Protein Kinase D3 (PKD3) Contributes to Prostate Cancer Cell Growth and Survival Through a PKCε/PKD3 Pathway Downstream of Akt and ERK 1/2

Prostate cancer is the second leading cause of cancer-related deaths among men in the United States. The high morbidity and mortality associate with the metastatic progression of the cancer. Protein kinase C (PKC), the primary target of diacylglycerol (DAG) and phorbol esters (the pharmacologic analogues of DAG and potent tumor promoters) has been implicated in the development and progression of prostate cancer (1, 2). The PKC family comprises 10 isoforms that are classified into cPKC (α, βI/βII, and γ; Ca²⁺- and DAG-dependent), nPKC (δ, ε, η, and 𝛉; Ca²⁺-independent but DAG-dependent), and aPKC (ζ and ι/λ; both Ca²⁺- and DAG- independent; ref. 3). Isoforms of the PKC family are differentially expressed in prostate cancer cell lines and have been shown to regulate different cellular responses. Of particular interest, PKCε has been implicated in prostate cancer progression. PKCε expression correlates with the aggressiveness of human prostate cancer (4), and PKCε alone drives androgen independence in tumor cell lines and in mice (5). Although a number of intracellular targets have been identified for PKCε including caveolin-1, ILK, Akt, Bax, Stat3, etc. (4, 6, 7), the immediate signaling pathways through which PKCε promotes prostate cancer progression remain to be defined.

The PKD family belongs to a family of novel serine/threonine kinases that bind DAG and phorbol esters. It is now classified as a subfamily of the Ca-Calmodulin kinase superfamily (8, 9). Three PKD isoforms (1, 2, and 3) have been identified, and PKD1 is the most studied isoform. PKD has been implicated in a variety of cellular processes including regulation of Golgi function, cell proliferation, apoptosis, and cell migration (8, 10). PKD is capable of shuttling between different subcellular compartments and is activated in intact cells by phorbol esters, G-protein–coupled receptor agonists, and certain growth factors (10). In most cells, the activity of PKD is controlled by PKC. PKC isoforms, particularly the novel PKCs, directly bind, phosphorylate and activate PKD (11). Thus, PKC/PKD is the canonical pathway that leads to PKD activation (8). PKD has been shown to transduce mitogenic signals and promote cell proliferation (8, 10). The pro-proliferative effects of PKD in cells are linked to the modulation of the extracellular signal-regulated kinase (ERK)1/2 mitogen-activated protein kinase (MAPK) pathway by PKD (12, 13). PKD is prosurvival and antiapoptotic in many cellular systems (14). PKD1 has been shown to be critical for oxidative stress-induced cell survival response through activating nuclear factor-κB (NF-κB) signaling (15, 16). These studies mostly focus on PKD1, whereas the roles of other PKD isoforms in cell growth and survival remain largely unknown.

In this report, we showed a role of PKD3 in prostate cancer cell growth and survival. Our results indicated that PKD3 expression was elevated in human prostate tumors, and that disease progression correlated with an increase in PKD3 nuclear accumulation. PKCε acting upstream controlled PKD3 activity and nuclear localization, and the effects of PKD3 on prostate cancer cells coupled to the modulation of downstream ERK1/2 and Akt activities.

Tissue samples and PKD immunohistochemistry. Tissue microarray slides containing variable numbers of histologically confirmed human prostate cancer specimens and normal human prostate tissue samples were purchased from US Biomax. The sections were deparaffinized by xylene and rehydrated in reducing gradients of ethanol (100%, 90%, and 75%). Antigen retrieval was performed by simmering the slides at near boiling temperature for 30 min in 10 mmol/L sodium citrate buffer (pH 6.0), followed by cooling slowly to room temperature. After brief permeabilization in TBS containing 0.05% Tween 20 for 10 min, the tissue arrays were stained with PKD1 (N-20; Santa Cruz) or PKD3 (AP7025a; Abgent) antibodies at 4°C overnight, and then incubated with biotinylated goat anti-rabbit antibody (Vector Laboratories). The slides were then developed using Vectastain ABC Standard kit, followed by 3,3′-diaminobenzidine/Ni substrate kit according to the manufacturer's instructions (Vector Laboratories). The tissues cores were counterstained with hematoxylin.

The staining was scored independently by two experienced researchers according to the number and intensity of immunopositive cells in a blinded fashion. The intensity of immunohistochemical staining was categorized as negative if the reaction was indistinguishable from the background or <5% of tumor cells were stained; low if 5% to 30% of tumor cells were positively stained; high if 30% to 60% of tumor cells were positively stained; and very high if >60% were positively stained.

Cell culture, siRNA transfection, and adenoviral infection. LNCaP, PC3, and DU145 cells were obtained from American Type Culture Collection and cultured according to the manufacturer's recommendations. LNCaP cells were discarded after eight passages.

The nontargeting and validated PKD3 siRNAs were obtained from Invitrogen. The siRNAs were transfected into cells using DharmaFECT 1 or 2 reagents according to the manufacturer's instructions. The adenoviruses for PKD3 were generated as described (17). Viral infection was conducted as previously described (18).

Western blot analysis and in vitro kinase assay. Cells were harvested in lysis buffer containing 50 mmol/L Tris-HCI (pH 7.4), 150 mmol/L NaCl, 1 mmol/L MgCl₂, 1 mmol/L AEBSF, 5 mmol/L EGTA, 1 mmol/L NaVO₃, 10 mmol/L NaF, 20 μg/mL leupeptin, 10% Glycerol, 1 proteinase inhibitor tablet (Roche Diagnostics Corp.), and 1% Triton X-100. Equal amounts of protein were subjected to Western blotting as described (19).

PKD3–green fluorescent protein (GFP) or endogenous PKD3 were separately immunoprecipitated by a GFP monoclonal antibody or a PKD3 antibody preconjugated to protein A/G sepharose beads. After washing with RadioImmunoPrecipitation Assay (RIPA) buffer [50 mmol/L Tris-HCI (pH 7.4), 150 mmol/L NaCl, and 1% Triton X-100] and kinase buffer (30 mmol/L Tris-HCI, 10 mmol/L MgCl₂, and 1 mmol/L DTT), the beads were subjected to the in vitro kinase assay as described (19).

Apoptosis assay, cell proliferation assay, and cell cycle analysis. Apoptosis was assessed by TUNEL assay and 4′,6-diamidino-2-phenylindole (DAPI) staining as described (20). Nuclear DNA condensation was visualized by staining with 300 nmol/L DAPI in PBS for 5 min.

Cell proliferation was determined by counting the number of viable cells upon trypan blue staining. For cell cycle analysis, PC3 cells were fixed in ice-cold 70% ethanol overnight at 4°C, followed by propidium iodide labeling. The labeled cells were analyzed using a FACScan Benchtop Cytometer (BD Biosciences). Results were analyzed using ModFit LT software (Verity Software House).

Indirect immunofluorescence. Cells were seeded on chamber slides and allowed to attach overnight. Immunofluorescent staining was performed as previously described (21). Fixed and permeabilized cells were incubated for 1 h with the primary antibodies, followed by the FITC- or rhodamine-conjugated secondary antibodies. Stained cells were imaged under laser scanning confocal fluorescent microscope.

Statistical analysis. Statistical significance between the expression of PKD3 in normal and tumor tissues was analyzed using Fisher's exact test. Correlation between the PKD3 expression and pathologic grades or Gleason scores was assessed using χ² test. All statistical analysis was done using GraphPad Prism IV software. A P value of <0.05 was considered statistically significant.

PKD3 expression was elevated in human prostate cancers. PKD3 expression was analyzed by immunohistochemical staining in a total of 132 tumor specimens and 26 normal tissue cores, among which were 12 normal tissues adjacent to the tumors. As shown in Fig. 1A, PKD3 was mostly absent from the secretory epithelial cells of the normal prostate glands. Weak to moderate cytoplasmic staining for PKD3, with some instances of nuclear localization, was observed in the basal epithelial cell layer. Additionally, we observed some weak to moderate cytoplasmic staining for PKD3 in the fibromuscular stroma (Fig. 1A,, 1–2), and less PKD3 staining was observed in the stroma surrounding the high-grade tumors compared with those in the normal and the low-grade tumors (Fig. 2A,, 3–4 versus 1–2). Importantly, more frequent and intense PKD3 staining was observed in malignant tumors of the prostate (Fig. 1A,, 3–4). As shown in Fig. 1B, the percentage of PKD3 positively stained tissues increased from 23% for normal prostate to 85% for prostate cancer. The difference was highly significant (P < 0.0001). However, the up-regulation of PKD3 expression in the tumors did not significantly correlate with the increased pathologic grades or Gleason scores.

The expression of PKD1 was analyzed in a total of 180 tumor specimens and 25 normal tissue cores. The secretory epithelial cells of the normal prostate glands showed weak cytoplasmic and plasma membrane staining for PKD1. We also observed very weak cytoplasmic staining of PKD1, with some instances of nuclear localization, in the fibromuscular stroma of normal and tumor tissues. Quantitative analysis showed that PKD1 expression significantly increased in tumors compared with the normal tissues, 87% positive for tumor tissues versus 52% positive for normal tissues (P < 0.001; Supplementary Fig. S1). Similar to PKD3, no significant correlation was found between the expression of PKD1 and pathologic grades. Taken together, our data indicated that PKD3 as well as PKD1 was up-regulated in prostate cancer, implying a potential role of PKD isoforms in the pathogenesis of prostate cancer.

PKD3 nuclear retention correlated with pathologic grades of prostate cancer. In this study, the localization of PKD3 was analyzed in normal and malignant tissues of the prostate. PKD3 showed progressively enhanced nuclear distribution ranging from in the cytosol only, evenly in the cytosol and the nucleus, predominantly in the nucleus, to exclusively in the nucleus (Fig. 2A). Tumor tissues with >50% of tumor cells showing predominant nuclear staining of PKD3 were considered positive for nuclear localization. The increase in nuclear distribution of PKD3 was remarkable from being completely devoid of nuclear localization in the glandular epithelial cells of normal tissues to 39% positive nuclear staining in tumor samples, with the rest of the tumors showing predominant cytoplasmic distribution (negatively stained tumor tissues were not included; Fig. 2B). Moreover, as shown in Fig. 2A and C, increased nuclear localization significantly correlated with high tumor grades. Grade I and II tumors exhibited 11% and 19% positive nuclear PKD3 staining, whereas grade III and IV tumors were 56% and 41% positive for nuclear PKD3, respectively (P < 0.01). Similarly, we observed increased nuclear accumulation of PKD1 in high-grade tumors, although the number of tumor tissues with positive nuclear staining for PKD1 was considerably less compared with those stained for PKD3 (∼5% for grade II and III tumors and 14% for grade IV tumors). These results suggested that nuclear accumulation of PKD3 might contribute to prostate cancer progression.

PKD3 was differentially expressed and localized between normal and human prostate cancer cells. The expression and distribution of PKD isoforms were examined in the benign prostate epithelial cells—PrEC and a panel of prostate cancer cell lines, including LNCaP—a less metastatic androgen-dependent prostate cancer cell line; C81 and C4-2—two androgen-independent cell lines derived from LNCaP (22, 23); DU145 and PC-3—two androgen-independent more metastatic and aggressive prostate cancer cell lines; TRAMP-C2—a cell line derived from the late-stage tumors in the Transgenic Adenocarcinoma of Mouse Prostate (TRAMP) mice (24). Expression of various PKD and PKC isoforms was evaluated by Western blotting. As shown in Fig. 3A, members of the PKD family were differentially expressed in prostate cancer cells. PKD1 was detected in LNCaP, C81, and C4-2 cells, although almost absent in PrEC, PC3, DU145, and TRAMP-C2 cells (trace amount was observed under longer exposure). PKD2 was detected at higher levels in PrEC and LNCaP cells but low in DU145 and PC3 and only minimally expressed in C81, C4-2, and TRAMP-C2 cells. In comparison, PKD3 expression was barely detectable in PrEC, LNCaP, C81, and C4-2 cells, although markedly elevated in DU145 and PC3 cells. In TRAMP-C2, PKD3 was the only isoform abundantly expressed. Interestingly, the expression pattern of PKD3 correlated well with the expression pattern of PKCα or PKCε but not with that of PKCδ or PKCζ. Furthermore, the expression levels of PKD1 and PKD3 were similar in LNCaP and its two androgen-independent isogenic lines, C81 and C4-2, suggesting a less likely link between PKD1 and PKD3 expression and androgen sensitivity. Equal loading was confirmed by blotting for glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

The intracellular distribution of PKD1, PKD2, and PKD3 was analyzed by immunofluorescent staining. PKD1 in LNCaP cells was localized primarily to the cytoplasm and plasma membrane and less to the nucleus, which agreed with a previous report (Fig. 3B,, 1; ref. 25). In contrast, PKD3 was predominantly localized in the nucleus in DU145 (Fig. 3B,, 2) and PC3 (data not shown) cells. Furthermore, active PKD stained by the p-S744/748-PKD antibody that recognizes activation-loop phosphorylated PKD localized predominantly in the nucleus of PC3 (Fig. 3B,, 3) and DU145 cells (data not shown). We have also examined the distribution of PKD2 in LNCaP, DU145, and PC3. As shown in Fig. 3B (4–6), we observed granular and diffuse cytoplasmic PKD2 staining in all three cell types examined with less in the nucleus. Taken together, our data showed cell type– and isotype-specific expression of the PKD isoforms, and increased expression and nuclear accumulation of PKD3 in the more aggressive and metastatic prostate cancer cells.

PKD3 promoted growth and survival of prostate cancer cells. Phorbol ester treatment has been shown to induce apoptosis in LNCaP cells in a PKCα- and δ-dependent manner (26, 27). To test whether PKD3 plays a role in the process, LNCaP cells were infected with empty (Adv) or wild-type PKD3 (Adv-WT) adenoviruses at 5 multiplicity of infection (MOI). Apoptosis was triggered by incubation with 100 nmol/L PMA for 1 hour. As shown in Fig. 3C, PMA induced ∼20% apoptosis in LNCaP cells as examined by terminal deoxynucleotidyl-transferase–mediated dUTP nick-end labeling (TUNEL) assay and DAPI staining, which was comparable with a previous report (26). Overexpression of PKD3 caused a nearly 6-fold reduction in apoptosis as examined by both assays. Western blotting confirmed the overexpression of PKD3 and showed that ectopically expressed PKD3 did not alter the endogenous levels of PKD1 and PKD2 (Fig. 3C , top right). Taken together, these results indicated that overexpression of PKD3 blocked PMA-induced apoptosis in LNCaP cells, implying that increased PKD3 expression promotes cell survival.

Increased PKD3 may confer growth advantages to androgen-independent prostate cancer cells. To test this hypothesis, PC3 cells were transfected with a control nontargeting siRNA (si-nt) and two PKD3 siRNAs (si-PKD3-1 and si-PKD3-2) that target different regions of PKD3 sequence. Approximately 80% of endogenous PKD3 were depleted by si-PKD3-1 and si-PKD3-2. The si-PKD3-1 targeted PKD3 specifically because levels of endogenous PKD2 were not altered (Fig. 3D,, lane 2), whereas si-PKD3-2 also depleted ∼50% of endogenous PKD2 in addition to the knockdown of PKD3. Cell proliferation was determined by counting cell numbers for 6 consecutive days. As shown in Fig. 3D (left), proliferation of the cells transfected with either si-PKD3-1 or si-PKD3-2 was drastically inhibited compared with that of the cells transfected with nontargeting siRNA. The antiproliferative effect of si-PKD3-2 was more potent, possibly due to its dual effect at depleting both PKD3 and PKD2. Overall, our results support a significant role of PKD3 in the growth of prostate cancer cells.

To further evaluate the effect of PKD3 expression on cell growth, we performed cell cycle analysis on PC3 cells with altered PKD3 expression. PC3 cells depleted of endogenous PKD3 were subjected to Flow Cytometry analysis. As shown in Supplementary Fig. S2, depletion of PKD3 by the PKD3 siRNAs resulted in accumulation of cells with G₀-G₁ phase DNA content at the expense of a decrease in S phase cells. This effect was more prominent for si-PKD3-2, correlating to its greater effect at inhibiting cell proliferation (Fig. 3D). Taken together, our data suggested that PKD3 promoted cell growth by accelerating G₀-G₁ to S phase transition.

PKCε regulated PKD3 kinase activity and nuclear localization. PKD is activated via phosphorylation at the activation loop by PKC in intact cells. To identify the specific PKC isoforms involved in the activation of PKD3, PC3 cells were cotransfected with PKD3-GFP and constitutively active PKCα (CA-α), PKCε (CA-ε), and PKCζ (CA-ζ). PKD activation-loop phosphorylation was evaluated by blotting with the p-S744/748-PKD antibody. As illustrated in Fig. 4A (top left), CA-ε significantly enhanced the activation-loop phosphorylation of both ectopically expressed PKD3-GFP (lane 4, top band) and endogenous PKDs (lane 4, bottom band), whereas CA-α and CA-ζ had no effect. Expression of CA-α, CA-ε, and CA-ζ (truncated catalytic domain of PKCζ) was confirmed by Western blotting. Next, PKD3-GFP was immunoprecipitated by a GFP antibody and subjected to in vitro kinase assay using syntide-2 as substrate. PKD3-GFP coexpressed with CA-ε exhibited a 3-fold higher kinase activity compared with those expressed alone or coexpressed with CA-α and CA-ζ (Fig. 4A,, right graph). To evaluate the regulation of endogenous PKD by PKCε, PC3 cells were either pretreated with a pan-PKC inhibitor RO 31-8220 or transfected with a PKCε siRNA. Cell lysates were first analyzed for PKD activation-loop phosphorylation by the p-S744/748-PKD antibody. Reducing endogenous PKCε partially blocked, whereas pretreatment with RO 31-8220 almost completely abolished, basal PKD phosphorylation (Fig. 4B,, left). Endogenous PKD3 was then immunoprecipitated from PC3 cells and subjected to in vitro kinase assay. As shown in Fig. 4B (right), pretreatment with RO 31-8220 abolished >60% of basal endogenous PKD3 activity, whereas depleting PKCε resulted in >30% reduction of endogenous PKD3 activity. As controls, RO 31-8220 treatment did not alter endogenous PKCε expression, and PKCε siRNA depleted ∼90% endogenous PKCε. Taken together, these results indicated that PKCε selectively regulated PKD3 kinase activity in prostate cancer cells.

To examine whether PKCε also regulates the localization of PKD3, CA-ε was cotransfected with the wild-type (PKD3-GFP) or the kinase-inactive PKD3 (K605N-PKD3-GFP) in PC3 cells. The CA-ε–positive cells were identified by immunostaining using the PKCε antibody. The nucleus/cytoplasm ratios of PKD3-GFP with or without coexpressed CA-ε were quantified. As shown in Fig. 4C, PKD3-GFP alone distributed evenly throughout the cytoplasm and nucleus. Coexpressing CA-ε led to the accumulation of PKD3-GFP in the nucleus, causing an overall 3.5-fold increase in nuclear fluorescent intensity of PKD3-GFP. In contrast, CA-ε did not affect the nuclear distribution of K605N-PKD3-GFP, which was primarily cytoplasmic and excluded from nuclei of the PC3 cells (Fig. 4C , bottom left). Similar results were obtained in DU145 cells (data not shown). These results indicated that PKCε promoted nuclear accumulation of PKD3.

PKD3 modulated Akt and ERK1/2 activity in prostate cancer cells. In LNCaP cells, PMA-induced apoptosis is attributed in part to PMA-induced down-regulation of Akt activity (26). Accordingly, constitutive active Akt is capable of rescuing PMA-induced apoptosis in LNCaP cells (1, 26–28). The relevance of the antiapoptotic effects of PKD3 to Akt signaling was evaluated upon PKD3 overexpression in LNCaP cells. As shown in Fig. 5A, PMA treatment down-regulated phospho-Akt levels in cells infected with increasing MOIs of the control adenovirus [Adv; p-S473-Akt (2) blot; Fig. 5A; refs. 26, 27]. Increasing PKD3 expression by Adv-WT PKD3 adenovirus not only increased basal Akt activity but also blocked PMA-induced down-regulation of Akt activity [p-S473-Akt (1) blot; Fig. 5A], correlating to its effects at blocking PMA-induced apoptosis. Increased PKD3 expression also slightly enhanced PMA-induced ERK1/2 activation (p-ERK1/2 blot; Fig. 5A). In contrast, the levels of native Akt and ERK1/2 proteins were not altered.

The effects of PKD3 on the duration of ERK1/2 and Akt activation were further evaluated. As shown in Fig. 5B, PMA induced rapid and prominent ERK1/2 activation, which peaked at ∼40 minutes and gradually reduced to near basal level after 120 minutes of the PMA treatment. The increased PKD3 expression led to rapid ERK1/2 activation, which peaked at 15 minutes and sustained up to 2 hours after the phorbol ester treatment. On the other hand, PMA induced transient down-regulation of phospho-Akt in cells infected with the control adenovirus, which was visible at 5 minutes, peaked 20 minutes, and return to basal level after 90 minutes of PMA treatment, agreed with a previous report (26). In contrast, no visible down-regulation of p-S473-Akt was observed in cells infected to wild-type PKD3 adenovirus. As controls, total ERK1/2 and Akt expression was not altered. PKD3 expression was confirmed by blotting with PKD3 antibody. Similarly, the overexpression of PKD3 also prolonged PMA-induced ERK1/2 activation in PC3 and DU145 cells (Supplementary Fig. S3). Taken together, PKD3 modulated the magnitude and duration of Akt and ERK1/2 activities, potentially accounting for the growth/survival-promoting effects of PKD3 in prostate cancer cells.

To determine whether PKD3 regulates Akt activity in androgen-independent prostate cancer cells, PC3 cells were infected with Adv and Adv-WT PKD3 adenoviruses. As illustrated in Fig. 5C, increasing PKD3 expression dose-dependently enhanced basal levels of p-S473-Akt and p-ERK1/2 in PC3 cells. In contrast, increased PKD3 expression did not have any appreciable effect on the expression of native Akt and ERK proteins. Similar results were obtained in DU145 cells (data not shown).

The role of PKD3 in regulating basal Akt activity was evaluated by knocking down endogenous PKD3 using siRNA in PC3 cells. As shown in Fig. 5D, the basal p-S473-Akt level was significantly reduced in cells transfected with specific PKD3 siRNA compared with those in the control cells. Western blotting confirmed that endogenous PKD3 was effectively depleted (Fig. 5D , bottom). As controls, the expression of Akt and ERK proteins was not altered. Thus, endogenous PKD3 contributed in part to basal Akt activity in PC3 cells.

To determine whether the effect of PKD3 on Akt and ERK was specific to prostate cancer cells, we conducted the experiment in the benign prostate epithelial cells—PrEC. As illustrated in Supplementary Fig. S4, overexpression of PKD3 also increased Akt phosphorylation in PrEC. However, this response is less potent compared with those in PC3 or DU145, considering the high levels of PKD3 expressed in PrEC. Also, greater PKD3 expression (30 MOI) did not further enhance the levels of Akt activation, which was distinct from the responses observed in PC3 and DU145 (Figs. 5C and 6C). Interestingly, the levels of p-ERK1/2 were not altered upon PKD3 overexpression in PrEC. As controls, Akt and ERK1/2 expression remained the same. Thus, PKD3 also modulates Akt activities in benign PrEC cells, although the effects vary in magnitude.

Akt activation induced by increased PKD3 expression was dependent on PI3K and p38 but independent of ILK in PC3 and DU145 cells. Akt activity can be modulated by many kinases and phosphatases in intact cells. To identify those that contribute to the up-regulation of Akt by PKD3, the effects of multiple kinase and phosphatase inhibitors on the process were evaluated. PC3 and DU145 cells infected with Adv-WT PKD3 adenovirus were incubated with a range of inhibitors as indicated in Fig. 6A. Preincubation with the PI3K inhibitors wortmannin or LY294002 completely abolished PKD3-induced Akt activation, indicating that phosphatidylinositol-3-OH kinase (PI3K) is required for the response. Interestingly, the p38 inhibitor (SB203580) also produced a partial effect, whereas other protein kinase and phosphatase inhibitors had no significant effects.

To confirm the involvement of p38, PC3 and DU145 cells with overexpressed PKD3 were treated with increasing concentrations of SB203580. As shown in Fig. 6B, SB203580 abolished PKD3-induced Akt activation in a dose-dependent manner with 50% inhibition at ∼10 to 20 μmol/L for PC3 cells and from 5 to 10 μmol/L for DU145. As controls, Akt expression was not affected by the inhibitor treatment. The involvement of p38 in the process was further evaluated by a dominant-negative p38 (p38-DN). As shown in Fig. 6C, PKD3-induced Akt activation was significantly blunted by p38-DN. The effect was most apparent when comparing the phospho-Akt levels in the control vector and the p38-DN–transfected cells expressing comparable levels of wild-type PKD3 (Fig. 6C , lane 2 versus 9, 3 versus 10, 4 versus 11, and 5 versus 12). The expression of p38-DN and PKD3 was confirmed by Western blotting. These data provided further support for the input of p38 in the regulation of Akt by PKD3.

The potential roles of direct Akt regulators have also been examined, including the integrin-linked kinase (ILK) that has also been shown to directly phosphorylate Akt at S473 (29). As shown in Supplementary Fig. S5, increasing DN-ILK expression or depleting endogenous ILK did not significantly alter PKD3-induced Akt phosphorylation, indicating that ILK was not required for PKD3-induced Akt activation.

PKD, a family of novel diacylglycerol/phorbol ester receptors, has been implicated in hyperproliferative disorders and cancer. However, its role in prostate cancer remains obscure. In this report, the relevance of PKD3, a less known member of the PKD family, to prostate cancer was investigated. Our study identified a novel PKCε/PKD3 pathway that controlled the nuclear localization of PKD3 and regulated downstream ERK1/2 and Akt activities in prostate cancer cells. These findings support a potential role of PKD3 in the progression of prostate cancer, and PKD may be a novel therapeutic target for the prognosis and treatment of this disease.

Based on our data in tissue arrays and tumor cell lines, PKD3 was up-regulated in the malignant tumors and the more aggressive tumor cell lines relative to its expression in the secretory epithelial cells of the normal prostate glands and the benign PrEC cell line. Statistically significant correlation was found between nuclear retention of PKD3 and high-grade tumors, which coincided with the predominant nuclear distribution of PKD3 in the more metastatic prostate cancer cells. These findings indicate a potential role of PKD3 nuclear localization in the advancement of prostate cancer. It is well-recognized that isoforms of the PKD family shuttle between the cytoplasm and nucleus (30–32), and the nuclear accumulation of PKD couples to its activation by PKC (30, 31). However, the functional consequence of this intracellular redistribution event was not fully understood. This report is the first that links PKD3 nuclear retention to cancer progression. The potential functional relevance of PKD3 to prostate cancer progression was supported by its antiapoptotic and pro-proliferative effects in prostate cancer cells, as shown in our study. Additionally, in a preliminary xenograft study using a PC3 line with stable PKD3 knockdown, we found that depleting endogenous PKD3 in prostate tumor cells could significantly suppress the growth of tumors in mice, which supports our general conclusion that aberrant PKD3 expression/localization promotes prostate cancer progression.⁴

PKD1, similar to PKD3, was up-regulated in prostate cancer, although substantial positive PKD1 staining (52%) was also noted in normal tissues. PKD isoforms exhibited different patterns of expression in the benign PrEC and various prostate tumor cell lines, with PKD2 predominantly expressed in the benign PrEC, PKD1 in the less aggressive LNCaP and its isogenic lines, and PKD3 primarily in the more metastatic DU145 and PC3. Meanwhile, the localization of PKD isoforms differs among these cell lines. It is plausible that isoforms of PKD may be differentially expressed and localized in normal prostate and different stages of prostate cancer. Although these isoforms may function redundantly, each isoform could play stage-specific roles in prostate cancer genesis and progression.

The oncogenic effects of PKCε in prostate cancer have been well-documented. Of those relevant to our findings, studies by Terrian and colleagues (5) have shown that depleting endogenous PKCε by an antisense oligo inhibits cell proliferation in DU145 and PC3. In addition, work by Verma and colleagues show that knockdown of PKCε in DU145 cells blocks cell invasion, potentially by binding directly and activating signal transducers and activators of transcription 3 (Stat 3; ref. 4). Besides Stat 3, other targets such as caveolin-1, Bax, integrin signaling, and PI3K/Akt pathway have been identified as the downstream mediators for PKCε-induced cellular responses in prostate cancer (5–7, 33). Here, we have identified PKD3 as an additional immediate downstream effector of PKCε in prostate cancer cells. The expression and function of PKD3 coincided with those of PKCε, and PKCε directly regulated the activity and nuclear localization of PKD3 in tumor cells. PKD3 exhibited progressively elevated nuclear distribution in prostate tumors, possibly as the consequence of its constitutive activation by PKCε. These findings are consistent with the hyperactivation of a PKCε signaling cascade in prostate cancer (4, 6) and further indicate a potential role of a constitutively active PKCε/PKD3 signaling pathway in the advancement of prostate cancer. It is possible that the hyperactive nuclear PKD3 modulates gene expression critical for tumor transformation and progression. Nuclear targets of PKD3 that are important for cell growth and survival, such as the transcription factor cyclic AMP–response element binding protein (34), sphingosine kinase 2 (35), class IIa histone deacetylases (36, 37), have been described and their roles as downstream targets of PKD3 in prostate cancer progression remain to be determined, as well as the targets of PKCε described above. Another potential downstream target of PKD in prostate cancer is the NF-κB signaling pathway. PMA induces NF-κB activation, and PKD1 has been shown to promote cell survival through activating NF-κB signaling pathway in response to oxidative stress (16, 38). It is possible that the prosurvival effects of PKD3 in LNCaP cells in response PMA treatment is in part attributed to the activation of NF-κB.

The PI3K/Akt and the ERK1/2 MAPK pathways play important roles in the initiation and progression of prostate cancer (39–42). Here, we showed that PKD3 promoted Akt and ERK activities in prostate cancer cells. The activation of Akt by PKD3 seemed to require p38 because a p38 inhibitor or a dominant-negative p38 significantly reduced the PKD3-induced Akt activity. It has been reported that PI3K-dependent p38 kinase activation regulates Akt phosphorylation through MAPK-activated protein kinase 2 (43). Our findings provide the initial evidence for the possible involvement of this regulatory pathway in modulating Akt activity in prostate cancer. Overexpression of PKD3 also increased Akt activity in benign PrEC cells, indicating that this response does not seem to be restricted to prostate cancer cells. However, the magnitude of the response was much less in PrEC compared with those in tumor cells and did not result in ERK1/2 activation.

It is well-recognized that both the intensity and duration of ERK activation could be altered to affect ERK-dependent cellular responses (44, 45). Here, we showed that PKD3 modulated both the extent and duration of ERK1/2 activation in all three prostate cancer cell lines examined, which coincides with the finding that PKD1 promotes cell cycle progression and cell proliferation through prolonged ERK activation (12). The target of PKD3 in this process may include the Ras effector RIN1, a PKD substrate, which acts by enhancing Ras and downstream ERK1/2 activity (46).

In summary, we showed that PKD3 expression and nuclear localization was up-regulated in human prostate cancer. PKD3, through a PKCε/PKD3 pathway, regulated downstream Akt and ERK1/2 signaling and promoted growth and survival of prostate cancer cells. Our study highlights the potential value of PKD3 as a novel therapeutic target and biomarker for prostate cancer.

No potential conflicts of interest were disclosed.

Grant support: National Institute of Diabetes and Digestive and Kidney Diseases Grant 1 R01 DK066168-01 and the National Cancer Institute Grant CA115498-03.

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 Shiyuan Chen at the Department of Pathology at the University for Pittsburgh for his technical advice and assistance in immunohistochemical analysis and Dr. Chuanyue Wu in the Department of Pathology at the University of Pittsburgh for providing the DN-ILK adenovirus used in this study.