Dictyostelium Differentiation-inducing Factor-3 Activates Glycogen Synthase Kinase-3 (cid:1) and Degrades Cyclin D1 in Mammalian Cells*

In search of chemical substances applicable for the treatment of cancer and other proliferative disorders, we studied the signal transduction of Dictyostelium dif-ferentiation-inducing factors (DIFs) in mammalian cells mainly using HeLa cells. Although DIF-1 and DIF-3 both strongly inhibited cell proliferation by inducing G 0 /G 1 arrest, DIF-3 was more effective than DIF-1. DIF-3 suppressed cyclin D1 expression at both mRNA and protein levels, whereas the overexpression of cyclin D1 over-rode DIF-3-induced cell cycle arrest. The DIF-3-induced decrease in the amount of cyclin D1 protein preceded the reduction in the level of cyclin D1 mRNA. The decrease in cyclin D1 protein seemed to be caused by accelerated proteolysis, since it was abrogated by N -acetyl-Leu-Leu-norleucinal, a proteasome inhibitor. DIF-3-induced degradation of cyclin

Differentiation-inducing factors (DIFs) 1 were identified in Dictyostelium discoideum as the morphogens required for stalk cell differentiation of Dictyostelium (1). In the DIF family, DIF-1 (1-(3,5-dichloro-2,6-dihydroxy-4-methoxyphenyl)-1-hexanone) was the first to be identified, and DIF-3, the monochlorinated analogue of DIF-1, is a natural metabolite of DIF-1 in Dictyostelium (2). However, the actions of DIFs are not limited to Dictyostelium. They also have strong effects on mammalian cells. DIF-1 and/or DIF-3 strongly inhibit proliferation and induce differentiation in several leukemia cells, such as the murine erythroleukemia cell line B8, human leukemia cell line K562, and human myeloid leukemia cell line HL-60 (3,4). DIF-3 has been reported to have the most potent antiproliferative effect on mammalian leukemia cells among DIF analogues examined to date (5). Recently, we found that DIF-1 strongly inhibits proliferation and induces differentiation in human vascular smooth muscle cells, indicating that cells sensitive to DIFs are not limited to transformed cells (6).
However, the target molecule (receptor) of DIFs is unknown, and it is not clear even in Dictyostelium how DIFs induce an antiproliferative effect and cell differentiation. DIFs are small hydrophobic molecules and are therefore expected to be able to cross cell membranes without requiring channels or carriers. Also, the rapidity with which DIFs induce prestalk cell-specific gene expression suggests that they directly regulate gene expression. Therefore, the target molecule(s) for DIFs may be located in cytoplasm or nucleus (7). Although the precise mechanisms underlying their antiproliferative and differentiationinducing effects are not yet known, we found that DIF-1 induces cell cycle arrest at G 0 /G 1 phase by suppressing the expression of cyclin D1 (6). Cyclin D1 is synthesized early in G 1 phase and plays a key role in the initiation and progression of this phase. When cells enter the S phase, cyclin D1 is rapidly degraded by ubiquitin-proteasome-dependent proteolysis (8).
Therefore, in the present study, we investigated the mechanism underlying the DIF-induced inhibition of cyclin D1 expression. We particularly paid attention to the possible involvement of glycogen synthase kinase-3␤ (GSK-3␤), because this serine/threonine protein kinase has been shown to regulate not only cyclin D1 gene transcription by phosphorylating ␤-catenin but also cyclin D1 proteolysis by directly phosphorylating cyclin D1 itself (9 -12). Here we show for the first time that DIF-3, which is more effective than DIF-1 in inhibiting cell proliferation, induces the rapid degradation of cyclin D1 and the activation of GSK-3␤.
Cell Culture and Transfection-HeLa cells and bovine aortic endothelial cells (BAECs) were grown in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum, 100 units/ml penicillin G, and 0.1 g/ml streptomycin. The cells were plated on plastic tissue culture dishes or coverslips. Human umbilical vein endothelial cells were plated on 0.1% gelatin-coated dishes and maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with 20% fetal bovine serum, 5 ng/ml bovine fibroblast growth factor, 100 units/ml penicillin G, and 0.1 g/ml streptomycin. Wild-type human cyclin D1 cDNA was provided by Dr. K. Tamai (Medical and Biological Laboratories Co., Nagano, Japan) and subcloned into pcDNA3 (Invitrogen). Transfection was carried out using TransIT-LT1 (Mirus), and transfected cells were maintained in growth medium for 16 h before stimulation.
Cell Proliferation Assay-The cells were plated on 24-well plates (0.5 ϫ 10 4 cells/well) and treated with or without various amounts of DIF-1 or DIF-3 for given periods. Cells were harvested by the trypsin/ EDTA treatment and enumerated.
mRNA Expression Analysis-Total cellular RNA was extracted with Isogen (Nippon Gene). Using 1 g of the RNA, the expression of cyclin D1 mRNA was analyzed by reverse transcription-polymerase chain reaction (RT-PCR) using Ready-To-Go RT-PCR Beads (Amersham Biosciences) (6).
Purification of Nucleic and Cytoplasmic Proteins-Nucleic and cytoplasmic proteins were purified from cells cultured in 100-mm plates using NE-PER TM nuclear and cytoplasmic extraction reagents (Pierce). Five g of each sample was subjected to Western blot analysis.
Immunoblotting-Samples were separated by 12% SDS-PAGE and transferred to a polyvinylidene difluoride membrane using a semidry transfer system (1 h, 15 V). After blocking with 5% skim milk or 5% bovine serum albumin for 1 h, the membrane was probed with a first antibody. For the polyclonal anti-cyclin D1 antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), the polyclonal anti-cyclin D2 antibody (Santa Cruz Biotechnology), the polyclonal anti-phospho-GSK-3␤ (Ser 9 ) antibody (Cell Signaling Technology), the polyclonal anti-phospho-Akt (Ser 473 ) antibody (Cell Signaling Technology), and the polyclonal anti-phospho-p90 RSK (Ser 380 ) antibody (Cell Signaling Technology), the incubation was carried out overnight at 4°C. For the monoclonal anti-phosphotyrosine antibody (PY-20; Santa Cruz Biotechnology), the monoclonal anti-GSK-3␤ antibody (BD Transduction Laboratories), and the monoclonal anti-cyclin D3 antibody (Santa Cruz Biotechnology), the incubation was carried out for 1 h at room temperature. The membrane was washed three times and incubated with horseradish peroxidase-conjugated anti-rabbit IgG or anti-mouse IgG (Bio-Rad) for 1 h. Immunoreactive proteins on the membrane were visualized by treatment with a detection reagent (LumiGLO, Cell Signaling Technology). An optical densitometric scan was performed using Science Lab 99 Image Gauge Software (Fuji Photo Film).
Immunoprecipitation-HeLa cells (1 ϫ 10 6 cells) were incubated with or without DIF-3 for the periods indicated. Cells were lysed on ice for 1 h in 1 ml of the lysis buffer (50 mM NaCl, 5 mM NaF, 2 mM Na 3 VO 4 , 10 mM Tris/HCl, pH 7.4, 1 mM EDTA, 1% (v/v) Triton X-100, and 2 mM phenylmethylsulfonyl fluoride), and insoluble cell debris was removed by centrifugation at 5,000 rpm for 3 min. The cell lysate was precleared with protein A-Sepharose CL-4B (Amersham Biosciences) and then incubated with an anti-GSK-3␤ antibody (1 g) and protein A-Sepharose CL-4B at 4°C for 3 h. After incubation, proteins bound to the antibody/protein A-Sepharose complex were precipitated by centrifugation at 15,000 rpm for 5 min and washed three times with the lysis buffer.
GSK-3␤ Activity Assay-For the in vitro kinase assay, immunoprecipitated samples were washed twice with lysis buffer and twice with a kinase assay buffer (20 mM Tris/HCl, pH 7.4, 5 mM MgCl 2 , and 1 mM dithiothreitol). Kinase activity was measured by mixing immunoprecipitated GSK-3␤ with 50 l of kinase assay buffer containing 20 mM Tris/HCl (pH 7.4), 5 mM MgCl 2 , 1 mM dithiothreitol, 250 M ATP, 5 Ci of [␥-32 P]ATP (Amersham Biosciences), and 10 M GSK-3␤ substrate peptide (Upstate Biotechnology, Inc., Lake Placid, NY). The samples were incubated at 30°C for 30 min, and the reaction was terminated by adding 10 l of 50% trichloroacetic acid. Samples were then centrifuged at 15,000 rpm for 5 min, and 40 l of supernatant was spotted on P81 phosphocellulose filter paper (Whatman). The filters were washed five times with 0.75% phosphoric acid and twice with acetone and analyzed by scintillation counting.
Fluorescence Microscopy-Cells plated on CELL-TAK R (Collaborative Biomedical Products)-treated coverslips were incubated with or without DIF-3 (30 M) for given periods and washed with phosphatebuffered saline. The cells were fixed and permeabilized in ice-cold methanol/acetone (1:1) for 15 min at Ϫ20°C and then washed twice with phosphate-buffered saline. After blocking with 2% bovine serum albumin in phosphate-buffered saline for 30 min, the coverslips were incubated with polyclonal or monoclonal anti-GSK-3␤ antibody (1:200) overnight at 4°C. The coverslips were washed twice with phosphatebuffered saline and incubated with anti-mouse IgG or anti-rabbit IgG ϩ IgA ϩ IgM-biotin (Nichirei) for 1 h at room temperature followed by streptavidin-fluorescein isothiocyanate conjugate (1:50 dilution; Invitrogen) for 1 h at room temperature. The cells were examined under a fluorescence microscope (Olympus).

RESULTS
DIFs Inhibited HeLa Cell Proliferation-DIFs exhibit powerful antiproliferative effect in leukemia cells (3)(4)(5). In the present study, we first examined whether DIFs also inhibit the proliferation of HeLa cells. As shown in Fig. 1, A and B, DIF-1 and DIF-3 both strongly inhibited HeLa cell proliferation in a dose-dependent fashion, suggesting that DIFs are also effective in solid tumors. These antiproliferative effects were unlikely to be caused by cytotoxicity, because the number of dead cells indicated by the trypan blue exclusion test was not increased by the treatment with DIFs (data not shown). Consistent with the result obtained in leukemia cells (5), DIF-3 was more effective than DIF-1 in HeLa cells. Therefore, we used DIF-3 in the subsequent experiments. We next examined the cell cycle distribution using flow cytometry. Although the cell populations in S and G 2 /M phases decreased after the treatment with DIF-3, the population in G 0 /G 1 phase significantly increased ( Fig. 1C), indicating that DIF-3 induced G 0 /G 1 arrest in HeLa cells. This result was consistent with our previous study as to the effect of DIF-1 on vascular smooth muscle cell cycle (6).
DIF-3 Suppressed the Expression of Cyclins D1, D2, and D3 in HeLa Cells-We reported that DIF-1 induced G 0 /G 1 arrest, suppressing the expression of cyclins D1, D2, and D3 in human vascular smooth muscle cells (6). In HeLa cells, DIF-3 also reduced both the mRNA and protein levels of cyclin D1 (Fig. 2,  A and B). RT-PCR analyses showed that the cyclin D1 mRNA level was not significantly affected after a 1-h incubation with DIF-3 ( Fig. 2A). Although it began to slowly decrease from 3 h, a considerable amount of cyclin D1 mRNA was still expressed until 6 h. Despite this slow decrease in the level of mRNA, DIF-3 rapidly reduced the protein level of cyclin D1 (Fig. 2B). Cyclin D1 protein markedly decreased after 1 h of incubation with DIF-3 and nearly completely disappeared by 3 h. This rapid decrease in the amount of protein was not explained by suppression of mRNA expression. The protein levels of cyclins D2 and D3 also significantly decreased after 1 h of treatment with DIF-3 and almost disappeared after 6 h of treatment (Fig. 2C).

DIF-3-induced G 0 /G 1 Arrest Was
Rescued by the Overexpression of Cyclin D1-To clarify whether the overexpression of cyclin D1 is able to rescue cells from cell cycle arrest induced by DIF-3, wild-type human cyclin D1 cDNA was transfected to BAECs, since transfection efficiency was the highest in BAECs among mammalian cell species we examined including HeLa cells. As shown in Fig. 3A, the expression levels of cyclin D1 in untransfected cells and in cells transfected with empty pcDNA3 were reduced after 24 h of incubation with DIF-3, demonstrating that DIF-3 showed the same effect in BAECs as in HeLa cells. However, the expression level of cyclin D1 in cells transfected with pcDNA3/cyclin D1 was not significantly changed by DIF-3 treatment. We then examined the cell cycle distribution using untransfected and transfected cells. Although DIF-3 induced G 0 /G 1 arrest in untransfected and pcDNA3-transfected cells, there was no significant difference in pcDNA3/cyclin D1-transfected cells between the absence and presence of DIF-3 treatment (Fig. 3B). Therefore, DIF-3 was likely to induce cell cycle arrest by reducing the expression level of cyclin D1.

FIG. 6. DIF-3 activates GSK-3␤.
HeLa cells were incubated with or without DIF-3 (30 M) for the periods indicated. A, in vitro kinase assay. GSK-3␤ was immunoprecipitated from cell lysates and measured for kinase activity using a substrate peptide derived from the sequence of glycogen synthase. The results are means Ϯ S.E. of three independent experiments performed in duplicate. *, p Ͻ 0.01 compared with the control at time 0 (Student's t test). B, the effect of lithium chloride on GSK-3␤ activity. GSK-3␤ immunoprecipitated from cell lysates was measured for kinase activity using a substrate peptide derived from the sequence of glycogen synthase in the presence or absence of 10 mM lithium chloride. The results are means Ϯ S.E. of three independent experiments. C, phosphorylation of Ser 9 on GSK-3␤. Cell lysates were subjected to immunoblot analysis using an anti-phospho-GSK-3␤ (Ser 9 ) antibody. The levels of Ser 9 phosphorylation on GSK-3␤ were quantified by densitometry and shown as percentages of the levels in the control cells. Values are means Ϯ S.E. of three independent experiments. *, p Ͻ 0.01 compared with the control (Student's t test). D, tyrosine phosphorylation of GSK-3␤. Immunoprecipitated GSK-3␤ was subjected to immunoblot analysis using an anti-phosphotyrosine antibody (PY-20). The membrane was reprobed with an anti-GSK-3␤ antibody. The levels of tyrosine-phosphorylated GSK-3␤ were quantified and shown as percentages of the levels in the control cells. Values are means Ϯ S.E. of three independent experiments. *, p Ͻ 0.01 compared with the control (Student's t test).

DIF-3 Induced Cyclin D1
Proteolysis-To determine the mechanism for the decrease in cyclin D1 protein, we used N-acetyl-Leu-Leu-norleucinal, which inhibits ubiquitin-proteasome-dependent degradation of cyclins (14). As shown in Fig.  4A, N-acetyl-Leu-Leu-norleucinal prevented the DIF-3-induced loss of cyclin D1. We then performed a chase experiment to determine whether the turnover rate of cyclin D1 is changed by the treatment with DIF-3. As shown in Fig. 4B, DIF-3 significantly decreased the half-life of cyclin D1 in the presence of cycloheximide. These results indicated that DIF-3 rapidly reduced the amount of cyclin D1 protein by accelerating ubiquitin-proteasome-dependent proteolysis.
Lithium Inhibited DIF-3-induced Effects-It has been reported that cyclin D1 undergoes ubiquitin-proteasome-dependent proteolysis upon phosphorylation by GSK-3␤ (12). To investigate whether GSK-3␤ is involved in the DIF-3-induced degradation of cyclin D1 and inhibition of cell proliferation, we examined the effect of lithium chloride on these DIF-3-induced effects, because lithium ion is a specific inhibitor for GSK-3␤ (15). As shown in Fig. 5A, lithium chloride (20 mM) completely inhibited the degradation of cyclin D1 induced by DIF-3. We next examined the effect of lithium chloride on the inhibition of cell proliferation induced by DIF-3 (Fig. 5, B and C). Since 20 mM lithium chloride strongly inhibited cell proliferation on its own probably due to cytotoxicity, we reduced the concentrations of lithium chloride to 2-5 mM. Although 5 mM lithium chloride inhibited cell proliferation by about 30%, it partially prevented the effect of DIF-3 (Fig. 5B). Fig. 5C shows more clearly that lithium chloride dose-dependently inhibited the effect of DIF-3. Based on these results, we hypothesized that DIF-3 activates GSK-3␤ to induce proteolysis of cyclin D1 and antiproliferative effect.

DIF-3 Activated GSK-3␤-
The activity of GSK-3␤ was measured using an in vitro kinase assay. DIF-3 (30 M) elevated GSK-3␤ activity by 1.9-fold after 30 min incubation, and GSK-3␤ was still activated at 3 h (Fig. 6A). In the presence of lithium chloride, however, DIF-3 was not able to activate GSK-3␤ (Fig. 6B). Since GSK-3␤ is activated by the dephosphorylation of Ser 9 (9 -11), the level of GSK-3␤ Ser 9 phosphorylation was examined using an anti-phospho-GSK-3␤ (Ser 9 ) antibody. As shown in Fig. 6C, DIF-3 dramatically reduced the phosphorylation level of Ser 9 on GSK-3␤ after the incubation with DIF-3 for 30 min, and the phosphorylation level of Ser 9 was slowly recovered, looking like a mirror image of the time course of GSK-3␤ activity (Fig. 6A). Further, we examined the tyrosine phosphorylation level of GSK-3␤ , since the activity of this enzyme has been reported to be increased by phosphorylation of Tyr 216 (9 -11). As shown in Fig. 6D, DIF-3 significantly elevated the tyrosine phosphorylation level of GSK-3␤ by 2.0fold after the incubation with DIF-3 for 30 min. These results strongly indicated that DIF-3 activates GSK-3␤.
DIF-3 Induced Nuclear Translocation of GSK-3␤-GSK-3␤ is a cytosolic protein; however, it is translocated into the nucleus when activated (9 -11). GSK-3␤ thereby accumulated in the nucleus phosphorylates cyclin D1 and excludes it from nucleus, resulting in its degradation in the cytoplasm (12). To test whether DIF-3-activated GSK-3␤ is able to target cyclin D1, we examined the subcellular distribution of GSK-3␤ after stimulation with DIF-3. Immunofluorescent staining for GSK-3␤ revealed that GSK-3␤ was most present in the cytoplasm and that there was only a small amount in nuclei in unstimulated cells; however, it was markedly translocated into nuclei after stimulation with DIF-3 (Fig. 7A). Importantly, the time course of GSK-3␤ translocation into nuclei was similar to that of cyclin D1 degradation. As shown in Fig. 7B, a similar result was obtained using another anti-GSK-3␤ antibody. The nuclear translocation of GSK-3␤ was confirmed by immunoblotting of subcellular protein fractions. As shown in Fig. 7C, the treatment with DIF-3 markedly increased GSK-3␤ in the nuclear fraction. Its reduction in the cytosolic fraction may not be very clear, because we applied 5 g of each protein sample, although the amounts of cytosolic proteins extracted were ϳ5 times those of nuclear proteins. These results clearly demonstrated that cytosolic GSK-3␤ was translocated into the nucleus in response to DIF-3.

The Phosphatidylinositol 3-Kinase (PI3K)/Akt Pathway and the Mitogen-activated Protein Kinase (MAPK) Cascade Were Not Involved in DIF-3-induced Cyclin D1 Degradation-
GSK-3␤ has been reported to be phosphorylated on Ser 9 by several protein kinases (e.g. Akt, which is activated by PI3K, and p90 ribosomal S6 kinase (p90 RSK ), which is activated by the MAPK cascade) (16). We therefore examined the effect of DIF-3 on the phosphorylation levels of Akt and p90 RSK to determine whether DIF-3 inhibits these kinases. Although DIF-3 did not affect the phosphorylation level of Akt, it significantly induced phosphorylation on Ser 380 of p90 RSK (Fig. 8A). We also examined the effects of a PI3K inhibitor wortmannin and a mitogen-activated protein kinase/extracellular signalregulated kinase kinase inhibitor U0126 on cyclin D1 degradation induced by DIF-3. However, as shown in Fig. 8B, neither prevented the effect of DIF-3 on cyclin D1. Therefore, the PI3K/Akt pathway and the MAPK cascade did not seem to be involved in DIF-3-induced cyclin D1 degradation.
The Effect of DIF-3 on Untransformed Mammalian Cells-In human umbilical vein endothelial cells as well, DIF-3 inhibited cell proliferation (Fig. 9A), induced G 0 /G 1 arrest (Fig. 9B), and induced the degradation of cyclin D1 and the activation of GSK-3␤ (Fig. 9C), indicating that these effects of DIF-3 are not limited to transformed cells but common between transformed and untransformed cells. DISCUSSION In the present study, we showed that DIF-3 accelerated the degradation of cyclin D1 by activating GSK-3␤ mainly using HeLa cells. However, this effect of DIF-3 was also found in other cell species including transformed and normal (untransformed) cells. As an example, we have demonstrated the results obtained in vascular endothelial cells. Therefore, DIF-3 may ubiquitously activate GSK-3␤ in the wide variety of cell types. Since HeLa cells express human papilloma virus antigens E6 and E7, which inactivate retinoblastoma protein (8), it is not clear whether cyclin D1 is required for their cell cycle progres- sion. However, lithium chloride prevented both the degradation of cyclin D1 and the inhibition of cell proliferation induced by DIF-3. Therefore, cyclin D1 may be required for proliferation, and the disappearance of cyclin D1 may be responsible for the anti-proliferative effect of DIF-3 in HeLa cells. Furthermore, overexpression of cyclin D1 prevented DIF-3-induced cell cycle arrest in untransformed endothelial cells. Taken together, cyclin D1 degradation seemed to be a main mechanism for DIF-3 to induce cell cycle arrest.
DIF-3, naturally generated from DIF-1 as its first metabolite, is much weaker than DIF-1 in the ability to induce stalk cell differentiation in Dictyostelium (2). However, in contrast, the antiproliferative effect of DIF-3 was significantly stronger than that of DIF-1 in HeLa cells, consistent with a previous report that DIF-3 is more effective than DIF-1 at inhibiting proliferation in K562 human leukemia cells (5). This species difference in the sensitivity to DIFs may be caused by a difference in the nature of the target molecule (such as the affinity for DIFs) between mammalian and Dictyostelium cells.
We found that DIF-3 not only elevated the activity of GSK-3␤ but also induced nuclear translocation of the kinase. GSK-3␤ was initially considered to be a soluble protein expressed in cytoplasm; however, GSK-3␤ would have no access to nuclear proteins such as cyclin D1 if it resided in cytoplasm. Recent evidence has indicated that GSK-3␤ in nucleus phosphorylates nuclear proteins, such as cyclin D1 (12), nuclear factor of activated T-cells (17), heat shock factor-1 (18), and cAMP-response element-binding protein (19). Indeed, GSK-3␤ has been identified from the nuclei of cell cycle-arrested NIH-3T3 cells (12), cardiomyocytes stimulated with endothelin-1 (20), and heat shock-and staurosporine-treated SH-SY5Y human neuroblastoma cells (21). Therefore, DIF-3 seems to make it possible for GSK-3␤ to phosphorylate cyclin D1 by translocating GSK-3␤ into the nucleus.
Not only a rapid proteolysis but also a reduction in cyclin D1 mRNA expression was induced by DIF-3, although this latter effect took much longer time. Cyclin D1 gene expression is activated by ␤-catenin, the degradation of which is also initiated by GSK-3␤ (22). Therefore, activation of GSK-3␤ was expected to lead to a reduction in both protein and mRNA levels of cyclin D1 through independent pathways. Our results agreed well with this rationale. The proteolysis and mRNA reduction both seemed to be explained by DIF-3-induced GSK-3␤ activation.
The target molecule for DIFs is still unknown even in Dictyostelium. Since DIF-3 activates GSK-3␤, a target for DIF-3 may be a protein closely related to the regulation of GSK-3␤ activity. The activity of GSK-3␤ is decreased by phosphorylation of Ser 9 and increased by phosphorylation of Tyr 216 (9 -11). We found that DIF-3 decreased the phosphorylation level of Ser 9 and increased the phosphorylation level of Tyr 216 on GSK-3␤. Akt activated by PI3K and p90 RSK activated by the MAPK cascade are the candidate molecules to modulate GSK-3␤ activity by Ser 9 phosphorylation (16). Although DIF-3 did not affect the Ser 473 phosphorylation of Akt, it strongly induced phosphorylation on Ser 380 of p90 RSK . However, we were not able to elucidate the role of DIF-3-induced p90 RSK activation, since this kinase did not seem to be involved in cyclin D1 degradation induced by DIF-3 (Fig. 6B). DIF-3 also enhanced tyrosine phosphorylation on GSK-3␤. Recently, a novel nonreceptor tyrosine kinase, ZAK-1, has been found to directly activate GSK-3 in Dictyostelium (23). However, a ZAK-1 counterpart in mammals has not yet been discovered. A target molecule for DIF-3 might be a novel protein kinase or phosphatase controlling GSK-3␤ activity in mammalian cells. Recently, it has been reported that not only phosphorylation but also distribution regulates GSK-3␤ activity (21). Thus, DIF-3-induced accumulation of GSK-3␤ in nuclei might have an impor-tant role to regulate GSK-3␤ activity. In addition, if mammalian cells have a target molecule for DIFs as Dictyostelium, one could hypothesize that mammals also produce DIF-like substances to control growth and differentiation.
The Wnt signaling pathway is essential for embryonic development, cell proliferation, cell differentiation, microtubule dynamics, and cell motility. Several mutations have been identified in the components of the Wnt pathway in a variety of malignant tumors. For instance, most human colon cancers have mutations in the APC gene that result in the accumulation of ␤-catenin (24). Mutations in ␤-catenin, which amplify the accumulation of ␤-catenin itself, also have been identified in colon cancers (25), malignant melanomas (26), prostate cancers (27), and hepatocellular carcinomas (28). ␤-Catenin accumulated to an abnormal level would produce an excessive amount of cyclin D1 mRNA and promote tumor growth (29). Therefore, chemicals that activate GSK-3␤, such as DIFs, may be useful for the treatment of several cancers.
Recently, it has been reported that proapoptotic stimuli, such as heat shock and staurosporine, activate GSK-3␤ and induce its accumulation in nucleus (21). However, cytotoxic or proapoptotic agents damage not only cancer cells but also normal cells, which would cause severe adverse drug events. Distinct from staurosporine and other proapoptotic agents, DIF-3 did not induce the activation of caspase-3 (data not shown). Therefore, the antiproliferative effect of DIF-3 did not seem to be caused by the induction of apoptosis. DIF-3 is a unique compound that activates GSK-3␤ but does not induce apoptotic cell death.