Research Article
Oncostatin M Induces Growth Arrest by Inhibition of Skp2, Cks1, and Cyclin A Expression and Induced p21 Expression 1
1
2
2,3
Hartmut Halfter, Matthias Friedrich, Ansgar Resch, Michael Kullmann, 1 1 2,3 Florian Sto¨gbauer, E. Bernd Ringelstein, and Ludger Hengst 1
Department of Neurology, Westfa¨lische Wilhelms-Universita¨t Mu¨nster, Mu¨nster, Germany; 2Max Planck Institute for Biochemistry, Martinsried, Germany; and 3Division of Medical Biochemistry, Biocenter, Medical University of Innsbruck, Innsbruck, Austria
Abstract Oncostatin M has been characterized as a potent growth inhibitor for various tumor cells. Oncostatin M–treated glioblastoma cells cease proliferation and instigate astrocytal differentiation. The oncostatin M–induced cell cycle arrest in G1 phase is characterized by increased level of the cyclindependent kinase (CDK) inhibitory proteins p21Cip1/Waf1/Sdi1 and p27Kip1 . Induction of p21 protein corresponds to increased mRNA level, whereas p27 accumulates due to increased stability of the protein. Interestingly, stabilization of p27Kip1 occurs even in S phase, showing that p27 stabilization is a direct consequence of oncostatin M signaling and not a result of the cell cycle arrest. Degradation of p27 in late G1 and S phase is initiated by the ubiquitin ligase complex SCF-Skp2/Cks1. Oncostatin M inhibits expression of two components of this E3 ligase complex (Skp2 and Cks1). Although combined overexpression of Skp2 and Cks1 rescues p27 degradation in S phase, it can not override p27 accumulation in G1 phase and cell cycle arrest by oncostatin M. In addition to increasing Cdk inhibitor level, oncostatin M also impairs cyclin A expression. Cyclin A mRNA and protein level decline shortly after oncostatin M addition. The accumulation of two CDK inhibitor proteins and the repression of cyclin A expression may explain the broad and potent antiproliferative effect of the cytokine. (Cancer Res 2006; 66(13): 6530-9)
Introduction Oncostatin M is a member of the interleukin-6 (IL-6) family cytokines that modulate differentiation, proliferation, and apoptosis of a wide variety of cell types (1). The cytokine exerts its biological activities through heterodimeric receptors composed of gp130 and either the oncostatin M–specific OSMRh or the LIFR (1). Ligand-induced dimerization of the receptor components initiates the activation of members of the Janus family of protein kinases (Janus-activated kinase, JAK). Receptor-associated JAK kinases phosphorylate and thereby activate members of the signal transducer and activator of transcription (STAT) family of transcription factors, which translocate into the nucleus and regulate expression of cytokine-specific target genes (1). The response to oncostatin M depends on the cellular context, and it
Note: H. Halfter and M. Friedrich contributed equally to this work. M. Friedrich is currently at the Institute for Biochemistry II, Universita¨tsklinikum Frankfurt, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany. Requests for reprints: Ludger Hengst, Division of Medical Biochemistry, Innsbruck Medical University, Fritz Pregl Str. 3, Innsbruck, Austria 6020. Phone: 43512-507-3270; Fax: 43-512-507-2872; E-mail:
[email protected]. I2006 American Association for Cancer Research. doi:10.1158/0008-5472.CAN-04-3734
Cancer Res 2006; 66: (13). July 1, 2006
has been shown that oncostatin M inhibits proliferation of many tumor derived cells (1). In other cell types (e.g., fibroblasts), oncostatin M can stimulate proliferation (1). Previously, we observed that oncostatin M induces a potent growth arrest in human glioblastoma cells (2) and efficiently prevents tumor formation in a mouse glioblastoma model (3). Progression through the cell cycle is controlled by oscillating activities of cyclin-dependent kinases (CDK; refs. 4, 5). The catalytic activity of CDK/cyclin complexes can be controlled by various mechanisms, including cyclin synthesis and degradation, activating or inhibitory phosphorylation, localization, or by binding of CDK inhibitory proteins (6, 7). During G1 progression, two families of CDK inhibitory proteins play a key role in regulating CDK activity of mammalian cells: INK4 proteins specifically bind to CDK4 and CDK6, whereas members of the Cip/Kip family (p21Waf1/Cip1/Sdi1 , p27Kip1 , and p57Kip2 ) bind to and regulate a broader spectrum of CDK complexes (8). Although Ink4 proteins always inhibit CDK kinase activity, Cip/Kip proteins may also act as assembly factors for cyclin D/CDK4/6 complexes (6). Whereas expression of p21 is regulated at the level of mRNA expression (9), p27 levels are frequently regulated posttranscriptionally. p27 plays a central role in regulation of the unperturbed cell cycle by regulating Cdk activity during the G0-G1-S progression. Its expression is regulated by diverse mitogenic or antimitogenic signals (8, 10). Changes in p27 frequently involve regulation of its synthesis and proteolysis (11, 12). At least two distinct pathways regulate p27 stability at different stages of the cell cycle (13), which both involve the ubiquitin/proteasome pathway. p27 becomes increasingly unstable as cells approach S phase. Proteolysis of p27 is initiated at this time by a CDK-dependent phosphorylation of p27 at Thr187. Phosphorylated p27 is subsequently ubiquitinylated by the SCF-Skp2 complex. This E3 ubiquitin ligase is composed of Skp1, Cul1, p45Skp2, and ROC1/Rbx1 and requires association with Cks1 for its ubiquitin ligase activity towards p27 (14, 15). Cks1 is a member of the highly conserved Suc1/Cks protein family originally identified as subunits of CDK complexes and associates with p27 and Skp2 (16). To efficiently degrade p27, the Cul1 subunit needs to be modified with the ubiquitinrelated protein Nedd8 (17). This modification dissociates the inhibitory subunit p120-CAND1 from Cul1 (18), and deneddylation of Cul1 by the COP9 signalosome inhibits p27 degradation (19, 20). Although the COP9 signalosome complex can stabilize p27, overexpression of its subunit CSN5/Jab1 can stimulate p27 degradation through p27 nuclear export (21). A second pathway of p27 elimination during G0-G1 progression is independent of T187 phosphorylation but also involves the ubiquitin/proteasome pathway (22, 23). Monomeric p27 is ubiquitinated by the KPC1/ KPC2 ligase independently from T187 phosphorylation and subsequently degraded by the proteasome (24, 25).
6530
www.aacrjournals.org
Oncostatin M and Glioblastoma Cell Cycle Progression
p45Skp2 is a member of a large family of F-box proteins and rate limiting for p27 degradation at the G1-S transition. A complex of Skp2 and Cks1 interacts with Thr187-phosphorylated p27, leading to p27 ubiquitination by SCF-Skp2 complex (26). Although SCF-Skp2 has been reported to ubiquitinate a number of additional substrates, including hOrc 1p, E2F-1, or p130, its role in p27 degradation is central for its function as an oncogene (26). The antagonistic relationship between p27 and Skp2/Cks1 is reflected in knockout animals and human tumors. Although mice lacking p27 are larger than wild-type animals due to multiorgan hyperplasia (27, 28), mice nullizygous for Skp2 or Cks1 are smaller than their littermates and accumulate elevated level of p27 (14, 29). Many phenotypes of the Skp2 knockout animals are reverted to normal by loss of the single substrate p27 (30, 31), suggesting that p27 is the central substrate of this ubiquitin ligase, and that Skp2 plays a nonredundant role in p27 degradation. This hypothesis is supported by the observation that elevated expression of Skp2 correlates with low level of p27 in various human tumors (26). Skp2 itself is regulated during the cell cycle and degraded by the ubiquitin-proteasome pathway (32). Ubiquitin-dependent degradation of Skp2 in G1 phase is mediated by the APC/CCdh1 complex (33, 34). Skp2 regulation does also involve transcriptional mechanisms. Androgen treatment of human prostate carcinoma cells or detachment can lead to reduced Skp2 mRNA expression and p27 stabilization (35, 36). The p27-related Cdk inhibitor p21 is also an unstable protein and degraded by the 26S proteasome. However, in contrast to p27, ubiquitination was reported to be not essential for p21 proteasomal degradation (37, 38). Ubiquitination of p21 can also involve the SCF-Skp2 pathway during the S phase (39) and switches from the basal turnover to SCF-Skp2 dependency upon DNA damage (40). The aim of this study was to elucidate molecular mechanisms that allow for the broad antimitogenic function of oncostatin M. We found that oncostatin M induces the CDK inhibitors p21 and p27. Although p21 induction involves increased mRNA expression, p27 accumulates due to stabilization of the protein. Decreased degradation of p27 is based on reduced Skp2 and Cks1 expression. Whereas Skp2 overexpression is usually sufficient to eliminate p27 and promote S-phase entry, ectopic expression of Skp2 in oncostatin M–treated cells could surprisingly neither destabilize p27 nor rescue cell proliferation. Combined overexpression of Skp2 and Cks1 is capable of reducing p27 induction but fails to abolish the cell cycle arrest. The sustained inability of cells to enter S phase despite overexpression of Skp2 and Cks1 can be explained by reduced expression of cyclin A. These redundant pathways converging in inhibiting Cdk kinase activity can explain the strong antiproliferative potency of oncostatin M and its broad spectrum of susceptible tumor cells.
Materials and Methods Cell culture. Maintenance and treatment of the human glioblastoma cell lines 86HG39 was done as recently described (3). For stable expression of Skp2, glioblastoma cells were transfected with a pcDNA3-mycSkp2 (32). Single clones were analyzed for Skp2 expression after selection with 400 Ag/ mL neomycin. Stable cell lines expressing Skp2 and Cks1 were generated by transfecting Skp2 expressing clones with a Cks1 cDNA cloned into pIRES2a vector (Stratagene, Heidelberg, Germany) containing a hygromycin resistance gene. Stable expression of cyclin A was achieved by cloning the cDNA of human cyclin A in the pIRES2a/hygroR vector (Stratagene), and after transfection, cells were selected with hygromycin, and single clones were expanded.
www.aacrjournals.org
Cell cycle analysis. 86HG39 glioblastoma cells or its derivatives cells expressing Skp2, Cks1, and Skp2 or cyclin A were cultured for 12 hours in DMEM (Life Technologies, Gaithersburg, MD) supplemented with 5% FCS (Sigma, Deisenkirchen, Germany)/1% penicillin/streptomycin before oncostatin M was added to a concentration of 25 ng/mL. Cell cycle distribution was determined after bromodeoxyuridine (BrdUrd) incorporation by two-dimensional fluorescence-activated cell sorting (FACS) analysis. Cells (3 105) were seeded in 6-cm cell culture dishes. Before harvesting, cells were incubated in the presence 30 Ag/mL BrdUrd (Sigma) for 45 minutes. Cells were prepared and stained for BrdUrd incorporation using FITClabeled anti-BrdUrd antibody according to manufacturer’s instructions (Becton Dickinson, Mountain View, CA) and for DNA content with propidium iodide, and cells were analyzed using FACSCalibur and Cell Quest Software (Becton Dickinson). The thymidine-induced S phase arrest was confirmed by staining of the fixed cells with propidium iodide. Immunohistochemical analysis. For anti-BrdUrd immunocytochemistry, 2 104 cells were seeded on glass coverslips in 48-well plates. Treatment was done as described above. Before harvest cells were incubated with 3 ng/mL BrdUrd for 60 minutes. Fixation and staining was done according to 5-Bromodeoxyuridine Labeling and Detection kit I instructions (Roche, Mannheim, Germany). Before mounting on microscope slides, cell nuclei were stained with Hoechst 33258 (Molecular Probes, Eugene, OR). Samples were analyzed on Leitz DMRX fluorescence microscope, and pictures were digitally acquired with a Spot CCD camera (Visitron, Mu¨nchen, Germany). Quantification of BrdUrd-positive cells was done using ANALYSIS software (Soft Imaging System, Mu¨nster, Germany). Northern blot analysis. Total RNA was isolated from 1 106 glioblastoma cells using the Trizol reagent (Life Technologies). Northern blotting was done as described (41). Polyadenylated RNA was purified from total RNA using magnetic beads from Miltenyi Biotech (BergischGladbach, Germany) according to the manufacturer’s recommendation. The membranes were hybridized consecutively with 32P-labeled DNA probes consisting of a 700-bp BamHI fragment of the human Skp2 gene, a 870-bp PvuII-SphI fragment of the human cyclin A gene, and a 450-bp PCR fragment from the human GAPDH gene. p27 degradation assay. 86HG39 cells were cultured for 24 hours in the presence or absence of oncostatin M. For determination of p27 protein stability, 2 105 cells were seeded on 6-cm dishes. The following day, cycloheximide (Sigma) was added to the culture medium ( final concentration, 100 Amol/L) to inhibit protein synthesis. At times indicated following the addition of cycloheximide, cells were lysed, and p27 protein levels were determined by Western blot analysis. Western blot analysis. 86HG39 cells or derivatives thereof were seeded for most experiments in subconfluent cultures on 6-cm dishes and were treated as indicated. Cell lysates and Western blotting were done as described recently (42). The antibodies used in this study are as follows: anti-p27 (clone 57), anticyclin A (C88020), anti-cyclin D1 (DCS-6), anti-pRb (R68320), and p21(clone 70) were obtained from BD PharMingen (San Diego, CA); anti-CDK2 (sc-163), anti-CDK4 (sc-260), anti-CDK6 (sc-177), anti-cyclin E (sc-481), and anti Cks1 were from Santa Cruz Biotechnology (Santa Cruz, CA); anti-Skp2 (2B12) and anti-cullin-1 (Rb-042) were from Neomarkers (Fremont, CA); anti-phospho-T187-p27 (06-996) was obtained from Upstate (Lake Placid, NY); and anti-actin (clone Ac-15) and anti-h-tubulin (clone SAP. 4G5) were from Sigma (Deisenkirchen, Germany). Antibodies against cyclin A (T310) and cyclin E (HE 172) were described earlier (11, 12). For the quantitative expression analysis of p27, Western blots were processed as described before. The quantification of the blots was achieved by directly recording chemiluminescence emission with a ChemDoc XRS System (Bio-Rad, Munich, Germany), and the quantitative analysis was done using the Quantity One software (Bio-Rad). In vitro kinase assay and immunoprecipitation. Preparation of lysates for immunoprecipitation was done as described (42). Immunoprecipitates were washed thrice with lysis buffer and once with kinase reaction buffer [40 mmol/L HEPES (pH 7), 20 mmol/L MgCl2, 10 mmol/L EGTA, 1 mmol/L DTT] and resuspended in 35 AL reaction buffer containing 20 Amol/L ATP,
6531
Cancer Res 2006; 66: (13). July 1, 2006
Cancer Research 5 ACi [g-32P]ATP (NEN, Boston, MA), 2 Ag histone H1 (Calbiochem, La Jolla, CA), or 2 Ag pRb-GST (Santa Cruz Biotechnology). The reactions were incubated for 30 minutes at 30 Cj and terminated by addition of 8 SDS loading buffer and heating to 95jC for 5 minutes. The samples were separated on a 12% SDS-PAGE gel. The gel was transferred on Whatman 3MM paper (Schleicher & Schuell, Dassel, Germany), dried, and phosphorylated proteins were visualized by autoradiography. Immunoprecipitation for the detection of phospho-T187-p27 was done as described above. Lysates were prepared from 5 105 cells on a 10-cm dish either treated with oncostatin M or left untreated. The lysates were incubated with antibodies recognizing p27 and analyzed by Western blot with antibodies specific for p27 phosphorylated on Thr187. For the analysis of p21 and p27 complexes, polyclonal antibodies against p21 and p27 were coupled to Protein A agarose using dimethyl pimelimidate dihydrochloride (DMP) as described (43). In brief, 10 Ag
antibodies were bound to 20 AL of Protein A agarose and coupled to the Protein A with 1 mg of DMP. After blocking and washing of the beads, the lysates were added. Lysates were prepared from untreated cells or cells treated with oncostatin M for 48 hours as described for in vitro kinase assay. Each lysate was split into half, and each half was incubated with either coupled anti-p21 or anti-p27 antibodies. Binding to the beads occurred for 5 hours at 4jC followed by extensive washing with PBS, including 0.1% NP40. Bound proteins were eluted with loading buffer for SDS-PAGE. The precipitated complexes were separated by SDS-PAGE on a 12% gel, transferred to a polyvinylidene difluoride membrane, and analyzed as indicated. RNA interference. RNA oligonucleotides were obtained from Dharmacon Research (Lafayette, CO). Annealing and RNA interference analysis was described by Elbashir et al. RNA oligonucleotides with the sequence p21, 5¶CUUCGACUUUGUCACCGAGdTdT-3¶/5¶-CUCGGUGACAAAGUCGAAGdTdT-3¶; p27-1, 5¶-ACCCCUAGAGGGCAAGUACdTdT-3¶/5¶-GUACUUGCCCUCUAGGGGUdTdT-3; and p27-2, 5¶-CCCGGGACUUGGAGAAGCAdTdT3¶/5¶-UGCUUCUCCAAGUCCCGGGdTdT-3¶ were synthesized. As a control, a scrambled double-stranded RNA (dsRNA), available at Dharmacon Research, with the sequences 5¶-UCGCGCUUUGUAGGAUUCGdTdT-3¶/5¶CGAAUCCUACAAAGCGCGAdTdT-3¶, was used. This sequence lacks complementary sequences in the human genome. dsRNA (1 Ag) was transfected into 86HG39 glioblastoma cells (3 105 per 6-cm dish) with Oligofectamine (Invitrogen, San Diego, CA) according to the manufacturer’s recommendation. Forty-eight hours after transfection, the cells were split on six-well dishes (105 per well) and on glass discs in a 48-well dish (15,000 per well), so that one well could be treated with 25 ng/mL oncostatin M, and the others were left untreated. The cells from the sixwell dish were lysed after 48 hours for Western blot analysis as described above. The cells on the glass discs were further treated for the detection of BrdUrd incorporation by immunhistochemical analysis as described above.
Results Oncostatin M induces a cell cycle arrest in G0-G1 phase. Oncostatin M inhibits proliferation of glioblastoma cells. The cell cycle distribution of the 86HG39 cells was determined in a time course experiment after oncostatin M addition. Glioblastoma cells were treated with oncostatin M as indicated. Forty-five minutes before harvesting, cells were incubated with BrdUrd to label cells in S phase. The cell cycle phase distribution was determined by flow cytometry by determining the DNA contend with propidium iodine and percentage of cells in S phase by a FITC-conjugated anti-BrdUrd antibody. The percentage of cells with a G0-G1 DNA content (2C) increased from 57% in the absence of the cytokine steadily to 84% after 48 hours of oncostatin M treatment. At the
Figure 1. Oncostatin M (OSM ) induces cell cycle arrest in G1 phase. A, 86HG39 cells arrest in G0-G1 phase after oncostatin M treatment. Glioblastoma cells were incubated with oncostatin M (25 ng/mL) for 0, 12, 24, and 48 hours as indicated. Before harvesting, cells were incubated with BrdUrd for 45 minutes. The distribution of cells in the different cell cycle phases was determined by flow cytometry using FITC-labeled anti-BrdUrd antibodies and propidium iodine. Top, two-dimensional FACS profiles. Bottom, percentage of cells in G0-G1, S, and G2-M phases were determined and plotted over time. B, CDK2/cyclin E and CDK6/cyclin D become inactivated in oncostatin M–treated cells. Kinase complexes were precipitated from 86HG39 cell lysates with anti-cyclin E and anti-CDK6 antibodies, and their associated kinase activities towards histone H1 (CDK2) or pRB-GST (CDK6) were determined. Incorporation of [g-32P]ATP into the substrates was visualized by autoradiography after separation of the reaction mixes by SDS-PAGE on 12% acrylamide gels. C, expression of G1-specific CDK and cyclin proteins is not reduced in oncostatin M–treated cells. Glioblastoma cells were treated with oncostatin M as indicated; 25 Ag of total cell lysate were analyzed by Western blot using the respective antibodies as indicated. D, oncostatin M leads to hypophosphorylation of pRB; 25 Ag of cell lysate was separated by SDS-PAGE after treating the cells with oncostatin M for the time indicated. Western blot analysis was done using antibodies recognizing both hyperphosphorylated and hypophosphorylated pRB.
Cancer Res 2006; 66: (13). July 1, 2006
6532
www.aacrjournals.org
Oncostatin M and Glioblastoma Cell Cycle Progression
same time, the percentage of cells in S phase declined from 30% to 3.7% (Fig. 1A). As cells ceased proliferation, we concluded that oncostatin M arrested the glioblastoma cell in G0-G1 phase. Sustained expression of cyclin E1 and cyclin D1 proteins suggests that cells accumulate in G1 phase rather then in G0 (Fig. 1C). Because cell numbers increased only slightly over time, and because no increase in apoptosis was observed (data not shown), we concluded that oncostatin M blocked the G1-S transition in human glioblastoma cells. Oncostatin M induces p21 mRNA expression and enhances p21 and p27 protein level. Cell cycle progression depends on the sequential oscillation of CDK activities (44). To investigate whether the oncostatin M mediated G0-G1 arrest involves CDK inactivation, we compared G1 and G1-S phase–specific kinase activities from oncostatin M–treated and untreated cells. Oncostatin M led to a significant reduction of cyclin E –associated histone H1- and CDK6-associated pRb kinase activity, showing that oncostatin M leads to inactivation of these G1-S CDKs (Fig. 1B). One key substrate of G1 kinases in vivo is the retinoblastoma protein pRB. pRB becomes hyperphosphorylated as CDKs become activated. To confirm CDK inactivation in vivo, we investigated the phosphorylation status of pRB in oncostatin M–treated cells. Consistent with
Figure 2. Oncostatin M (OSM ) induces the CDK inhibitors p21 and p27. A, accumulation of p21 and p27 proteins in oncostatin M–treated glioblastoma cells; 25 Ag of lysates prepared from 86HG39 cells harvested at times indicated were subjected to SDS-PAGE and analyzed by Western-blot using anti-p27, anti-p21, and anti-CDK2 antibodies. Level of a-actin is shown as a loading control. B, oncostatin M induces p21 mRNA level. Northern blot analysis of RNA prepared from oncostatin M–treated glioblastoma cells. Total RNA was extracted from glioblastoma cells after oncostatin M treatment for the indicated time, and 5 Ag of the RNA was analyzed by Northern blotting using cDNA probes hybridizing to p27, p21, and GAPDH mRNA. C, silencing of p21 expression by siRNA fails to abolish p27 induction. Glioblastoma cells were transfected with an siRNA designed to reduce p21 expression followed by oncostatin M treatment for 48 hours. Expression of p21 and p27 was detected by Western blot analysis.
www.aacrjournals.org
the lack of G1 CDK kinase activity, pRb became hypophosphorylated upon oncostatin M treatment (Fig. 1D). Interestingly, protein level of CDK subunits 2, 4, and 6 and levels of cyclins E and D1 protein were not reduced significantly by oncostatin M (Fig. 1C). Loss of CDK activity might, therefore, result from increased CDK inhibitor level. Increased level of p27 has previously been observed in oncostatin M–treated human melanoma cells (45). Similarly, p27Kip1 increased steadily in 86HG39 cells to a plateau 24 hours after oncostatin M stimulation (Fig. 2A). This increase of p27 expression was posttranscriptional because p27 mRNA level remained largely unchanged (Fig. 2B). In addition to p27, p21 was also induced by oncostatin M with similar kinetics (Fig. 2A). In contrast to p27, this increase in p21 protein was paralleled by an increase of p21 mRNA (Fig. 2B). To investigate if the induction of p27 might be a consequence of the precedent p21 induction, we blocked the expression of p21 by small interfering RNA (siRNA) transfection. Knockdown of p21 was successfully achieved as shown in Fig. 2C, whereas a control siRNA oligonucleotide (SC) had no effect. The strong reduction of p21 expression had no significant effect on the induction of p27 by oncostatin M. This observation excludes that the p27 accumulation is a secondary consequence of elevated p21 expression. Oncostatin M stabilizes the p27 protein. To show oncostatin M regulation of p27 stability, we compared degradation of p27 in cycloheximide-treated cells in the presence or absence of oncostatin M. In untreated cells, the amount of p27 protein declined steadily and became undetectable at 8 hours of cycloheximide treatment, whereas p27 level declined much slower in oncostatin M–treated cells (Fig. 3A). This confirmed that p27 accumulates due to significantly increased stability of the protein. p27 accumulation by oncostatin M is independent from the cell cycle position. Stabilization of p27 might be a consequence of the oncostatin M–induced cell cycle arrest because p27 is relatively stable in G1 and becomes increasingly unstable as cells progress into the S phase (11). To determine if p27 accumulation was merely a result of oncostatin M–induced growth arrest or mechanistically involved in mediating the arrest, glioblastoma cells were arrested in S phase using a thymidine block. If p27 accumulation would be a consequence of the cell cycle arrest, no accumulation would be expected in S-phase cells, where p27 is very unstable (Fig. 3B; in the absence of oncostatin M). However, treatment of S phase–arrested cells with oncostatin M led to a keen increase of p27, without interfering with the S-phase arrest (Fig. 3B and C). These data showed that p27 accumulated independently from the cell cycle position and indicated that the increase in p27 protein is a primary consequence of oncostatin M signaling. Interestingly, the expression of p21 was also induced by oncostatin M in S phase–arrested glioblastoma cells (Fig. 7C). Oncostatin M reduces Skp2 mRNA and protein level. Degradation of p27 in S phase requires its preceding phosphorylation on T187. To test whether oncostatin M prohibits T187 phosphorylation, we compared expression level of p27 with the abundance of the T187 phosphorylated isoform of the inhibitor. In the presence of oncostatin M, T187 phosphorylation was strongly increased (Fig. 4A). Lack of T187 phosphorylation is, therefore, not responsible for stabilization of p27. Because T187-phosphorylated p27 is a substrate of the SCF-Skp2 ligase, we next analyzed the expression of subunits of the SCF-Skp2 complex. The Cul1 protein was expressed at increased level in oncostatin M–arrested cells. A mobility shift of a Cul1 isoform that could be detected by SDS-PAGE was not inhibited by oncostatin M treatment (Fig. 4B).
6533
Cancer Res 2006; 66: (13). July 1, 2006
Cancer Research
Ectopic expression of Skp2 fails to destabilize p27 in oncostatin M–arrested glioblastoma cells. Skp2 is rate limiting for p27 degradation and proliferation (46). To investigate whether lack of Skp2 expression accounts for the oncostatin M–induced cell cycle arrest and the stabilization of p27, myc-tagged Skp2 protein was constitutively expressed in 86HG39 cells. Two different stable clones were selected for further analysis: one clone showed ectopic expression comparable with the endogenous Skp2 protein in untreated cells, whereas the other clone more severely overexpressed Skp2 (Fig. 5A). Addition of oncostatin M to both cell lines led to a decline in endogenous Skp2 protein (lower band, Fig. 5A) but did not reduce level of the slightly larger myc-tagged
Figure 3. Oncostatin M (OSM ) increases p27 protein stability even in S phase–arrested cells. A, p27 stability is enhanced by oncostatin M. 86HG39 glioblastoma cells were incubated in absence or presence of 25 ng/mL oncostatin M as indicated. After 24 hours of oncostatin M incubation, 150 Ag/mL cycloheximide was added. Total cell lysates were prepared at the indicated time points and analyzed for the expression of p27 by Western blot. h-Tubulin was detected as a loading control. B and C, oncostatin M stabilizes p27 in S-phase cells. B, glioblastoma cells were incubated with 2 mmol/L thymidine for 24 hours to arrest the cells in S phase of the cell cycle. If indicated, 25 ng/mL oncostatin M was then added, and the cells were further incubated in the presence of both thymidine and oncostatin M. Expression of p27 in cells was compared in the presence or absence of thymidine and oncostatin M. No p27 expression was observed in the absence of oncostatin M, whereas p27 accumulated in the presence of oncostatin M. C, FACS analysis of cells treated with thymidine and oncostatin M confirms that these cells remained arrested in S phase, whereas the untreated cells (control) showed the normal cell cycle phase distribution.
This suggested that oncostatin M does not inhibit expression or modification of Cul1. In contrast, oncostatin M induced a dramatic decline of Skp2 protein (Fig. 4B). Declining Skp2 level corresponded well with the accumulation of p27 (Fig. 4B), suggesting that the decline in Skp2 protein is responsible for the accumulation of p27. To investigate the mechanism of Skp2 repression by oncostatin M, we first determined Skp2 mRNA level. Although well detectable in proliferating cells, Skp2 mRNA expression was lost in cells treated with oncostatin M (Fig. 4D). Already 1 hour after oncostatin M addition, a decrease in Skp2 mRNA was detected, and the level remained low in oncostatin M–treated cells (Fig. 4D). The immediate decline of Skp2 mRNA after oncostatin M addition suggested that this was a primary consequence of the cytokine treatment and not a result from a cell cycle arrest. Reduced Skp2 protein expression might in addition involve regulation of the protein stability. As we observed that level of overexpressed myc-tagged Skp2 protein remained unchanged by oncostatin M addition (see below; Fig. 5A), we concluded that reduction in Skp2 protein is result of impaired mRNA expression. One mechanism of p27 stabilization is, therefore, consequence of lack of Skp2 expression.
Cancer Res 2006; 66: (13). July 1, 2006
Figure 4. Oncostatin M (OSM ) impairs p27 degradation through Skp2 and Cks1. A, p27 phosphorylation on Thr187 is not inhibited by oncostatin M. 86HG39 glioblastoma cells were treated with oncostatin M, and the increase in p27 was followed in a Western blot (p27). The same extracts were analyzed for Thr187-phosphorylated p27 (p-p27 ). B, Skp2 protein expression is decreased by oncostatin M. Protein samples were analyzed in Western blots for Skp2, cyclin D1, p27, and Cul1 expression. p27 protein level increase as Skp2 expression declines. A form of Cul-1 (arrow ) with slightly lower mobility may represent Cul-1 modified by Nedd8. C, down-regulation of Skp2 and Cks1 by oncostatin M. Lysates of oncostatin M–treated 86HG39 cells were analyzed for protein expression of the SCF components Skp2, Skp1, and Cks1by Western blots using specific antibodies. The blot was consecutively treated with the various antibodies. D, Skp2 mRNA levels are decreased by oncostatin M; 10 Ag total RNA (left) or 1 Ag polyadenylated RNA (right ) isolated from oncostatin M–treated 86HG39 cells were analyzed by Northern blotting with a radioactively labeled Skp2 cDNA and a Cks1 cDNA probe. GAPDH was subsequently hybridized as a loading control.
6534
www.aacrjournals.org
Oncostatin M and Glioblastoma Cell Cycle Progression
Figure 5. Skp2 overexpression is not sufficient for p27 destabilization and does not protect from oncostatin M (OSM )–induced cell cycle arrest. A, overexpression of Skp2 cannot destabilize p27 in oncostatin M–treated cells. 86HG39 glioblastoma cell lines were generated that stably overexpress Skp2 from a CMV promoter. Two clones of different Skp2 expression levels were selected. Expression in one clone (Skp2-6) approximately matched the endogenous level in proliferating cells, whereas the second clone (Skp2-10) severely overexpressed Skp2. These clones, parental cells or cells transfected with the empty vector (pcDNA3.1) were analyzed in Western blots for Skp2, p27, p21, and cyclins E and D1 level. Actin and tubulin are shown as loading controls. The larger myc-tagged Skp2 protein (indicated by the arrow ) separates above endogenous Skp2. B, Skp2 overexpression is not sufficient to abrogate the oncostatin M–induced cell cycle arrest. To investigate if Skp2 overexpression can prevent the oncostatin M–induced cell cycle arrest, the cell cycle phase distribution of 86HG39 cells and stable clones cells was determined by two-dimensional FACS analysis. The distribution in the cell cycle phases was determined and is represented for each time point. Black columns, cells in G0-G1 phase; white columns, cells in S phase; gray columns, cells in G2-M phase. C, Skp2 overexpression fails to rescue reduced Cks1 expression. 86HG39 glioblastoma cells (left) or 86HG39 glioblastoma cells overexpressing Skp2 (right ) were incubated with oncostatin M for the indicated times. Protein extracts were isolated, separated by SDS-PAGE on a 16 % acrylamide gel, blotted, and analyzed for expression of Skp2, Cks1, Skp1, and h-tubulin.
exogenous Skp2 (Fig. 5A, arrow). Surprisingly, expression close to the physiologic level and even dramatic overexpression of Skp2 (in clone Skp2-10) failed to destabilize p27 in oncostatin M–arrested cells (Fig. 5A). In addition to its inability to destabilize p27, Skp2 overexpression was also not sufficient to abrogate the antiproliferative effect of oncostatin M. Independent of the Skp2 level, all cell
www.aacrjournals.org
lines arrested readily upon oncostatin M addition (Fig. 5B). Lack of increased proliferation of oncostatin M–treated cells suggested that a second component of the p27 degradation pathway was impaired by oncostatin M signaling. Decreased Cks1 expression in oncostatin M–treated glioblastoma cells and partial attenuation of oncostatin M–induced p27 induction by ectopic expression of Skp2 and Cks1. Expression of basic SCF components like Skp1 or Cul1 was not inhibited by oncostatin M treatment, whereas expression of the Cks1 protein was strongly reduced by oncostatin M (Fig. 4C and Fig. 5C). The Cks1 protein is a cofactor of the SCF-Skp2 ubiquitin ligase complex and is required for degradation of p27. As overexpression of Skp2 did not rescue Cks1 expression (Fig. 5C), we speculated that loss of Cks1 protein was responsible for the accumulation of p27 even in the presence of overexpressed Skp2. As for Skp2, reduction of Cks1 protein corresponded to reduction of Cks1 mRNA expression (Fig. 4D). Reduced Skp2 and Cks1 mRNA level seems to be a characteristic for oncostatin M–responsive cell lines, as the oncostatin M–resistant glioblastoma cell line A172 (47) did not down-regulate either mRNA, whereas oncostatin M–responsive melanoma cells showed reduced Skp2 and Cks1 mRNA level upon oncostatin M treatment (data not shown). These observations suggested that inhibition of Skp2 and Cks1 expression is one redundant mechanism that could be responsible for the oncostatin M–induced cell cycle arrest. We speculated that combined expression of Skp2 and Cks1 might overcome the oncostatin M–induced growth inhibition. To investigate this hypothesis, we generated stable cell lines that simultaneously overexpressed Skp2 (myc tagged) and Cks1 (HA tagged; Fig. 6A). Whereas p27 level increased within 24 hours after oncostatin M addition in wild type and Skp2-overexpressing cells, combined overexpression of Skp2 and Cks1 clearly delayed the accumulation of p27. There was no detectable change in p27 level in 86HG39-SKP2-Cks1 cells 24 hours after oncostatin M addition (Fig. 6A). Still, at 48 hours after oncostatin M treatment, these cells accumulated p27 to level that were only slightly reduced compared with those of wild-type cells or cells expressing exclusively Skp2 and Cks1. To further investigate if the S phase–dependent degradation of p27 was restored by combined expression of Cks1 and Skp2, we examined the oncostatin M–dependent accumulation of p27 in S phase–arrested cells. Oncostatin M treatment over 24 hours was now unable to accumulate p27 in S phase–arrested cells (Fig. 6C). Consistent with the observation that overexpression of Skp2 and Cks1 only delayed the oncostatin M–induced p27 accumulation in asynchronous cells, oncostatin M–treated glioblastoma cells were still able to arrest in G1 phase, although somewhat less efficient than wild-type cells (Fig. 6B). Knockdown of p27 by siRNA could not affect the cell cycle arrest by oncostatin M in Skp2/Cks1–expressing glioblastoma cells. Because the concurrent overexpression of Skp2 and Cks1 could not abrogate the oncostatin M–induced p27 expression, we investigated if a knockdown of p27 by siRNA might abolish the G1 arrest in Skp2/Cks1–overexpressing cells. 86HG39 glioblastoma cells and two cell lines derived from this parental line were transfected with siRNA molecules designed to knock down p27 gene expression. Significant reduction of p27 level was achieved by both different siRNA oligonucleotides in all three cell lines, whereas an unrelated control oligonucleotide (SC) did not affect the p27 expression (Fig. 7A). 86HG39 cells but also cell lines overexpressing Skp2 or a combination of Skp2 and Cks1 still
6535
Cancer Res 2006; 66: (13). July 1, 2006
Cancer Research
at this residue (48), and recent studies proposed an additional noncatalytic function of cyclin A in the ubiquitination of p27 (49). Although cyclin E level remained unchanged after oncostatin M treatment, we found severely reduced expression of cyclin A in oncostatin M–treated glioblastoma cells (Fig. 8A). Similarly to Skp2 and Cks1, the reduction of cyclin A protein was preceded by a decrease in cyclin A mRNA level that clearly preceded the cell cycle arrest (Fig. 8B): already within 3 hours of oncostatin M treatment, a decline of cyclin A mRNA level could be observed (Fig. 8B). Oncostatin M treatment of S phase–arrested cells still resulted in a decline of the S-phase cyclin (Fig. 8C), suggesting that the decline in cyclin A is not a consequence of the G1 phase arrest by oncostatin M but rather cell cycle phase independent. Stable overexpression of cyclin A could not overcome the oncostatin M–mediated accumulation of p21 and p27 (Fig. 8D) nor could it rescue the growth inhibition by oncostatin M (data not shown). This suggests that stabilization of CKIs and inhibited expression of cyclin A act independently on inhibiting cell cycle progression. Reduced cyclin A expression is unlikely limiting p27
Figure 6. Overexpression of Skp2 and Cks1 in 86HG39 cells. A, top, cell clones coexpressing Skp2 and Cks1 were generated and analyzed for expression of Cks1 (HA, indicating HA-tagged Cks1; endo, endogenously expressed Cks1) and Skp2 and p27 and p21. B, FACS analysis of Cks1 and Skp2 overexpressing 86HG39 cells. Black, cells in G1 phase; white, S-phase cells; gray, G2-M indicates the fraction of cells in G2-M phase. C, expression of p27 in Cks1 and Skp2 overexpressing 86HG39 cells. Stable clones of 86HG39 cells overexpressing Skp2 and Cks1 exhibit diminished p27 induction in response to oncostatin M (OSM ). The cells were treated with thymidine for 24 hours. As indicated, oncostatin M was added, and cells were incubated for another 24 hours. Extracts of cells were investigated in Western blots for p27 and h-actin. D, coexpression of Cks1 and Skp2 prevents an accumulation of p27 in oncostatin M–treated S-phase cells. Quantification of p27 was done as described in Materials and Methods by direct measurement of the chemiluminescence of the Western blot.
arrested in the absence of p27 (Fig. 7B). Loss of p27 in Skp2 and Cks1 overexpressing cells is still not sufficient to rescue proliferation in the presence of oncostatin M. Oncostatin M induced down-regulation of cyclin A. Ubiquitination of p27 by the SCF-Skp2 complex requires its preceding phosphorylation on the Thr187. Cyclin A–, cyclin E–, and cyclin B1–associated Cdk complexes can phosphorylate p27
Cancer Res 2006; 66: (13). July 1, 2006
Figure 7. siRNA-medated knockdown of p27 in glioblastoma cells fails to rescue proliferation in the presence of oncostatin M (OSM ), irrespective of Skp2 and Cks1 overexpression. A, p27 knockdown following siRNA transfection. 86HG39 glioblastoma cells and their derivates overexpressing either Skp2 or Skp2 and Cks1 were transfected with siRNA targeting p27. The p27-directed siRNA reduced p27 expression. SC, control siRNA. Level of p27 and p21 were determined by Western blot analysis. B, knockdown of p27 expression does not prevent the cell cycle arrest by oncostatin M. Following siRNA transfection as described in (A ), glioblastoma cells were incubated with oncostatin M for 48 hours. One hour before harvesting, BrdUrd was added to determine the percentage of cells in S phase. Incorporation of BrdUrd was analyzed immunohistochemically using a FITC-labeled anti-BrdUrd antibody. Percentage of BrdUrd-positive cells is displayed. Silencing of p27 did not alter the ability of oncostatin M to prevent cell proliferation, even in the presence of overexpressed Skp2 and Cks1.
6536
www.aacrjournals.org
Oncostatin M and Glioblastoma Cell Cycle Progression
T187 phosphorylation, as we observed strong T187 phosphorylation in the presence of oncostatin M (Fig. 4A). To determine the distribution of p21 and p27 in different Cdk/ cyclin complexes, we immunoprecipitated p21 and p27 complexes and determined level of associated cyclins A, D1, or E by Western blotting. Significantly more cyclin E and cyclin D1 were associated with p21 and p27 after oncostatin M addition. Consistent with the reduced cyclin A protein level, less cyclin A was found in complex with p27 (Fig. 8E), whereas level of cyclin A bound to p21 remained largely unaltered.
SCF-Skp2 ubiquitin ligase complex. This E3 ligase complex includes the cullin Cul1/Cdc53, Skp1, the F-Box protein Skp2, the Ring finger protein Rbx1/Roc1, and the Cks1 protein. Although Cul1, Skp1, and Rbx/Roc1 are mutual components of other SCF complexes, Skp2 and Cks1 act in substrate targeting and interact with T187-phosphorylated p27 (46). p27 stabilization by oncostatin M involves inhibition of Skp2 and Cks1 expression. Downregulation of Skp2 and Cks1 mRNA was observed immediately after oncostatin M addition, strongly suggesting that inhibition of Skp2/Cks1 transcription or decreased Skp2/Cks1 mRNA stability
Discussion The cytokine oncostatin M can inhibit proliferation of a broad range of different human tumor cell lines (1). We aimed to investigate the molecular basis by which this cytokine prevents cell proliferation. Cell cycle arrest by oncostatin M coincides with inactivation of CDK activities. Reduced CDK activity results from reduced expression of cyclin A and increased expression of CDK inhibitors p21 and p27. Oncostatin M or the related cytokine IL-6 have previously been reported to induce transcription of CDK inhibitors; p21 transcription was induced in osteosarcoma and p27 transcription in melanoma cells (45, 50). In contrast, we found here that p27 mRNA level remained unchanged in glioblastoma cells. Instead, degradation of p27 was blocked by oncostatin M. To exclude that stabilization of p27 was a result of the cell cycle arrest in G1 phase where p27 is more stable (11, 12), we determined the stability of p27 in S phase–synchronized cells. Although p27 is usually very unstable in S phase (Fig. 3B; refs. 11, 12), oncostatin M dramatically increased p27 stability in S phase–arrested cells. This indicated that degradation of p27 in S phase was inhibited by oncostatin M. Consistently, we found that oncostatin M prevents expression of two subunits of the E3 ligase that triggers degradation of p27 in S phase. Degradation of p27 follows the ubiquitin-proteasome pathway (26). At least two distinct ubiquitin ligases regulate p27 degradation during the cell cycle. Although the degradation of p27 during G0-G1 progression is independent from T187 phosphorylation (22, 23) and depends on ubiquitination by the KPC complex rather than SCF-Skp2 (51), degradation of the inhibitor in S phase requires T187 phosphorylation of p27 and an active
Figure 8. Down-regulation of cyclin A by oncostatin M (OSM ). A, downregulation of cyclin A protein parallels the decrease of Skp2 after oncostatin M treatment. 86HG39 glioblastoma cells were treated with oncostatin M for the indicated time periods, and lysates were prepared for Western blot analysis. The blot was successively incubated with the antibodies. B, Northern blot analysis of RNA isolated from oncostatin M–treated glioblastoma cells. Total RNA (left ) and polyadenylated RNA (right ) were hybridized with cDNA fragments of cyclin A. As a loading control, the blot was subsequently hybridized with a labelled GAPDH probe (bottom ). C, oncostatin M impairs expression of cyclin A protein in S phase–arrested 86HG39 cells. Glioblastoma cells were either left untreated ( ), treated with oncostatin M for 48 hours (OSM ), arrested in S phase with thymidine (2 mmol/L) for 48 hours (Thy), arrested with thymidine as before followed by oncostatin M treatment for 24 hours (Thy + OSM ). Western blot analysis was done by consecutive use of the indicated antibodies. D, overexpression of cyclin A in glioblastoma cells. 86HG39 glioblastoma cells were transfected with the cyclin A cDNA under control of the CMV promoter. Stable clones were selected for constitutive expression of cyclin A and then treated with oncostatin M for the indicated time points. Lysates were subjected to SDS gel electrophoresis, and blots were incubated with the antibodies for detection of cyclin A, p21, and p27. E, distribution of Cdks and cyclins in p21 and p27 complexes. p21 and p27 were immunoprecipitated from lysates treated with oncostatin M for 24 hours or left untreated. The bound proteins were subjected to SDS gel electrophoresis and analyzed for the presence of the cyclins A, D1, and E.
www.aacrjournals.org
6537
Cancer Res 2006; 66: (13). July 1, 2006
Cancer Research
is a primary consequence of oncostatin M signaling. Interestingly, oncostatin M treatment failed to reduce Skp2 and Cks1 expression in oncostatin M–resistant glioblastoma cell lines (i.e., glioblastoma cell lines that still initiate Jak/STAT signaling upon oncostatin M treatment but fail to cease proliferation; data not shown), suggesting that SCF-Skp2 regulation is a central component in the antiproliferative response to oncostatin M. Stabilization of p27 in the absence of Skp2 was also observed in mice, as homozygous deletion of Skp2 resulted in increased p27 (29). Regulation of Skp2 has been observed after shifting cells from adherent growth into suspension (35), in androgen-treated prostate carcinoma cells (36), after ectopic expression of PTEN in glioblastoma cells or mouse embryonic stem cells, which lack the endogenous PTEN gene (52, 53). Interestingly, in all of these studies, overexpression of Skp2 was sufficient to reverse the p27 accumulation and rescued proliferation. 86HG39 glioblastoma cells investigated in this study express PTEN endogenously (and grow adherent); however, overexpression of Skp2 was not sufficient to rescue p27 degradation after oncostatin M treatment. Ablation of the Cks1 gene in mice leads to a similar phenotype as deletion of the Skp2 gene, including a smaller body size, indicating impaired cell proliferation (14). A prominent role of Cks1 in controlling cell proliferation was suggested by the observation that Cks1 is induced in human tumors (54). Whereas Skp2 is regulated by multiple mechanisms, data on regulation of Cks1 by mitogenic or antimitogenic signals have been scarce. In mink lung epithelial cells, transforming growth factor-h treatment leads to downregulation of Cks1 mRNA and protein followed by up-regulation of p27 protein (55). This suggested that decreasing Cks1 mRNA is sufficient for stabilization of p27 protein and inhibition of cell proliferation. These conclusions are consistent with our observation that Skp2 overexpression was not sufficient to rescue p27 degradation and proliferation of oncostatin M–treated cells. Only combined expression of Skp2 and Cks1 clearly delayed the accumulation of p27 in the presence of oncostatin M (Fig. 6A). However, even the coexpression of Skp2 and Cks1 was not sufficient to stably override the oncostatin M–induced growth
References 1. Grant SL, Begley CG. The oncostatin M signalling pathway: reversing the neoplastic phenotype? Mol Med Today 1999;5:406–12. 2. Halfter H, Lotfi R, Westermann R, et al. Inhibition of growth and induction of differentiation of glioma cell lines by oncostatin M (OSM). Growth Factors 1998;15: 135–47. 3. Friedrich M, Hoss N, Stogbauer F, et al. Complete inhibition of in vivo glioma growth by oncostatin M. J Neurochem 2001;76:1589–92. 4. Sherr CJ. G1 phase progression: cycling on cue. Cell 1994;79:551–5. 5. Morgan DO. Cyclin-dependent kinases: engines, clocks, and microprocessors. Annu Rev Cell Dev Biol 1997;13:261–91. 6. Sherr CJ, Roberts JM. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev 1999;13:1501–12. 7. Ekholm SV, Reed SI. Regulation of G(1) cyclindependent kinases in the mammalian cell cycle. Curr Opin Cell Biol 2000;12:676–84. 8. Hengst L, Reed SI. Inhibitors of the Cip/Kip family. Curr Top Microbiol Immunol 1998;227:25–41. 9. Gartel AL, Tyner AL. Transcriptional regulation of the p21((WAF1/CIP1)) gene. Exp Cell Res 1999;246:280–9.
Cancer Res 2006; 66: (13). July 1, 2006
arrest and to prevent accumulation of p27 in asynchronous cells. It restored, however, p27 degradation in oncostatin M–treated S phase–arrested cells (Fig. 6C). As discussed below, alternative oncostatin M–induced pathways lead to a cell cycle arrest in G1 phase, which then would lead to p27 accumulation, as a result of the cell cycle arrest. This model would be consistent with the delayed accumulation of p27 in cells coexpressing Skp2 and Cks1. This cell cycle arrest involves increased p21 and reduced cyclin A. Increased p21 protein may be sufficient to inactivate Cdks in G1 phase even if p27 is not stabilized. Reduced cyclin A leads to lack of S-phase Cdk activity and a shift of Cdk inhibitors to other Cdk complexes. In addition, reduced cyclin A may also directly contribute to p27 accumulation. Cyclin A has been shown to form a complex with Skp1 and Skp2 (56), and the cyclin has been shown to present p27 to ubiquitination more efficiently than cyclin E (48). Recent reports indicate that cyclin A has an additional noncatalytic function in p27 degradation (49), suggesting that cyclin A mediates p27 degradation independently from its role in p27-T187 phosphorylation. The broad potency of oncostatin M to arrest multiple tumor cells (where different growth inhibitory pathways may be impaired due to oncogenic transformation) may rely in part on the redundancy of pathways that lead to CDK inactivation. If one single pathway is inactivated in a specific tumor cell, the alternative pathways may be sufficient to sustain growth arrest by oncostatin M. It will be interesting to further explore the different signaling pathways triggered by oncostatin M that converge in inhibiting Cdk kinase activity.
Acknowledgments Received 10/18/2004; revised 4/5/2006; accepted 4/20/2006. 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 W. Krek and Christiane Wirbelauer (Friedrich-Miescher-Institute, Basel, Switzerland) for providing us with the pCDNA-Skp2myc vector, the members of our laboratories for discussions, and M. Apel and R. Bru¨ffer for technical assistance.
10. Slingerland J, Pagano M. Regulation of the cdk inhibitor p27 and its deregulation in cancer. J Cell Physiol 2000;183:10–7. 11. Pagano M, Tam SW, Theodoras AM, et al. Role of the ubiquitin-proteasome pathway in regulating abundance of the cyclin-dependent kinase inhibitor p27. Science 1995;269:682. 12. Hengst L, Reed SI. Translational control of p27Kip1 accumulation during the cell cycle. Science 1996;271: 1861–4. 13. Hengst L. A second RING to destroy p27(Kip1). Nat Cell Biol 2004;6:1153–5. 14. Spruck C, Strohmaier H, Watson M, et al. A CDKindependent function of mammalian Cks1: targeting of SCF(Skp2) to the CDK inhibitor p27Kip1. Mol Cell 2001; 7:639–50. 15. Ganoth D, Bornstein G, Ko TK, et al. The cell-cycle regulatory protein Cks1 is required for SCF(Skp2)mediated ubiquitinylation of p27. Nat Cell Biol 2001;3: 321–4. 16. Hao B, Zheng N, Schulman BA, et al. Structural basis of the Cks1-dependent recognition of p27(Kip1) by the SCF(Skp2) ubiquitin ligase. Mol Cell 2005;20:9–19. 17. Podust VN, Brownell JE, Gladysheva TB, et al. A Nedd8 conjugation pathway is essential for proteolytic targeting of p27Kip1 by ubiquitination. Proc Natl Acad Sci U S A 2000;97:4579–84.
6538
18. Liu J, Furukawa M, Matsumoto T, et al. NEDD8 modification of CUL1 dissociates p120(CAND1), an inhibitor of CUL1-1 binding and SCF ligases. Mol Cell 2002;10:1511–8. 19. Lyapina S, Cope G, Shevchenko A, et al. Promotion of NEDD-CUL1 conjugate cleavage by COP9 signalosome. Science 2001;292:1382–5. 20. Yang X, Menon S, Lykke-Andersen K, et al. The COP9 signalosome inhibits p27(kip1) degradation and impedes G1-S phase progression via deneddylation of SCF Cul1. Curr Biol 2002;12:667–72. 21. Tomoda K, Kubota Y, Kato J. Degradation of the cyclin-dependent-kinase inhibitor p27Kip1 is instigated by Jab1. Nature 1999;398:160–5. 22. Malek NP, Sundberg H, McGrew S, et al. A mouse knock-in model exposes sequential proteolytic pathways that regulate p27Kip1 in G1 and S phase. Nature 2001; 413:323–7. 23. Hara T, Kamura T, Nakayama K, et al. Degradation of p27(Kip1) at the G(0)-G(1) transition mediated by a Skp2-independent ubiquitination pathway. J Biol Chem 2001;276:48937–43. 24. Hara T, Kamura T, Kotoshiba S, et al. Role of the UBL-UBA protein KPC2 in degradation of p27 at G1 phase of the cell cycle. Mol Cell Biol 2005;25:9292–303. 25. Kotoshiba S, Kamura T, Hara T, et al. Molecular dissection of the interaction between p27 and Kip1
www.aacrjournals.org
Oncostatin M and Glioblastoma Cell Cycle Progression
ubiquitylation-promoting complex, the ubiquitin ligase that regulates proteolysis of p27 in G1 phase. J Biol Chem 2005;280:17694–700. 26. Bloom J, Pagano M. Deregulated degradation of the cdk inhibitor p27 and malignant transformation. Semin Cancer Biol 2003;13:41–7. 27. Fero ML, Rivkin M, Tasch M, et al. A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27(Kip1)deficient mice. Cell 1996;85:733–44. 28. Nakayama K, Ishida N, Shirane M, et al. Mice lacking p27(Kip1) display increased body size, multiple organ hyperplasia, retinal dysplasia, and pituitary tumors. Cell 1996;85:707–20. 29. Nakayama K, Nagahama H, Minamishima YA, et al. Targeted disruption of Skp2 results in accumulation of cyclin E and p27(Kip1), polyploidy and centrosome overduplication. EMBO J 2000;19:2069–81. 30. Nakayama K, Nagahama H, Minamishima YA, et al. Skp2-mediated degradation of p27 regulates progression into mitosis. Dev Cell 2004;6:661–72. 31. Kossatz U, Dietrich N, Zender L, et al. Skp2dependent degradation of p27kip1 is essential for cell cycle progression. Genes Dev 2004;18:2602–7. 32. Wirbelauer C, Sutterluty H, Blondel M, et al. The F-box protein Skp2 is a ubiquitylation target of a Cul1based core ubiquitin ligase complex: evidence for a role of Cul1 in the suppression of Skp2 expression in quiescent fibroblasts. EMBO J 2000;19:5362–75. 33. Bashir T, Dorrello NV, Amador V, et al. Control of the SCF(Skp2–1) ubiquitin ligase by the APC/C(Cdh1) ubiquitin ligase. Nature 2004;428:190–3. 34. Wei W, Ayad NG, Wan Y, et al. Degradation of the SCF component Skp2 in cell-cycle phase G1 by the anaphase-promoting complex. Nature 2004;428:194–8. 35. Carrano AC, Pagano M. Role of the F-box protein Skp2 in adhesion-dependent cell cycle progression. J Cell Biol 2001;153:1381–90.
www.aacrjournals.org
36. Lu L, Schulz H, Wolf DA. The F-box protein SKP2 mediates androgen control of p27 stability in LNCaP human prostate cancer cells. BMC Cell Biol 2002;3:22. 37. Sheaff RJ, Singer JD, Swanger J, et al. Proteasomal turnover of p21Cip1 does not require p21Cip1 ubiquitination. Mol Cell 2000;5:403–10. 38. Touitou R, Richardson J, Bose S, et al. A degradation signal located in the C-terminus of p21WAF1/CIP1 is a binding site for the C8 alpha-subunit of the 20S proteasome. EMBO J 2001;20:2367–75. 39. Bornstein G, Bloom J, Sitry-Shevah D, et al. Role of SCFSkp2 ubiquitin ligase in the degradation of p21Cip1 during S-phase. J Biol Chem 2003;278:25752–7. 40. Bendjennat M, Boulaire J, Jascur T, et al. UV irradiation triggers ubiquitin-dependent degradation of p21(WAF1) to promote DNA repair. Cell 2003;114: 599–610. 41. Weis J, Schonrock LM, Zuchner SL, et al. CNTF and its receptor subunits in human gliomas. J Neurooncol 1999;44:243–53. 42. Halfter H, Friedrich M, Postert C, et al. Activation of Jak-Stat and MAPK2 pathways by oncostatin M leads to growth inhibition of human glioma cells. Mol Cell Biol Res Commun 1999;1:109–16. 43. Harlow E, Lane D. Antibodies. A laboratory manual. Cold Spring Harbour (NY): Cold Spring Harbour Press; 1988. 44. Sherr CJ. Cancer cell cycles. Science 1996;274:1672–7. 45. Kortylewski M, Heinrich PC, Mackiewicz A, et al. Interleukin-6 and oncostatin M-induced growth inhibition of human A375 melanoma cells is STATdependent and involves upregulation of the cyclindependent kinase inhibitor p27/Kip1. Oncogene 1999; 18:3742–53. 46. Nakayama KI, Nakayama K. Regulation of the cell cycle by SCF-type ubiquitin ligases. Semin Cell Dev Biol 2005;16:323–33.
6539
47. Halfter H, Postert C, Friedrich M, et al. Activation of the jak-stat- and MAPK-pathways by oncostatin M is not sufficient to cause growth inhibition of human glioma cells. Brain Res Mol Brain Res 2000; 80:198–206. 48. Montagnoli A, Fiore F, Eytan E, et al. Ubiquitination of p27 is regulated by Cdk-dependent phosphorylation and trimeric complex formation. Genes Dev 1999;13: 1181–9. 49. Zhu XH, Nguyen H, Halicka HD, et al. Noncatalytic requirement for cyclin A-cdk2 in p27 turnover. Mol Cell Biol 2004;24:6058–66. 50. Bellido T, O’Brien CA, Roberson PK, et al. Transcriptional activation of the p21(WAF1,CIP1,SDI1) gene by interleukin-6 type cytokines: a prerequisite for their pro-differentiating and anti-apoptotic effects on human osteoblastic cells. J Biol Chem 1998;273: 21137–44. 51. Kamura T, Hara T, Matsumoto M, et al. Cytoplasmic ubiquitin ligase KPC regulates proteolysis of p27(Kip1) at G1 phase. Nat Cell Biol 2004;6:1229–35. 52. Gottschalk AR, Basila D, Wong M, et al. p27Kip1 is required for PTEN-induced G1 growth arrest. Cancer Res 2001;61:2105–11. 53. Mamillapalli R, Gavrilova N, Mihaylova VT, et al. PTEN regulates the ubiquitin-dependent degradation of the CDK inhibitor p27(KIP1) through the ubiquitin E3 ligase SCF(SKP2). Curr Biol 2001;11:263–7. 54. Inui N, Kitagawa K, Miwa S, et al. High expression of Cks1 in human non-small cell lung carcinomas. Biochem Biophys Res Commun 2003;303: 978–84. 55. Wang W, Ungermannova D, Jin J, et al. Negative regulation of SCF(Skp2) ubiquitin ligase by TGF-beta signaling. Oncogene 2003;23:1064–75. 56. Zhang H, Kobayashi R, Galaktionov K, et al. p19Skp1 and p45Skp2 are essential elements of the cyclin A-CDK2 S phase kinase. Cell 1995;82:915–25.
Cancer Res 2006; 66: (13). July 1, 2006
