Identification of JTP‐70902, a p15^INK4b‐inductive compound, as a novel MEK1/2 inhibitor

Abstract

The INK4 family members p16^INK4a and p15^INK4b negatively regulate cell cycle progression by inhibition of cyclin‐dependent kinase (CDK) 4/6. Loss of p16^INK4a functional activity is frequently observed in tumor cells, and is thought to be one of the primary causes of carcinogenesis. In contrast, despite the biochemical similarity to p16^INK4a, the frequency of defects in p15^INK4b was found to be lower than in p16^INK4a, suggesting that p15^INK4b‐inductive agents may be useful for tumor suppression. Here we report the discovery of a novel pyrido‐pyrimidine derivative, JTP‐70902, which exhibits p15^INK4b‐inducing activity in p16^INK4a‐inactivated human colon cancer HT‐29 cells. JTP‐70902 also induced another CDK‐inhibitor, p27^KIP1, and downregulated the expression of c‐Myc and cyclin D1, resulting in G₁ cell cycle arrest. MEK1/2 was identified by compound‐immobilized affinity chromatography as the molecular target of JTP‐70902, and this was further confirmed by the inhibitory activity of JTP‐70902 against MEK1/2 in kinase assays. JTP‐70902 suppressed the growth of most colorectal and some other cancer cell lines in vitro, and showed antitumor activity in an HT‐29 xenograft model. However, JTP‐70902 did not inhibit the growth of COLO320 DM cells; in these, constitutive extracellular signal‐regulated kinase phosphorylation was not detected, and neither p15^INK4b nor p27^KIP1 induction was observed. Moreover, p15^INK4b‐deficient mouse embryonic fibroblasts were found to be more resistant to the growth‐inhibitory effect of JTP‐70902 than wild‐type mouse embryonic fibroblasts. These findings suggest that JTP‐70902 restores CDK inhibitor‐mediated cell cycle control by inhibiting MEK1/2 and exerts a potent antitumor effect. (Cancer Sci 2007; 98: 1809–1916)

Abbreviations: 

BSA

bovine serum albumin

CDK

cyclin‐dependent kinase

CKI

CDK‐inhibitor

DMEM

Dulbecco's modified Eagle's medium

DTT

dithiothreitol

EDTA

ethylene diamine tetra‐acetic acid

ELISA

enzyme‐linked immunosorbent assay

ERK

extracellular signal‐regulated kinase

INK4

inhibitor of CDK4

log GI₅₀

logarithm of the 50% growth inhibition value

MAPK

mitogen‐activated protein kinase

MBP

myelin basic protein

MEF

mouse embryonic fibroblasts

MEK

MAPK/ERK kinase

pRb

retinoblastoma protein

SDS‐PAGE

sodium dodecylsulfate–polyacrylamide gel electrophoresis

TGF

transforming growth factor

Disruption of cell cycle control, reported in numerous tumor cells, is thought to be one of the primary causes of malignant transformation. CDK are key regulators of the cell cycle, and their activities are tightly controlled in a phase‐specific manner during cell cycle progression.^{( 1 )} Cyclin D–CDK4/6 and cyclin E–CDK2 play important roles in the G₁ to S phase transition of the cell cycle by phosphorylating the tumor suppressor pRb, which forms a complex with E2F.^{( 1} , ^{2 )} The phosphorylation of pRb by CDK causes release and activation of E2F, allowing transcription of genes necessary for the progression to S phase.^{( 3 )}

The loss of G₁/S transition control in tumor cells often arises from aberrant activation of cell‐proliferative signaling pathways or inactivation of cell cycle‐regulating proteins. Of these pathways, the Ras–MAPK pathway has been investigated extensively. The classical MAPK cascade, comprising c‐Raf or B‐Raf, MEK1/2, and ERK1/2, is constitutively activated in various malignant tumors, often through gain‐of‐function mutations of Ras and Raf family members.^{( 4 )} These activating mutations cause increased expression of cyclin D, leading to CDK activation and aberrant cell proliferation.^{( 5} , ^{6 )}

The cell cycle progression through cyclin–CDK activation is negatively regulated by CKI. Two families of CKI have been identified in mammalian cells: one includes p21^WAF1, p27^KIP1 and p57^KIP2, each of which inhibit cyclin E/A–CDK2 complexes, and the other family, referred to as the INK4 family, is composed of p16^INK4a, p15^INK4b, p18^INK4c and p19^INK4d, which inhibit cyclin D–CDK4/6. p16^INK4a is known to function as a tumor suppressor. Mutation or inactivation of p16^INK4a has been detected in many human malignant tumors,^{( 7 )} and mice deficient in p16^INK4a are highly susceptible to malignant tumors.^{( 8} , ^{9 )} In contrast, it has been reported that the frequency of mutation or defects in p15^INK4b of tumor cells is lower than in p16^INK4b.^{( 10 )} TGFβ inhibits cell proliferation at G₁ phase and induces the expression of p15^INK4b approximately 30‐fold in human keratinocyte HaCaT cells, suggesting that p15^INK4b may act as an effector of TGFβ‐mediated cell cycle arrest.^{( 11 )} It has also been demonstrated that overexpression of p15^INK4b in cancer cells induces cell cycle arrest at G₁ phase.^{( 10} , ^{12 )} These observations suggest that p15^INK4b may function as a replacement for p16^INK4a, and that agents with the ability to induce p15^INK4b may provide new strategies for the prevention or therapy of malignancies, especially against tumors that lack p16^INK4a function.^{( 13} , ¹⁴ , ¹⁵ , ^{16 )}

Here we report that the pyrido‐pyrimidine derivative JTP‐70902, identified by screening for p15^INK4b‐inducing activity, induces cell cycle arrest at G₁ phase in the p16^INK4a‐inactivated human colon cancer cell line HT‐29.^{( 17} , ^{18 )} Analysis of the p15^INK4b‐inducing mechanism revealed that JTP‐70902 binds to MEK1/2 and inhibits its kinase activity. We will discuss the importance of CKI induction in the suppression of tumor growth by MEK1/2 inhibition.

Materials and Methods

Chemicals and materials.  JTP‐70902 (N‐[3‐{5‐(4‐bromo‐2‐fluorophenylamino)‐3‐cyclopropyl‐8‐methyl‐2,4,7‐trioxo‐3,4,7,8‐tetrahydro‐2H‐pyrido(2,3‐d)pyrimidin‐1‐yl}‐phenyl]‐methanesulfonamide) was synthesized at Japan Tobacco (Osaka, Japan). Screening of our compound library for p15^INK4b‐inducing activity was conducted by a branched‐DNA method, using a QuantiGene High Volume Kit for the Direct Quantitation of Cellular mRNA (Bayer Corporation, Tarrytown, NY, USA) with p15^INK4b‐specific probes. U0126 was purchased from Cell Signaling Technology (Beverly, MA, USA). Antibodies against p15^INK4b, p18^INK4c, p27^KIP1, cyclin A, cyclin E, CDK2, CDK4, CDK6, c‐Myc and α‐tubulin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies for MEK1/2, ERK1/2, phospho‐ERK1/2, p38MAPK, phospho‐p38MAPK, Akt, phospho‐Akt, pRb, phospho‐pRb (Ser780) and cyclin D1 were from Cell Signaling Technology. Horseradish peroxidase‐labeled secondary antibodies and ECL Western Blot Detection Reagent were purchased from GE Healthcare Bio‐Science (Piscataway, NJ, USA). Recombinant kinases of MEK1 (inactive), MEK2 (inactive), ERK2 (inactive), c‐Raf (truncated form, active) and B‐Raf (V599E, active) were obtained from Upstate (Lake Placid, NY, USA).

Cell culture and growth inhibition assay.  Human colorectal cancer cell line HT‐29 (American Type Culture Collections, Manassas, VA, USA) and COLO320 DM (Health Science Research Resources Bank, Osaka, Japan) were maintained in McCoy's 5 A medium and DMEM‐low glucose, respectively, both of which were supplemented with 10% fetal bovine serum and antibiotics. Cells were precultured for 24 h and then exposed to JTP‐70902. Cell growth was determined by Trypan‐blue dye exclusion test or evaluated using an In Vitro Toxicology Assay Kit, Sulforhodamine B Based (Sigma‐Aldrich, St Louis, MO, USA). Evaluation of growth inhibition against the panel of 39 human cell lines and COMPARE analysis were carried out in accordance with the methods of Yamori et al.^{( 19 )} Wild‐type and p15^INK4b‐deficient (p15^INK4b–/–) MEF,^{( 20 )} a kind gift from Dr Paul Krimpenfort (Netherlands Cancer Institute, Amsterdam, the Netherlands), were maintained in DMEM supplemented with 10% fetal bovine serum, 2 mM glutamine, 0.1 mM minimum essential medium (MEM) non‐essential amino acids, 55 mM 2‐mercaptoethanol and antibiotics. Wild‐type and p15^INK4B–/– MEF were plated at a density of 4 × 10⁴ cells in 96‐well plates. After incubation with JTP‐70902 for 4 days, cell viability was assessed using Cell Counting Kit‐8 (Dojindo, Kumamoto, Japan).

Western blot analysis and mRNA quantitation.  Western blot analysis of HT‐29 and COLO320 DM cell lysates was carried out according to a standard method. Peroxidase‐based chemiluminescence was developed by addition of ECL and the bands were detected using an LAS 3000 Image Analyzer (Fuji Film, Tokyo, Japan). Tumor tissues were homogenized in lysis buffer (1% Nonidet P‐40, 50 mM Tris‐HCl [pH 7.4], 150 mM NaCl, 0.2 mM Na₃VO₄, 1 mM EDTA) containing the protease inhibitor cocktail Complete (Roche). The resultant homogenate was centrifuged (20 000g, 10 min) and the supernatant was used for western blot analysis. p15^INK4b and p27^KIP1 mRNA levels in tumor tissues were assessed by quantitative reverse transcription–polymerase chain reaction. Total RNA was extracted from the tumor tissue using a GenElute Mammalian Total RNA Miniprep Kit (Sigma‐Aldrich). The reverse transcription reaction using 2.5 µg of the total RNA was carried out using a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA), and quantitative polymerase chain reaction was carried out using Applied PCR mixture (Applied Biosystems) and an ABI PRISM 7700 Sequence Detector (Perkin‐Elmer Biosystems, CA, USA). Oligonucleotide primers and TaqMan probes were designed using Primer Express version 1.5 software (Applied Biosystems), as follows: p15^INK4b forward primer 5′‐CCGCCCACAACGACTTTATT‐3′, reverse primer 5′‐CAGCCTTCATCGAATTAGGTG‐3′, probe 5′‐(FAM)TCTTACCCAATTTCCCACCCCCAC(TAMRA)‐3′; p27^KIP1 forward primer 5′‐AAAGACTGATCCGTCGGACAG‐3′, reverse primer 5′‐GTAGAAGAATCGTCGGTTGCA‐3′, probe 5′‐(FAM)AGCAATGCGCAGGAATAAGGAAGCGAC(TAMRA)‐3′.

Cell cycle analysis.  Unsynchronized HT‐29 cells were exposed to JTP‐70902 for 24 h. The nuclear DNA was stained with propidium iodide using a Cycle Test Plus kit (Becton Dickinson, Mountain View, CA, USA). The propidium iodide fluorescence was measured using a FACSort (Becton Dickinson) flow cytometer, and the data were analyzed using FlowJo software (Tree Star, San Carlos, CA, USA). The proportions of cells in G₀/G₁, S and G₂/M phases were determined using the Watson pragmatic algorithm.

Compound‐immobilized affinity chromatography.  JTP‐70902 was chemically modified by adding a spacer that bridged the compound and surface of the resin. Activated CH Sepharose 4B (GE Healthcare Bio‐Science) suspended in coupling buffer (0.25 mM NaHCO₃, 50% dioxane [pH 8.0]) was incubated with the modified JTP‐70902 overnight at room temperature, and then stirred in 0.1 mM Tris‐HCl (pH 8.0) for 1 h to block the unreacted active group. After blocking, the resin was washed twice alternately with acetate buffer (pH 4.0) and 0.1 M Tris‐HCl (pH 8.0). The HT‐29 cytosol fraction was prepared in accordance with the method of Hiramoto et al.^{( 21 )} One milliliter of the cytosol fraction (containing 0.1–1% Tween‐20) was added to 10 µL of the compound‐immobilized resin and incubated for 1 h at 4°C with gentle mixing. The resin was collected and washed with wash buffer (20 mM HEPES [pH 7.9], 20% glycerol, 0.1 M KCl, 0.5 mM EDTA, 0.1% Tween‐20, 1 mM DTT), then resuspended in 2× SDS‐PAGE sample buffer. After heating for 5 min at 95°C, the resin was loaded onto a 4/20% SDS‐PAGE gel and the gel was stained using the 2D Silver‐Staining Reagent ‘Daiichi’ (Daiichi Pure Chemicals, Tokyo, Japan). The protein bands were excised from the gel and provided for protein mass spectrometry analysis (Apro Life Science Institute, Tokushima, Japan).

Protein kinase assays.  A Raf–MEK–ERK cascade kinase assay consisting of active Raf (c‐Raf or B‐Raf), inactive MEK (MEK1 or MEK2) and inactive ERK2 was carried out in accordance with the method of Mallon et al.^{( 22 )} with some modifications. Two micrograms of dephosphorylated MBP (Upstate) in 0.2 M sodium carbonate (pH 9.4) was coated onto 96‐well plates (Immulon 1B ‘U’‐bottom plates; Thermolabsystems, Franklin, MA, USA) at 4°C overnight. After blocking with 1% BSA, the inactive forms of MEK1 (2 ng) or MEK2 (4 ng) with inactive ERK2 (50 ng) diluted in assay buffer (20 mM MOPS [pH 7.4], 0.1% 2‐mercaptoethanol, 0.1% BSA) were preincubated in the plates with test compounds at room temperature for 10 min. The kinase reaction was started by the addition of active B‐Raf (V599E) or c‐Raf, with 10 µM ATP and 12.5 mM MgCl₂. After incubation at 30°C for 30 min, the phosphorylation of MBP by ERK2 was detected with peroxidase‐labeled antiphosphorylated MBP antibody.

Nude mouse HT‐29 xenograft model.  Female BALB/c‐nu/nu mice obtained from Charles River Japan (Astugi, Japan) were used. On day 0, HT‐29 cells suspended in ice‐cold Hanks’ Balanced Salt Solution (HBSS) (–) were inoculated subcutaneously into the right flank of mice at 5.0 × 10⁶ cells/100 µL/site. JTP‐70902 or vehicle (0.5% methylcellulose) was administered orally twice daily for 21 days from day 5. The tumor length (L [mm]) and width (W [mm]) were measured using a microgauge twice a week after commencement of dosing, and tumor volume was calculated using the following formula:

tumor volume (mm³) = L × W × W/2.

The tumor volume in treated groups was calculated as a percentage of the tumor volume in the vehicle groups and expressed as treated/control (%). All procedures relating to the use of animals in this study were reviewed and approved by the Institutional Animal Care and Use Committee at Japan Tobacco.

Statistical analysis.  Results were expressed as mean ± SD. In the HT‐29 xenograft model, the significance of difference between the vehicle and JTP‐70902‐treated groups was analyzed by one‐way analysis of variance (anova) followed by Dunett's test. In the MEF growth inhibition assay, the significance of difference between the wild‐type MEF and p15^INK4b‐deficient MEF was analyzed using Student's t‐test. Differences were considered to be statistically significant when P < 0.05.

Results

JTP‐70902, identified as a potent p15^INKb inducer, modulates p27^KIP1, cyclin D1, cyclin A and c‐Myc protein levels and induces G₁ arrest in HT‐29 cells.  We established a high‐throughput branched‐DNA assay and screened a total of 160 000 compounds for their ability to induce the expression of p15^INK4b mRNA. As a result, a pyrido‐pyrimidine derivative, JTP‐70902, was identified to have a high p15^INK4b‐inducing activity (Fig. 1). In a time‐course analysis using human colon cancer HT‐29 cells, 100 nM JTP‐70902 slightly induced expression of the p15^INK4b protein after 8 h exposure, and the effect was pronounced after 24 h exposure (Fig. 2a). In analysis of other cell cycle regulatory proteins, p27^KIP1 was upregulated after 8 h exposure, and cyclin D1 and cyclin A were found to be downregulated 4 and 24 h, respectively, after the addition of 100 nM JTP‐70902 (Fig. 2a). The expression of p18^INK4c, CDK4, CDK6, CDK2 and cyclin E was not significantly affected. We next examined the effect of JTP‐70902 on c‐Myc, which has been reported to repress expression of p15^INK4b and p27^KIP1,^{( 23} , ²⁴ , ^{25 )} and found that JTP‐70902 drastically reduced its level after 2 h of exposure, preceding the other events observed (Fig. 2a). When HT‐29 cells were exposed to various concentrations of JTP‐70902 for 24 h, the alterations in p15^INK4b, p27^KIP1 and cyclin A protein levels were observed at a minimum concentration of 10 nM, whereas obvious reduction in the level of cyclin D1 was detected at 100 nM or more (Fig. 2b, top panel). The protein level of c‐Myc was also reduced at a concentration of 10 nM after 2 h exposure (Fig. 2b, bottom panel).

Figure 1.

Chemical structure of JTP‐70902.

Figure 2.

Effects of JTP‐70902 on the expression of cell cycle regulatory proteins and cell proliferation. (a) HT‐29 cells were treated with 0.1% dimethyl sulfoxide alone or 100 nM of JTP‐70902 for the indicated time, and expression of cyclin‐dependent kinase (CDK)‐inhibitor, cyclins, CDK and c‐Myc and the phosphorylation of retinoblastoma protein (pRb) were analyzed by western blotting. α‐Tubulin was chosen as a control. (b) HT‐29 cells were treated with various concentrations of JTP‐70902 for 24 h (top panel) or 2 h (bottom panel). The expression level of each protein and the phosphorylation status of pRb were analyzed by western blotting. (c) Asynchronized HT‐29 cells were treated with various concentrations of JTP‐70902 for 24 h and the DNA content of the cells was determined by flow cytometry. (d) HT‐29 cells were treated with JTP‐70902 at various concentrations, and the number of viable cells was determined by Trypan‐blue dye exclusion test. The symbols indicate mean ± SD values (n = 3): closed circle, 0 nM; open triangle, 1 nM; open circle, 5 nM; open square, 25 nM; cross, 100 nM.

As expected from the induction of p15^INK4b and p27^KIP1 and reduction of cyclin D1 and cyclin A, conversion of hyperphosphorylated pRb to the hypophosphorylated form was detected after 8 h exposure to 100 nM JTP‐70902, coincident with the timing of alterations in p15^INK4b and p27^KIP1 levels (Fig. 2a). The phosphorylation of pRb at Ser780 by CDK4/6 was slightly diminished after 8 h exposure, and became undetectable thereafter. These changes in the phosphorylation status of pRb occurred with 10 nM JTP‐70902 (Fig. 2b). In accordance with these observations, cell cycle analysis by flow cytometry after 24 h exposure revealed that JTP‐70902 induced G₁ arrest at a concentration of 10 nM (Fig. 2c). It is likely that the cell cycle was arrested after the induction of p15^INK4b and p27^KIP1 as the increment in cell number in G₁ phase was not detected after 8 h exposure to even 100 nM JTP‐70902 (data not shown). In addition, JTP‐70902 exerted growth‐inhibitory effects from a concentration of 5 nM in a proliferation assay of HT‐29 cells, and showed complete inhibition at 100 nM (Fig. 2d). Overall, these results suggest that JTP‐70902 induces p15^INK4b and p27^KIP1 expression and reduces c‐Myc, cyclin D1 and cyclin A levels, and consequently inhibits the growth of HT‐29 cells by arresting the cell cycle at G₁ phase.

JTP‐70902 binds specifically to MEK1/2.  The rapid downregulation of c‐Myc led us to investigate whether JTP‐70902 influences growth‐signaling pathways. In the evaluation of the phosphorylation status of ERK1/2, p38MAPK and Akt in HT‐29 cells, only the phosphorylation of ERK1/2 was inhibited by JTP‐70902 at a concentration of 10 nM or more, whereas phosphorylation of p38MAPK and Akt was not affected even at 1000 nM (Fig. 3a). This inhibition of ERK1/2 signaling occurred within 1 h of exposure to JTP‐70902, prior to the downregulation of c‐Myc, cyclin D1 and cyclin A and induction of p15^INK4b and p27^KIP1 (2, 3). These results indicate that the molecular target of JTP‐70902 may be involved in the Raf–MEK–ERK pathway.

Figure 3.

Effects of JTP‐70902 on the mitogen‐activated protein kinase/extracellular signal‐regulated kinase (ERK) kinase (MEK)–ERK pathway and identification of MEK1/2 as a JTP‐70902‐binding protein. (a) HT‐29 cells were treated with various concentrations of JTP‐70902 for 2 h, and the phosphorylation of ERK1/2, p38MAPK and Akt was analyzed by western blotting. (b) HT‐29 cells were treated with 0.1% dimethyl sulfoxide alone or 100 nM of JTP‐70902 for the indicated time, and phosphorylation of ERK1/2 was analyzed by western blotting. (c,d) The HT‐29 cytosolic fraction was supplemented with increasing concentrations of JTP‐70902 prior to loading on the resin, on which a JTP‐70902 analog was immobilized. The proteins bound to the resin were analyzed by sodium dodecylsulfate–polyacrylamide gel electrophoresis followed by silver staining (c) or western blotting with anti‐MEK1/2 antibody (d). (c) The 46‐kDa protein (asterisked arrow) and a slightly smaller non‐specific protein (arrow) are indicated.

To identify the molecular target of JTP‐70902 directly, we carried out compound‐immobilized affinity chromatography. The HT‐29 cytosol fraction was loaded onto JTP‐70902‐immobilized resin. Among the proteins recovered from the resin, only the band representing the 46‐kDa protein clearly disappeared when increasing concentrations of JTP‐70902 were added to the cytosolic fraction prior to loading (Fig. 3c, indicated by asterisked arrow), indicating specific binding of the 46‐kDa protein to JTP‐70902. Liquid chromatography–tanden mass spectrometry (LC‐MS/MS) analysis of the excised 46‐kDa band presented several peptide sequences of MEK1 and MEK2 (data not shown). Furthermore, western blot analysis showed that the 46‐kDa protein was stained with anti‐MEK1/2 antibody, and the band intensity detected was also reduced by addition of JTP‐70902 into the cytosolic fraction (Fig. 3d). These results strongly suggest that MEK1/2 is the intracellular molecular target of JTP‐70902.

JTP‐70902 is a potent inhibitor of MEK1/2 kinase.  The direct inhibition by JTP‐70902 against the kinase activity of MEK1/2 was examined using an ELISA‐based Raf–MEK–ERK cascade kinase assay consisting of active Raf (c‐Raf or B‐Raf), inactive MEK (MEK1 or MEK2) and inactive ERK2. JTP‐70902 inhibited the phosphorylation of MBP by the cascade kinase reaction regardless of the isotype of Raf or MEK used, and the inhibitory activity was much more potent than that of the known MEK inhibitor U0126 (Table 1). JTP‐70902 did not inhibit the activity of c‐Raf, B‐Raf or ERK1/2 even at 10 µM, indicating that the cascade kinase reactions were inhibited as a consequence of MEK1/2‐inhibition (Supplementary Table S1). In addition, JTP‐70902 had no inhibitory activities to the other 23 Tyr kinases and 13 Ser/Thr kinases (Supplementary Table S1). As previous studies have shown that HT‐29 cells express an active mutant of B‐Raf (V599E) that is responsible for constitutive activation of the MEK–ERK pathway and tumorigenic proliferation,^{( 26 )} it was concluded that the direct inhibition of MEK1/2 by JTP‐70902 results in growth inhibition in HT‐29 cells. Furthermore, it was observed that known MEK inhibitors, U0126 and PD98059, also induced p15^INK4b and p27^KIP1 expression in HT‐29 cells as well as JTP‐70902, corroborating that inhibition of MEK causes p15^INK4b and p27^KIP1 induction (data not shown).

Table 1.

Inhibition of mitogen‐activated protein kinase/extracellular signal‐regulated kinase (ERK) kinase (MEK) enzyme activity by JTP‐70902

MEK	Activator	IC50 value (nM)
		JTP‐70902	U0126
MEK1	B‐Raf	11	1200
	c‐Raf	47	2200
MEK2	B‐Raf	7.2	130
	c‐Raf	6.0	220

Raf–MEK–ERK cascade assays were carried out as described in Materials and Methods. IC₅₀ values of JTP‐70902 and U0126 in various combinations of Raf and MEK are shown.

JTP‐70902 shows a unique inhibition profile in comparison with conventional anticancer drugs in vitro, and antitumor activity in vivo.  We next investigated the in vitro growth inhibition profile of JTP‐70902 on a panel of 39 human cancer cell lines, for comparison with those listed in a drug sensitivity database.^{( 19} , ²⁷ , ²⁸ , ²⁹ , ^{30 )} It has been reported that agents targeting the same molecule or sharing the same mechanism of action produce a similar antitumor spectrum with high correlation coefficient in COMPARE analysis.^{( 30} , ³¹ , ^{32 )} A growth inhibition parameter, log GI₅₀, of JTP‐70902 is shown in Fig. 4a. The mean log GI₅₀ value of JTP‐70902 was –5.86, which lay in the middle of the value range of standard anticancer drugs in the panel.^{( 19 )} Most of the colorectal cancer cell lines and some of the breast, lung, melanoma, ovarian and stomach cancer cell lines were sensitive to JTP‐70902. The effective growth inhibition in colorectal cancer cell lines caused by JTP‐70902 agreed with frequent activation of the Ras–Raf–MEK–ERK pathway in these cells and their sensitivities to known MEK inhibitors.^{( 33} , ³⁴ , ³⁵ , ^{36 )} The COMPARE analysis revealed that there was no conventional anticancer drug showing a high correlation coefficient with JTP‐70902. These results are in accordance with the findings that the molecular target of JTP‐70902 is MEK1/2.

Figure 4.

Antitumor activities of JTP‐70902 in vitro and in vivo. (a) Log GI₅₀ values for 39 human cancer cell lines treated with JTP‐70902 for 48 h are presented. Columns extending to the right or left indicate the sensitivity or resistance of the cells to JTP‐70902, respectively. MG‐MID, the mean of the logarithm of the 50% growth inhibition values for 39 cell lines; Delta, the logarithm of the difference between the MG‐MID and GI₅₀ of the most sensitive cell line; Range, the logarithm of the difference between the GI₅₀ of the most resistant cell line and that of the most sensitive one. (b) HT‐29 cells were inoculated subcutaneously into the right flank of nude mice, and JTP‐70902 or vehicle was administered orally twice daily for 21 days starting from 5 days after the inoculation (closed circle, vehicle; open triangle, 10 mg/kg twice daily; open circle, 30 mg/kg twice daily; open square, 100 mg/kg twice daily). Tumor volumes were measured twice a week. The results are expressed as mean ± SD (n = 6). *P < 0.05, **P < 0.01, ***P < 0.001 versus vehicle (Dunnett's test). (c,d) Single oral dose of 100 mg/kg JTP‐70902 or vehicle was administered to nude mice bearing the HT‐29 xenograft (n = 3). The tumor tissues were excised, the phosphorylation level of ERK1/2 at 4 h after dosing was analyzed by western blotting (c), and the expression levels of p15^INK4b and p27^KIP1 mRNA at 4 and 8 h after dosing were determined by quantitative reverse transcription–polymerase chain reaction (d). (d) The data are expressed as mean ± SD. White column, vehicle treatment group; black column, JTP‐70902 treatment group.

To confirm the in vivo efficacy of JTP‐70902, we evaluated the antitumor activity of JTP‐70902 in the nude mouse HT‐29 xenograft model. HT‐29 cells were inoculated subcutaneously into the right flank of nude mice, and JTP‐70902 was administered orally twice daily for 21 days starting from 5 days after the inoculation. Tumor growth was significantly suppressed in mice treated with 100 mg/kg JTP‐70902 during the course of the treatment. The treated/control values at the end of the observation period (day 25) were 100, 80 and 43% in the 10, 30 and 100 mg/kg groups, respectively (Fig. 4b). Neither significant changes in bodyweight nor gross lesions at autopsy were observed in the mice treated with the maximum dose of JTP‐70902. Furthermore, by single oral dosing of 100 mg/kg JTP‐70902, the phosphorylation of ERK1/2 in the established tumor tissues was completely inhibited (Fig. 4c), and both p15^INK4b and p27^KIP1 mRNA levels were upregulated in parallel (Fig. 4d). These results indicate that JTP‐70902 exerts an antitumor effect in vivo by the mechanisms of action that were demonstrated in the in vitro studies.

JTP‐70902 cannot inhibit cell proliferation independently of the MEK–ERK pathway, and requires induction of p15^INK4b for its full activity.  Lastly, we examined the correlation between MEK1/2 inhibition and growth inhibition by JTP‐70902, and the involvement of p15^INK4b in its effect. We found that JTP‐70902 could not inhibit the growth of human colon cancer COLO320 DM cells even at 10 µM (Fig. 5a). Western blot analysis revealed that constitutive phosphorylation of ERK1/2 did not occur in COLO320 DM cells, although the abundance of ERK1/2 in COLO320 DM cells was higher than in HT‐29 cells (Fig. 5b). Moreover, the induction of p15^INK4b and p27^KIP1 by JTP‐70902 was not observed in COLO320 DM cells (Fig. 5b). These results suggest that COLO320 DM proliferation is independent of the MEK–ERK pathway and this is the reason for the insensitivity to JTP‐70902. These observations also support the idea that the antitumor activity and p15^INK4b‐ and p27^KIP1‐inducing activities of JTP‐70902 are dependent on MEK1/2 inhibition. To reveal whether p15^INK4b contributes to the growth‐inhibitory activity of JTP‐70902, we compared the sensitivities of p15^INK4b–/– and wild‐type MEF to JTP‐70902. The p15^INK4b–/– MEF were more resistant to the growth‐inhibitory effect of JTP‐70902 than wild‐type MEF (Fig. 5c), suggesting that p15^INK4b at least partially contributes to the growth‐inhibitory effect of JTP‐70902.

Figure 5.

Growth‐inhibitory effects of JTP‐70902 against COLO320 DM cells and p15^INK4b‐deficient mouse embryonic fibroblasts (MEF). (a) HT‐29 cells (closed circle) and COLO320 DM cells (open circle) were treated with JTP‐70902 for 72 h, and cell viability was measured by Sulforhodamine B assay. The results are shown as percentage of dimethyl sulfoxide (DMSO) treatment (mean ± SD, n = 3). (b) HT‐29 and COLO320 DM cells were treated with JTP‐70902 at the concentrations indicated for 24 h. The phosphorylation of ERK1/2 and the expression of p15^INK4b and p27^KIP1 were analyzed by western blotting. (c) Wild‐type (gray column) or p15^INK4b‐deficient MEF (white column) were treated with JTP‐70902 at the concentrations indicated for 4 days, and cell viability was determined as described in Materials and Methods. The results are shown as percentage of DMSO (mean ± SD, n = 3). *P < 0.05 versus p15^INK4b wild type (Student's t‐test).

Discussion

Our present work is based on the findings that p15^INK4b has similar functions to p16^INK4a as a tumor suppressor,^{( 7 )} but its mutation or defect in tumor cells is less frequent than in p16^INK4a.^{( 10 )} These reports suggest that p15^INK4b may act as a replacement for p16^INK4a, and a p15^INK4b‐inductive agent might enable development of a new strategy for the prevention or therapy of malignancies.^{( 13 )} Previous studies demonstrating that several antiproliferative agents, such as trichostatin‐A, indole‐3‐carbinol and 15‐deoxy‐Δ^12,14‐PGJ₂, inhibit cell proliferation accompanied by p15^INK4b induction also support this view.^{( 14} , ¹⁵ , ^{16 )}

We identified JTP‐70902, which has p15^INK4b‐inducing activity and inhibits the growth of HT‐29 cells. Using compound‐immobilized affinity chromatography, the molecular target of JTP‐70902 was identified as MEK1/2, and JTP‐70902 inhibited its kinase activity. Because HT‐29 cells were reported to have a transcriptional defect in p16^INK4a due to promoter methylation,^{( 17} , ^{18 )} it is possible for p15^INK4b induced by JTP‐70902 to complement the CKI function of p16^INK4a. Indeed, we found that the p15^INK4b–/– MEF were more resistant to the growth‐inhibitory effect of JTP‐70902 than wild‐type MEF (Fig. 5c), suggesting that p15^INK4b at least partially plays a role in growth arrest by a MEK inhibitor. As the possible mechanism of the partial growth suppression by JTP‐70902 in p15^INK4b–/– MEF cells, there is the possibility that JTP‐70902 may have p27^KIP1‐inducing activity or cyclin D‐ and c‐Myc‐reducing activities, as previously reported for other known MEK inhibitors.^{( 37} , ³⁸ , ^{39 )} As other possible mechanisms, unknown pathways might exist that have growth‐inhibitory effects on MEF cells. A recent report showed that the upregulation of p27^KIP1 was necessary for cell cycle arrest of pancreatic cell lines by the MEK inhibitor U0126, as determined by a knockdown study using antisense DNA and small interfering RNA.^{( 37 )} In contrast, it has been demonstrated that p27^KIP1 induction and inhibition of CDK2 were dispensable, but CDK4/6 inhibition was required for growth arrest by U0126 in several human colon cancer cell lines,^{( 38 )} indicating that p15^INK4b induction may be necessary for CDK4/6 inhibition in these cells. Furthermore, our observations of activities in a panel of 10 human colorectal cancer cell lines revealed that either p15^INK4b or p27^KIP1 or both were upregulated in nine cell lines in which proliferation was repressed by JTP‐70902 analogs, but not in COLO320 DM cells (data not shown). COLO320 DM cells were insensitive to JTP‐70902, and there was no alteration in the expression of p15^INK4b or p27^KIP1 in the cells (Fig. 5b). These findings indicate that inhibition of MEK1/2 may elicit cell cycle arrest by distinct mechanisms involving p15^INK4b or p27^KIP1, depending on the cancer cell lines. The precise contribution of each CKI induction to the antitumor effects should be further investigated.

The current study is the first report demonstrating that p15^INK4b can be induced by MEK1/2 inhibition. In contrast, although the MEK–ERK pathway was originally regarded to promote cellular proliferation and downregulation of CKI, such as p21^WAF1 and p27^KIP1, several studies reported that activation of the ERK pathway was required for induction of p15^INK4b.^{( 40} , ⁴¹ , ^{42 )} The ERK signaling pathway triggered by TPA and Saikosaponin A was involved in p15^INK4b induction and growth inhibition in HepG2 cells,^{( 41 )} and further evidence demonstrated that sustained ERK activation is associated with cell cycle arrest.^{( 43 )} Together with our observations in the panel of colorectal cancer cell lines described above, it is most likely that induction of p15^INK4b depends on the intensity and duration of MEK–ERK activation: a robust and persistent activation of ERK by agents such as TPA promotes p15^INK4b induction, whereas in the case of the MEK–ERK pathway involved in autonomous cell proliferation, inhibition of the MEK–ERK pathway leads to induction of p15^INK4b. Furthermore, JTP‐70902 downregulated c‐Myc, which is one of the target proteins of the MEK–ERK pathway. Phosphorylation by ERK1/2 was reported to enhance the stability of c‐Myc protein,^{( 44} , ^{45 )} and the expression of p15^INK4b and p27^KIP1 is repressed by c‐Myc.^{( 23} , ²⁴ , ^{25 )} These findings raise the possibility that inactivation of ERK1/2 causes rapid degradation of c‐Myc protein and results in the induction of p15^INK4b and p27^KIP1 in HT‐29 cells.

Although mutation or defects in MEK1/2 itself were rarely detected in human malignant tumors, constitutive activation of the upstream molecules of MEK1/2, such as K‐Ras or B‐Raf, was often observed.^{( 4 )} A recent report demonstrated that cells harboring the B‐Raf mutation have higher sensitivity to the MEK inhibitor CI‐1040 (also called PD 184352) than either Ras mutant cells or Ras/B‐Raf wild‐type cells, implying that B‐Raf mutant cells are much more reliant on the MEK–ERK pathway.^{( 39 )} Our data were in agreement with this, showing that HT‐29 expressing mutant B‐Raf was one of the most sensitive cell lines to JTP‐70902. With respect to Ras mutant cell lines, we found a discrepancy in the sensitivities because human colon cancer HCT 116 cells displayed the same sensitivity as HT‐29 cells, but A549 was resistant to JTP‐70902 (Fig. 4a). Further studies are required to establish the relationship between Ras mutation and c‐Myc reduction, p15^INK4b‐ or p27^KIP1‐induction by MEK inhibitors.

In conclusion, we discovered a p15^INK4b‐inductive compound, JTP‐70902, that exerts potent antitumor effects in vitro and in vivo, and found that its molecular target is MEK1/2. JTP‐70902 also induces expression of p27^KIP1 and downregulates cyclin D1, cyclin A and c‐Myc in tumor cells, as previously reported for other known MEK inhibitors.^{( 37} , ³⁸ , ^{39 )} We also found that p15^INK4b at least partially contributes to the growth‐inhibitory effect of JTP‐70902, and p15^INK4b induction is dependent on MEK1/2 inhibition. These findings suggest that restoration of CKI‐mediated cell cycle control is implicated in growth‐inhibitory effects of MEK inhibitors and would be a potentially beneficial approach for clinical treatment of malignancies.