Protein Kinase Cε Interacts with Signal Transducers and Activators of Transcription 3 (Stat3), Phosphorylates Stat3Ser727, and Regulates Its Constitutive Activation in Prostate Cancer

Prostate cancer is the most common type of cancer in men and ranks second only to lung cancer in cancer-related deaths. The management of locally advanced prostate cancer is difficult because the cancer often becomes hormone insensitive and unresponsive to current chemotherapeutic agents. Knowledge about the regulatory molecules involved in the transformation to androgen-independent prostate cancer is essential for the rational design of agents to prevent and treat prostate cancer. Protein kinase Cε (PKCε), a member of the novel PKC subfamily, is linked to the development of androgen-independent prostate cancer. PKCε expression levels, as determined by immunohistochemistry of human prostate cancer tissue microarrays, correlated with the aggressiveness of prostate cancer. The mechanism by which PKCε mediates progression to prostate cancer remains elusive. We present here for the first time that signal transducers and activators of transcription 3 (Stat3), which is constitutively activated in a wide variety of human cancers, including prostate cancer, interacts with PKCε. The interaction of PKCε with Stat3 was observed in human prostate cancer, human prostate cancer cell lines (LNCaP, DU145, PC3, and CW22rv1), and prostate cancer that developed in transgenic adenocarcinoma of mouse prostate mice. In reciprocal immunoprecipitation/blotting experiments, prostatic Stat3 coimmunoprecipitated with PKCε. Localization of PKCε with Stat3 was confirmed by double immunofluorescence staining. The interaction of PKCε with Stat3 was PKCε isoform specific. Inhibition of PKCε protein expression in DU145 cells using specific PKCε small interfering RNA (a) inhibited Stat3Ser727 phosphorylation, (b) decreased both Stat3 DNA-binding and transcriptional activity, and (c) decreased DU145 cell invasion. These results indicate that PKCε activation is essential for constitutive activation of Stat3 and prostate cancer progression. [Cancer Res 2007;67(18):8828–38]

Prostate cancer is the second leading cause of cancer-related deaths in men (1). The risk of prostate cancer increases rapidly after age 50, with two thirds of all prostate cancer cases found in men after age 50. Prostate cancer first manifests as an androgen-dependent disease and can be treated with androgen deprivation therapy. Despite the initial success of androgen ablation therapy, prostate cancer progresses from androgen dependent to androgen independent. The hormone-refractory invasive prostate cancer is the end stage and accounts for the majority of prostate cancer patient deaths (2–4). Defining the molecular mechanisms linked to the transition of androgen-dependent prostate cancer to an androgen-independent prostate cancer is essential for planning strategies in the prevention and treatment of prostate cancer. Both protein kinase Cε (PKCε) and signal transducers and activators of transcription 3 (Stat3) have been shown to play roles in the development of androgen-independent prostate cancer (5, 6).

Stats comprise a family of six [Stat1 (α and β isoforms), Stat2 and Stat3 (α and β isoforms), Stat4, Stat5 (α and β isoforms), and Stat6] latent transcription factors, which reside in the cytoplasm and are encoded by seven distinct genes (7). Stat activation is linked to cell proliferation, differentiation, apoptosis, embryogenesis, and immune responses (8–11). Stats exhibit functional divergence in their roles in oncogenesis. Stat3 and Stat5 promote cell survival, whereas Stat1 has been associated with growth-inhibitory effects (12, 13). Constitutively activated Stats, particularly Stat3, are found in several human cancers [e.g., head and neck squamous cell carcinoma (SCC), breast, ovary, prostate, and lung; refs. 14–18]. Because naturally occurring mutations of Stat3 have not been observed, constitutive activation of Stat3 seems to be mediated by aberrant growth factor signaling (7, 8). Tyrosine phosphorylation of Stat3 (Tyr⁷⁰⁵) is mediated by a wide variety of polypeptides and is essential for Stat3 dimerization and nuclear translocation. Stat3 also has a conserved Ser⁷²⁷ residue, which is a target for phosphorylation (19). Evidence indicates that cooperation of both tyrosine and serine phosphorylations is necessary for full activation of Stat3 (20). The identity of the protein kinase responsible for Stat3Ser727 phosphorylation in prostate cancer is unknown.

PKC is a major intracellular receptor for the mouse skin tumor promoter 12-O-tetradecanoylphorbol-13-acetate (TPA; ref. 21). PKC represents a large family of phosphatidylserine (PS)-dependent serine/threonine kinases (22–25). Based on structural similarities and cofactor dependency, 11 PKC isoforms have been classified into three subfamilies: classic (cPKC), novel (nPKC), and atypical (aPKC) isoforms. The cPKCs (α, βI, βII, and γ) are dependent on PS, diacylglycerol (DAG), and Ca²⁺. The nPKCs (δ, ε, η, and 𝛉) retain responsiveness to DAG and PS but do not require Ca²⁺ for full activation. The aPKCs (λ and ς) only require PS for their activation (22). PKC isoforms exhibit functional specificity in their signals to oncogenesis (22). PKCε participates in the regulation of diverse cellular functions, including gene expression, neoplastic transformation, cell adhesion, mitogenicity, and cellular motility. Overexpression of PKCε in rodent fibroblasts led to increase in growth rates, anchorage independence, and tumor formation in nude mice. Additionally, PKCε overexpression transformed nontumorigenic rat colonic epithelial cells (reviewed in ref. 22). PKCε is a transforming oncogene (26) and a predictive biomarker of various human cancers (27), including prostate cancer (5). We found PKCε is linked to the development of SCC elicited either by the 7,12-dimethylbenz(a)anthracene-TPA protocol (28–30) or by repeated UV radiation (UVR) exposures (31, 32). PKCε is also overexpressed in human prostate cancer (33). Overexpression of PKCε transformed androgen-dependent LNCaP tumor cells to androgen-independent cells (5). The transformation of androgen-dependent LNCaP cells to an androgen-independent variant was associated with increased cell proliferation and resistance to apoptosis (34).

We and others have shown, using both in vivo experimental animal models and human cancer–derived cell lines, that PKCε-mediated oncogenic activity is linked to its ability to promote cell survival (16, 26, 34). However, the mechanisms by which PKCε signals cell survival remain elusive. We present here for the first time in prostate cancer that Stat3, which is constitutively activated in a wide variety of human cancers, is a protein partner of PKCε. PKCε interacts with Stat3, phosphorylates Stat3Ser727, and increases both DNA-binding and transcriptional activity of Stat3. PKCε-mediated Stat3Ser727 phosphorylation seems to be essential for constitutive activation of Stat3 and prostate cancer cell invasion.

Chemicals, antibodies, and assay kits. The antibodies and sources of the antibodies used in this study were as follows: PKCε, Stat3, phosphorylated Stat3Tyr705 (pStat3Tyr705), phosphatidylinositol 3-kinase (PI3K; p85), PI3K (p110), p21, p27, Bcl-2, Bcl-xL, survivin, cyclooxygenase-2 (COX-2), β-actin, donkey anti-goat immunoglobulin (IgG)-FITC for PKCε, and donkey anti-rabbit IgG-rhodamine for Stat3 (Santa Cruz Biotechnology); phosphorylated AKT (pAKT; Ser⁴⁷³), pAKT (Thr³⁰⁸), AKT, phosphorylated Janus-activated kinase (pJAK) 1 (Tyr^1022/1023), and pJAK2 (Tyr^1007/1008; Cell Signaling Technology); and phosphorylated Stat3Ser727 (pStat3Ser727; BD Biosciences). Blocking peptides for PKCε and Stat3 were obtained from Santa Cruz Biotechnology. Immunocomplex PKC assay kit was from Biotrak assay system (Amersham Biosciences). Double-stranded Stat3 consensus DNA-binding motif 5′-GATCCTTCTGGGAATTCCTAGATC was obtained from Santa Cruz Biotechnology. PKCε small interfering RNA (siRNA) and siRNA transfection reagents were purchased from Dharmacon, Inc. Dual-Glo luciferase reporter assay kit was purchased from Promega Corp. Cell Invasion Assay kit was purchased from Chemicon International. Plasmid pLucTKS3 was obtained from Dr. James Turkson (University of Central Florida, Orlando, FL).

Human specimens. Prostate cancer specimens were obtained at radical prostatectomy. The University of Wisconsin Human Subjects Committee approved the use of these human prostate specimens. Tissue microarrays were obtained from Folio Biosciences.⁴

Tissue microarrays were produced by relocating tissue from conventional histologic paraffin blocks (Folio Biosciences). This was done using a needle to biopsy a standard histology section and placing the core into an array on a recipient paraffin block. Optimal sectioning of arrays was obtained with ∼5-μm sections. Using this technology, each tissue is treated in an identical manner to avoid slide-to-slide variations in the conventional sections. The tissue microarrays allowed the entire cohort to be analyzed in one batch on a single slide. Specificities of the tissue microarrays were as follows: fixative, formalin; core size, 1.5 mm; section thickness, 5 μm; control quality, pathology of every 10th slide confirmed by board-certified pathologist associated with the company; >95% tissue core retention; and validation, validated for immunohistochemistry.

Animals. Transgenic adenocarcinoma of mouse prostate (TRAMP) mice were obtained both from Dr. Norman Greenberg (Fred Hutchinson Cancer Research Center, Seattle, WA; ref. 35) and from The Jackson Laboratory and screened for the presence of the SV40 large T antigen gene by PCR as detailed on The Jackson Laboratory Web site.⁵

FVB/N mice were obtained from Harlan Sprague Dawley. The mice used in our experiments were either C57BL/6 or [C57BL/6XFVB] F1. Animal care and handling was conducted in accordance with established human guidelines and protocol approved by the University of Wisconsin, School of Medicine and Public Health Animal Care Committee.

Cell lines. Human prostate cell lines LNCaP, DU145, PC3, CW22rv1, and RWPE-1 were obtained from the American Type Culture Collection.

Histology. Prostate specimens were fixed for 1 h in 10% neutral buffered formalin, transferred to PBS (pH 7.4), and then embedded in paraffin. Sections (4 μm thick) of each specimen were cut for immunohistochemical study.

Western blot analysis. Indicated prostate specimens were homogenized in immunoprecipitation lysis buffer [50 mmol/L HEPES (pH 7.5), 150 mmol/L NaCl, 10% glycerol, 1% Triton X-100, 1.5 mmol/L MgCl₂, 10 μg/mL aprotinin, 10 μg/mL leupeptin, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF), 200 mmol/L Na₃VO₄, 200 mmol/L NaF, 1 mmol/L EGTA]. The homogenate was centrifuged at 14,000 × g for 30 min at 4°C. Whole-cell lysate (25 μg) was fractionated on 10% or 15% Tris-glycine SDS-polyacrylamide gels for Western blot analysis as described before (16).

Determination of PKCε and Stat3 localization by immunofluorescence staining. Paraffin-fixed prostate samples and prostate cancer cell line DU145 were used to determine the localization of PKCε and Stat3. Tissue sections (5 μm thick) of prostate samples were cut to determine the localization of PKCε and Stat3 as described before (16).

PKCε immunocomplex kinase assay. Prostate tissue was placed in 0.5 mL of immunoprecipitation lysis buffer, homogenized using a glass Teflon tissue homogenizer, agitated for 30 min at 4°C, and centrifuged at 14,000 rpm in a microcentrifuge for 15 min at 4°C. The clear supernatant was used for immunoprecipitation using polyclonal antibody to PKCε. PKCε immunoprecipitate (25 μL) was assayed for kinase activity as described before (28).

PKCε siRNA transfection. Eighty percent confluent prostate cancer DU145 cells were starved by incubation for 18 to 24 h before assay in the serum-free medium. The transfection was done as per the manufacturer's instructions (Dharmacon). A set of four pooled nontargeting siRNAs (2 μmol/L/100-mm Petri dish) was used as a control. A set of four PKCε-specific siRNA (2 μmol/L/100-mm Petri dish) was used to silence PKCε. After 48 h of siRNA transfection, the cells were harvested and used for Western blot analysis, electrophoretic mobility shift assay (EMSA), and luciferase assay.

Electrophoretic mobility shift assay. Indicated amount of protein extract was incubated in a final volume of 20 μL of 10 mmol/L HEPES (pH 7.9), 80 mmol/L NaCl, 10% glycerol, 1 mmol/L DTT, 1 mmol/L EDTA, and 100 μg/mL poly(deoxyinosinic-deoxycytidylic acid) for 15 min. A γ-³²P–radiolabeled double-stranded Stat3 consensus binding motif 5′-GATCCTTCTGGGAATTCCTAGATC (Santa Cruz Biotechnology) probe was then added and incubated for 20 min at room temperature. The protein-DNA complexes were resolved on a 4.5% nondenaturing polyacrylamide gel containing 2.5% glycerol in 0.25× Tris-borate EDTA at room temperature, and gels were dried and autoradiographed. Stat3 DNA-binding activities were determined.

Luciferase assay. To determine the influence of PKCε on Stat3-associated gene expression, a sensitive Dual-Glo luciferase reporter assay using plasmids pLucTKS3 and pTKRenillaLuc was used (pGEM-luc vector, Promega).

Cell invasion assay. The cell invasion was assayed using a 24-well Collagen-Based Cell Invasion Assay kit as per the manufacturer's instructions. Briefly, DU145 cells at 80% confluency were serum starved for 18 to 24 h before the assay. Cells were harvested and the pellet was gently resuspended in serum-free medium. In the upper chamber, 0.5 × 10⁶ cells per well were plated in triplicates and incubated for 2 h at 37°C in a humidified incubator with 5% CO₂ before PKCε siRNA transfection. Both the insert and the holding well were subjected to the same medium composition with the exception of serum. The insert contained no serum, whereas the lower well contained 10% fetal bovine serum that served as a chemoattractant. PKCε siRNAs were obtained from Dharmacon. The transfection of siRNA was done as per the manufacturer's instructions. A set of four pooled nontargeting siRNAs (described earlier) was used as a control. A set of four PKCε siRNA (described earlier) was used to silence PKCε. Forty-eight hours after siRNA transfection, the cell invasion assay was done as per the manufacturer's instructions. The cells in the insert were removed by wiping gently with a cotton swab. Migrated cells sticking to the bottom side of the insert were stained with Cell Stain. Invading cells on the bottom side of the membrane were photographed using a light inverted microscopy (Nikon Eclipse TS 100) at ×20 magnification. In addition, the number of cells migrated to the bottom side was estimated by colorimetric measurements at 560 nm according to assay instructions. Mean ± SE was calculated from three independent experiments.

Expression levels of PKCε and Stat3 and their in vivo colocalization were analyzed in human prostate cancer (Fig. 1). In the analysis illustrated in Fig. 1, prostate tumor specimens were obtained after prostatectomy and fixed in 10% neutral buffered formalin. Sections (5 μm thick) were cut for immunohistochemical staining of PKCε and Stat3. Small samples of prostate cancer were examined in a blinded fashion by two pathologists (T.D.O. and W.Z.) without previous knowledge of Gleason score and classified according to both Gleason pattern and WHO criteria (36, 37). Expression levels of PKCε in prostate are shown in Fig. 1A. Low levels of PKCε expression were detected in the cytoplasm of benign prostate epithelial cells. Atrophic epithelium showed trace to moderate cytoplasmic labeling for PKCε. Areas of prostate intraepithelial neoplasia showed light cytoplasmic (Fig. 1A,, double arrow) and nuclear (Fig. 1A,, arrowhead) expression levels of PKCε. Cytoplasm of the well-differentiated prostate adenocarcinoma showed only trace levels of PKCε staining. Cytoplasm of moderately differentiated prostate carcinoma showed moderate to strong expression levels of PKCε labeling. Poorly differentiated prostate adenocarcinoma showed moderate cytoplasmic and nuclear (arrow) PKCε expression levels (Fig. 1A). Immunohistochemical staining of human prostate carcinoma specimens showed increased nuclear staining of Stat3 in prostate cancer specimens compared with benign prostate specimens (Fig. 1B). Benign prostate epithelium showed low level of cytoplasmic Stat3. Moderately to poorly differentiated prostate adenocarcinoma showed moderate cytoplasmic and intense nuclear staining (Fig. 1B). Inclusion of appropriate blocking peptide before immunostaining of human SCC specimens completely prevented PKCε and Stat3 staining, indicating the immunostaining was specific (Fig. 1C).

The localization of PKCε with Stat3 was determined by double immunofluorescence staining. In this experiment (Fig. 1D), 5-μm-thick sections from paraffin-fixed human prostate cancer specimen were used. The sections were incubated with a mixture of PKCε (goat polyclonal, 1:50) and Stat3 (rabbit polyclonal, 1:50) primary antibodies. The sections were subsequently incubated with the mixture of two secondary antibodies (donkey anti-goat IgG-FITC for PKCε and donkey anti-rabbit IgG-rhodamine for Stat3). The stained sections were examined using an Olympus microscope. PKCε and Stat3 localization are indicated by the presence of green and red fluorescence, respectively. The yellow fluorescence indicates colocalization of both PKCε and Stat3 (Fig. 1D). Merged images clearly illustrate significant colocalization of PKCε and Stat3 (Fig. 1D , white arrow).

In a more detailed analysis, PKCε and Stat3 levels were determined by immunohistochemistry of human prostate tissue microarrays prepared from paraffin-fixed benign, hyperplastic, and carcinoma (grade 1–4) specimens (Table 1). For each section, staining was assessed as weak or faint (+), moderate or intermediate (++), and strong or intensive (+++) staining when present in >50% of total cells evaluated in each tissue core. As compared with benign prostate tissue, PKCε expression level in prostate cancer grade 2 to 4 was significantly (P < 0.001) increased. The level of expression of PKCε seemed to correlate with aggressiveness of prostate cancer (Table 1).

Stat3 staining was moderate to strong in human prostate cancer specimens compared with benign prostate specimens, which exhibited very weak staining (Table 1). Cytoplasmic as well as nuclear Stat3 staining was found in prostate cancer tissue specimens. Lower-grade or well-differentiated human prostate cancer specimens had more cytoplasmic Stat3 staining, whereas higher-grade or poorly differentiated human prostate cancer specimens showed more nuclear staining. As compared with benign prostate tissue, Stat3 expression level in prostate cancer grade 2 to 4 was significantly (P < 0.001) elevated (Table 1).

To precisely determine the modulation of PKCε, Stat3, and related signaling components in prostate cancer, we used prostate cancer from TRAMP mice. The TRAMP mice spontaneously develop progressive prostate cancer that is invasive and capable of metastatic spread to distant sites, primarily in pelvic lymph nodes and the lungs (35, 38, 39). It has been reported that TRAMP mice castrated at 12 weeks of age develop androgen-independent prostate cancer (38). In addition, TRAMP mice develop spontaneous mutations in the androgen receptor (AR; ref. 39). Thus, prostate cancer in TRAMP mice closely mimics clinical prostate cancer with respect to progression, androgen independence, and biochemical profiles (35, 38, 39). The mouse genetic background makes a difference in prostate cancer progression (39). Prostate cancer progression in TRAMP mice on a mixed C57BL/6XFVB background is different from that of a C57BL/6 background. The C57BL/6 TRAMP mice can survive longer (∼52 weeks) than [C57BL/6 TRAMP × FVB] F1 mice (∼32 weeks; ref. 35). In addition, prostate cancer in C57BL/6 TRAMP mice often invades into seminal vesicles. Like the human prostate cancer disease, prostate cancer in TRAMP mice is also multifocal and heterogenous. In the experiments described here, both C57BL/6 and [C57BL/6 TRAMP × FVB] F1 male TRAMP mice were used.

PKC isoforms and Stat3 expression levels. In this experiment (Fig. 2), [C57BL/6 TRAMP × FVB] F1 mice (n = 13) were sacrificed at 26 weeks of age and prostate cancer was excised. A portion of the prostate cancer specimen was fixed in formalin, whereas another part was homogenized in lysis buffer to extract total protein for analysis of protein expression levels of PKC isoforms, Stat3, pStat3Tyr705, and pStat3Ser727 as well as PKCε activity and Stat3 DNA-binding activity. Normal prostate tissue specimens from nontransgenic mice were processed in parallel.

Typical PKCε and Stat3 immunostaining results are illustrated in Fig. 2A. Increased protein expression levels of PKCε in prostate cancer (Fig. 2A) correlated with increased PKCε activity (Fig. 2D). In addition, constitutively activated Stat3 in prostate cancer (Fig. 2A) seems to have increased DNA-binding activity (Fig. 2D). We also determined the level of expression of various PKC isoforms in prostate cancer from TRAMP mice. Western blot analysis of PKC isoforms in the normal prostate from nontransgenic mice and prostate cancer from 26-week-old TRAMP mice is shown in Fig. 2B. PKCγ, PKC𝛉, and PKCλ were not detected in either normal or prostate cancer samples. PKCε expression was elevated in prostate cancer with reciprocal decreases in the expression levels of PKCδ, PKCβII, PKCη, and PKCζ. PKCα and PKCβI remain unaltered in prostate cancer (Fig. 2B). It is noteworthy that these results are consistent with the expression of PKC isoforms in human prostate adenocarcinomas as reported by Cornford et al. (33). In human prostate cancer specimens, the level of PKCε was elevated, whereas the levels of expression of PKCδ and PKCβII were decreased or undetected (33).

Western blot analysis revealed increased expression levels of pStat3Tyr705, pStat3Ser727, and Stat3β in prostate cancer from 26-week-old TRAMP mice compared with normal prostate from age-matched nontransgenic mice (Fig. 2C).

Expression level of cytokines and Stat3 signaling components in prostate cancer from TRAMP mice. Several cancer-promoting agents induce the release of proinflammatory cytokines [interleukin (IL)-1, IL-6, IL-7, IL-15, granulocyte macrophage colony-stimulating factor, and tumor necrosis factor-α (TNF-α)], chemotactic cytokines (IL-1 and TNF-α), and cytokines regulating immunity (IL-10, IL-12, and IL-18; ref. 40). Overwhelming evidence indicates a relationship between cytokines and the development and progression of many cancers. Elevated serum levels of specific cytokines have been reported in prostate cancer specimens and in prostate cancer cell lines (41, 42). Several reports indicate high concentrations of IL-6 in men with hormone-refractory prostate cancer compared with normal controls (43, 44). We also found elevated levels of IL-6 in prostate cancer from [C57BL/6 TRAMP × FVB] F1 mice (Fig. 3A). Increased IL-6 expression levels accompanied increased expression of pJAK1 and pJAK2 (Fig. 3B).

To find clues of the kinase cascade(s) involved with increased IL-6 expression in prostate cancer, we examined the expression of various signaling kinases and found increased expression of PI3K (p85), PI3K (p110), phosphorylated AKTSer473, phosphorylated AKTThr308, and AKT (Fig. 3C).

Activation of the PKCε and Stat3 pathway has been shown to inhibit apoptosis and promote survival of prostate cancer cells (7, 13, 34). We determined the expression of genes associated with apoptosis (Bcl-2, Bcl-xL, and survivin), cell cycle regulation (p21 and p27), and proliferation (COX-2) in prostate cancer from TRAMP mice. As shown in Fig. 3D, increased expression of PKCε and Stat3 in prostate cancer from TRAMP mice accompanied decreased expression of p21 and p27 and increased expression of Bcl-xL, Bcl-2, survivin, and COX-2.

To determine whether PKCε associates with Stat3, reciprocal immunoprecipitation/blotting experiments were done using prostate tissue specimens from wild-type (WT) and TRAMP mice (Fig. 4A) and human prostate cancer cell lines (Fig. 4B). In these experiments, the protein extract was immunoprecipitated with antibodies against PKCε or Stat3. The immunoprecipitated samples were subjected to immunoblot analysis using antibodies against PKCε, Stat3, pStat3Tyr705, and pStat3Ser727. The results of PKCε interaction with Stat3 in prostate tissue are shown in Fig. 4A. Stat3, pStat3Tyr705, and pStat3Ser727 coimmunoprecipitated with PKCε. In the reciprocal immunoprecipitation/blotting experiments, PKCε coimmunoprecipitated with Stat3 (Fig. 4A). PKCε interaction with pStat3Ser727 was more dramatic in prostate cancer from TRAMP mice than normal prostate tissue from WT mice. PKCε interaction with Stat3 in human prostate cancer cell lines is shown in Fig. 4B. Analysis of steady-state expression levels of PKCε revealed increased PKCε expression level in DU145 cells (Fig. 4B,, top). PKCε interaction with Stat3 was observed in LNCaP, DU145, PC3, and CW22rv1 cell lines. However, PKCε interaction with Stat3 was not detectable in RWPE-1 cell line. The co-localization of PKCε with Stat3 in prostate cancer from TRAMP mice (Fig. 4C,, left) and in DU145 cells (Fig. 4C , right) was confirmed by double immunofluorescence staining.

The interaction of Stat3 with other PKC isoforms was also determined. In this experiment (Fig. 4D), prostate cancer protein extract from TRAMP mice was immunoprecipitated with Stat3 antibody. The immunoprecipitated samples were subjected to immunoblot analysis using antibodies against various PKC isoforms. PKCε coimmunoprecipitated with Stat3. However, other PKC isoforms showed little (α and βI) or no (βII, γ, δ, λ, ζ, η, and 𝛉) association with Stat3 (Fig. 4D).

To determine whether PKCε is linked to Stat3Ser727 phosphorylation and whether this phosphorylation is essential for the DNA-binding and transcriptional activity of Stat3, we used siRNAs to silence PKCε in DU145 cells (Fig. 5A and B). A pool of four specific siRNA oligonucleotides directed against PKCε mRNA was transfected into DU145 cells to inhibit PKCε synthesis (Fig. 5A). Inhibition of PKCε expression resulted in dramatic inhibition of Stat3Ser727 phosphorylation but minimally inhibited Stat3Tyr705 phosphorylation and total Stat3 levels (Fig. 5A). Furthermore, siRNA-mediated inhibition of PKCε inhibited Stat3 DNA-binding activity as determined by EMSA (Fig. 5B,, left). In addition, siRNA-mediated PKCε deficiency in DU145 cells inhibited Stat3-regulated gene expression as determined using the luciferase reporter gene assay (Fig. 5C , right).

The possibility was explored whether PKCε is linked to prostate cancer progression. In this experiment (Fig. 5D), PKCε expression was inhibited using PKCε siRNA. PKCε inhibition, which accompanies inhibition of Stat3Ser727 phosphorylation and Stat3 DNA-binding and transcriptional activities, decreased the invasive ability of DU145 cells (Fig. 5D,, top). siRNA-mediated inhibition of DU145 cells invasion was dependent on the dose of siRNA (Fig. 5D , bottom).

We report here that both PKCε and constitutively activated Stat3 are overexpressed in prostate cancer and their expression levels correlate with prostate cancer aggressiveness. PKCε overexpression in prostate cancer accompanied (a) increased in the levels of IL-6 and pJAK1 and pJAK2, (b) increased phosphorylation of PI3K and AKT, (c) decreased expression of cyclin-dependent protein kinase inhibitors (p21 and p27), and (d) up-regulation of antiapoptotic (Bcl-2, Bcl-xL, and survivin) and proliferative (COX-2) markers. PKCε interacts with Stat3 and phosphorylates Stat3Ser727. PKCε-mediated Stat3Ser727 phosphorylation correlated with Stat3 DNA-binding and transcriptional activities as well as prostate cancer progression.

PKCε was overexpressed in human prostate cancer and prostate cancer developed either in C57BL/6 or [C57BL/6 TRAMP × FVB] F1 TRAMP mice (Figs. 1 and 2). The fact that PKCε expression is significantly elevated in prostate cancer and correlates with prostate cancer aggressiveness (Figs. 1 and 2; Table 1) implies that PKCε is probably linked to the maintenance of androgen-independent prostate cancer. In this context, the pioneering work of Terrian and his associates on the role of PKCε in prostate carcinogenesis, using prostate cancer–derived cell lines, is noteworthy (5, 34, 45). In their reports, PKCε overexpression transformed androgen-dependent LNCaP tumor cells to androgen-independent cells (5). The transformation of androgen-dependent LNCaP cells to an androgen-independent variant was associated with increased cell proliferation and resistance to apoptosis. Antisense experiments established that endogenous PKCε plays an important role in regulating the growth and survival of androgen-independent prostate cancer cells, suggesting that PKCε expression may be sufficient to maintain prostate cancer growth and survival after androgen ablation (5).

PKCε, which is overexpressed in prostate cancer, seems to be constitutively active. This conclusion is supported by two lines of evidence. First, increased PKCε expression correlated with PKCε activity in prostate cancer from TRAMP mice (Fig. 2), and second, in an immunohistochemical analysis, PKCε was observed to translocate to both plasma and nuclear membranes (Figs. 1 and 2). The mechanism by which PKCε remained constitutively active during progression of prostate cancer can be speculated. The generally accepted paradigm for PKC activation (22) consists of two major biological events. The initial event is a well-ordered sequential phosphorylation followed by allosteric activation mediated by two lipid signals. Priming phosphorylation on the activation loop residue threonine by phospholipid-dependent kinase 1 initiates autophosphorylation on the turn motif threonine residue and hydrophobic motif serine residues. Most PKC in unstimulated cells is in the triple-phosphorylated mature form. The mature, phosphorylated species are released to the cytosol and the pseudosubstrate gains access to the active site, thus maintaining the enzyme in an autoinhibited state (22). As the mature PKC positions near the membrane for rapid access to DAG, the translocation of the mature enzyme from the cytosolic soluble fraction to particulate plasma membrane fraction is mediated by the binding of a lipid second messenger. Subsequent association with the membrane allows PKC to carry out phosphorylation of target substrates. The constitutive activation of PKCε is perhaps due to dysregulation of PKCε synthesis, activation, and/or degradation.

Constitutively activated PKCε regulates activation of the signaling network linked to cell survival essential for the maintenance of androgen-independent prostate cancer (5). As observed in prostate cancer from TRAMP mice, PKCε expression accompanied up-regulation of phosphorylated PI3K and AKT (Fig. 3), major components of the cell survival pathway (Fig. 6). Consistent with these findings, using CWR22 xenografts, it was shown by proteomic analysis that the association of PKCε with Bax may neutralize apoptotic signals propagated through the mitochondrial death signaling pathway. In addition, integrin signaling links PKCε to the protein kinase B/AKT survival pathway in recurrent prostate cancer cells (34). Our results (Fig. 3), using the in vivo TRAMP mouse model, are consistent with these findings (34). PKCε overexpression, which correlates with prostate cancer aggressiveness (Table 1), accompanied an increase in proteins that modulate apoptosis (survivin, Bcl-2, and Bcl-xL) and cell survival (Stat3; Fig. 3).

Several lines of evidence support the link between PKCε and Stat3. First, both PKCε and Stat3 are overexpressed in prostate cancer (Figs. 1 and 2; Table 1). Second, using reciprocal immunoprecipitation experiments, it was observed that PKCε associates with Stat3 (Figs. 4 and 5). Finally, double immunofluorescence studies confirm their colocalization (Figs. 1 and 4). PKCε depletion prevents PKCε association with Stat3 and Stat3Ser727 phosphorylation. Together, these biochemical, immunologic, and enzymology studies indicate that PKCε interacts with Stat3 and is linked to Stat3Ser727 phosphorylation.

Depending on the cellular context, Stat3 is a substrate for several protein kinases (46). Stat3 is an in vitro substrate for mitogen-activated protein kinase (MAPK) and/or extracellular signal-regulated kinases (ERK; ref. 47). Pioneering research from Dr. Zigang Dong's laboratory (Hormel Institute, University of Minnesota, Austin, MN) has shown that UV-induced Stat1 and Stat3 (Ser⁷²⁷) phosphorylation involves the integration of multiple kinase pathways (48, 49). Recently, we have reported that PKCε interacts with Stat3 in mouse skin. PKCε interaction with Stat3 was increased after exposure of skin with UVR (16). We also observed in vivo using PKCε knockout mice and in vitro in an immunocomplex kinase assay that PKCε phosphorylated Stat3 at Ser⁷²⁷ residue, implying that PKCε is a Stat3Ser727 kinase (16). Our results (Figs. 4 and 5) indicate that PKCε is also an initial signal that is linked to phosphorylation of Stat3Ser727 in prostate cancer. Dependent on the cellular context and stimulus, Stat3 may be a direct or indirect substrate of PKCε.

The functional specificity of each PKC isoform is determined in part by the differential localization of the isozyme-specific RACKS. RACKS, which anchor activated PKC isozymes in close proximity to its selective substrates, are not PKC substrates. Proteins that interact with PKC include proteins localizing PKC to a specific intracellular site (e.g., RACKS and RICKS), protein kinases (e.g., Raf1, Btk, S6 kinase, PKD, PDK1, PKF, and Fyn), substrate proteins (e.g., MARCKS and STICKS), and other proteins with unique functions (25, 50–54). Several proteins have specifically been shown to interact with PKCε (51, 54). Our results (Figs. 4 and 5) indicate that PKCε interacts with Stat3 and phosphorylates Stat3Ser727. PKCε-mediated phosphorylation of Stat3Ser727 may be an essential component of constitutive activation of Stat3 in a wide variety of human cancers (14–18).

The mechanism by which Stat3 is constitutively activated in prostate cancer can be possibly explained as shown in Fig. 6. Stat3 activation, which involves dimerization, nuclear translocation, DNA binding, and transactivation of transcription, requires phosphorylation of both Tyr⁷⁰⁵ and Ser⁷²⁷ (8–12). Stat3Tyr705 phosphorylation is mediated by a wide variety of growth factors (e.g., IL-6). IL-6 has been characterized as a prostate exocrine gene product that interacts with its receptors, activates AR, and regulates prostate cancer cell proliferation. IL-6 signaling is mediated through JAK, Stat3, as well as MAPK, which induces AR-mediated gene activation in prostate cancer (12, 13). JAK-Stat is the classic pathway that has been shown to mediate cellular responses to a variety of cytokines, including IL-6. In response to IL-6, Stat3 is transiently associated with gp130 and subsequently phosphorylated by JAKs on Tyr⁷⁰⁵ of Stat3. Stat3 represents an important regulator of IL-6–targeted gene expression. PKCε-mediated Stat3 Ser⁷²⁷ phosphorylation is essential for both optimal DNA-binding and transcriptional activities of Stat3 (Fig. 6).

In summary, PKCε and constitutively activated Stat3 are key components in the mechanisms involved in transition of androgen-dependent prostate cancer to androgen-independent prostate cancer (45). We have made the novel observation that PKCε associates with Stat3 and regulates Stat3 activation (Figs. 1, 4, and 5). PKCε activation transduces multiple signals involving inhibition of apoptotic pathways and promotion of cell survival pathways (Figs. 3 and 6). PKCε-mediated cell survival pathway involves constitutive activation of Stat3. PKCε is an initial signal that regulates activation of Stat3. PKCε and Stat3, the proteins with oncogenic traits, should be explored as potential targets for prevention of androgen-independent prostate cancer.

Grant support: Department of Defense grant W81XWH and NIH grant CA35368 (A.K. Verma).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank David B. Pierce for expert technical assistance.