T Cell Receptor–initiated Calcium Release Is Uncoupled from Capacitative Calcium Entry in Itk-deficient T Cells

A murine strain 129 genomic clone encoding the Itk kinase domain (exons 9, 10, 11, 12, and 13 [see Fig. 1]) was isolated. A 2.6-kb HindIII fragment extending from the middle of exon 10 to beyond exon 12 was replaced by the 1.8-kb pGK-neo cassette; these deleted sequences encode a portion of the Itk kinase domain (amino acids 476–543). The targeting vector also contained 4.8 kb and 4.2 kb of flanking sequences at the 5′ and 3′ ends, respectively, and a 2.0-kb thymidine kinase cassette (pGK-tk). The construct was linearized and transfected into J1 (gift of Dr. A. Sharpe, Harvard Medical School, Boston, MA) and CCE (gift of Dr. E. Robertson, Harvard University, Cambridge, MA) embryonic stem (ES) cell lines. 360 clones resistant to Geneticin (Sigma Chemical Co., St. Louis, MO) and FIAU [1-(2′-deoxy-2′-fluro-β-d arabinofuranosyl)- 5-iodouracil; gift of Dr. A. McMahon, Harvard University, Cambridge, MA] were screened, yielding two homologous integrants, one from each ES cell line. The ES clones were injected into C57Bl/6 blastocysts to generate chimeric mice. Chimeras (male) were backcrossed to C57Bl/6 females and progeny were screened for the mutated Itk allele. Heterozygous Itk^+/− mice were intercrossed to generate Itk^−/− mice. Identical phenotypes were observed for Itk^−/− mice generated from both of the two original ES cell lines; these lines were used interchangeably. Mice were maintained under specific pathogen-free conditions.

For CD4⁺ T cell purification, lymph node (inguinal, axillary, brachial, mandibular, and mesenteric) cells from multiple wild-type or Itk-deficient mice were isolated and were purified on columns (Biotex Laboratories, Alberta, Canada) according to the manufacturer's protocol. The resulting cells were 90–97% pure (Itk^+/+ or Itk^+/−) or 80–90% pure (Itk^−/−) CD4⁺ T cells.

Before antibody staining 10⁶ lymph node cells were plated in wells precoated with 5 μg/ ml goat anti–hamster IgG(H+L) followed by 5 μg/ml hamster anti-mouse CD3ε (PharMingen, San Diego, CA). After 18–20 h of stimulation, cells were stained with the following antibodies: biotinylated anti-CD3ε (PharMingen, San Diego, CA), biotinylated anti-CD69 (PharMingen), biotinylated anti-CD25 (PharMingen), biotinylated anti-CD62L (PharMingen), FITC-conjugated anti– TCR-α/β (PharMingen), PE-conjugated anti-CD4 (GIBCO BRL, Bethesda, MD), and CD8-Red (GIBCO BRL); secondary reagent was streptavidin-FITC (Southern Biotechnology Associates, Birmingham, AL). Cells were analyzed on a FACScan^® using the CellQuest software (Becton Dickinson, Mountain View, CA). Immunoprecipitations and Western blotting were done with antibodies to the following proteins: Itk (monoclonal antibodies 7F10 and 10B2; reference 6), Tec (gift of Dr. B. Tang, St. Jude Children's Research Hospital, Memphis, TN), TCR-ζ (Santa Cruz Biotech, Santa Cruz, CA), ZAP-70 (Santa Cruz Biotech), and PLC-γ1 (Santa Cruz Biotech and Upstate Biotechnology, Lake Placid, NY). Anti-phosphotyrosine antibody 4G10 was a gift from Dr. B. Druker (Oregon Health Sciences Center, Portland, OR).

For T cell stimulations, 5 × 10⁴ purified CD4⁺ T cells were plated in wells precoated with goat anti–hamster antibody (5 μg/ml) followed by hamster anti– mouse CD3ε (0.5 μg/ml). Where indicated, a 1:8 dilution (determined to be saturating by cell surface staining) of anti-CD28 antibody supernatant (37N.D1; a gift from Dr. J. Allison (University of California at Berkeley, Berkeley, CA; reference 20) was added. As a control, cells were also stimulated with PMA (1 ng/ ml) plus ionomycin (500 ng/ml). After 24 h of stimulation, the supernatants were assessed for IL-2 concentration using HT-2 indicator cells as previously described (21).

5 × 10⁶/ml purified CD4⁺ cells were incubated with 3 μg/ml fluo-3, 5 μg/ml Fura-Red, 0.5% pluronic acid (Molecular Probes, Eugene, OR), and 2% FCS in RPMI for 30–50 min. Dye-loaded cells were washed twice with RPMI and left at room temperature for 30 min in the dark before use. 1.5 × 10⁶ (in 300 μl) dye-loaded cells were placed in 1 ml of serum-free RPMI and warmed to 37°C. Stimulated cells were analyzed on a FACScan^®. The mean fluorescence ratio (FL-1:FL-3) was calculated using the FACSAssistant^® program.

Purified CD4⁺ T cells were incubated in serum-free RPMI for 2 h at 37°C before stimulation. Cells were then washed twice, resuspended in 120 μl of RPMI containing 25 μg/ml biotinylated anti-CD3ε antibody, and incubated on ice for 10 min. 50 μg/ml of streptavidin (Sigma Chemical Co.) was then added, and cells were incubated at 37°C for the indicated times; signaling was terminated by diluting the samples into 1 ml of ice-cold PBS containing 20 mM NaF and 1 mM Na₃VO₄. Cells were then lysed in lysis buffer (0.5% Brij97, 20 mM Tris, pH 8, 100 mM NaCl, 1 mM EDTA, 20 mM NaF, and 1 mM Na₃VO₄) supplemented with protease inhibitors. Immunoprecipitations were done with lysate from 3 × 10⁶ cells. Immunoprecipitated proteins were resolved by 8 or 12% SDS-PAGE, transferred to Immobilon-P membranes (Millipore, Bedford, MA), and blocked and probed as previously described (22).

8 × 10⁶ purified CD4⁺ T cells were starved in serum-free RPMI for 1 h at 37°C before stimulation. Cells were stimulated as for immunoprecipitation studies, but at 3.2 × 10⁷/ml; signaling was terminated by adding 50 μl ice-cold 100% Trichloroacetic acid. IP3 levels were assessed using the IP3 Radioreceptor Assay Kit (NEN Life Science Products, Boston, MA) following the manufacturer's protocol. Controls included medium alone or mock-stimulated (no primary antibody) cells.

To determine the function of Itk in T cell signaling, we generated Itk-deficient mice by gene targeting. Exons of the Itk gene encoding critical regions of the kinase domain were deleted, and replaced with the neo^R cassette (Fig. 1,a). Mice homozygous for the mutated Itk allele (Itk^−/−; see Fig. 1,b) fail to express any Itk protein in lymph nodes, spleen, or thymus (Fig. 1 c). Consistent with a previous report on Itk-deficient mice (19), T cell development is abnormal in the absence of the Itk protein. Specifically, the size and general histology of the thymus are normal, but CD4⁺ T cell maturation is diminished in Itk^−/− mice. The number of CD4⁺ T cells in peripheral lymphoid organs is reduced to ∼50% of that found in wild-type mice; in contrast, the number of CD8⁺ T cells is unchanged (data not shown).

To investigate the role of Itk in TCR signaling, we examined the responses of Itk^−/− CD4⁺ T cells to TCR stimulation. First, Itk^−/− lymph node cells were stimulated with plate-bound anti-CD3 antibody, and cells were assessed 18–20 h later for changes in surface marker expression. CD25 and CD69 were upregulated and CD62L (L-selectin) was downregulated after TCR stimulation on both mutant and wild-type cells (Fig. 2,a). The only difference we consistently observed was a somewhat higher degree of CD25 upregulation on Itk^+/+ cells compared with Itk^−/− cells; this difference is most likely due to the IL-2– induced increase in IL-2 receptor α chain (CD25) expression, which is lacking in Itk^−/− T cells (see below). These data indicate that the TCR-dependent signaling pathways leading to changes in activation marker and adhesion molecule expression are intact in Itk^−/− T cells. Second, we examined IL-2 secretion in response to TCR stimulation. When purified CD4⁺ Itk^−/− T cells were stimulated with anti-CD3 antibody, no IL-2 could be detected in the supernatants after overnight culture (Fig. 2 b). Stimulation with anti-CD3 plus anti-CD28 antibodies, which dramatically enhances IL-2 production from wild-type cells, resulted in detectable, but still reduced, IL-2 secretion from Itk^−/− T cells. This signaling defect is proximal to the TCR, as stimulation with PMA plus ionomycin, which bypasses the early stages of TCR signaling, elicits comparable IL-2 secretion by Itk^−/− and wild-type T cells.

The functional defect in IL-2 production after stimulation of Itk^−/− T cells suggested a defect in the TCR signaling pathway leading to IL-2 transcription. Since IL-2 promoter activity is highly dependent on TCR-induced sustained increases in cytoplasmic free calcium (23–25), we examined calcium responses after TCR stimulation of CD4⁺ T cells. In wild-type cells, TCR cross-linking elicits the expected large, transient peak of increased calcium in the cytoplasm, followed by a sustained elevation in intracellular calcium (Fig. 3,a). In contrast, Itk^−/− T cells have a substantially reduced initial peak of elevated cytoplasmic calcium and do not sustain elevated calcium levels. This response requires high concentrations of the stimulating anti-CD3ε antibody (10 μg/ml); when lower concentrations of antibody are used, wild-type T cells display a slower and more flattened elevation in intracellular calcium, whereas Itk^−/− T cells fail to respond altogether (data not shown). These differences are not due to defective dye loading in the mutant cells, as ionomycin induces equivalent large increases in cytoplasmic calcium in both cell types (Fig. 3 a). Although the defective calcium influx observed with high concentrations of stimulating antibody is novel, the more severe defect observed with lower concentrations of stimulating antibody parallels the defect observed upon surface immunoglobulin cross-linking of Btk^−/− chicken B cells (18). It is not clear if the partial defect observed at high concentrations of stimulating antibody would also be observed in the B cell system.

In normal cells, the overall pattern of calcium elevation results from two components of the calcium response. The initial transient peak is due to the release of calcium from intracellular stores (26); the resulting depletion of these stores triggers the opening of the plasma membrane calcium channels, generating a sustained elevation in intracellular calcium (27). To determine whether Itk^−/− T cells are defective in one or both of these components of the calcium response, we chelated the extracellular calcium with EGTA, eliminating the influx of calcium and thereby revealing the release of calcium from intracellular stores. In the presence of EGTA, the transient calcium increase induced by TCR cross-linking is similar in both wild-type and mutant cells (Fig. 3 b), suggesting that the intracellular calcium stores are released (and therefore depleted) to a similar extent in both cell types. Nonetheless, when extracellular Ca²⁺ is restored, allowing the influx of calcium across the plasma membrane, wild-type cells sustain a substantially higher level of intracellular calcium than do mutant cells. These data indicate that the signaling pathways opening plasma membrane calcium channels in response to store depletion are defective in Itk^−/− T cells, resulting in the reduced level of calcium elevation after TCR stimulation. Since sustained calcium elevation is required for the sustained nuclear translocation of nuclear factor of activated T cells (NFAT) and for IL-2 transcription (23–25), the failure of Itk^−/− cells to sustain intracellular calcium levels is likely to be a direct cause of the reduced IL-2 production observed after TCR stimulation.

To determine whether the depletion of intracellular calcium stores is competent to induce the opening of plasma membrane calcium channels, wild-type and mutant T cells were treated with the drug thapsigargin. Thapsigargin, a specific inhibitor of the Ca²⁺-ATPase in the endoplasmic reticulum membrane, prevents the uptake of cytosolic calcium and results in a leak that depletes intracellular calcium stores and triggers calcium entry (28, 29). In contrast to anti-CD3 antibody cross-linking, thapsigargin treatment initiates a substantial calcium influx in both wild-type and mutant T cells, demonstrating that store-operated calcium entry is normal in Itk^−/− T cells (Fig. 3 c). It is possible that distinct depletion-sensitive signaling pathways are induced depending on the mechanism of store depletion; however, we favor a model in which the channel-operating calcium stores are depleted to a lesser extent in the Itk^−/− cells. This latter hypothesis is consistent with the reduced production of IP3 observed after TCR stimulation of Itk^−/− T cells (see below).

In the presence of extracellular EGTA, thapsigargin induces similar levels of calcium elevation in mutant and wild-type T cell (Fig. 3 d), suggesting that the sizes of the thapsigargin-sensitive intracellular calcium stores are similar in both cell types. Notably, the thapsigargin-stimulated calcium release observed in these naive T cells is unexpectedly small, yet capable of stimulating calcium influx. In light of this result, it is surprising that the large calcium release observed upon TCR stimulation in Itk-deficient T cells fails to trigger calcium influx. This apparent contradiction could be explained in at least three ways. First, the extent of calcium elevation in the presence of thapsigargin may be a poor measure of store depletion, as rates of store refilling may differ between wild-type and Itk^−/− T cells. Second the similar extent of store release observed between wild-type and Itk^−/− cells may underestimate the actual difference in total calcium released if the rate at which calcium is cleared from the cyotosol increases as the concentration of cytosolic calcium increases; this phenomenon has been documented previously in T cells (30). Third, it is possible that the thapsigargin-sensitive intracellular stores are a distinct and small subset of the TCR-coupled intracellular calcium stores, and are uniquely competent to trigger calcium influx in response to depletion. In this latter case, Itk would be required to couple the TCR to these specific channel-operating calcium stores.

It should be noted that the amplitude of the thapsigargin-induced calcium elevation is always slightly diminished in Itk-deficient T cells (Fig. 3 c). When purified cell populations are gated to exclude contaminating CD4⁻ cells, the same difference is observed (data not shown); therefore, this difference is not due to disparities in the purities of the CD4⁺ T cell preparations from mutant and wild-type lymph nodes. The absence of Itk could affect peak intracellular calcium levels by multiple mechanisms, potentially reducing the frequency of calcium channel opening, decreasing the electrochemical gradient driving calcium entry, or by facilitating the upregulation of the plasma membrane calcium pump. At present we do not know which, if any, of these mechanisms result in the reduction in maximum intracellular calcium levels induced by thapsigargin.

TCR-induced calcium responses in T cells are initiated by the second messenger, IP3. IP3 is generated by the PLC-mediated hydrolysis of the membrane phospholipid, phosphatidylinositol 4,5-bisphosphate (PIP2). After PIP2 cleavage, IP3 diffuses through the cytoplasm and binds to IP3 receptors, inducing the release of calcium from internal stores (for review see reference 31). In addition, IP3 is thought to play a role in the subsequent calcium signal that causes the opening of the plasma membrane calcium channels (32). Therefore, we assessed the levels of IP3 generated after TCR stimulation of wild-type and Itk^−/− T cells (Fig. 4,a). In wild-type cells, IP3 levels begin to rise by 45 s after TCR stimulation, remain at peak levels between 1 and 2 min, and return to basal levels by 5 min (Fig. 4,a and data not shown). In contrast, no significant increase in IP3 levels over background can be detected in Itk^−/− T cells at any time point (Fig. 4 and data not shown). Parekh et al. have demonstrated that calcium release can be triggered by IP3 concentrations 50-fold below those required to trigger calcium influx (32). The minimal calcium-releasing doses of IP3 identified by Parekh are at the limit of detection of our assay system, and it is likely that TCR stimulation in Itk^−/− T cells actually induces small amounts of IP3 that activate calcium release without triggering calcium influx. It is unclear why this threshold effect exists. As proposed by Parekh et al., it could result either from a low affinity IP3-receptor present on a specialized channel-coupled calcium store, or from the need for IP3 concentrations sufficient to allow diffusion to outpace the catabolism of IP3 (32).

The PLC isoform PLC-γ1 is abundant in T cells and tyrosine phosphorylated after TCR stimulation. Since this PLC isoform is activated by tyrosine phosphorylation (33), we examined whether reduced tyrosine phosphorylation of PLC-γ1 could account for the defective TCR-induced IP3 production in Itk^−/− T cells. As shown in Fig. 4,b, immunoblotting of PLC-γ1 with an antiphosphotyrosine antibody indicated that the tyrosine phosphorylation of PLC-γ1 from stimulated Itk^−/− T cells is markedly reduced at 0.5 and 1 min. This reduced phosphorylation is not due to decreased levels of PLC-γ1, as demonstrated by anti-PLC-γ1 immunoblotting (Fig. 4 b). In contrast, the TCR-induced phosphorylation of the receptor-proximal signaling molecules ZAP-70 and TCR-ζ is similar in both wild-type and Itk^−/− T cells. Furthermore, the overall pattern of tyrosine phosphorylation induced by TCR stimulation, with minor exceptions, is not markedly different between Itk^+/+ and Itk^−/− T cells (data not shown); in addition, some downstream functions of TCR signaling are intact (e.g., changes in activation marker expression, see above). Since ZAP-70 activation is critical for TCR signaling (34, 35), our data suggest that ZAP-70 activation is normal in the absence of Itk. These data indicate that Itk is not upstream of either TCR-ζ or ZAP-70 phosphorylation, but is required for optimal PLC-γ1 activation after TCR cross-linking. This placement of Itk relative to ZAP-70 is consistent with that proposed by Kurosaki et al. for Btk and Syk (17).

Here we demonstrate that Itk^−/− T cells have a defect in TCR signaling that leads to reduced IL-2 secretion after TCR cross-linking. Although we have not ruled out additional signaling defects in Itk^−/− T cells, the impaired influx of calcium across the plasma membrane is likely to be responsible for the loss of TCR-induced IL-2 production. Since calcium influx enables the sustained increases in cytoplasmic calcium required for transcription factor activation (23–25), this defect could explain the inability of Itk^−/− T cells to secrete IL-2. As calcium release from intracellular stores appears normal in the Itk^−/− T cells, these data suggest that Itk might be involved in the signaling pathways coupling store depletion to calcium influx. However, we found this explanation unlikely because thapsigargin induces calcium influx in the absence of Itk. This finding suggests that a functional depletion-sensing pathway can be decoupled from calcium release. Defective coupling could result from impaired PLC-γ1 activation in Itk^−/− T cells by at least two mechanisms. The dramatically reduced production of IP3 after TCR stimulation in Itk^−/− T cells could fail to trigger a low affinity IP3 receptor on a channel-operating subset of intracellular calcium stores, while fully activating calcium release from the remaining stores. Alternatively, the low levels of IP3 produced in the Itk^−/− T cells could fail to deplete intracellular calcium stores to the extent required to trigger calcium influx. In this case, we would be forced to conclude that the observed store release is a poor measure of actual store depletion, either because we are underestimating release in the wild-type T cells, or because the rates at which the stores refill differ between the wild-type and Itk^−/− T cells. We believe that this is the first demonstration that the TCR-induced influx of extracellular calcium can be uncoupled from the TCR-induced release of intracellular calcium, while preserving functional, depletion-sensitive calcium influx mechanisms. Thus, these experiments suggest that Itk^−/− T cells will provide an interesting model system for detailed studies of the precise mechanisms regulating intracellular calcium levels.

We thank Dr. Arlene Sharpe and Dr. Elizabeth Robertson for embryonic stem cell lines and technical advice; Jennifer Noble for generating the chimeric mice; Dr. Brian Druker for providing 4G10 antibody; Dr. B. Tang for providing the anti-Tec antibody; Dr. Andrew McMahon for providing FIAU; and Dr. Daniel Thomis for helpful discussions and critical reading of the manuscript.

This work was supported by grants from the National Institutes of Health (AI-37584) and the Council for Tobacco Research-USA, Inc.

ES

embryonic stem

PI3

inositol 1,4,5 tris-phosphate

PLC

phospholipase C