Inhibition of calcineurin-NFAT signaling by the pyrazolopyrimidine compound NCI3

Authors


Abstract

Dephosphorylation of NFAT by the Ca2+-calmodulin-dependent Ser/Thr protein phosphatase calcineurin is a bottleneck of T cell receptor-dependent activation of T cells. In dimeric complexes with immunophilins, the immunosuppressants cyclosporine A (CsA) and tacrolimus (FK506) block this process by inhibition of the enzymatic activity of calcineurin. We have identified the pyrazolopyrimidine compound NCI3 as a novel inhibitor of calcineurin-NFAT signaling. Similar to CsA and FK506, NCI3 inhibits dephosphorylation and nuclear translocation of NFAT, IL-2 production and proliferation of stimulated human primary T cells with IC50 values from 2 to 4.5 µM. However, contrary to CsA and FK506, NCI3 neither blocks calcineurin`s phosphatase activity nor requires immunophilins for inhibiting NFAT activation. Our data suggest that NCI3 binds to calcineurin and causes an allosteric change interfering with NFAT dephosphorylation in vivo but not in vitro. NCI3 acts not only on the endogenous calcineurin but also on a C-terminally truncated, constitutively active version of calcineurin. The novel inhibitor described herein will be useful in better defining the cellular regulation of calcineurin activation and may serve as a lead for the development of a new type of immunosuppressants acting not by direct inhibition of the calcineurin phosphatase activity.

Abbreviations:
CK1:

casein kinase 1

CsA:

cyclosporin A

Cyp18:

cyclophilin18

FK506:

tacrolimus

GSK3β:

glycogen synthase kinase 3β

NCI:

NFAT-calcineurin-signaling inhibitor

PKA:

protein kinase A

pNPP:

p-nitrophenyl phosphate

PP1:

Ser/Thr protein phosphatase type 1

PP2A:

protein phosphatase type 2A

PP2B:

protein phosphatase type 2B or calcineurin

PP2C:

protein phosphatase 2C

RII phosphopeptide:

19-residue phosphopeptide of the regulatory subunit of type II cAMP-dependent protein kinase

Introduction

Ligation of T cell receptors (TCR), in conjunction with costimulatory receptors, triggers a complex signaling network that activates numerous transcription factors controlling the gene expression program of T cell activation. In this process, members of the prominent nuclear factor of activated T cells (NFAT), nuclear factor κB (NF-κB) and AP1 transcription factor families regulate the expression of proteins, such as IL-2 and CD25 (IL-2 α receptor), which are essential for T cell proliferation 1. NFAT factors have been shown to be useful targets for the inhibition of T cell activation and proliferation by the immunosuppressants cyclosporin A (CsA) and tacrolimus (FK506) which are widely used in transplantation medicine to prevent transplant rejection 2, 3.

Four of the five NFAT family members, designated as NFATc1–c4, are structurally related cytoplasmatic proteins that are activated upon Ca2+ mobilization. Three of them, NFATc1 (also designated as NFATc or NFAT2), NFATc2 (NFATp or NFAT1) and NFATc3 (NFATx or NFAT4), are expressed in lymphoid cells, particularly in T lymphocytes 4. NFAT proteins are regulated by a reversible cycle of dephosphorylation and rephosphorylation. A TCR ligation-induced rise in intracellular free calcium levels triggers the activation of the Ca2+-calmodulin-dependent Ser/Thr-specific protein phosphatase calcineurin, which is unique in its property to dephosphorylate NFAT proteins 5. Once activated, NFAT proteins translocate into the nucleus, bind to regulatory elements of NFAT target genes, interact with other nuclear proteins and control transcription. When calcineurin is inactivated by a decrease in free Ca2+ level, NFAT factors are rapidly re-phosphorylated and exported from the nucleus to the cytoplasm 6.

Calcineurin is a highly conserved enzyme composed of two subunits, the regulatory subunit containing the Ca2+ binding sites and the catalytic subunit bearing the active center of the phosphatase and the calmodulin binding site 7. The immunosuppressants CsA and FK506, in dimeric complexes with their cellular partners, the immunophilins, block the access to the active center of calcineurin and therefore inhibit its enzymatic activity toward all physiological substrates 811. Although NFAT factors are the most prominent targets of calcineurin, other transcription factors, such as TORC2, a CREB coactivator 12, have also been described to be under calcineurin control and therefore are affected by CsA and FK506. This might explain that CsA and FK506, in addition to their immunosuppressive action, show a number of side effects in other tissues which can lead to severe kidney failure and further clinical complications. It is suggested that compounds interfering solely with the interaction between calcineurin and NFAT factors, but leaving other calcineurin substrates unaffected, exert less side effects than CsA and FK506 13, 14.

This assumption prompted us to screen for novel calcineurin inhibitors. Here, we describe the identification and molecular characterization of the pyrazolopyrimidine NFAT-calcineurin-signaling inhibitor (NCI)3 as a non-cytotoxic, low-molecular-weight inhibitor of NFAT activation that blocks NFAT signaling via calcineurin without interfering with its phosphatase activity.

Results

NCI1, NCI2, and NCI3 differ in their binding and inhibiting properties

By testing a substance library of heterocyclic compounds synthesized chemically, we identified the compound NCI1 (Fig. 1) as a substance that is able to inhibit the dephosphorylation of a 19-residue phosphopeptide of the regulatory subunit of type II cAMP-dependent protein kinase (RII phosphopeptide) by recombinant human calcineurin 15 (Fig. 2A). To determine the structure-function relationship of pyrazolopyrimidine derivatives we tested 76 compounds, in more detail the compounds NCI1, NCI2, and NCI3. These three compounds have a similar structure (Fig. 1) but differ in the character of the side chain and a methyl group that is lacking in NCI1. In contrast to NCI1, NCI2 and NCI3 exerted only a weak or no inhibitory activity, respectively, on calcineurin's phosphatase activity against the RII phosphopeptide (Fig. 2A). Next, we tested whether, similar to the CsA/cyclophilin18 (Cyp18) complex, the pyrazolopyrimidines affect dephosphorylation of the small synthetic phosphatase substrate p-nitrophenyl phosphate (pNPP) by calcineurin, reflecting conformational changes of calcineurin induced by inhibitor binding. While NCI1 did not affect the phosphatase activity against pNPP, NCI2 and NCI3 (10 µM) inhibited slightly the dephosphorylation of pNPP by 10–20% (Fig. 2B).

Figure 1.

Structures of the pyrazolopyrimidine derivatives NCI1, NCI2, and NCI3. NCI1, NCI2, and NCI3 differ in the character of the side chain, and the methyl group is lacking in NCI1.

Figure 2.

Inhibition of the calcineurin phosphatase activity by the pyrazolopyrimidine derivatives. (A) NCI1 (triangles), NCI2 (squares), and NCI3 (circles) were incubated with calcineurin (1.32 nM) and calmodulin (50 nM) at 22°C for 30 min. The calcineurin phosphatase activity was determined by dephosphorylation of biotinylated RII phosphopeptide or (B) pNPP (lower panel). Activity is expressed as percent related to the calcineurin activity of an inhibitor-free sample. Each data point represents the mean of triplicates ± SD.

NCI1 and 2, but not NCI3, inhibit calcineurin phosphatase activity

In its active site, calcineurin shares several common structural features with other Ser/Thr protein phosphatases 16. Therefore, we were interested to see whether the NCI compounds inhibit other common Ser/Thr protein phosphatases. Using phospho-casein for the detection of protein phosphatase type 2A (PP2A), calcineurin (or protein phosphatase type 2B; PP2B) and protein phosphatase type 2C (PP2C) activity, and phosphorylase a for Ser/Thr protein phosphatase type 1 (PP1) phosphatase activity, we did not observe any inhibitory activity of NCI3. At concentrations of 10 µM, however, NCI1 inhibited strongly and NCI2 more moderately the PP2B/calcineurin activity but did not affect the activity of any other Ser/Thr protein phosphatase (Table 1).

Table 1. Inhibitory effect of NCI compounds (10 µM) on the enzymatic activity of the major classes of Ser/Thr phosphatases
 Phosphatase activity (% of control)a)
  1. a) The results are expressed as % of the respective phosphatase activity without inhibitor (SD ⩽10%). The activity was measured against phospho-casein, except for PP1.

  2. b) PP1 phosphatase activity was measured against phosphorylase a.CaN, calcineurin.

 NCI1NCI2NCI3
PP1b)949597
PP2A9810195
PP2B (CaN)258097
PP2C9710499

NCI3 acts as an immunosuppressant

To evaluate the immunosuppressive capacity of the three NCI compounds and to determine their cytotoxic effects, first we controlled the survival of stimulated primary human T cells treated with these compounds for 24 h. NCI1 and, to a lower degree, NCI2 induced cell death (Fig. 3A), whereas at concentrations up to 30 µM, NCI3 did neither influence the survival of stimulated nor unstimulated T cells (Fig. 3A, and data not shown). Therefore, all subsequent experiments were performed with NCI3 only.

Figure 3.

NCI3 inhibits intracellular IL-2 production and proliferation of stimulated human T cells and is not cytotoxic. (A) Isolated human PBMC were stimulated with immobilized anti-CD3/CD28 antibodies for 24 h in the presence of NCI1 (triangles), NCI2 (squares), NCI3 (circles), or DMSO as control. Dead cells were directly analyzed after staining with propidium iodide by flow cytometry. The frequency of dead cells in the DMSO control was 15.6%. Data are means of triplicates ± SD. (B) Sorted human CD4+ T cells stimulated with anti-CD3/CD28 antibodies for 8 h (closed circles, solid line) or with PMA/ionomycin for 5 h (open circles, dashed line) in the presence of NCI3 or DMSO were treated with brefeldin A for the last 3 h before fixation. Intracellular IL-2 production was determined by staining with anti-IL-2-PE antibody and measured by flow cytometry. The percentage of IL-2-producing cells relative to the DMSO control was plotted against the concentration of NCI3. Data are means of triplicates ± SD. (C) Sorted human CD4+ T cells were labeled with CFSE and stimulated with immobilized anti-CD3/CD28 antibodies in the presence of NCI3 or DMSO as control. After 5 days, the proliferation was measured by flow cytometry. The percentage of proliferating cells from total cells was plotted against the used concentrations of NCI3. Data are means of triplicates ± SD. (D) Histograms of CFSE-stained cells show the proliferation of selected samples.

In a concentration-dependent manner, NCI3 inhibited IL-2 production of human PBMC stimulated by PMA/ionomycin or by anti-CD3/CD28 antibodies. Using intracellular cytokine staining, we determined 3 µM as the half maximal concentration of NCI3 (IC50 value) to inhibit the number of IL-2-producing cells with PMA/ionomycin stimulation and 4 µM with CD3/CD28 stimulation (Fig. 3B). Likewise, an IC50 value of 4.5 µM was determined for the inhibition of proliferation of sorted primary human CD4+ T cells induced by anti-CD3/CD28 antibodies (Fig. 3C). NCI3 lowered both the number of proliferating cells and cell divisions in a concentration-dependent manner. The inhibitory effect of NCI3 on proliferation could be partially abrogated by addition of IL-2 to the culture medium (Fig. 3D). The similar IC50 values observed for the two different T cell stimulations suggest that in T cells the action of NCI3 is downstream of calcium mobilization and PKC activation.

NCI3 blocks specifically TCR signaling

IL-2 gene transcription is dependent on the activation of numerous inducible transcription factors, including NFAT, NF-κB, and AP1 factors 17. Their activation is the result of ligation of TCR, of costimulatory receptors such as CD28, or of both types of receptors. Ionomycin and PMA can mimic this signaling to a large extent by the direct activation of calcium- and PKC-dependent pathways. Using reporter genes driven by multiple factor binding sites in transient transfection assays, we studied the effect of NCI3 on NFAT, NF-κB, and AP1 transcription factors in Jurkat T cells.

After PMA/ionomycin stimulation, Jurkat cells transfected with luciferase reporter gene plasmids responded in a dose-dependent manner to NCI3 pretreatment (Fig. 4). In the tested concentration range of 0.1 up to 20 µM, NCI3 did not affect luciferase activity in cells transfected with AP1-reporter plasmids or the constitutively active reporter pGL3-luc plasmid (data not shown). Comparison of the NCI3-mediated suppression of the NFAT- and NF-κB-luciferase reporter gene expression with the CsA-mediated effects indicated that the pathways inducing the activity of these two transcription factors exhibited different sensitivity to NCI3 and CsA. Whereas CsA inhibited the expression of both reporter gene constructs with similar efficiencies (IC50 of 2.0 and 4.0 nM) 18, NCI3 exhibited IC50 values of 2 ± 0.1 µM for the NFAT and 7 ± 1 µM for the NF-κB-driven construct. Interestingly, the TNF-α-mediated stimulation of the NF-κB reporter plasmid in Jurkat cells remained unaffected by CsA and NCI3 (Table 2), suggesting that in this signaling pathway calcineurin is not involved.

Figure 4.

NCI3 impairs the transcriptional activity of NFAT and NF-κB. Jurkat T cells transfected with luciferase reporter gene plasmids driven by four NFAT or five NF-κB binding sites were incubated with NCI3 or DMSO as control and then stimulated with PMA/ionomycin for 5 h. Data are expressed as % of luminescence activity of an inhibitor-free stimulated control. Data are means of triplicates ± SD.

Table 2. Comparison of inhibitory effects of CsA/Cyp18 and NCI3
 CsA or CsA/Cyp18NCI3
  1. a) The numbers are IC50 values, if not further specified.

  2. b) “–” indicates that no inhibitory effect was observed in the tested concentration range.

  3. c) The numbers are relative values compared to the control without inhibitor. Concentrations were 2.5 µM CsA/Cyp18 and 20 µM NCI3.

  4. d) The kind of T cell stimulation is specified in brackets.

  5. e) The tested concentrations were up to 500 nM.

  6. f) The tested concentrations were up to 20 µM.CaN, calcineurin; n.d., not determined.

CaN activity – RII peptide100 nMa)b)
CaN activity – pNPPc)200%70%
IL-2 production (PMA/ionomycin)d)5 nM3 µM
IL-2 production (anti-CD3/CD28)n.d.4 µM
Proliferation (anti-CD3/CD28)5 nM4.5 µM
NFAT-luciferase (PMA/ionomycin)2 nM2 µM
NF-κB-luciferase (PMA/ionomycin)4 nM7 µM
AP1-luciferase (PMA/ionomycin)e)f)
NF-κB-luciferase (TNF-α)
Inhibition of Cyp18yesno

In order to delineate the mechanism of NCI3 action, we examined its effect on the nuclear translocation of NFATc2. As shown in Fig. 5A, upon stimulation of human CD4+ T cells with PMA/ionomycin, NFATc2 translocates rapidly into the nucleus. This process can be blocked by both CsA (50 nM) and NCI3 (10 µM). The translocation of NFATc2 from the cytosol to the nucleus depends on the dephosphorylation of the NFAT regulatory domain. Consequently, we investigated next whether the NCI3-dependent block in nuclear translocation of NFATc2 was accompanied by the inhibition of its dephosphorylation. Jurkat cells were stimulated by ionomycin with or without pretreatment with inhibitors, and NFATc2 was identified in whole cell lysates by immunoblotting 19. As shown in Fig. 5B, CsA (50 nM) and NCI3 (10 µM) totally inhibited the dephosphorylation of NFATc2 in stimulated Jurkat cells. The same result was obtained in CD3/CD28-stimulated human CD4+ T cells.

Figure 5.

NCI3 inhibits dephosphorylation of NFATc2. (A) NFATc2 translocation into the nucleus is inhibited by NCI3. Cellular localization of NFATc2 was detected in human CD4+ T cells by immunocytochemistry. Cells were stimulated with PMA/ionomycin in the absence of a compound, or in the presence of CsA (50 nM) or NCI3 (10 µM). The unstimulated sample without inhibitors served as control. (B) NCI3 suppresses dephosphorylation of NFATc2 upon stimulation of Jurkat T cells and sorted human CD4+ T cells. Cells were pretreated for 1 h with DMSO as control, NCI3 or CsA as indicated. Jurkat cells were then stimulated with ionomycin for 1 h in the presence of inhibitors. CD4+ cells were stimulated with anti-CD3/CD28 antibodies for 2 h. Whole cell protein extracts were prepared, and NFATc2 was immunodetected in Western blots.

NCI3 partially interferes with the calcineurin VIVIT peptide interaction

The NFAT analogue VIVIT peptide was shown to bind efficiently to calcineurin 20. We evaluated whether NCI3 inhibits this interaction by using FITC-labeled calcineurin, biotinylated 17mer VIVIT peptide and streptavidin-coupled microspheres. By flow cytometry, the VIVIT peptide-bound calcineurin was detected as particle-associated fluorescence. The specificity of the test was shown by a backshift of the fluorescence signal by unlabeled calcineurin or unbiotinylated VIVIT peptide (data not shown). In addition, the biotinylated reversed VIVIT peptide TIVIV did not induce any specific fluorescence signal, verifying that unspecific binding of FITC-calcineurin to biotinylated VIVIT peptide does not play a role (Fig. 6A). NCI3 inhibits the binding of FITC-labeled calcineurin to VIVIT peptide in a concentration-dependent manner, but only to a maximum of 55% (Fig. 6A). As expected, the calcineurin phosphatase inhibitor CsA did not have any inhibitory effect on VIVIT peptide binding to calcineurin (data not shown).

Figure 6.

NCI3 partially interferes with the binding of calcineurin to the VIVIT peptide but does not inhibit the dephosphorylation of NFATc2 and Elk-1 by calcineurin in vitro. (A) FITC-labeled calcineurin was incubated with biotinylated peptides and then added to streptavidin-coated Sepharose beads. Via flow cytometry, the peptide-bound calcineurin was detected as particle-associated fluorescence. In samples with inhibitor, FITC-labeled calcineurin was pre-incubated with NCI3. Left, data are from one representative experiment out of eight. Right, data are from two experiments in triplicates each and are shown as means ± SD. (B) Lysates from HEK293 cells overexpressing NFATc2 were incubated with recombinant human calcineurin in the absence or presence of 30 µM TIVIV or VIVIT peptide, or NCI3 or combinations of them. (C) GST-Elk-1 (aa 307–428) was phosphorylated by Erk1 and incubated with recombinant human calcineurin with or without 30 µM NCI3. Unphosphorylated GST-Elk-1 served as a positive control. GST-Elk-1 and NFATc2 were immunoblotted. The results were consistent in three separate experiments.

NCI3 does not activate the kinase activities of cAMP-dependent protein kinase and GSK3β

In principle, hyperactivation of NFATc2 kinases by small molecular compounds could oppose NFAT dephosphorylation by calcineurin and therefore lead to inhibition of T cell activation 21. However, NCI3 did not influence phosphorylation of recombinant NFATc2 by cAMP-dependent protein kinase (PKA) and glycogen synthase kinase 3 (GSK3) in vitro (data not shown). Moreover, hyperactivation of GSK3 in human CD4+ cells by NCI3 was not observed in immunoblotting with a phospho-specific antibody against GSK3β (data not shown). CK1 failed to phosphorylate recombinant NFATc2 in our assay.

NCI3 does not inhibit NFAT and Elk-1 dephosphorylation by calcineurin in vitro

As shown, NCI3 only minimally interferes with the dephosphorylation of the small molecular compound pNPP and does not block dephosphorylation of the RII phosphopeptide and the phosphorylated protein casein by recombinant human calcineurin. Next, we tested the in vitro effect of NCI3 on full-length NFATc2 and Elk-1; both are physiological substrates of calcineurin 22. In contrast to the VIVIT peptide or 50 nM CsA (not shown), NCI3 does not inhibit the dephosphorylation of phospho-NFATc2 up to the measured concentration of 30 µM (Fig. 6B). The reverse VIVIT peptide, TIVIV, is not inhibitory, confirming the specificity of the VIVIT peptide to block dephosphorylation of NFAT by calcineurin. As shown in Fig. 6C, NCI3 fails to inhibit dephosphorylation of recombinant phospho-Elk-1 by calcineurin in vitro. All results were confirmed using calcineurin isolated from bovine brain.

NCI3 affects even the truncated, constitutively active calcineurin in cells

In order to delineate the calcineurin domain(s) to which NCI3 might bind, we studied the effect of NCI3 on NFAT activity in reporter gene assays in Jurkat cells that were transiently transfected with DNA constructs encoding the full-length calcineurin subunit A (aa 1–521) or a C-terminally truncated, constitutively active version of calcineurin subunit A (aa 1–396) lacking its autoinhibitory domain. As shown in Fig. 7, overexpression of both calcineurin constructs partially rescued the NFAT-driven luciferase activity by 1 µM NCI3. NCI3 at 5 µM efficiently inhibits the activation of NFAT by both full-length and truncated calcineurin, suggesting that the putative binding site(s) of NCI3 is located within the N-terminal aa 1–396 of calcineurin. Further substantial truncation of calcineurin is not possible because the linker region (aa 335–347) has just been shown to be essential for the binding of calcineurin to NFAT 23.

Figure 7.

NCI3 inhibits NFAT activation by both full-length calcineurin and a truncated, constitutively active version of calcineurin. Jurkat T cells cotransfected with DNA plasmids encoding full-length calcineurin or a truncated, constitutively active version of calcineurin, or control plasmid pEGFP-N3 and NFAT-luciferase reporter plasmid were pre-incubated with 1 and 5 µM NCI3 or DMSO as control, followed by stimulation with PMA/ionomycin for 5 h. Data are expressed as % of luminescence activity of control plasmid-transfected cells stimulated without inhibitors. Data are means of triplicates ± SD.

Discussion

In a search for novel calcineurin inhibitors that affect specifically NFAT activation, we have discovered NCI3 which represents a new type of inhibitor of calcineurin-NFAT signaling. In contrast to the well-known immunosuppressants CsA and FK506 which suppress the phosphatase activity of calcineurin, NCI3, although blocking NFAT dephosphorylation in cells, does not affect the phosphatase activity of calcineurin, neither against peptides nor against proteins. However, NCI3 slightly inhibits the phosphatase activity of calcineurin toward the synthetic substrate pNPP. These properties of NCI3 differ remarkably from those of CsA which, in dimeric complexes with Cyp18, inhibits the calcineurin-mediated dephosphorylation of the RII phosphopeptide but elevates twofold its phosphatase activity toward pNPP (Table 2) 2, 18, 24. Thus, both NCI3 and the CsA/Cyp18 complex modulate calcineurin's phosphatase activity against pNPP, probably by binding to and changing the conformation of calcineurin.

Our results show that NCI3 inhibits NFAT signaling and prevents T cell activation. First, we demonstrated that NCI3 suppresses the dephosphorylation of NFATc2 in CD4+ and Jurkat T cells. This has been shown by immunodetection of the phosphorylated, slowly migrating version of NFATc2 in whole cell protein extracts from CD4+ or Jurkat T cells stimulated by antibodies or ionomycin, respectively, in the presence of NCI3 (Fig. 5B). Second, using NFATc2-specific antibodies in fluorescence microscopy, we could show that the nuclear translocation of NFATc2 is inhibited by NCI3 in primary human T cells. Third, we demonstrated that IL-2 synthesis, which depends on NFAT activation, is blocked by NCI3 with an IC50 value of ∼3 µM. Consequently, T cell proliferation is also inhibited by NCI3, albeit exhibiting a slightly higher IC50 value (4.5 µM).

Our experiments show that NCI3 is able to penetrate the cell membrane and to act in both Jurkat T cells and primary human T cells. Inhibitory effects on NFAT activation caused by general cytotoxic effects of NCI3 can be ruled out, because NCI3 did not induce cell death in primary human T cells, as verified by a low number both of dead cells using propidium iodide staining (Fig. 3A) and apoptotic cells using annexin V staining (data not shown). Furthermore, luciferase reporter gene assays indicate that NCI3 does not diminish the activity of constitutively expressed luciferase.

The results of our experiments with primary human T cells show that NCI3 acts independently of the stimulation method, i.e. it suppressed both anti-CD3/CD28 antibody- and PMA/ionomycin-mediated T cell activation with comparable IC50 values. Both stimuli are mimicking the antigen-specific stimulation of T cells, albeit at different levels. Anti-CD3/CD28 antibodies target TCR and costimulatory receptors at the cell surface, whereas PMA and ionomycin appear to act downstream of the receptors by activating directly protein kinase C and Ca2+ influx into the cell, respectively. Therefore, we conclude that NCI3 is operating downstream of cytosolic Ca2+ level/PKC.

Not only inhibition of NFAT dephosphorylation but also enhanced NFAT rephosphorylation by selected kinases can lead to an abrogation of NFAT-mediated gene expression by export of the phosphorylated protein into the cytosol. We could rule out a NCI3-induced activation of the prominent NFAT kinases PKA and GSK3β by NCI3 (data not shown). However, it remains open whether NCI3 enhances the activity of other kinases like casein kinase 1 (CK1), p38 mitogen-activated protein kinase (MAPK), or dual-specificity tyrosine phosphorylation-regulated kinase 1 (DYRK1), which might oppose calcineurin-NFAT signaling.

There are some hints for direct interaction of NCI3 and calcineurin. First, NCI3 partially interferes with calcineurin VIVIT interaction in the binding assay. Second, the structurally related compounds NCI1 and NCI2 specifically inhibit calcineurin but not the other Ser/Thr protein phosphatases PP1, PP2A, and PP2C, all of them sharing similar active centers. This points to binding of NCI compounds to remote sites. Third, NCI3 interferes slightly with dephosphorylation of the synthetic substrate pNPP by calcineurin. Fourth, NCI3 is acting downstream of cytosolic Ca2+ level/PKC and at or upstream of NFAT dephosphorylation. Calcineurin is located directly at this position in the TCR signaling cascade. Furthermore, calcineurin is unique for its activation by Ca2+/calmodulin as well as its ability to dephosphorylate cellular NFAT 5. Fifth, NCI3 inhibits TCR-dependent NFAT and NF-κB signaling, which are both affected by calcineurin 25. Sixth, NCI3 effects are partially rescued by ectopically expressed calcineurin 26 in Jurkat T cells.

The molecular mechanism of NCI3 action is unclear. The inhibitory pattern of NCI3 and CsA in T cells is comparable (Table 2). It is known, that CsA acts on calcineurin and inhibits its activity. The underlying molecular mechanism of inhibition, however, differs. In contrast to CsA, NCI3 does not inhibit NFATc2 dephosphorylation by calcineurin in vitro. Our data (Fig. 6A) from the VIVIT-calcineurin binding assay show that NCI3 interacts with calcineurin and thereby partially interferes with its binding properties towards the VIVIT peptide. However, data from the in vitro dephosphorylation assay (Fig. 6B) rule out that this is the main effect of NCI3 within T cells, since NCI3 cannot rescue NFAT dephosphorylation in samples containing VIVIT peptide. We assume that NCI3 binding modifies the conformation of calcineurin and therefore might not only change the NFAT-binding region but also interfaces for other important proteins. The changes in these other calcineurin regions might be more drastic and could confer the main effect of NCI3 in vivo by modulating the binding of other proteins beside NFAT to calcineurin. Thereby NCI3 might disturb the interaction between calcineurin and other partner proteins. These proteins could have the function of a cofactor or anchor in the cell. In vitro these proteins might not influence the dephosphorylation of NFAT by calcineurin because they are, under these conditions, either not in the proper concentration or localization. Exemplarily, for Bcl-2 it was already shown 27 that it translocates calcineurin to the endoplasmatic reticulum and mitochondria and thereby away from NFAT while leaving the phosphatase activity intact. Thus, Bcl-2 inhibits NFAT dephosphorylation by calcineurin only in an intact cell system but not in an in vitro assay with cell lysates.

NCI3 might have fewer side effects than the commonly used immunosuppressive drugs CsA and FK506, because a compound blocking the calcineurin-NFAT signaling and not the general phosphatase activity of calcineurin is expected to have a lower toxicity 14, 20. However, it is unclear to what extent the toxicity of CsA and FK506 is due to inhibition of NFAT, to interference with dephosphorylation of other calcineurin substrates, to inhibition of the peptidyl-prolyl-cis/trans-isomerase activity of the immunophilins, or to other non-calcineurin-dependent effects. The compound NCI3 might allow us to address these questions directly. In contrast to the VIVIT peptide and even specifically designed peptide derivatives, NCI3 is able to enter the cells and has no transport limitations through the cell membrane. Furthermore, NCI3 might be used as a lead to develop novel drugs useful in treating not only transplant rejection but also autoimmune diseases.

Materials and methods

Materials

PP1 (recombinant rabbit muscle α-isoform) was purchased from Calbiochem (Bad Soden, Germany) and PP2C (recombinant human α-isoform) from Upstate Biotechnology (Biomol, Hamburg, Germany). PP2A was isolated from porcine brain and purified to homogeneity as described previously 28. The catalytic subunit of bovine heart PKA, calmodulin, buffers, salts, and calcineurin isolated from bovine brain were obtained from Sigma (Steinheim, Germany). GSK3β, CK1, recombinant GST-Elk-1 and the anti-Elk-1 antibody were purchased from Cell Signaling (NEB, Frankfurt, Germany), and recombinant human activated Erk1 from Calbiochem.

Synthesis of pyrazolopyrimidine derivatives

The pyrazolopyrimidines NCI1, NCI2, and NCI3 were synthesized according to standard procedures. Cyclization of ethyl benzoylacetate with 3-amino-pyrazoles furnished the 7-hydroxy-pyrazolopyrimidine core. Subsequent reaction with chlorinating reagent, such as phosphorus oxychloride, affords the corresponding chloro-derivatives which were reacted with piperazine, 3-amino-propanol and 5-amino-pentanol to give the target compounds NCI1, NCI2 and NCI3, respectively (M.K., PhD thesis, Humboldt-University, Berlin 2005).

Stock solutions of all compounds were prepared in DMSO (Sigma) and added to each cell sample to a final concentration of 0.5% DMSO. Control samples with DMSO alone had always a final concentration of 0.5% DMSO.

Preparation of calcineurin and synthesis of peptides

Expression and purification of recombinant human calcineurin α was performed as published 15, 29. The biotinylated and non-biotinylated VIVIT peptide (MAGPHPVIVITGPHEE) 18, the reversed VIVIT peptide (EEHPGTIVIVPHPGAM), and the biotinylated RII peptide (DLDVPIPGRFDRRVSVAAE-OH) were synthesized by solid-phase peptide synthesis (Quartett, Berlin, Germany).

Inhibition of calcineurin and other protein Ser/Thr phosphatases

The biotinylated RII peptide was phosphorylated by PKA 15. Calcineurin activity was measured using a scintillation proximity assay or pNPP-based assay as previously described 15. In short, pre-incubation of calmodulin (50 nM), calcineurin (1.32 nM), and inhibitor at the required concentrations in the phosphatase assay buffer (40 mM Tris/HCl pH 7.5, 100 mM NaCl, 6 mM MgCl2, 0.5 mM dithiothreitol, 1 mM CaCl2, 0.1 mg/mL bovine serum albumin) was carried out at 22°C for 30 min in a 96-well microtiter plate (Costar, Bodenheim, Germany). Of biotinylated [33P]RII phosphopeptide, 10 pmol were added to each well in a total assay volume of 100 µL. After dephosphorylation of RII phosphopeptide by calcineurin at 30°C for 20 min, 90 µL of the reaction mixture were transferred to a scintillation well coated with streptavidin (PerkinElmer Life Sciences). Biotinylated RII phosphopeptide was allowed to bind to streptavidin for 20 min at 22°C. The well was washed once with water, and the RII phosphopeptide-associated 33P was measured in a MicroBeta top counter (Wallac, Turku, Finland).

The phosphatase activity of the four major Ser/Thr protein phosphatases PP1, PP2A, PP2B, and PP2C against protein substrates and their inhibition was assayed with biotinylated 33P-labeled phospho-casein or, in the case of PP1, with biotinylated 33P-labeled phosphorylase a, using a scintillation proximity assay as described previously 15, 18. The protein phosphatase concentrations were adjusted to an activity level of approximately 80% dephosphorylation of the substrate within 20 min at 30°C.

Determination of cytotoxic effects of the pyrazolopyrimidines

Human PBMC (2 × 106 cells/mL) were stimulated with plate-bound anti-CD3 + anti-CD28 antibody (1 and 4 µg/mL, respectively) in the presence of different concentrations of NCI1, NCI2, and NCI3, or DMSO as control. The cells were cultured in 24-well plates for 24 h and directly analyzed with propidium iodide on viable cells by flow cytometry.

Determination of proliferation and intracellular cytokine production

CD4+ T cells were isolated by positive selection from human PBMC using magnetic Multisort-MicroBeads (Miltenyi Biotech, Bergisch Gladbach, Germany) 30 to a purity of >97%. The sorted T cells were labeled with 5-(and 6-)carboxyfluorescein diacetate succinimidylester (CFSE) followed by stimulation with bead-coupled anti-CD3 + anti-CD28 antibody (1 and 4 µg/mL, respectively, for 2 × 106 beads/mL) in the presence of inhibitors, or DMSO as control in 24-well plates (2 × 106 cells/mL/well). The cells were cultured for 5 days and directly analyzed for viable cells by flow cytometry. For intracellular cytokine staining, PBMC (2 × 106/mL) were pre-incubated in 24-well plates with NCI3 at different concentrations of CsA (Calbiochem) or DMSO at 37°C for 30 min. PBMC were stimulated with 10 ng/mL PMA and 1 µg/mL ionomycin (both Sigma) for 5 h. Brefeldin A (5 µg/mL) was added for the last 3 h of stimulation. Cells were then fixed in 2% paraformaldehyde for 20 min, permeabilized by washing in PBS supplemented with 0.5% saponin and 1% FCS, incubated with an anti-IL-2 PE-conjugated antibody and measured by flow cytometry 18.

Plasmids

Human calcineurin α subunit A was cloned from the vector pETCNα, kindly provided by J. O. Liu 29. Full-length calcineurin and a truncated variant, consisting of aa 1–396 and lacking the C-terminal calmodulin binding and the autoinhibitory domains 31, were cloned into the pEGFP-N3 vector (Clontech, Heidelberg, Germany), using Hind III and Not I as restriction sites. The identity of plasmid constructions was confirmed by DNA sequencing.

Luciferase reporter gene assays

Jurkat T cells were electroporated according to an Amaxa protocol (Amaxa, Cologne, Germany) with NF-κB-, AP1- (Stratagene, Amsterdam, The Netherlands), NFAT- or pGL3-luciferase reporter gene plasmids (1–1.5 µg) (Promega, Mannheim, Germany). The transfected cells were cultured in RPMI 1640 with 10% FCS and 2 mM L-glutamine for 16 h at 37°C in 5% CO2. Aliquots of electroporated cells were pre-incubated with the inhibitors or 0.5% DMSO for 30 min, followed by stimulation with 10 ng/mL PMA + 1 µg/mL ionomycin or 100 ng/mL TNF-α for 5 h. After cell lysis, the level of the extracted luciferase was determined by bioluminescence measurement using the luciferase assay system (Promega). The constitutively active pGL3-luciferase plasmid was used to test toxic effects of the inhibitors.

For calcineurin overexpression, Jurkat T cells were cotransfected with a mixture of 1.5 µg NFAT-luciferase reporter plasmid and 1 µg calcineurin construct per 4 × 106 cells by electroporation (Amaxa). To equalize DNA amounts and as a negative control, the vector pEGFP-N3 was used.

Confocal microscopy

Sorted human CD4+ T cells were pre-incubated with DMSO (1 µL/mL) or inhibitors for 15 min, followed by stimulation with 10 ng/mL PMA and 1 µg/mL ionomycin for 60 min. Cells were washed, fixed in 3% formaldehyde, permeabilized (0.5% Triton-X 100) and stained with an NFATc2-specific monoclonal antibody (G1-D10; Santa Cruz Biotechnology, Santa Cruz, CA) followed by an anti-mouse IgG1-Alexa 488 antibody (Molecular Probes, Eugene, OR). Stained cells were centrifuged onto slides and analyzed by confocal microscopy (Microscope TCS SL; Leica, Heidelberg, Germany).

NFAT dephosphorylation in CD4+ and Jurkat T cells

Sorted human CD4+ or Jurkat T cells (5 × 106) kept in 0.5% FCS-containing RPMI for 40 h were pre-incubated at 37°C for 1 h with 50 nM CsA or various concentrations of NCI3 (2, 5 and 10 µM). Then, CD4+ cells were stimulated with bead-coupled anti-CD3 + anti-CD28 antibody and Jurkat cells with 0.5 µM ionomycin at 37°C for 3 h in the presence or absence of inhibitors. After centrifugation, cells were resuspended in lysis buffer (100 mM HEPES pH 7.4, 10 mM KAc, 2 mM MgAc, 2 mM EGTA, 0.05% NP40), followed by immunoblotting of NFATc2 using a polyclonal antibody raised against a recombinant human GST-NFATc2 protein.

Flow cytometry-based measurement of calcineurin VIVIT peptide interaction

FITC-labeled calcineurin (25 nM) was incubated with NCI3 in assay buffer (25 mM TrisCl pH 7.4, 3 mM MgCl2, 1 mM EGTA, 200 µM TCEP, 0.05% BSA) for 1 h at 4°C. Then, biotinylated VIVIT or TIVIV peptide was added to a final concentration of 300 nM and incubated for 15 min at 4°C. Streptavidin-coated Sepharose beads (Amersham Biosciences, Uppsala, Sweden) equilibrated in assay buffer were transferred into the assay mixture and incubated for an additional 10 min at 4°C. Bead-associated calcineurin was measured as fluorescence intensity by flow cytometry.

Kinase activities

Purified GST-NFATc2 from lysates of E. coli strain BL21(DE3) transfected with pGEX-4T-3-GST-NFATc2 was incubated with PKA and GSK3β in T4 DNA ligase buffer (Fermentas, St. Leon-Rot, Germany) or with CK1 in CK1 buffer (NEB) in the presence of 30 µM NCI3 or DMSO for 30 min at 30°C. The phosphorylation of GST-NFATc2 was assessed by immunoblotting with a mouse monoclonal antibody for NFATc2 (BD Biosciences, Heidelberg, Germany).

Sorted human CD4+ T cells were pre-incubated with NCI3 or DMSO for 30 min and then stimulated with PMA/ionomycin (10 and 1000 ng/mL) for 30 min at 37°C. Cell lysates were immunoblotted with anti-GSK3 and anti-phospho-GSK3 antibodies (Cell Signaling).

Dephosphorylation of NFATc2 and GST-Elk-1 by calcineurin in vitro

Lysates from HEK293 cells transfected with murine NFATc2 were used for the in vitro dephosphorylation experiments. Recombinant calcineurin (0.05 mg/µL) or calcineurin isolated from bovine brain were pre-incubated in dephosphorylation buffer (40 mM TrisCl pH 7.5, 100 mM NaCl, 6 mM MgCl2, 0.5 mM DTT, 1 mM CaCl2, 0.1 ng/mL BSA, 15 µg/mL calmodulin) at 30°C for 20 min. Then the substrates, either lysates with phospho-NFATc2 or phospho-Elk-1, were incubated at 30°C for 30 min. The phosphorylation/dephosphorylation of the substrates was assessed by immunoblotting with anti-Elk-1 rabbit polyclonal antibody and with mouse monoclonal anti-NFATc2 antibody.

Acknowledgements

We thank Vladimir Pavlovic, Katrin Moser, Stefanie Gross, and Henju Marjuki for their help and discussion. We are also grateful to Regina Schuck and Martina Heidler for excellent technical assistance. The work was supported by grants of the State of Berlin to A.R. and R.B., DRFZ Berlin, by the German Ministry of Education and Research (BMBF) within the National Genome Research Network NGFN-2 (01GS0413), and the Deutsche Forschungsgemeinschaft to J.L. (548/12-1), G.F. (455/8-1), and E.S. (469/15-1).

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

  7. 7

    WILEY-VCH

Ancillary