion of Numb leads to lethality on embryonic day 11.5 (34, 37), we used Cre/lox technology to target deletion of the Numb gene specifically to developing T cells. To achieve this we crossed Numbflox/floxmice (34) with mice expressing the Cre recombinase under the control of the Lck proximal and CD4 promoters. These promoters have been shown to effect deletion at the DN2/3 and DN3/4 stages of thymic development, respectively (3, 33, 38). We generated mice in which one allele of Numb was already deleted by first crossing Numb+/– (34) mice to Cre transgenic mice and subsequently crossing progeny from this cross to Numbflox/flox mice. To assess the extent of deletion of Numb, we extracted total genomic DNA from unfractionated thymocytes and examined the Numb locus by PCR using primers that distinguish wild-type (WT), floxed, and deleted Numb alleles (34). Fig. 5A shows a representative example of the deletion of Numb. In thymocytes from Lck Cre– Numbfl/+ mice, both the WT and floxed alleles are detected; however, in the Lck-Cre+ Numbfl/– (Numb-deficient) littermate, the deleted allele is the predominant product with a very low level of floxed allele detected. This residual Numb could be due to the presence of DN cells that have not yet deleted Numb, or it could be the result of incomplete deletion of Numb. We next examined Numb protein expression in both unfractionated thymocytes and purified peripheral T cells by immunoblotting (Fig. 5B). We estimate that the deletion of Numb protein in thymocytes and mature T cells from Numb-deficient mice is >95%, with deletion occurring to a similar extent in Numb-deficient mice generated under the Lck-proximal and CD4 promoters. We also observed the presence of a lower m.w. band in lysates from Numb+/– and Numb-deficient mice that was not present in lysates from Numb+/+ or floxed mice (Fig. 5B and data not shown). Sequence analysis of RT-PCR products from these mice indicates that this minor band corresponds to a splice around the targeted exons of Numb (data not shown). This mutant Numb protein is unlikely to be functional given that it lacks a significant portion of the PTB domain and an independently generated Numb-deficient mouse displays a similar embryonic phenotype (34, 37).
FIGURE 5. Deletion of Numb. A, Analysis of genomic DNA by PCR in unfractionated thymocytes from control Lck-Cre–Numbfl/+ and Lck-Cre+Numbfl/– (Numb-deficient) mice. Arrows indicate the presence of WT, floxed, and targeted Numb alleles. B, Analysis of Numb protein in unfractionated thymocytes and CD3+ peripheral LNC from Lck-Cre- Numbfl/+ and Lck-Cre+Numbfl/– (Numb-deficient) mice. Lysates (106 cell equivalents/lane) were separated on a 10% SDS-PAGE gel and blotted with polyclonal rabbit antisera generated against the C terminus of Numb (aa 489–524).
T cell development in the absence of Numb
We examined T cell development in mice in which Numb deletion was driven by CRE expressed under the control of the Lck proximal and CD4 promoters. In Lck-Cre+ Numbfl/ –mice, we observed a modest block between DN3 and DN4 that could be attributed to the presence of the Lck Cre transgene because it was also observed in Lck-Cre+ Numb+/+ controls (data not shown). Similarly, a reduction in the number of peripheral T cells and a skew in the ratio of CD4/CD8 T cells in Lck-Cre+ Numbfl/ –mice could also be attributed to the presence of the Lck Cre transgene alone (Fig. 6, A and B). In addition, we did not observe any defects in thymic development or in the periphery of Numb-deficient mice generated under the CD4 promoter (Fig. 6, C and D, and data not shown). Taken together these data suggest that T cell development can proceed normally in the absence of Numb.
FIGURE 6. Peripheral lymphocyte subsets in Numb-deficient mice. B/T (A and C) and CD4/8 (B and D) ratios in Lck–Cre+Numb-deficient mice (Lck–Cre+Numbfl/–) and controls (Lck–Cre–Numbfl/–, Lck–Cre–Numbfl/+, Lck–Cre+Numb+/+) and in CD4–Cre+Numb-deficient mice (CD4–Cre+Numbfl/–) and controls (CD4+Cre–Numbfl/+). The B/T ratio is based on the number of spleen cells staining positively for CD19 and CD3 as determined by flow cytometry. The CD4/CD8 ratio is based on the number of spleen cells staining positively for CD4 and CD8 as determined by flow cytometry. Points represent values from individual mice, and bars represent mean values. All data are from mice <30 wk of age.
Peripheral T cell function in the absence of Numb
Given our data implicating Numb in TCR signaling, we examined next the ability of Numb-deficient T cells to respond to signals through the TCR. As shown in Fig. 7A, Numb-deficient T cells proliferate equally well in response to plate-bound anti-CD3 as do Numb WT T cells. To examine the ability of Numb-deficient T cells to respond to a more physiologic stimulus, we assessed the in vitro recall response to OVA after immunization with OVA in CFA. As shown in Fig. 7B, there is no significant difference in the ability of Numb-deficient T cells to proliferate in response to Ag stimulation relative to Numb WT littermate controls. We also found no significant differences in the production of IFN- by Numb-deficient T cells relative to littermate controls (data not shown). Finally, we examined the ability of Numb-deficient T cells to respond to homeostatic signals that regulate T cell survival and drive proliferation in response to lymphopenia (reviewed in Refs. 39 and 40). We transferred CFSE-labeled peripheral T cells isolated from Numb WT or Numb-deficient mice to irradiated recipients and analyzed the ability of T cells to undergo homeostatic proliferation by comparing CFSE dilution profiles. We found that Numb-deficient T cells and Numb WT T cells undergo homeostatic division to a similar degree (Fig. 8). We also examined the phosphorylation of TCR-, which is constitutively phosphorylated in normal T cells in response to environmental survival signals (41), and found that its phosphorylation is similar in Numb-deficient T cells compared with T cells from littermate controls (data not shown). Taken together, these data suggest that Numb deficiency does not affect peripheral T cell function and homeostasis.
FIGURE 7. Numb-deficient T cells respond normally to stimulation. A, Proliferative response of Numb-deficient T cells to plate-bound anti-CD3. Purified CD3+ T cells were cultured in triplicate with the indicated amounts of anti-CD3. [3H]thymidine was added at 48 h, and cells were harvested 16 h later. Data are shown as the mean cpm of triplicate wells, where cpm is the mean cpm in test wells – the mean cpm in wells with medium only. An independent experiment gave similar results. B, Recall response to OVA. LNCs were harvested from immunized mice and tested in triplicate over a dose response of 0.1–100 μg/ml OVA. [3H]thymidine was added at 48 h, and cells were harvested 16 h later. Data are shown as the mean cpm of triplicate wells, where cpm is the mean cpm in test wells – the mean cpm in wells with medium only.
FIGURE 8. Homeostatic proliferation of Numb-deficient T cells. CD3+ lymph node T cells (106) from Numb-deficient mice and littermate controls (Ly5.2) were labeled with 2 μM CFSE and transferred i.v. to irradiated B65.1 (Ly5.1) mice. Histograms show the CFSE profile of either CD4+Ly5.1– or CD8+Ly5.1– cells. Each panel represents an individual B65.1 recipient.
Expression of Numblike in T cells
The lack of an appreciable effect of Numb deficiency on thymocyte development and peripheral T cell responses despite evidence for functional associations between Numb and the TCR signaling apparatus raises the possibility of functional redundancy by a related protein. Numblike is highly homologous to Numb (20), and low levels of Numblike have been shown to functionally compensate for Numb deficiency (42). We examined the expression of Numblike in both thymocytes and peripheral T cells and found that Numblike is expressed in both thymocytes and peripheral T cells (Fig. 9 and data not shown), providing a possible explanation for the normal T cell development and functional responses we observe in Numb-deficient mice.
FIGURE 9. Numblike expression in T cells. Numblike message was amplified by RT-PCR from CD3+ lymph node T cells. Amplification of -actin message from the same sample is shown as a control.
Discussion
The importance of the TCR and Notch signaling pathways in T cell development together with evidence for functional interplay between these two pathways prompted us to investigate how these two signaling pathways are linked. In this study we provide evidence that both Notch and Numb are recruited to the T cell/APC interface and cocluster with T cell signaling components in both immature and mature T cells upon exposure to Ag. We also provide evidence from coimmunoprecipitation experiments that Numb interacts with multiple components of the TCR signaling apparatus. Taken together, our data suggest that Notch and Numb may be part of the TCR signalosome in thymocytes and mature T cells. A recent report that Notch interacts with Lck in T cells also fits with this view (43).
A number of recent studies point to a role for Notch in modulating TCR signaling. In some studies, Notch activity was found to enhance TCR-induced proliferation by potentiating IL-2R expression (28) and NF-B activity (29), whereas another study found an inhibitory effect of Notch on TCR signaling (30). Notch has also been implicated in regulating the balance of Th1/Th2 cell development (31, 32). For thymocytes, we have found that Notch synergizes with TCR signals to regulate gene expression in thymocytes (44), whereas others have reported that Notch activity dampens TCR signals (45). Together, these studies point to a complex functional interplay between the TCR and Notch signaling pathways. Our data showing physical associations among TCR, Notch, and Numb provide a framework for further examination of how Notch and TCR signals could be influencing one another.
In addition to its role in regulating Notch, Numb has been implicated in the cellular processes of endocytosis and ubiquitination (16, 17, 46). Two recent reports suggest that these functions of Numb may be related to its ability to regulate Notch. First, Numb was found to associate with the ubiquitin ligase Itch, thereby promoting the ubiquitination and degradation of membrane-associated Notch receptor (16). Second, Numb was found to associate with -adaptin, a protein involved in receptor-mediated endocytosis, and this interaction was found to be important in down-regulating Notch expression (17). Given these considerations and our data suggesting Numb as a link between the TCR and Notch signaling pathways, several nonmutually exclusive models of how these pathways operate in T cells emerge. In the first model, Numb regulates Notch activity in T cells, and Notch acts directly to modulate T cell fate. Although this model fits well with classic studies of Numb and Notch in other systems, it does not take into account our data showing the interaction of Numb with components of the T cell signaling apparatus or the recent studies suggesting regulation of TCR signaling by Notch (discussed above). In a second model, Numb functions as an adapter protein in T cells and regulates T cell signaling as a part of the TCR signalosome. This model fits with the idea of Numb as an adapter protein and our interaction data, but does not incorporate the well-established role of Numb in regulating Notch. In a third model, Numb regulates Notch signaling, and Notch then regulates signals through the TCR. In considering these last two models, Numb could regulate TCR, Notch, or both by targeting them for ubiquitination and degradation (16, 17). Our observation that the ubiquitin ligase c-Cbl is part of the complex containing Numb fits with this idea. Although we did not observe any difference in the levels of cell surface TCR or Notch expression in Numb-deficient T cells (data not shown), it is still possible that Numb targets a subset of activated TCR or Notch receptors for endocytosis and degradation, in line with previous suggestions (16, 17). The extent to which any of these models may operate in T cells will be the subject of future experiments.
Despite these indications for a connection between Notch signaling and TCR signaling, we did not see an appreciable effect of Numb deficiency on peripheral T cell responses. One possible explanation for this is that Numblike compensates for lack of Numb. Numblike is highly homologous to Numb, and low level expression of Numblike has been shown to be functionally redundant with Numb during neural development (42) (Y.-N. Jan and W. Zhong, unpublished observations). Because, Numblike is expressed in thymocytes (data not shown) and peripheral T cells, it is likely that Numblike compensates for the absence of Numb in Numb-deficient T cells. This may explain the lack of a thymic defect in Numb-deficient mice. Another possibility is that the T cells that are seeded to the periphery in Numb-deficient mice represent those that have best adapted to function normally in the absence of Numb. One way to resolve these issues is to examine development and function in T cells deficient for both Numb and Numblike.
In summary, we have shown that the TCR and Notch signaling pathways may be connected in T cells via the adapter protein Numb. Given the indications that Notch signaling affects multiple steps in T cell development and the possibility of complex interplay between the Notch and TCR signaling pathways, it is not surprising that many different effects have been attributed to Notch in T cells. Clarification of the role of Notch in T cell development and peripheral T cell function may come from a better understanding of which effects of Notch are direct consequences of Notch target gene expression vs modulation of TCR signals. Our data provide a framework for future work in deciphering how Notch mediates its effects in both immature and mature T cells.
Acknowledgments
We acknowledge Mimi Mong and Xiao Jun-Kang for production of the Notch monoclonal 8G10, Alice Liao and Rakesh Ahuja for assistance with genotyping mice, Hector Nolla for assistance with cell sorting, and B. J. Fowlkes, Astar Winoto, and members of the Robey laboratory for comments on this manuscript.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Heath Grant AI32985 (to E.R.). A.C.A. is supported by National Institutes of Heath Postdoctoral Fellowship F132AI50415-01.
2 A.C.A. and E.A.K. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Ana C. Anderson, Center for Neurologic Diseases, Brigham and Women’s Hospital and Harvard Medical School, HIM, 77 Avenue Louis Pasteur, Room 784, Boston, MA 02115. E-mail address: aanderson@rics.bwh.harvard.edu
4 Abbreviations used in this paper: PTB, phosphotyrosine binding; DN, double negative; LNC, lymph node cell; PCC, pigeon cytochrome c; WT, wild type.
Received for publication May 11, 2004. Accepted for publication November 3, 2004.
References
von Boehmer, H., H. J. Fehling. 1997. Structure and function of the pre-T cell receptor. Annu. Rev. Immunol. 15:433.
Marrack, P., J. Kappler. 1997. Positive selection of thymocytes bearing - T cell receptors. Curr. Opin. Immunol. 9:250.
Wolfer, A., A. Wilson, M. Nemir, H. R. MacDonald, F. Radtke. 2002. Inactivation of Notch1 impairs VDJ rearrangement and allows pre-TCR-independent survival of early lineage thymocytes. Immunity 16:869.
Washburn, T., E. Schweighoffer, T. Gridley, D. Chang, B. Fowlkes, D. Cado, P. Salmon, E. Robey. 1997. Notch activity influences the vs. T cell lineage decision. Cell 88:833.
Schmitt, T. M., J. C. Zuniga-Pflucker. 2002. Induction of T cell development from hematopoietic progenitor cells by -like-1 in vitro. Immunity 17:749.
Robey, E., D. Chang, A. Itano, D. Cado, H. Alexander, D. Lans, G. Weinmaster, P. Salmon. 1996. An activated form of Notch influences the choice between CD4 and CD8 T cell lineages. Cell 87:483.
Yasutomo, K., C. Doyle, L. Miele, R. N. Germain. 2000. The duration of antigen receptor signaling determines CD4+ versus CD8+ T-cell lineage fate. Nature 404:506
Robey, E. A., B. J. Fowlkes. 2002. A reassessment of the effect of activated Notch1 on CD4 and CD8 T cell development. J. Immunol. 169:1817.
MacDonald, H. R., A. Wilson, F. Radtke. 2001. Notch1 and T-cell development: insights from conditional knockout mice. Trends. Immunol. 22:155.
Anderson, A. C., E. A. Robey, Y. H. Huang. 2001. Notch signaling in lymphocyte(Ana C. Anderson, Elizabet)