Discovery of a novel, potent, and Src family-selective tyrosine kinase inhibitor. Study of Lck- and FynT-dependent T cell activation.

Here, we have studied the activity of a novel protein-tyrosine kinase inhibitor that is selective for the Src family of tyrosine kinases. We have focused our study on the effects of this compound on T cell receptor-induced T cell activation, a process dependent on the activity of the Src kinases Lck and FynT. This compound is a nanomolar inhibitor of Lck and FynT, inhibits anti-CD3-induced protein-tyrosine kinase activity in T cells, demonstrates selectivity for Lck and FynT over ZAP-70, and preferentially inhibits T cell receptor-dependent anti-CD3-induced T cell proliferation over non-T cell receptor-dependent phorbol 12-myristate 13-acetate/interleukin-2 (IL-2)-induced T cell proliferation. Interestingly, this compound selectively inhibits the induction of the IL-2 gene, but not the granulocyte-macrophage colony-stimulating factor or IL-2 receptor genes. This compound offers a useful new tool for examining the role of the Lck and FynT tyrosine kinases versus ZAP-70 in T cell activation as well as the role of other Src family kinases in receptor function.

Study of the roles of the Src family kinases p56 lck and p59 fynT in T cell receptor (TcR) 1 function has been hampered by a lack of specific pharmacological inhibitors. Here, we describe and utilize a novel, potent, and Src family-selective small molecule inhibitor to further study the role of these kinases in T cell activation.
The lck gene, which encodes a lymphocyte-specific, membrane-associated protein-tyrosine kinase of the nonreceptor type (1,2), was first identified in retrovirally induced murine T cell leukemias (2)(3)(4). Lck contains a unique N-terminal sequence (5) that directs its specific interaction with the cytoplasmic domains of the CD4 and CD8 glycoproteins (6 -10). This interaction is required for antigen-specific responses of several different T cell hybridomas (11,12). Lck has also been reported to associate with the IL-2R ␤-chain via a distinct interaction (13), although its role in IL-2R function remains unclear. Several reports have demonstrated that Lck plays an important role in both T cell maturation and activation. Loss of Lck expression in the human Jurkat T cell line abrogates its response to anti-TcR antibodies (14), and inactivation of the lck gene or overexpression of a dominant negative lck transgene in mice leads to an arrest of thymocyte development at a stage prior to the expression of CD4, CD8. and TcR (15,16).
Whereas the role of Lck in T cell development and function is well established, the role of Fyn is less well defined. Small amounts of Fyn are found associated with the TcR complex following mild detergent extraction of T lymphocytes (17). Furthermore, activation of T cells through the TcR results in increased enzymatic activity of Fyn (18). Additional evidence that Fyn is involved in lymphocyte activation comes from experiments in transgenic animals. Twenty-fold overexpression of FynT in transgenic thymocytes results in enhanced responsiveness to anti-CD3 antibody as measured by the stimulation of tyrosine phosphorylation in whole cells, Ca 2ϩ accumulation, and proliferation (19). Furthermore, overexpression of a catalytically inactive form of FynT in the thymocytes of transgenic mice substantially inhibited TcR-mediated T cell activation (19). Finally, gene-knockout mice that lack either p59 fynT or p59 fynB demonstrate defects in TcR/CD3-or alloantigen-mediated signaling (20,21), suggesting that Fyn plays a role in TcR-mediated signaling in vivo.
Here, we show that unlike previously described protein-tyrosine kinase inhibitors, PP1 inhibits Lck and FynT in vitro at concentrations significantly lower than those required to inhibit ZAP-70, JAK2, the EGF-R kinase, and protein kinase A. It inhibits whole cell tyrosine phosphorylation and proliferation in T cells stimulated with anti-CD3 and mitogens. Finally, it selectively inhibits IL-2 gene expression over GM-CSF and IL-2R gene induction in human T cells. Thus, this compound appears to dissect a component of TcR signaling not distinguished by other inhibitors of TcR signaling such as FK506. This compound is a powerful new tool to study the role of Src family protein-tyrosine kinases in lymphocyte function.

EXPERIMENTAL PROCEDURES
Cells and Culture Conditions-Human peripheral blood lymphocytes (PBL) were obtained from healthy donors by gradient separation of whole blood using Histopaque (22) and were cultured in RPMI 1640 medium containing 10% fetal calf serum. Anti-CD3-induced proliferation experiments were carried out using 1 ϫ 10 5 PBL cells/well in 96-well assay plates (Costar Corp.) that were precoated with 20 g/ml rabbit anti-mouse antiserum (Jackson Laboratories, Bar Harbor, MA) in PBS for 4 h. PBL were treated with varying concentrations of compound and either anti-CD3 (50 ng/ml; Becton Dickinson) or PMA (10 ng/ml; Sigma) plus PHA (2 g/ml; Wellcome). Proliferation was assessed by the addition of 1 Ci/well [ 3 H]thymidine (DuPont NEN) at 48 h, followed by harvesting the cells at 72 h using a Skatron harvester. Results are reported as an average IC 50 , determined from a plot of the percent inhibition of proliferation from media control, derived from separate experiments run in triplicate (see Table II). Influenza-induced T cell proliferation (see Table II, Ag-Sp (which represents specific an-* 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. ‡ To whom correspondence should be addressed: Pfizer Central Research, Eastern Point Rd., Groton, CT 06340. Tel.: 203-441-5334; Fax: 203-441-4111. 1 The abbreviations used are: TcR, T cell receptor; IL-2, interleukin-2; IL-2R, IL-2 receptor; EGF-R, epidermal growth factor receptor; GM-CSF, granulocyte-macrophage colony-stimulating factor; PBL, peripheral blood lymphocyte(s); PBS, phosphate-buffered saline; PMA, phorbol 12-myristate 13-acetate; PHA, phytohemagglutinin; PAGE, polyacrylamide gel electrophoresis. tigen)) was assessed by combining 1 ϫ 10 5 PBL in triplicate with 160 l of RPMI 1640 medium containing 10% fetal calf serum with 20 l of either diluted test compound or media alone in a 96-well microtiter plate (Costar Corp.). Antigen (influenza virus vaccine Fluzone, Connaught Laboratories) was prepared by centrifuging and washing (three times) 2 ml of vaccine through a Centricon-3 concentrator (Amicon, Inc.) to remove preservative and diluting the remaining material to 40 ml (1:20). Twenty microliters of antigen was then added to each well, and the plates were incubated for 72 h at 37°C in 5% CO 2 . [ 3 H]Thymidine was then added (0.5 Ci/well), and the plates were incubated for an additional 18 h at 37°C. Cells were harvested with a 96-well harvester (Tomtec), and the amount of incorporated [ 3 H]thymidine was determined using a Pharmacia Biotech ␤-plate counter. Concentrations that cause 50% inhibition of proliferation (IC 50 ) were determined from a plot of the percent inhibition of proliferation from media control versus concentration of test compound added. Results are presented as the mean IC 50 from repeated experiments (see Table II). T cell proliferation in the one-way mixed lymphocyte reaction was assessed by combining in triplicate 5 ϫ 10 4 fresh PBL, 5 ϫ 10 4 irradiated (5000 rads) pooled stimulator PBL, and diluted test compound in RPMI 1640 medium in each well of a 96-well assay plate. After 18 h of incubation at 37°C in 5% CO 2 , 0.5 Ci of [ 3 H]thymidine was added to each well, and the cells were incubated for another 18 h. The cells were then harvested using a Pharmacia Biotech ␤-plate system. Percent inhibition was determined by the following equation: % inhibition ϭ 1 Ϫ (mean cpm of drugtreated cells/mean cpm of control stimulated cells) ϫ 100. IC 50 values represent the concentration of drug that caused 50% inhibition of the control response. Results are presented as the mean IC 50 from repeated triplicate experiments (see Table II). Purified human T cells, used in the whole cell phosphotyrosine analyses, were isolated by the T-Kwik method (One Lambda, Canoga Park, CA). T cell purity, as determined by flow cytometric analysis, was generally Ͼ90% CD3 ϩ T cells, with the major contaminant being CD16 ϩ cells (natural killer cells). Jurkat T cells (American Type Culture Collection) were maintained in RPMI 1640 medium, 10% fetal calf serum. All cells were maintained at 37°C in a humidified 5% CO 2 atmosphere.
Immune Complex Enzyme Assays-The enolase substrate (Sigma) used for measuring Lck and FynT catalytic activity (see Fig. 2 and Table I) was prepared as described (23). The acid-treated enolase was diluted 1:20 with 1 ϫ PBS before aliquoting 100 l/well into a Nunc 96-well high protein binding assay plate. Assay wells were then aspirated; blocked with 0.5% bovine serum, 1 ϫ PBS for 1 h at 37°C; and then washed five times with 300 l of 1 ϫ PBS/well. The source of Lck was either LSTRA cells or Lck expressed in HeLa cells using a vaccinia expression system (8,24). FynT (a gift of R. Perlmutter, University of Washington, Seattle) was expressed in HeLa cells using the vaccinia system (24,25). Cells (12.5 ϫ 10 6 /ml) were lysed in lysis buffer (20 mM Tris, pH 8.0, 150 mM NaCl, 0.5% Nonidet P-40, and 23 trypsin inhibitory units/ml aprotinin), and the lysates were clarified by centrifugation at 14,000 cpm for 15 min at 4°C in an Eppendorf tube. The Lck antibody was produced by immunizing rabbits with a synthetic peptide containing residues 41-54 of the N-terminal domain of Lck. The anti-Fyn antibody was obtained from Upstate Biotechnology, Inc. The clarified lysates were then incubated with the appropriate anti-kinase antibody at 10 g/ml for 2 h at 4°C. Protein A-Sepharose beads (Pharmacia Biotech Inc.; prepared as a 50% (w/v) suspension) were added to the antibody/lysate mixture at 250 l/ml and allowed to incubate for 30 min at 4°C. The beads were then washed twice in 1 ml of lysis buffer and twice in 1 ml of kinase buffer (25 mM HEPES, 3 mM MnCl 2 , 5 mM MgCl 2 , and 100 M sodium orthovanadate) and resuspended to 50% (w/v) in kinase buffer. Twenty-five microliters of the bead suspension was added to each well of the enolase-coated 96-well high protein binding plate together with an appropriate concentration of compound and [␥-32 P]ATP (DuPont NEN; 25 l/well of a 200 Ci/ml solution in kinase buffer). After incubation for 20 min at 20°C, 60 l of boiling 2 ϫ solubilization buffer (26) containing 10 mM ATP was added to the assay wells to terminate the reactions. Thirty microliters of the samples was removed from the wells, boiled for 5 min, and run on a 7.5% SDSpolyacrylamide gel. The gels were subsequently dried and exposed to Kodak X-AR film (see Fig. 2A). For quantitation, films were scanned using a Molecular Dynamics laser scanner, and the optical density of the major substrate band, enolase p46, was determined. Concentrations of compound that caused 50% inhibition of enolase phosphorylation (IC 50 ) were determined from a plot of the density versus concentration of compound (see Fig. 2B). In companion experiments for measuring the activity of compounds against Lck (see Fig. 2C), the assay plate was washed with two wash cycles on a Skatron harvester using 50 mM EDTA, 1 mM ATP. Scintillation fluid (100 l) was then added to the wells, and 32 P incorporation was measured using a Pharmacia Biotech micro-␤-counter. Concentrations of compound that caused 50% inhibition of enzyme activity (IC 50 ) were determined from a plot of the percent inhibition of enzyme activity versus concentration of compound. Since there was good correlation between the gel and plate assays, subsequent repeat experiments for both Lck and FynT were performed using scintillation counting (see Table I). EGF-R activity was measured by immunoprecipitation of EGF-R from A-431 cells obtained from the American Type Culture Collection. Cell lysates were prepared by adding 4 ml of lysis buffer to a T-75 flask that contained a confluent layer of cells. The lysates were clarified by centrifugation as described above and then incubated with 10 g/ml anti-EGF-R (Upstate Biotechnology, Inc.) for 2 h at 4°C. Protein A-Sepharose beads were added to the antibody/lysate mixture at 250 l/ml and allowed to incubate for 30 min at 4°C. The beads were then washed twice in 1.0 ml of lysis buffer and twice in 1.0 ml of kinase buffer (as described above) and finally resuspended to 50% (w/v) in kinase buffer. To each 1.5-ml assay tube was aliquoted 50 l of bead suspension, which was then spun for 15 s at 14,000 rpm in an Eppendorf microcentrifuge, and the supernatant was discarded. To the bead pellet was then added 5 l of the appropriate compound dilution, 5 l of EGF (Upstate Biotechnology, Inc.) to a final concentration of 100 pM, and 5 l of a 33 Ci [␥-32 P]ATP/ml solution in kinase buffer. After incubation for 20 min at 20°C, the beads were washed once with 1.0 ml of lysis buffer and once with 1.0 ml of 1 ϫ PBS. To the bead pellet was added 60 l of boiling 2 ϫ solubilization buffer (26) containing 10 mM ATP. Samples were run on 7.5% SDS-polyacrylamide gels, which were subsequently dried and exposed using BAS-III imaging plates. Labeled EGF-R protein bands were visualized, and 32 P incorporation was quantitated using the BAS-2000 BioImaging analyzer (Fuji Medical Systems USA, Stamford, CT). Concentrations of compound that caused 50% inhibition of enzyme activity (IC 50 ) were determined from a plot of the percent inhibition of enzyme activity by different concentrations of compound. Murine JAK2 was produced in baculovirus and supplied as an immune complex bound to protein A-Sepharose beads (Upstate Biotechnology, Inc.). JAK2 beads (2.5 l) were resuspended in 20 l of kinase buffer (10 mM HEPES, pH 7.4, 50 mM NaCl, 5 mM MgCl 2 , 5 mM MnCl 2 , 0.1 mM Na 3 VO 4 , 0.25 mCi/ml [␥-32 P]ATP) for 10 min at room temperature. The beads were then washed, and JAK2 autophosphorylation was measured by eluting the labeled proteins into SDS-PAGE buffer and was analyzed on a 7.5% polyacrylamide gel. Bands corresponding to JAK2 were quantitated using the Fuji BAS-1000 phosphoimager. IC 50 values were determined as described above. Full-length ZAP-70 kinase (27)(28)(29) was produced using baculovirus expression (30). Lysates from Sf9 cells infected 48 h previously with a human ZAP-70 recombinant virus were prepared as described above for Lck, and a 1:100 dilution was used in a soluble kinase assay. Briefly, kinase activity was quantitated by measuring the incorporation of ␥-32 P into the substrate p62 (31), using SDS-PAGE to resolve phosphorylated p62 and a phosphoimager to quantitate radioactivity. ZAP-70-specific activity was assessed by subtracting p62 phosphorylation obtained using Sf9 cell lysates infected with nonrecombinant baculovirus. IC 50 values were determined as described above.
Whole Cell Phosphotyrosine Measurements-Inhibition of anti-CD3stimulated tyrosine phosphorylation in purified human peripheral blood T cells was measured as follows. All incubations were carried out at 37°C in an Eppendorf Thermomixer 5436 at a mixing setting of 11. Cells (1 ϫ 10 6 in 100 l of RPMI 1640 medium) were incubated for 15 min with drug prior to a 6-min incubation with 1 g of anti-CD3/ml (anti-leu4, 100 g/ml; Becton Dickinson). The final volume of the reaction was 115 l. Reactions were terminated by the addition of 57.5 l of 3 ϫ solubilization buffer (26) incubated at 100°C prior to its addition. Samples were mixed, boiled for 5 min, and stored at Ϫ70°C. Western blots of these cell lysates, run on 10% SDS-polyacrylamide gels, were probed with a polyclonal anti-phosphotyrosine antibody, and immune complexes were detected with 125 I-labeled protein A (ICN) (26). For quantitation, films were scanned using a Molecular Dynamics laser scanner, and the optical densities of the major substrate band, p70, were quantitated in the presence of anti-CD3 (in the presence and absence of drug). Percent inhibition was calculated as follows: (1 Ϫ (p70 optical density units in presence of drug/p70 units in absence of drug)) ϫ 100. IC 50 equals the concentration of compound at which 50% inhibition was measured.
cDNA Probes-Plasmid DNAs were prepared as described (32). cDNA inserts were isolated by digestion with the appropriate restriction enzymes, preparative agarose gel electrophoresis, electroelution, and passage over a G-50 spin column (33). The cDNA probes were labeled using a DNA labeling kit (Pharmacia Biotech Inc.). The human cDNAs for IL-2, GM-CSF, IL-2R␣, and glyceraldehyde-3-phosphate de-hydrogenase were obtained from the American Type Culture Collection.
Northern Blots-Total RNA was isolated from 1 ϫ 10 7 fresh isolated human PBL using acid guanidinium/phenol/chloroform (34). Ten to fifteen micrograms of total RNA/lane was electrophoresed on a 6% formaldehyde, 1.5% agarose gel in buffer containing 5 mM NaHPO 4 , pH 4.45, 15 mM Na 2 HPO 4 , pH 8.85, and 1.5% formaldehyde. The RNA was capillary-blotted to GeneScreen (DuPont NEN) and UV-cross-linked using a UV Stratalinker 8000. The immobilized RNA was hybridized to probe DNA (4 ϫ 105 dpm/ml) in 50% deionized formamide, 1% SDS, 1 M NaCl, and 10% dextran sulfate. The blots were incubated, with agitation, overnight at 42°C. The blots were subsequently washed with constant agitation, twice in 2 ϫ SSC (1 ϫ SSC: 0.15 M NaCl, 0.015 M sodium citrate) for 5 min at room temperature, twice for 30 min in 2 ϫ SSC containing 1% SDS at 60°C, and then twice for 30 min each in 0.1 ϫ SSC at room temperature prior to autoradiography. For each probe, a single blot was stripped and hybridized to cDNA probes specific for each mRNA.
Promoter-Reporter Plasmid Transfections-The promoter-luciferase reporter plasmids were constructed as described previously (32). Briefly, the promoters from the human IL-2 (positions Ϫ327 to ϩ51) and IL-2R␣ (positions Ϫ479 to ϩ109) genes were cloned into pUC13 upstream of the firefly luciferase gene and SV40 polyadenylation signal. The human Jurkat T cell line was transfected as described previously, and luminescence was measured using a Dynatech ML1000 Luminometer. All experiments were run in triplicate, and the data are presented as mean light units. Standard deviations greater than 0.1 are shown.

Selective Inhibition of Src Family Kinases in Vitro-
The pyrazolopyrimidine PP1 (Fig. 1) was synthesized as one of a series of compounds used for inhibition of p56 lck and p60 c-src , based on a parent compound first discovered in tyrosine kinase inhibitor screens. When PP1 was first examined in vitro for its ability to inhibit tyrosine phosphorylation of enolase by p56 lck , it was found to be a potent inhibitor of this enzyme (Fig. 2). p56 lck isolated from LSTRA was incubated with enolase and various concentrations of PP1 in a 96-well plate. Solubilization buffer was then added, and the reactions were run out on an SDS-polyacrylamide gel. As shown in Fig. 2A, PP1 inhibited phosphorylation of the 46-kDa enolase protein in a dose-dependent manner. The 46-kDa enolase band was quantitated using a Molecular Dynamics laser scanner, and the IC 50 for inhibition of enolase was ϳ5 nM (Fig. 2B). Duplicate reactions were run in triplicate in the 96-well plate in the presence of various concentrations of PP1 and quantitated using a Pharmacia Biotech micro-␤-plate reader (Fig. 2C). The IC 50 for inhibition of Lck in these duplicate reactions using the plate reader was 20 nM. A repeat experiment comparing the gel assay with the plate reader yielded IC 50 determinations of 8 and 4 nM, respectively (data not shown). Similar results were obtained using a peptide substrate containing the ITAM sequence derived from the human TcR -chain (data not shown). Since the gel and plate assays provided similar results and the plate reader was more facile for quantitating enolase phosphorylation, subsequent repeat experiments were performed using the plate reader to obtain average IC 50 values for inhibition of Lck and FynT. In three separate subsequent experiments, PP1 inhibited p56 lck activity with an average IC 50 of 5 nM (Table I). Similar results were obtained with p56 lck expressed in HeLa

FIG. 2. Inhibition of Lck catalytic activity in vitro.
A, Lck phosphorylation of the enolase substrate was analyzed in the presence of increasing concentrations of PP1 using SDS-PAGE as described under "Experimental Procedures." Protein standards (in kilodaltons) are shown on the right. B, phosphorylation of the enolase substrate in A was quantitated using a Molecular Dynamics laser scanner, and the results are plotted for determination of an IC 50 value as described under "Experimental Procedures." C, duplicate reactions run in parallel with A were quantitated using scintillation counting as described under "Experimental Procedures," and the results are plotted to determine an IC 50 value. cells using a vaccinia virus expression system (data not shown) (8,24). The compound was also a potent inhibitor of a second lymphocyte Src family kinase, p59 fynT (IC 50 ϭ 6 nM), expressed using the vaccinia system (Table I). A closely related pyrazolopyrimidine, PP2 (Fig. 1), was similarly effective in the inhibition of Lck and FynT (Table I). In further selectivity tests using other Src family protein-tyrosine kinases, PP1 also inhibited Src (170 nM) and Hck (20 nM), while PP2 demonstrated potent inhibition of Hck (5 nM) (data not shown). In contrast, PP1 and PP2 were both 50 -100-fold less active in the inhibition of A-431 epidermal growth factor receptor autophosphorylation (IC 50 ϭ 0.25 and 0.48 M, respectively). Further specificity for inhibition of Lck and FynT was demonstrated when it was found that PP1 and PP2 were essentially inactive for inhibition of ZAP-70 and JAK2 (Table I) and protein kinase A (data not shown). Since the activity of the ZAP-70 enzyme may be enhanced following phosphorylation at residue 493 by Lck (35), we also examined whether inhibition of ZAP-70 by PP1 was altered when coexpressed in insect cells together with Lck as described previously (35). Although coexpression of ZAP-70 with the catalytic domain of Lck consistently led to a 3-4-fold elevation in the specific activity of ZAP-70 for the p62 substrate, PP1 was still unable to inhibit this enzyme up to concentrations of 100 M (data not shown). For comparison purposes, we also examined the activity of staurosporine and genistein, two previously described tyrosine kinase inhibitors. The fermentation product, staurosporine, has previously been demonstrated to be a potent but nonselective protein kinase inhibitor (36,37). In the experiments reported here (Table I), staurosporine was found to be a nanomolar inhibitor of p56 lck and p59 fynT as well as a low micromolar inhibitor of the EGF-R kinase. However, unlike PP1 and PP2, it was also a potent inhibitor of the ZAP-70 and JAK2 tyrosine kinases. For further comparison, the naturally occurring isoflavone genistein (36) was tested for its ability to inhibit the four tyrosine kinases. As expected, it was the least potent inhibitor (Table I). Thus, relative to other reported tyrosine kinase inhibitors, the novel compounds PP1 and PP2 demonstrated potent and selective inhibition of the Src family kinases, such as p56 lck and p59 fynT .
Repression of Early Signaling Events in Human T Cells-One of the earliest events in TcR/CD3 triggering of T cells is the stimulation of the tyrosine phosphorylation of multiple substrates (38). Previous studies have demonstrated that tyrosine kinase inhibitors are capable of inhibiting TcR-induced T cell activation (39,40). Therefore, PP1 was tested for its ability to inhibit anti-CD3-stimulated tyrosine phosphorylation in purified human T cells (Fig. 3) 1 and 2). These substrates included proteins of approximately 100, 70, 50 -60, and 35 kDa. The phosphorylation of all substrates was significantly reduced by 1, 10, and 100 M PP1 (Fig. 3, lanes 3-5, respectively), with 50% inhibition occurring at ϳ0.5 M PP1 as measured by quantitating the p70 band from two separate experiments (data not shown). PP2 was equipotent with PP1 for inhibition of anti-CD3-induced tyrosine phosphorylation (data not shown). Thus, treatment of human T cells with this potent Lck and FynT inhibitor blocks the increase in whole cell tyrosine phosphorylation observed following treatment with anti-CD3. These results are consistent with participation of these kinases in early T cell signal transduction; however, they do not rule out a role for other Src family kinases that may be present in T cells.
Inhibition of TcR-induced T Cell Proliferation by PP1-Since PP1 blocked anti-CD3-induced tyrosine phosphorylation, we next examined whether it could inhibit the proliferation of human T cells in response to different stimuli in repeated experiments using different donors (Table II). Human PBL were isolated using Histopaque and plated in the presence of the appropriate compound and stimulus. Proliferation was assessed using [ 3 ing different human PBL donors. An examination of the activity of PP1 yielded an average IC 50 of 0.5 M when PBL were treated with anti-CD3 in 10 different donors and 26 M when proliferation was induced by PMA/IL-2 in seven donors (Table II). Similar results were obtained using purified T cells (data not shown). An examination of other T cell proliferative signals revealed somewhat reduced potency for PP1 (Table II). Fresh human T cells derived from five donors stimulated in a one-way mixed leukocyte reaction were also inhibited with an average IC 50 of 3.9 M. T cells derived from healthy donors and stimulated with influenza virus vaccine (Ag-Sp) were inhibited with an average IC 50 of 5.2 M (Table II). A similar profile was observed with PP2. Thus, PP1 and PP2 demonstrated inhibition of tyrosine kinase-dependent TcR-induced T cell proliferation. However, they were less effective inhibitors of proliferative signals that by-pass the T cell receptor complex (e.g. PMA/ IL-2). As a control, staurosporine potently inhibited T cell proliferation in response to both alloantigen (MLR) and specific antigen (Ag-Sp) (Table II), in line with its reported potent broad inhibitory properties (36). Genistein, a previously characterized tyrosine kinase inhibitor (36,37), demonstrated less potent inhibition of T cell proliferation using all stimuli examined (Table II). Selective Inhibition of IL-2 Gene Induction by PP1-Since PP1 selectively depressed TcR-dependent proliferation of T cells, we tested whether this compound would demonstrate specificity for inhibition of lymphokine genes involved in T cell proliferation. Human PBL were isolated from a normal donor and were either left untreated (Fig. 4, lane 1 FK506 was used at 0.1 M as a control (Fig. 4, lane 5). IL-2 mRNA induction was almost completely inhibited by both 1 and 5 M PP1 (Fig. 4, lanes 3 and 4, respectively). However, the GM-CSF, IL-2R, and glyceraldehyde-3-phosphate dehydrogenase mRNA levels were not significantly affected by these concentrations of PP1. Interestingly, IL-2R mRNA levels were slightly enhanced in the presence of PP1. In contrast, FK506, a powerful immunosuppressive that is an inhibitor of the Ca 2ϩdependent phosphatase calcineurin (41), potently suppressed IL-2 and GM-CSF mRNA induction (Fig. 4, lane 5), as previously reported (42). IL-2R mRNA induction was also partially repressed by this concentration of FK506. These results suggest that PP1 may selectively inhibit signaling events required for the activation of the IL-2 gene and that the signaling events leading to GM-CSF and IL-2R induction may be distinct.
To further test this hypothesis, we examined whether PP1 could selectively inhibit IL-2 over IL-2R gene expression at the level of transcription using reporter plasmids containing either the IL-2 or IL-2R promoter linked to the firefly luciferase gene (Fig. 5). Human Jurkat T cell lines were transfected with these reporter constructs, and the cells were stimulated with PHA (2 g/ml) and PMA (10 ng/ml) for 18 h prior to assay for luciferase activity as described previously (32). As shown in Fig. 5A, both the IL-2 and IL-2R promoter constructs were induced to express luciferase in response to PHA ϩ PMA (compare uninduced (U) and stimulated (S) bars for each reporter construct). PP1 demonstrated dose-dependent inhibition of the IL-2 reporter construct (Fig. 5B, squares) with an IC 50 of ϳ1 M. In contrast, the IL-2R reporter was not inhibited by PP1 up to concentrations of 35 M (Fig. 5B, circles). These results are consistent with the Northern data and suggest that compound PP1 is capable of selectively inhibiting signaling events required for IL-2 gene induction. DISCUSSION In this study, we have disclosed the structure and activity of a novel tyrosine kinase inhibitor (PP1) that potently inhibits Lck and FynT, anti-CD3-induced protein tyrosine phosphorylation, and subsequent IL-2 gene activation in T lymphocytes. Moreover, this compound shows selectivity for the Src family over other families of tyrosine kinases including ZAP-70, JAK2, and the EGF-R kinase.
Several properties distinguish PP1 from previously identified tyrosine kinase inhibitors. Its combination of potency (nanomolar in vitro and low micromolar in intact cells) and selectivity for members of the Src family of tyrosine kinases is unprecedented. Earlier studies have reported the ability of other tyrosine kinase inhibitors to inhibit signal transduction in T lymphocytes. Prolonged treatment (12-16 h) with 1 M herbimycin A, a benzoquinoid ansamycin antibiotic, was shown to inhibit TcR-mediated tyrosine phosphorylation, inositol phospholipid hydrolysis, and calcium elevation (40). The lengthy incubation required to observe inhibition and the covalent interaction of herbimycin A with sulfhydryl groups on protein-tyrosine kinases (37) have limited its use. The isoflavone genistein (43) has been also shown to block T cell receptor signal transduction and early activation events (44). In those studies, incubation with 40 M genistein prior to receptor crosslinking with anti-CD3 significantly inhibited Lck activity and TcRphosphorylation as well as activation of the IL-2 gene, but was unable to inhibit IL-2 secretion induced by Ca 2ϩ ionophore and PMA, agents that by-pass the TcR. The high concen-  trations of genistein required to observe inhibition in vitro and in whole cells and its lack of specificity in protein kinase inhibition (36) contrast with the potency and specificity we have observed with PP1. A styryl-based protein-tyrosine kinase inhibitor, 67B-83-A (45), was previously shown to be selective for Lck as compared with EGF-R (Ͼ100-fold), and it was also selective for inhibition of Lck when compared with other Src family kinases (from Ͼ10-fold to Ͼ200-fold). However, the compound had an IC 50 of 7 M for inhibition of Lck in vitro, several orders of magnitude higher than that which we have observed for PP1 using similar protocols for kinase inhibition. Staurosporine, a member of the indolecarbazole group of antibiotics, is also a potent inhibitor of Src family tyrosine kinases in vitro. However, in our hands, staurosporine demonstrated less specificity than PP1 and potently inhibited Lck, FynT, ZAP-70, and JAK2 (Table II). Other investigators have similarly reported that staurosporine is a potent inhibitor of Src family kinases (IC 50 ϭ 90 -200 nM) and was also effective for inhibition of nonreceptor tyrosine and serine/threonine protein kinases (36,37). A new quinolone derivative, WIN 61651, has also been described as an inhibitor of p56 lck (46). This compound is significantly less potent than PP1 for inhibition of Lck (18 -24 M) and appears to be less selective since it demonstrates equal potency for the platelet-derived growth factor receptor. Finally, the tyrphostins have been extensively studied for their ability to inhibit various classes of protein-tyrosine kinases (47,48). Tyrphostins generally show selectivity for protein-tyrosine kinases over other classes of kinases such as protein kinase A, protein kinase C, or Ca 2ϩ /calmodulin-dependent kinases. However, less information is available concerning the activities of these compounds on T cell kinases and function. Although studies of tyrosine kinase inhibitors are confounded by the lack of standardized systems for comparison of specificity and potency of protein kinase inhibition, our data support the conclusion that PP1 is the most potent and selective inhibitor of Src family tyrosine kinases such as Lck and FynT reported to date.
The ability of PP1 to dissect signaling pathways was most evident in two of our experiments. First, PP1 was effective at blocking anti-CD3-induced T cell activation events, while it was less effective at inhibiting the TcR-independent proliferation induced by PMA and IL-2. Second, using Northern and reporter assays, we found that PP1 was more effective at inhibiting IL-2 gene expression than either GM-CSF or IL-2R gene induction. These results suggest that Lck and FynT may play a specific role in IL-2 gene expression required for TcRinduced T cell proliferation, but not in the induction of the GM-CSF or IL-2R genes. In contrast, FK506 inhibited anti-CD3-induced expression of IL-2 and GM-CSF mRNAs. Thus, the inhibitory effects of PP1 appear to be more specific than those of FK506. FK506 forms a complex with FKBP, and the FK506-FKBP complex competitively binds to and inhibits the Ca 2ϩ /calmodulin-dependent phosphatase calcineurin (41). This results in the inhibition of genes that respond to TcR Ca 2ϩ -dependent signaling such as IL-2 and GM-CSF (42). Since PP1 inhibits Ca 2ϩ flux in anti-CD3-stimulated T cells (data not shown), it was anticipated that FK506 and PP1 would show similar patterns for inhibition of TcR-induced genes in T cells. However, FK506 and PP1 differentially affected GM-CSF mRNA induction. This suggests that a Ca 2ϩ -independent signal may emanate from the TcR to induce GM-CSF expression and that FK506, but not PP1, inhibits this signal. The nature of this other signal and the mechanism by which FK506 may inhibit this other pathway remain to be defined.
The potency of PP1 in inhibiting intracellular activities (e.g. T cell tyrosine phosphorylation and proliferation) was considerably reduced relative to inhibition of isolated kinase activity. This difference may be attributed in part to permeability of the compound and its distribution within the cell, but may also be a reflection of the kinetic characteristics of the inhibitor itself. Preliminary studies with PP1 (data not shown) show complex kinetics for inhibition of Lck; however, at certain concentrations of ATP, the compound appears to be competitive with ATP. It therefore is possible that the intracellular millimolar concentrations of ATP found within the cell (49) would act to decrease the potency of the inhibitor in vivo. Studies are underway to carefully define the kinetics of Lck inhibition by PP1 and related compounds. Although this property of PP1 and its related compounds appears to limit their usefulness as pharmacological agents in the treatment of T cell-mediated disease, their discovery represents a significant advance in the use of tyrosine kinase inhibitors as tools to study the role of Lck and FynT in T cell signaling.