Blockade of Protein Geranylgeranylation Inhibits Cdk2-Dependent p27^Kip1 Phosphorylation on Thr187 and Accumulates p27^Kip1 in the Nucleus: Implications for Breast Cancer Therapy

Abstract

We describe the design of a potent and selective peptidomimetic inhibitor of geranylgeranyltransferase I (GGTI), GGTI-2418, and its methyl ester GGTI-2417, which increases the levels of the cyclin-dependent kinase (Cdk) inhibitor p27^Kip1 and induces breast tumor regression in vivo. Experiments with p27^Kip1 small interfering RNA in breast cancer cells and p27^Kip1 null murine embryonic fibroblasts demonstrate that the ability of GGTI-2417 to induce cell death requires p27^Kip1. GGTI-2417 inhibits the Cdk2-mediated phosphorylation of p27^Kip1 at Thr187 and accumulates p27^Kip1 in the nucleus. In nude mouse xenografts, GGTI-2418 suppresses the growth of human breast tumors. Furthermore, in ErbB2 transgenic mice, GGTI-2418 increases p27^Kip1 and induces significant regression of breast tumors. We conclude that GGTIs' antitumor activity is, at least in part, due to inhibiting Cdk2-dependent p27^Kip1 phosphorylation at Thr187 and accumulating nuclear p27^Kip1. Thus, GGTI treatment might improve the poor prognosis of breast cancer patients with low nuclear p27^Kip1 levels.

Members of the Ras and Rho family of GTPases are signal transducers that regulate many biological processes, including cell cycle progression, cell survival, and cell differentiation (15, 51). When persistently activated, they promote several oncogenic events, including uncontrolled proliferation, resistance to apoptosis, sustained angiogenesis, invasion, and metastasis (15, 51). The ability of the GTPases to mediate these tumorigenic events requires their posttranslational modification with farnesyl and geranylgeranyl lipids (54). The enzymes responsible for these modifications are farnesyltransferase (FTase) and geranylgeranyltransferase I (GGTase I), and inhibitors of FTase (FTIs) and GGTase I (GGTIs) have been developed as potential anticancer drugs (22, 40-42). While FTIs are presently undergoing phase II/III clinical trials, GGTIs are in advanced preclinical stages but have not yet entered clinical trials. We have shown that GGTIs accumulate tumor cells in the G₀/G₁ phase (50) and upregulate the cyclin-dependent kinase (Cdk) inhibitor p21^Cip1 at the transcriptional level, resulting in inhibition of Cdks (1) and retinoblastoma protein (pRb) hypophosphorylation (47), and we have suggested this as a mechanism by which GGTIs block cells in G₁ (1, 47). This is consistent with the fact that RhoA downregulates p21^Cip1 and that inhibition of RhoA geranylgeranylation, which suppresses its function, increases p21^Cip1 levels (14, 50). GGTIs also induce tumor cell death, and a possible mechanism for this is inhibition of Akt activation and suppression of the levels of survivin (17). However, since ectopic expression of constitutively activated Akt or forced overexpression of survivin rescues cancer cells only partially from GGTI effects (17), other mechanisms may be involved. Rho proteins also decrease the level of another cell cycle regulator, p27^Kip1 (p27), by enhancing its degradation (16, 26, 27). Therefore, a possible mechanism of action for GGTIs could be the inhibition of Rho geranylgeranylation and function resulting in increased levels of p27.

In normal cells, p27 inhibits nuclear Cdk activities and is thus considered a tumor suppressor (for recent reviews, see references 7, 8, and 13). p27 levels, function, and subcellular localization are regulated by phosphorylation on multiple sites. Thus, phosphorylation at Ser10 at the G₀-G₁ transition by protein kinase KIS translocates a major portion of nuclear p27 to the cytoplasm (25, 28, 29), where it is then degraded (30). The remaining nuclear p27 is regulated by Cdk2-mediated phosphorylation at Thr187 in mid- to late G₁ phase, which triggers its degradation by the ubiquitin-proteasome system.

The facts that Rho proteins, such as RhoA and RhoC, are geranylgeranylated and persistently activated (21) and that nuclear p27 levels are very low in human breast cancer (5, 10, 11, 24) prompted us to investigate whether inhibition of protein geranylgeranylation inhibits Thr187 phosphorylation of p27 and increases nuclear p27 levels and whether this is required for GGTI effects. We chose to address this important issue in breast cancer, since in this disease, low levels of nuclear p27 are associated with poor prognosis, resistance to chemotherapy, and shortened patient life expectancy (2).

MATERIALS AND METHODS

In vitro FTase and GGTase I activity assays.

GGTase I and FTase activities from supernatants of human Burkitt's lymphoma (Daudi) cells (American Type Culture Collection [ATCC], Rockville, MD) centrifuged at 60,000 × g were assayed as described previously (34). Inhibition and kinetics studies were performed by determining the ability of peptidomimetics or CVLL peptide to inhibit the transfer of [³H]geranylgeranyl and [³H]farnesyl from [³H]geranylgeranylpyrophosphate (Perkin Elmer, Wellesley, MA) and [³H]farnesylpyrophosphate (Amersham Biosciences, Piscataway, NJ) to recombinant H-Ras-CVLL and H-Ras-CVLS, respectively.

Cells and cell culture.

Human MDA-MB-468 and MDA-MB-231 breast cancer cells and murine NIH 3T3 cells were obtained from the ATCC and cultured in Dulbecco modified Eagle medium (DMEM). Human SK-Br3 and BT-474 breast cancer cells (ATCC) were cultured in McCoy's 5a medium and Hybri-Care medium, respectively. Mouse embryonic fibroblast (MEF) cells from p27 wt and p27 null mice (4) were grown in DMEM. All media were supplemented with 10% fetal calf serum, 10 units/ml penicillin, and 10 μg/ml streptomycin. Subconfluent cells were treated with different concentrations of GGTI, FTI, or dimethyl sulfoxide (DMSO) vehicle for specific time periods. After treatment, cells were harvested by trypsinization.

Membrane and cytosolic fractions were prepared using the Mem-PER eukaryotic membrane protein extraction kit from Pierce (Rockford, IL) according to the manufacturer's instructions.

siRNA-mediated knockdown of p27 in MDA-MB-468 cells.

Predesigned small interfering RNA (siRNA) to p27 (catalog number 118714) and the negative control (catalog number 4611) were purchased from Ambion (Austin, TX). Twenty-four hours before transfection, MDA-MB-468 cells were plated onto 12-well plates in fresh DMEM containing 10% fetal bovine serum and no antibiotics. Transient transfection of siRNA was carried out using the Oligofectamine reagent (Invitrogen, Carlsbad, CA), following the manufacturer's instructions. In brief, 5 nM of p27 siRNA or control siRNA was mixed with Opti-MEM medium (Invitrogen), in such a way that the total volume went to 90 μl, and then complexed with a mixture of 2 μl of Oligofectamine and 8 μl Opti-MEM; the total volume of the RNA-Oligofectamine complex was 100 μl. The RNA-Oligofectamine complex was then incubated for 20 min at room temperature before being added to the cells. Before transfection, the old medium was discarded, cells were washed once with fresh Opti-MEM, and 400 μl of fresh Opti-MEM placed into each well before adding the RNA-Oligofectamine complex to each well. The final diluted volume in each well was 500 μl. After 8 h, 500 μl DMEM containing 30% fetal bovine serum was added to each well, and cells were further incubated for 40 h.

Immunoprecipitation of p27 protein.

MDA-MB-468 cells were lysed using the CelLytic M cell lysis reagent (Sigma-Aldrich, St. Louis, MO) containing a protease inhibitor cocktail, 2 mM phenylmethylsulfonylfluoride (PMSF), 2 mM Na₃VO₄, and 6.4 mg/ml p-nitrophenylphosphate (Sigma-Aldrich). Lysates were incubated overnight with p27 antibody (BD Biosciences, San Jose, CA) at 4°C while rocking; afterward, incubation protein A anti-immunoglobulin G agarose beads were added and incubated for 2 h at 4°C and then washed four times with an excess of lysis buffer. Samples were then boiled at 100°C for 10 min in 2× sodium dodecyl sulfate sample buffer and analyzed by Western blotting as described below.

Western blot analysis.

To prepare whole-cell lysates, cells were trypsinized, washed twice with phosphate-buffered saline (PBS), and lysed in 30 mM HEPES (pH 7.5), 10 mM NaCl, 5 mM MgCl₂, 25 mM NaF, 1 mM EGTA, 1% Triton X-100, 10% glycerol, protease inhibitor cocktail, 2 mM PMSF, 2 mM Na₃VO₄, and 6.4 mg/ml p-nitrophenylphosphate. Lysates were cleared by centrifugation at 12,000 × g for 15 min, and the supernatants were collected as whole-cell extracts. The protein concentration was determined by the Bradford assay. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes, which were then blotted with antibodies specific for whole Rap1 or unprenylated Rap1, H-Ras, RhoA, the C-terminal region of pRb, Akt (Santa Cruz Biotechnology, Santa Cruz, CA), phospho-Akt(S473) (Cell Signaling, Danvers, MA), p27 (BD Biosciences, San Jose, CA), phospho-p27(T187) (Invitrogen, Carlsbad, CA), HDJ-2 (Lab Vision Corporation, Fremont, CA), phosphotyrosine (Millipore, Billerica, MA), poly(ADP-ribose) polymerase (PARP; Roche, Indianapolis, IN), and β-actin (Sigma-Aldrich, St. Louis, MO). Select bands were quantified using AlphaEaseFC densitometry software (Alpha Innotech, San Leandro, CA).

Trypan blue exclusion assay.

Adherent cells were harvested by trypsinization and diluted with 0.4% trypan blue dye and counted on a hemacytometer. Cells excluding the dye were scored as live cells, whereas cells absorbing the dye were scored as dead cells. Cell proliferation was determined by dividing the number of live cells in the treated sample by the number of live cells in the control, and the degree of cell death was expressed as the percentage of dead cells out of the total cell number.

Cell cycle analysis.

Cell cycle analysis based on DNA content was performed as described previously (36). At each time point, cells were harvested, counted, and washed twice with PBS. Cells (2 × 10⁶ to 3 × 10⁶) were suspended in 0.5 ml PBS, fixed in 5 ml of 70% ethanol overnight at −20°C, centrifuged, resuspended again in 1 ml of a propidium iodide staining solution (50 μg/ml propidium iodide, 100 units/ml RNase A, and 1 mg/ml of glucose in PBS), and incubated at room temperature for 30 min. The cells were then analyzed for cell cycle distribution using FACScan (BD Biosciences) and ModFit LT cell cycle analysis software (Verity Software, Topsham, ME).

Immunofluorescence.

MDA-MB-468 cells were plated onto a glass coverslip and treated with 50 μM GGTI-2417 or DMSO as a control for 48 h. Medium was aspirated, and the cells were washed twice with ice-cold, sterile Dulbecco's PBS (DPBS) solution (Invitrogen, Carlsbad, CA), fixed with 4% paraformaldehyde at room temperature for 15 min, and then treated with 1% Nonidet P-40 for 30 min at room temperature. After blocking with 5% bovine serum albumin for 30 min at room temperature, cells were incubated with anti-p27 antibody (1:500) at room temperature for 1 h. After washing three times with DPBS solution, cells were incubated with fluorescein isothiocyanate-conjugated secondary antibody (1:1,000) for 1 h at room temperature. Cells were then washed three times with sterile DPBS solution, mounted using the Vectashield mounting reagent (Vector Laboratories, Inc., Burlingame, CA) containing 4′,6′-diamidino-2-phenylindole (DAPI) for nuclear visualization, and analyzed using a Zeiss Axiovert Z-1 imager microscope (Carl Zeiss, Oberkochen, Germany) at 525 nM (fluorescein isothiocyanate) and 420 nM (DAPI). Images were analyzed using Image-Pro Plus 6.2 software (Media Cybernetics, Bethesda, MD).

Antitumor activity in the nude mouse tumor xenograft model.

Nude mice (nu/nu; Charles River, Wilmington, MA) were maintained in accordance with the Institutional Animal Care and Use Committee (IACUC) procedures and guidelines. MDA-MB-231 cells were harvested via trypsinization, pelleted at 805 × g for 5 min, and resuspended in a 50:50 mixture of sterile PBS and Matrigel (BD Biosciences) at 10 × 10⁶ cells per 100 μl. A volume of 100 μl (∼10 × 10⁶ cells) was injected orthotopically into the upper mammary fat pads of nude mice. The tumor xenografts were monitored by electronic caliper measurements and tumor volume (V) was calculated using the formula V = W²L/2, where width (W) is the largest diameter and length (L) is the smallest diameter. When the tumors reached ∼100 mm³, the animals were randomized and injected intraperitoneally daily with vehicle (70% DMSO), 100 mg/kg GGTI-2417 or GGTI-2418, or every third day with GGTI-2418 (200 mg/kg). Tumors were measured every third day for 12 to 18 days.

Antitumor activity in the ErbB2 transgenic mouse model.

A breeding pair of homozygous FVB/N-Tg (MMTVneu) 202Mul/J mice (JAX no. 002376; Bar Harbor, ME) was used to generate a transgenic colony in accordance to IACUC protocols and procedures. Phenotypically, these mice overexpress the receptor tyrosine kinase ErbB2 driven by the mouse mammary tumor virus promoter, which results in tumor formation at ∼200 days of age. Once tumors reached sizes above 800 mm³, osmotic minipumps (Alzet, Cupertino, CA) were loaded with either vehicle (80% polyethylene glycol 300, 20% DMSO) or GGTI-2418 (100 mg/kg daily) and implanted subcutaneously under the dorsal surface of the mouse, between the shoulder blades. Vehicle or GGTI-2418 was constantly delivered for 14 days, during which tumor volumes were measured and calculated as described above. Tumor biopsy samples were snap frozen in liquid N₂ and stored at −80°C until further processed. Weighted tumor samples were mixed with the appropriate amount of the T-PER tissue protein extraction reagent (Pierce, Rockford, IL) containing protease inhibitor cocktail, 2 mM PMSF, 2 mM Na₃VO₄, and 6.4 mg/ml p-nitrophenylphosphate and homogenized with the PCR tissue homogenizing kit (Fisher Scientific). The homogenate was centrifuged at 13,000 × g for 30 min at 4°C, and the supernatant was subjected to Western blotting with the indicated antibodies.

RESULTS

GGTI-2418 is a highly potent, competitive, and selective inhibitor of GGTase I in vitro.

Protein prenyltransferases such as GGTase I and FTase modify the carboxyl-terminal cysteines of proteins terminating with CAAX tetrapeptide sequences where C is cysteine, A is aliphatic residues, and X is any amino acid. GGTase I prefers proteins where X is a leucine, whereas FTase prefers methionine or serine at the X position (54). We designed Cys-A-A-Leu peptidomimetics where cysteine was replaced with methyl imidazole and the dipeptide “A-A” was replaced by 3-aryl-piperazin-2-one derivatives (Fig. 1A). The synthesis of these peptidomimetics has been described elsewhere (38). In the manuscript, we sought to characterize the biochemical activities of some of these peptidomimetics and determine the mechanism by which they induce cell death in breast cancer. Figures 1A and B show the structures and activities of the most potent and selective of these compounds. In vitro, GGTI-2418 inhibited GGTase I and FTase activities with 50% inhibitory concentrations (IC₅₀s) of 9.5 ± 2.0 nM and 53 ± 11 μM, respectively, a 5,600-fold selectivity toward inhibition of GGTase I versus FTase. Two other compounds substituting either p-fluorphenyl (GGTI-2432) or naphthyl (GGTI-2430) group for the phenyl group of GGTI-2418 showed similar potencies against GGTase I in vitro (IC₅₀s of 7.1 ± 4.3 nM and 14 ± 6.4 nM, respectively). The FTase inhibition activities of these two compounds indicate that GGTI-2432 and GGTI-2430 were 18,000- and 340-fold more selective for GGTase I than they were for FTase, respectively. The K_m and V_max values obtained for H-Ras-CVLL were 2.8 μM and 0.07 pmol/min, respectively. GGTI-2418 demonstrated competitive inhibition of GGTase I against the H-Ras-CVLL protein with a K_i value of 4.4 ± 1.6 nM (Fig. 1C).

FIG. 1.

GGTI-2418 and its prodrug GGTI-2417 are potent and selective inhibitors of GGTase I activity in vitro and in whole cells. (A) Structures of GGTI-2417/GGTI-2418, GGTI-2431/GGTI-2432, and GGTI-2429/GGTI-2430. (B) Increasing concentrations of GGTI-2418, GGTI-2432, and GGTI-2430 were incubated in the presence of H-Ras-CVLL protein and [³H]geranylgeranylpyrophosphate to determine inhibition of GGTase I activity (•) and H-Ras-CVLS protein and [³H]farnesylpyrophosphate to determine inhibition of FTase activity (○), as described in Materials and Methods. (C) In vitro GGTase I competition assay showing the activity against H-Ras-CVLL alone (⋄) or in the presence of 10 nM GGTI-2418 (•) and 20 nM GGTI-2418 (▴). (D) H-Ras-transformed NIH-3T3 cells were treated in the absence or presence of GGTI compounds for 48 h and processed by Western blotting with antibodies to Rap1 and H-Ras. U indicates the band for unprenylated H-Ras or Rap1 protein, and P indicates the band for fully prenylated H-Ras or Rap1 protein.

GGTI-2417 selectively inhibits protein geranylgeranylation in H-Ras-transformed NIH 3T3 cells.

As with previous FTIs and GGTIs developed by us and others, we employed a methyl ester prodrug strategy (Fig. 1A), which allows the compounds to more easily cross the plasma membrane (34, 35), thus facilitating delivery of these drugs to cells. NIH 3T3 cells stably transformed with GTP-locked H-Ras (61L) (H-Ras/3T3) were treated with increasing concentrations of GGTIs. Figure 1D shows that GGTI-2417 exhibited the most potent inhibitory activity against Rap1 geranylgeranylation, with an IC₅₀ of 400 ± 100 nM, followed by GGTI-2429 and GGTI-2431 with IC₅₀s of 600 nM and 700 nM, respectively. In contrast, GGTIs had minimal effect on H-Ras farnesylation at concentrations as high as 50 μM. Each compound thus showed high selectivity for inhibition of cellular GGTase I compared to for FTase, the most selective being GGTI-2417, with >125-fold selectivity (Fig. 1D). Therefore, we used GGTI-2417 and GGTI-2418 as the most potent and selective inhibitors for the remainder of the study.

GGTI-2417 increases p27 protein levels and induces accumulation in the G₀/G₁ phase as well as apoptotic cell death in breast cancer cells.

Low levels of nuclear p27 are associated with poor prognosis for various cancers, especially breast cancer (2, 5, 10, 11, 24). Since RhoA is geranylgeranylated and downregulates nuclear p27 (16, 26, 27), we determined whether inhibition of protein geranylgeranylation affects p27 levels. We treated MDA-MB-468 cells for 48 h with different concentrations of GGTI-2417, vehicle control or 25 μM FTI-2153, a highly selective FTase, but not GGTase I, inhibitor (46). GGTI-2417 inhibited the geranylgeranylation of Rap1, but not the farnesylation of HDJ-2, an exclusively farnesylated protein, in a dose-dependent manner (Fig. 2A). Importantly, GGTI-2417 treatment led to accumulation of RhoA in the cytosol, which was paralleled by a decrease of RhoA in the plasma membrane, indicating that GGTI-2417 inhibited RhoA geranylgeranylation (Fig. 2B). GGTI-2417 increased p27 protein levels in a concentration-dependent (Fig. 2A) and time-dependent manner starting after 12 h of exposure to GGTI-2417 (Fig. 2C). As expected, FTI-2153 completely inhibited HDJ-2 processing, did not inhibit Rap1 geranylgeranylation, and failed to increase p27 (Fig. 2A, lane 7). Together with our earlier observation that 1 μM FTI-2153 effectively prevents H-Ras prenylation (46), these results suggest that the increase in p27 correlates with inhibition of protein geranylgeranylation, but not with that of protein farnesylation.

FIG. 2.

GGTI-2417 induces a concentration-dependent increase of p27 protein levels, G₀/G₁ phase accumulation, inhibition of proliferation, and cell death in breast cancer cells. MDA-MB-468 breast cancer cells were treated with compound or vehicle control for 48 h and then processed for various assays. (A) Effect of increasing concentrations of GGTI-2417 or 25 μM FTI-2153 on unprenylated Rap1 (U-Rap1), HDJ-2, and p27 levels. The U-Rap1 and p27 bands were quantified by densitometry. The results are given above the images as percent unprenylated Rap1 (% U) compared to the maximum seen at 100 μM GGTI-2417 or as fold change compared to the control in lane 1, respectively. (B) Effect of increasing concentrations of GGTI-2417 on the distribution of RhoA between plasma membrane and cytosolic fractions. (C) Changes in p27 levels in response to GGTI-2417 over time. The p27 bands were quantified by densitometry, and the change is indicated above the image. (D) Trypan blue exclusion assays to determine cell proliferation and cell death. Standard deviations from three independent experiments are shown with error bars. (E) Effect of increasing concentrations of GGTI-2417 on PARP cleavage. (F) MDA-MB-468, MDA-MB-231, SK-Br3, and BT-474 breast cancer cells were treated with 50 μM of GGTI-2417 or vehicle and then tested by Western blot assays for unprenylated Rap1, HDJ-2, and p27. U and P designate unprocessed and processed forms of the prenylated proteins, respectively.

GGTI-2417 inhibited cell proliferation in a dose-dependent manner compared to control cells, with an IC₅₀ of approximately 4 μM (Fig. 2D, upper panel). These results suggest that inhibition of proliferation induced by GGTI-2417 correlates closely with the concentration-dependent increase in p27 protein levels as seen in Fig. 2A. This was most likely due to accumulation in the G₀/G₁ phase of the cell cycle; 50 μM GGTI-2417 increased the fraction of cells in G₀/G₁ from 51 to 78%. In addition to inhibiting proliferation, GGTI-2417 induced tumor cell death, as determined by an increase in trypan blue-positive (dead) cells in a GGTI-2417 concentration-dependent manner (Fig. 2D, lower panel). Furthermore, Fig. 2E shows that GGTI-2417 treatment of MDA-MB-468 cells resulted in PARP cleavage, as demonstrated by the appearance of an 85-kDa band derived from PARP. Since PARP cleavage is an early marker of apoptosis (31), this finding suggests that the type of cell death induced by GGTI-2417 is apoptosis.

To examine whether GGTI-2417 has similar effects on other human breast cancer cell lines, we also exposed MDA-MB-231, SK-Br3, and BT-474 breast cancer cells to 50 μM GGTI-2417. Similar to effects seen with the MDA-MB-468 cells, which were included as a control, this treatment inhibited the processing of Rap1A, but not HDJ-2, induced p27 protein levels (Fig. 2F), inhibited cell proliferation, increased the fraction of cells in the G₀/G₁ phase as well as the percentage of cell death in all three cell lines (Table 1).

TABLE 1.

Effects of 50 μM GGTI-2417 on proliferation, cell death, and cell cycle distribution

Cell line	Presence of GGTI-2417	% Proliferation	% Cells in G0/G1 phase	% Cell death
MDA-MB-468	−	100	49 ± 9.3	7 ± 4
	+	52 ± 5.5	75 ± 4.7	21 ± 10
MDA-MB-231	−	100	62 ± 8.7	18 ± 7
	+	30 ± 8.2	86 ± 8.1	68 ± 6
SK-Br3	−	100	62 ± 4.9	9 ± 2
	+	75 ± 5.5	69 ± 4.2	26 ± 8
BT-474	−	100	69 ± 8.6	15 ± 3
	+	42 ± 12	78 ± 8.7	32 ± 5

Knockdown of p27 protein levels by siRNA results in resistance to GGTI-2417-induced cell death in MDA-MB-468 cells.

To determine whether the inhibition of proliferation and/or induction of tumor cell death requires p27, we knocked down p27 levels by using an siRNA approach. Figure 3A shows that in the presence of p27 siRNA, the ability of GGTI-2417 to induce p27 expression was blocked. While cotreatment with p27 siRNA did not significantly affect the ability of GGTI-2417 to inhibit proliferation in MDA-MB-468 cells (Fig. 3B), it rescued the cells from GGTI-2417-induced cell death (Fig. 3C). In the absence of p27 siRNA, GGTI-2417 increased tumor cell death from 10.3 to 37.2%. In contrast, in the presence of 5 nM p27 siRNA, GGTI-2417 was unable to induce tumor cell death. However, these p27-deficient cells could still die in response to other known apoptotic stimuli, e.g., taxol (Fig. 3C), suggesting that cell death in response to GGTI-2417, and not cell death per se, requires p27. Taken together, these results indicate a pivotal role for p27 in the ability of GGTI-2417 to induce breast tumor cell death.

FIG. 3.

p27 is required for the induction of cell death by GGTI-2417. (A to C) siRNA-mediated silencing of p27. MDA-MB-468 cells were treated with 50 μM GGTI-2417, 1 μM taxol, or vehicle (DMSO) in the absence or presence of p27 siRNA. (A) Increased p27 levels in response to GGTI-2417 and loss of p27 expression and induction in the presence of p27 siRNA. Treated cells were counted via trypan blue exclusion assay to determine inhibition of proliferation (B) and induction of cell death (C). (D to F) MEFs lacking p27 expression are unable to die in response to GGTI-2417. p27 wt and p27 null MEFs were treated with 50 μM GGTI-2417 for 72 h and processed by further assays. (D) GGTI-2417 increases p27 levels in wt, but not null, MEFs and induces pRb hypophosphorylation in both p27 wt and p27 null MEFs. Cells were counted via trypan blue exclusion assay for inhibition of proliferation (E) and induction of cell death (F). The results represent the averages of two independent experiments, each done in triplicates. +, present; −, absent.

GGTI-2417 does not induce cell death in MEF cells lacking p27.

To further study the role of p27 in the induction of cell death by GGTI-2417, we exposed wt MEFs and MEFs lacking p27 (p27 null) to 50 μM of GGTI-2417. In wt MEFs, as in human breast cancer cells, GGTI-2417 inhibited Rap1A geranylgeranylation and increased the levels of p27 (Fig. 3D). As expected, in p27 null MEFs, there was no basal p27 expression. GGTI-2417 also induced pRb hypophosphorylation in p27 wt as well as p27 null MEFs, suggesting that the ability of GGTI-2417 to arrest cells in G₁ (Table 1) does not depend on p27 (Fig. 3D, lower panel).

GGTI-2417 treatment was associated with potent inhibition of proliferation in p27 wt MEFs (89.0% ± 3.1%) and MEFs lacking p27 (61.6% ± 8.1%) (Fig. 3E), which is consistent with the pRb phosphorylation state reported above. In p27 wt MEFs, GGTI-2417 also strongly induced cell death from 22.4% ± 7.2% in vehicle-treated cells to 53.4% ± 7.0% (Fig. 3F). p27 null cells showed slightly higher basal levels of cell death (27.5% ± 2.1%); however, treatment with GGTI-2417 did not increase the number of dead cells (Fig. 3F).

GGTI-2417 prevents the degradation of nuclear p27.

Since the RhoA-dependent downregulation of p27 appears to be mediated by Cdk2-mediated phosphorylation (27), which in turn depends on prior phosphorylation in two Tyr residues (12, 23), we next determined whether GGTI-2417-induced increase in p27 levels was associated with altered p27 phosphorylation. Indeed, Fig. 4A shows that GGTI-2417 inhibited the phosphorylation of p27 at Tyr74 and/or Tyr88 and, even more dramatically, at Thr187. Accordingly, cycloheximide did not prevent GGTI-2417-dependent increase in p27 levels (Fig. 4B), suggesting that GGTI-2417 stabilizes rather than induces p27. Immunofluorescent staining revealed that in MDA-MB-468 cells, nuclear levels of p27 increased sixfold in response to GGTI-2417, whereas the cytosolic p27 levels were not affected significantly (Fig. 4C and D).

FIG. 4.

GGTI-2417 inhibits phosphorylation events that are required for subsequent p27 degradation and accumulates nuclear p27. (A) MDA-MB-468 cells were treated with 50 μM GGTI-2417 and processed as described in the legends for Fig. 2 and 3. Western blot analysis of immunoprecipitated p27 with a p27 or phosphotyrosine antibody, followed by densitometric analysis, revealed that p27 levels increased 11.9-fold, while Tyr phosphorylation increased by only 3.5-fold, suggesting a specific downregulation of Tyr74 and/or Tyr88 phosphorylation (top panel). There are three Tyr residues in p27, in positions 74, 88, and 89, but only Tyr74 and Tyr88 are phosphorylated (12, 23). GGTI-2417 caused an even greater downregulation of Thr187 phosphorylation in p27. (B) Inhibition of protein synthesis with cycloheximide (10 μg/ml) 2 h prior to, and during, exposure to 50 μM GGTI-2417 does not prevent the increase in p27 levels. The changes compared to the control, as determined by densitometry, are given above the image. (C) Representative immunofluorescence images showing that a 48-h exposure of MDA-MB-468 cells to GGTI-2417 upregulates nuclear p27, where it can function as a Cdk inhibitor. (D) Quantitative analysis of the relative cellular amounts of p27 in the nucleus and cytoplasm. Five hundred ninety-five vehicle-treated cells in six fields and 447 GGTI-2417-treated cells in 11 fields were analyzed. The plotted values were computed with the following formula: number of pixels × intensity/number of cells. The value for vehicle-treated cells was set at 100%. The columns represent the mean ± standard error of the mean. +, present; − absent.

GGTI-2418 significantly inhibits growth of breast tumors in two animal models.

To determine whether GGTI-2418 inhibits tumor growth in nude mice, human MDA-MB-231 breast cancer cells were implanted orthotopically in the mammary fat pads. Tumors from vehicle-treated mice grew to an average size of 860 ± 87 mm³. In contrast, tumors from mice treated with 100 mg/kg GGTI-2418 daily or 200 mg/kg every third day grew to only average tumor sizes of 139 ± 17 or 276 ± 26 mm³, respectively (Fig. 5A), corresponding to tumor growth inhibitions of 94% and 77%, respectively (P < 0.005 for both). Treatment with 100 mg/kg GGTI-2417 daily resulted in growth inhibition of 76% (P < 0.005) (data not shown). Taken together these data clearly indicate that GGTI-2418 potently inhibits the growth of breast tumor xenografts with either daily or intermittent dosing.

FIG. 5.

GGTI-2418 significantly inhibits the growth of breast tumor xenografts and induces regression of ErbB2-driven mammary tumors in transgenic mice. (A) Nude mice implanted with MDA-MB-231 breast cancer tumors in the mammary fat pads were injected intraperitoneally with either vehicle daily (⧫), 100 mg/kg GGTI-2418 daily (□), or 200 mg/kg GGTI-2418 every third day (▵). (B to C) Effect of GGTI-2418 treatment on mammary tumor progression in ErbB2 transgenic mice. (D) Effect of GGTI-2418 on p27 in vivo. Tumor biopsies were obtained from ErbB2 transgenic mice before and after vehicle or GGTI-2418 treatment and prepared for Western blot analyses of select proteins. As expected, GGTI-2418 accumulates unprenylated Rap1 and prevents the activation of Akt. GGTI-2418 also upregulates p27 levels in vivo. The numbers above the Western blot indicate the fold change in posttreatment samples, as determined by densitometric analysis. The change was 0.63 ± 0.19-fold in vehicle-treated mice and 2.68 ± 0.63-fold in GGTI-2418-treated mice (P = 0.03).

To determine whether GGTI-2418 is also effective in an entirely different tumor model, we used ErbB2 transgenic mice, which developed mammary tumors ranging in size from 818 to 3,903 mm³ (see Table S1 in the supplemental material). In the absence of treatment, tumor growth was rapid, at 103 ± 12 mm³ per day. In contrast, treatment with GGTI-2418 at 100 mg/kg/day not only halted tumor growth, but also actually induced massive tumor regression within a few days. Figures 5B and C show a representative example of a tumor that decreased by 76% following GGTI-2418 treatment. The degree of regression was independent of the size of the tumor before initiation of treatment; in seven mice with a total of 17 tumors, treatment with 100 mg/kg GGTI-2418 resulted in tumor regression between 34 and 100%, with an average of 60% ± 4% (standard error of the mean) (see Table S1 in the supplemental material).

To evaluate whether the tumor regression described above was associated with changes in p27 levels as well as the known molecular targets of GGTI-2418, Rap1, and phospho-Akt (17) in vivo, we performed tumor biopsies in several mice before and after initiation of GGTI-2418 treatment. As shown in Fig. 5D, GGTI-2418 therapy inhibited the geranylgeranylation of Rap1 and caused a dramatic decrease in S473 phosphorylation of Akt. Most importantly, in three out of three vehicle-treated mice, p27 levels decreased or did not change during tumor progression, whereas in five out of five GGTI-treated mice, p27 levels were upregulated between 1.4- and 5-fold.

DISCUSSION

Rho proteins are overexpressed and/or persistently activated in breast cancer, and this is associated with poor patient prognosis (9, 39). One possible mechanism by which Rho proteins may contribute to breast cancer oncogenesis is by downregulating the levels of the CDK inhibitor p27. Indeed, in cultured cells, Rho proteins have been shown to decrease the protein levels of p27 (26), possibly through a mechanism involving the protein kinase ROCK, a downstream effector of Rho proteins (16). The fact that low p27 levels are associated with poor prognosis, resistance to chemotherapy, and shorter life expectancy of breast cancer patients (5, 10, 11, 24) supports the hypothesis that Rho proteins may contribute to breast cancer oncogenesis by keeping p27 levels persistently low. In the manuscript, we provide evidence that in breast cancer cells, inhibition of the important posttranslational modification geranylgeranylation, which is required for Rho function, results in a robust upregulation of p27.

Although we have observed before that GGTI increases p27 levels (20), this is the first study addressing the mechanism by which GGTIs increase p27 levels. Our results suggest that GGTI-2417 increases the levels of p27 by inhibiting its Thr187 phosphorylation by Cdk2, its degradation, and nuclear accumulation. Furthermore, the GGTI-2417 effects on p27 may depend at least partially on preventing proper RhoA location and therefore, activation. Our findings are consistent with the work of others who showed that inhibiting Rho function upregulates p27 levels (26, 27, 52). In this study, we have found that p27 is required for GGTI-2417 to induce breast tumor cell death. Using either p27 null cells or p27 siRNA, in cells expressing no or very little p27, GGTI-2417 was unable to induce cell death. On the other hand, GGTI-2417 could still induce hypophosphorylation of pRb and inhibit the proliferation of cells expressing very little or no p27 (Fig. 3B and E), indicating that GGTI-mediated G₀/G₁ arrest may depend on molecules other than p27. Nonetheless, our data suggest the GGTI-induced increase in nuclear p27 function is required for the ability of GGTI-2417 to induce breast tumor cell death. p27 protein levels and subcellular distribution are regulated by reversible phosphorylation on multiple sites. Phosphorylation at Thr187 by Cdk2/cyclin E in G₁ creates a so-called phosphodegron that is recognized by the SCF^Skp2 ubiquitin ligase, which recruits p27 to ubiquitin-dependent proteolysis. Our results suggest that GGTI-2417 inhibits Thr187 phosphorylation of p27 by Cdk2 and its subsequent degradation. This is consistent with our previous report demonstrating that a first-generation GGTI analog, GGTI-298, inhibits Cdk2 activity (1). However, it has been a puzzling problem to solve how p27 can be effectively phosphorylated by Cdk2, which is bound to, and kept inactive by, p27. Recently, a possible explanation to this conundrum has been provided by the finding that p27 is phosphorylated on Tyr74 and Tyr88, which partially activates p27-bound Cdk2/cyclin E complexes and thus appears to convert p27 from a Cdk2 inhibitor to a Cdk2 substrate (12, 23). Considering these recent data, our results (Fig. 4A) are also consistent with the idea that GGTI-2417 downregulates Tyr phosphorylation of p27. Whether this contributes to the even greater loss of phosphorylation in Thr187 remains to be determined.

Furthermore, GGTI-2417 did not increase cytoplasmic p27. This is important, since more-recent data suggest that cytoplasmic p27 has tumor-promoting activities. For example, in cancer cells, persistently activated Akt phosphorylates cytoplasmic p27 at both Thr198 and Thr157, which stabilizes p27 and prevents it from being reimported to the nucleus, respectively (32, 36, 37, 44, 49). RNA interference experiments suggest that cytoplasmic p27 increases cell migration, survival, and tumorigenicity of human glioma cells (53), and human metastatic melanoma cells have high levels of cytoplasmic p27 (18). Most strikingly, a p27 mutant unable to bind cyclins or Cdks, accumulates in the cytoplasm and has oncogenic properties in a mouse knock-in model (6). Together, these data suggest that both expression levels and subcellular distribution of p27 are important for tumor prognosis and therapeutic strategies (7, 8). Given that GGTI-2417 also prevented Akt activation (Fig. 5D), it is possible that GGTI-2417 may be able to promote p27's cytoplasmic degradation and/or relocation into the nucleus.

Breast tumors aberrantly overexpress several genes that are known to activate Rho proteins that in turn downregulate p27. For example, the receptor tyrosine kinases EGFR and ErbB2 are overexpressed in a large number of breast cancers (for reviews, see references 3 and 48). These receptors activate Ras, which in turn activates Rho proteins. Our findings that GGTI-2418 suppresses the growth in nude mice of MDA-MB-231 breast tumors (which contain a K-Ras mutation) and that GGTI-2418 causes tumor regression in a transgenic animal model where breast tumors are driven by ErbB2 further support our hypothesis that GGTIs will have potent antitumor activity in breast cancers in which signaling pathways lead to activation of Rho and subsequent downregulation of p27. This is also consistent with recent studies demonstrating that ErbB2 antibodies such as Herceptin modulate p27 via multiple signaling pathways (33).

The fact that inhibition of Rho geranylgeranylation upregulates p27 levels is not surprising since Rho proteins are known to downregulate p27 (16, 26, 27). Furthermore, given that p27 is important for coordinating cell cycle progression and cell survival (7, 8, 13, 19), it is not surprising either that p27 is required for GGTI to induce tumor cell death. Whether p27 plays a pro- or antiapoptotic role depends on the cell type, cell status (transformed versus normal), subcellular localization, and cleavage status of p27 (see reference 8 and the references cited therein). Our data suggest that p27 is critical for cell death and that tumors must keep its nuclear level low to survive. This is consistent with the fact that low levels of p27 are required for assembling cyclin D/Cdk4 and cyclin D/Cdk6 complexes whereas high levels inhibit Cdk activities (43). Taken together, our studies have identified inhibition of p27 phosphorylation and the subsequent accumulation of nuclear p27 as a key mediator in the mechanism of GGTI antitumor activity and demonstrate that inhibition of protein geranylgeranylation may be an effective approach to breast cancer therapy. The recent finding that a targeted deletion of the β subunit of GGTase I reduces tumor formation and improves the survival of mice with K-Ras-expressing lung tumors (45) further validates GGTIs as potential antitumor agents. This, coupled with the fact that GGTI-2418 causes significant regression of breast tumors driven by ErbB2, a prevalent poor prognostic factor in this disease, gives strong support to evaluating GGTIs in breast cancer patients whose tumors contain low levels of nuclear p27 and/or express high levels of ErbB2.