Tbx3 impinges on the p53 pathway to suppress apoptosis ... - Nature

Oncogene (2002) 21, 3827 ± 3835 ã 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

ORIGINAL PAPERS

Tbx3 impinges on the p53 pathway to suppress apoptosis, facilitate cell transformation and block myogenic dierentiation Hanqian Carlson1,2, Sara Ota1, Youngsup Song1, Yanwen Chen1,2 and Peter J Hurlin*,1,2 1

Shriners Hospitals for Children, Oregon Health Sciences University, 3101 Sam Jackson Park Road, Portland, Oregon 97201, USA; 2Department of Cell and Developmental Biology, Oregon Health Sciences University, 3101 Sam Jackson Park Road, Portland, Oregon 97201, USA

Tbx3 is a member of the T-box family of transcription factors. Mutations in Tbx3 cause ulnar-mammary syndrome, an autosomal dominant disorder characterized by upper limb defects, apocrine-gland defects including mammary hypoplasia, and tooth, hair and genital defects. In cell culture, Tbx3 and its close relative Tbx2 are capable of immortalizing mouse embryo ®broblasts. We show that expression of Tbx3 together with Myc or oncogenic Ras (H-RasVal17) leads to ecient transformation of mouse embryo ®broblasts. Oncogene cooperation by Tbx3 correlates with an ability of Tbx3 to suppress the induction of p19ARF and p53 that is typically caused by overexpression Myc and Ras, and to protect against Myc-induced apoptosis. Whereas Tbx3 is capable of interfering with apoptosis caused by excessive Myc levels, a Tbx3 mutant lacking its C-terminal repression domain shows no anti-apoptotic activity and fails to repress levels of p19ARF or p53. Consistent with an ability to suppress p53 pathway function, we ®nd that Tbx3, but not a Tbx3 C-terminal mutant, eciently blocks myogenic dierentiation of C2C12 myoblasts. Our results support the idea that deregulation and/or excessive levels of Tbx3 may have oncogenic potential in vivo. Oncogene (2002) 21, 3827 ± 3835. DOI: 10.1038/sj/ onc/1205476 Keywords: T-box; Tbx3, Myc, Ras, p19ARF; p53 Introduction The Tbx family of transcription factors perform a variety of functions critical to early embryonic development (Smith, 1999). The prototypical Tbx protein brachyury is required for posterior, axial and paraxial mesoderm formation as well as formation of the notochord (Smith, 1999). Whereas homozygous mutations in Brachyury cause early embryonic lethality, heterozygous Brachyury mutations are responsible for the classical mouse mutant tailess (Herrmann et al., 1990). Although mutations in Brachyury have

*Correspondence: PJ Hurlin; E-mail: [email protected] Received 23 January 2002; revised 7 March 2002; accepted 12 March 2002

not been linked to any human disease or birth defects, mutations in a number of other Tbx genes have. These include Tbx1 (DiGeorge syndrome, see Schinke and Izumo (2001) for review), Tbx3 (ulnar-mammary syndrome, Bamshad et al., 1997), Tbx5 (Holt-Oram syndrome, Basson et al., 1997; Li et al., 1997) and Tbx22 (X-linked cleft palate and ankyloglossia, Braybrook et al., 2001). Consistent with known functions of Brachyury, a common theme amongst the human disease syndromes caused by mutations in Tbx family genes is the disruption of regional epithelial/mesodermal interactions and associated defects in mesoderm induction and dierentiation events. Ulnar-mammary syndrome (UMS), caused by heterozygous mutations in Tbx3 (Bamshad et al., 1997, 1999), is an autosomal dominant disorder characterized by de®ciencies and duplications of posterior elements of the forelimb, by mammary hypoplasia and apocrine gland de®ciency/dysfunction and by tooth, hair and sex organ defects. Tbx3 is related to Brachyury and other Tbx family members by their shared DNA-binding motif known as the T-domain. Some mutations in Tbx3 that cause UMS disrupt the T-domain and are thought to function as null alleles (Bamshad et al., 1997), but others occur downstream from the Tdomain and may aect the transcriptional repression activity or other functions of Tbx3 (Bamshad et al., 1999; He et al., 1999; Carlson et al., 2001). The range and nature of defects associated with UMS, together with sites of Tbx3 expression (Chapman et al., 1996; Gibson-Brown et al., 1998) are consistent with a key role of Tbx3 as a mediator of regional mesoderm induction and dierentiation. However, very little is known concerning the underlying molecular mechanisms responsible for such activity. Whereas the relationship between most Tbx family members is limited to the conserved T-domain, Tbx2 and Tbx3 show much more extensive sequence relationship. Indeed the Tbx2/Tbx3 gene pair and the more distantly related Tbx4/Tbx5 gene pair is thought to have evolved from an ancestral gene through a twostep process involving an initial tandem duplication event followed by a parallel duplication event (Agulnik, 1996). Beyond the sequence similarities, it is now known that Tbx2 and Tbx3 both behave as transcriptional repressors and contain a conserved repression domain in their C-terminus (He et al.,

Tbx3 effects on cell growth and differentiation H Carlson et al

3828

Oncogene

1999; Carlson et al., 2001). Tbx3 and Tbx2 also share similar, but not identical expression patterns during embryonic development (Chapman et al., 1996; Gibson-Brown et al., 1998). However, mutations in Tbx2 have not been found associated with UMS or any other disease syndrome and the apparent haploinsuciency basis for UMS supports the notion that Tbx2 and Tbx3 do not have absolutely redundant functions during embryonic development. Nonetheless, the close sequence relationship between Tbx3 and Tbx2 suggests that these proteins may share redundant biological activities. Recent cell culture assay systems demonstrate that Tbx2 has functions that impinge on the replicative lifespan of mouse and human ®broblasts (Jacobs et al., 2000). Tbx2 was found capable of bypassing the premature senescence phenotype of mouse embryo ®broblasts lacking the polycomb group protein Bmi1 (Jacobs et al., 2000). Further, Tbx2 and, more recently, Tbx3, was found capable of immortalizing primary mouse embryo ®broblasts (Jacobs et al., 2000; Carlson et al., 2001; Brummelkamp et al., 2002). The underlying mechanism for the ability of Tbx2 and Tbx3 to prevent senescence and immortalize cells appears to be through transcriptional repression of the Cdkn2a (p19ARF) gene (Jacobs et al., 2000; Brummelkamp et al., 2002). Since cells lacking p19ARF fail to senesce in culture and can be propagated inde®nitely (Kamijo et al., 1997), downregulation of p19ARF by Tbx2 and Tbx3 provides a molecular explanation for its ability to immortalize cells. p19ARF regulates the abundance of p53 tumor suppressor protein through its physical interaction with Mdm2 oncoprotein, a ubiquitin ligase that mediates the degradation of p53 (Sherr and Weber, 2000; Sharpless and DePinho, 1999). Binding of p19ARF to Mdm2 blocks Mdm2 nucleo-cytoplasmic shuttling and sequesters it to the nucleolus, thus preventing its targeting of p53 for destruction (Tao and Levine, 1999). It is therefore predicted, but has not yet been demonstrated, that downregulation of p19ARF and immortalization by Tbx2 and Tbx3 should correlate with decreased levels of p53 and suppression of its cell cycle arrest/senescence functions. In addition to its functions in regulating cellular lifespan, the p19ARF/Mdm2/p53 pathway plays an important role in regulating apoptosis (Sherr and Weber, 2000; Sharpless and DePinho, 1999). One way that the apoptotic function of the p19ARF/MDM2/ p53 pathway is activated, is in response to excessive proliferation stimuli, such as that caused by oncogene activation/overexpression. For example, oncogenes such as Myc and Ras induce p19ARF expression and consequently allow stabilization and accumulation of p53 (Serrano et al., 1997; Zindy et al., 1998; Palmero et al., 1998). Such increased p53 levels promote apoptosis and/or cell cycle arrest and this function is hypothesized to serve as a default mechanism to promote the destruction, or interfere with the progression of cancer prone cells (Evan and Vousden, 2001). Furthermore, mutations that disrupt p53-mediated apoptosis appear central to cancer development. Thus the ability of Tbx2

to downregulate p19ARF, together with the ®nding that it is ampli®ed in a subset of mammary carcinomas, has raised the possibility that it may function as an oncogene (Jacobs et al., 2000). In this study, we demonstrate that Tbx3 downregulates p19ARF and p53 levels in cells, can suppress apoptosis and facilitate cell transformation by Myc and Ras oncogenes and interfere with myogenic dierentiation. Our results suggest that events leading to excessive expression levels of Tbx3 may contribute to tumorigenesis in vivo, and that defective regulation of p53 caused by mutations in Tbx3 may be at least partly responsible for ulnar-mammary syndrome.

Results Tbx3 cooperates with Myc and Ras to transform cells We previously showed that ectopic expression of Tbx3 was capable of immortalizing primary MEFs (Carlson et al., 2001). Since cell immortalizing events, such as mutation or loss of p19ARF or p53 are required for oncogenic transformation or progression caused by oncoproteins such as Myc and oncogenic H-Ras (Serrano et al., 1997; Zindy et al., 1998; Palmero et al., 1998), we tested whether expression of Tbx3 could cooperate with Myc and H-RasV12 to transform cells. Primary MEFs at passage 1 were transfected with Tbx3 alone, Myc alone or H-RasV12 alone or transfected with equal amounts of Tbx3 and Myc or H-RasV12. Whereas Tbx3 alone, Myc alone and H-RasV12 alone produced only rare or no foci of transformed cells, coexpression of Tbx3 with either Myc or H-RasV12 produced a large number of transformed foci (Figure 1a,b). Further, coexpression of Tbx3 with Myc and H-RasV12 resulted in the production of far more foci than observed with Myc and H-RasV12 alone (Figure 1a,b). These results suggest that functions associated with the ability of Tbx3 to immortalize cells (Carlson et al., 2001), allows Tbx3 to facilitate cell transformation by Myc and HRasV12 oncoproteins. In addition to focus-formation, cells infected with Tbx3 and Myc or Tbx3 and H-RasV12 were tested for the ability to form colonies in soft agar. Transfected MEFs were allowed to grow for 2 weeks, at which time transformed foci became visible in Tbx3 plus Myc, Tbx3 plus H-RasV12 and H-RasV12 plus Myc populations. Cells were pooled from individual transfectant populations, plated in soft agar and the number of colonies formed were counted 3 weeks later. Tbx3 expression was found to have little or no ability to promote soft agar colony formation in the presence of H-RasV12, Myc or H-RasV12 plus Myc (data not shown). Tbx3 downregulates p19ARF and p53 and protects against Myc-induced apoptosis Thus the ability of Tbx3 to immortalize MEFs and facilitate cell transformation by ectopic expression of


A

Figure 1 Focus formation assay showing that Tbx3 cooperates with Myc and H-RasV12 to transform primary cells. Mouse embryo ®broblasts (MEFs) at passage 1 were transfected with plasmids expressing the indicated proteins. Plates on the right side were cotransfected with Tbx3 expression vector. Growth medium was replenished every 2 days for 14 ± 18 days and cells stained with methylene bluetohighlightcoloniesthatformeddensefoci (a). Similar results were achieved in each of two independent experiments (b)

Myc and H-RasV12, suggested that Tbx3 might exert its activity by interfering with p53 pathway functions. Further, the closely related Tbx2 protein was previously demonstrated to downregulate the p19ARF gene (Jacobs et al., 2000). We examined the expression of p19ARF, Mdm2, p53 and levels of the p53 target gene product p21, in cells infected with Tbx3 and in cells sequentially transfected with Tbx3 and c-Myc or H-Ras (Figure 2a). Coinfection eciencies using this protocol were typically in the range of 25 ± 50%, based on the infection eciency of a control green-¯uorescent protein-expressing virus (not shown). Levels of p19ARF were found to be downregulated by Tbx3 alone and the elevated p19ARF levels caused by Myc or H-Ras overexpression were suppressed in the presence of Tbx3 (Figure 2a). Consistent with downregulation of p19ARF levels, the elevated levels of p53 and p21 in cells overexpressing Myc and Ras were downregulated in the presence of Tbx3 (Figure 2a). In contrast to p19ARF, levels of p16 (Cdkn1a) were not aected by Tbx3, but were upregulated by H-Ras as previously reported (Serrano et al., 1997). Examination of total pRb levels demonstrated the presence of a slower mobility species that appeared most abundant in Tbx3-expressing cells. The presence of a slower mobility form(s) of pRb is consistent with hyper-phosphorylation and inactivation of Rb. Using antibodies that speci®cally recognize phosphorylated forms of pRb, a form of pRb speci®cally phosphorylated at Ser 807/810 was detected. Phosphorylation of pRb at 807/811 within the pRb `C-pocket' has been demonstrated to cause release of bound Abl tyrosine kinase (Welch and Wang, 1993, 1995). Unlike the 807/811-phosphorylated form of pRb, we were unable to consistently show the presence of other phosphorylated forms, nor were we able to demonstrate increased activity of cyclin-dependent kinase 2, 4 or 6 (data not shown). Primary cells overexpressing Myc are prone to apoptosis under serum (survival factor) ± limiting condition (Evan and Vousden, 2001) and several studies indicate the disruption of p53 pathway functions associated with apoptosis is required, or greatly facilitates Myc-induced oncogenic transformation (Zindy et al., 1998; Eischen et al., 1999; Jacobs et al., 1999; Maestro et al., 1999). Therefore we next examined the ability of Tbx3 to suppress apoptosis and its aects on the cell cycle. MEFs expressing Tbx3 alone or in combination with Myc or Ras were subjected to serum-limiting conditions (0.2% fetal bovine serum) for 48 h and apoptotic cells identi®ed by FACS analysis of propidium iodide-stained cells. As shown in Figure 2b, overexpression of Myc induced apoptosis as indicated by a dramatic increase in the percentage of the sub-G1 fraction of cells when in 0.2% serum (see arrow). Coexpression of Tbx3 with Myc led to a decrease in the level of apoptosis to near that of control cells infected with empty virus or Tbx3 alone (Figure 2b, see arrow). Coexpression of Tbx3 with H-Ras appeared to cause a decrease in the

3829

Oncogene


3830

slightly elevated level of apoptosis seen in cells overexpressing H-Ras in 0.2% serum and also caused a slight increase in the percentage cells in S phase compared to cells overexpressing H-Ras alone (Figure 2b). Since the sequential viral infection protocol used in the above experiments allowed only a fraction of the target cells to be infected, a second approach was used to examine the extent to which Tbx3 could protect cells against Myc-dependent apoptosis. For these experiments, we took advantage of an immortal ®broblast cell line developed by Mateyak et al. (1999) called HO ± Myc, which lacks endogenous cMyc, but expresses ectopic, deregulated c-Myc at levels that are moderately elevated compared to its parental Rat1a cell line. Under normal serum levels (10%) these cells grow similarly to wild-type cells,

but when deprived of serum, are highly prone to apoptosis (Mateyak et al., 1999). We used retroviral infection to introduce Tbx3 and a mutant form of Tbx3 lacking the entire C-terminus downstream from the T-domain (Tbx3-300, Carlson et al., 2001) into HO-Myc cells and examined the apoptotic response following reduction of serum levels to 0.2% (Figure 3a,b). Retroviral infection eciencies using HO-Myc cells were typically 65 ± 80% (not shown). When infected with either empty vector or Tbx3-300, HOMyc cells showed a strong apoptotic response. In contrast, HO-Myc cells infected with Tbx3 virus showed signi®cant resistance to apoptosis (Figure 3a,b). However, by 96 h, many of the Tbx3-infected population had succumbed to apoptosis (Figure 3a,b). In comparison, overexpression of Bcl2 in HO-Myc cells resulted in at least a 2-fold greater

Figure 2 Eects of Tbx3 expression on levels of p53 pathway proteins and apoptosis. Primary MEFs at passage 1 were infected with empty virus, Tbx3, Myc, or Ras virus alone. For co-expression of Tbx3 with Myc and with Ras, Tbx3 infected cells were subsequently infected with either empty virus or by Myc or Ras viruses. For the singly expressing Tbx3, Myc and Ras cell populations, a second infection procedure was also performed, but with empty virus. Infected cells were harvested 2 days after the second infection for Western blot analysis (a) or for FACS analysis (b). For Western blots, equal amounts of cell extracts were used to determine the relative levels of the indicated proteins using antibodies listed in Materials and methods. Tbx3 and Tbx3-300 proteins contain Ha-antibody epitopes at their C-termini (Carlson et al., 2001). For FACS analysis, cells were ®xed in 70% ethanol and stained with propidium iodide. Relative DNA content of individual cells is shown in (b), with the growth phase 1 (G1, diploid DNA content) and growth phase 2/mitosis (G1/M, tetraploid DNA content) populations indicated, along with the percentage of apoptotic cells (exhibiting a sub-GI DNA content). Arrows point to cells infected with Myc and with Tbx3 plus Myc that show a sub-G1 DNA content characteristic of cells undergoing apoptosis. Note the decrease in the population of apoptotic cells that occurs when Tbx3 is coexpressed with Myc. The percentage of apoptosis present in each of the cell populations, based on sub-G1 DNA content is indicated Oncogene


3831

Figure 3 Tbx3 suppresses Myc-dependent apoptosis. HO ± Myc cells were infected with viruses expressing the indicated proteins and 2 days later growth medium (10% serum) was replaced with low serum (0.2%) medium. Cells were photographed at the indicated times following lowering the serum (a). The appearance of highly refractile, rounded and picnotic cells are characteristic of apoptosis. To quantitate apoptosis, all cells were collected from empty virus, Tbx3 ± 300 and Tbx3 infected populations at the indicated times, ®xed in 70% ethanol, stained with propidium iodide and DNA content measured by FACS (b). The percentage of apoptotic cells determined by the number of cells showing a sub-G1 (diploid) DNA content is indicated. Represented are the results of two separate experiments performed in duplicate (error bars indicate standard deviations). (c) Reduced expression levels of p53pathway components caused by Tbx3 in HO ± Myc cells. HO ± Myc cells infected with empty virus, Tbx3 or Tbx3 ± 300 were harvested 2 days following viral infection, lysed and equal amounts of cell extracts used for Western blot analysis of the indicated proteins as described in Materials and methods

number of cells protected from apoptosis under identical conditions (not shown). These results suggest that Tbx3 signi®cantly delays the robust apoptotic response of these cells and may protect some HO-Myc cells from death, but does not completely disrupt pathways required for apoptosis. To examine the eects of Tbx3 on the p53 pathway in HO-Myc cells, levels of p19ARF, p53, Mdm2 and p21 were compared in uninfected cells and in cells infected with Tbx3 and Tbx3-300. Levels of p19ARF were consistently found to be signi®cantly lower in Tbx3 cells, as was p53, and the p53 target gene products Mdm2 and p21 (Figure 3c). In contrast,

Tbx3-300 expression appeared to cause a slight upregulation of p19ARF, p53, Mdm2 and p21 (Figure 3c). Thus overexpression of Tbx3 downregulates, but does not shut o expression of p19ARF and p53, and such downregulation is consistent with the observation that Tbx3 partially protects against apoptosis of HOMyc cells. Suppression of myogenic differentiation by Tbx3 The p53 and pRb pathways play critical roles in the control of myogenic dierentiation (Walsh and Perlman, 1997; Cerone et al., 2000). The ability of Tbx3 Oncogene


3832

to regulate p19ARF and p53 levels and to induce phosphorylation and possibly inactivate Rb (Figures 2a and 3c) led us to test whether ectopic expression of Tbx3 in the myogenic cell line C2C12 could prevent dierentiation. For these experiments, pools of C2C12 cells expressing Tbx3, Tbx3-300 or no Tbx3 protein were generated by stable transfection using G418 selection. These dierent populations were tested for their ability to dierentiate into myotubes when the culture medium was changed from growth medium (15% bovine calf serum) to dierentiation medium (2% horse serum). Whereas vector control cells and Tbx3-300 cells showed an abundance of myotubes following 5 days in dierentiation medium, very few myotubes were observed in the Tbx3 expressing population (Figure 4). Consistent with a disrupted dierentiation response, Tbx3 expressing cells failed to upregulate the muscle dierentiation marker myosin light chain (MLC) and the myogenic regulatory factor myogenin (Figure 5). Interestingly, the basal levels of both myogenin and myoD were reduced in Tbx3-expressing cells compared to vector control cells, as was MLC, and their expression remained suppressed following exposure to dierentiation medium for 5 days (0 days, Figure 4).

Figure 5 Northern blot comparing the induction of the muscle dierentiation marker myosin light chain and expression levels of the muscle regulatory factors myogenin and MyoD in control, empty virus infected C2C12 cells and C2C12 cells stably expressing Tbx3. Glyceraldehyde phosphate dehydrogenase (GFAP) served as a loading control

It is important to note that Tbx3 expression appeared to be toxic to a sub-population of C2C12 cells and that another sub-population was capable of undergoing dierentiation into myotubes following more extended time in dierentiation medium (not shown). One possibility is that these dierent responses are caused by dierent expression levels of Tbx3 in the pooled populations of transfected cells used in these experiments. Nonetheless, we conclude that Tbx3 expression provides a strong inhibitory signal for myogenic dierentiation. Discussion

Figure 4 Tbx3 blocks myogenic dierentiation. C2C12 myogenic cells were transfected with pFB, pFB ± Tbx3 or pFB ± Tbx3 ± 300 and stable transfectants selected by resistance to the G418 resistance conferred by the virus. Pooled populations, growing in 15% fetal bovine serum, were established and examined for their ability to dierentiate when placed in dierentiation medium (2% horse serum). Cells were photographed (406) following incubation in dierentiation medium for the indicated times Oncogene

A probable mechanism underlying the ability of Tbx3 to cooperate with Myc and Ras oncogenes to transform cells (Figure 1) and to inhibit myogenic dierentiation (Figure 4) is reduction of p53 levels caused by downregulation of p19ARF by Tbx3 (Figures 2a and 3c). Based on results obtained with Tbx2 (Jacobs et al., 2000) and more recently with Tbx3 (Brummelkamp et al., 2002), downregulation of p19ARF by these factors is direct. The downregulation of p19ARF by Tbx3 is predicted to decrease cellular levels of p53 by freeing the p19ARFinteracting protein Mdm2, leading to increased Mdm2-mediated turnover of p53 (Sherr and Weber, 2000; Sharpless and DePinho, 1999). Indeed, we ®nd


that p53 levels decrease in response to ectopic expression of Tbx3, and that levels of the p53 target gene products Mdm2 (functioning as part of a feedback mechanism regulating p53 abundance) and p21 also decrease (Figures 2a and 3c). The ability of Tbx3 to downregulate p19ARF, p53 and p53 target gene levels, is consistent with its ability to interfere with myogenic dierentiation, which has been shown to be dependent on p53 activity (Cerone et al., 2000). Similarly, the ability of Tbx3 to negatively regulate p53 levels is consistent with its capacity to participate in cell transformation. In contrast to Tbx3, both Myc and Ras stimulate increased p19ARF levels and typically make cells prone to apoptosis and cell cycle arrest respectively (Sherr and Weber, 2000; Sharpless and DePinho, 1999). A likely manifestation of the decreased levels of p53 caused by Tbx3 expression is partial suppression of apoptosis, as we show using primary MEFs overexpressing Myc (Figure 2b) and the HO ± Myc cell line (Figure 3a,b). HO ± Myc cells are derivatives of the immortal Rat1a cell line that express deregulated Myc on a myc-null background (Mateyak et al., 1999) and, like many cell types expressing elevated Myc levels, are highly sensitive to apoptosis when serum-deprived. Although Myc-dependent apoptosis was strongly suppressed by Tbx3 in primary MEFs (Figure 2b), suppression of apoptosis in HO-Myc cells was less impressive (Figure 3a,b). These results raise the possibility that primary cells may be more responsive to Tbx3 eects on apoptosis than immortal cell lines. Alternatively, the HO ± Myc assay system may provide a more accurate assessment of the protective eects of Tbx3 expression. Nevertheless, we speculate that even a partial disruption of p53-dependent apoptosis caused by downregulation of p19ARF in primary MEFs, is sucient for Tbx3 to tip the balance in favor of growth promotion and cell transformation by Myc. It is important to recognize that Myc-induced apoptosis is manifest much more readily in serum-limiting (survival factor-limiting) conditions and that the cell transformation experiments using Tbx3 in combination with Myc and Ras were performed in normal growth conditions (as these types of experiments are typically performed). Thus, partial disruption of p53mediated apoptosis by Tbx3, combined with the presence of survival factors above some threshold that buers cells from the apoptotic and cell cycle arrest activities of excessive Myc and Ras levels, likely work together to promote cell transformation by Myc, as well as Ras oncogenes. Indeed, the ability of Tbx3 to cooperate with Myc to transform cells, yet only partially block Myc-dependent apoptosis, suggests that additional activities of Tbx3, or additional activities associated with downregulation of p19ARF, promote cell transformation. The ability of Tbx3 to impinge on p53 pathway function, facilitate cell immortalization and transformation, and inhibit myogenic dierentiation, is consistent with the idea that Tbx3 may participate in tumorigenic progression in vivo. Interestingly, the

bHLH protein Twist has properties similar to those described here for Tbx3 and there is evidence that it functions as an oncogenic factor in rhabdomyosarcoma (Maestro et al., 1999). Jacobs et al. (2000), who made the original observation that the Tbx2 protein can immortalize MEFs and can repress p19ARF gene transcription, showed that the Tbx2 gene is ampli®ed and overexpressed in some breast cancers. Consistent with the notion that Tbx3 may be capable of promoting neoplastic growth, null and/or inactivating mutations in Tbx3 in humans leads to hypoproliferation of a number of dierent tissues, including breast tissue in ulnarmammary syndrome (Schinzel, 1987). Taken together with the Tbx2 results and work presented here, a more comprehensive examination of Tbx2 and Tbx3 involvement in tumorigenesis is warranted. Although the present observations point to a potential role for Tbx3 in tumorigenesis, our results may also shed some light on how mutations in Tbx3 lead to ulnar-mammary syndrome. In experiments using a mutant Tbx3 lacking its C-terminus, a mutant form of Tbx3 that is similar to those predicted to be produced by a number of Tbx3 mutations found in ulnar ± mammary syndrome (Bamshad et al., 1999; Carlson et al., 2001), we found that it failed to interfere with apoptosis (Figures 2b and 3a,b) and failed to block myogenic dierentiation (Figure 4). Thus, perhaps UMS mutations lead to hypoplasia as a consequence of failure to adequately protect against apoptotic death of key cell populations at critical times during embryonic development. Alternatively or coincidentally, mutant Tbx3 proteins and/or Tbx3 de®ciency, may lead to the premature or unscheduled dierentiation of key target tissues. However, even though hypoplasia is evident in many tissues of UMS patients, UMS is a highly complex syndrome that does not appear to be a disease of cell proliferation per se, but more a disease of defective epithelial/ mesenchymal interactions (Bamshad et al., 1997; Smith, 1999). Furthermore, regulation of p19ARF levels and p53 pathway activity and possibly pRb pathway activity (Figures 2a, 3c) is likely to be but one aspect of Tbx3 function and it will be important to determine other Tbx3 target genes and biological activities. However, it is notable that mutations in the p53 family member p63, cause EEC syndrome (ectrodactyly, ectodermal dysplasia, and cleft lip/ palate) (Celli et al., 1999; van Bokhoven et al., 2001), the related Hays ± Wells syndrome (McGrath et al., 2001), and are found in a subset of individuals with non-syndromic split hand ± split foot malformation (SHFM) (Ianakiev et al., 2000; van Bokhoven et al., 2001) and limb-mammary syndrome (van Bokhoven et al., 2001). All of these diseases share some overlapping features with UMS, supporting the possibility that disrupted regulation of p53 and perhaps p63 in some aected tissues may indeed be an important underlying mechanism in this syndrome.

3833

Oncogene


3834

Materials and methods Focus formation assays Primary mouse embryo ®broblasts (MEFs) were isolated from embryonic day 13.5 mouse (C57Bl/6) embryos as previously described (Spector et al., 1998) and grown in DMEM supplemented with 10% FBS. Analogous to experiments pioneered by Land et al. (1983), focus formation (transformation) assays were performed by transfecting MEFs at passage 1 or 2 with combinations of pBabe ± Myc (a kind gift from C Grandori) pEJ ± Hras and pCS2 ± Tbx3 plasmids (Carlson et al., 2001) using the calcium phosphate precipitation method (Chen and Okayama, 1987). Transfected MEFs were fed every 2 days with growth medium and stained with methylene blue after 2 ± 3 weeks. Focus formation experiments were performed in duplicate and repeated with dierent batches of MEFs at least two times with essentially identical results. For soft agar assays, instead of staining cells used in focus formation assays, they were pooled and plated in duplicate at 5000 cells/plate in soft agar as previously described (Hurlin et al., 1987). Soft agar assays were performed twice with reproducible results. Retroviral infection and apoptosis assays Retroviral infections of MEFs and HO-Myc (a kind gift of J Sedivy) cells were carried out for 24 h in DMEM/10% FBS using viral supernatants produced in transiently transfected Ecopak 293 cells (Clonetech). Infections were performed using pFB retrovirus (Stratagene), pFB ± Tbx3 ± Ha, or pFB ± Tbx3 ± 300 (Carlson et al., 2001), pBabe ± Myc and pBabe ± H-Ras (a kind gift from Scott Lowe). Virus titers were determined by infecting NIH3T3 cells with a pFB ± GFP (green ¯uorescent protein) retrovirus and positive (green ¯uorescent) cells counted. Typical infection eciencies were 50 ± 80%. For apoptosis assays, infected cells were plated at 50 000 cells per 30 mm dish 24 h after infection (or sequential infections) and serum levels were lowered 12 h later from 10% to 0.2% serum. Cells were photographed and collected daily, ®xed in 70% ethanol, stained with propidium iodide and sorted for DNA content by ¯uorescence-activated cell sorting (FACS). Assays were performed at least three times.

Myogenic differentiation assays and Northern blotting C2C12 cells (ATCC) were maintained in DMEM supplemented with 15% fetal calf serum (FCS) and grown in humidi®ed incubators at 378C and 10% CO2. To induce dierentiation, cells were grown to 90% con¯uence and growth medium switched to 2% horse serum (dierentiation medium) for 5 days. Cells were photographed daily. Northern blots were carried out using total RNA by standard methods and hybridized with radiolabelled probes for myoD, myogenin and myosin light chain (gifts from M Thayer) and GFAP. Western blot analysis For retrovirus-infected cells, cells were harvested 48 h after infection and lysed with NP-40 buer (150 mM NaCl, 20 mM Tris HCl pH 8, 1 mM EDTA, 1% NP-40, 10% glycerol, and 10 mg/ml Aprotinin, 1 mM PMSF, and 25 mg/ml leupeptin added fresh). Equal amounts of protein were resolved by SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene di¯uoride membranes (Millipore Corp., Bedford, MA, USA). Membranes were incubated in TTBS (10 mM Tris pH 7.6, 150 mM NaCl, 0.1% Tween-20) containing 5% non-fat dry milk overnight at 48C. Membranes were probed using the following antibodies: p19ARF (Abcam), p53, p21, p16, c-Myc, tubulin (Santa Cruz Biotechnology), Mdm2 (A Levine, Princeton), pRb, and phospho-speci®c Rb (New England Biolabs), Ha epitope (Covance). Antibody incubations and washes were performed according to conditions provided by the supplier.

Acknowledgments We would like to thank Scott Lowe, Arnold Levine, John Sedivy, Carla Grandori and Robert Eisenman for reagents. This work was supported by a grant from Shriners Hospitals for Children to PJ Hurlin.

References Agulnik SI, Garvey N, Hancock S, Ruvinsky I, Chapman DL, Agulnik I, Bollag R, Papaioannou V and Silver LM. (1996). Genetics, 144, 249 ± 254. Bamshad M, Lin RC, Law D, Watkins WC, Krakowiak PA, Moore ME, Franceschini P, Lala R, Holmes LB, Gebuhr TC, Bruneau BG, Schinzel A, Seidman JG, Seidman CE and Jorde LB. (1997). Nat. Genet., 16, 311 ± 315. Bamshad M, Le T, Watkins WS, Dixon ME, Kramer BE, Roeder AD, Carey JC, Root S, Schinzel A, Van Maldergem L, Gardner RJ, Lin RC, Seidman CE, Seidman JG, Wallerstein R, Moran E, Sutphen R, Campbell CE and Jorde LB. (1999). Am. J. Hum. Genet., 64, 1550 ± 1562. Basson CT, Bachinsky DR, Lin RC, Levi T, Elkins JA, Soults J, Grayzel D, Kroumpouzou E, Traill TA, LeblancStraceski J, Renault B, Kucherlapati R, Seidman JG and Seidman CE. (1997). Nat. Genet., 15, 30 ± 35. Braybrook C, Doudney K, Marcano AC, Arnason A, Bjornsson A, Patton MA, Goodfellow PJ, Moore GE and Stanier P. (2001). Nat. Genet., 29, 179 ± 183. Oncogene

Brummelkamp TR, Kortlever RM, Lingbeek M, Trettel F, MacDonald ME, van Lohuizen M and Bernards R. (2002). J. Biol. Chem., 277, 6567 ± 6572. Carlson H, Ota S, Campbell CE and Hurlin PJ. (2001). Hum. Mol. Genet., 10, 2403 ± 2413. Cerone MA, Marchetti A, Bossi G, Blandino G, Sacchi A and Soddu S. (2000). Cell Death Dier., 7, 506 ± 508. Celli J, Duijf P, Hamel BC, Bamshad M, Kramer B, Smits AP, Newbury-Ecob R, Hennekam RC, Van Buggenhout G, van Haeringen A, Woods CG, van Essen AJ, de Waal R, Vriend G, Haber DA, Yang A, McKeon F, Brunner HG, van Bokhoven H. (1999). Cell, 99, 143 ± 153. Chapman DL, Garvey N, Hancock S, Alexiou M, Agulnik SI, Gibson-Brown JJ, Cebra-Thomas J, Bollag RJ, Silver LM and Papaioannou VE. (1996). Dev. Dyn., 206, 379 ± 390. Chen CA and Okayama H. (1988). Biotechniques, 6, 632 ± 638. Eischen CM, Weber JD, Roussel MF, Sherr CJ and Cleveland JL. (1999). Genes Dev., 13, 2658 ± 2669.


Evan GI and Vousden KH. (2001). Nature, 411, 342 ± 348. Gibson-Brown JJ, I Agulnik S, Silver LM and Papaioannou VE. (1998). Mech. Dev., 74, 165 ± 169. He Ml, Wen L, Campbell CE, Wu JY and Rao Y. (1999). Proc. Natl. Acad. Sci. USA, 96, 10212 ± 10217. Herrmann BG, Labeit S, Poustka A, King TR and Lehrach H. (1990). Nature, 343, 617 ± 622. Hurlin PJ, Fry DG, Maher VM and McCormick JJ. (1987). Cancer Res., 47, 5752 ± 5757. Ianakiev P, Kilpatrick MW, Toudjarska I, Basel D, Beighton P and Tsipouras P. (2000). Am. J. Hum. Genet., 67, 59 ± 66. Jacobs JJ, Scheijen B, Voncken JW, Kieboom K, Berns A and van Lohuizen M. (1999). Genes Dev., 13, 2678 ± 2690. Jacobs JJ, Keblusek P, Robanus-Maandag E, Kristel P, Lingbeek M, Nederlof PM, van Welsem T, van de Vijver MJ, Koh EY, Daley GQ and van Lohuizen M. (2000). Nat. Genet., 26, 291 ± 299. Kamijo T, Zindy F, Roussel MF, Quelle DE, Downing JR, Ashmun RA, Grosveld G and Sherr CJ. (1997). Cell, 91, 649 ± 659. Land H, Parada LF and Weinberg RA. (1983). Nature, 304, 596 ± 602. Li QY, Newbury-Ecob RA, Terrett JA, Wilson DI, Curtis AR, Yi CH, Gebuhr T, Bullen PJ, Robson SC, Strachan T, Bonnet D, Lyonnet S, Young ID, Raeburn JA, Buckler AJ, Law DJ and Brook JD. (1997). Nat. Genet., 15, 1 ± 9. Maestro R, Dei Tos AP, Hamamori Y, Krasnokutsky S, Sartorelli V, Kedes L, Doglioni C, Beach DH and Hannon GJ. (1999). Genes Dev., 13, 2207 ± 2217. Mateyak MK, Obaya AJ and Sedivy JM. (1999). Mol. Cell. Biol., 19, 4672 ± 4683. McGrath JA, Duijf PH, Doetsch V, Irvine AD, de Waal R, Vanmolkot KR, Wessagowit V, Kelly A, Atherton D, Griths WA, Orlow SJ, van Haeringen A, Ausems MG, Yang A, McKeon F, Bamshad MA, Brunner HG, Hamel BC and van Bokhoven H. (2001). Hum. Mol. Genet., 10, 221 ± 229.

Palmero I, Pantoja C and Serrano M. (1998). Nature, 395, 125 ± 126. Serrano M, Lin AW, McCurrach ME, Beach D and Lowe SW. (1997). Cell, 88, 593 ± 602. Schinke M and Izumo S. (2001). Nat. Genet., 27, 238 ± 240. Schinzel A. (1987). J. Med. Genet., 24, 778 ± 781. Sharpless NE and DePinho RA. (1999). Curr. Opin. Genet. Dev., 9, 22 ± 30. Sherr CJ and Weber JD. (2000). Curr. Opin. Genet. Dev., 10, 949 ± 953. Smith J. (1999). Trends Genet., 15, 154 ± 158. Spector D, Goldman RD and Leinwand LA. (1998). Cells, a laboratory manual, Vol. 1. New York: Cold Spring Harbor Press. Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA, 96, 6937 ± 6941. van Bokhoven H, Hamel BC, Bamshad M, Sangiorgi E, Gurrieri F, Duijf PH, Vanmolkot KR, van Beusekom E, van Beersum SE, Celli J, Merkx GF, Tenconi R, Fryns JP, Verloes A, Newbury-Ecob RA, Raas-Rotschild A, Majewski F, Beemer FA, Janecke A, Chitayat D, Crisponi G, Kayserili H, Yates JR, Neri G and Brunner HG. (2001). Am. J. Hum. Genet., 69, 481 ± 492. Walsh K and Perlman H. (1997). Curr. Opin. Genet. Dev., 7, 597 ± 602. Welch PJ and Wang JY. (1993). Cell, 75, 779 ± 790. Welch PJ and Wang JY. (1995). Mol. Cell. Biol., 15, 5542 ± 5551. Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Sherr CJ and Roussel MF. (1998). Genes Dev.,12, 2424 ± 2433.

3835

Oncogene