Cell cycle phenotype-based optimization of G₂-abrogating peptides yields CBP501 with a unique mechanism of action at the G₂ checkpoint

Introduction

There are numerous ways for normal cells to be transformed into cancer cells. The common feature of most cancer cells is that they exhibit genomic instability (1). Most cancer cells exhibit impaired regulation at their cell cycle G₁ checkpoint. Abnormalities in classic oncogenes and in tumor suppressors, such as ras, c-myc, p53, and Rb, all impair G₁ checkpoint regulation (2). The unique dependency of most cancer cells on G₂ checkpoint to survive with DNA damage makes “G₂ checkpoint abrogation” an attractive strategy to selectively kill cancer cells (3). Because there are multiple G₂ checkpoint machineries in cells and some of them are defective in tumor cells, it would be ideal to disrupt only certain G₂ checkpoint machineries on which cancer cells are dependent to selectively kill cancer cells (4).

Some of the mechanisms of the cell cycle G₂ checkpoint are conserved between yeast and humans (5). Damage to DNA sequentially activates ATM/ATR or p38 and then CHK1, CHK2, and/or MAPKAP-K2 (6). This activation leads to phosphorylation of CDC25 at various sites, including Ser²¹⁶ of human CDC25C (7), which keeps CDC25 sequestered in the cytosol (8) by inducing the binding of 14-3-3 to this site and/or directly reduces its phosphatase activity (9). Curiously, a MARK family kinase, C-Tak1/KP78, can also phosphorylate Ser²¹⁶ of CDC25C (10, 11). Phosphorylation of CDC25 prevents it from activating CDC2/cyclin B, the master control switch of the transition from G₂ to M phase (12, 13). There are many other molecules in the G₂ checkpoint signal cascade to be considered, such as CDC25B (14), Polo-like kinases (15), Claspin (16), TopBP1 (17), BRCA1 (18), p21 (19), GADD45 (20), c-Abl (21, 22), WEE1 (23), and proliferating cell nuclear antigen (24); still more may be identified in the future. Some of these molecules are involved in independent G₂ checkpoint signal cascades, and each tumor cell is likely to depend on different sets of G₂ checkpoint signal cascades.

Nonetheless, the critical roles of CDC25C and other factors that interact with Ser²¹⁶ are suggested in work of Peng et al. (7), which examined G₂ checkpoint control in a cell line containing a site-directed mutant of CDC25C in which Ala replaces the Ser²¹⁶ residue. We applied this result to the design of short peptides, TAT-S216 and TAT-S216A (25), to competitively inhibit the activity of the protein kinases CHK1 and CHK2 in vitro. The successful inhibition of G₂ checkpoint control and the selectively enhanced sensitivity of cancer cells to DNA-damaging agents led us to further optimize the peptides using a phenotype-based (cell cycle distribution pattern) screening method. This approach led to the identification of the peptide CBP501, which inhibits the activity of multiple Ser²¹⁶-specific kinases, such as MAPKAP-K2, C-Tak1, and CHK1, and also shows cancer cell–selective enhancement of the cytotoxicity of DNA-damaging anticancer medicine. Here, we report that CBP501 is a novel anticancer drug candidate with G₂ checkpoint-abrogating activity.

Materials and Methods

Cell Culture and Reagents

Cells were cultured in various media supplemented with 10% FCS (Equitech-Bio, Kerrville, TX) at 37°C with 5% CO₂/air. The media used were RPMI 1640 (Sigma-Aldrich, St. Louis, MO) for Jurkat cells, McCoy's 5A (Invitrogen, Carlsbad, CA) for HCT116 cells, RPMI 1640 supplemented with 0.1% phytohemagglutinin (Sigma-Aldrich) and 1 unit/mL interleukin-2 (Hemagen Diagnostics, Inc., Columbia, MD) for normal T cells, and DMEM (Dainihonseiyaku Co., Osaka, Japan) with 2.5% horse serum (Invitrogen) for MIAPaCa2 cells. CBP501 was manufactured by Peptide Institute, Inc. (Osaka, Japan) and UCB-Bioproducts (Braine-L'Alleud, Belgium). TAT-S216A and CBP004 were manufactured by Sigma-Aldrich. Cisplatin (CDDP), colchicine, and bleomycin were purchased from Nihonkayaku (Tokyo, Japan), Sigma-Aldrich, and Wako Pure Chemicals (Osaka, Japan), respectively.

Cell Cycle Analysis

For cell cycle analysis, cells were stained with Krishan's buffer (0.1% sodium citrate, 50 μg/mL propidium iodide, 20 μg/mL RNase A, 0.5% NP40) followed by flow cytometry (Becton Dickinson, Franklin Lakes, NJ).

Phosphorylation Inhibition Analysis

Phosphorylation inhibition analyses were done by CycLex Co. Ltd. (Nagano, Japan). Sources of enzymes (full-length human recombinant proteins), measurement methods, and reaction buffers are as follows (all incubations were done at 30°C): AKT1 and AKT2: SF-9 cells, ELISA, 1× Mg/Mn kinase buffer [20 mmol/L HEPES-KOH (pH 7.5), 1 mmol/L DTT, 80 μg/mL bovine serum albumin, 10 mmol/L MgCl₂, 10 mmol/L MnCl₂]; CHK1, CHK2, and C-Tak1: SF-9 cells for CHK1 and C-Tak1 and Escherichia coli for CHK2, ELISA, 1× Mg kinase buffer [20 mmol/L HEPES-KOH (pH 7.5), 1 mmol/L DTT, 80 μg/mL bovine serum albumin, 10 mmol/L MgCl₂]; Polo-like kinase-1 and MAPKAP-K2: glutathione S-transferase fusion from E. coli, ELISA, and 1× Mg kinase buffer; protein kinase A: E. coli (catalytic subunit), ELISA, and 1× protein kinase A reaction buffer [20 mmol/L Tris-HCl (pH 7.0), 3 mmol/L MgCl₂]; p38: E. coli, RIA, and 1× Mg kinase buffer; TrkA: SF-9 (catalytic subunit), ELISA, and 1× Mg buffer; c-Abl: SF-9 (catalytic subunit), ELISA, and 1× Mg/Mn kinase buffer; and protein kinase C: rat brain, ELISA, and 1× protein kinase C reaction buffer [20 mmol/L Tris-HCl (pH 7.0), 3 mmol/L MgCl₂, 2 mmol/L CaCl₂, 50 μg/mL phosphatidylserine].

Computer-Assisted Docking Model Analysis

Potential binding sites for CBP501 were identified in the three-dimensional structures of G₂ checkpoint proteins of known structure of CHK1 (26) using SHOSITES (Fazix Co., New York, NY) as described previously (27). A molecular model of CBP501 was generated with Quanta software (Accelrys, Inc., San Diego, CA), and this model peptide was fit into SHOSITES pseudodensity using O software (Uppsala Software Factory, Uppsala, Sweden; ref. 28). The initial fit into this site was refined by converting the SHOSITES pseudodensity to crystallographic Fourier structure factors and doing standard crystallographic refinement (CNX, Accelrys), treating both the real and imaginary parts of these complex-valued structure factors as “experimental” restraints.

Detection of Phosphoproteins

Jurkat cells were treated with or without bleomycin plus or minus CBP501 at the indicated concentrations for the indicated times. The cells were lysed in buffer A [100 mmol/L NaCl, 10 mmol/L Tris-HCl (pH 8.0), 1 mmol/L DTT, 0.2% NP40, 10 mmol/L NaF, 10 mmol/L Na₃VO₄, 500 nmol/L okadaic acid, proteinase inhibitors], and 50 μg of lysate were run on 10% SDS-PAGE gel and analyzed by Western blot using the following antibodies: anti-phosphorylated Ser²¹⁶ of CDC25C (Cell Signaling Technology, Beverly, MA), anti-CDC25C (Santa Cruz Biotechnology, Santa Cruz, CA), anti-phosphorylated Tyr¹⁵ of CDC2 (Cell Signaling Technology), anti-phosphorylated Ser¹⁰ of histone H3 (Upstate Biotechnology, Uppsala, Sweden), anti-phosphorylated Ser³⁴⁵ of CHK1 (Cell Signaling Technology), or anti-phosphorylated Thr⁶⁸ of CHK2 (Cell Signaling Technology) antibodies.

Colony Formation Analysis

The cells were seeded at 300 per six-well plate in a triplicate manner, treated with compounds, and cultured for 7 to 8 days. The colonies were fixed and stained with crystal violet (Sigma-Aldrich).

Xenograft Model

Six-week-old male severe combined immunodeficient mice (Charles River Laboratories, Wilmington, MA) were injected s.c. in the flank with a suspension of HCT116 or NCI-H460. Tumor size was measured thrice weekly using a pair of calipers. The volumes were calculated using the following formula: volume (cm³) = [width² (mm) × length (mm)] / 2,000. The relative tumor volume was expressed as the V_t/V₀ index, where V_t is the tumor volume on a given day and V₀ is the volume of the same tumor just before first treatment (i.e., initial tumor volume). Mean relative tumor sizes ± SE were plotted. Body weight was measured thrice weekly starting at the first treatment. The percentage maximal weight loss compared with the weight just before the initial treatment was measured. For the survival analysis, 8-week-old male severe combined immunodeficient mice were transplanted i.p. with 8.5 × 10⁶ cells per animal of HCT116 cells (n = 10). Animals were housed in accordance with guidelines from the Association for the Assessment and Accreditation of Laboratory Animal Care International, and the protocols were approved by institutional animal care committee of CanBas Co. Ltd.

Statistical Analysis

The statistical significance of the differences between groups was determined by Student's t tests. Survival tests were analyzed by the Kaplan-Meier method and compared using the log-rank test.

Results and Discussion

Generation of CBP501

The proposed mechanism of action of G₂ checkpoint-abrogating peptides TAT-S216 and TAT-S216A is depicted in Fig. 1A . The sequence of amino acid residues 211 to 221 of CDC25C was connected to the 11 amino acids of the HIV-TAT transduction sequence so that the 211 to 221 portion could competitively inhibit CHK1/CHK2 in live cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Optimization of TAT-S216A. A, schema of G₂ checkpoint signal cascades and the proposed mechanism of action of TAT-S216A at a time it was designed. B, cell cycle phenotype-based screening protocol was used to identify candidate peptides that produced the cell cycle accumulation shifts. Line graph, fluorescence-activated cell sorting patterns of propidium iodide–stained phytohemagglutinin- and interleukin-2-activated normal T cells or Jurkat leukemic T cells cultured for 24 h with (red) or without (black) a candidate compound ± bleomycin or colchicine. X axis, DNA content/cell; Y axis, cell count. C, sequences, ED₅₀, and IC₅₀ of TAT-S216A, CBP004, and CBP501. ED₅₀ values indicate the concentration at which a compound decreases the G₂ accumulation at 24 h of bleomycin-treated Jurkat cells to half that of cells treated with bleomycin alone. IC₅₀ values were obtained by in vitro phosphorylation inhibition analysis with recombinant proteins. Amino acids in capital letter symbols indicate l-type amino acid, whereas those given as small letters indicate d-type amino acids. d-Bpa, d-benzoylphenylalanyl; d-Phe-F5, d-pentafluorophenylalanyl; d-Cha, d-cyclohexylalanyl.

As a first step in the optimization of TAT-S216 and TAT-S216A, amino acid residues of TAT-S216 (LYRSPSMPENL), corresponding to residues 211 to 221 of CDC25C, except for the position of Ser²¹⁶ (italicized S), were substituted randomly with a variety of amino acids. Out of ∼2,000 modified peptides, those peptides phosphorylated most efficiently in vitro by CHK1 were selected (data not shown). The second step was the phenotype-based optimization. The compounds that met the following four criteria with highest activity and selectivity were chosen: (a) alone, the compound does not change the cell cycle distribution of Jurkat cells, a human T-cell lymphoma cell line, or the distribution of phytohemagglutinin- and interleukin-2-activated human primary normal T cells (T cells), which are dividing as rapidly as Jurkat cells; (b) it does not change the cell cycle distribution of T cells that are simultaneously treated with candidate peptides plus the DNA-damaging agent bleomycin or the microtubule-disrupting agent colchicine; (c) it does not disturb the M phase accumulation of Jurkat cells when cells are simultaneously treated with peptide plus colchicine; and (d) it decreases the accumulation of G₂ phase Jurkat cells in response to simultaneous treatment with bleomycin. A typical cell cycle pattern is shown in Fig. 1B. The dose-response curve for each variable was also analyzed in actual screening. Following multiple rounds of amino acid substitution, the phenotype-based analysis and the structure-activity relationship analyses resulted in the identification of CBP501, a synthetic peptide with 12 d-type amino acid residues (Fig. 1C). The ED₅₀ of CBP501 in reducing the G₂ population of bleomycin-treated Jurkat cells was 0.1 μmol/L whereas that of TAT-S216A was 100 μmol/L. Up to 12.5 μmol/L CBP501, by itself or in the presence of colchicine, did not change the cell cycle distribution of Jurkat cells (data not shown). Despite the original conception of the mechanism of action of TAT-S216 and TAT-S216A, in vitro kinase inhibition analysis of the three peptides (including an intermediately optimized peptide, CBP004) indicated that the inhibition of CHK1 and CHK2 by these peptides did not directly correlate with the ED₅₀ values (Fig. 1C).

Mechanism of Action of CBP501

A panel analysis of the inhibition of various kinases by CBP501 showed selectivity of CBP501 inhibition for the kinases that phosphorylate Ser²¹⁶ of CDC25C, such as MAPKAP-K2, C-Tak1, and CHK1 (Fig. 2A ). MAPKAP-K2 was not known to be involved in G₂ checkpoint at the time of the optimization. Its identification illustrates an advantage of the phenotype-based screening protocol, which allows detection of inhibitors of unknown target molecules.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Inhibition of Ser²¹⁶-CDC25C phosphorylating kinases by CBP501. A, IC₅₀ values determined by in vitro phosphorylation inhibition analysis with recombinant kinases. B, SHOSITES-based hypothetical structural model for the CBP501-hChk1 complex. A molecular model of CBP501 (yellow) fitted into a groove in hChk1 (purple). Portion of the groove known to bind the substrate CDC25C and the extended acidic groove. C, CBP501 dose-dependent suppression of the phosphorylation at Ser²¹⁶ of CDC25C in the cell lysate of Jurkat cells treated for 6 h with CBP501 (indicated concentrations) with or without 40 μg/mL bleomycin. Top, phosphorylated Ser²¹⁶ of CDC25C; bottom, corresponding band for total CDC25C. D, Western blot analysis of Jurkat cell lysate of cells treated with 40 μg/mL bleomycin alone (−) or bleomycin plus 10 μmol/L CBP501 (+) for the indicated time.

The mode of interaction between CBP501 and CHK1 was predicted by in silico docking model analysis (Fig. 2B). This analysis supported a substrate mimic model for the binding of the NH₂-terminal half of CBP501 to CHK1, as this portion was predicted to fit into the substrate groove. Moreover, this analysis predicted that the basic COOH-terminal half of CBP501 could also fit into the extended acidic groove of CHK1. In agreement with the in vitro inhibitory activity of CBP501 against kinases that phosphorylate Ser²¹⁶ of CDC25C, Ser²¹⁶ phosphorylation was decreased in Jurkat cells by CBP501 treatment in dose- and time-dependent manners (Fig. 2C and D). The Ser³⁴⁵ of CHK1 and Thr⁶⁸ of CHK2 were not reduced by CBP501 treatment, suggesting that CBP501 does not inhibit kinases upstream of the cascade, such as ATM and ATR (Fig. 2D). The decreased phosphorylation of Tyr¹⁵ on CDC2 at 6 and 9 h and increased phosphorylation of Ser¹⁰ on histone H3 at 9 h (Fig. 2D) showed that the cells entered M phase in the presence of bleomycin (indicating G₂ checkpoint abrogation).

We have not explored the possibility of S phase checkpoint abrogation by CBP501 in this study. Because a known inhibitor of CHK1, UCN-01, can abrogate both S and G₂ checkpoints and S checkpoint abrogation but not G₂ checkpoint abrogation was shown to be the reason of enhanced clonogenic inhibition by UCN-01 on top of CPT-11 and S phase checkpoint inhibition by UCN-01 increased checkpoint response (29), which is similar to what we observed here as the increased phosphorylation of CHK2 by CBP501 treatment (Fig. 2D), further study is warranted to explore the possibility of S phase checkpoint inhibition by CBP501.

CBP501 Enhances Cytotoxicity of CDDP and Bleomycin against Cancer Cells In vitro

Treatment of cancer-derived cell lines, such as HCT116, a human colon cancer cell line, and MIAPaCa2, a human pancreatic cancer cell line, with bleomycin or CDDP induces G₂ arrest. CBP501 by itself did not affect cell cycle distribution of these cells up to 25 μmol/L (data not shown). Simultaneous treatment of these cells with CBP501 and bleomycin or CDDP decreased the accumulation of cells at G₂ phase, which did not occur in (normal) human umbilical vascular endothelial cell. Representative cell cycle distributions of MIAPaCa2 and human umbilical vascular endothelial cells treated with CDDP plus or minus CBP501 are shown in Fig. 3A . Figure 3B shows photomicrographs of MIAMaCa2 cells treated with CBP501, CDDP, or both and also control cells (no addition). The combined treatment with CBP501 and CDDP induced massive cell death. Colony formation analysis confirmed that CBP501 plus bleomycin or CDDP resulted in enhanced growth inhibition and/or killing of HCT116 and MIAPaCa2 cells (Fig. 3C and D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Activity of CBP501 in vitro. A, fluorescence-activated cell sorting analysis of cells treated with the indicated compound(s) for 3 h and cultured for a further 45 h in fresh medium. The numbers following CBP501 are concentration levels in μmol/L. B, fluorescence-activated cell sorting (FACS; top) and photomicrograph (bottom) of MIAPaCa2 treated with nothing (CONT), 10 μmol/L CBP501, 10 μg/mL CDDP, or 10 μmol/L CBP501 plus 10 μg/mL CDDP for 3 h and cultured for a further 45 h in fresh medium. X axis, DNA content/cell; Y axis, cell count. C and D, colony formation assay of MIAPaCa2 (C) and HCT116 (D) cells treated with bleomycin (BLM) or CDDP with or without the indicated doses of CBP501. X axis, dose of bleomycin or CDDP; Y axis, counted colonies at the end of study (n = 3).

In Xenograft Models, CBP501 Augments the Antitumor Activity of CDDP and Bleomycin without Increasing Adverse Effects

In the HCT116 xenograft model, i.v. injection of 5 mg/kg CBP501 alone showed only marginal antitumor activity, but adding 5 mg/kg of CBP501 and 5 mg/kg bleomycin together significantly enhanced antitumor activity; up to 10 mg/kg of bleomycin alone resulted in no increase in activity (Fig. 4A ). The body weight change of the animals indicated that the addition of 5 mg/kg CBP501 and 5 mg/kg bleomycin was less toxic than the treatment with 10 mg/kg bleomycin alone (Fig. 4B). As shown in Fig. 4C, the combination of the two cytotoxic agents bleomycin and CDDP showed strong antitumor activity; however, this combination severely increased adverse physiologic effects as shown by the weight loss of the mice (18.1%; Fig. 4D). On the other hand, although the antitumor activity of CBP501 plus bleomycin was as effective as CDDP plus bleomycin (Fig. 4C), the adverse effect was much less (9.1%; Fig. 4D). Figure 4E shows the enhanced antitumor activity of CDDP by CBP501 in NCI-H460, a human lung cancer xenograft model. There is a dose-dependent enhancement of CDDP cytotoxicity by CBP501. The tumor-selective enhancement of the cytotoxicity of anticancer medicine of CBP501 was further confirmed by the survival analysis of severe combined immunodeficient mice in a xenograft model (Fig. 4F). The survival difference between bleomycin and bleomycin plus CBP501 treatment was statistically significant (P < 0.01). These results suggest that CBP501 significantly increases the therapeutic index of bleomycin and CDDP, which could have significant effect clinically.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

CBP501 enhances the antitumor activity of bleomycin and CDDP without increasing their general toxicity to mice. A and B, growth inhibition (A) and maximum body weight change (B) of HCT116 s.c. xenograft tumor model in mice treated with CBP501 and bleomycin. Relative sizes of tumors were plotted versus the number of days after initiation of the treatment. The median size of the tumor was 0.1 cm³ at day 1 (n = 6). Arrows, days of treatment (days 1 and 15). Mice were pretreated with 10 mg/kg diphenhydramine 30 min before each treatment. B, maximal body weight loss relative to the original weight just before the first treatment is indicated with SE. C and D, growth inhibition of HCT116 s.c. xenograft tumors (C) and maximum body weight change (D) in mice treated with the combination of CBP501, bleomycin, and CDDP. Relative sizes of tumors were plotted versus the number of days after initiation of the treatment. The median size of the tumor was 0.1 cm³ at day 1 (n = 6). Mice were pretreated with 10 mg/kg diphenhydramine 30 min before each treatment. Arrows, days of treatment (days 1, 8, and 18). D, maximal body weight loss relative to the original weight just before the first treatment was indicated with SE. E, growth inhibition of NCI-H460 s.c. xenograft tumors by CBP501 and CDDP. Relative sizes of tumors were plotted versus the number of days after the initiation of the treatment. The median size of the tumor was 0.1 cm³ at day 1 (n = 6). Arrows, days of treatment (days 1, 4, 8, 11, 15, and 18 for CBP501 alone and days 1, 4, 15, and 18 for CBP501 plus CDDP). The CDDP dose was 3 mg/kg in all indicated groups. F, Kaplan-Meier graph of the survival analysis of the i.p. transplanted HCT116 xenograft mouse model. After the inoculation of HCT116 tumor cells at day 0, treatments were initiated at day 4. Mice were pretreated with 10 mg/kg diphenhydramine (DPH) s.c. 30 min before each treatment. The differences between bleomycin alone and bleomycin + CBP501 were statistically significant by log-rank examination (P < 0.01).

Although the data are not shown here, CBP501 entered cells within 5 min and its levels did not drop for over 20 h, indicating a high degree of stability in vitro. Further, CPB501 activity was not attenuated by treatment with human or mouse serum for 1 h at 37°C. The ability of CPB501 to enhance the antitumor activity of cancer drugs, combined with its ability to mitigate the adverse effects of the drugs and its stability, makes CBP501 a novel and promising clinical candidate as a G₂ checkpoint abrogator.