Introduction

Top Abstract Introduction Methods Results Discussion References

An improved understanding of oncogenesis and the roles that oncogenes and tumor suppressors play in the regulation of cell proliferation has led to more rational approaches for the design and development of neoplastic-specific, target-directed, anticancer drugs. Signal transduction pathways in general, and growth factor-mediated signaling in particular, have become prime targets for novel antiproliferative agents. It is generally assumed that agents aimed to correct aberrant signaling will have a distinct advantage over traditional anticancer therapies by selectively affecting growth of tumor cells over normal tissues.

Growth factors and cytokines play a pivotal role in regulating cell proliferation, cell cycle progression and cell survival. Many tumors overproduce growth factors, and autocrine stimulation appears to be a major factor in the establishment and maintenance of the malignant phenotype. In addition, growth factors are key regulators of the cell cycle, and the mechanisms that link extracellular signals to transcriptional activation and cell cycle regulation are now being uncovered (Hill and Treisman, 1995).

The most prevalent mechanism used by cells to regulate growth factor signal transduction is reversible protein phosphorylation and dephosphorylation by kinases and phosphatases (Sun and Tonks, 1994). Both classes of enzymes are currently being explored as potential anticancer targets (Wipf et al., 1997; Mohammadi et al., 1997; Dudley et al., 1995; Baratte et al., 1992; Chen et al., 1996a). Special emphasis has been placed on the development of nonelectrophilic, cell active, small molecules that inhibit signal transduction, as these agents should have several desirable attributes such as stability, potential oral availability, diffusibility, nonimmunogenicity lacking in large molecules such as antibodies or peptides.

Using parallel chemistry, we have recently synthesized on solid support a library of small molecule, nonelectrophilic oxazoles that were modeled after antiphosphatase natural products, such as okadaic acid and calyculin A. Several members of the natural product-based library inhibited the growth of MDA-MB-231 breast cancer cells in culture, and one member, SC-9 (table 1), accumulated cells in the G0/G1 phase of the cell cycle (Wipf et al., 1997). In vitro, SC-9 effectively inhibited both the protein tyrosine phosphatase PTP1B and Cdc25 dual specificity phosphatases, but did not affect serine/threonine phosphatases, the DSPase CL100 or alkaline phosphatase (Rice et al., 1997). We have, however, no information concerning intracellular actions of this novel compound or the biochemical basis of its antiproliferative activity.

Any potential anticancer agent should display selectivity for the malignant phenotype. Thus, we have used cellular transformation by SV40 as a relevant model to investigate potential intracellular mechanisms of SC-9 growth inhibition and its selectivity toward a malignant phenotype. SV40 is a complete transforming agent (Ray et al., 1990; Ray and Kraemer, 1993), unlike other oncogenes (e.g., ras) that, in primary fibroblasts, induce a senescence-like phenotype and require other genetic alterations to achieve their full transforming potential (Weinberg, 1997; Serrano et al., 1997). SV40 transformation has been reported to elevate levels or activities of key mitotic regulators, such as cyclin A, cyclin B and Cdc2 (Chang et al., 1997), and to increase expression of Cdc25B (Nagata et al., 1991) in human diploid fibroblasts. SV40 directly transactivates Cdc2 promoter/reporter constructs and increases levels of Cdc2 mRNA in monkey kidney cells (Chen et al., 1996b). In addition to its effects on cell cycle regulators, SV40 enhances promoter activity of the IGF-1 gene (Porcu et al., 1994). Cells transfected with SV40 have high levels of IGF-1 mRNA, and secrete IGF-1 into the growth medium (Baserga, 1993). This reduces growth factor requirements and is likely to be important for SV40 transforming ability (Baserga, 1993). IGF-1 is also required for Cdc2 expression (Surmacz et al., 1992). Thus, SV40 transformation results in a model system that owes its transformed phenotype to both autocrine stimulation and overexpression of cell cycle regulators. Using this model system, we found that SC-9 is selectively toxic to the virally transformed cells and relate the growth inhibition to disrupted IGF-1 signaling pathways and decreased Cdc2 levels. The SC-9 pharmacophore may thus be useful in the development of agents in the treatment of tumors whose growth or survival depends on autocrine stimulation.

	Methods

Top Abstract Introduction Methods Results Discussion References

Chemical compounds. The general synthesis of compounds SC-9, SC-09, SC-109 has been previously described (Wipf et al., 1997; Rice et al., 1997). A slightly modified strategy was used for the synthesis of the new compounds SC-6III and SC-4II. Briefly, rac-glutamate was selectively side-chain esterified with trimethylsilylchloride in allyl alcohol, N-protected with ,,-trichloroethoxycarbonyl chloride, and methylated to give fully protected N-Troc-allyl-methylglutamate. The diester was de-allylated, coupled with protected ethylene diamine and reacted with 2,5-diphenyl-oxazole-4-carboxylic acid in the presence of PyBroP as described (Frerot et al., 1991). After attachment of the oxazole moiety, the Troc group was removed with zinc in acetic acid, and a carbodiimide-mediated coupling provided substrates SC-6III and SC-4II. All intermediates and products were purified by chromatography on SiO₂ and characterized by nuclear magnetic resonance and high resolution mass spectrometry as described previously (Wipf et al., 1997; Rice et al., 1997).

Determination of cLog P values. Computation of cLog P, the calculated logarithm of the octanol-water partition coefficient, was performed on an Indigo2 R4400 workstation according to the protocol by Villar (Alkorta and Villar, 1992). Extended conformations of compounds were fully optimized using the semiempirical method PM3. Charges and other parameters for the regression analysis were also obtained with the PM3 module on Spartan 5.0 (Wavefunction, Inc., Irvine, CA).

Cell culture. MEF were isolated from fetuses of 14.5-day pregnant mice (129 Ola × C57Bl/6) using previously described methods (Kondo et al., 1995). Cells were maintained in Dulbecco's minimum essential medium containing 20% FBS (HyClone, Logan, UT), and 1% penicillin-streptomycin (GIBCO BRL) in a humidified atmosphere of 5% CO at 37°C. Primary cultures of MEF were never extended beyond passage 15 to avoid entering of crisis. MEF cells were transformed using the plasmid pCC5 (a kind gift from Dr. Stephen Strom, University of Pittsburgh) expressing SV40 large T antigen under the control of its own promoter by means of a cationic lipid (Lipofectamine, GIBCO BRL) according to manufacturer's instructions. SV40 transformed MEF were grown in DMEM supplemented with 5% FBS and 1% penicillin-streptomycin (GIBCO BRL) to attempt to control for similar growth and plating efficiencies compared with MEF. As indicated, in some experiments both cell lines were grown and treated in DMEM containing 10% FBS and 1% penicillin-streptomycin.

Assay for antiproliferative activity. We used our previously described MTT microtiter assay (Kondo et al., 1995) to determine the antiproliferative activity of the newly synthesized compounds. Cells were plated at 2000 cells/well in 96-well plates. After a 24-hr incubation at 37°C, cells were exposed continuously for 48 hr to each compound, incubated with MTT for 3 hr and total cell number determined by colorimetric quantitation of the blue formazane dye at 540 nm in DMSO as previously described (Kondo et al., 1995).

Colony formation. Long-term survival of MEF and SV40 MEF was determined in a clonogenic assay essentially as described (Freshney, 1994). Briefly, cells were plated (200 cells/well) in six-well plates and treated the next day with vehicle (DMSO) or inhibitors without media change to not disturb the cell attachment process. After 10 to 12 days in culture with continuous exposure to drug, colonies were exposed to staining solution containing 0.25% crystal violet and 10% formalin (35% v/v) in 80% methanol for 30 min, washed with water and counted. Plating efficiency was determined as the fraction of cells that attached to the support and grew into colonies larger than 1 mm in diameter.

Western blotting. Cells were grown to subconfluency in 100-mm dishes, harvested, and lysed in lysis buffer (30 mM HEPES, pH 7.5, 1% Triton-X 100, 10% glycerol, 5 mM MgCl₂, 25 mM NaF, 1 mM EGTA, 10 mM NaCl, 2 mM Na₃VO₄, 10 µg/ml trypsin inhibitor, 10 µg/ml aprotinin, 25 µg/ml leupeptin, 2 mM PMSF, 6.4 mg/ml Sigma104 phosphatase substrate). Lysates were electrophoresed on 4-20% gradient gels (NOVEX, San Diego, CA), transferred to nitrocellulose and immunoblotted with antibodies against SV40 (Ab-2, Oncogene Science, Manhasset, NY), Cdc25A (144), Cdc25B (C-20), Cdc25C (C-20), Cdc2 (17), or IGF-1 receptor beta subunit (C-20, all from Santa Cruz Biotechnology, Santa Cruz, CA), or antiphosphotyrosine (PY20, Transduction Laboratories, Lexington, KY). Hyperphosphorylated and hypophosphorylated Cdc2 were separated on a large 10% polyacrylamide gel. For determination of Erk activation, lysates were separated on 15% SDS-PAGE and immunoblotted with anti-Erk1 (K23, Santa Cruz Biotechnology, recognizes Erk1 and Erk2) or Erk2 (Upstate Biotechnology, Lake Placid, NY) antibodies. Positive antibody reactions were visualized using peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA) and an enhanced chemiluminescence detection system (Renaissance, NEN, Boston, MA) according to manufacturer's instructions. Equal loading was ensured by reblotting with an anti-actin antibody (H-196, Santa Cruz Biotechnology). For quantitation of protein expression levels, X-ray films were scanned on a Molecular Dynamics personal SI densitometer and analyzed using the ImageQuant software package (Ver. 4.1, Molecular Dynamics, Sunnyvale, CA).

In vitro phosphatase assays. Phosphatase-active GST-Cdc25B₂ was expressed and isolated from a plasmid (pGEX2T-KG) encoding GST-Cdc25B₂ fusion protein in Escherichia coli strain BL21 (DE3) as described (Rice et al., 1997). The activity of GST-Cdc25B₂ was measured in our recently described fluorescence-based microtiter plate assay (Rice et al., 1997) except that 3-O-methylfluorescein monophosphate (Sigma Chemical Co., St Louis, MO) was used as a substrate. Briefly, 100 to 250 ng of enzyme were incubated with the substrate for 5 min at room temperature in 150 µl assay buffer containing 30 mM Tris (pH 8.5), 50 mM NaCl, 1.5 mM EDTA, 0.033% bovine serum albumin, and 1 mM DTT. Inhibition studies were carried out at concentrations of 3-O-methylfluorescein that represented apparent K_m values (i.e., 40 µM). Inhibitors were dissolved in DMSO and added to the reaction mixture before the addition of enzyme. All reactions including controls were performed at a final concentration of 7% DMSO.

	Results

Top Abstract Introduction Methods Results Discussion References

Antiproliferative and antiphosphatase activity of a selected member of a targeted array library. We previously described the generation of a library of compounds modeled after calyculin A, microcystin LR, and okadaic acid, natural product inhibitors of PSTPases (Wipf et al., 1997). Surprisingly, in vitro inhibition studies for antiphosphatase activity revealed a number of compounds that were potent inhibitors of PTP1B (i.e., >50% inhibition at 3 µM), but had little effect on the PSTPases PP1 and PP2A or the closely related DSPase CL100 at concentrations as high as 100 µM (Rice et al., 1997). SC-9, one of the most potent inhibitors of PTP1B, also inhibited Cdc25 A, B, and C (K_i ~ 10 µM) (Rice et al., 1997) and was cytotoxic to MDA-MB-231 breast cancer cells in culture (Wipf et al., 1997). We have now expanded the targeted array library by other structural analogs and chosen five closely related congeners to assess structural requirements for growth inhibitory activity. Table 1 shows the structures of the library members, which contain substituents of varying hydrophobicity and steric bulk in the R₂, R₃ and R₄ positions. Compounds SC-9, SC-09 and SC-109 were synthesized as discrete compounds by traditional solution-phase chemistry as described (Wipf et al., 1997; Rice et al., 1997). The novel oligoether compounds SC-6III and SC-4II were synthesized by a slightly modified procedure as described in the Methods Section.

SV40 transformation resulted in elevated levels of Cdc25B and Cdc2 and increased tyrosine phosphorylation of the IGF-1 receptor. To generate a relevant model system for the analysis of the activity and selectivity of SC-9 against the malignant phenotype, we transfected primary MEF with SV40. As expected, these cells grew in soft agar, exhibited reduced serum-dependence and achieved higher saturation densities on plastic surfaces than the parental cells (data not shown). The presence of SV40 was confirmed by Western blotting (fig. 1A). In accordance with previously published data (Chang et al., 1997; Nagata et al., 1991), SV40 MEF showed a dramatic increase in Cdc25B and Cdc2 protein levels after SV40 transformation (fig. 1B and D), whereas levels of Cdc25A were unchanged (fig. 1C). The appearance of three higher molecular bands in the Cdc25 A immunoblot has been observed previously (Galaktionov et al., 1995), but their identity is unclear. We were unable to detect Cdc25C by Western blotting with a commercially available antibody (data not shown). Increased signaling through the IGF-1 receptor was assessed by sequential immunoblotting with antibodies to phosphotyrosine and the IGF-1 receptor. Figure 1E shows that the antiphosphotyrosine antibody detected a major band of approximately 100 kDa. Reblotting with an anti-IGF-1 receptor antibody confirmed the identity of the 100 kDa band and also demonstrated equivalency in protein loading (fig. 1F). Thus, the transformed cells exhibited higher levels of IGF-1 receptor tyrosine phosphorylation presumably reflecting autocrine stimulation by IGF-1 (Baserga, 1993), whereas levels of the IGF-1 receptor itself were similar in both cell lines.

View larger version (32K): [in this window] [in a new window]   Fig. 1.   SV40 large T antigen transformation results in increased levels of mitotic regulators and enhanced IGF-1 receptor autophosphorylation. Cell lysates were isolated from exponentially growing MEF and SV40 MEF, separated by SDS-PAGE on 4 to 20% gradient gels, and immunoblotted with antibodies against A, SV40; B, Cdc25B; C, Cdc25A; D, p34Cdc2; E, phosphotyrosine (PY20) and F, IGF-1 receptor as described in "Methods." Numbers indicate the positions of molecular weight markers in kDa. Data are representative of three independent experiments.

Structural requirements for cytotoxicity and in vitro inhibition of Cdc25B. The five library congeners from table 1 containing structural variations in the R₂, R₃ and R₄ positions were assayed for their antiproliferative activity against SV40 MEF using the MTT assay as described in "Methods." Within compounds bearing a C₉ alkyl chain in the R₄ position, growth inhibitory activity decreased with decreasing bulk in the R₃ position (SC-09), and was abolished after replacement of the phenyl group in R₂ with a sterically less demanding methyl substituent (SC-109) (table 2). Replacement of the highly hydrophobic C₉ alkyl chain in SC-9 with a more polar but sterically similar oligoether group (SC-6III) also resulted in a loss of antiproliferative activity. Shortening of the oligoether residue in the R₄ position accentuated loss of activity (SC-4II). Taken together, these results indicated a requirement for hydrophobic substituents, especially in the R₂ and R₄ position, for antiproliferative activity. To address the question whether the observed biological activity of the five congeners was merely due to their overall hydrophobic character, we calculated log P values (cLogP) for all compounds based on energy-minimized extended conformations by the method of Villar (Alkorta and Villar, 1992). Because the compounds are present as carboxylates under cell culture and in vitro phosphatase assay conditions, cLogP values were calculated for the compounds in their free acid and carboxylate forms. As expected, cLogP values were much lower for the compounds in their ionized forms. Irrespective of the charge characteristics of the compounds, however, both sets of values indicated that there was no obvious correlation between overall hydrophobicity and inhibition of cell proliferation. For example, SC-09 was more toxic than SC-109, even though their cLog P values were similar. Furthermore, the oligoether compound SC-6III was about 1000 times more polar than SC-09, yet both compounds were comparable in their antiproliferative activities. Thus, even though SC-9, the most active compound, was also the most hydrophobic, overall hydrophobicity was not the sole determinant of biological activity. Furthermore, the decrement in antiproliferative activity did not readily correlate with loss of in vitro Cdc25B inhibition (table 2).

SC-9 decreased Cdc2 levels in SV40 transformed MEF. One of the putative substrates for both Cdc25B and Cdc25C is the cyclin dependent kinase Cdc2 (Sebastian et al., 1993). Thus, we treated SV40 MEF continuously for 48 hr with increasing concentrations of SC-9 and analyzed cellular lysates by immunoblot analysis with an anti-Cdc2 antibody as described in "Methods." Hyperphosphorylation of Cdc2 results in the appearance of a slower migrating band on SDS-PAGE (Draetta and Beach, 1988). We found that a 48 hr exposure to 30 or 60 µM SC-9 markedly decreased levels of Cdc2 (fig. 2, lanes 3 and 4). Reprobing of the Cdc2 immunoblot with an anti-actin antibody demonstrated approximately equal protein loading (fig. 2B). Densitometric scanning of both the upper (phosphorylated) and lower (unphosphorylated) bands of Cdc2 indicated that, at 10 and 30 µM, SC-9 did not alter the phosphorylation status of Cdc2. We hypothesized that the decrease in Cdc2 protein levels was a result of an intracellular action of SC-9. Thus, we investigated other events associated with a loss of Cdc2 that would account for SC-9's ability to inhibit cell growth and the previously observed G0/G1 phase accumulation.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   SC-9 decreases levels of Cdc2 in SV40 MEF. Cells were grown to 50% confluency in 100-mm dishes and treated for 48 h with vehicle or increasing concentrations of SC-9. A, Lysates were separated on 10% SDS-PAGE and immunoblotted with an anti-Cdc2 antibody. B, Actin was immunolabeled as a loading control. Lane 1, vehicle control; lanes 2 to 4, 10, 30 and 60 µM SC-9. P and U denote the phosphorylated and non-phosphorylated forms of Cdc2.

SC-9 decreased IGF-1 receptor tyrosine phosphorylation and receptor expression. One of the known regulators of Cdc2 is IGF-1 (Surmacz et al., 1992). Because it had previously been shown that SV40 transformation enhances levels of IGF-1 mRNA and increases IGF-1 secretion (Baserga, 1993), we investigated whether the decrease in Cdc2 levels caused by SC-9 correlated with a reduction in IGF-1 signaling in SV40 transformed cells. SV40 MEF were treated with various concentrations of SC-9 for 48 hr, lysates separated on SDS-PAGE and immunoblotted with antibodies against phosphotyrosine and the IGF-1 receptor. Figure 3A shows that SC-9 decreased phosphotyrosine levels on a 100-kDa protein (lanes 2-4). Reprobing of the same blot with an anti-IGF-1 receptor antibody revealed that the decrease in phosphotyrosine levels at the highest concentration of SC-9 was, in part, due to a decrease in IGF-1 receptor expression (fig. 3B). Thus, SC-9 interfered with IGF-1-mediated receptor signaling by reducing both receptor tyrosine phosphorylation and, unexpectedly, IGF-1 receptor levels. Concomitant with reduced receptor autophosphorylation, we also observed inactivation of MAPK (Erk2), a downstream target in the IGF-1 receptor signaling cascade. Figure 3C shows an anti-Erk2 immunoblot after separation of the identical lysates into the lower, unphosphorylated and the slower migrating, phosphorylated species as described in "Methods." SC-9 decreased Erk2 phosphorylation in a concentration-dependent manner (lanes 2-4). To determine whether MAPK inhibition alone was sufficient for inhibition of cell proliferation, we treated cells with PD-98059, a specific inhibitor of MEK, the direct upstream activating kinase of Erk (Dudley et al., 1995). PD-98059 (50 µM) completely inhibited MAPK phosphorylation (fig. 3C, lane 5) but in contrast to SC-9 did not affect IGF-1 receptor expression (fig. 3B, lane 5) or cell proliferation (data not shown) and had only a partial effect on receptor tyrosine phosphorylation (fig. 3A, lane 5). Densitometric analysis of the immunoblots in figures 2 and 3 indicated that IGF-1 receptor autophosphorylation, inactivation of Erk2 and the decrease in Cdc2 were all concentration-related, with an IC₅₀ of approximately 30 µM (figs. 3D and 6A), whereas loss of receptor was only observed at the highest concentration tested.

View larger version (36K): [in this window] [in a new window]   Fig. 3.   SC-9 disrupts IGF-1 signaling. SV40 MEF were treated for 48 hr with vehicle or inhibitors and lysed as described in "Methods." Lysates were separated on 4-20% SDS-PAGE and immunoblotted with A, antiphosphotyrosine (PY20); B, anti-IGF-1 receptor or C, anti-Erk2 antibodies. D, Blots were analyzed for -actin to ensure equal protein loading. Lane 1, vehicle control; lanes 2 to 4, 10, 30 and 60 µM SC-9; lane 5, 50 µM PD-98059. P and U denote the unphosphorylated and phosphorylated forms of Erk2. E, Graphical representation of the results from A to D and figure 2. Protein bands were quantitated by densitometric scanning and plotted as percent of vehicle control. MAPK phosphorylation was expressed as the percentage of the upper (phosphorylated) band on SDS-PAGE compared to total MAPK (upper and lower band), normalized to vehicle-treated control.

Less toxic library congeners and vanadate did not inhibit IGF-1 receptor signaling. We next examined the effects of the less toxic library congeners SC-09 and SC-109 on IGF-1 signaling. We found that SC-09 partially affected receptor tyrosine phosphorylation, and that SC-109 was inactive (fig. 4A, lanes 2 and 3). Neither SC-09 nor SC-109 markedly reduced IGF-1 receptor or Cdc2 levels (fig. 4B and E, lanes 2 and 3). At concentrations that were cytotoxic, vanadate markedly increased receptor tyrosine phosphorylation (fig. 4A, lane 4), suggesting a distinct mechanism of action from SC-9. Furthermore, SC-09 and SC-109 did not affect MAPK activation (fig. 4C and D, lanes 2 and 3). The apparent slight increase in MAPK phosphorylation and IGF-1 receptor autophosphorylation by SC-109 (fig. 4A, C and D, lane 3) was not reproducible. PD-98059 (50 µM) caused essentially complete inhibition of both Erk1 and Erk2, but only partially decreased receptor autophosphorylation and levels of Cdc2 (fig. 4A and E, lane 5). These data indicate that, although the decrease in Erk phosphorylation by SC-9 correlated with an inhibition of IGF-1 signaling, inactivation of MAPK alone was not sufficient to cause a complete loss of Cdc2 or inhibition of cell proliferation.

View larger version (48K): [in this window] [in a new window]   Fig. 4.   Less toxic congeners of SC-9 do not affect IGF-1 signaling. Lysates from vehicle or inhibitor-treated SV40 MEF were immunoblotted with A, antiphosphotyrosine (PY20); B, anti-IGF-1 receptor; C, anti-Erk2; D, Erk1 or E, p34Cdc2 antibodies. Lane 1, vehicle control; lane 2, 100 µM SC-09; lane 3, 100 µM SC-109; lane 4, 30 µM sodium vanadate; lane 5, 50 µM PD-98059.

SC-9, but not okadaic acid or vanadate, was selectively toxic to SV40 transformed MEF. Having demonstrated a profound effect on intracellular signaling by SC-9, we next examined the sensitivity of MEF and SV40 MEF to SC-9 and two classic antiphosphatases, vanadate and okadaic acid. Using a clonogenic assay, we found that SV40 MEF were two to three times more sensitive to SC-9 based on the concentration required for a 50% decrease in plating efficiency compared to the wild-type MEF (fig. 5). This selectivity was reproduced in an MTT assay (fig. 6A) where it was seen also when both cell lines were grown in 10% FBS (data not shown). We then treated both normal and SV40 transformed MEF with okadaic acid and vanadate, known inhibitors of PSTPases and PTPases, respectively (Mumby and Walter, 1993), and found that only SC-9 preferentially affected growth of the transformed cells (fig. 6A) whereas both okadaic and vanadate were equally effective in normal or transformed MEF (fig. 6B and C), even though okadaic acid was about 1000 times more potent than SC-9 or vanadate. These results indicated a transformation-specific intracellular action of SC-9 compared to okadaic acid or vanadate.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   SC-9 preferentially inhibits clonogenic growth of SV40 transformed cells. MEF () and SV40 MEF () were plated in triplicate in 6 well plates and treated with vehicle or inhibitor as described in "Methods." After 10 to 12 days of continuous exposure to drug, cells were stained with crystal violet and colonies >1 mm in diameter counted. Data are expressed as % of colonies compared to vehicle treated control and are representative of five independent experiments. Bars = S.E.M. Absolute plating efficiencies were 10 ± 3% for MEF and 14 ± 6% for SV40 MEF.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Cytotoxicity of SC-9, okadaic acid and vanadate on normal and SV40 transformed MEF. MEF () and SV40 MEF () were plated in 96-well plates and after 24 hr treated with various concentrations of A, SC-9; B, okadaic acid or C, vanadate. Cell survival was determined 48 hr later using the MTT assay as described in "Methods." Each value is the mean of three to five independent experiments performed in quadruplicate. Bars = S.E.M.

SC-9 decreased IGF-1 receptor tyrosine phosphorylation and levels of Cdc2 in both MEF and SV40 MEF. Finally, we examined whether SC-9 differentially affected IGF-1 signaling and levels of Cdc2 in MEF or SV40 MEF. To control for the known effects of serum on Cdc2 expression (Surmacz et al., 1992), both cell lines were plated and treated with SC-9 in the presence of 10% FBS. After 48 hr, lysates were separated on SDS-PAGE and immunoblotted with antibodies against phosphotyrosine, IGF-1 receptor, and Cdc2. Figure 7 shows that SV40 MEF had much higher levels of Cdc2 and displayed higher phosphotyrosine content on the IGF-1 receptor, consistent with the results from figure 1. Equal loading was ensured by immunoblotting with an anti-actin antibody (fig. 7D). As judged by cell numbers and morphology, SC-9 under these conditions was about 2- to 3-fold less toxic to normal MEF (data not shown), confirming the selective toxicity of SC-9 seen in figures 5 and 6. Interestingly, SC-9 caused downregulation of Cdc2 (fig. 7C) and inhibition of IGF-1 receptor signaling (fig. 7A and B) in both cell lines, although the absolute decrement in Cdc2 levels and receptor tyrosine phosphorylation was much larger in the transformed cells. These results indicate that decreased Cdc2 levels and inhibition of IGF-1 signaling are not secondary to inhibition of cell proliferation, but suggest that SC-9 specifically affects a cellular target associated with IGF-1 signaling.

View larger version (61K): [in this window] [in a new window]   Fig. 7.   Inhibition of IGF-1 signaling and Cdc2 expression in MEF and SV40 MEF. Cells were grown to 50% confluency in medium containing 10% FBS and treated with the indicated concentrations of SC-9 for 48 hr. Western blotting was performed with antibodies against A, antiphosphotyrosine (PY20); B, IGF-1 receptor; C, p34Cdc2 and D, -actin (loading control) as described in "Methods." Data are representative of two independent experiments.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Independence from growth factor control and loss of cell cycle regulation is a common feature of human malignancies. Many tumors overexpress growth factor receptors and depend on autocrine or paracrine stimulation (Baselga and Mendelsohn, 1994). Most growth factors use mechanisms involving protein phosphorylation and dephosphorylation to transduce extracellular signals to the nucleus. The roles of growth factors like platelet-derived growth factor, epidermal growth factor and IGF-1 in cell cycle progression from G0 to S-phase have been studied extensively (for a recent review see Winkles, 1998).

Mechanistic studies using SV40 large T antigen transformation have established a role of the IGF-1 receptor in the malignant phenotype. SV40 transformation causes an increase in IGF-1 mRNA expression and protein secretion (Baserga, 1993). Cells with a targeted disruption of the IGF-1 receptor gene are no longer transformed by SV40 (Sell et al., 1993). Because a number of human tumors have been shown to overexpress the IGF-1 receptor, aberrant signaling through the IGF-1 receptor has become a target for anticancer drug design (Baserga, 1996).

Growth factors also regulate cell cycle dependent kinases. Serum stimulation of quiescent cells results in the expression of Cdc2, a key component of the mitosis-promoting factor comprising p34Cdc2 and cyclin B. In cells expressing high numbers of IGF-1 receptors, Cdc2 can be induced by IGF-1 alone, and this induction can be suppressed by an antisense oligodeoxynucleotide against the IGF-1 receptor gene (Surmacz et al., 1992). Cdc2 is activated by dephosphorylation on threonine and tyrosine residues by the cell cycle phosphatases Cdc25B and Cdc25C (Sebastian et al., 1993), and Cdc2 as well as Cdc25B are overexpressed in SV40 transformed cells.

We have recently described the design and synthesis of a unique, small molecule, targeted array library that has antiphosphatase elements (Wipf et al., 1997). Several members of this library inhibited PTP1B and Cdc25, but not PSTPases at low micromolar concentrations in vitro. One compound, SC-9, selectively inhibited growth of transformed cells in culture. This selective toxicity was not shared by the phosphatase inhibitors vanadate and okadaic acid. Thus, SC-9 had a different activity profile than two classical antiphosphatase compounds, suggesting the antiproliferative effects may be associated with other pharmacological activities resident in this basic pharmacophore. Furthermore, although SC-9 was a potent inhibitor of Cdc25 in vitro, we saw no increased phosphorylation of Cdc2, a known substrate for Cdc25B and Cdc25C, in cultured cells, and growth inhibition was not strictly correlated with in vitro inhibition of Cdc25B. Thus, although we cannot formally exclude phosphatase inhibition as the cause of growth inhibition for SC-9, it seems more likely that these compounds exert their antiproliferative effects through another mechanism.

Structure-activity relationship studies using congeners with only slight structural modifications demonstrated that sterically demanding residues in the R₂ and R₃ positions or a hydrophobic alkyl chain in the R₄ position enhanced cytotoxicity. Log P values, however, calculated for the five compounds in an extended energy-minimized conformation showed that antiproliferative activity was not a result of overall hydrophobicity, even though the most biologically active compound (SC-9) was also the most hydrophobic. These results are consistent with SC-9 having a specific intracellular site of action.

SC-9 decreased Cdc2 levels and this was an interesting result because only few pharmacological agents, e.g., interferon- (Saunders and Jetten, 1994), retinoic acid (Zhu et al., 1997) and mezerein (Jiang et al., 1995) have been reported to reduce Cdc2 levels. Butyrolactone-I, a Cdc2 kinase inhibitor, decreased Cdc2 protein levels after prolonged exposure (Nishio et al., 1996). We have, however, seen no in vitro inhibition of Cdc2 kinase by SC-9 (Rice et al., 1997). The Cdc2 depletion is unlikely to be a result of cell cycle inhibition because Cdc2 levels, as with those of many other cyclin-dependent kinases, appear to remain stable throughout the cell cycle (Draetta and Beach, 1988; Morgan, 1995), and are down-regulated only in serum-deprived or senescent cells (Surmacz et al., 1992; Richter et al., 1991). Thus, the growth arrest caused by SC-9 is reminiscent of quiescence, senescence or differentiation.

Consistent with the previously reported ability of IGF-1 to induce Cdc2, the loss of Cdc2 correlated with an inhibition of IGF-1 receptor signaling and MAPK activation. Interestingly, however, selective disruption of IGF-1 signaling at the MAPK level by the MEK-specific inhibitor PD-98059 only led to a partial decline in Cdc2 levels and did not result in inhibition of cell proliferation. This most likely reflects direct transactivation of Cdc2 by SV40 which would be independent of MAPK activation. In contrast to PD-98059, SC-9, at concentrations that caused complete dephosphorylation of Erk, abolished Cdc2 expression. The precise mechanisms for the Cdc2 down-regulation are unknown, and could be due to decreased transcription, translation or increased protein degradation. Finally, we investigated possible mechanisms for SC-9 selectivity toward the transformed cells and found that the absolute reduction in Cdc2 levels caused by SC-9 was greater in SV40 MEF than in the parental cells. We propose that the increased sensitivity of the transformed cells to SC-9 is a result of their increased dependence on elevated Cdc2 expression and IGF-1 signaling. We suggest the SC-9 pharmacophore could prove useful in the further development of antiproliferative agents for the treatment of growth factor dependent tumors.