![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Vol. 17, Issue 12, 5141-5152, December 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
-dependent Mechanism
*Institut National de la Santé et de la Recherche Médicale U752, Collège de France, 75005 Paris, France; Department of Pathology-Neurooncology, Hopital Sainte-Anne, 75674, Paris Cedex 14, France; Institut National de la Santé et de la Recherche Médicale U691, Collège de France, 75005 Paris, France; Institut National de la Santé et de la Recherche Médicale U421, Faculté de Médecine, 94010 Creteil, France; ||Institut National de la Santé et de la Recherche Médicale U542, Hopital Paul Brousse, 94807 Villejuif Cedex, France; and ¶Cancer Research Center of Hawaii, University of Hawaii at Manoa, Honolulu, HI 96813
Submitted November 22, 2005;
Revised August 31, 2006;
Accepted September 7, 2006
Monitoring Editor: Anne Ridley
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
mediated PEA-15 inhibition of astrocyte migration. PEA-15–/– astrocytes constitutively expressed a 40-kDa form of PKC that was down-regulated upon PEA-15 reexpression. Together, these data reveal a new function for PEA-15 in the inhibitory control of astrocyte motility through a PKC-dependent pathway involving the constitutive expression of a catalytic fragment of PKC. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). In contrast to astrocytes, gliomal cells exhibit high motility. Their ability to infiltrate rapidly the normal brain parenchyma limits the efficacy of surgical resection and targeted radiotherapy. Thus, understanding the intracellular mechanisms underlying cell migration within the CNS is of great importance for prognosis and development of therapeutics.
Control of cell motility has been shown to depend on protein kinase C (PKC) activity. PKCs form a family of at least 11 isoforms. Based on their structural and biochemical properties, these PKC isoforms can be divided into three major groups: 1) classical PKCs (
, 
1, 2, and 
), which are activated by diacylglycerol (DAG) and are Ca2+ dependent; 2) novel PKCs (nPKC: , 
, 
, 
, and µ), which are activated by DAG but are Ca2+ independent; and 3) the atypical PKCs (
, 
, and 
), which do not respond to either DAG or calcium. PKC isozymes have been implicated in the three independent but highly coordinated cellular processes involved during tumor cell migration: 1) cell attachment to extracellular matrix or basement membrane (Fagerholm et al., 2002); 2) cell motility, which involves the reorganization of the actin cytoskeleton (Iwabu et al., 2004); and 3) cell invasion, through extracellular matrix degradation by proteolytic enzymes (Woo et al., 2004). Accordingly, inhibition of PKC has been shown to inhibit cell motility and invasiveness (Kermorgant et al., 2001). Interestingly, phosphoprotein enriched in astrocytes-15 kDa (PEA-15), a major small cytoplasmic astrocytic phosphoprotein, is a PKC substrate. Furthermore, PEA-15–induced insulin resistance in type 2 diabetes results from a dysregulation of the balance between the activities of PKC and (Condorelli et al., 2001).
PEA-15 was first identified as an abundant phosphoprotein in brain astrocytes (Araujo et al., 1993); subsequently, it was shown to be widely expressed in different tissues and highly conserved among vertebrates (Danziger et al., 1995; Estelles et al., 1996). It is composed of an N-terminal death effector domain and a C-terminal tail of irregular structure that contains the serines phosphorylated by PKC (S104) or by Akt/calcium/calmodulin-dependent protein kinase II (CaMKII) (S116) (Renault et al., 2003). PEA-15 inhibits apoptosis (Estelles et al., 1999; Condorelli et al., 1999; Kitsberg et al., 1999; Hao et al., 2001). In human malignant glioma cell lines, PEA-15 expression and phosphorylation on its PKC site are required for resistance to tumor necrosis factor-related apoptosis-inducing ligand-mediated apoptosis (Hao et al., 2001). PEA-15 also modifies extracellular signal-regulated kinase (ERK) signaling by controlling ERK subcellular localization (Formstecher et al., 2001; Whitehurst et al., 2004) and thereby restricts cell proliferation. Recently, the phosphorylation of PEA-15 was reported to determine whether PEA-15 binds ERK or FADD (Krueger et al., 2005; Renganathan et al., 2005). Therefore, PEA-15 occurs as a regulator protein that controls cell apoptosis as well as proliferation, two cell functions dysregulated in cancer. PEA-15 involvement in cancer is complex. In transformed and metastatic murine squamous carcinoma cells, the pea-15 gene is up-regulated (Dong et al., 2001) and PEA-15 overexpression favors skin tumors (Formisano et al., 2005). In contrast, a tumor suppressor function was recently reported for PEA-15 in cellular models of breast and ovary tumors (Gaumont-Leclerc et al., 2004; Bartholomeusz et al., 2006). This led us to investigate human tumors arising from astrocytes.
Here, examining human glioblastomas, we observed ex vivo that cells migrating away from the tumor core express low levels of PEA-15, regardless of the expression level in the original tumor. This prompted us to explore the role of PEA-15 in the motility of astrocytes. We report that loss of PEA-15 expression results in an increased astrocyte motility by a PKC-dependent mechanism related to the constitutive expression of a novel 40-kDa form of PKC.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
was obtained from AbCys (Paris, France). PEA-15 (Ab-7) rabbit polyclonal antibody has been characterized previously (Sharif et al., 2004). PKC (C20), PKC (C15), PKC (C15), PKCµ (C20), and PKC (C18) polyclonal antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). PKC mouse monoclonal antibody (clone 21, ref 610085) was from BD Biosciences Transduction Laboratories (Lille, France). PKC mouse antibody (2059), phospho-PKC (Thr505), myosin light chain (MLC) and phospho-MLC (serine 19) antibodies were from Cell Signaling Technology (Ozyme, St. Quentin-en-Yvelines, France). Mouse green fluorescent protein (GFP) antibody was from Roche Diagnostics (Mannheim, Germany). MIB (Ki-67) and glial fibrillary acidic protein antibodies were from Dako Denmark (Glostrup, Denmark).
Plasmids
GFP-PKC Construct. Rat PKC cDNA was provided by Prof. Peter Parker (CRUK Institute for Cancer Studies, London Research Institute, London, United Kingdom). Rat PKC open reading frame was amplified by polymerase chain reaction (PCR) by using 5'-CGGAATTCATGGCACCGTTCCTGCGCA-3' as forward primer and 5'-GGGGGTACCCTATTCCAGGAATTGCTC-3' as reverse primer. The product of PCR was then digested with EcoRI and KpnI restriction enzymes and was cloned into pEGFP C2 vector (Clonetech, Mountain View, CA). GFP-PEA-15 (Kitsberg et al., 1999) and GFP-CF-PKC (DeVries et al., 2002) have been described previously.
Cell Culture
Glial cultures were prepared as described previously (Araujo et al., 1993). Briefly, culture dishes were coated with 1.5 µg/ml poly-L-ornithine (mol. wt. 40,000; Sigma). Striatal and cortical cells from 1-d-postnatal wild-type or PEA-15–/– mice were dissociated and plated in culture defined minimal essential medium (MEM)-F-12 medium consisting of a 1:1 mixture of MEM and F-12 nutrient (Invitrogen, Cergy Pontoise, France) supplemented with 33 mM glucose, pH 7.4, 2 mM glutamine, 3 mM sodium bicarbonate, 5 mM HEPES, 50 IU/ml penicillin, 5 µg/ml streptomycin (all from Invitrogen), and 10% fetal bovine serum (Perbio Science France, Brebières, France). The culture medium was changed every 3 d. When cells were almost confluent, around day 7, cells were treated by 5 µM cytosine arabinoside for 2 d.
Tumor Explants Collection and Culture
Glioblastomas (n = 11) were collected by an anatomopathologist in the surgical room and examined immediately by smears technique (Beuvon et al., 2000) to make sure that sample contained a majority of tumor cells. These samples were either directly fixed by formolzinc solution (formaldehyde 2% with 8 g/l NaCl, 3 g/l ZnSO4) for an histopathological control or cut by 1 cm3 in a transport medium and then minced in 1-mm3 fragments and maintained on surgical sponge of gelatin (Gelfoam; Pharmacia and Upjohn, Guyancourt, France; also available upon request) in RPMI 1640 medium (Sigma) containing 2 g/l sodium bicarbonate, 100 IU/ml penicillin, 100 IU/ml streptomycin supplemented with 6% fetal calf serum at 37°C, and 5% CO2 from 10 to 30 d. The blocks of gelatin containing tumor explant were then fixed and paraffin embedded, before being cut into 4-µm-thick sections by using a microtome (Microm HM340E; Electron Microscopy Sciences, Fort Washington, PA).
Incubation of the Ab-7 Antibody with Glutathione S-Transferase (GST) or GST-PEA-15 Proteins
PEA-15 Ab-7 antibody diluted 1/1000 in phosphate-buffered saline (PBS) buffer (without Ca2+ or Mg2+, pH 7.4; Sigma) was incubated 24 h at 4°C, with rotation, with 1.5 µg of GST-PEA-15 or GST adsorbed on glutathione-Sepharose beads.
Immunohistochemistry
Tissue sections were first deparaffinized before being incubated in 0.3% H2O2 in methanol for 5 min. Tissue sections were then washed twice in PBS before being immunostained for 1 h at room temperature with the indicated antibodies diluted in PBS containing 0.3% Triton X-100. Immunohistochemical detection was achieved using the avidin–biotin complex immunoperoxidase technique and the diaminobenzidine chromogen (Vector Laboratories, Burlingame, CA). Hematoxylin and eosin (Mayer hematoxylin; Merck, Darmstadt, Germany) staining was then performed to visualize cells nuclei and bodies, respectively.
Immunocytochemistry and Confocal Laser Scanning Microscopy
PEA-15 immunocytochemistry was performed as described previously (Sharif et al., 2004). Actin staining was performed using phalloidin coupled to Alexa 488 (Invitrogen). Nuclei were stained with TO-PRO-3 iodide (Invitrogen) according to manufacturer's specifications. Cells were examined using a Leica TS2 (Leica, Wetzlar, Germany) confocal microscope with appropriate filters.
Wound Scratch Assay
Astrocytes were replated between day 11 and 13 in vitro on polyornithine-coated 12-well culture dishes (Falcon; BD Biosciences Discovery Labware, Bedford, MA). The next day, a scratch was done in the confluent monolayer by using a sterile pipette tip. Then, cells were washed three times with PBS buffer (without Ca2+ or Mg2+, pH 7.4; Sigma) and replaced in MEM-F-12 medium supplemented with 10% fetal calf serum or 50 ng/ml TGF for the indicated times. Pictures of marked fields were taken using an inverted phase contrast microscope (Nikon Diaphot) at different time intervals. By using Lucia software (Laboratory Imaging, Hostiva, Czek Republic), areas of fields that were not yet colonized by cells were measured.
Astrocyte Adhesion Assay
Immulon-2 96-well plates were coated overnight (4°C) with 10 µg/ml human fibronectin. Uncoated control wells were blocked with 2% heat-inactivated bovine serum albumin (BSA) (Sigma) for 4 h (4°C). Wild-type and knockout astrocytes were subsequently added (2 x 105 cells/well) in serum-free media and incubated for 60 min (37°C). Wells were washed twice with PBS before be fixed for 20 min with 2% glutaraldehyde (Electron Microscopy Service Laboratories, Westmont, NJ). The fixed cells were washed once with PBS and then stained with 0.1% crystal violet (Serva Biochemicals, Hauppauge, NY) for 45 min. Adherent astrocytes were washed three times with PBS, solubilized in 0.1 N sodium citrate (Sigma) for 30 min, and absorbance at 595 nm was determined using an ELISA plate reader (Molecular Devices, Sunnyvale, CA). To determine 100% of cells for quantitation, a set of wells was fixed and stained without being washed. Data represent experiments done in triplicate.
Transwell Assay
The underside of the polycarbonate membranes of transwells (6.5 mm in diameter, 10 µm in thickness, 8-µm pores; Corning Life Sciences, Acton, MA) were coated with 10 µg/ml human fibronectin for 3 h at 37°C. Primary astrocytes were trypsinized at day 13 in vitro, stained with trypan blue, and counted using a hemocytometer. Cells (2 x 104) were plated onto the uncoated topside of the filter in MEM-F-12 medium in presence of drugs for 45 min (for 2 h with leptomycin B), before adding 50 ng/ml TGF. The inferior well contained same concentrations of drugs and TGF. After 12 h of migration at the incubator (37°C, 5% CO2), cells were fixed 20 min at room temperature with paraformaldehyde (4% in PBS, pH 7.5). The topside of the transwell was washed, and each filter was swabbed with a cotton-tipped applicator to remove cells that did not migrate through the filter. Cells were stained using Hoechst dye (sanofi-aventis, Bridgewater, NJ) according to manufacturer's specifications. Filters were cut out and mounted on slides in Fluoromount-G medium (Southern Biotech, United Kingdom, obtained through Cliniscience, Maubeuge, France). By using Lucia software (Laboratory Imaging) connected to an epifluorescence microscope, 15 fields per membrane were counted. For PMA-induced PKC down-regulation experiments, astrocytes were treated 48 h by 1 µM PMA before migration assay.
Purification of Protein Transducer Domain (PTD)-Fusion Proteins
PTD-PEA-15 and PTD-GFP constructs have been described previously (Caron et al., 2001; Embury et al., 2001). The isolation and purification were done as described previously (Embury et al., 2001). Titration of proteins obtained was performed comparing increasing volumes of PTD-fusion protein with standards of BSA loaded on a 15% polyacrylamide gel and then stained with Coomassie blue. Astrocytes were treated with 4 µM PTD-fusion proteins for 2 h before trypsinization and during the duration of migration.
Immunoblotting
After treatment, cells were rinsed twice with PBS and then scraped on ice in lysis buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% NP-40, 2 mM EDTA, 2 mM EGTA, and 2 mM sodium orthovanadate, supplemented with phosphatase inhibitor cocktails 1 and 2 (Sigma) and Mini Complete protease inhibitor cocktail without EDTA (Roche Diagnostics) following manufacturer's specifications. Cell lysates were passed thrice through a 26-gauge needle and then centrifuged at 13,000 x g during 20 min at 4°C. After protein titration by microBCA protein assay (Pierce Chemical, Rockford, IL), supernatants were mixed with sample buffer (Laemli), boiled, and then separated on 10% polyacrylamide gels (8 and 13% for blotting PKCµ and PEA-15, respectively). After transfer to a polyvinylidene difluoride membrane (Immobilon-P; Millipore, Billerica, MA) blots were probed with indicated primary antibodies before visualizing with horseradish peroxidase-conjugated secondary antibodies followed by development with an enhanced ECL kit (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). In the seven glioblastoma biopsies examined, a strong PEA-15-immunoreactive signal was observed in tumor cells of the tumor solid core (Figure 1A). However, in agreement with the well-known heterogeneity of glioblastomas, we also observed some areas of the tumor core that were not PEA-15 positive. The tumor cells were identified by a histopathologist with regard to their morphological characteristics and their positive MIB immunolabeling, the latter identifying proliferating cells (Figure 1A, inset). Specificity of this labeling was confirmed through the preabsorption of the PEA-15 antibody with a GST-PEA-15 fusion protein that suppressed any immunohistochemical staining (Figure 1B), whereas preabsorption with GST alone did not extinguish the signal (Figure 1C). In the parenchyma surrounding the tumor solid core, which corresponds to the infiltrative component of the tumor, both tumor cells and reactive astrocytes were positive for PEA-15 (Figure 1D). Whereas in reactive astrocytes the immunohistochemical staining was always localized in the cytoplasm, some tumoral cells exhibited in addition a nuclear localization of the signal. Because in these tumors only a few tumor cells exhibited a positive MIB immunolabeling, it remained difficult to identify every malignant cell inside the tumor infiltrative component, thus preventing the in vivo study of PEA-15 expression in tumor cells.
|
), whereas they allow monitoring of the robust invasive capacities of the tumor cells that characterize high-grade gliomas (Sorour et al., 1975). Accordingly, tumor cells, identified by their MIB-positive immunolabeling (Figure 2C), migrated away from the explant and infiltrated the depths of the sponges in 11 of the 12 samples studied. Macrophages, distinguished from tumor cells by their very different morphology (round and voluminous cytoplasm) as well as their CD68-immunohistochemical staining also infiltrated the depths of the sponges, as observed in tumors in vivo (our unpublished data). The level of the PEA-15–immunoreactive signal in the glioma explants varied from high to low, but it was always detectable (Figure 2A). Regardless of the intensity of the PEA-15–immunoreactive signal in the explant, all cells migrating away exhibited weak or no PEA-15 immunoreactivity (Figure 2B). These observations led us to test whether PEA-15 expression can alter cell motility.
|
). We thus compared motility of astrocytes grown in primary culture from either PEA-15 knockout or wild-type mice. Cellular motility was first analyzed at the edge of a scratch wound made across an astrocyte monolayer in the presence of fetal calf serum (FCS). As expected, wounding induced migration of the remaining cell sheet into the gap (Figure 3A). Wild-type astrocytes extended lamellipodial protrusions in the direction of migration (Figure 3B, left). Wounding induced much more drastic morphological changes in PEA-15–/– astrocytes, which emitted long oriented protrusions (Figure 3B, right). PEA-15–/– astrocytes recolonized the wound faster than their wild-type counterparts (Figures 3A, right, and C) (p < 0.05 at 32 h and p < 0.001 at 48 h).
|
Because some discrepancies concerning the effect of other proteins on migration, e.g., p27 (Assoian, 2004), were likely related to the experimental procedures used, we further confirmed the PEA-15 effect on astrocyte migration, by using another migration assay, the transwell assay. In the transwell assay, in agreement with the results obtained using the wound scratch assay, in the presence of 10% FCS PEA-15–/– astrocytes presented a significant 1.6 ± 0.58-fold enhanced migration compared with wild-type astrocytes (Figure 4A). On a second type of stimulation, in the presence of 50 ng/ml TGF, which has been previously described as a strong stimulator of astrocyte motility (Faber-Elman et al., 1996), PEA-15–/– astrocytes migrated 1.9 ± 0.65-fold more than their wild-type counterparts (Figure 4A). Considering the more robust effect of PEA-15 in the presence of TGF, further studies were done using TGF as a migration inducer. Whereas the wound assay mimics a directed migration of cells still contacting their neighboring cells, the transwell assay, which requires trypsinization of cells, measures haptokinesis toward fibronectin of independent cells. The influence of PEA-15 on cell migration thus seems to be relatively independent of the extracellular context. The transwell assay was used for further studies because this migration assay reveals a difference of migratory capability between both types of astrocytes in less time (12 h) compared with the wound scratch assay.
|
). It was therefore necessary to rule out the possibility that the enhanced migration of PEA-15–/– astrocytes was related to an epiphenomenon, rather than to a direct effect of PEA-15. We thus analyzed PEA-15–/– astrocyte motility by using the transwell assay after reintroducing PEA-15. The poor efficiency of cDNA transfection for astrocytes in primary cultures led us to restore PEA-15 expression by using a fusion protein composed of the PTD and PEA-15 (PTD-PEA-15) (Embury et al., 2001). The 11 amino acids of the PTD sequence allow the transduction into living cells of fused proteins dissolved in aqueous solution (Schwarze et al., 1999). Neither the fusion protein control PTD-GFP, at any concentration used, nor a low dose of 0.4 µM PTD-PEA-15 had a significant effect on astrocyte motility. Treatment by 4 µM PTD-PEA-15, although having no significant effect on wild-type astrocyte migration, led to a major reduction of the extent of the migration of PEA-15–/– astrocytes (Figure 4B). Western blotting experiments confirmed the efficiency of the transduction by PTD-PEA-15 (Figure 4C). It is noteworthy that although the theoretical weight of PTD-PEA-15 is 16 kDa, the apparent molecular weight on a polyacrylamide gel of the protein is 26 kDa, probably due to the presence of numerous charged amino acids in the PTD sequence. Together these data demonstrate that expression of PEA-15 inhibits astrocyte motility. In addition, transient transfection of NIH-3T3 cells by GFP-PEA-15 cDNA drastically reduced migration on transwell assay, whereas transfection by GFP alone had no effect (Figure 4D), thus indicating that the control exerted by PEA-15 on motility is not restricted to astrocytes.
The Antiproliferative and Antimigratory Functions of PEA-15 Are Distinct
PEA-15 inhibits astrocyte proliferation (Formstecher et al., 2001). The increased number of migrating PEA-15–/– astrocytes could thus be a consequence of increased cell division. To investigate this hypothesis, migration assays were done in the presence of 10 µg/ml aphidicholin, a mitotic inhibitor reported to have no effect on the motility of astrocytes stimulated by basic fibroblast growth factor (Milner et al., 1999). 5-Bromo-2-deoxyuridine incorporation experiments confirmed that aphidicholin completely inhibited both wild-type and PEA-15–/– astrocyte proliferation for at least 48 h (our unpublished data), a period of time longer than the duration of the transwell experiments (12 h). Inhibition of proliferation by aphidicholin did not alter the migration of either wild-type or PEA-15–/– astrocytes (Figure 5A). This result demonstrates that the enhanced migration of PEA-15–/– astrocytes is not a consequence of their increased proliferation.
|
). PEA-15 contains a nuclear export sequence required for its capacity to dynamically localize ERK in the cytoplasm (Formstecher et al., 2001). As reported previously, treatment of wild-type astrocytes with 10 ng/ml leptomycin B, a potent inhibitor of the nuclear export machinery protein CRM1 (Nishi et al., 1994; Ossareh-Nazari et al., 1997) led to the relocalization of PEA-15 in the nucleus, as observed by confocal microscopy (our unpublished data). In the presence of leptomycin B, the motility of wild-type astrocytes was not modified (Figure 5B). It thus seems that the mechanisms whereby PEA-15 controls cell migration differ from those involved in its control of cell proliferation.
ERK/Mitogen-activated Protein Kinase (MAPK), Phosphatidylinositol 3-Kinase/Akt, and CaMKII Are Not Required for the PEA-15 Effect on Migration
Our cell motility assays use FCS or TGF as inducers of motility. Both FCS and TGF are known to activate MAPK/ERK and PI3K/Akt pathways. We therefore tested whether these pathways are involved in the effect of PEA-15 on cell migration. The presence of 10 µM U0126, a specific inhibitor of mitogen-activated protein kinase kinase (Favata et al., 1998) during migration experiments reduced similarly by 
50% the migration extent of both cell types (Figure 6) (52 ± 13.5 and 48 ± 6.0% for wild-type and PEA-15–/– astrocytes, respectively). Therefore, the migration of astrocytes in presence of TGF depends on ERK/MAPK activity, but this signaling pathway is not involved in the specific effect of PEA-15. PI3K inhibition by 30 µM LY294002 decreased the migration extent of wild-type astrocytes by 15.4 ± 3.6% and did the same for PEA-15–/– astrocytes (–21.7% ± 2.5%) (Figure 6), demonstrating that Akt activity is also required for full astrocyte motility but that it does not interfere with the specific effect of PEA-15 on cell migration. Next, we tested the CaMKII, which can regulate cell migration (Pauly et al., 1995). The inhibition of CaMKII by 10 µM KN-62 did not significantly modify the migration of either wild-type or PEA-15–/– astrocytes. (Figure 6), suggesting that CaMKII is not involved in astrocyte motility.
|
). We thus investigated their role in the effect of PEA-15 on astrocyte migration. In the presence of 5 µM bisindoylmaleimide (BIM), a broad-spectrum inhibitor of PKC, the migration of PEA-15–/– astrocytes was reduced by 38.3 ± 11.9%, matching the extent of migration of wild-type astrocytes, whereas the migration of these latter was not significantly affected (Figure 7A). This result implicates a PKC in the enhanced migration observed for PEA-15–/– astrocytes. The isoforms of PKC are classified in three families based on their activation requirements. Long-term treatment (48 h) with 1 µM PMA is known to desensitize PKCs belonging to the classical and novel PKC families, through down-regulation of PKC protein levels (Blackshear, 1988). PMA treatment inhibited dramatically the migration of PEA-15–/– astrocytes by 45 ± 4.9% (Figure 7B), allowing PEA-15–/– astrocytes to return to the same migration rate as their wild-type counterparts. This result indicates that the PKC isoform implicated belongs to either the classical or the novel PKC family. Treatment of astrocytes with 100 nM Gö6976, previously shown to inhibit selectively the family of classical PKCs (Martiny-Baron et al., 1993), did not alter the migration of either PEA-15–/– or wild-type cells (Figure 7C), therefore excluding the participation of these classical PKCs and thereby demonstrating that PEA-15 inhibition of astrocyte motility is mediated by a PKC belonging to the novel PKC family.
|
Mediates PEA-15 Inhibition of Astrocyte Migration; Emoto et al., 1995; Datta et al., 1997; Endo et al., 2000), we therefore explored nPKC expression by using antibodies recognizing epitopes located in the C-terminal part of the enzymes. All nPKC isoforms were expressed in wild-type astrocytes. PKC, PKCµ, and PKC were similarly expressed in wild-type and PEA-15–/– astrocytes (Figure 8A). Western blot analysis revealed a weak down-regulation of endogenous 78-kDa PKC in PEA-15–/– astrocytes in comparison with their wild-type counterparts (ratio knockout/wild type = 67 ± 8%, n = 3; Figure 8B). In addition, we observed a strong down-regulation of PKC in PEA-15–/– astrocytes that was confirmed using a second antibody recognizing an epitope located in the N-terminal domain of the enzyme (our unpublished data).
|
, one isoform corresponding to the holoenzyme, and the other isoform corresponding to the low-molecular-weight C-terminal fragment. Regarding PKC, use of the antibody directed against the PKC phospho-threonine 505, located in the catalytic moiety, revealed the presence only in PEA-15–/– astrocytes of a 40-kDa form of PKC.
PKC and PKC have both been reported to control cell migration (Tang et al., 1997, 1999; Kruger and Reddy, 2003; Li et al., 2003; Iwabu et al., 2004); therefore, we determined their implication in the control exerted by PEA-15 on astrocyte migration.
We first performed dose-response experiments in presence of rottlerin, initially described as a specific inhibitor of PKC (Gschwendt et al., 1994) that was further shown to inhibit PKC (Villalba et al., 1999). Very low doses of rottlerin, 0.1 and 0.3 µM, had a significant inhibitory effect only on the migration of PEA-15–/– astrocytes, and they abolished the difference between wild-type and PEA-15–/– astrocyte motilities (Figure 9A). The down-regulation of PKC in PEA-15–/– astrocytes implies that the enhanced motility of PEA-15–/– astrocytes relies on an increased enzymatic activity associated to PKC. This was further confirmed by coapplication of 300 nM bryostatin-1, known to specifically protect PKC during PMA-induced down-regulation, contrary to all other PMA-sensitive PKC isoforms that are down-regulated (Szallasi et al., 1994a, b; Lu et al., 1997). In this condition, PEA-15–/– astrocytes conserved their enhanced migration (Figure 9B). Together, these results indicate that PEA-15 inhibits astrocyte migration by inhibiting a PKC-dependent stimulatory pathway.
|
in wild-type and PEA-15–/– astrocytes. Probably because of low expression levels of PKC in astrocytes, we were unable to detect endogenous PKC by immunocytochemistry. We thus transfected GFP-PKC into wild-type and PEA-15–/– astrocytes and confocal analysis revealed a mainly cytoplasmic localization in both cases (Figure 9C). In addition, GFP-PKC did not colocalize with the cytoskeletal proteins actin and tubulin or with the focal adhesion protein vinculin (our unpublished data).
To more precisely test the involvement of the 40-kDa form of PKC observed in PEA-15–/– astrocytes in their enhanced migration, we examined the effect of PTD-PEA-15 transduction of PEA-15–/– astrocytes. Whereas treatment with PTD-GFP had no effect, reexpression of PEA-15 led to the total disappearance of the 40-kDa form of PKC in PEA-15–/– astrocytes (Figure 9D).
Phosphorylation of the threonine 505 located in the activation loop is considered as a reliable marker of PKC activation (Stempka et al., 1999; Rybin et al., 2003). Lack of a specific immunoprecipitating antibody against the 40-kDa isoform prevented a direct kinase assay to precisely quantify the enzymatic activity of the 40-kDa-PKC in PEA-15–/– astrocytes lysates. We next tried to quantify the total PKC enzymatic activity in astrocyte lysates to demonstrate a global increase of the enzymatic activity of PKC in PEA-15–/– astrocytes. Unfortunately, we were unable to immunoprecipitate the protein from astrocyte lysates, thus preventing the determination of this enzymatic activity. This was probably due to the low endogenous expression level of PKC in astrocytes, since we were able to immunoprecipitate PKC from lysates of Chinese hamster ovary cells overexpressing PKC (our unpublished data), wherein the catalytic fragment of PKC is not present.
We thus searched for indirect evidence of modified PKC activity in PEA-15–/– cells. PKC has been shown to mediate a signaling cascade during epidermal growth factor (EGF)-induced cell migration, leading to the phosphorylation of MLC on its serine 19 (Iwabu et al., 2004). Accordingly, we observed by Western blotting an enhanced phosphorylation of the MLC on its serine 19 in PEA-15–/– astrocytes in comparison with their wild-type counterparts (Figure 9E).
Together, these results strongly suggest that the presence of the 40-kDa form of PKC in PEA-15–/– astrocytes is responsible for the enhanced migration of these cells.
A 41-kDa fragment of PKC (catalytic fragment [CF]-PKC) exhibiting a constitutive enzymatic activity has been reported after induction of apoptosis (Emoto et al., 1995; Ghayur et al., 1996; Steinberg, 2004). CF-PKC results from caspase 3-dependent cleavage of the holoenzyme and is proapoptotic. We transfected wild-type and PEA-15–/– astrocytes with a cDNA encoding this CF-PKC fused to GFP. Both types of astrocytes died after apoptosis (our unpublished data). Contrary to CF-PKC, the 40-kDa PKC observed is constitutively expressed in PEA-15–/– astrocytes, and no sign of apoptosis was observed, suggesting they represent two different forms of the kinase.
We envisaged that the 40-kDa PKC expressed in PEA-15–/– astrocytes resulted from a cleavage by a caspase. PEA-15–/– cells were treated by a caspase 3 inhibitor, DEVD (50 µM), or a pan-caspase inhibitor, ZVAD (50 µM), during 12 h, a time sufficient for the disappearance of this fragment upon treatment by PTD-PEA-15 (Figure 9D). Both specific caspase 3 and broad-spectrum caspase inhibition did not modify the expression level of the 40-kDa PKC in PEA-15–/– astrocytes (Figure 10A), whereas the strong decrease of the 40-kDa fragment of PKC observed in both types of astrocytes was used as an internal control of inhibitors efficiency (our unpublished data). Furthermore, broad-spectrum inhibition of serine and thiol proteases by 100 µM leupeptin had no effect on the expression level of the 40-kDa PKC in PEA-15–/– astrocytes (Figure 10B). These results strongly suggested an absence of cleavage of PKC in PEA-15–/– astrocytes. To further test this hypothesis, we transfected a cDNA encoding GFP-PKC, wherein the GFP is fused to the N-terminal extremity of PKC. No cleaved form of GFP-PKC could be detected up to 72 h posttransfection in PEA-15–/– astrocytes (Figure 10C).
|
observed in PEA-15–/– astrocytes does not result from a proteolytic cleavage. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell migration involves two sets of factors. Intracellular factors determine intrinsic cell properties, adhesion, and cell migration, whereas external factors, including extracellular matrix (ECM) and cell–cell interaction, control tissue permissivity to cell migration. A rich field of research revealed the complexity of brain ECM components as well as some of their remodeling during development or pathologies. For example, fibronectin deposited in the ECM of brain tumors strongly promotes the migration of glioma cells (Ohnishi et al., 1998). Interestingly, PEA-15 has been shown to block H-Ras–initiated inhibition of integrin activation (Ramos et al., 1998). PEA-15 binding to ERK is required for this modulation (Chou et al., 2003). Increased integrin activity is known to inhibit migration and would be consistent with PEA-15 inhibition of cell migration. However, comparing wild-type and PEA-15–/– astrocytes we have not seen any difference in integrin activation (Ramos, unpublished data), a result in agreement with the observation that both types of astrocytes have similar adhesion on fibronectin. Further studies should reveal if PEA-15 expression modulates cell migration differently according to the matrix composition.
Most of PKC isozymes have been implicated in cell migration. Western blotting experiments as well as pharmacological manipulations on PKC isozymes allowed us to demonstrate that PEA-15 inhibits astrocyte migration by a PKC-dependent mechanism. PKC has been particularly implicated in cell migration (Kruger and Reddy, 2003; Li et al., 2003; Iwabu et al., 2004). PKC is indeed responsible for a major part of the EGF-induced fibroblast contractile force generation, because RNA interference-mediated PKC depletion prevents MLC phosphorylation, which in turn promotes motility (Iwabu et al., 2004). Increased PKC has also been correlated with enhanced metastatic potential in mammary tumor cells (Kiley et al., 1999).
The expression of a 40-kDa form of PKC (40-kDa-PKC) in PEA-15–/– astrocytes seems to account for their enhanced migration. Indeed, reexpression of PEA-15 in PEA-15–/– astrocytes upon transduction with PTD-PEA-15 simultaneously decreased astrocyte migration and expression of this 40-kDa-PKC. The strong phosphorylation of threonine 505 of 40-kDa-PKC in PEA-15–/– astrocytes suggests a constitutively active form, because this phosphorylation has been reported as a reliable marker of PKC activation (Stempka et al., 1999; Rybin et al., 2003). The enhanced phosphorylation of myosin light chain in PEA-15–/– astrocytes is in agreement with an increased PKC activity. However, the biological effect of this 40-kDa-PKC may also result from phosphorylation of specific substrates that are not affected by the holoenzyme.
Contrary to the proapoptotic CF-PKC previously observed upon apoptosis induction (Steinberg, 2004), the 40-kDa-PKC is constitutively expressed in PEA-15–/– astrocytes. Although PEA-15–/– astrocytes exhibit an increased sensitivity to various inducers of apoptosis (Kitsberg et al., 1999; Renault et al., 2003), no sign of apoptosis was observed in the cells in any of the culture conditions used in this study. Furthermore, apoptosis was induced after transfection of PEA-15–/– astrocytes with GFP-CF-PKC, indicating that they did not develop a resistance to apoptosis and further supporting that the 40-kDa-PKC expressed in these cells differs from CF-PKC. Accordingly, whereas the CF-PKC results from the proteolytic cleavage of the holoenzyme by caspase 3 (Emoto et al., 1995; Ghayur et al., 1996), the lack of effect of the inhibition of various proteases as well as the absence of cleavage of transfected GFP-PKC in PEA-15–/– astrocytes suggest that the 40-kDa-PKC does not result from proteolytic cleavage. The size and the enzymatic activity of the 40-kDa-PKC indicate that it is not encoded by the two alternative splicing variants of PKC that have already been described in mammals (Ueyama et al., 2000; Sakurai et al., 2001). The molecular mechanisms linking PEA-15 expression and 40-kDa-PKC generation remain to be elucidated.
In L6 skeletal muscle cells, PEA-15 action on glucose transport is mediated by an enhanced activation of the classical PKC that in turn inhibits PKC (Condorelli et al., 2001). In astrocytes migration, the inefficiency of Gö6976, the inhibitor of classical PKC, to reverse the difference between wild-type and PEA-15–/– astrocyte motilities, rules out the involvement of PKC. It is noteworthy that, contrary to the desensitization of PKC induced by PMA, which is actually a down-regulation, the inhibition of PKC activity by BIM had no significant effect on wild-type astrocyte migration. This suggests that down-regulation of PKC by PMA abrogates its inhibition of PKC, this latter isoform being known to stimulate astrocyte polarization (Etienne-Manneville and Hall, 2001, 2003) and cell migration (Crean et al., 2004; Petit et al., 2005). Contrary to PMA, BIM should inhibits both PKC and PKC, although less efficiently for this latter, thus possibly explaining the different effects exerted by PMA and BIM on wild-type astrocyte migration. Whether PEA-15 expression modulates PKC signaling, i.e., PKC or PKC, through a common molecular mechanism or whether it is specific of each isoform needs further investigation.
It remains also to be determined whether phosphorylation of PEA-15 can modulate its control of cell migration. The observation that CaMKII or PI3K/Akt inhibition does not interfere with PEA-15 effect on astrocyte migration suggests that phosphorylation of PEA-15 on its serine 116 is not involved. Because PEA-15 contains a serine substrate of PKC (serine 104) (Kubes et al., 1998), a cross-regulation between PKC and PEA-15 may tune astrocytes motility.
In consideration of the different known cellular functions of PEA-15 in different chronic disorders, including cancer (Eramo et al., 2005; Stassi et al., 2005) and diabetes (Condorelli et al., 1998; Condorelli et al., 2001; Vigliotta et al., 2004), the novel observation that PEA-15 inhibits cellular motility may have a very significant impact on future investigation of key areas such as metastasis and diabetes complications.
|
ACKNOWLEDGMENTS |
|---|
construct, Dr. Ricardo Pastori for gift of PTD-PEA-15 construct, Dr. Jacques Bertoglio for gift of PTD-GFP construct, Dr. Denis Hervé for gift of BIM, and Eric Etienne for confocal assistance. We are grateful to Drs. Charles-Felix Calvo, Etienne Formstecher, Laurent Muller, and Catherine Monnot for fruitful discussions. This research was supported by the Association pour la Recherche contre la Cancer (ARC, Grants 3500 and 4621 to H.C.). F.R.M. is a recipient from study fellowships from French Ministry of Research and from the Académie Nationale de Médecine. J.W.R. is supported by National Cancer Institute Grant CA-93849 from the National Institutes of Health. |
Footnotes |
|---|
Address correspondence to: Hervé Chneiweiss (herve.chneiweiss{at}college-de-france.fr)
Abbreviations used: CaMKII, calcium/calmodulin-dependent protein kinase II; CF-PKC, catalytic fragment of protein kinase C; ERK, extracellular signal-regulated kinase; GST, glutathione S-transferase; MAPK, mitogen-activated protein kinase; MLC, myosin light chain; PEA-15, phosphoprotein enriched in astrocytes-15 kDa; PI3K, phosphatidylinositol 3-kinase; PKC, protein kinase C; PMA, phorbol 12-myristate 13-acetate; PTD, protein transducer domain; TGF, transforming growth factor
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Assoian, R. K. (2004). Stopping and going with p27kip1. Dev. Cell 6, 458–459.[CrossRef][Medline]
Bartholomeusz, C., Itamochi, H., Nitta, M., Saya, H., Ginsberg, M. H., Ueno, N.T. (2006). Antitumor effect of E1A in ovarian cancer by cytoplasmic sequestration of activated ERK by PEA15. Oncogene 25, 79–90.[Medline]
Beuvon, F., Varlet, P., Fallet-Bianco, C., Daumas-Duport, C. (2000). The "smears" technique for the extemporaneous examination: diagnostic contribution to neurosurgical pathology. Ann. Pathol 20, 499–506.[Medline]
Blackshear, P. J. (1988). Approaches to the study of protein kinase C involvement in signal transduction. Am. J. Med. Sci 296, 231–240.[Medline]
Caron, N. J., Torrente, Y., Camirand, G., Bujold, M., Chapdelaine, P., Leriche, K., Bresolin, N., Tremblay, J. P. (2001). Intracellular delivery of a Tat-eGFP fusion protein into muscle cells. Mol. Ther 3, 310–318.[CrossRef][Medline]
Chou, F. L., Hill, J. M., Hsieh, J. C., Pouyssegur, J., Brunet, A., Glading, A., Uberall, F., Ramos, J. W., Werner, M. H., Ginsberg, M. H. (2003). PEA-15 binding to ERK1/2 MAPKs is required for its modulation of integrin activation. J. Biol. Chem 278, 52587–52597.
Condorelli, G., Vigliotta, G., Cafieri, A., Trencia, A., Andalo, P., Oriente, F., Miele, C., Caruso, M., Formisano, P., Beguinot, F. (1999). PED/PEA-15, an anti-apoptotic molecule that regulates FAS/TNFR1-induced apoptosis. Oncogene 18, 4409–4415.[CrossRef][Medline]
Condorelli, G., et al. (1998). PED/PEA-15 gene controls glucose transport and is overexpressed in type 2 diabetes mellitus. EMBO J 17, 3858–3866.[CrossRef][Medline]
Condorelli, G., Vigliotta, G., Trencia, A., Maitan, M. A., Caruso, M., Miele, C., Oriente, F., Santopietro, S., Formisano, P., Beguinot, F. (2001). Protein kinase C (PKC)-alpha activation inhibits PKC-zeta and mediates the action of PED/PEA-15 on glucose transport in the L6 skeletal muscle cells. Diabetes 50, 1244–1252.
Crean, J. K., Furlong, F., Finlay, D., Mitchell, D., Murphy, M., Conway, B., Brady, H. R., Godson, C., Martin, F. (2004). Connective tissue growth factor [CTGF]/CCN2 stimulates mesangial cell migration through integrated dissolution of focal adhesion complexes and activation of cell polarization. FASEB J 18, 1541–1543.
Danziger, N., Yokoyama, M., Jay, T., Cordier, J., Glowinski, J., Chneiweiss, H. (1995). Cellular expression, developmental regulation, and phylogenic conservation of PEA-15, the astrocytic major phosphoprotein and protein kinase C substrate. J. Neurochem 64, 1016–1025.[Medline]
Datta, R., Kojima, H., Yoshida, K., Kufe, D. (1997). Caspase-3-mediated cleavage of protein kinase C theta in induction of apoptosis. J. Biol. Chem 272, 20317–20320.
DeVries, T. A., Neville, M. C., Reyland, M. E. (2002). Nuclear import of PKCdelta is required for apoptosis: identification of a novel nuclear import sequence. EMBO J 21, 6050–6060.[CrossRef][Medline]
Dong, G., Loukinova, E., Chen, Z., Gangi, L., Chanturita, T. I., Liu, E. T., Van Waes, C. (2001). Molecular profiling of transformed and metastatic murine squamous carcinoma cells by differential display and cDNA microarray reveals altered expression of multiple genes related to growth, apoptosis, angiogenesis, an
