Pentameric procyanidin from Theobroma cacao selectively inhibits growth of human breast cancer cells

Breast cancer is the most prevalent malignancy in women and constitutes 30% of all cancers reported in the United States (1). Current treatments in patients with early stage (localized), as well as late stage (metastatic) disease may involve surgery, followed by radiation and hormonal and/or cytotoxic chemotherapy. One of the major problems associated with drug treatments of cancer is their toxic side effect. Most of the currently used chemotherapeutic agents do not optimally discriminate between normal and tumor cells and induce cell death in all actively proliferating cells. As a result, new drug development approaches are aimed at developing compounds that specifically attack tumor cells while sparing normal cells.

Early genetic changes associated with malignancy involve genes that regulate cell cycle progression, and often these changes result in a loss of G₁ checkpoint in tumor cells due to defects in retinoblastoma protein (pRb) and p53 cell cycle pathways (2). Differences in these pathways between normal and tumor cells can be exploited in tumor cells by novel chemopreventive and chemotherapeutic approaches that selectively block the cell cycle progression.

Similarly, the mitochondrion, a cellular organelle responsible for energy production for all cellular functions, has been reported to function differently in cancer cells compared with normal cells, exhibiting a higher mitochondrial membrane potential approximately 60 mV higher than that of control cells (3). Altered mitochondrial function in tumor cells contributes to a wide array of tumor phenotypes, including higher rates of glycolysis (4), glucose transport (5), gluconeogenesis (6), and modified amino acid metabolism (7). Furthermore, various tumor cell lines exhibit differences in the number, size, and shape of their mitochondria relative to slower growing tumors and normal controls (8). Such differences between mitochondria of normal and cancer cells should enable the development of more selective and efficient cancer therapies capable of altering mitochondrial functions in cancer cells while sparing normal cells.

In a search for novel compounds that could be beneficial for cancer prevention and treatment, procyanidin and prodelphinidins (9, 10) have attracted a great deal of attention due to their wide range of biological activities. There is a growing body of evidence suggesting that these compounds act as potent antioxidants in vitro, ex vivo, and in vivo (11). Our interest has focused on the evaluation of several representative compounds within the procyanidin class for activity against breast cancer. Initial observations have shown that procyanidin-rich fractions prepared from the seeds of Theobroma cacao, which are used to prepare cocoa, inhibit in vitro growth of several human breast cancer cell lines (12). These studies have revealed oligomeric procyanidins, such as the pentameric procyanidin (referred to hereafter as pentamer) to be cytotoxic to breast cancer cells. The structure of the naturally occurring pentamer has been confirmed by total synthesis (12). Comparative in vitro studies, using the MDA MB-231 cell line, indicated growth inhibition caused by pentamer to be the result of cell cycle arrest at G₀/G₁. Moreover, a series of control experiments showed that direct actions of the compound itself caused the G₀/G₁ arrest, rather than any potential effects caused by hydrogen peroxide formed by autooxidation (12).

Collectively, these results prompted us to examine the molecular changes associated with pentamer-induced cell cycle arrest and cytotoxicity against human breast cancer cells. Our results provide evidence that pentamer selectively inhibits growth of human breast cancer cell lines, whereas normal and immortalized human mammary epithelial cells (HMEC) exhibit resistance. The mechanism seems to involve decreased phosphorylation of integral growth regulatory proteins, including p53. These results indicate the potential for pentamer to be further evaluated as a selective chemopreventive and/or chemotherapeutic of human breast cancer and possibly other human cancers.

MDA MB-231, MDA MB-436, MDA MB-468, SKBR-3, and MCF-7 cell lines were obtained from the Lombardi Comprehensive Cancer Center, Tissue Culture Shared Resource and grown in complete medium containing DMEM (Life Technologies, Gaithersburg, MD) supplemented with 10% fetal bovine serum (Quality Biological, Gaithersburg, MD). Medium for MCF-7 also contained nonessential amino acids (Life Technologies).

Normal HMEC at passage 8 were purchased from Clonetics-Bio Whittaker, Inc. (Walkersville, MD) and maintained according to the supplier's instructions in mammary epithelial basal medium (Clonetics-Bio Whittaker) supplemented with 52 μg bovine pituitary extract/mL, 10 ng/mL human epidermal growth factor, 5 μg/mL insulin, and 0.5 μg/mL hydrocortisone and grown in 37°C incubators with low (0.1–0.2%) CO₂ settings. The spontaneously immortalized, nontumorigenic human mammary epithelial cell line MCF-10A and its modified counterparts MCF-10A-ErbB2/Ras (13) and MCF-10A-c-Myc (14) were maintained in F-12/DMEM supplemented with 5% horse serum (Life Technologies), 20 ng/mL epidermal growth factor (Upstate Biotechnology Incorporated, Lake Placid, NY), 10 μg/mL insulin (Biofluids, Rockville, MD), and 500 ng/mL hydrocortisone. The 184A1N4 (referred to hereafter as A1N4) and 184B5 cell lines are nontumorigenic cell lines derived from primary cultures of HMECs that were immortalized with benzo(a)pyrene (15) and provided by Dr. M.R. Stampfer (University of California, Berkeley, CA). A1N4 were maintained in Improved Minimum Essential Medium with 0.5% fetal bovine serum, 0.5 μg/mL hydrocortisone (Biofluids), 5 μg/mL insulin, and 10 ng/mL epidermal growth factor. 184B5 were maintained in mammary epithelial growth medium (Clonetics-Bio Whittaker).

All antibodies used are listed in Table 1. Enhanced chemiluminescence detection reagent was obtained from Amersham (Arlington Heights, IL). HPLC-purified oligomeric pentameric procyanidin was provided by Masterfoods USA, Hackettstown, NJ.

For breast cancer cell lines, 1 × 10³ cells per well were seeded in 96-well dishes (Corning Costar, Cambridge, MA). Primary cultures of HMECs, A1N4 and 184B5 cells were evaluated at two different cell densities (1 × 10³ and 2 × 10³ per well). After 24 hours, cells in triplicate wells were treated with pentamer at a concentration of 100 μg/mL or equal volume of DMSO vehicle. At indicated time points up to 10 days, cells were stained with crystal violet as described previously (16) and the absorbance at 550 nm was analyzed on an Ultramark Microplate Imaging System (Bio-Rad, Philadelphia, PA). Two to three independent experiments were done in triplicate. The unpaired Student's t test was used to determine significance of differences.

Mitochondrial membrane potential was evaluated using the ApoAlert mitochondrial membrane sensor kit (Clontech, Inc., Palo Alto, CA). Cells were first grown to 60% to 70% confluence in 25 cm² dishes. The next day, the medium was replaced with fresh medium containing pentamer (25 or 100 μg/mL), DMSO or medium only. At the indicated time, floating and adherent cells were combined, and 1 × 10⁶ cells analyzed according to the manufacturer's instructions. Samples were analyzed at the Lombardi Comprehensive Cancer Center, Core Flow Cytometry Shared Resource Facility.

Multilayered dot blot is a recently developed method for screening liquid protein samples in a high-throughput manner using multiwell plates (17–19) and was conducted in collaboration with 20/20 Gene Systems, Inc., (Rockville, MD) as previously described (18). Briefly, MDA MB-231 cells were treated with vehicle or 100 μg/mL pentamer for 48 or 72 hours, and cell pellets lysed in multilayered dot blot lysis buffer (20/20 Gene Systems). Protein concentrations were determined by using a bicinchoninic acid protein assay kit (Pierce, Rockford, IL). Ten and 20 μg of each protein sample were loaded in duplicate into a Bio-Dot Microfiltration Apparatus (Bio-Rad) and transferred through a five-membrane stack as recommended by the manufacturer (20/20 Gene Systems). After transfer, membranes were biotinylated in 1 mg/mL EZ-Link Sulfo-NHS-Biotin solution (Pierce), then incubated overnight at 4°C with 45 primary antibodies (listed in Table 1). After washing, the membranes were incubated with a mixture of FluoroLink Cy 5-labeled streptavidin (Amersham) and fluorescein-labeled secondary anti-mouse or anti-rabbit antibody (Molecular Probes, Eugene, OR).

Protein pellets from vehicle or pentamer-treated MDA MB-231, MDA MB-468, and MCF-7 cells were lysed in PBS with 1% SDS and sonicated, then protein concentrations analyzed as above. Ten micrograms of each protein sample was separated on 4% to 20% gradient gels (Bio-Rad) and transferred to polyvinylidene difluoride (PVDF) membranes (Bio-Rad) or P-FILM membrane (stacks of 10 membranes, 20/20 Gene Systems) according to manufacturer's instructions. Membranes were incubated overnight at 4°C in primary antibodies (Table 1), followed by horseradish peroxidase-conjugated secondary antibody, developed with Enhanced Chemiluminescence Plus reagent (Amersham) and signals visualized on BIOMAX MR film (Kodak, Rochester, NY). glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was visualized by secondary antibody conjugated to alkaline phosphatase DuoLux substrate (Vector Laboratories, Burlingame, CA). PVDF membranes were stripped using a protocol provided by Cell Signaling (Beverly, MA) and reprobed as indicated.

The protein samples used for PVDF membranes (described previously) were used in these experiments. Ten micrograms of each protein sample from vehicle or pentamer (100 μg/mL)–treated cells were separated by PAGE on 4% to 20% gradient gels (Bio-Rad) and transferred onto a stack of 10 multilayered dot blot membranes as above; the antibodies used are listed in Table 1. The secondary antibody used in all cases was anti-rabbit at 1:2,000 dilution (Amersham). Image analysis for both proteins was done with Kodak one-dimensional image analysis software. Background-corrected intensity values were normalized for loading by using GAPDH intensities. Numerical analysis was done in Excel.

Previous studies examined the effects of procyanidins derived from cocoa on the cell cycle of human breast cancer cells (12). The results indicated that the pentamer, but not the lower oligomeric compounds, caused G₀/G₁ cell cycle arrest. The pentameric form was therefore subjected to further study in order to determine whether its inhibitory action was selective for tumor cells and to gain insight into the mechanism(s) responsible for this activity.

We first used the crystal violet assay to evaluate the effects of pentamer on the growth of a panel of human breast cancer cell lines (MDA MB-231, MDA MB-436, MD MB-468, and SKBR3) compared with human immortalized (MCF-10A, 184B5, and A1N4) and normal primary HMECs. All experiments were done using 100 μg/mL of pentamer, a dose shown to be effective in growth inhibition, as previously described (12). The data showed that the growth of MDA MB-231 and MDA MB-436 was significantly inhibited (5.1-fold and 4.6-fold, respectively) compared with DMSO-treated controls (Fig. 1A and B; Table 2). SKBR-3 cells were growth-inhibited 5.5-fold and MDA MB-468 cells were 4.2-fold more growth-inhibited in comparison to DMSO-treated controls (Table 2).

In order to test the selectivity of pentamer for cancer cells, the sensitivity of spontaneously immortalized MCF-10A cells and normal mortal HMECs was tested as described above. Whereas MCF-10 cells were completely resistant to the growth-inhibitory effects of pentamer (Fig. 1C), slight inhibition of normal primary HMEC growth was observed (about 1.45-fold, Fig. 1D). HMECs were 2- to 4-fold less sensitive to the growth-inhibitory effects of pentamer compared with cancer cells (Fig. 1A, B, and D; Table 2).

However, when benzo(a)pyrene immortalized 184B5 cells were tested, they exhibited greater sensitivity (2.4-fold, Table 2) to pentamer than HMECs or MCF-10A. Interestingly, benzo(a)pyrene-immortalized A1N4 cell line exhibited very high sensitivity to pentamer (6.8-fold growth inhibition) compared with controls (Table 2). The level of sensitivity of A1N4 cells was similar to that of human breast cancer cells. The results of pentamer treatment–mediated growth inhibitory effects at day 10 post-treatment on all cell lines tested are summarized in Table 2.

In order to determine whether pentamer resistance of MCF-10A cells and normal HMECs was due to the slow proliferation rate of these cells, we next evaluated the effects of pentamer on the growth of parental MCF-10A cells (population doubling time, 31.4 hours) compared with highly proliferating MCF-10A-c-Myc (MCF-10A cells transduced with c-Myc) with a population doubling time of 20.4 hours (14) and MCF-10A-ErbB2/Ras (with a 4-fold more rapid doubling time compared with parental MCF-10A cells; ref. 13) using crystal violet assay as above. We observed that, similar to parental MCF-10A cells, both MCF-10A-c-Myc and MCF-10A-ErbB2/Ras were resistant to pentamer-induced growth inhibition (data not shown). Based on these results, we conclude that the resistance of MCF-10A to pentamer does not result from slower proliferation rate.

The MDA MB-231, MDA MB-436, MDA MB-468, and SKBR-3 breast cancer cell lines tested (Table 2) all lack estrogen receptor expression and have mutant p53. In order to test whether p53 mutations or estrogen receptor status influence response to pentamer, we analyzed the growth of pentamer-treated estrogen receptor–positive MCF-7 breast cancer cells bearing wild-type p53. As shown in Table 2, pentamer treatment caused significant growth inhibition of MCF-7 as well, indicating that growth-inhibitory effects of pentamer against human breast cancer cells are not determined by the status of estrogen receptor or p53.

Based on the differential sensitivities of breast cancer cells compared with MCF-10A or normal HMECs in response to pentamer, we hypothesized that differences in mitochondrial function between breast cancer cells and their normal or immortalized mammary epithelial cell counterparts might be involved. We first tested the pentamer for its effects on mitochondrial membrane potential of MDA MB231 tumor cells compared with the MCF10A cell line. Preliminary studies indicated that 6-hour treatment achieved maximal mitochondrial membrane potential loss and was used for subsequent studies. MDA MB-231 cells exhibited a high degree of mitochondrial membrane depolarization (52%) in response to pentamer, compared with 5% to 10% in DMSO-treated controls (data not shown). By contrast, pentamer treatment caused depolarization of the mitochondrial membrane in only 19% of the MCF-10A cells, where the DMSO background was only 9% (data not shown). Analysis of other tumor cell lines and the immortalized A1N4 cell line were more variable; although in MDA MB-436, MDA MB-468, and A1N4 cells pentamer also caused significant depolarization of the mitochondrial membrane (5.7-, 5.0-, and 3.2-fold, respectively), neither SKBR-3 nor MCF-7 tumor cells showed consistently significant depolarization in response to pentamer (data not shown). As a result, we are unable to draw a conclusion about the effect of pentamer on mitochondrial membrane potential at this time.

To identify the molecular targets that could provide insight into the mode of inhibitory action of pentamer, we further investigated protein samples of DMSO and pentamer-treated MDA MB-231 human breast cancer cells by proteomic multilayered dot blot analysis (18). The proteomics approach provides information on both expression and activity of proteins. MDA MB-231 cells were selected because they are highly susceptible to the growth-inhibitory effects of the compound (Fig. 1A). The proteins selected as potential targets of pentamer for detailed screenings were those commonly involved in the control of cellular growth, proliferation, survival, and apoptosis (Table 1). A total of 46 different antibodies were used in this screening with emphasis placed on the relative phosphorylation status of potentially active proteins. The results indicated that the phosphorylation status for eight proteins [Cdc2, forkhead transcription factor (FKHR), p53, protein kinase C (PKC)-δ, PKC-[thetas]-, pRb, stress-activated protein kinase/c-Jun-NH₂-kinase, and Stat 5] had been affected by pentamer, whereas total protein levels were unaffected (Fig. 2A and data not shown). We further noted that phosphorylation status for five out of these eight proteins was affected by pentamer within 48 hours (Cdc2, PKCδ, pRb, stress-activated protein kinase/c-Jun-NH₂-kinase, Stat5); decrease ranged from 22% to 36% (data not shown). Phosphorylation was significantly decreased in all eight proteins after 72 hours of pentamer treatment, ranging from 26% (stress-activated protein kinase/c-Jun-NH₂-kinase) to 63% in the case of FKHR (data not shown). Ward hierarchical clustering analysis showed clear segregation between control and treated groups based on the behavior of proteins examined.

Using immunoblotting, we confirmed the multilayered dot blot results on phosphorylated forms of Cdc2 (Tyr¹⁵), FKHR (Ser²⁵⁶), p53 (Ser³⁷, Ser⁴⁶, and Ser³⁹²) and pRb (Ser⁷⁹⁵ and Ser^807/811; Fig. 2B). Total protein expression was not affected by pentamer as evidenced by multilayered dot blot analysis (Fig. 2A), and Western analysis confirmed equal expression of total Cdc2, p53, and FKHR (Fig. 2C). However, levels of both total and phosphorylated pRb were decreased by pentamer (Fig. 2C). Equal loading was confirmed by GAPDH expression (Fig. 2B and C). Using current experimental immunoblotting conditions and PVDF membranes, we could not confirm the multilayered dot blot results for PKC-δ, PKC-[thetas], stress-activated protein kinase/c-Jun-NH₂-kinase, and Stat5.

Both p53 and pRb contain a number of phosphorylation sites, and each site determines specific protein function. Because pentamer elicits a G₀/G₁ cell cycle growth arrest (12) and both p53 and pRb play crucial roles in controlling this phase of the cell cycle, we did high-resolution functional profiling of these two proteins in MDA MB-231 cells treated with 100 μg/mL of pentamer or DMSO vehicle using P-FILM technology (20/20 Gene Systems).

The levels of endogenous p53 expression in MDA MB-231 cells were slightly decreased in pentamer-treated samples after only 48 hours (Fig. 3A). Densitometry suggests that this decrease was not significant (data not shown). Profiling the serine residues of p53 revealed only slight decreases in phosphorylation of Ser¹⁵ residue and Ser⁴⁶ at 48 hours post-treatment (Fig. 3A and B), whereas significant dephosphorylation of p53 Ser³⁹² residue was detected in response to pentamer at both 48 and 72 hours post-treatment (Fig. 3A). Additional confirmation of the p53 results was obtained by traditional immunoblotting (Fig. 3B). Phosphorylation of the Ser²⁰ residue was decreased at 48 hours, but was not affected in samples treated with pentamer for 72 hours (Fig. 3A).

To further evaluate p53 dephosphorylation on Ser³⁹² by pentamer, we next analyzed p53 status in the MDA MB-468 (mutated p53) and the MCF-7 (wild-type p53) cell lines following treatment with pentamer. Similar to MDA MB-231 cells, expression of p53 in MDA MB-468 was not significantly affected by pentamer treatment (Fig. 3B). However, the percentage of p53 phosphorylated at Ser³⁹² was decreased in both MDA MB-231 (53% compared with untreated controls) and MD MB-468 cells (33% compared with untreated controls; Fig. 3B), whereas neither Ser¹⁵ or Ser⁴⁶ was significantly affected in MDA MB-468 cells (Fig. 3B). Under similar experimental conditions, we could not detect any signal for endogenous wild-type p53 or phosphorylated p53 in MCF-7 cells (Fig. 3B). This result was expected because it has been shown that endogenous levels of wild-type p53 are not easily detected by immunoblotting, whereas mutated p53 detection is more readily demonstrable due to its overexpression (20). The equality of loading in all cell lines tested was confirmed by equal expression of GAPDH protein levels (Fig. 3A and B).

In order to perform profiling of pRb phosphorylation in response to pentamer, the samples were analyzed for expression of total pRb and for the phosphorylation status of pRb at Ser⁷⁸⁰, Ser⁷⁹⁵, and Ser^807/811 in MDA MB-231 cells. We found that pentamer treatment caused a decrease in protein expression of endogenous pRb, possibly resulting in the decrease of phosphorylation on all tested residues (Fig. 3C). The expression of GAPDH protein was not changed, indicating equal loading of samples.

The goal of the present study was to test the selectivity of oligomeric pentameric procyanidin from cocoa (pentamer) against a panel of human breast cancer cells and to examine the molecular mechanism(s) by which it causes G₀/G₁ cell cycle arrest in these cells (12). This is, to our knowledge, the first demonstration that pentamer selectively inhibits growth of human breast cancer cell lines independent of their p53 or estrogen receptor status, whereas normal HMECs and nontransformed, immortalized MCF-10A cells are highly resistant. The resistance of normal HMECs to pentamer's cytotoxicity relative to highly sensitive human breast cancer cells is significant because it could serve as a basis for in vivo evaluation of pentamer as a potential chemopreventive and chemotherapeutic agent in both animal models and human disease.

We first evaluated whether pentamer's inhibitory effects involve depolarization of the mitochondrial membrane. The MDA MB-231 cells were highly sensitive, whereas the mitochondria of MCF-10A cells were only slightly affected by pentamer. Similarly, the induction of mitochondrial depolarization by structurally different, small molecular weight polyphenolics have recently been reported (21). However, whereas both the SKBR-3 and MCF-7 tumor cell lines were highly growth-inhibited by pentamer (Table 2), neither consistently showed significant depolarization of the mitochondrial membrane. It has previously been shown that pentamer can exert two distinct effects on cancer cells, growth arrest and cytotoxicity (12), which may contribute to the net growth inhibition observed in all tumor cell lines in our study. The mitochondrial effects caused by pentamer could be expected to trigger either growth arrest or apoptotic or nonapoptotic cell death processes. Indeed, recent studies using the novel mitochondriotoxic small molecule F16 suggested that compounds causing mitochondrial depolarization inhibited tumor growth of breast cancer cells by G₁ cell cycle arrest (22), apoptosis (23), or necrosis (23), depending on the genetic background of the cell. The lack of sensitivity to mitochondrial membrane depolarization in response to pentamer in certain tumor cell lines may therefore reflect the involvement of mitochondrial membrane potential loss in only one of the two inhibitory activities associated with pentamer, growth arrest or cell death. Future studies will be required to precisely address the role of mitochondrial membrane depolarization in the inhibitory action of pentamer.

Similar to breast cancer cells, benzo(a)pyrene-immortalized A1N4 were highly sensitive to pentamer-mediated growth inhibition, whereas 184B5 cells exhibited a lesser degree of sensitivity (Table 2). The underlying cellular basis for pentamer sensitivity of these two lines is unknown. Our results in 184B5 cells are consistent with previously reported sensitivity to growth inhibition of these cells to tea polyphenol (−) epigallocatechin gallate and soy isoflavone genistein (24).

We next sought to identify the molecular events that contribute to the inhibitory effects of pentamer on breast cancer cells. We therefore selected the highly sensitive MDA MB-231 cells for further molecular analysis. Initial multilayered dot blot screen using pentamer-treated MDA MB-231 cells revealed that pentamer decreases phosphorylation of eight proteins involved in control of cell cycle and apoptosis (Fig. 2A and data not shown), although immunoblotting confirmed these results for only Cdc2, p53, pRb, and FKHR (Fig. 2B). Cdc2 can promote both cell cycle progression and apoptosis (25), and its phosphorylation at Tyr¹⁵ is inhibitory for these activities (26); furthermore, ErbB2 phosphorylation of Cdc2 at Tyr¹⁵ causes resistance to taxol-induced apoptosis (26). The pentamer-induced dephosphorylation of Cdc2 at Tyr¹⁵ is therefore of particular interest for its potential to facilitate apoptosis. Similarly, pentamer caused a decrease in the Akt/protein kinase B–dependent phosphorylation site of FKHR (Ser²⁵⁶; Fig. 2A and B). This is of importance, because Akt phosphorylation of FKHR Ser²⁵⁶ is critical to the ability of growth factors to suppress FKHR-dependent transcription, thereby preventing FKHR from inducing cell cycle arrest and apoptosis through transcriptional regulation of p27 (kip1; ref. 27) and FKHR-controlled Fas ligand expression (28–30).

We further showed that in MDA MB-231 and MDA MB-468 cells, pentamer specifically dephosphorylates p53 at Ser³⁹² without affecting the other serine residues we evaluated (Ser⁶, Ser⁹, Ser¹⁵, Ser²⁰, Ser⁴⁶, and Ser³⁷). The phosphorylation of Ser³⁹², in particular, activates specific DNA binding functions by stabilizing p53 tetramer formation (31), and several stress stimuli are reported to phosphorylate Ser³⁹² (Ser³⁸⁹ in mouse; ref. 32). Furthermore, increased phosphorylation of p53 at Ser³⁹² is frequent in human tumors (33, 34) and correlates with increased aggressiveness (33), indicating that pentamer-mediated dephosphorylation of Ser³⁹² may be of importance.

Detailed evaluation of pRb status and its phosphorylation in MDA MB-231 revealed that pentamer decreased the protein expression of pRb itself, a likely explanation for observed decreased phosphorylation of this protein at Ser⁷⁸⁰, Ser⁷⁹⁵, and Ser^807/811, all residues critically involved in G₁-S transition (35, 36). Similar to pentamer, transforming growth factor-β and retinoic acid have been reported to cause G₁ cell cycle arrest by causing dephosphorylation of pRb at Ser⁷⁸⁰, Ser⁷⁹⁵, and Ser^807/811 (37, 38). A more detailed time course of the kinetics of pRb decreased phosphorylation relative to its overall down-regulation requires further study.

In summary, we provide novel evidence that pentamer selectively inhibits proliferation of human breast cancer cell lines. We have further identified a number of critical proteins known to be activated (phosphorylated) in cancer cells whose phosphorylation status has been normalized in the presence of pentamer. Future studies will be aimed at elucidating the precise role of these proteins in pentamer's inhibitory action, particularly p53, which has recently been shown to translocate to mitochondria (39). Of particular interest will be investigations of possible connections between pentamer's effects on cell cycle and its effects on mitochondria; it should be noted that the mitochondriotoxic molecule F16 is also capable of eliciting growth arrest and cell death that is associated with dephosphorylation of specific growth regulatory proteins (22). Furthermore, it will be important to establish the relevance of observed in vitro anticancer effects of pentamer in in vivo preclinical breast cancer models. Such studies will be helpful for the evaluation of pentamer as an effective chemopreventive and chemotherapeutic agent against human breast cancer.

We thank Dr. Karen Creswell of the Lombardi Comprehensive Cancer Center at Georgetown University for performing flow cytometry analyses, technical personnel (Shuangju Lu, Lisa S. Cahill, Feng Zhu, and Victoria A. North) of the Lombardi Comprehensive Cancer Center Tissue Culture Shared Resources (which is partially supported by NIH Grant 3P30-CA-51008), as well as Doug Gallo for technical input. We also thank Drs. James Rae and Marc Lippman (Department of Internal Medicine, University of Michigan, Ann Arbor, MI) for helpful advice early in the course of this study.