|Home | About | Journals | Submit | Contact Us | Français|
Glioblastoma multiforme, a highly aggressive tumor of the central nervous system, has a dismal prognosis that is due in part to its resistance to radio- and chemotherapy. The protein kinase C (PKC) family of serine threonine kinases has been implicated in the formation and proliferation of glioblastoma multiforme. Members of the protein kinase D (PKD) family, which consists of PKD1, -2 and, -3, are prominent downstream targets of PKCs and could play a major role in glioblastoma growth. PKD2 was highly expressed in both low-grade and high-grade human gliomas. The number of PKD2-positive tumor cells increased with glioma grading (P < .001). PKD2 was also expressed in CD133-positive glioblastoma stem cells and various glioblastoma cell lines in which the kinase was found to be constitutively active. Inhibition of PKDs by pharmacological inhibitors resulted in substantial inhibition of glioblastoma proliferation. Furthermore, specific depletion of PKD2 by siRNA resulted in a marked inhibition of anchorage-dependent and -independent proliferation and an accumulation of glioblastoma cells in G0/G1, accompanied by a down-regulation of cyclin D1 expression. In addition, PKD2-depleted glioblastoma cells exhibited substantially reduced tumor formation in vivo on chicken chorioallantoic membranes. These findings identify PKD2 as a novel mediator of glioblastoma cell growth in vitro and in vivo and thereby as a potential therapeutic target for this devastating disease.
Malignant gliomas are the most frequent tumors originating in the central nervous system of adults. Glioblastoma multiforme is the most aggressive variant of glioma, and this tumor is largely refractory to current chemotherapy and radiotherapy regimens. Therefore, the discovery of novel therapeutic approaches for this disease is paramount. Serine threonine kinases of the protein kinase C (PKC) family play a role in the formation of brain tumors by mediating apoptosis resistance1 and inducing proliferation.2,3 Conversely, PKC inhibitors have been shown to block growth, migration, and invasion and to induce apoptosis of glioblastoma cells.4,5 The phospholipase C/PKC signalling cascade is also a major signalling module in the autocrine VEGF loop that regulates glioblastoma cell cycle progression, viability, and radioresistance via vascular endothelial growth factor receptor 2 (VEGFR2).6 To date, however, selective inhibitors of PKC isoforms, such as PKCβ, do not appear to have enough single-agent activity to be useful, at least as monotherapy.7 Members of the protein kinase D (PKD) family of serine threonine kinases are prominent downstream targets of PKCs.8–10 The PKD family belongs to the calcium-/calmodulin-dependent protein kinase superfamily11 and comprises PKD1/PKCµ, PKD2, and PKD3/PKCν.12 The catalytic domain of these kinases differs from that of the classic and atypical members of the PKC family. PKDs are activated by various stimuli, including phorbol esters, G protein coupled receptors, and reactive oxygen species,12,13 and also by PKCs that directly activate PKDs via phosphorylation at 2 critical serine residues in the activation loop of their catalytic domain.14
Evidence of the role of PKDs in tumors has thus far been contradictory. In many gastric cancers, PKD1 is down-regulated by CpG hypermethylation, and this could lead to increased metastasis.15 PKD1 also inhibits migration of breast cancer cells by phosphorylation of slingshot.16,17 In contrast, IGF-I promotes myeloma cell migration via activation of PKDs.18 In prostate cancer, there is no CpG island hypermethylation of PKD1.19 PKD1 has been reported to be down-regulated in androgen-independent prostate cancer cells.20 However, more recent studies have reported over-expression of PKD1 and particularly of PKD3 in aggressive prostate cancers compared with normal prostate tissue and have found that PKDs contribute to tumor growth, survival, migration, and invasion.21–23 Other reports have found that PKD1 phosphorylates β-catenin and represses β-catenin–mediated transcriptional activity and prostate cancer cell proliferation.24
The data on a potential role of PKD2 in cancer are more uniform. PKD2 expression/activity correlates positively with the state of de-differentiation in lymphoma,25 and PKD2 has been identified as a kinase involved in HeLa and pancreatic cancer cell survival.26 We previously showed that PKD2 mediates activation of NF-κB in chronic myelogenous leukemia cells.27 In keratinocytes, in which phorbol esters are major tumor promoters, PKDs stimulate proliferation and prevent differentiation.28–30 These data suggest that PKDs—and in particular, PKD2—could play a role in phorbol ester–susceptible tumors, such as skin tumors, and potentially also in glioblastoma.
Here, we examined the expression of PKD2 in low- and high-grade human gliomas and found a positive correlation between the number of PKD2-positive tumor cells and glioma grading, with maximum PKD2 expression in human glioblastoma. We found that PKD2 expression was not only restricted to glioblastoma cells but also detectable in CD133-positive tumor–initiating cells.31–33 In addition, PKD2 was highly expressed in all commonly used glioblastoma cell lines and was found to be constitutively active in these cells. Our data show—to our knowledge, for the first time—that pharmacological inhibition of PKDs or selective depletion of PKD2 by specific siRNAs results in a marked decrease in anchorage-dependent and -independent proliferation of glioblastoma cells accompanied by down-regulation of cyclin D1, but not by apoptosis. In addition, silencing of PKD2 in human glioblastoma cells growing on the chicken chorioallantoic membrane (CAM) blocked tumor formation in vivo. These data show that PKD2 plays a role in glioblastoma tumor formation and could be an interesting novel target for the treatment of this particular tumor family.
Human glioma samples were obtained from the tissue bank of the Department of Pathology, University of Ulm (Ulm, Germany). The samples were pseudonymized to comply with the German law for correct use of archival tissue for clinical research.34 Samples were fixed in 10% buffered formalin and paraffin embedded. The glioma sample set consisted of the following: 9 pilocytic astrocytomas, 9 diffuse astrocytomas, 9 anaplastic astrocytomas, 9 oligodendrogliomas, 10 anaplastic oligodendrogliomas, and 7 glioblastomas (Supplementary Table S1). Human gliomas were classified neuropathologically according to the World Health Organization (WHO) central nervous tissue tumor classification35 and then analyzed for expression of PKD2. Immunohistochemical analysis was performed on 3-mm-thick paraffin-embedded sections using the biotin streptavidin peroxidase LSAB kit (Vector Labs via Biozol). After microwave (20 min) antigen retrieval (antigen-retrieval solution pH 6.0; Dako), an antibody against PKD2 (Epitomics, rabbit monoclonal; 1:200; Biomol) was applied at room temperature, followed by application of the biotin streptavidin reagents in accordance with the manufacturer's instructions and labeled using the alkaline phosphatase substrate kit 1 for PKD2 (Vector Labs). Each section was then counterstained with hematoxylin. A section with skipped primary antibody for each glioma served as a negative control. Tissues were analyzed and photographed using a Leica microscope (Leica Camera).
To assess the intensity of PKD2 staining in different gliomas, we stained a collective of 53 samples for PKD2 via immunohistochemistry. Four samples of normal brain were also investigated. To ensure equal exposure time, the substrate (same lot for all gliomas) was incubated for exactly 10 min. The reaction was stopped by washing with distilled water. Staining intensity of glioma cells was semiquantitatively assessed as follows: 1, low; 2, intermediate; and 3, high. Clinical parameters (age, sex, diagnosis, and relapse) were obtained from the tissue bank databases of the Department of Pathology and Neuropathology, University Hospital of Ulm. Microscopic analysis was performed by a board-certified experienced neuropathologist (D.R.T.).
To assess the percentage of PKD2-positive tumor cells in different human gliomas, the same samples as described above were analyzed. The following groups of PKD2-positive tumor cells were used: <10% positive cells, 10%–50% positive cells, 50%–90% positive cells, and >90% positive cells. Evaluation was performed by an experienced neuropathologist (D.R.T.).
Cytospin preparations of cells on glass slides were air-dried, and surface antigens were fixed with 4% formaldehyde at room temperature for 30 min. Cells were washed with phosphate-buffered saline (PBS) and blocked with 20% normal goat serum (PAA Laboratories) diluted in PBS for 30 min at room temperature. The cells were then incubated with antibody against CD133 (CD133/1[W6B3C1], Miltenyi Biotec; dilution, 1:50) for 2 h at room temperature, washed 3 times with PBS, and incubated with Alexa Fluor 488 goat anti-mouse immunoglobuline (Ig) G antibody (Invitrogen; dilution, 1:400) for 1 h at room temperature. After 3 washes with PBS, the cells were permeabilized with 0.1% Triton diluted in PBS for 10 min, washed 3 times with PBS, and blocked with 20% normal goat serum for 30 min at room temperature. Thereafter, cells were incubated with antibody against PKD2 (Orbigen; dilution, 1:200) overnight at 4°C, washed 3 times with PBS, and incubated with Alexa Fluor 568 goat anti-rabbit IgG antibody (Invitrogen; dilution, 1:400) for 1 h at room temperature. After 3 washes with PBS, nuclei were counterstained with 1 µg/mL 4′,6-diamidino-2-phenylidene for 15 min and mounted in Dako fluorescent mounting medium (Dako). All antibody dilutions were performed in PBS containing 10% normal goat serum. Controls for immunofluorescence included the second-step antibody alone to control for nonspecific binding. All control preparations gave consistently negative results.
Glioblastoma cell lines U87, U138, and U251 were obtained from the American Type Culture Collection and cultured in DMEM (Invitrogen) supplemented with 10% fetal calf serum (FCS) (PAA Laboratories), 1% penicillin/streptomycin (Invitrogen) as described previously. Primary glioblastoma cell cultures (designated GBM) and samples were obtained from patients undergoing surgery in accordance with a protocol approved by the institutional review board. Tumor spheres were cultivated as described elsewhere.32 The inhibitors Gö6976 and Gö6983, as well as the phorbol ester phorbol 12-myristate 13-acetate (PMA), were purchased from Calbiochem; GF109203x was from Sigma.
Primary normal human astrocytes (Applied Cell Biology Research Institute) were cultured in astrocyte medium (Gibco) containing N2-supplement, DMEM basal medium, and fetal bovine serum (One Shot Astrocyte-Certified FBS; Gibco). Cells were cultured on basement membrane matrix mixture-coated plates (laminin, collagen IV, entactin, heparin sulfate proteoglycan; Geltrex, Gibco) and were passaged every 3 days. All described experiments were performed using low-passage (<6) primary normal human astrocytes.
Glioblastoma cells (5 × 105) were seeded in 6-well tissue culture dishes and grown in DMEM/10% FCS. After 24 h, this medium was replaced with FCS-free medium, and cells were transfected on 2 consecutive days with 800 ng of siPKD2 (either siPKD2 Stealth-1761-oligo A [AAUGACCUUAACUGCCACGUCCCGG] or siPKD2 Stealth-283-oligo B [CCUGAGUGUGGCUUCUACG GCCUUU]; both Invitrogen, Karlsruhe, Germany) using TransMessenger reagent (Qiagen) in accordance with the manufacturer's recommendations. Scrambled control siRNA was from Qiagen (#1022076; UUCUCCGAACGUGUCACGU).
Whole-cell extracts (50–100 µg) were subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membranes (Millipore Corp., Bedford, MA, USA). Membranes were blocked using 5% dry milk in PBS containing 0.2% Tween 20. For subsequent washes, 0.2% Tween 20 in PBS was used. Membranes were incubated overnight with specific antibodies at 4°C under shaking conditions. To detect phosphorylated PKD, we used a pPKD Ser744/748 antibody (Cell Signaling, Frankfurt, Germany). The anti-PKD2 antibody was from Upstate (Schwalbach, Germany), and cyclin D1, caspase 3 and PARP antibodies were from Cell Signaling. β-Actin antibody used as a loading control was from Sigma-Aldrich. Goat anti-rabbit secondary antibody was from BioRad Laboratories. Chemiluminescence was detected using a ChemiSmart Chemiluminescence 5000 documentation system (PeqLab Biotechnologie).
Whole-cell extracts (2000 μg) were incubated with 2 µg of PKD2 antibody (Bethyl Laboratories) at 4°C for 1 h, and then 80 μL of protein A sepharose beads was added and incubated for 1 h. Immobilized proteins were washed extensively, resuspended in Laemmli buffer, and subjected to SDS-PAGE and Western blotting, as described above.
Two milliliters of serum-supplemented medium and 0.5% agar (Becton Dickinson) containing either 3 µM Gö6976 or 5 µM Gö6983 or solvent were added at 40°C in a 35-mm culture dish and allowed to solidify (base agar). Next, on top of the base layer, a mixture of serum-supplemented medium and 0.35% agar (total of 2 mL) containing 10,000 glioblastoma cells was added together with the inhibitors or solvent and allowed to solidify (top agar). In the event of PKD2 knockdown, PKD2 was depleted by 2 consecutive rounds of siPKD2 transfection 48 h and 24 h before incubation of cells in the soft agar. Formation of glioblastoma cell colonies was followed for the next 12 days, at which time the final number of colonies was assessed.
Cells were harvested using trypsin/EDTA solution and washed with PBS. For cell cycle analyses, cells were fixed overnight using ice-cold 70% ethanol and then spun down and incubated for 1 h at 37°C in the presence of 50 µg/mL propidium iodide and 100 µg/mL RNAse A in PBS. The cell cycle distribution was recorded by flow cytometry (FACS Calibur; Becton Dickinson) and analyzed using ModFit LT software (Verify Software House).
Shells of fertilized chicken eggs were opened on day 4, and silicon rings (diameter, 5 mm) were applied onto the chorioallantoic membrane. After 2 consecutive siRNA deliveries, 1 × 106 of glioblastoma cells transfected with scrambled siRNA or PKD2-siRNA were transplanted on day 8 to the CAM. After 48 h, tumors were harvested and photographed. Formalin-fixed tumors were further embedded in paraffin, and immunostaining of Ki-67 (Clone MIB-1, Dako) and PKD2 (Epitomics) was performed. Cells were scored as PKD2 positive when they exhibited intense PKD2 immunostaining. To determine the number of PKD2-positive cells, 300 cells in 4 distinct, randomly chosen areas within the tumor were examined.
Data are given as means ± standard errors of the mean (SEM) of 3 independent experiments, unless otherwise stated. To test for statistically significant differences, the paired t test was applied using Sigma Plot software, version 11.0. For categorical data analysis, the Fisher exact test was calculated using PASW-statistics software, version 18.0. Differences were considered statistically significant at either P< .05 or P< .01, as indicated in the figure legends.
PKD2 is expressed in multiple tissues.9,36,37 We examined the expression of PKD2 by immunohistochemistry in human normal brain (Fig. 1) and a set of 53 human gliomas comprising all WHO grades (Fig. 2A and Supplementary Table S1). In normal human brain there was a moderate expression of PKD2 in the cortical neuropil of all layers (I–VI) of the occipital neocortex. Neuronal somata and the white matter exhibited a pronounced PKD2 immunoreactivity (Fig. 1A–C). High-power magnification of layer III of the occipital cortex revealed that PDK2 was moderately detectable in the nerve cell somata (arrows, Fig. 1B) but not in the nuclei. Cortical glial cells did not exhibit a distinct PKD2 expression (Fig. 1B). In the white matter, PDK2 expression was marked in neuronal perikarya (Fig. 1B), axon/myelin sheet–associated material (open arrows, Fig. 1C), and the somata of oligodendroglia and microglia cells but not in the nuclei (arrowheads, Fig. 1C). Astrocytes did not exhibit marked PKD2 immunoreactivity (Fig. 1C). Astroglial, oligodendroglial, and microglial cells were identified by their morphology and nuclear structure. In the cerebellum, there was moderate PDK2 expression in the neuropil of the molecular layer (Fig. 1D) and the granule cell layer, particularly in the cytoplasm of granule cells (Fig. 1D), but not in cerebellar neurons. Purkinje cells in the Purkinje cell layer (Fig. 1D) showed moderate PKD2 expression in their perikaryon. In the cerebellar white matter, there was moderate expression of PKD2.
Next, we sought to determine PKD2 immunoreactivity in gliomas of different grades. Staining intensity of glioma cells was semiquantitatively assessed as low (1), intermediate (2), and high (3). Fig. 2A shows representative photographs of various human gliomas (oligodendroglioma , anaplastic oligodendroglioma , pilocytic astrocytoma , diffuse astrocytoma , anaplastic astrocytoma , and glioblastoma ). PKD2 immunoreactivity was largely detectable in the cytoplasm of the glioma cells (Fig. 2A). In addition, there was strong PKD2 expression in the tumor-associated endothelium, particularly in microvascular proliferations of glioblastomas (Fig. 2A, glioblastoma; Supplementary Table S1). For statistical analysis, the tumors were assigned to a staining intensity group as described in Materials and Methods (Fig. 2B). In most gliomas, there was intermediate to strong PKD2 expression without major differences in PKD2 immunoreactivity with grading (P= .386, by the Fisher exact test). To determine the percentage of PKD2-positive tumor cells, we defined 4 groups with <10%, 10%–50%, 50%–90%, and >90% PKD2-positive tumor cells and assigned all gliomas to these groups (Fig. 2C). Statistical analysis revealed an increase in the percentage of PKD2-positive tumor cells with increasing WHO grade (Supplementary Table S2; P< .001, by the Fisher exact test; P< .001, by the trend test).
PKD2 expression was also detectable in various established glioblastoma cell lines. PKD2 expression was detectable in primary normal human astrocytes but was substantially higher in the human glioblastoma cell lines (Fig. 3A). To further confirm expression of PKD2 in human glioblastoma, we examined PKD2 expression in primary glioblastoma cells and compared it with expression in primary normal human astrocytes. Samples were obtained from patients, propagated in adherent culture, and subsequently analyzed by western blot for PKD2 expression. All but 1 primary glioblastoma cell line exhibited PKD2 expression. Again, PKD2 expression was higher in 4 of 6 primary glioblastomas, compared with primary human astrocytes (Fig. 3B). These results were confirmed by PKD2 immunostaining in adherent cultured glioblastoma cells (data not shown).
Recent studies have demonstrated that tumors are organized, as are normal tissues, in a hierarchical manner. In particular, glioblastomas31,38,39 are believed to originate from and be maintained by cancer stem cells (CSCs). CSCs are capable of continuous self-renewal and multilineage differentiation and have an enhanced capacity to initiate tumor formation in vivo.31,38,39 Therefore, we examined the expression of PKD2 in glioblastoma CSCs. CD133 has been shown to label the CSC population in glioblastomas, and nonadherent growth in neurospheres has been shown to facilitate CSC self-renewal.33 Double-staining for PKD2 and CD133 in sphere-cultured glioblastoma cell lines revealed PKD2 expression in CD133-positive CSCs (Fig. 3C). Confocal microscopic examination of tumor spheres confirmed PKD2 expression in CD133-positive cells (Supplementary Material, Film S1). Thus, PKD2 is expressed in both primary glioblastomas and glioblastoma CSCs.
Having established that PKD2 is expressed in human glioblastoma, we were interested in the activity of the kinase in glioblastoma cells. Activity of PKDs can be determined by assessing the phosphorylation at 2 sites in the activation loop of the catalytic domain, Ser706 and Ser710.9 To examine whether PKD2 was phosphorylated/active in glioblastoma cell lines, PKD2 was immunoprecipitated from lysates of glioblastoma cells and primary normal human astrocytes that were incubated with solvent or PMA, and the immunoprecipitates were further analyzed by the activation-specific phospho-PKD antibody. Indeed, we found that PKD2 was already phosphorylated/active in the activation loop in primary normal human astrocytes and glioblastoma cells under basal conditions, and its phosphorylation was further increased by PMA (Fig. 3D). Gö6976, a selective inhibitor of classic PKCα and PKDs,40 abrogated basal and PMA-stimulated phosphorylation of PKDs (Fig. 4A). The effect of Gö6976 on PMA-induced PKD2 phosphorylation was concentration dependent, with a maximum inhibitory effect at 3 µM of Gö6976 (data not shown).
PKCs—in particular, PKCε and PKCη—are major upstream regulators of PKDs, and they phosphorylate PKDs at the critical serine residues in the activation loop.9,14 To establish PKCs as upstream regulators of PKD2 in glioblastoma cell lines as well, we used the selective inhibitor Gö6983, which inhibits the catalytic activity of classic and novel PKCsη, θ, δ, and ε. Incubation of cells with Gö6983 completely abrogated basal and PMA-stimulated activation of PKD, demonstrating that PKCs also act as upstream regulators of PKDs in glioblastoma cells. Phosphorylation of PKCε in response to PMA was blocked by the PKC inhibitors Gö6983 and the bisindolylmaleimide GF109203X (Fig. 4B). In addition, ectopic expression of PKCε in glioblastoma cells resulted in increased PKD2 phosphorylation (Fig. 4C). Thus, PKD2 is constitutively active in glioblastoma cells and PKCs contribute to its activation.
Next, we were interested in whether PKC and PKD activities contribute to the proliferation of U87, U138, and U251 glioblastoma cells. As shown in Fig. 5A, Gö6983 at 5 µM and GF109203X at 3.5 µM reduced proliferation of U87 and U138 cells by 30%–50% and proliferation of U251 cells by 15%–30%. Interestingly, incubation of cells with 3 µM of the selective PKD inhibitor Gö6976 completely prevented proliferation of all 3 cell lines. This suggests a major role for PKDs in glioblastoma proliferation that goes beyond the role of a mere PKC downstream target. Gö6983 and Gö6976 also significantly reduced colony growth of all 3 glioblastoma cell lines (Fig. 5B and C). Again, in U138 and U251 cells, inhibition of PKDs was more efficient in blocking colony formation than was inhibition of PKCs. Thus, inhibition of PKDs results in significantly reduced anchorage-dependent and -independent proliferation that is even more pronounced than what occurs upon inhibition of PKCs.
Gö6976 is a selective inhibitor of all PKDs. However, only PKD2 is highly expressed in various tumors, suggesting a role of this kinase in carcinogenesis and/or tumor maintenance (Azoitei and Seufferlein, unpublished observations; and Azoitei et al.37). To establish whether depletion of PKD2 alone would mimic the effects of Gö6976, we used specific siRNA oligonucleotides that efficiently reduced PKD2 protein expression by at least 50% in all 3 cell lines. A scrambled, control oligonucleotide had no effect on PKD2 expression (Fig. 6A and B). PKD2 depletion reached a maximum at ~70 h after transfection of the glioblastoma cells (Fig. 6C and data not shown) and had no effect on the expression of PKD1 or PKD3 (Fig. 6D, arrows).
Silencing of PKD2 led to a marked reduction of cell proliferation in all glioblastoma cell lines, comparable to the effect of Gö6976. A scrambled siRNA oligonucleotide had no effect on the proliferation of U87, U138, and U251 cells, compared with untransfected control cells (Control; Fig. 7A). Depletion of PKD2 also significantly reduced the number of glioblastoma colonies growing in soft agar and thereby anchorage-independent growth of the tumor cells. Again, the effect of PKD2 depletion by siRNA was even more pronounced than the pharmacological inhibition of PKD2 by Gö6976. A scrambled control oligonucleotide had no effect on colony formation (Fig. 7B and C).
Our data showed that PKD2 depletion inhibits anchorage-dependent and -independent growth of glioblastoma cells. Interestingly, PKD2 phosphorylation is up-regulated in the M phase of the cell cycle.41 These data suggest that the kinase could play a role in cell cycle progression. To further analyze a potential effect of PKD2 on the cell cycle of glioblastoma cells, we performed cell cycle analyses in the presence and absence of PKD2. Depletion of PKD2 in glioblastoma cells resulted in an accumulation of cells in the G0/G1 phase and a decrease of cells in the S phase of the cell cycle. This suggests an arrest of the cell cycle in G0/G1 upon depletion of PKD2 (Fig. 8A). In line with this finding, depletion of PKD2 resulted in a substantial reduction of cyclin D1 expression in the glioblastoma cells (Fig. 8A). In contrast, depletion of PKD2 did not result in increased caspase 3 or PARP cleavage as indicators of apoptosis (Fig. 8B). Thus, PKD2 regulates cell cycle progression of human glioblastoma cells—in particular, G1/S progression—most likely via inducing cyclin D1 expression, but not by inhibiting apoptosis.
Having determined that PKD2 is required for glioblastoma cell growth in vitro, we evaluated the effect of PKD2 silencing in glioblastoma tumor formation in vivo using the CAM model, an established model of in vivo tumor formation.42 U87 cells were transfected with either a scrambled oligonucleotide or PKD2 siRNA and subsequently seeded on the CAM of chicken embryos and allowed to form tumors. There are only 2 PKD isoforms in chicken tissues corresponding to human PKD1 and PKD3, but no isoform with close similarity to human PKD2.43 Owing to the substantial differences in the nucleotide sequence between human and chicken PKDs, the oligonucleotide used exclusively targets human PKD2 and has no effect on chicken PKDs (data not shown). U87 cells transfected with a scrambled oligonucleotide formed substantial tumors on the CAM (Fig. 9A). The specific siRNA oligonucleotide efficiently reduced PKD2 expression in U87 glioblastoma cells during the course of the experiment (Fig. 9C and D). In contrast to the scrambled oligonucleotide, depletion of PKD2 in U87 cells largely prevented the formation of glioblastoma tumors on the CAM in vivo (Fig. 9A and B). This corresponded to a significantly reduced proliferation index, as evidenced by the number of Ki-67–positive tumor cells (Fig. 9E and F).
Glioblastoma is the most common human brain tumor and is often lethal. Phorbol ester–susceptible PKC isoforms have been implicated in invasion and chemotherapy resistance of glioblastoma and have been proposed as novel therapeutic targets for this disease.1–5 Serine threonine kinases of the PKD family also act as phorbol ester receptors and constitute, at the same time, major downstream effectors of PKC signaling.44 PKD2 is highly expressed in diffuse large B cell and anaplastic large cell lymphoma25 and has been implicated in tumor cell survival.26,27 However, the precise role of PKD2 in tumor biology is still elusive. Our data demonstrate that PKD2 is expressed in gliomas of various grades, primary and established glioblastoma cell lines, and CSCs. Furthermore, PKDs—and in particular, PKD2—were found to be constitutively phosphorylated/active in glioblastoma cell lines. Pharmacological inhibition of PKD activity blocked glioblastoma cell proliferation even more efficiently than PKC inhibitors. This could indicate that, in glioblastoma, PKDs are not only PKC downstream targets but can also play a PKC-independent role in signaling. However, pharmacological inhibition of protein kinases using allegedly selective inhibitors has certain disadvantages. For example, we observed that Gö6976, which inhibited PKD phosphorylation slightly less efficiently than did Gö6983, was more potent than this compound in inhibiting cell growth and colony formation. Therefore, the data obtained with pharmacological inhibitors were confirmed by experiments using PKD2-specific siRNAs. These experiments showed that inhibition of colony formation by selective knockdown of PKD2 was even more pronounced than inhibition of PKCs or PKDs using pharmacological inhibitors. These results demonstrated that PKD2 alone is required for both anchorage-dependent and -independent growth of glioblastoma cells in vitro. At the level of the cell cycle, depletion of PKD2 in glioblastoma cells induced arrest of tumor cells in the G1 phase of the cell cycle and down-regulation of cyclin D1. At the same time—and in contrast to selective PKC inhibitors, such as enzastaurin—depletion of PKD2 did not lead to the induction of apoptosis in glioblastoma cells. PKD2 was required not only for glioblastoma cell growth in vitro but also for tumor formation on the CAM in vivo, indicating a potentially critical role for this kinase in glioblastoma tumorigenesis. To date, data in the literature regarding the expression of different PKD isoforms in various tumors and their role in tumor growth, survival, migration, and invasion are controversial,15,17–21,23–30,45 indicating that the action of PKDs in tumor biology is likely to be both tumor specific and isoform specific. The fact that specific depletion of PKD2 was sufficient to block tumor growth in vitro and in vivo also demonstrates that the PKD isoforms do not act in a redundant fashion but play distinct roles in cellular signalling, despite a comparatively high sequence homology. This conclusion is supported by data from Yeamen et al.,46 who demonstrated that the different PKD isoforms fulfil specific functions in the regulation of intracellular protein transport. Our in vitro data, the fact that PKD2 promotes tumorigenesis in the CAM model, and the positive correlation of PKD2-expressing tumor cells with tumor grading point to a growth advantage for glioblastoma cells when they express PKD2 and suggest that the kinase could play a role in both carcinogenesis as well as tumor maintenance. Thus, PKD2 could constitute an interesting biomarker and a novel therapeutic target in a substantial subgroup of highly malignant gliomas.
This article is dedicated to the late Heinrich Baust, a longstanding collaborator and friend who also contributed to this article.
Work in the laboratory of T.S. is supported by the DFG (SE676/9-1 and SFB 518/B3). A.K. is supported by a fellowship provided by the Medical Faculty of Ulm University (Bausteinprogramm, L.SBR.0011). We would like to thank Ralf Köhntop for excellent technical assistance and Dr. Matthias Oertel (University of Giessen, Neurosurgery) for providing primary normal human astrocytes.