Resveratrol inhibits Src and Stat3 signaling and induces the apoptosis of malignant cells containing activated Stat3 protein

Resveratrol (trans-3,5,4′-trihydroxystilbene), a naturally-occurring phytoalexin with antioxidant and antiinflammatory properties, is abundant in red grapes and grape products and is believed to confer cardioprotective effects associated with red wine consumption (1). Initial studies show growth inhibition of human breast cancer (2), oral cancer (3, 4), prostate cancer (5, 6), and leukemia cell lines (7), blockade of angiogenesis, and protection against carcinogenesis (8, 9) by resveratrol, suggesting the potential anticancer properties of this compound. The molecular mechanisms for these resveratrol biological effects are not fully defined, although a few reports indicate that it interferes with the mitogen-activated protein kinase and protein kinase C pathways (7, 9–13), inhibits ribonucleotide reductase and DNA synthesis (14), and disrupts cyclooxygenase activity (15).

The signal transducer and activator of transcription (STAT) family of cytoplasmic proteins is important for promoting the proliferation, survival, and other biological processes triggered by cytokines and growth factors (16–18). STAT activation is initiated upon phosphorylation of a critical tyrosine residue by growth factor receptors, Janus kinases, or the Src family kinases, and involves dimerization between two phosphorylated STAT monomers followed by translocation of the dimers into the nucleus. In the nucleus, STAT dimers bind to specific DNA response elements in the promoters of target genes and regulate their expression. Whereas normal STAT activation is highly regulated and transient, one member of the STAT family, Stat3, is constitutively activated in diverse human tumors largely due to hyperactive tyrosine kinases. Constitutively-active Stat3 induces oncogenic processes, such as dysregulated growth, survival, angiogenesis, and immune modulation, and thereby contributes to malignant transformation and progression (18–22). Given that the Src tyrosine kinase activates Stat3 and the collective roles of both Src and Stat3 in many human tumors (23–25), and the potential for resveratrol to modulate signal transduction pathways, we investigated the effects of resveratrol on Src-Stat3 signaling. We report that resveratrol inhibits both Src tyrosine kinase and downstream Stat3 activation in malignant cells and induces cell growth inhibition and apoptosis of malignant cells harboring constitutively-active Stat3.

Normal mouse fibroblasts (NIH3T3) and their v-Src-transformed counterparts (NIH3T3/v-Src) have been previously described (23, 26). The human breast cancer MDA-MB-231, MDA-MB-453, MDA-MB-468, pancreatic cancer Panc-1 and Colo-357, and prostate cancer LNCaP and DU145 cell lines were obtained from the American Type Culture Collection (Manassas, VA). Cells were grown in DMEM or in RPMI 8226 with 10% heat-inactivated fetal bovine serum. Trans-resveratrol, RNase, DMSO, propidium iodide, and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay kit were from Sigma-Aldrich (St. Louis, MO), and terminal nucleotidyl transferase–mediated nick end labeling assay kit was from BD Biosciences PharMingen (San Diego, CA). Each treatment condition is a single dose of resveratrol at the indicated concentration or 0.1% DMSO (vehicle) as control.

Nuclear extracts were prepared from cells and used for electrophoretic mobility shift assay, as previously described (23, 25–28). In some cases, cells were pretreated with resveratrol (30–50 μmol/L, 24 hours) prior to harvesting for nuclear extract preparation. Nuclear extracts (normalized for equal total protein) were incubated with the ³²P-labeled oligonucleotide probe hSIE (high-affinity sis-inducible element, m67 variant, 5-AGCTTCATTTCCCGTAAATCCCTA) that binds both Stat1 and Stat3 (28, 29). Supershift assays were done with rabbit polyclonal antibodies specific for COOH-terminal amino acid residues of Stat3 (750–769) or Stat1 (685–710; Santa Cruz Biotechnology, Santa Cruz, CA). Protein-DNA complexes were resolved by nondenaturing gel electrophoresis and detected by autoradiography.

Cell lysates were prepared with ice-cold radioimmunoprecipitation assay buffer, clarified by centrifugation, and protein concentrations of supernatants were determined by Bio-Rad protein assay (Hercules, CA). Equivalent amounts of total cellular proteins were separated by SDS-10% PAGE and then transferred onto nitrocellulose membranes. Probing of nitrocellulose membranes with primary antibodies and detection of horseradish peroxidase–conjugated secondary antibodies by enhanced chemiluminescence (Amersham, Piscataway, NJ) were done as previously described (24, 26, 27). The probes used were antibodies against Mcl-1, Bcl-x_L, cyclin D1, and Stat3 (Santa Cruz Biotechnology), pY-Stat3 and pY-Src (Cell Signaling Technology, Beverly, MA), and Src GD11 monoclonal antibody (Upstate Biotechnologies, Waltham, MA).

NIH3T3/v-Src cells were harvested, lysed in radioimmunoprecipitation assay buffer, and protein concentrations were determined. Equal amounts of total protein (150 μg) were incubated for 2 hours with 5 μL of Src monoclonal antibody (Upstate Biotechnologies). Protein G-agarose was added and incubation continued for an additional 4 hours. Immunoprecipitated proteins were washed thrice in radioimmunoprecipitation assay buffer and once in kinase buffer (0.1 mol/L HEPES, pH 7.5, 0.02 mol/L MgCl₂, 0.01 mol/L MnCl₂, and 0.02 mol/L DTT). Equal amounts of immunoprecipitated proteins were incubated in kinase reaction mixture containing kinase buffer, 2 μg denatured enolase (Sigma-Aldrich), 10 μmol/L cold ATP, and 2 μCi (1 Ci = 37 GBq) [γ-³²]ATP, and incubated at 30°C for 30 minutes. The reaction was stopped with gel loading buffer and the proteins were separated by 10% PAGE and subjected to autoradiography.

The indirect peroxidase-antiperoxidase test was done on cytospins prepared from cell lines (control and treated with 30 or 50 μmol/L resveratrol for 24 hours). Immunostaining for pY-Stat3 was done with a rabbit antihuman polyclonal pY-Stat3 antibody (Cell Signaling Technology) with rabbit immunoglobulins as negative controls using avidin-biotin-peroxidase complexes (Vector Laboratories, Burlingame, CA). After incubation and blocking of endogenous peroxidase and nonspecific background staining in 3% hydrogen peroxide and methanol for 20 minutes, slides were washed with PBS, treated with 1.5% normal serum and 3% bovine serum albumin, and incubated with the pY-Stat3 primary antibody overnight at 4°C. Slides were rinsed in PBS and incubated with a biotinylated secondary antibody (Vector Laboratories) for 60 minutes, followed by incubation with avidin-biotin-peroxidase complex for 1 hour at room temperature. Chromagen was developed with Nova-Red. All slides were counterstained with hematoxylin for 30 seconds before dehydration and mounting. PhosphoY-Stat3-positive cells were stained red (due to Nova-Red) and pY-Stat3-negative cells were stained blue (due to hematoxylin).

Proliferating fibroblasts treated (0–50 μmol/L resveratrol) or untreated (0.1% DMSO) for different times were trypsinized and counted by phase-contrast microscopy for viable cells (using trypan blue exclusion). In other studies, resveratrol-containing medium was removed 48 hours after incubation and replaced with fresh resveratrol-free medium for 24 hours. Cell numbers were then counted as above. For MTT assay, the seeding density of the breast, pancreatic, and prostate cancer cell lines ranged from 8 × 10³ to 2 × 10⁴/well in 96-well plates. After 12 hours of adherence, cells were treated with 0 to 100 μmol/L of resveratrol for different times, then exposed to the MTT dye (5 mg/mL) and incubated at 37°C for 3 hours. The resulting formazan crystals were solubilized with DMSO and the absorbance of each well was measured at 540 nm using a multiscanner autoreader (Dynatech MR 5000, Chantilly, VA).

Cells (mouse fibroblasts, human breast, pancreatic, and prostate cancer) were treated with 0 to 50 μmol/L of resveratrol or with DMSO for the indicated times. Cells were then fixed with 70% cold ethanol and analyzed after propidium iodide staining. DNA content was determined by a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Cell cycle phase distribution was analyzed using the Cell-Fit program. Data acquisition was gated to exclude cell doublets. For resveratrol recovery experiments, NIH3T3 cells were treated with 20 μmol/L resveratrol or DMSO (control) for 72 hours. Cells were then washed twice with fresh DMEM and cultured in fresh DMEM for 24 hours at 37°C, processed and subjected to cell cycle analysis as described above.

Human breast, pancreatic, and prostate cancer cells, were treated with 0 to 50 μmol/L of resveratrol or with DMSO (control) for 72 hours. Cell suspensions were fixed with 1% paraformaldehyde and 70% ethanol, washed with the APO-DIRECT Wash Buffer (BD Biosciences PharMingen), and incubated with terminal nucleotidyl transferase enzyme and FITC-deoxyuridine 5-triphosphate, and analyzed by flow cytometry. Normal mouse fibroblasts (NIH3T3) and v-Src-transformed counterparts (NIH3T3/v-Src) were treated with or without the caspase inhibitor, zVAD (50 μmol/L; MP Biomedicals, LLC, Solon, OH), for 2 hours prior to being treated with 20 μmol/L resveratrol for 36 hours. Cells were then harvested, stained with Annexin V and 7-amino-actinomycin D (BD Biosciences PharMingen) according to the manufacturer's protocol, and analyzed by flow cytometry to quantify the percentage of apoptosis.

Previous reports suggested that resveratrol interferes with Src tyrosine kinase activity (9, 13). We investigated the effects of resveratrol on Src activation by subjecting v-Src-transformed mouse fibroblasts (NIH3T3/v-Src) to resveratrol treatment (5–20 μmol/L resveratrol for 1–24 hours) and preparing cell lysates for SDS-PAGE and immunoblotting with specific antibodies against phosphorylated Src (pY-Src) and Src protein. Compared with levels in control (DMSO-treated) cells, treatment with resveratrol diminished pY-Src level in a time- and dose-dependent manner, whereas total Src protein levels were not affected in treated cells (top, Fig. 1A). Reduction of pY-Src levels was evident by 1 hour for 20 μmol/L resveratrol [lane 4, Fig. 1(ii)] and persisted for up to 24 hours, suggesting the direct inhibition of Src by resveratrol. In in vitro kinase assay with exogenous enolase substrate and Src protein immunoprecipitated from lysates of resveratrol-treated cells, we confirm that resveratrol inhibited intracellular Src activity (Fig. 1B). These results show that resveratrol directly inhibits Src tyrosine kinase activity.

We have previously shown that Stat3 is constitutively-activated in v-Src-transformed mouse fibroblasts (NIH3T3/v-Src) and breast carcinoma cell lines in part due to Src kinase activity (23, 25, 26). To investigate if inhibition of Src activity by resveratrol modulates Stat3 phosphorylation, malignant cells were treated with resveratrol and were examined for Stat3 activation. In v-Src-transformed mouse fibroblasts (NIH3T3/v-Src), treatment with resveratrol results in time- and dose-dependent decreases in pY-Stat3 but not total Stat3 levels, as measured by Western blot analysis (bottom, Fig. 1A). The decreases in preexisting pTyr-Stat3 levels paralleled reduction in pY-Src levels and occurred as early as 1 hour after resveratrol treatment and persisted for >24 hours (Fig. 1A; data not shown). While this article was being reviewed, a study was published which showed that resveratrol inhibits Stat3 phosphorylation in endothelial cells in response to interleukin-6 (30), which is partly consistent with our findings. Reduction in pTyr Stat3 by resveratrol is further confirmed by immunohistochemical staining in NIH3T3/v-Src and in human prostate and pancreatic cancer cell lines that harbor constitutively-active Stat3 (Fig. 2B). Furthermore, electrophoretic mobility shift assay analysis of nuclear extracts prepared from cells shows a decrease in Stat3 DNA-binding activity in NIH3T3/v-Src, and in human breast, prostate, and pancreatic tumor cell lines harboring constitutively-active Stat3 treated with resveratrol (Fig. 2A). In supershift analysis, Stat3 DNA-binding activity was blocked and supershifted by anti-Stat3 antibody but not by anti-Stat1 antibody (Fig. 2A), confirming that the protein/DNA complex contained Stat3. These results show that resveratrol inhibits Src tyrosine kinase activity and downstream Stat3 signaling in malignant cells.

Given that Src and Stat3 are involved in the growth and viability of malignant cells (25, 26, 31, 32) and previous reports that resveratrol inhibits the growth of human tumor cells (2–7), we examined the effects of resveratrol on v-Src-transformed fibroblasts (NIH3T3/v-Src) as well as human breast cancer MDA-MB-231, MDA-MB-468, prostate cancer DU145, and pancreatic cancer Colo-357 and Panc-1 cell lines that harbor constitutively-active Stat3. Proliferation and viability were measured by trypan blue exclusion with phase-contrast microscopy (Fig. 3A-D) and MTT assay (Fig. 3E-G), respectively. In malignant cells harboring activated Stat3 (NIH3T3/v-Src, MDA-MB-231, MDA-MB-468, DU145, Panc-1, and Colo-357), resveratrol induced dose-dependent growth inhibition and loss of viability (Fig. 3B, C, E-G) with extensive fragmentation of cells (data not shown). By contrast, only decreased growth rate with no significant loss of viability is observed with normal mouse fibroblast (NIH3T3, Fig. 3A and C; ref. 26), human breast carcinoma (MDA-MB-453, Fig. 3E; ref. 25) prostate cancer (LNCaP, Fig. 3F; ref. 33) cell lines that do not harbor constitutively-active Stat3 at 5–50 μmol/L of resveratrol treatment. Removal of resveratrol restores the growth rate to normal, as observed for normal NIH3T3 (fifth and eight columns, Fig. 3D; data not shown), suggesting that the decrease in growth rate was reversible. The IC₅₀ values for loss of viability reveal malignant cells harboring aberrant Stat3 signaling which are two to three times more sensitive to resveratrol than cells that do not (Table 1). Together, these data indicate that resveratrol inhibits Src kinase activity and Stat3 activation, and results in loss of viability of malignant cells that harbor aberrant Stat3 signaling.

Both Src and Stat3 proteins have key roles in cell cycle progression, whereas inhibition of constitutively-active Stat3 induces cell cycle arrest (25, 31). Flow cytometric analyses were done to further probe the malignant cell growth inhibition induced by resveratrol. In v-Src-transformed mouse fibroblasts, treatment with resveratrol induces an irreversible cell cycle block and accumulation of cells at the G₀-G₁ phase by 24 hours, compared with nontreated cells (Fig. 4B). This is in contrast with a reversible block at the G₀-G₁ phase of normal NIH3T3 fibroblasts treated with resveratrol (Fig. 4A), which is reversed after the removal of resveratrol (bottom, Fig. 4A). By comparison, the irreversible block at G₀-G₁ in the resveratrol-treated v-Src-transformed fibroblasts precedes an increased sub-G₁ peak and decreased G₁, S, and G₂-M phases at 48 to 72 hours, suggesting DNA fragmentation and apoptosis (Fig. 4B). In human tumor cell lines that harbor constitutively-active Stat3, resveratrol treatment induces cell cycle arrest at the G₀-G₁ phase with concomitant decrease of cells at the S or G₂-M phase of human breast cancer (MDA-MB-231), pancreatic cancer (Panc-1), and prostate cancer (DU145) cell lines or an arrest at the S phase with a concomitant decrease of cells in the G₂-M phase of human breast cancer (MDA-MB-468) and pancreatic cancer (Colo-357) cell lines (Table 2). Although all these cell lines have activated Stat3, we observed differential patterns of change in cell cycle phase distribution in response to resveratrol. At the concentrations used, resveratrol treatment, however, does not significantly affect the cell cycle profile of LNCaP cells which do not harbor Stat3 activity (Table 2), whereas the breast carcinoma MDA-MB-453 cell line is blocked at the S phase (Table 2). These findings suggest that resveratrol-induced inhibition of Src-Stat3 signaling leads to irreversible growth arrest of malignant cells, consistent with the role of aberrant Stat3 signaling in cell cycle and growth control.

With the known importance of Src and Stat3 in survival, we further examined the effects of resveratrol on the survival of malignant cells. Analyses by terminal nucleotidyl transferase–mediated nick end labeling show that treatment with resveratrol induces apoptosis in the breast (MDA-MB-231 and MDA-MB-468), pancreatic (Colo-357 and Panc-1), and prostate (DU145) cancer cell lines that harbor constitutive Stat3 activity, in contrast with the minimal apoptosis observed in cell lines (MDA-MB-453 and LNCaP) that lack aberrant Stat3 signaling (Fig. 5A-C). Furthermore, Annexin V staining with flow cytometry analysis show that resveratrol selectively induces apoptosis in v-Src-transformed mouse fibroblasts (NIH3T3v-Src) which harbor constitutively-active Stat3 but not in normal mouse fibroblasts (NIH3T3, Fig. 5D). The apoptosis of NIH3T3v-Src fibroblasts is partially blocked by a general caspase inhibitor, zVAD (Fig. 5D), suggesting that resveratrol-induced apoptosis may in part be mediated through the caspase pathway, as previously reported (34).

To further understand the mechanism by which resveratrol-mediated inhibition of constitutive Stat3 activation leads to cell growth inhibition and apoptosis, we examined the cell cycle regulator, cyclin D1, and the antiapoptotic proteins Bcl-x_L and Mcl-1, all of which are target genes of Stat3. In the human prostate cancer DU145 cell line and v-Src-transformed mouse fibroblasts (NIH3T3/v-Src) treated with resveratrol, Western blot analysis shows significant reductions in the cyclin D1 as well as Bcl-x_L and Mcl-1 expression levels (Fig. 6), which correlate with the inhibition of Stat3 phosphorylation and activity by resveratrol (Figs. 1A and 2A and Fig. 6). The activation of mitogen-activated protein kinases, p42/44, is not altered by resveratrol treatment (Fig. 6). These data suggest that resveratrol induces cell growth inhibition and apoptosis of tumor cells at least in part by inhibiting Stat3-depedent induction of cyclin D1, Bcl-x_L, and Mcl-1.

Current studies provide new insights into the biological and molecular mechanisms for the antitumor effects of resveratrol and related stilbene compounds (35). We show that resveratrol inhibits the Src kinase activity, as previously indicated (9, 12) and represses downstream aberrant Stat3 signaling in malignant cells. Inhibition of Src and Stat3 by resveratrol is rapid, occurring in the early hours after treatment. Accordingly, resveratrol induces cell cycle arrest and apoptosis of malignant cells harboring constitutively-active Stat3. Similar biological effects of resveratrol have been reported in the androgen-independent prostate cancer (DU145 or PC-3; refs. 5, 10, 35, 36) and glioma (U251; ref. 37) cell lines, which harbor constitutively-active Stat3 (33, 38). By contrast, minimal apoptosis is observed in cell lines that do not harbor constitutively-active Stat3 at concentrations of resveratrol that otherwise inhibit Stat3 activation. Altogether, our findings suggest that the anticancer properties of resveratrol are in part due to blockade of Src and constitutive Stat3 activities, and are among the first to show cell cycle arrest, growth inhibition, and apoptosis of human cancer cell lines by resveratrol via abrogation of Src-Stat3 signaling.

The potential for resveratrol to modulate gene expression has been previously reported (6, 39). It is conceivable that changes in gene regulation by this compound contribute to its biological effects. Our studies show that resveratrol represses cyclin D1, Bcl-x_L, and Mcl-1, whereas others have similarly reported that resveratrol inhibits survivin induction (40), all of which are known Stat3-regulated genes (41–44), suggesting that modulation of Stat3-regulated cell cycle control and survival genes constitutes part of its mechanisms of antitumor cell activity. Moreover, given the role of Stat3 in repressing p53 expression and function (45), and in NO production in the tumor microenvironment (46), inhibition of constitutive Stat3 activity by resveratrol could induce antitumor cell effects in part by enhancing p53 function and NO activity, as shown by a previous report (47).

In the LNCaP cell line, which lacks constitutively-active Stat3, reports show that low resveratrol concentrations promoted DNA synthesis and progression into S phase (48), whereas apoptosis occurred at high concentrations (100–1,000 μmol/L; ref. 36) or in parallel with modulation of other factors, such as prostate-specific antigen (5, 6, 49). Previous studies also indicate that resveratrol-mediated apoptosis of breast cancer and endometrial adenocarcinoma cells may be dependent or independent of the estrogen receptor status (2, 50–52). These and our findings together raise the potential for resveratrol to induce differential mechanisms for mediating antitumor cell activity depending on the concentration and the molecular background of the tumor cell (12, 53–56). Consistent with this, previous studies showed that resveratrol inhibits the nuclear factor κB pathway and induces apoptosis of leukemia cell lines via caspases (34). We note in our study that resveratrol-induced apoptosis in v-Src-transformed fibroblasts is partly sensitive to the inhibition of caspases.

Repression of Src-Stat3 by resveratrol parallels similar effects on vascular endothelial growth factor (9), mitogen-activated protein kinase (13, 57, 58), and protein kinase C (10), and together provide evidence that modulation of signal transduction pathways also contribute to resveratrol-induced biological effects (12). Although additional work remains to fully understand resveratrol and its antitumorigenic effects, our studies provide novel potential applications for resveratrol as a small-molecule inhibitor of Stat3 signaling with antitumor cell activity (21, 59–65).

We thank colleagues and members of our laboratory for stimulating discussion. The authors also acknowledge the Molecular Imaging and Flow Cytometry Core Groups of the Moffitt Cancer Center for assistance with images and flow cytometry.