Penta-1,2,3,4,6-O-galloyl-β-d-glucose induces p53 and inhibits STAT3 in prostate cancer cells in vitro and suppresses prostate xenograft tumor growth in vivo

Prostate cancer is the most commonly diagnosed cancer that affects one in six American men (1). Notable features of prostate cancer that make chemoprevention a rational and cost-effective approach include its ubiquity and its long latency between premalignant lesions and clinically evident cancers (2). The induction of tumor suppressor signaling such as the p53 axis (3) and/or an inhibition of oncogenic signaling such as signal transducer and activator of transcription 3 (STAT3; refs. 4, 5) in precancerous or cancer cells could be potential molecular mechanisms of prostate cancer primary or secondary chemoprevention.

p53 is the first identified and the best known tumor suppressor that controls cell cycle checkpoints and apoptosis and DNA repair (6). p53 is mutated in over half of all human tumors (7). Activation of p53 signaling through transcriptional and post-transcriptional modifications such as phosphorylation could be an effective approach against precancerous and cancer cells bearing the wild-type p53. Therefore, p53 is considered to be a rational molecular target not only for cancer therapy but also for chemoprevention. Indeed, several phytochemical cancer chemopreventive agents such as epigallocatechin-3-gallate (8), resveratrol (9), and curcumin (10), have been reported to induce apoptosis through the p53-dependent signaling pathway.

In contrast to normal cells, the transformed cells can become “oncogene addicted.” STAT proteins, identified originally as key components of cytokine signaling pathways, are a family of latent cytoplasmic transcription factors that act downstream of Janus-activated kinase or Src kinase activation and mediate intracellular signaling from a wide variety of cytokines, growth factors, and hormones (11, 12). Constitutive activation of STAT3 and other family members has been detected in a wide variety of human tumor specimens and cancer cell lines, including prostate cancer (12). STAT3 is activated by phosphorylation at Tyr⁷⁰⁵, which induces dimerization, nuclear translocation, and DNA binding (13). Transcriptional activation activity of STAT3 is further regulated by phosphorylation at Ser⁷²⁷. Key downstream targets of STAT3 include cyclin D1 for cell cycle regulation, Bcl-XL, and the related Mcl-1 for cell survival (14). It has been shown that, in DU145 cells, which have constitutively active STAT3 (15), inhibition of STAT3 signaling by small interfering RNA (siRNA) or other genetic approaches was sufficient to induce caspase-mediated apoptosis (16, 17). Interleukin-6 (IL-6) is a multifunctional cytokine that regulates the growth and differentiation of many cancers. Elevated expression of IL-6 has been implicated in prostate cancer androgen-independent progression, neuroendocrine differentiation, and drug resistance, which are associated with activation of STAT3 (18–20). Compounds that inhibit STAT3 signaling may therefore be good candidates for chemoprevention of prostate cancer (4).

Polyphenolic compounds are among the various chemopreventive agents that have shown promise for inhibiting prostate cancer carcinogenesis in pre-clinical models (21). Penta-1,2,3,4,6-O-galloyl-β-d-glucose (PGG; Fig. 1A) is a naturally occurring gallotannin polyphenolic compound in Oriental herbs such as Galla rhois, the gallnut of Rhus chinensis Mill, and the root of peony Paeonia suffruticosa Andrews. Several cell culture studies with prostate cancer cells and other cancer cell lines have shown multibiological activities of PGG that confer potential chemopreventive effect. For example, PGG inhibited growth of LNCaP prostate cancer cells and the secretion of androgen-regulated gene product, prostate-specific antigen (PSA; ref. 22). In the test tube, PGG was shown to inhibit rat liver microsomal 5α-reductase activity, which converts testosterone to the more active dihydrotestosterone (22). Such results suggest a potential efficacy of PGG to inhibit the androgen-dependent growth of prostate cancer. In addition to prostate cancer cells, PGG was shown to induce apoptosis of leukemia and breast cancer cells (23) and estrogen receptor signaling (24). PGG also was reported to inhibit invasion-related molecules such as matrix metalloproteinase-9 in melanoma cells (25). A recent study by coauthor Kim's group (26) has revealed that PGG possesses potent antiangiogenic and proapoptotic activities and exerts a strong in vivo antitumor growth effect in a mouse Lewis lung cancer (LLC) allograft model. The signaling mechanisms for these biological effects of PGG still remain poorly defined, although a few in vitro studies showed that it may inhibit activator protein-1 activity in melanoma cells (25), nuclear factor-κB activity in macrophages (27), and cyclooxygenase-2 activities in vascular endothelial cells (26). Although these reported in vitro and in vivo effects suggest a probable anticancer activity of PGG against prostate cancer, its in vivo efficacy has not been established to date.

In this report, we investigated the effects of PGG on p53 and STAT3 signaling in apoptosis induction and established for the first time in vivo efficacy of PGG to inhibit DU145 prostate cancer xenograft growth in an athymic nude mouse model. Our data support the hypothesis that PGG may be an in vivo active prostate cancer chemopreventive agent by targeting multiple critical molecular and cellular pathways.

PGG (molecular weight = 986) was isolated from G. rhois, the gallnut of R. chinensis Mill (26). The purity was ∼98%. The general caspase inhibitor (zVADfmk) was purchased from MP Biomedicals. Antibodies specific for phospho-p53 Ser¹⁵, Bax, p21^Cip1, Bcl-XL, Mcl-1, cleaved poly(ADP-ribose) polymerase (PARP; p89), caspase-3, -8, and -9, cleaved caspase-3 and -9 and phospho-STAT3 were purchased from Cell Signaling Technology. An antibody for cyclin D1 was purchased from Santa Cruz Biotechnologies. N-acetyl-l-cysteine and an antibody for β-actin were purchased from Sigma. Hydroethidine and 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate were purchased from Molecular Probes. Manganese(III) tetrakis(N-methyl-2-pyridyl)porphyrin, a superoxide dismutase mimetic chemical, was purchased from Alexis Biochemicals.

LNCaP and DU145 cell lines were obtained from the American Type Culture Collection. LNCaP cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum without antibiotics. DU145 cells were grown in MEM supplemented with 10% fetal bovine serum without antibiotics. At 24 to 48 h after plating when cells were 50% to 60% confluence, the medium was changed before starting the treatment with PGG or the other agents. To standardize all PGG/drug exposure conditions, cells were bathed in culture medium at a volume to surface area ratio of 0.2 mL/cm² (e.g., 15 mL for a T75 flask and 5 mL for a T25 flask). For the experiments in which caspase inhibitor was used, the inhibitor and PGG were given to the cells at the same time. For the experiment in which antioxidants were used, the cells were exposed to N-acetyl-l-cysteine or manganese(III) tetrakis(N-methyl-2-pyridyl)porphyrin for 2 h before initiating treatment with PGG. DMSO (≤2 μL/mL) was added as a vehicle solvent to the control culture that did not receive the inhibitor. This concentration of DMSO did not cause any adverse morphologic response.

For the evaluation of overall inhibitory effect of PGG on cell number, the cells were treated with PGG daily for 4 days. After treatment, the culture medium was removed and the cells were fixed in 1% glutaraldehyde solution in PBS for 15 min. The fixed cells were stained with 0.02% aqueous solution of crystal violet for 30 min. After washing with PBS, the stained cells were solubilized with 70% ethanol. The absorbance at 570 nm with the reference filter 405 nm was evaluated using a microplate reader (Beckman Coulter).

Apoptosis was assessed by multiple methods. The first was a cell death detection ELISA kit purchased from Roche Diagnostics. The second was Annexin V staining of externalized phosphatidylserine in apoptotic cells by flow cytometry using Annexin V/FITC Staining Kit from MBL International. The third method was immunoblot analysis of PARP cleavage. All these methods were as described previously (28, 29).

The cell lysate was prepared in ice-cold radioimmunoprecipitation assay buffer as described previously (28). Immunoblot analyses were essentially as described (28), except that the signals were detected by enhanced chemofluorescence with a Storm 840 scanner (Molecular Dynamics).

The siRNAs for human p53, p21, or Bak were purchased from Santa Cruz Biotechnologies. The cells were transfected with 20 nmol/L siRNAs using INTERFERin transfection reagent (Polyplus Transfection) for 24 h and then were used for subsequent experiments with PGG treatment.

Generation of intercellular reactive oxygen species (ROS) was measured by flow cytometry following staining with hydroethidine and 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate, which have been shown to specifically detect superoxide and hydrogen peroxide (30, 31), as we have recently described (32, 33). The hydroethidine is a reduced form of ethidium. Upon oxidation by superoxide, red fluorescent ethidium accumulates in the nucleus. 6-Carboxy-2′,7′-dichlorodihydrofluorescein diacetate is a reduced form of 2,7-dichlorofluorescein. Oxidation by hydrogen peroxide can be detected by monitoring the increase in green fluorescence with a flow cytometer. At 30 min before harvest, hydroethidine and 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate were added to the medium to a concentration of 2 and 5 μmol/L, respectively. The cells were collected as described above, and the ethidium and 2,7-dichlorofluorescein fluorescence was measured using a Becton Dickinson flow cytometer.

The animal use protocol was approved by the Institutional Animal Care and Use Committee of the University of Minnesota, and carried out at The Hormel Institute's Animal Facility. PGG was given daily by i.p. injection (20 mg/kg body weight) starting 7 days before s.c. inoculation of cancer cells (“prevention” setting) for 45 days. This dose was chosen based on earlier work with LLC allograft (26). To establish cancer xenograft, 2 × 10⁶ DU145 cells were mixed with Matrigel (50%) (Becton Dickinson) and injected s.c. into the right flank of 7- to 8-week-old male BALB/c athymic nude mice (NxGen BioSciences). Tumors were measured twice per week with a caliper and tumor volumes were calculated using the following formula: 1 / 2 (w₁ * w₂ * w₂), where w₁ is the largest tumor diameter and w₂ is the smallest tumor diameter.

Immunostaining for pSTAT3 Tyr⁷⁰⁵ was done with an antibody from Cell Signaling Technology. Antigen retrieval was done after deparaffinization and rehydration of the tissue sections (4 μm) by microwave for 10 min in 1 mmol/L EDTA. Sections were cooled to room temperature, treated with 3% hydrogen peroxide in methanol for 10 min, and blocked with 6% horse serum for 40 min at room temperature. Sections were then incubated with the primary antibody to phospho-STAT3 (diluted 1:150; Cell Signaling Technology) at 4°C overnight. Sections were washed in PBS and incubated with the secondary antibody (biotinylated goat anti-rabbit; 1:150; Vector Laboratories) for 30 min. After further washes, color was developed by indirect avidin/biotin-enhanced horseradish peroxidase method with the DAKOCytomation ABCcomplex/Horseradish Peroxidase kit and with 3,3′-diaminobenzidine tetrahydrochloride as substrate.

For semiquantitation, 10 representative ×200 power photomicrographs were taken with a digital camera, avoiding gross necrotic areas. The positively stained cancer epithelial cells within each photomicrograph were counted. The counting of total cancer cells was aided with the ImagePro+ image processing program.

Numerical data were expressed as mean ± SE. Statistical analyses were carried out with Prism and Sigma plot softwares, and P < 0.05 was considered statistically significant. The data were analyzed by ANOVA followed by Bonferroni t test for pairwise multiple comparisons or other appropriate tests. When only two groups were in an experiment, Student's t test was used.

To assess the direct inhibitory effect of PGG on prostate cancer cells, we exposed sparsely seeded LNCaP (p53 wild-type, androgen-dependent) and DU145 (mutant p53, androgen independent) cells to daily changes of fresh medium containing PGG (25-75 μmol/L) for 4 days. This schedule simulates a pill-a-day regimen. As shown in Fig. 1A, PGG exerted a strong dose-dependent inhibitory effect on both cell lines. In PC-3 cells (p53-null and androgen-independent), the growth-inhibitory efficacy was identical to DU145 cells (data not shown). These data therefore indicated that PGG inhibited prostate cancer cell growth regardless of their p53 status and androgen dependence status.

To determine whether cell cycle arrests contributed to the growth-inhibitory action of PGG, we analyzed LNCaP and DU145 cells treated with PGG for different duration by flow cytometry (Table 1). We observed a rapid enrichment of G₁ cells at the expense of G₂ cells in LNCaP and DU145 cell lines from 6 to 12 h exposure to 50 μmol/L PGG (Table 1). However, from 24 to 48 h, cells accumulated in the S phase, with near elimination of G₂ cells (Table 1). In PC-3 cells, the same patterns of cell cycle arrests were observed (data not shown). These data indicated that PGG inhibited cell proliferation by arresting cells not only in G₁ phase but also in S phase.

In addition to cell cycle arrests, morphologic observations of PGG-treated DU145 and LNCaP cancer cells suggested that an induction of cell death contributed to the overall growth inhibition. To determine whether cell death induced by PGG involved caspase activation, we analyzed cleavage of PARP using Western blot (Fig. 1B). In LNCaP cells, caspase-mediated cleavage of PARP was time dependent (detected as early as 12 h) and was PGG concentration dependent (48 h time point). Because PGG has been reported to suppress PSA secretion into the conditioned medium of LNCaP cells (22), we analyzed the abundance of PSA and androgen receptor proteins in the cell lysate of PGG-treated LNCaP cells (Fig. 1B). We detected a decrease of androgen receptor and PSA expression at 24 h of PGG treatment. By 48 h, a PGG concentration-dependent decrease of both proteins was observed, but the effects were not dissociable from apoptosis as indicated by cleaved PARP. Therefore, the reported effect of PGG on androgen receptor and PSA (22) was likely secondary to cell death.

The PARP cleavage in LNCaP cells was accompanied by the cleavage activation of the executioner caspase-3 (19-kDa active forms) and the apical caspases of the mitochondria pathway (caspase-9, 37-kDa active form) and death receptor pathway (caspase-8, 43/41-kDa cleavage intermediates; Fig. 1B). In DU145 cells (Fig. 1C), PARP cleavage was delayed in comparison with LNCaP cells. By 48 h of PGG treatment, cleavage of caspase-3, -9, and -8 were detected in a PGG concentration-dependent manner. The results indicated that caspase activation through both intrinsic and extrinsic cascades was involved in PGG-induced apoptosis in these two cell lines, although the time kinetics showed slower activation in DU145 cells than in LNCaP cells.

To test whether caspase activation was critical for cell death induced by PGG in both cell lines, we assessed the effect of a pan-caspase inhibitor on PGG-induced cell death measured by a cell death ELISA kit for oligonucleosomes in the apoptotic cells. As shown in Fig. 1D, 40 μmol/L pan-caspase inhibitor completely blocked PGG-induced cell death in both DU145 and LNCaP cells. This conclusion was further confirmed by Annexin V staining assay (data not shown). Taken together, PGG induced predominantly caspase-mediated apoptosis in LNCaP (p53 wild-type and androgen-dependent) and DU145 (mutant p53, androgen-independent) prostate cancer cells.

It has been shown that the cytotoxic effects of some polyphenolic compounds such as gallic acid and green tea polyphenol epigallocatechin-3-gallate are due to hydrogen peroxide generated by these compounds in the culture medium (34–36). Because PGG is a glucose ester of five gallic acid molecules, we compared the production of hydrogen peroxide in the cell-free medium by PGG with gallic acid in equal molar concentration of phenol groups. As shown in Fig. 2A, gallic acid induced a significant production of hydrogen peroxide, peaking at 1 h (squares), which was quenched by catalase added to the cell culture medium (triangles). PGG did not induce cell-free hydrogen peroxide production within the time frame shown (Fig. 2A) and throughout 24 h (data not shown). These data support PGG as a distinct entity from gallic acid or epigallocatechin-3-gallate to induce cell cycle G₁ and S arrests (Table 1) and caspase-mediated apoptosis. Indeed, it has been reported that, in DU145 cells, gallic acid induces S-phase arrest (6-12 h), progressing to G₂-M arrest (24 h) with associated changes in DNA damage response/ATM activation and checkpoint kinase Chk2 activation (37).

In LNCaP cells, as in many other cell types, sustained activation of p53 is often associated with caspase-mediated apoptosis (38). We examined whether the observed apoptosis effect of PGG in LNCaP cells was accompanied by an activation of this pathway. As shown in Fig. 2B (left), PGG induced a rapid and persistent p53 Ser¹⁵ phosphorylation in LNCaP cells, some 6 h ahead of the up-regulation of p21^Cip1 protein, which is the best known transcriptional target of p53 (39). PGG also decreased the abundance of cyclin D1 as early as 6 h of exposure and persisted through 48 h, suggesting a possible mechanism for G₁ arrest due to a consequential decrease of CDK4/CDK6 activities to drive G₁ cell cycle progression. Activation of p53 through Ser¹⁵ phosphorylation by PGG correlated with the induction of not only p21^Cip1 but also Bax, which is another known transcriptional target of p53 (40), in a PGG concentration-dependent manner at 24 h of treatment (Fig. 2B, right). These results indicated that the transcriptional activity of p53 was enhanced by PGG treatment.

To determine the role of p53 activation in PGG-induced apoptosis in LNCaP cells, we tested the effect of attenuating p53 by siRNA. As shown in Fig. 2C (left), p53 siRNA caused a significant reduction of p53 protein expression, which correlated with an inhibition of p21^Cip1 (lane 4 versus lane 2) and the specificity of p21^Cip1 down-regulation was verified by p21 siRNA (lane 3). Knocking down of p53 by siRNA led to a >50% decrease of PGG-induced apoptosis in LNCaP cells (Fig. 2C, graph). The role of p53 in PGG-induced apoptosis was further tested by an ectopic expression of a dominant-negative mutant p53 DN-P151S (32, 41). The inhibition of p53 transcriptional function by dominant-negative mutant was evident by the reduced p21^Cip1 induction with PGG treatment (Fig. 2D, left, lane 4 versus lane 2). The attenuation of p53 activation led to a 45% reduction of PGG-induced apoptosis in DN-P151S cells compared with that in vector control cells (Fig. 2D, graph). The results from both types of genetic manipulation supported a major (but not sole) mediator role for p53 in caspase-mediated apoptosis by PGG in LNCaP cells.

We hypothesized that the activation of p53 and caspase-mediated apoptosis by PGG in LNCaP cells might involve ROS generation. We measured intracellular ROS by flow cytometry following 2′,7′-dichlorofluorescein (preferentially detecting hydroperoxide) and DHE (preferentially detecting superoxide) staining as described in Materials and Methods. A significant increase of DHE-positive cells was observed at 6 h in LNCaP cells after PGG treatment (Fig. 3A) followed by a strong increase of 2′,7′-dichlorofluorescein-positive cells at 24 h (Fig. 3B).

We tested the effects of two antioxidants on apoptosis induction by PGG (Fig. 3C). The apoptosis induction by PGG was significantly reduced by either manganese(III) tetrakis(N-methyl-2-pyridyl)porphyrin, a cell-permeable superoxide dismutase mimetic, or N-acetyl-l-cysteine, a precursor of intracellular glutathione synthesis and ROS scavenger. Attenuation of ROS production by either manganese(III) tetrakis(N-methyl-2-pyridyl)porphyrin or N-acetyl-l-cysteine significantly decreased PGG-induced p53 Ser¹⁵ phosphorylation and PARP cleavage (>90%; Fig. 3D). Taken together, the results support a critical role of intracellular ROS generation in PGG-induced p53 activation and caspase-mediated apoptosis in LNCaP cells.

Because DU145 cells contain nonfunctional mutant p53 (42), PGG must induce caspase-mediated apoptosis also through p53-independent targets and pathways. To test whether STAT3 might be targeted by PGG in DU145 cells, we analyzed the effect of PGG on STAT3 phosphorylation status (Fig. 4A). PGG treatment caused a rapid (as early as 6 h) and persistent (throughout 48 h) inhibition of STAT3 Tyr⁷⁰⁵ phosphorylation, ahead of PARP cleavage by >24 h (Fig. 4A). Furthermore, inhibition of phospho-STAT3 by PGG dose-dependently correlated with a reduction of Bcl-XL, a key mitochondria protective protein of the Bcl-2 family and of cyclin D1, both being transcriptional targets of STAT3 (ref. 14; Fig. 4B).

Whereas most articles have reported that LNCaP cells do not contain constitutively active STAT3 (43, 44), some have reported the contrary (e.g., ref. 16). Presumably, such discrepancy is due to variations in cell passage status and culture conditions among different laboratories. We did not observe phosphorylation of STAT3 in LNCaP cells under conditions where DU145 cells showed strong STAT3 Tyr⁷⁰⁵ phosphorylation (data not shown). We therefore tested whether IL-6-induced STAT3 activation could be inhibited by PGG in LNCaP cells. IL-6 caused a rapid (30 min; Fig. 4C) and persistent (24 and 48 h) STAT3 Tyr⁷⁰⁵ phosphorylation (Fig. 4D) and induction of Bcl-XL and Mcl-1, which are known transcriptional targets of STAT3 (14). Treatment with PGG led to a significant decrease of IL-6-induced STAT3 Tyr⁷⁰⁵ phosphorylation (Fig. 4C and D) and Bcl-XL/Mcl-1 expressions in LNCaP cells (Fig. 4D). These results suggest that PGG was capable of inhibiting both constitutive and IL-6-induced STAT3 Tyr⁷⁰⁵ phosphorylation in prostate cancer cells, contributing to apoptosis signaling and cell cycle arrests.

To evaluate the functional significance of down-regulation of STAT3-target Bcl-XL in PGG-induced caspase-mediated apoptosis, we tested the effect of an overexpression of Bcl-XL on apoptosis induction by PGG in DU145 cells (Fig. 5A). Apoptosis by PGG in the DU145 cells overexpressing Bcl-XL were decreased by >72% in comparison with the vector control cells (55.7% versus 15.5%). Western blot analysis showed a corresponding decrease of PARP cleavage (Fig. 5A), supporting decreased Bcl-XL contributed critically to PGG-induced caspase-mediated apoptosis in DU145 cells.

Down-regulation of Bcl-XL and Mcl-1 through suppression of STAT3 activity by PGG could be expected to free their binding partners Bak and Bax, which in turn lead to mitochondrial permeability change and the activation of caspases in the intrinsic pathway. We tested the effect of knocking down Bak on PGG-induced apoptosis in DU145 cells, which do not express Bax (Fig. 5B). Knocking down of Bak decreased PGG-induced PARP cleavage by half. These results implicated mitochondria-dependent caspase activation to play a major role in PGG-induced apoptosis.

The strong growth-inhibitory effects of PGG against these prostate cancer cells prompted us to evaluate the in vivo growth-inhibitory effect of PGG using the DU145 xenograft model. PGG was given by i.p. injection (20 mg/kg body weight) starting 7 days before s.c. inoculation of cancer cells (“prevention” setting). This dose was chosen based on earlier work with mouse LLC allograft model by coauthor Kim's group (26). As shown in Fig. 6A, treatment with PGG led to a significant inhibition of tumor growth (ANOVA, P < 0.01) and decreased the final tumor weight by 53% (P < 0.05). The PGG-treated group did not show decreased body weight of the mice (Fig. 6B). Immunohistochemistry staining of excised tumor sections showed strong nuclear localization of phospho-STAT3 Tyr⁷⁰⁵ (Fig. 6C). The percent nuclei positive for phospho-STAT3 staining was decreased by one third (Fig. 6C). Western blot analyses of selected frozen tumors showed the same trend of decreased phospho-STAT3 abundance in PGG-treated group (Fig. 6D). Overall, the data suggest that in vivo targeting of phospho-STAT3 by PGG contributed to its in vivo efficacy against DU145 prostate cancer xenograft growth.

Given carcinogenesis is a multistage dynamic process involving alterations of numerous oncogenic signaling pathways and inactivation of tumor suppressor pathways (45), the lack of success with single-target approach for treatment and prevention is not surprising. Development of agents that can affect diverse molecular targets and cellular processes is urgently needed to improve prostate cancer chemoprevention and therapy. Here, we have shown that PGG might be one such novel agent that induces caspase-mediated apoptosis as well as G₁- and S-phase arrests in prostate cancer cells of diverse p53 and androgen dependence status. The decreased cyclin D1 in PGG-treated LNCaP cells (Fig. 2A) and DU145 cells (Fig. 4B) could be important to account for the rapid induction of G₁ arrest by PGG through diminished CDK4/CDK6 activity, which requires cyclin D1 (46). The mechanisms responsible for the G₁ and S arrests are currently under investigation.

We provided strong evidence that PGG could activate the p53 tumor suppressor signaling pathway and inhibit the STAT3 oncogenic signaling cascade to induce caspase-mediated apoptosis in prostate cancer cells. In particular, we found that PGG induced a rapid and persistent p53 phosphorylation (Fig. 2), which was temporally hours ahead of PARP cleavage indicative of the activation of caspases (Fig. 1). The genetic approach with either siRNA (Fig. 2C) or dominant-negative mutant (Fig. 2D) to inactivate p53 signaling supports a major role (although not a sole mediator) of p53 activation in PGG-induced apoptosis in LNCaP cells. We identified ROS generation as a major mediator of p53 activation and apoptosis because two antioxidants effectively blocked p53 activation and caspase activation and overall apoptosis (Fig. 3). Taken together, our results support the induction of intracellular ROS by PGG leads to p53 activation, which in turn stimulates Bax and mitochondria-dependent activation of caspases as a predominant apoptosis signaling pathway in LNCaP cells. Apoptosis induction by PGG in LNCaP cells is faster than DU145 cells and may be associated with a rapid p53 phosphorylation.

The growth-inhibitory effect of PGG on LNCaP cells was similar to that on DU145 and PC-3 cells, which do not have functional p53 (42), suggesting that PGG possesses additional targets. We focused on STAT3 for several reasons: (a) Janus-activated kinase/STAT3 activation is involved in the development of hormone-refractory prostate cancer by either ligand-independent activation of androgen receptor or through direct transcriptional mechanisms (18, 47, 48). (b) DU145 and PC-3 cells are androgen-independent prostate cancer cells with constitutively active STAT3 (16). (c) Block of constitutive STAT3 signaling results in growth inhibition and apoptosis of androgen-independent prostate cancer cells in vitro and in vivo (16, 17), supporting their dependence on and addiction to constitutive STAT3 signaling for growth and survival. In this study, we found that PGG inhibited constitutively active STAT3 in androgen-independent DU145 prostate cancer cells (Fig. 4A), which was correlated with down-regulation of its transcriptionally regulated targets such as Bcl-XL and cyclin D1 (Fig. 4B). Given the important role of IL-6-signaling in STAT3 activation and prostate cancer progression to androgen independence (18, 43), we tested the effect of PGG on IL-6-induced STAT3 activation using LNCaP cells, which do not have constitutively active STAT3. We found that PGG also has the capacity of inhibiting IL-6-induced STAT3 activation and Bcl-XL expression in LNCaP cells (Fig. 4C and D).

To further study the functional role of inactivation of STAT3/Bcl-XL axis in PGG-induced apoptosis, we tested the effect of overexpression of Bcl-XL or knockdown of Bcl-XL binding partner Bak on apoptosis by PGG (Fig. 5). The results of both manipulations support the important role of STAT3/Bcl-XL-mediated mitochondrial activation in PGG-induced apoptosis in DU145 cells. Together, these data show, for the first time, PGG as a potent anti-STAT3 agent that targets both constitutive and IL-6-induced STAT3 activation (Tyr⁷⁰⁵ phosphorylation) in human prostate cancer cell lines. Furthermore, an in vivo inhibitory effect of PGG on STAT3 activation was observed in xenograft tumors (Fig. 6C and D). How PGG decreases STAT3 phosphorylation will be investigated in future work.

The xenograft study is, to our knowledge, the first to show the in vivo efficacy of PGG against the growth of human prostate cancer cells (Fig. 6A). The data confirmed the in vivo efficacy of PGG against the mouse LLC allograft model (26) and extended it by showing that the anti-prostate cancer activity did not require immunocompetence because of the nude mice are severely immunocompromised. In terms of the route of PGG administration, we used i.p. injection to deliver PGG in the current study as in the LLC model (26) for proof of principle. To address the issue of oral bioavailability of PGG, we have in a separate study with the LLC allograft model tested the efficacy of oral gavage of 10 mg PGG/kg body weight starting 12 days after cancer inoculation. We observed a 60% inhibition of tumor weight after 7 days of PGG therapeutic treatment (final tumor weight in g/mouse: vehicle 1.35 ± 0.25 versus PGG 0.55 ± 0.15; n = 7 mice; t test, P < 0.001). The demonstration of oral efficacy is important for the practical use of PGG as a chemopreventive agent because oral intake is a noninvasive means to deliver PGG to internal target organ sites such as the prostate.

In summary, PGG can activate p53 tumor suppressor and inactivate STAT3 oncogenic signaling to mediate its proapoptotic and antiproliferative effects in prostate cancer cells. These actions may, at least in part, account for the inhibition of human xenograft tumor growth in vivo. Taken together, our data strongly suggest that PGG has an excellent potential for prostate cancer chemoprevention or treatment.

All authors have no personal or financial conflict of interest and have not entered into any agreement that could interfere with our access to the data on the research or on our ability to analyze the data independently, to prepare articles, and to publish them.

We thank Prof. Peter Daniel (Humboldt University) for the stable DU145 cell lines transfected with the pIRES-neo-Bcl-xL or the pIRES-neo control plasmid, Prof. Ralph W. deVere White (Department of Urology, University of California) for generously providing the p53 dominant-negative mutant LNCaP cells, Todd Schuster for performing flow cytometry and Annexin V staining, the animal care staff of the Hormel Institute Animal facility for excellent animal care and health monitoring, and Andria Hansen for submission of the article.