Review

Introduction

Transforming growth factor- (TGF-) is a pluripotent cytokine that controls multiple cellular responses including the induction of cell growth inhibition, differentiation, cellular senescence, wound healing and apoptosis (Derynck et al., 2001; Shi and Massague, 2003; Siegel and Massague, 2003). These cellular responses are thought to define the role of TGF- as a tumor suppressor. Indeed, deregulated or aberrant TGF- signaling has been strongly implicated in the pathogenesis of human solid tumors, while less is known about the role of this pathway in leukemia and lymphoma pathogenesis (Derynck et al., 2001; Shi and Massague, 2003; Siegel and Massague, 2003).

The TGF- pathway has been well characterized and involves TGF- binding to the TGF- receptor type II (TRII), which results in the recruitment of TGF- receptor type I (TRI) and leads to assembly of the heterodimeric receptor complex. TRII then phosphorylates TRI in the 'GS sequence' located upstream from the kinase domain, and activates TRI kinase activity (Shi and Massague, 2003). The receptor-activated Smads (R-Smads), which are composed of Smad2 and Smad3, are then phosphorylated by TRI. This phosphorylation releases R-Smads from the receptor complex and enables R-Smads to form a heterodimeric complex with Smad4 (Co-Smad), which then translocates into the nucleus where the TGF- target genes are regulated (Shi and Massague, 2003). In the nucleus, Smad proteins bind to their cognate DNA binding sites with low affinity, which greatly increases as they interact with transcriptional coactivators or transcription repressors (Shi and Massague, 2003). Among TGF- target genes, p21 and p15 are the best characterized. These cyclin kinase inhibitors are upregulated upon TGF- stimulation, and mediate TGF- growth inhibition (Derynck et al., 2001; Shi and Massague, 2003; Siegel and Massague, 2003). In addition, TGF--induced growth suppression can also be achieved through downregulation of the c-myc oncoprotein. Downregulation of myc expression gives rise to a positive feedback loop which further amplifies p15 and p21 expression, as c-myc is known to be a negative regulator for TGF-induced p15 and p21 expression (Derynck et al., 2001; Shi and Massague, 2003; Siegel and Massague, 2003).

The inhibitory protein Smad7 negatively regulates TGF- signaling by recruiting the HECT-domain-containing E3 ubiquitin ligases Smurf1 or Smurf2 to the Smad7–TR complex, in turn facilitating TGF- receptor ubiquitination and subsequent degradation (Kavsak et al., 2000; Ebisawa et al., 2001). Another inhibitory Smad, termed Smad6, antagonizes TGF- signaling by preventing Smad2 interaction with Smad4 (Imamura et al., 1997). The transcriptional induction of Smad6 and Smad7 by TGF- provides a negative feedback loop to attenuate TGF- signaling.

Apart from its role as a tumor suppressor, TGF- can also promote tumor metastasis in the late stage of tumor development, suggesting that TGF- serves dual roles in both tumor suppression and cancer progression (Derynck et al., 2001; Siegel and Massague, 2003). How TGF- regulates cancer progression and metastasis is presently not well defined. Some key proteins involved in tumor invasion and metastasis have recently been identified as targets in the TGF- pathway, such as Snail, Slug and SIP1 (Romano and Runyan, 2000; Comijn et al., 2001; Martinez-Alvarez et al., 2004). Snail, Slug and SIP1 are positively regulated by TGF- and are known to mediate epithelial–mesenchymal transition (EMT), a process that can alter cell–cell adhesion, which is accompanied by the emergence of migratory and invasive properties that would therefore favor tumor invasion and metastasis (Romano and Runyan, 2000; Comijn et al., 2001; Martinez-Alvarez et al., 2004). Further investigation is warranted to determine the physiologic relevance these proteins have in TGF- mediated metastasis.

Early endosome pathway and the lipid-raft pathways in TGF- signaling

The internalization of membrane bound receptors occurs mostly through the lipid-raft and clathrin endocytic pathways (Le Roy and Wrana, 2005). It has been demonstrated that receptor endocytosis through the clathrin–endosome pathway mediates the turnover of membrane receptors, such as epidermal growth factor receptor (EGFR), leading to subsequent ablation of the receptor signal (Le Roy and Wrana, 2005). This endocytotic pathway causes internalization of membrane receptors into the early endosome, where they are sorted and subsequently transferred to the late endosome/lysosome for degradation. Alternatively, receptors that enter into the early endosome can also be recycled to the plasma membrane for maintenance of receptor availability for ligand binding and reactivation of the receptor signal transduction. In contrast, the lipid-raft pathway is thought to play a critical role in various physiologic processes, such as protein trafficking, cell polarization and signal transduction (Le Roy and Wrana, 2005). Numerous reports have demonstrated that the lipid-raft pathway is crucial for signal transduction in several pathways such as the T-cell receptor-dependent signaling cascade (Langlet et al., 2000), B cell receptor signaling (Langlet et al., 2000), H-ras-mediated Raf activation (Parton and Hancock, 2004), tumor necrosis factor (TNF) receptor-mediated nuclear factor (NF)-B activation (Legler et al., 2003), and interferon- (IFN-)-induced signal transducer and activator of transcription (STAT) signaling (Sehgal et al., 2002). Therefore, the lipid rafts represent critical sites for organizing and transmitting the extracellular signal into the cells.

TGF- receptor internalization is also mediated through both of these critical endocytic pathways (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005). In the steady state, TGF- receptors are partitioned into the early endosomes and lipid-raft compartments with relatively equal distribution (Di Guglielmo et al., 2003). In contrast to EGF-mediated EGFR internalization, TGF- does not seem to affect the partition of the receptors into these two compartments (Di Guglielmo et al., 2003). Although the clathrin–endosome pathway traditionally plays a role in membrane receptor regulation (i.e. EGF mediated EGFR internalization), this pathway plays a crucial role in the TGF- signaling cascade (Figure 1a). Intriguingly, instead of execution of receptor downregulation, the clathrin/early endosome pathway is critical for activation of TGF- signal cascades (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005). Studies have shown that K⁺ depletion and a dominant-negative mutant of dynamin, which interferes with clathrin-dependent trafficking (Hayes et al., 2002), inhibit TGF- receptor trafficking into the early endosome antigen-1 (EEA1)-containing early endosomes and also abrogate Smad2 phosphorylation and nuclear translocation. A further study by Di Guglielmo et al. (2003) showed that TGF- receptor type I/II (TRI/II) were found in EEA1-containing early endosomes and Rab11-positive recycling endosomes, but not in p62-positive late endosomes, suggesting that TRI/II localized in the early endosome is recycled back to the plasma membranes, but is not further sorted into the late endosme/lysosome for degradation. Taken together, these studies highlight the important role of the early endosome in TGF- signal transduction pathways.

Figure 1.

Regulation of TGF- signaling mediated by clathrin–early endosome- and lipid-raft-mediated endocytosis. (a) cPML is required for localization of TGF- receptors and SARA in the early endosome. cPML interacts with the TGF- receptor/SARA/Smad2/3 complexes and recruits them into the EEA1-enriched early endosome, leading to activation of TGF- signal transduction. (b) Promotion of TGF- receptor degradation by the Smad7/Smurf2 complex. TGF- receptors associate with the Smad7/Smurf2 complex and are recruited to the caveolin-1-enriched lipid rafts, leading to the ubiquitin-dependent degradation of the receptor

Full figure and legend (124K)

Smad-anchor for receptor activation (SARA) was originally identified as a Smad2 interaction protein, which can also associate with the TGF- receptor complex and present Smad2 to the receptor complexes for phosphorylation and activation (Tsukazaki et al., 1998). Recently, it became clear as to how SARA can regulate TGF- signaling. SARA, which contains a FYVE-domain that is responsible for the binding of SARA to phosphatidylinositol 3-phoshphate (PtdIns3P) in the membrane (Itoh et al., 2002), is highly enriched in EEA1-containing early endosomes. It was reported that SARA may promote TGF- receptor trafficking into the early endosome (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005), in turn leading to activation of Smad2 phosphorylation (Figure 1a). Another protein, hepatocyte-growth factor-regulated tyrosine-kinase substrate (HRS), which is localized in the early endosomes, can also positively regulate TGF- signaling presumably by cooperation with SARA in recruitment of TGF- receptor trafficking into the early endosome (Miura et al., 2000).

Besides trafficking into the early endosomes, TGF- receptor is also internalized into the lipid-raft compartments (Figure 1b). The lipid-raft pathway for TGF- receptor internalization negatively regulates TGF signaling by inducing TGF- receptor degradation (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005). Caveolin-1, which is highly enriched in the lipid rafts, has been shown to interact with TRI (Razani et al., 2001), and its overexpression leads to promotion of TGF- receptor degradation through the lipid-raft compartment (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005). Smad7 and Smurf2, which are also present in the caveolin-1-positive vesicles (Di Guglielmo et al., 2003; Le Roy and Wrana, 2005), may form a complex and mediate TGF- receptor ubiquitination in this compartment (Figure 1b), leading to its subsequent degradation by the proteasome-dependent pathway. However, the exact compartment in which the receptors are degraded has not yet been determined.

Role of cytoplasmic promyelocytic leukemia PML (cPML) in TGF- signaling

Acute promyelocytic leukemia (APL) is associated with reciprocal translocations, which always involve the RAR locus on chromosome 17, which variably translocates and fuses to the PML, PLZF, NPM, NuMA or STAT5B genes (Pandolfi, 2001; Piazza et al., 2001). PML gene is a member of a family of proteins, which share a novel zinc-finger binding motif termed RING and one or two additional Cys/His rich regions (B-Boxes) followed by a predicted coiled-coil domain (Zhong et al., 2000; Salomoni and Pandolfi, 2002). The RING and B-Boxes are located at the PML N-terminal end, followed by an -helical domain, and a coiled-coil region, at its C-terminal end, a serine/proline-rich region where PML is phosphorylated (Zhong et al., 2000; Salomoni and Pandolfi, 2002). The RING finger, B-boxes and coiled-coil domains are retained in all PML isoforms so far identified (Jensen et al., 2001). Although the PML RING domain can mediate protein–protein interactions, no clear biochemical function has yet been attributed to this moiety, which in other proteins is known to mediate ubiquitination (Pickart, 2001). Moreover, whether PML displays DNA-binding ability remains unclear. The PML coiled-coil region has been shown to be responsible for the formation of stable PML homodimers (Zhong et al., 2000; Salomoni and Pandolfi, 2002).

PML modulates the activity of key tumor suppressive pathways (e.g. p53 and Rb) (Zhong et al., 2000; Salomoni and Pandolfi, 2002), and has been demonstrated to regulate important tumor suppressive functions: (i) cell cycle control and growth inhibition; (ii) a proapoptotic role upon a number of stimuli; (iii) induction of premature cellular senescence upon oncogenic Ras (Salomoni and Pandolfi, 2002). Pml inactivation in mice results in susceptibility to cancer and accelerates leukemogenesis in PML-RAR transgenic models of APL (Wang et al., 1998; Rego et al., 2001).

PML is found in macromolecular structures associated with the nuclear matrix variably termed nuclear bodies (NBs) or PML oncogenic domains (PODs), where PML colocalizes with a number of nuclear proteins (Zhong et al., 2000; Salomoni and Pandolfi, 2002). The biochemical role of PML NBs remains unclear, although a number of NB proteins have been implicated in transcriptional regulation (Zhong et al., 2000). In the t(15;17) promyelocytic cell line NB4, as well as in promyelocytes derived from APL patients or mouse models of APL, the PML NBs are disrupted and aberrant structures incorporating PML, RAR and the PML/RAR fusion protein colocalize (Zhong et al., 2000; Salomoni and Pandolfi, 2002). Therefore, PML/RAR has the ability to direct PML, RAR and presumably other nuclear antigens into these aberrant structures, thereby diverting them from their physiologic processes.

A number of cytoplasmic PML isoforms have been identified over the last decade (Fagioli et al., 1992; Jensen et al., 2001); however, their roles in cellular functions seem to have been overlooked. One study showed that a PML mutant lacking the nuclear localization signal (NLS), rendering PML cytosolic, was profoundly impaired in its ability to suppress transformation of NIH3T3 induced by neu oncogene (Le et al., 1996). Another study concluded that nuclear localization of PML is required for its growth-suppressive function. In this study, a PML cytoplasmic isoform did not yield an appreciable effect on cell growth inhibition when compared with a PML nuclear isoform (Fagioli et al., 1998). Owing to lack of any known functions for the cytoplasmic isoforms by these studies, research on PML in this decade has focused exclusively on the nuclear isoforms.

Recently, a cytosolic function of PML was identified and this function involves the regulation of TGF- signaling (Le Roy and Wrana, 2004, 2005; Lin et al., 2004). In Pml^-/- primary cells, TGF- induced cell growth inhibition and cellular senescence are profoundly impaired (Lin et al., 2004). Strikingly, restoration of a cytoplasmic PML isoform (cPML) in these cells, but not a nuclear PML isoform, fully rescues these TGF- defects (Lin et al., 2004), suggesting that cPML is essential for TGF- functions. Further analysis of TGF- signaling revealed that Smad2/3 phosphorylation and nuclear translocation are also markedly reduced in Pml^-/- primary cells (Lin et al., 2004). cPML interacts and colocalizes with Smad2/3 and SARA and may serve as an adaptor for Smad2/3 and SARA association (Lin et al., 2004). As previously mentioned, TGF- receptor internalization through the clathrin/early endosome pathway is critical for transmitting TGF- signal transduction. Immunofluorescence studies and sucrose flotation assays revealed that cPML is localized in the early endosome and that TRI/II and SARA trafficking into the early endosome was remarkably reduced in Pml^-/- primary cells (Lin et al., 2004). These results point to the critical role of cPML in the regulation of TGF- signaling through affecting the internalization of the TRI/II/SARA/Smad2/3 complex into the early endosome. On the basis of these findings, we propose a model to explain how cPML may regulate TGF- signaling (Figure 1a): cPML may promote the trafficking of the TRI/II/SARA/Smad2/3 complex into the early endosome; alternatively, cPML may be essential to retain the TRI/II/SARA/Smad2/3 complex in the early endosome.

Role of TGF- signaling in the hematopoietic stem cells (HSCs)/progenitors

HSC and their progenitors orchestrate hematopoiesis, a process that involves self-renewal, proliferation and differentiation (Jordan and Guzman, 2004; Warner et al., 2004). Aberrant regulation of hematopoiesis can give rise to hematopoietic diseases and leukemia (Jordan and Guzman, 2004; Warner et al., 2004). HSCs possess two properties: a cell renewal function that can generate daughter cells with identical functions as the parent cells and the ability to differentiate into progenitor cells, which can further differentiate to more mature cells, thereby establishing a hierarchy of the hematopoietic system (Jordan and Guzman, 2004; Warner et al., 2004). While in vitro studies have shown that TGF- primarily exerts an inhibitory role during hematopoiesis, the in vivo role of TGF- in this process is still under debate (Fortunel et al., 2000). A myriad of studies using in vitro models for hematopoietic differentiation have shown that TGF- is able to suppress cell growth and colony formation of early progenitors (Fortunel et al., 2000). TGF- maintains HSC/progenitors in a quiescent state, whereas the lineage-restricted progenitors are generally less sensitive to its cytostatic effects (Fortunel et al., 2000). For example, it was observed that primitive CD34⁺CD38^- cells show a high response to TGF- growth inhibition, while more mature CD34⁺CD38⁺ cells are far less sensitive (Van Ranst et al., 1996). The cytostatic effects of TGF- on these cells may partly depend on its ability to upregulate cell cycle inhibitors, such as p21, p15, and p57, and to downregulate c-myc (Kim and Letterio, 2003; Scandura et al., 2004).

Goey et al. (1989) examined the in vivo role of TGF- in hematopoiesis by administering TGF- to mice and found that TGF- strongly inhibits proliferation of early progenitors. Taken together, these studies suggest that TGF- can act as a negative regulator of hematopoiesis. One potential problem with these studies is that the high doses of TGF- utilized may not occur in physiologic conditions, raising doubt about the role of TGF- in hematopoiesis. To circumvent this problem, a hematopoietic conditional TRI knockout mouse model was generated (Larsson et al., 2003). Consistent with the inhibitory role of TGF-, HSCs from TRI-null mice showed increased proliferation when cultured in vitro (Larsson et al., 2003). Surprisingly, TGF- deficiency did not affect the size of the progenitor cells and showed normal differentiation and maturation of hematopoietic cells. Furthermore, transplanted TGF- deficient HSCs have normal regenerative and self-renewal ability in vivo (Larsson et al., 2003), arguing that disruption of TGF- signaling in vivo is not sufficient to affect hematopoiesis. This experiment, however, does not rule out the possible involvement of TGF- signaling in hematopoiesis.

Role of TGF- signaling in leukemogenesis

Leukemia is generally viewed as a stem cell disease (Jordan and Guzman, 2004; Warner et al., 2004). Like other neoplasms, leukemia arises from the clonal expansion of a single cell. Leukemic stem cells (LSCs) that preserve the properties of HSC (namely cell renewal capacity and ability to differentiate into progenitors) can be isolated from leukemia patients (Jordan and Guzman, 2004; Warner et al., 2004). LSCs isolated from leukemia patients can be serially transplanted and engrafted into NOD/SCID mice. The resulting leukemic grafts were highly similar to the original disease, exhibiting identical blast morphology and dissemination profiles (Jordan and Guzman, 2004; Warner et al., 2004). Thus, an improved understanding of the distinct molecular features and differences between LSCs and HSCs may lead to the development of therapies that can target LSCs directly, and hopefully improve treatment for patients with leukemia.

Disruption of TGF- signaling, either by mutational inactivation of signaling components or by downregulation of their expression, is known to play an important role in cancer development (Massague et al., 2000; Derynck et al., 2001). Several TGF- signaling components are bona fide tumor suppressors with the ability to constrain cell growth and inhibit cancer development at its early stages, and inactivation of these components (e.g. TRII, TRI, Smad2 and Smad4) are associated with some solid cancers such as breast, pancreatic and colon (Massague et al., 2000; Derynck et al., 2001).

Although TGF- is proposed to act as a potent endogenous negative regulator of hematopoiesis (Fortunel et al., 2000, 2003), its role in leukemogenesis remains largely unknown until recently. Wolfraim et al. (2004) reported that Smad3 is an important tumor suppressor in T-cell lineage acute lymphoblastic leukemia (T-cell ALL). Smad3 protein level was not detected in leukemic cells from children with T-cell ALL, despite the fact that Smad3 messenger level was normal (Wolfraim et al., 2004). Using a Smad3 deficient mouse model, they further showed that the loss of Smad3 alone was insufficient to induce leukemia. However, when coupled with loss of the cyclin kinase inhibitor p27^kip1, leukemia developed in Smad3-deficient mice (Wolfraim et al., 2004). In acute myeloid leukemia (AML), there were two distinct mutations (a missense mutation and a frameshift mutation) in Smad4 that resulted in disruption of its ability to potentiate TGF- transcriptional activity (Imai et al., 2001). Furthermore, loss of TRI and TRII expression has been observed in patients with myeloid or lymphocytic leukemia (Le Bousse-Kerdiles et al., 1996; DeCoteau et al., 1997). The AML1-ETO and AML1-EVI-1 fusion proteins, which are involved in AML and chronic myeloid leukemia (CML), respectively, associate with Smad3, inhibiting TGF- functions (Jakubowiak et al., 2000; Mitani, 2004). These results clearly highlight the importance of TGF- signaling in leukemogenesis. However, it is not likely that disruption of TGF- signaling alone is sufficient for the development of leukemia. Support for this notion arises from the fact that the hematopoietic conditional TRI knockout mice did not exhibit any sign of leukemia (Larsson et al., 2003). Furthermore, Smad3-null mice, which exhibit profound defects in TGF- signaling, also do not develop leukemia throughout their lifespan (Datto et al., 1999; Yang et al., 1999; Larsson et al., 2003). Therefore, these studies suggest that disruption of TGF- signaling may not play an important role in leukemia initiation, but instead may be involved in the progression of leukemia.

Disruption of TGF- signaling in APL

In APL blasts where PML-RAR is overexpressed, TGF--induced cell differentiation is blocked. Remarkably, retinoid acid (RA) treatment, which induces PML-RAR degradation, resensitizes the cells to TGF-, suggesting that the presence of PML-RAR overexpression in APL blasts may cause TGF- resistance (Grignani et al., 1993; Testa et al., 1993, 1994). A recent study attempted to decipher the molecular mechanisms underlying this resistance and revealed a marked defect in TGF- signaling including Smad2/3 phosphorylation and nuclear translocation, which is similar to that in Pml^-/- primary cells (Lin et al., 2004). Of note, degradation of PML-RAR by RA fully restores these TGF- defects (Lin et al., 2004). These results indicate that PML-RAR overexpression in APL blasts ablates the TGF- signaling pathway. Given the fact that PML-RAR is present both in the cytosol and nucleus (Kastner et al., 1992; Lin and Pandolfi, unpublished observation), it is plausible that PML-RAR may inhibit TGF- signaling through direct inhibition of cPML to participate in TGF- signaling. This hypothesis can be supported by the observation that overexpression of PML-RAR abrogated the interaction between Smad3 and cPML (Lin et al., 2004). Furthermore, in APL blasts, the interaction of cPML with Smad2/3 is lost, and degradation of PML-RAR by RA restores this interaction (Lin et al., 2004).

On the basis of these results, we speculate that PML-RAR may disrupt TGF- signaling through two modes (Figure 2). In the cytosol, PML-RAR can sequester cPML away from the TRI/II/SARA/Smad2/3 complex, thus preventing this complex from trafficking into the early endosome. In the nucleus, PML-RAR may also prevent nuclear PML from cooperating with Smad3/4 to activate TGF- target genes, given the fact that the nuclear PML isoform PML4 can interact and cooperate with Smad3 to transactivate TGF- transcriptional activity (Lin and Pandolfi, unpublished observation).

Figure 2.

Suppression of TGF- signaling by PML-RAR. PML-RAR is localized in the cytoplasm and nucleus. (a) In the cytoplasm, PML-RAR can sequester cPML away from the TGF- receptor/SARA/Smad2/3 complexes, which may in turn prevent the internalization of the TGF- receptor/SARA/Smad2/3 complexes into the early endosome, thereby blocking TGF- signaling. (b) In the nucleus, PML-RAR may interrupt the interaction of nuclear PML (nPML) with Smad2/3/Smad4 complex, leading to blockade of TGF- transcriptional activity

Full figure and legend (85K)

Other signaling pathways involved in TGF- resistance in leukemia

Previous studies have found that numerous TGF- negative regulators are overexpressed in some subsets of leukemia patients. The best examples are the c-Ski oncoproteins, which have been shown to transform bone marrow cells in vitro and are overexpressed in leukemic patients (Dahl et al., 1998; Kronenwett et al., 2005). c-Ski oncoproteins negatively regulate TGF- signaling through their ability to interfere with the formation of the R-Smad and Smad4 complex and by recruiting transcription corepressors (N-CoR) and histone deacetylase (HDAC) (Liu et al., 2001). The EVI-1 oncoprotein, which is highly upregulated in myeloid leukemia, was also shown to suppress TGF- signaling by interacting with Smad3 and interfering with its function (Kurokawa et al., 1998; Mitani, 2004). Furthermore, the AML1/ETO and AML1/EVI-1 fusion proteins, which are involved in AML and CML, respectively, associate with Smad3, resulting in blockade of TGF- functions (Jakubowiak et al., 2000; Mitani, 2004). In addition, the Tax protein, encoded by human T-cell lymphotropic virus type 1 (HTLV-1) and thought to be involved in leukemogenesis in adult T-cell leukemia, is a potent repressor of Smad-mediated TGF- transcriptional activation by interaction with CBP/p300 (Mori et al., 2001; Lee et al., 2002). The cell cycle inhibitor p15, which mediates cytostatic effects of TGF-, is frequently silenced due to promoter hypermethylation in AML (Krug et al., 2002). Other signaling pathways that induce Smad6/7 expression via Stat1, RAS-MAP kinase, NF-B may potentially antagonize the TGF- signaling during the leukemogenesis. Figure 3 summarizes the potential signals impinging on the cells to cause TGF- resistance.

Figure 3.

Mechanisms of TGF- resistance underlying leukemia pathogenesis. Growth factors and cytokines activate various signaling pathways, which then induce Smad6/7 expression, leading to suppression of TGF- signaling. c-Ski oncoproteins negatively regulate TGF- signaling through their ability to interfere with the formation of R-Smad and Smad4 complex and by recruiting the transcription corepressors (N-CoR) and histone deacetylase (HDAC). EVI-1, AML1/EVI-1 and AML1/ETO fusion proteins associate with Smad3 and thereby block TGF- functions. The Tax protein, encoded by HTLV-1, is a potent repressor of Smad-mediated TGF- transcriptional activation by interaction with CBP/p300

Full figure and legend (74K)

Conclusion and future directions

PML and TGF- are tumor suppressors that exert similar biological functions including cell growth inhibition, induction of cellular senescence and apoptosis. Loss of PML expression is frequently detected in a variety of human cancers such as colon, prostate and breast (Gurrieri et al., 2004), where TGF- signaling is also frequently impaired (Massague et al., 2000; Derynck et al., 2001). Recently, we have identified cPML as a critical regulator of TGF- signaling through regulation of TRI/II and SARA accumulation in the early endosome, a crucial step for TGF- signal transduction (Lin et al., 2004). Future studies are needed to determine the detailed mechanism as to how cPML could regulate this process. For instance, as numerous studies have proposed that ubiquitin modification of proteins is critical for endocytosis of numerous membrane receptors such as EGFR (Le Roy and Wrana, 2005), it will be of great interest to test in the future whether ubiquitination of TGF- receptor provides a signal for triggering its internalization and whether or not cPML may affect the levels of receptor ubiquitination.

As TGF- signaling is impaired in APL (Lin et al., 2004), it will be of great interest to evaluate whether loss of TGF- signaling indeed contributes to APL development. Intercrosses of TRI null mice with transgenic mouse models of APL (e.g. PML-RAR transgenic mice) will address this question. Nevertheless, since loss of TGF- signaling is not sufficient to cause leukemia, it is apparent that initiating oncogenic molecular events are needed to trigger leukemogenesis. In their presence, however, deregulation of TGF- signaling could have devastating consequences. In this respect, it is important to underline that several of the fusion oncoproteins associated with the pathogenesis of hemopoietic malignancies share the ability to interfere simultaneously with multiple molecular pathways towards oncogenic transformation.

Finally, as aforementioned, TGF- signaling serves dual roles not only in tumor suppression but also in promotion of tumor progression and metastasis. Interestingly, in leukemia patients, TGF- expression is often highly elevated (Raza et al., 1992; Mori et al., 2000). Given the fact that the TGF- cytostatic effect in leukemogenesis may be impaired due to diverse mechanisms summarized in Figure 3, upregulation of TGF- may instead promote leukemia progression and invasion. This question can be addressed in established leukemia mouse models by blocking TGF- function using a specific TGF- inhibitor or inducing dominant-negative TRII expression through an inducible promoter (i.e. Tet-on system) in the compound mutant mice, and evaluating the effect of TGF- inhibition on leukemia progression and invasion. Answering these fundamental questions may shed new light, and allow further understanding on the physiologic role of TGF- signaling and its role in leukemogenesis, ultimately providing novel therapeutic modalities for the treatment of leukemia.

References

1. Comijn J, Berx G, Vermassen P, Verschueren K, van Grunsven L, Bruyneel E, Mareel M, Huylebroeck D & van Roy F. (2001) Mol. Cell 7: 1267−1278. | Article | PubMed | ISI | ChemPort |
2. Dahl R, Kieslinger M, Beug H & Hayman MJ. (1998) Proc. Natl. Acad. Sci. USA 95: 11187−11192. | Article | PubMed | ChemPort |
3. Datto MB, Frederick JP, Pan L, Borton AJ, Zhuang Y & Wang XF. (1999) Mol. Cell. Biol. 19: 2495−2504. | PubMed | ISI | ChemPort |
4. DeCoteau JF, Knaus PI, Yankelev H, Reis MD, Lowsky R, Lodish HF & Kadin ME. (1997) Proc. Natl. Acad. Sci. USA 94: 5877−5881. | Article | PubMed | ChemPort |
5. Derynck R, Akhurst RJ & Balmain A. (2001) Nat. Genet. 29: 117−129. | Article | PubMed | ISI | ChemPort |
6. Di Guglielmo GM, Le Roy C, Goodfellow AF & Wrana JL. (2003) Nat. Cell Biol. 5: 410−421. | Article | PubMed | ISI | ChemPort |
7. Ebisawa T, Fukuchi M, Murakami G, Chiba T, Tanaka K, Imamura T & Miyazono K. (2001) J. Biol. Chem. 276: 12477−12480. | Article | PubMed | ISI | ChemPort |
8. Fagioli M, Alcalay M, Pandolfi PP, Venturini L, Mencarelli A, Simeone A, Acampora D, Grignani F & Pelicci PG. (1992) Oncogene 7: 1083−1091. | PubMed | ISI | ChemPort |
9. Fagioli M, Alcalay M, Tomassoni L, Ferrucci PF, Mencarelli A, Riganelli D, Grignani F, Pozzan T, Nicoletti I & Pelicci PG. (1998) Oncogene 16: 2905−2913. | Article | PubMed | ChemPort |
10. Fortunel NO, Hatzfeld A & Hatzfeld JA. (2000) Blood 96: 2022−2036. | PubMed | ISI | ChemPort |
11. Fortunel NO, Hatzfeld JA, Monier MN & Hatzfeld A. (2003) Oncol. Res. 13: 445−453. | PubMed |
12. Goey H, Keller JR, Back T, Longo DL, Ruscetti FW & Wiltrout RH. (1989) J. Immunol. 143: 877−880. | PubMed | ChemPort |
13. Grignani F, Ferrucci PF, Testa U, Talamo G, Fagioli M, Alcalay M, Mencarelli A, Peschle C, Nicoletti I & Pelicci PG. (1993) Cell 74: 423−431. | Article | PubMed | ISI | ChemPort |
14. Gurrieri C, Capodieci P, Bernardi R, Scaglioni PP, Nafa K, Rush LJ, Verbel DA, Cordon-Cardo C & Pandolfi PP. (2004) J. Natl. Cancer Inst. 96: 269−279. | Article | PubMed | ChemPort |
15. Hayes S, Chawla A & Corvera S. (2002) J. Cell Biol. 158: 1239−1249. | Article | PubMed | ISI | ChemPort |
16. Imai Y, Kurokawa M, Izutsu K, Hangaishi A, Maki K, Ogawa S, Chiba S, Mitani K & Hirai H. (2001) Oncogene 20: 88−96. | Article | PubMed | ChemPort |
17. Imamura T, Takase M, Nishihara A, Oeda E, Hanai J, Kawabata M & Miyazono K. (1997) Nature 389: 622−626. | Article | PubMed | ISI | ChemPort |
18. Itoh F, Divecha N, Brocks L, Oomen L, Janssen H, Calafat J, Itoh S & Dijke P, Pt. (2002) Genes Cells 7: 321−331. | Article | PubMed | ISI | ChemPort |
19. Jakubowiak A, Pouponnot C, Berguido F, Frank R, Mao S, Massague J & Nimer SD. (2000) J. Biol. Chem. 275: 40282−40287. | Article | PubMed | ChemPort |
20. Jensen K, Shiels C & Freemont PS. (2001) Oncogene 20: 7223−7233. | Article | PubMed | ChemPort |
21. Jordan CT & Guzman ML. (2004) Oncogene 23: 7178−7187. | Article | PubMed | ChemPort |
22. Kastner P, Perez A, Lutz Y, Rochette-Egly C, Gaub MP, Durand B, Lanotte M, Berger R & Chambon P. (1992) EMBO J. 11: 629−642. | PubMed | ISI | ChemPort |
23. Kavsak P, Rasmussen RK, Causing CG, Bonni S, Zhu H, Thomsen GH & Wrana JL. (2000) Mol. Cell 6: 1365−1375. | Article | PubMed | ISI | ChemPort |
24. Kim SJ & Letterio J. (2003) Leukemia 17: 1731−1737. | Article | PubMed | ChemPort |
25. Kronenwett R, Butterweck U, Steidl U, Kliszewski S, Neumann F, Bork S, Blanco ED, Roes N, Graf T, Brors B, Eils R, Maercker C, Kobbe G, Gattermann N & Haas R. (2005) Oncogene in press. | Article | PubMed | ChemPort |
26. Krug U, Ganser A & Koeffler HP. (2002) Oncogene 21: 3475−3495. | Article | PubMed | ChemPort |
27. Kurokawa M, Mitani K, Irie K, Matsuyama T, Takahashi T, Chiba S, Yazaki Y, Matsumoto K & Hirai H. (1998) Nature 394: 92−96. | Article | PubMed | ISI | ChemPort |
28. Langlet C, Bernard AM, Drevot P & He HT. (2000) Curr. Opin. Immunol. 12: 250−255. | Article | PubMed | ISI | ChemPort |
29. Larsson J, Blank U, Helgadottir H, Bjornsson JM, Ehinger M, Goumans MJ, Fan X, Leveen P & Karlsson S. (2003) Blood 102: 3129−3135. | Article | PubMed | ISI | ChemPort |
30. Le Bousse-Kerdiles MC, Chevillard S, Charpentier A, Romquin N, Clay D, Smadja-Joffe F, Praloran V, Dupriez B, Demory JL, Jasmin C & Martyre MC. (1996) Blood 88: 4534−4546. | PubMed | ChemPort |
31. Le Roy C & Wrana JL. (2004) Nature 431: 142. | Article | PubMed | ChemPort |
32. Le Roy C & Wrana JL. (2005) Nat. Rev. Mol. Cell. Biol. 6: 112−126. | Article | PubMed | ChemPort |
33. Le XF, Yang P & Chang KS. (1996) J. Biol. Chem. 271: 130−135. | Article | PubMed | ISI | ChemPort |
34. Lee DK, Kim BC, Brady JN, Jeang KT & Kim SJ. (2002) J. Biol. Chem. 277: 33766−33775. | Article | PubMed | ChemPort |
35. Legler DF, Micheau O, Doucey MA, Tschopp J & Bron C. (2003) Immunity 18: 655−664. | Article | PubMed | ISI | ChemPort |
36. Lin HK, Bergmann S & Pandolfi PP. (2004) Nature 431: 205−211. | Article | PubMed | ISI | ChemPort |
37. Liu X, Sun Y, Weinberg RA & Lodish HF. (2001) Cytokine Growth Factor Rev. 12: 1−8. | Article | PubMed | ChemPort |
38. Martinez-Alvarez C, Blanco MJ, Perez R, Rabadan MA, Aparicio M, Resel E, Martinez T & Nieto MA. (2004) Dev. Biol. 265: 207−218. | PubMed | ChemPort |
39. Massague J, Blain SW & Lo RS. (2000) Cell 103: 295−309. | Article | PubMed | ISI | ChemPort |
40. Mitani K. (2004) Oncogene 23: 4263−4269. | Article | PubMed | ChemPort |
41. Miura S, Takeshita T, Asao H, Kimura Y, Murata K, Sasaki Y, Hanai JI, Beppu H, Tsukazaki T, Wrana JL, Miyazono K & Sugamura K. (2000) Mol. Cell. Biol. 20: 9346−9355. | Article | PubMed | ISI | ChemPort |
42. Mori A, Wada H, Okada M, Takatsuka H, Tamura A, Fujimori Y, Okamoto T, Takemoto Y, Kanamaru A & Kakishita E. (2000) Acta Haematol. 103: 220−223. | Article | PubMed | ChemPort |
43. Mori N, Morishita M, Tsukazaki T, Giam CZ, Kumatori A, Tanaka Y & Yamamoto N. (2001) Blood 97: 2137−2144. | Article | PubMed | ISI | ChemPort |
44. Pandolfi PP. (2001) Hum. Mol. Genet. 10: 769−775. | Article | PubMed | ISI | ChemPort |
45. Parton RG & Hancock JF. (2004) Trends Cell Biol. 14: 141−147. | Article | PubMed | ChemPort |
46. Piazza F, Gurrieri C & Pandolfi PP. (2001) Oncogene 20: 7216−7222. | Article | PubMed | ChemPort |
47. Pickart CM. (2001) Annu. Rev. Biochem. 70: 503−533. | Article | PubMed | ISI | ChemPort |
48. Raza A, Yousuf N, Abbas A, Umerani A, Mehdi A, Bokhari SA, Sheikh Y, Qadir K, Freeman J, Masterson M, Miller MA, Lampkin B, Browman G, Bennett J, Goldberg J, Grunwald H, Larson R, Vogler R & Preisler H. (1992) Blood 79: 1037−1048. | PubMed | ChemPort |
49. Razani B, Zhang XL, Bitzer M, von Gersdorff G, Bottinger EP & Lisanti MP. (2001) J. Biol. Chem. 276: 6727−6738. | Article | PubMed | ISI | ChemPort |
50. Rego EM, Wang ZG, Peruzzi D, He LZ, Cordon-Cardo C & Pandolfi PP. (2001) J. Exp. Med. 193: 521−529. | Article | PubMed | ISI | ChemPort |
51. Romano LA & Runyan RB. (2000) Dev. Biol. 223: 91−102. | Article | PubMed | ISI | ChemPort |
52. Salomoni P & Pandolfi PP. (2002) Cell. 108: 165−170. | Article | PubMed | ISI | ChemPort |
53. Scandura JM, Boccuni P, Massague J & Nimer SD. (2004) Proc. Natl. Acad. Sci. USA 101: 15231−15236. | Article | PubMed | ChemPort |
54. Sehgal PB, Guo GG, Shah M, Kumar V & Patel K. (2002) J. Biol. Chem. 277: 12067−12074. | Article | PubMed | ChemPort |
55. Shi Y & Massague J. (2003) Cell 113: 685−700. | Article | PubMed | ISI | ChemPort |
56. Siegel PM & Massague J. (2003) Nat. Rev. Cancer 3: 807−820. | Article | PubMed | ISI | ChemPort |
57. Testa U, Grignani F, Barberi T, Fagioli M, Masciulli R, Ferrucci PF, Seripa D, Camagna A, Alcalay M & Pelicci PG. (1994) Cancer Res. 54: 4508−4515. | PubMed | ISI | ChemPort |
58. Testa U, Masciulli R, Tritarelli E, Pustorino R, Mariani G, Martucci R, Barberi T, Camagna A, Valtieri M & Peschle C. (1993) J. Immunol. 150: 2418−2430. | PubMed | ISI | ChemPort |
59. Tsukazaki T, Chiang TA, Davison AF, Attisano L & Wrana JL. (1998) Cell 95: 779−791. | Article | PubMed | ISI | ChemPort |
60. Van Ranst PC, Snoeck HW, Lardon F, Lenjou M, Nijs G, Weekx SF, Rodrigus I, Berneman ZN & Van Bockstaele DR. (1996) Exp. Hematol. 24: 1509−1515. | PubMed | ChemPort |
61. Wang ZG, Delva L, Gaboli M, Rivi R, Giorgio M, Cordon-Cardo C, Grosveld F & Pandolfi PP. (1998) Science 279: 1547−1551. | Article | PubMed | ISI | ChemPort |
62. Warner JK, Wang JC, Hope KJ, Jin L & Dick JE. (2004) Oncogene 23: 7164−7177. | Article | PubMed | ChemPort |
63. Wolfraim LA, Fernandez TM, Mamura M, Fuller WL, Kumar R, Cole DE, Byfield S, Felici A, Flanders KC, Walz TM, Roberts AB, Aplan PD, Balis FM & Letterio JJ. (2004) N. Engl. J. Med. 351: 552−559. | Article | PubMed | ChemPort |
64. Yang X, Letterio JJ, Lechleider RJ, Chen L, Hayman R, Gu&