Myostatin inhibits rhabdomyosarcoma cell proliferation ... | Nature

Oncogene (2004) 23, 524–534

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $25.00 www.nature.com/onc

Myostatin inhibits rhabdomyosarcoma cell proliferation through an Rb-independent pathway Brett Langley1,2, Mark Thomas1, Craig McFarlane1, Stewart Gilmour2, Mridula Sharma1 and Ravi Kambadur*,1 1 Animal Genomics, AgResearch, Private Bag 3123, East Street, Hamilton, New Zealand; 2Faculty of Medicine and Health Science, University of Auckland, Private Bag 92019, Auckland, New Zealand

Rhabdomyosarcoma (RMS) tumors are the most common soft-tissue sarcomas in childhood. In this investigation, we show that myostatin, a skeletal muscle-specific inhibitor of growth and differentiation is expressed and translated in the cultured RMS cell line, RD. The addition of exogenous recombinant myostatin inhibits the proliferation of RD cells cultured in growth media, consistent with the role of myostatin in normal myoblast proliferation inhibition. However, unlike normal myoblasts, upregulation of p21 was not observed. Rather, myostatin signalling resulted in the specific downregulation of both Cdk2 and its cognate partner, cyclin-E. The analysis of Rb reveals that there was no change in its phosphorylation status with myostatin treatment, consistent with D-type-cyclin– Cdk4/6 complexes being active in the absence of p21. Moreover, the activity of Rb appeared to be unchanged between treated and nontreated RD cells, as determined by the ability of Rb to bind E2F1. The examination of NPAT, a substrate of cyclin-E–Cdk2 involved in the transcriptional activation of replication-dependent histone gene expression, revealed that it undergoes a loss of phosphorylation with myostatin treatment. Supporting this, a downregulation in H4-histone gene expression was observed. These results suggest that myostatin could potentially be used as an inhibitor of RMS proliferation and define a previously uncharacterized, Rb-independent mechanism for the inhibition of muscle precursor cell proliferation by myostatin. Oncogene (2004) 23, 524–534. doi:10.1038/sj.onc.1207144 Keywords: myostatin; NPAT; retinoblastoma; histone

rhabdomyosarcoma;

Introduction Rhabdomyosarcoma (RMS) is the most common childhood soft-tissue sarcoma and the third most common extracranial childhood solid tumor (Dagher and Helman, 1999; Merlino and Helman, 1999). This tumor consists of several different subtypes, depending primarily on their characteristic histology, with alveolar *Correspondence: R Kambadur; E-mail: [email protected] Received 5 March 2003; revised 17 August 2003; accepted 2 April 2003

(RMS-A) and embryonal (RMS-E) subtypes being the two most common forms. The genesis of RMS-A usually involves a structural rearrangement of chromosome 13 with chromosomes 2 or 1 (Turc-Carel et al., 1986). This is thought to result in the fusion of the fkhr gene (a member of the forkhead family of transcription factors) with pax3 or pax7, respectively (Galili et al., 1993; Shapiro et al., 1993; Davis et al., 1994). It is hypothesized that the product of this fusion inappropriately activates the transcription of genes that contribute to the transformed phenotype (Bennicelli et al., 1995; Scheidler et al., 1996). In contrast to RMSA, no primary genetic rearrangement analogous to fkhrpax has been identified in RMS-E. Instead, a high frequency of loss of heterozygosity at 11p15 has been observed in RMS-E, suggesting an unknown tumor suppressor gene (Scrable et al., 1989). This region of chromosome 11 also contains a number of imprinted genes implicated in oncogenesis, including h19, igf2 and p57 (Feinberg, 1999). In addition, cytogenetic analyses indicate a general pattern of chromosomal gain leading to hyperdiploidy in RMS-E (Pandita et al., 1999), but the genomic regions and the effects from these gains have yet to be determined. RMSs are malignant tumors of mesenchymal origin thought to arise from cells that are committed to a skeletal muscle lineage, but fail to complete the differentiation program (Dagher and Helman, 1999; Merlino and Helman, 1999). Controlling the process of myogenic specification and differentiation are the family of basic helix–loop–helix factors, MRFs, which include MyoD, Myf-5, myogenin and MRF4 (Tapscott and Weintraub, 1991; Olson and Klein, 1994; Ludolph and Konieczny, 1995; Megeney and Rudnicki, 1995; Perry and Rudnick, 2000). Indeed, all of the MRFs have the ability to convert nonmyogenic cells to a muscle phenotype (Braun et al., 1989; Edmondson and Olson, 1989; Weintraub et al., 1989; Choi et al., 1990; Miner and Wold, 1990). In RMS, even though MyoD and myogenin are expressed and are able to bind DNA, rhabdomyoblasts fail to complete the differentiation program, suggesting a deficiency of a factor required for MyoD or myogenin activation (Bouche et al., 1993; Tapscott et al., 1993). Consistent with this, Puri et al. (2000) showed that p38 MAP kinase (MAPK), which is essential for skeletal muscle differentiation, is deficient in RMS cells, and moreover, the forced expression of

Myostatin inhibits rhabdomyosarcoma cell proliferation B Langley et al

525

p38 MAPK restored MyoD function, leading to cell arrest and differentiation. Transforming growth factor-beta (TGF-b) has also been implicated in the proliferation or myogenic differentiation of RMS. Using an established cell line, RD which was derived from an RMS-E tumor (Houghton et al., 1981); the complete removal of TGF-bsignalling by the expression of a dominant-negative type II TGF-b receptor, induces growth arrest but does not trigger differentiation. In contrast, a reduction of TGFb to the range of 0.14–0.20 102 ng/ml was shown to induce myogenic differentiation (Bouche et al., 2000). This finding supports the cell cycle arrest and differentiation of RD cells by 12-O-tetradecanoylphorbol-13acetate (TPA), which is suggested to occur through the TPA-induced loss of urokinase plasminogen activator (uPA) activity resulting in a reduction in the conversion of latent TGF-b to active TGF-b (Aguanno et al., 1990; Bouche et al., 1993, 2000). Myostatin, a growth and differentiation factor (GDF8) belonging to the TGF-b superfamily, acts as a negative regulator of skeletal muscle mass. The deletion of myostatin in mice, mutations in the myostatin gene in Belgian Blue and Piedmontese cattle or the perturbation of myostatin function all result in increased skeletal muscle mass (Kambadur et al., 1997; McPherron et al., 1997; McPherron and Lee, 1997; Grobet et al., 1998; Szabo et al., 1998; Zhu et al., 2000; Yang et al., 2001). Like other TGF-b family members, myostatin appears to be synthesized as a precursor molecule, which is subsequently cleaved to generate an N-terminal propeptide or latency-associated peptide (LAP) and a Cterminal active peptide or mature myostatin. Mature myostatin forms a homodimer and has been shown to bind Activin type II B, and with less efficiency Activin type II A receptors, wherein it presumably elicits its biological function (Lee and McPherron, 2001). The Nterminal LAP functions bind to the mature myostatin preventing this receptor binding, thus regulating myostatin activity. Consistent with its role as a negative regulator of skeletal muscle growth, the mature signalling portion of myostatin has been shown to be biologically active in repressing the proliferation and differentiation of cultured myoblasts (Thomas et al., 2000; Langley et al., 2002). Fundamental to the proliferation of myoblasts is the cell cycle. Progression through the cell cycle and cell cycle arrest is largely controlled by cyclin-dependent kinases (Cdks), cyclins and Cdk-inhibitors (CKIs). The principle Cdks responsible for G1-phase progression and entry into the S phase are Cdk4/6 and Cdk2, which bind and become activated by their regulatory subunits, the D-type cyclins (cyclin-D1, -D2 and -D3) and cyclin-E, respectively. As active complexes cyclin–Cdks phosphorylate retinoblastoma (Rb), inactivating it, releasing transcription factors required for the G1–S-phase progression from their negative restraint. Regulating the G1-cyclin–Cdk complexes are two classes of CKIs, the Cip/Kip (or p21) family and the INK4 (or p16) family. Myostatin has been shown to inhibit myoblast proliferation by specifically upregulating p21 and down-

regulating Cdk2, resulting in a loss of cyclin-E–Cdk2 activity and cell cycle arrest (Thomas et al., 2000). In this study, we have investigated whether myostatin can inhibit RMS cell proliferation. We demonstrate that myostatin inhibits the proliferation of cultured RD cells by the downregulation of both cyclin-E and Cdk2, a kinase complex important for the activation factors promoting G1- to S-phase transition and S-phase progression. Unlike that reported for normal myoblasts, the CKI, p21, was not upregulated in RD cells treated with myostatin. Furthermore, no change in Rb phosphorylation or activity was observed with the treatment of myostatin. Supporting RD cell proliferation inhibition by the downregulation of cyclin-E–Cdk2, a pathway involving NPAT is disrupted with myostatin treatment. NPAT has been shown to be a substrate of cyclin-E–Cdk2 that is crucial for histone gene expression and cell cycle progression at the G1/S boundary. These results suggest that myostatin could potentially be used as an inhibitor of RMS proliferation and define a previously uncharacterized, Rb-independent mechanism for the inhibition of muscle precursor cell proliferation by myostatin.

Results Myostatin is expressed and translated in RMS cells To examine if RMS cells express myostatin, reverse transcriptase–PCR (RT–PCR) using myostatin-specific primers was performed on total RNA extracted from cultured RD cells. As can be seen in Figure 1a, the RT–PCR amplification revealed that myostatin is expressed in RD cells. Furthermore, sequence analysis of the RT–PCR products revealed that no mutations that would result in an inactive myostatin protein are present. In addition to myostatin being expressed in RD cells, Western blot analysis and immunocytochemistry (ICC) of cultured RD cells using antimyostatin antibodies revealed that myostatin protein was also synthesized and processed (Figure 1b and c). These results demonstrate that RMS tumors are unlikely to arise from a loss of myostatin, consistent with the fact that myostatin-null animals such as Belgian blue cattle do not display any increased incidence of tumorigenesis (Arthur, 1995). RMS cell proliferation decreases with increasing concentrations of myostatin and the inhibitory effect is reversible Myostatin is characterized as a specific negative regulator of skeletal muscle growth. Consistent with this, recombinant myostatin has been shown to inhibit primary bovine and cultured mouse C2C12 myoblasts in a concentration-dependent manner (Thomas et al., 2000). Since rhabdomyoblasts are cells committed to a myogenic lineage, by analogy with normal myoblasts, it was asked whether RD cell proliferation is inhibited by exogenously added myostatin. We used RD (embryonic) Oncogene


526

Figure 1 Myostatin is expressed in RMS cells. (a) Ethidium bromide-stained agarose gel photograph showing the 1128 bp myostatin cDNA, RT–PCR amplified from RD cells. The control reaction contained no template. (b) Western blot showing the levels of precursor, LAP and mature forms of myostatin from RD cells and C2C12 myoblasts. Myostatin immunoreactive protein was detected using antimyostatin antibodies. (c) Micrographs showing cultured RD cells immunostained using antimyostatin antibody (a) or anti-rabbit IgG negative-control (b) and counterstained with Gill’s hematoxylin. Bar equals 50 mM

and RMS130 (alveolar) RMS cells to assess the effect of myostatin. Cells were cultured in growth media (containing 10% FCS) with varying concentrations of exogenous myostatin and analysed for their proliferation potential by methylene blue assay. Consistent with the growthinhibitory effect of myostatin on C2C12 and primary myoblasts, myostatin inhibited the growth of both lines of RMS in a dose-dependent manner (Figure 2a and b). The myostatin treatment did not result in myogenic differentiation and myotube formation (Figure 2c). To address the question of whether the myostatin growth inhibition of RMS cells is reversible, RMS cells were incubated with a growth-inhibitory dose of myostatin (8 mg/ml) for 48 h, after which myostatin was removed and cells were incubated in growth media. Proliferation of the myoblasts was assessed at 48-h intervals by the methylene blue assay. Control RMS cells, incubated in growth media without myostatin, showed a steady increase in cell number during the course of the experiment (Figure 3). When myostatin was added, the total cell number remained constant for the entire duration of the experiment. Conversely, when myostatin was subsequently removed from the cultures at 48 h, myoblasts resumed growth and the total cell number increased (Figure 3). Oncogene

Figure 2 Myostatin inhibits RMS cell proliferation. Graph showing the proliferation of RD (a) and RMS13 (b) RMS cells cultured in growth media with 0–10 mg/ml myostatin for 72 h. Myoblast proliferation was measured by methylene blue assay. The optical density at 655 nm is directly proportional to myoblast number. Bars represent the average of eight replicates from two independent experiments, 7standard error. All bars are significantly different (Po0.01). (c) Micrographs showing RD cells cultured in growth media without (a) or with (b) 8 mg/ml myostatin for 72 h. Cells were stained with Gill’s hematoxylin and eosin. Bar equals 50 mM. (d) Western Blot shows the overexpression of cyclinE and Cdk2 in RD cells. Ponceau S staining indicates equal loading of the proteins. (e) Graph showing the partial rescue of RD cell growth inhibition by myostatin. Control or cyclin-E and/or Cdk2 overexpressing RD cells were incubated with 2 mg/ml myostatin for 72 h and RD cell proliferation was measured by methylene blue assay. The inhibition of control RD cells by myostatin is considered as 100%. The percentage of growth inhibition of cyclin-E and/or Cdk2 overexpressing cells was calculated by comparing with the inhibition of controls cells by myostatin

Taken together, these results suggest that myostatin can inhibit the proliferation of RMS cells in a dosedependent and reversible manner such as was described for C2C12 and primary bovine myoblasts (Thomas et al., 2000). Myostatin treatment causes cell cycle arrest in RMS cells Flow cytometric analysis was used to determine if RMS cells were accumulating in the gap (G1 and G2) phases of


527

Figure 3 RMS cell growth inhibition by myostatin is reversible. Graph showing the proliferation of RD cells cultured for 192 h with myostatin in the media for 48 h and then removed (——) or for the duration of the experiment (—J—). Growth of control RD cells in the absence of myostatin throughout the experiment is also shown (—.—). Myoblast proliferation was measured by methylene blue assay. The optical density at 655 nm is directly proportional to myoblast number

the cell cycle as previously shown for C2C12 myoblasts (Thomas et al., 2000). Unfortunately, since RMS cells are aneuploidy and contain several chromosomal translocations (Shapiro et al., 1991; Pappo et al., 1993; Niggli et al., 1994), no meaningful data could be attained with respect to cell cycle distribution. To overcome this problem, BrdU staining was used to examine if any DNA synthesis was occurring in RMS cells treated with myostatin. To this end, RD cells were methionine-starved to synchronize them in the G0/G1 phase of the cell cycle. Following the release from this methionine-minus block, the RD cells were incubated in growth media with or without myostatin in the presence of BrdU. ICC analysis shows that nearly all (95.0070.47%) of the untreated myoblasts progressed through the S phase of the cell cycle, as determined by BrdU incorporation during DNA synthesis (Figure 4a and b). In contrast to the control cells, very few (6.8570.77%) of the myostatin-treated cells progressed through the S phase (Figure 4a and b). These values are significantly different (Po0.001). The observation that myostatintreated RD cells fail to incorporate BrdU suggests that, like C2C12 and bovine primary myoblasts, myostatin arrests RD cells in the G1 phase of the cell cycle. Unlike normal myoblasts, myostatin treatment does not upregulate p21, or other CKIS in RMS cells The treatment of RMS cells with myostatin results in cell proliferation arrest and a concomitant loss of DNA synthesis, similar to that previously described for C2C12 myoblasts. To examine if the CKI, p21, or other CKIs, p15, p16 and p27 were contributing to cell cycle arrest

Figure 4 Myostatin arrests RMS cells in the G1 phase of the cell cycle. (a) Micrographs showing BrdU incorporation in RD cells cultured in growth media without (a) or with (b) 10 mg/ml myostatin after synchronization in G1. RD cells were immunostained using anti-BrdU antibody and counterstained with Gill’s hematoxylin. Bar equals 50 mM. (b) Graph showing the percentage of BrdU positive RD cells from (a, a) and (a, b). Bars represent the average BrdU positive cell number of five random fields of view for three independent experiments, 7standard error. Values are significantly different (Po0.01)

and proliferation inhibition, their expression was examined after myostatin treatment. Actively growing RMS cells were cultured with or without myostatin in the media for 6, 12, 18 and 24 h before harvesting for total protein. Western blot analyses using anti-p21 antibodies showed that p21 protein was not upregulated at any of the time points after treatment with myostatin (Figure 5). To examine if myostatin upregulated any other CKI belonging to either the p21- or p16-families, further Western blot analyses were performed on the myostatin-treated or nontreated RMS protein extracts. Figure 5 shows that neither p15 nor p16 (of the p16family) is upregulated in response to myostatin treatment. Similarly, p27 and p57 (of p21-family) levels were unchanged between the myostatin-treated and nontreated RMS cells (Figure 5). These results suggest that although RMS cell growth is inhibited by myostatin, the exact molecular mechanism for this inhibition is different from that of C2C12 myoblast inhibition. Myostatin treatment significantly downregulates Cdk2 and its cognate partner cyclin-E, but not Cdk4 and cyclin-D1 Since the levels of the G1-cyclin–Cdks can regulate the cell cycle, it is possible that the observed inhibition of Oncogene


528

Figure 5 Myostatin does not alter the levels of the CKIs p21, p15, p16 or p27 in RMS cells. Western blots showing the levels of p21, p15, p16 and p27 protein in RD cells cultured with ( þ ) or without () myostatin for 24 and 48 h. p21, p15, p16 and p27 immunoreactive proteins were detected using their respective antibodies. Tubulin protein levels, detected by antitubulin antibodies, are included to show even loadings

RMS cell proliferation by myostatin could be due to downregulated Cdks and/or cyclins. To examine if myostatin downregulates cyclin–Cdk levels, Western blot analyses were performed on total protein extracts from RMS cells cultured with or without myostatin in the media for 24 and 48 h. Figure 6 shows that the level of Cdk2 was indeed downregulated at the 24- and 48-h time points in response to myostatin treatment. As seen with C2C12 myoblasts, the levels of Cdk4 were not appreciably different between myostatin-treated and -untreated RMS cells (Figure 6). Western analyses examining the levels of cyclin-E and cyclin-D1 showed that cyclin-E expression is dramatically reduced in RMS cells treated with myostatin (Figure 6). The levels of cyclin-D1 did not change appreciably (Figure 6). Taken together, these results suggest that myostatin may inhibit RMS cell proliferation by the downregulation of the cell cycle regulatory complex, cyclin-E–Cdk2. To further prove this fact, we tried to rescue RMS cells from myostatin inhibition by stably overexpressing either cyclin-E alone or Cdk2 and cyclin-E together in the RMS cells. As shown in Figure 2d, stable transfection resulted in the ectopic expression of cyclin-E and Cdk2 in the RMS cells. When these cells were treated with myostatin, RD cells expressing cyclin-E alone are more resistant to myostatin-mediated repression and thus have increased proliferation as compared to control RD cells treated with myostatin. Similarly, cells ectopically expressing cyclin-E and Cdk2 also showed an increasing rate of proliferation as compared to Oncogene

Figure 6 Myostatin downregulates cyclin-E and Cdk2, but not cyclin-D1 and Cdk4 in RMS cells. Western blots showing the levels of cyclin-E, cyclin-D1, Cdk2 and Cdk4 protein in RD cells cultured with ( þ ) or without () myostatin for 24 and 48 h. Cyclin-E, cyclin-D1, Cdk2 and Cdk4 immunoreactive proteins were detected using their respective antibodies. Tubulin protein levels, detected by antitubulin antibodies, are included to show even loadings

control RD cells when treated with myostatin (Figure 2e). These results confirm that myostatin inhibits RMS cells’ proliferation by downregulating the expression of cyclin-E–Cdk2. Myostatin treatment does not alter Rb phosphorylation status in RMS cells Rb protein, the major substrate of G1-Cdks, acts by binding to and repressing the activity of certain transcription factors, such as the heteromeric E2F/ DP1 complex. Recent studies suggest that Cdk4 and Cdk6 in conjunction with D-type cyclins, which are expressed early in G1, partially phosphorylate Rb, allowing for the transcriptional activation of Cdk2 and cyclin-E. The subsequent phosphorylation of Rb by cyclin-E–Cdk2 fully inactivates Rb allowing the transcription of S-phase-specific genes and G1- to S-phase cell cycle progression. The observed lack of p21 upregulation and normal expression of cyclin-D1– Cdk4 in RMS cells treated with myostatin suggested that the Rb protein would not accumulate in a hypophosphorylated form. To examine the phosphorylation status of Rb in myostatin-treated RMS cells, Western blot analysis was performed, using an antibody that detects both the hypo- and hyperphosphorylated forms of Rb, on total proteins extracted from RMS cells or control rat L6 myoblasts incubated with or without myostatin protein for 12 or 24 h. As seen in Figure 7a, the treatment of control rat L6 myoblasts resulted in Rb accumulating in


529

Figure 7 Myostatin signalling does not result in active hypophosphorylated Rb protein in RMS cells. (a) Western blot showing the phosphorylation status of Rb protein in rat L6 myoblasts and RD cells cultured with ( þ ) or without () myostatin for 48 h. A monoclonal anti-Rb antibody that recognizes the hypophosphorylated (pRb) and hyperphosphorylated (pRbPP) forms of Rb protein was used. (b) Western blot showing the levels of E2F1 protein coimmunoprecipitated with Rb protein from RD cells cultured with ( þ ) or without () myostatin for 48 h. Coimmunoprecipitations were performed using anti-Rb antibodies and E2F1 immunoreactive protein was detected using anti-E2F1 antibodies. Control represents no anti-Rb antibody during immunoprecipitation. Equal loading of the protein is denoted by Ponceau S staining

through, the S phase. Furthermore, cyclin-E and Cdk2 have been shown to control cell cycle independent of Rb. Since both cyclin-E and Cdk2 are downregulated by myostatin treatment, the inhibition of RMS cell proliferation may, at least in part, be due to the loss of cyclin-E–Cdk2 substrate-specific phosphorylation other than Rb. To ascertain whether the downregulation of cyclinE–Cdk2 results in the loss of substrate-specific phosphorylation, the phosphorylation and activity of NPAT was examined in RD cells cultured in growth media with or without myostatin. Since the specific phosphorylation of NPAT by cyclin-E–Cdk2 has previously been reported to result in a mobility shift in SDS–polyacrylamide gel electrophoresis (Ma et al., 2000), Western blot analysis was used to determine the phosphorylation of NPAT. Two NPAT immunoreactive bands that migrated at different rates, consistent with nonphosphorylated NPAT and phosphorylated NPAT, were observed in protein extracts from RD cells cultured in growth media without myostatin (Figure 8a). In contrast to this, only a single NPAT-immunoreactive band, consistent with nonphosphorylated NPAT, was observed in protein extracts from RD cells cultured in growth media with myostatin (Figure 8a). These results suggest that the loss of NPAT phosphorylation does indeed occur with myostatin treatment and the downregulation of cyclin-E–Cdk2. The role of NPAT, after activation by cyclin-E–Cdk2 phosphorylation, is to promote replication-dependent histone gene transcription (Ma et al., 2000; Zhao et al.,

a hypophosphorylated state, consistent with the published results (Thomas et al., 2000). In contrast to this, there was no apparent difference in the hyper- and hypophosphorylated forms of Rb between the treated and nontreated RMS cells. Rb activity is not altered by myostatin treatment The examination of Rb phosphorylation suggested no difference in the phosphorylation status in myostatintreated and nontreated RD cells. To examine if myostatin treatment altered Rb activity, the ability of Rb to bind E2F1 was examined in myostatin-treated and nontreated RD cells. The immunoprecipitation of Rb using anti-Rb antibodies and subsequent Western blot analysis for E2F1, using anti-E2F1 antibodies, revealed that there was no difference in the levels of E2F1 coimmunoprecipitated with Rb between the myostatin-treated and nontreated RD cells (Figure 7b). These results suggest that in RD cells, Rb phosphorylation and activity are unchanged between myostatin treatment and nontreatment, despite the loss of cyclin-E and downregulation of Cdk2. Furthermore, cell cycle arrest appears to be independent of Rb. Downregulation of Cdk2–cyclin-E results in the loss of NPAT phosphorylation and histone H4 gene activation Cyclin-E–Cdk2 is expressed in late G1 and has many substrates fundamental for entry into, and progression

Figure 8 Myostatin signalling results in the loss of NPAT phosphorylation and activity. (a) Western blot showing the phosphorylation of NPAT protein in RD cells cultured with ( þ ) or without () myostatin for 24 h. A monoclonal anti-NPAT antibody that recognizes the phosphorylated (NPATPP) and nonphosphorylated (NPAT) forms of NPAT protein was used. (b) Northern blot showing the expression of histone H4 in RD cells cultured with ( þ ) or without () myostatin for 24 or 48 h. The Northern blot was probed with a 339 bp histone H4 probe. The histone H4 mRNA transcript is shown. Ethidium bromide-stained formaldehyde/agarose gel showing 28S and 18S rRNA is also included Oncogene


530

Figure 9 A model for the Rb-independent role of myostatin in muscle growth. During embryonic myogenesis, Myf-5 and MyoD specify cells to adopt the myoblast fate. Myoblasts then migrate and proliferate. In response to myostatin signalling, cyclin-E–Cdk2 levels and activity are reduced, causing the loss of cyclin-E–Cdk2 substrate-specific phosphorylation (for example, NPAT) required for G1 to S-phase progression. Thus, myoblasts arrest in the G1 phase of the cell cycle, and cell number is regulated

2000). Thus, to examine the activity of NPAT, Northern blot analysis was used to determine the expression of histone H4 in RD cells cultured in growth media with or without myostatin. As can be seen in Figure 8b, histone H4 mRNA transcripts of the expected molecular size were detected at both 24 and 48 h time points in RD cells cultured in growth media without myostatin. By contrast, the histone H4 mRNA transcripts detected in RD cells cultured in growth media with myostatin for 24 and 48 h were lower, consistent with a loss of NPAT activity (Figure 8b).

Discussion Myostatin, a member of the TGF-b superfamily, is a key regulator of skeletal muscle growth and development. In this investigation, we show that myostatin lacks any inactivating sequence mutations and is expressed, translated and processed in the RMS cell line, RD. We also show that like normal myoblasts, RD cell proliferation is inhibited by the addition of exogenous recombinant myostatin. The molecular mechanism for this inhibition, however, differs from that described for the proliferation inhibition of normal myoblasts. Specifically, no increase in the CKI, p21, is observed in RD cells treated with myostatin; cyclin-E and Cdk2 are downregulated and proliferation inhibition appears to be Rb independent (Figure 9). Myostatin is expressed in RD cells In mammals, myostatin is expressed in myogenic precursor cells and skeletal muscle, although low levels of expression have been detected in other tissues such as adipose (McPherron et al., 1997), mammary (Ji et al., 1998) and cardiac tissue (Sharma et al., 1999). Since RMS tumors are thought to arise from mesenchymal cells committed to a skeletal muscle lineage (Dagher and Helman, 1999; Merlino and Helman, 1999), it was examined whether the loss or perturbation of myostatin synthesis contributes to RMS tumorigenesis. The results presented here show that myostatin is indeed expressed Oncogene

(Figure 1a), translated (Figure 1b and c) and processed (Figure 1b) in RD cells. Furthermore, sequence analysis of the expressed myostatin cDNA revealed that no mutations were detected that would inactivate myostatin function or signalling. RD cell growth is inhibited by myostatin The role of myostatin as a negative regulator of skeletal muscle growth is now very well established. Inactivating mutations in cattle and mice or the deletion of myostatin in mice all result in increased skeletal musculature due to hypertrophy and/or hyperplasia. Furthermore, myostatin has been shown in cell culture systems to inhibit the proliferation of both primary bovine and mouse C2C12 myoblasts (Thomas et al., 2000). This inhibition was dose dependent and did not cause increased apoptosis or myogenic differentiation. The examination of RD cell proliferation in response to myostatin revealed that like normal myoblasts, RD cell proliferation is inhibited by myostatin treatment in a dose-dependent manner (Figure 2a and b). The inhibition to proliferation is reversible upon removal of myostatin from the culture media (Figure 3). Flowcytometric analysis previously performed on normal C2C12 myoblasts treated with myostatin showed a significant reduction of cells in the S phase and an accumulation of cells in the G1- and G1/M-phases of the cell cycle (Thomas et al., 2000). Unfortunately, with RD cells, no meaningful data could be attained by flowcytometric analysis with respect to cell cycle distribution. However, examination of BrdU incorporation in myostatin-treated RD cells determined that, like normal C2C12 myoblasts, there was a significant reduction in the proportion of cells in the S phase of the cell cycle (Figure 4). Myostatin has been shown to inhibit the proliferation of normal myoblasts by altering the levels and activity of members of the cell cycle machinery (Thomas et al., 2000). Specifically, in response to myostatin treatment, the cyclin-dependent kinase inhibitor p21 is upregulated and the level of Cdk2 is slightly reduced resulting in a loss of cyclin-E–Cdk2 activity. This loss of activity results in the loss of Rb phosphorylation causing cell


531

cycle arrest. In contrast to C2C12 myoblasts, when RD cells were treated with myostatin, no upregulation of p21 was observed (Figure 5). Rather, myostatin signalling resulted in the specific downregulation of both Cdk2 and its cognate partner, cyclin-E (Figure 6). The analysis of Rb reveals that there was no change in its phosphorylation status with myostatin treatment (Figure 7a), consistent with D-type-cyclin–Cdk4/6 complexes being active in the absence of p21. Moreover, the activity of Rb appeared to be unchanged between treated and nontreated RD cells, as determined by the ability of Rb to bind E2F1 (Figure 7b). Despite the phosphorylation and activity of Rb, RD cells are arrested by the treatment of myostatin. The downregulation of both Cdk2 and cyclin-E appears to be central to this proliferation inhibition since the ectopic expression of cyclin-E and Cdk2 results in partial rescue of myostatin-mediated inhibition of RMS cells. Cyclin-E–Cdk2 is expressed in late G1 and has many substrates including Rb, E2Fs, SWI/SNF (chromatin remodelling factors), splicesome-associated proteins, NPAT (histone gene transcriptional activator), p27 and CDC25A (activators of cyclin-E–Cdk2 itself), as well as unknown substrates involved in chromosome duplication and DNA replication initiation (Ewen, 2000). The majority of these are fundamental for entry into, and progression through, the S phase. Furthermore, cyclin-E and Cdk2 have been shown to control cell cycle independent of Rb. The inactivation of cyclinE has been shown to result in a G1 arrest in cells in which Rb function has been inactivated by the expression of SV40 large T antigen (Ohtsubo et al., 1995). Similarly, the inhibition of Cdk2 activity by the ectopic expression of Cdk inhibitors, such as p21 and p27, or a dominant-negative Cdk2 mutant can cause cell cycle arrest without the functions of Rb (van den Heuvel and Harlow, 1993; Sherr and Roberts, 1995; Hofmann and Livingston, 1996; Ewen, 2000). To investigate if the downregulation of cyclin-E– Cdk2 was functionally significant in the arrest of RD cell proliferation by myostatin, the activation and activity of NPAT were examined. NPAT, a Cdk2 substrate, has been identified as a transcription factor involved in the activation of replication-dependent histone gene expression. Multiple copies for each histone subtype (H1, H2A, H2B, H3 and H4) exist to accommodate the large demand for histone synthesis during DNA replication (Heintz, 1991; Osley, 1991; Stein et al., 1992). While replication-independent histone genes are constitutively expressed at low levels, the majority of vertebrate histone genes are replicationdependent and expressed during the S phase (Albig and Doenecke, 1997). Thus, the phosphorylation of NPAT by cyclin-E–Cdk2 coordinates DNA replication with histone gene synthesis (Ma et al., 2000; Zhao et al., 2000). The analysis of NPAT phosphorylation in RD cells treated with myostatin revealed that it undergoes a loss of phosphorylation, consistent with the downregulation of cyclin-E–Cdk2 (Figure 8a). Supporting this loss of phosphorylation, the activity of NPAT also appeared to be reduced, as a downregulation in h4-

histone gene expression was observed in RD cells treated with myostatin (Figure 8b). In agreement with the results described here, whereby RD cell proliferation is inhibited by myostatin, is a report showing that TGF-b can inhibit the proliferation of RD cells in a dose-dependent manner (Bouche et al., 2000). While the exact molecular mechanism for this inhibition in RD cells is not reported, other lines of evidence showing that TGF-b1 downregulates cyclin-E and cyclin-A mRNA and protein levels in epithelial cells without altering cyclin-D1 levels (Geng and Weinberg, 1993; Hu and Zuckerman, 2001) also support the findings described here. In addition, the treatment of epithelial cells by TGF-b1 has been shown to inhibit cdk2 and cdk4 mRNA induction (Geng and Weinberg, 1993; Hu and Zuckerman, 2001). Unlike this observation, however, no downregulation of Cdk4 was observed in RD cells with myostatin treatment. It is interesting that p21 is not upregulated in RD cells in response to myostatin, as seen in C2C12 cells. It has been well documented that the transcriptional activation of p21 during differentiation is lost in RD cells (Otten et al., 1997); however, this appears to be a result of the inability of MyoD to activate transcription rather than the inherent dysfunction of p21 itself. Certainly, the restoration of MyoD function by enforced induction of p38 MAPK results in p21 upregulation and terminal myogenic differentiation (Puri et al., 2000). The loss of p21 upregulation may also partially contribute to the loss of Rb hypophosphorylation, as p21 is responsible for inhibiting cyclin–Cdk phosphorylation of Rb. Cell cycle regulation in the G1 phase has attracted a great deal of attention as a promising target for the research and treatment of cancer. Many of the important genes associated with G1-regulation have been shown to play a key role in proliferation, differentiation and oncogenic transformation and programmed cell death (apoptosis). Currently, a variety of ‘cytostatic’ agents that affect G1 progression and/or G1/ S transition are being evaluated in clinical trials. Indeed, the antitumor sulfonamide E7070, causing a cellular accumulation in the G1 phase, has been shown to suppress the activation of Cdk2 and cyclin-E expression in HCT116 colorectal cancer cell line highly sensitive to the drug (Owa et al., 2001). Based on these data and the results described here, where the treatment of RD cells with myostatin results in the downregulation of cyclinE–Cdk2 and proliferation inhibition, myostatin could potentially be useful as a ‘cytostatic’ agent for the treatment of RMS. Another feature making myostatin potentially useful is its specificity. Like other TGF-b family members, the mature C-terminal protein binds to the N-terminal remnant of the original precursor molecule (LAP) resulting in a latent complex (Lee and McPherron, 2001; Thies et al., 2001; Yang et al., 2001). Myostatin then has to be released from this latent complex to be active. Indeed, the specificity of myostatin is evident in myostatin-null animals, where the altered phenotype is restricted exclusively to skeletal muscle (Kambadur et al., 1997; McPherron et al., 1997; Oncogene


532

McPherron and Lee, 1997; Grobet et al., 1998; Zhu et al., 2000).

Materials and methods Reagents Myostatin cDNA generation, subcloning into the pET protein expression system, and expression and purification of myostatin protein in Escherichia coli have been described previously (Thomas et al., 2000). The generation and characterization of polyclonal antimyostatin antibodies have been described previously (Sharma et al., 1999). Monoclonal and polyclonal anti-NPAT antibodies were a generous gift from Dr Jiyong Zhao (Massachusetts General Hospital Cancer Center, Charlestown, MA, USA). Cells and cell culture RD cells were obtained from the ATCC (Rockville, MD, USA). RD cells were grown prior to assay in Dulbecco’s modified Eagle’s medium (DMEM; Invitrogen), buffered with 41.9 mM NaHCO3 (Sigma Cell Culture Ltd) and 5% gaseous CO2. A measure of 7.22 nM phenol red (Sigma) was used as a pH indicator. Penicillin (1 105 IU/l) (Sigma), 100 mg/l streptomycin (Sigma) and 10% fetal bovine serum (FBS; Invitrogen) were added to the media. RD cells were seeded on Thermonox coverslips or 6-cm plates (Nunc) at a density of 20 000 cells/cm2. Following a 16-h attachment period, test media consisting of DMEM media containing 10% FBS (DMEM/10% FBS) with or without 0–10 mg/ml of recombinant myostatin were added. Plates were incubated at 371C, 5% CO2 for a further 24, 48 or 72 h. After the incubation period, proliferating RD cells were harvested for protein or RNA or assayed using a methylene blue photometric endpoint assay as previously described (Oliver et al., 1989). In this assay, absorbance at 655 nm is directly proportional to the final cell number. Results are presented here as the mean and standard error of eight replicates. The position of the samples on the plate was randomly assigned and all samples were run in replicates of eight. The results presented in this paper are representative of at least two independent experiments. To examine the reversibility of myostatin inhibition on RD, cells were seeded onto a Nunc 96-well plate at a density of 18 000 cells/cm2. After a 16-h attachment period, DMEM/10% FBS media with or without 6 mg/ml myostatin were added. Cultures were incubated at 371C, 5% CO2 for 48 h, after which the media were replaced with fresh media at 96 and 144 h. Cells were fixed for methylene blue assay at 192 h. Cultures to have myostatin inhibition ‘reversed’ received DMEM/10% FBS media without myostatin at 48 h, with further media changes at 96 and 144 h. Cdk2, cyclin-E expression plasmids (Zhao et al., 1998) were stably transfected into RD cells using lipofectamine 2000 according to the manufacturer’s protocol. Immunocytochemistry RD cells were fixed with 70% ethanol : formaldehyde : glacial acetic acid (20 : 2 : 1) for 30 s, then rinsed three times with PBS. Cells were blocked overnight at 41C in TBS (0.05 M Tris-HCl, pH 7.6 (Sigma); 0.15 M NaCl) containing 1% normal sheep serum (NSS). Cells were incubated with primary antibody, 1 : 100 dilution antimyostatin antibody (Thomas et al., 2000), in TBS/1% NSS for 1 h. Rabbit IgG (5 mg/ml; Dako) was used Oncogene

as a negative control. Cells were washed (3 5 min) with TBST (TBS; 0.05% Tween-20) and incubated with secondary antibody, 1 : 300 dilution biotinylated donkey anti-rabbit IgG (RPN1004; Amersham) in TBS/1% NSS for 30 min. Cells were washed as before and incubated with a tertiary antibody, 1 : 100 dilution of streptavidin–biotin peroxidase complex (RPN1051; Amersham) in TBS/1% NSS for 30 min. Myostatin-immunostaining was visualized using 3,3-diaminobenzidine tetrahydrochloride (DAB; Invitrogen) enhanced with 0.0375% CoCl. Myostatin-immunostained cultures were lightly counterstained with Gill’s hematoxylin and photographed using an Olympus BX50 microscope (Olympus Optical Co., Germany) fitted with a DAGE-MTI DC-330 color camera (DAGE-MTI Inc., IN, USA). BrdU incorporation For BrdU incorporation, RD cells were seeded onto Nunc Thermonox coverslips at a density of 25 000 cells/cm2 and cultured as described above. After a 16-h attachment period, RD cells were arrested in G0 by culturing in methionine-free DMEM (Invitrogen) containing 1% FBS. After 48 h of incubation, the methionine-free media were replaced by growth media containing 1 : 1000 BrdU labelling reagent (Amersham) and either 0 or 10 mg/ml myostatin. RD cells were cultured for a further 30 h. RD cells were fixed in 70% ethanol for 5 min and blocked overnight at 41C in TBS containing 1% NSS. Cells were washed two times with PBS, incubated with 4 N HCl for 10 min, rinsed two times with 0.1 M (pH 8.5) and rinsed with PBS. Primary antibody incubations were with 1 : 1000 anti-BrdU antibody (Sigma) for 1 h in TBS containing 1% NSS. Cells were washed (3 5 min) with TBST and incubated with secondary antibody, 1 : 100 dilution of biotinylated sheep anti-mouse IgG (RPN1001; Amersham) in TBS/1% NSS for 30 min. Cells were washed as before and incubated with a tertiary antibody, 1 : 100 dilution of streptavidin–biotin peroxidase complex (RPN1051; Amersham) in TBS/1% NSS for 30 min. BrdU-immunostaining was visualized using DAB (Invitrogen) enhanced with 0.0375% CoCl. BrdU-immunostained cultures were counterstained with Gill’s hematoxylin and photographed. Immunoprecipitation and immunoblot analysis For quantitative CKI, Cdk, cyclin and Rb immunoblot analyses, mouse C2C12 myoblasts were cultured as described above in six-well plates or 100 mm dishes according to treatment. Cells (B3 106) were resuspended in 200 ml lysis buffer (50 mM Tris pH 7.6; 250 mM NaCl; 5 mM EDTA; 0.1% Nonidet P-40; complete protease inhibitor (Roche Molecular Biochemicals)) and sonicated or passed through a syringe. The cell extracts were centrifuged to pellet cell debris. The Bradford Reagent (Bio-Rad Laboratories) was used to estimate the total protein content to ensure equal loadings. Immunoprecipitations were performed by incubating 200 mg of total protein extract with 1 mg of monoclonal anti-Rb antibody (G3-245; PharMingen) in 100 ml of extraction buffer for 1 h at 41C. Protein A–Agarose (Invitrogen) (50 ml of 50% – washed twice with lysis buffer) was added for 1 h at 41C, followed by centrifugation to pellet immunoprecipitated complexes. After centrifugation, pellets were washed five times with lysis buffer. Pellets were then resuspended in 20 ml 4 NuPAGE sample buffer (Novex) and boiled for 5 min. The total protein (15 mg) or immunoprecipitations were fractionated by 4–12% SDS–PAGE (Novex) and transferred onto a nitrocellulose membrane by electroblotting. The membranes were blocked in TBST/5% milk at 41C overnight,


533 followed by incubation with the primary antibody for 3 h at room temperature. The following primary antibodies were used for immunoblotting; myostatin: 1 : 2000 dilution of rabbit polyclonal antimyostatin antibody (Thomas et al., 2000); p15, 1 : 400 dilution of purified rabbit polyclonal anti-p15 antibody (sc-613; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA); p16, 1 : 400 dilution of purified rabbit polyclonal antip16 antibody (sc-1207; Santa Cruz Biotechnology Inc.); p21, 1 : 400 dilution of purified mouse monoclonal anti-p21 antibody (SX118; PharMingen, CA, USA); p27, 1 : 400 dilution of purified mouse monoclonal anti-p27 antibody (sc-1641; Santa Cruz Biotechnology Inc.); Cdk2, 1 : 400 dilution of purified mouse monoclonal anti-Cdk2 antibody (sc-2648; Santa Cruz Biotechnology Inc.); Cdk4, 1 : 400 dilution of purified rabbit polyclonal anti-Cdk4 antibody (sc-601; Santa Cruz Biotechnology Inc.); cyclin-D1, 1 : 400 dilution of purified mouse monoclonal anticyclin-D1 antibody (sc-8396; Santa Cruz Biotechnology Inc.); cyclin-E, 1 : 400 dilution of purified rabbit polyclonal anticyclin-E antibody (sc-481; Santa Cruz Biotechnology Inc.); a-tubulin, 1 : 3000 dilution of purified mouse monoclonal anti-a-tubulin antibody (DM 1A; Sigma); Rb, 1 : 200 dilution of purified mouse monoclonal anti-pRb antibody (G3-245; PharMingen); E2F1, 1 : 250 dilution of purified mouse monoclonal anti-E2F1 antibody (KH95; PharMingen); NPAT, 1 : 250 dilution of purified mouse monoclonal antiNPAT antibody (C27; Transduction Laboratories) or 1 : 2000 rabbit polyclonal anti-NPAT antibody (Zhao et al., 2000). The membranes were washed (5 5 min) with TBST and further incubated with anti-mouse IgG HRP conjugate, 1 : 2000 dilution (W402B; Promega Corp.) or anti-rabbit IgG HRP conjugate, 1 : 1000 dilution (P0448; Dako) secondary anti-

bodies for 1 h at room temperature. The membranes were washed as above, and HRP activity was detected using Renaissance Western blot chemiluminescence (NEL104; NEN Life Science Products Inc.). Northern blot analysis and RT–PCR RNA was isolated from cultured RD cells using Trizol reagent (Invitrogen). Northern analysis was performed using 12 mg of total RNA. The membrane was hybridized with 32P-labelled histone H4 cDNA probe in Church and Gilbert hybridization buffer at 601C, overnight. The membrane was washed at 601C for 15 min each with 2 SSC, 0.5% SDS, and then 1 SSC, 0.5% SDS. The myostatin and histone H4 cDNAs were obtained by RT–PCR. First-strand cDNA was synthesized in a 20 ml RT reaction from 5 mg total RNA (from RD cell or C2C12 myoblast total RNA) using SuperScript II preamplification kit (Invitrogen). PCR was performed with 2 ml of the RT reaction at 941C for 20 s, 521C (or 551C for histone H4) for 20 s,and 721C for 1 min for 35 cycles using Taq polymerase (Roche). This was followed by a single 721C extension step for 5 min. The primers used for Myostatin were 50 -ATGCAAAAACTGCAATCTCTG-30 and 50 -TCA TGAGCACCCACAGCGATC-30 (1128 bp); histone H4 were 50 -ATGTCCGGCTGTG GAAAG-30 and 50 -TCGAAACG TGCAAAGCTG-30 (339 bp). The histone H4 probe cDNA was radioactively labelled using a32P-(dCTP) (Amersham) and the Rediprime II labelling kit (Amersham) according to the manufacturer’s protocol.

References Aguanno S, Bouche M, Adamo S and Molinaro M. (1990). Cancer Res., 50, 3377–3382. Albig W and Doenecke D. (1997). Hum. Genet., 101, 284–294. Arthur PF. (1995). Aust. J. Agric. Res., 46, 1493–1515. Bennicelli JL, Fredericks WJ, Wilson RB, Rauscher III FJ and Barr FG. (1995). Oncogene, 11, 119–130. Bouche M, Canipari R, Melchionna R, Willems D, Senni MI and Molinaro M. (2000). FASEB J., 14, 1147–1158. Bouche M, Senni MI, Grossi AM, Zappelli F, Polimeni M, Arnold HH, Cossu G and Molinaro M. (1993). Exp. Cell. Res., 208, 209–217. Braun T, Buschhausen-Denker G, Bober E, Tannich E and Arnold HH. (1989). EMBO J., 8, 701–709. Choi J, Costa ML, Mermelstein CS, Chagas C, Holtzer S and Holtzer H. (1990). Proc. Natl. Acad. Sci. USA, 87, 7988–7992. Dagher R and Helman L. (1999). Oncologist, 4, 34–44. Davis RJ, D’Cruz CM, Lovell MA, Biegel JA and Barr FG. (1994). Cancer Res., 54, 2869–2872. Edmondson DG and Olson EN. (1989). Genes Dev., 3, 628–640. Ewen ME. (2000). Genes Dev., 14, 2265–2270. Feinberg AP. (1999). Cancer Res., 59, 1743s–1746s. Galili N, Davis RJ, Fredericks WJ, Mukhopadhyay S, Rauscher III FJ, Emanuel BS, Rovera G and Barr FG. (1993). Nat. Genet., 5, 230–235. Geng Y and Weinberg RA. (1993). Proc. Natl. Acad. Sci. USA, 90, 10315–10319. Grobet L, Poncelet D, Royo LJ, Brouwers B, Pirottin D, Michaux C, Menissier F, Zanotti M, Dunner S and Georges M. (1998). Mamm. Genome, 9, 210–213. Heintz N. (1991). Biochim. Biophys. Acta, 1088, 327–339.

Hofmann F and Livingston DM. (1996). Genes Dev., 10, 851–861. Houghton JA, Houghton PJ, Brodeur GM and Green AA. (1981). Int. J. Cancer, 28, 409–415. Hu X and Zuckerman KS. (2001). J. Hematother. Stem Cell Res., 10, 67–74. Ji S, Losinski RL, Cornelius SG, Frank GR, Willis GM, Gerrard DE, Depreux FF and Spurlock ME. (1998). Am. J. Physiol., 275, R1265–R1273. Kambadur R, Sharma M, Smith TP and Bass JJ. (1997). Genome Res., 7, 910–916. Langley B, Thomas M, Bishop A, Sharma M, Gilmour S and Kambadur R. (2002). J. Biol. Chem., 277, 49831–49840. Lee SJ and McPherron AC. (2001). Proc. Natl. Acad. Sci. USA, 98, 9306–9311. Ludolph DC and Konieczny SF. (1995). FASEB J., 9, 1595–1604. Ma T, Van Tine BA, Wei Y, Garrett MD, Nelson D, Adams PD, Wang J, Qin J, Chow LT and Harper JW. (2000). Genes Dev., 14, 2298–2313. McPherron AC, Lawler AM and Lee SJ. (1997). Nature, 387, 83–90. McPherron AC and Lee SJ. (1997). Proc. Natl. Acad. Sci. USA, 94, 12457–12461. Megeney LA and Rudnicki MA. (1995). Biochem. Cell Biol., 73, 723–732. Merlino G and Helman LJ. (1999). Oncogene, 18, 5340–5348. Miner JH and Wold B. (1990). Proc. Natl. Acad. Sci. USA, 87, 1089–1093. Niggli FK, Powell JE, Parkes SE, Ward K, Raafat F, Mann JR and Stevens MC. (1994). Br. J. Cancer, 69, 1106–1110. Oncogene


534 Ohtsubo M, Theodoras AM, Schumacher J, Roberts JM and Pagano M. (1995). Mol. Cell. Biol., 15, 2612–2624. Oliver MH, Harrison NK, Bishop JE, Cole PJ and Laurent GJ. (1989). J. Cell Sci., 92, 513–518. Olson EN and Klein WH. (1994). Genes Dev., 8, 1–8. Osley MA. (1991). Annu. Rev. Biochem., 60, 827–861. Otten AD, Firpo EJ, Gerber AN, Brody LL, Roberts JM and Tapscott SJ. (1997). Cell. Growth Differ., 8, 1151–1160. Owa T, Yoshino H, Yoshimatsu K and Nagasu T. (2001). Curr. Med. Chem., 8, 1487–1503. Pandita A, Zielenska M, Thorner P, Bayani J, Godbout R, Greenberg M and Squire JA. (1999). Neoplasia, 1, 262–275. Pappo AS, Crist WM, Kuttesch J, Rowe S, Ashmun RA, Maurer HM, Newton WA, Asmar L, Luo X and Shapiro DN. (1993). J. Clin. Oncol., 11, 1901–1905. Perry RL and Rudnick MA. (2000). Front. Biosci., 5, D750–D767. Puri PL, Wu Z, Zhang P, Wood LD, Bhakta KS, Han J, Feramisco JR, Karin M and Wang JY. (2000). Genes Dev., 14, 574–584. Scheidler S, Fredericks WJ, Rauscher III FJ, Barr FG and Vogt PK. (1996). Proc. Natl. Acad. Sci. USA, 93, 9805–9809. Scrable H, Cavenee W, Ghavimi F, Lovell M, Morgan K and Sapienza C. (1989). Proc. Natl. Acad. Sci. USA, 86, 7480–7484. Shapiro DN, Parham DM, Douglass EC, Ashmun R, Webber BL, Newton Jr WA, Hancock ML, Maurer HM and Look AT. (1991). J. Clin. Oncol., 9, 159–166. Shapiro DN, Sublett JE, Li B, Downing JR and Naeve CW. (1993). Cancer Res., 53, 5108–5112. Sharma M, Kambadur R, Matthews KG, Somers WG, Devlin GP, Conaglen JV, Fowke PJ and Bass JJ. (1999). J. Cell Physiol., 180, 1–9.

Oncogene

Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149–1163. Stein GS, Stein JL, van Wijnen AJ and Lian JB. (1992). Curr. Opin. Cell. Biol., 4, 166–173. Szabo G, Dallmann G, Muller G, Patthy L, Soller M and Varga L. (1998). Mamm. Genome, 9, 671–672. Tapscott SJ, Thayer MJ and Weintraub H. (1993). Science, 259, 1450–1453. Tapscott SJ and Weintraub H. (1991). J. Clin. Invest., 87, 1133–1138. Thies RS, Chen T, Davies MV, Tomkinson KN, Pearson AA, Shakey QA and Wolfman NM. (2001). Growth Factors, 18, 251–259. Thomas M, Langley B, Berry C, Sharma M, Kirk S, Bass J and Kambadur R. (2000). J. Biol. Chem., 275, 40235–40243. Turc-Carel C, Lizard-Nacol S, Justrabo E, Favrot M, Philip T and Tabone E. (1986). Cancer Genet. Cytogenet., 19, 361–362. van den Heuvel S and Harlow E. (1993). Science, 262, 2050–2054. Weintraub H, Tapscott SJ, Davis RL, Thayer MJ, Adam MA, Lassar AB and Miller AD. (1989). Proc. Natl. Acad. Sci. USA, 86, 5434–5438. Yang J, Ratovitski T, Brady JP, Solomon MB, Wells KD and Wall RJ. (2001). Mol. Reprod. Dev., 60, 351–361. Zhao J, Dynlacht B, Imai T, Hori T and Harlow E. (1998). Genes Dev., 12, 456–461. Zhao J, Kennedy BK, Lawrence BD, Barbie DA, Matera AG, Fletcher JA and Harlow E. (2000). Genes Dev., 14, 2283–2297. Zhu X, Hadhazy M, Wehling M, Tidball JG and McNally EM. (2000). FEBS Lett., 474, 71–75.