| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, North Carolina 27710
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
Deregulation of TGF-ß signaling is implicated in the pathogenesis of many diseases including arthritis, atherosclerosis, glomerulonephritis, human hereditary telangiectasia, and carcinogenesis. Loss of cellular sensitivity to TGF-ß-mediated growth inhibition may contribute directly to these pathological states, specifically carcinogenesis.
With the demonstrated importance of TGF-ß signaling in a variety of biological processes, and loss of TGF-ß responsiveness as an important correlate of certain diseases, a tremendous effort has been undertaken in the last decade to elucidate the mechanisms by which TGF-ß propagates its signal. An important step in understanding TGF-ß signaling came with the identification of three cell surface proteins that bind to TGF-ß ligand with high affinity. These were called type I, II, and III receptors based on their molecular weight. The type I and type II receptors belong to a large family of receptor serine/threonine kinases. Upon TGF-ß ligand binding to type II receptor, type I receptor is recruited into a complex containing both receptors and ligand. This causes the phosphorylation and subsequent kinase activation of type I receptor by the constitutively active type II receptor kinase. Currently, activated type I receptor kinase is thought to be sufficient to modulate most TGF-ß downstream signals, although it is possible that type II receptor may also contribute to downstream signaling. The type III receptor is not essential for signal transduction, but may serve to present TGF-ß ligand to the type I and II receptors. In cells with lower signaling receptor affinities for a particular TGF-ß ligand, the presence of a large amount of type III receptor on the cell surface may serve to promote a productive TGF-ß signal.
The purpose of this review is to focus on the current body of knowledge of downstream signaling events resulting from TGF-ß treatment. Recently, relevant substrates and effectors of the TGF-ß receptor kinases have been identified at a rapid pace and will be the subject of this review. The TGF-ß ligands and receptors responsible for propagating this signal will only be reviewed briefly, as an in-depth review of the ligands and receptors can be found elsewhere in the literature (1, 2, 3, 4). We will focus our discussion on one family of TGF-ß receptor substrates, the Smads, and discuss the recent advances in our understanding of Smad signaling (also reviewed in Refs. 7, 8, 9, 10). Due to the importance of TGF-ß-mediated growth inhibition on a wide variety of cell types, we will discuss in detail the molecular mechanisms by which TGF-ß can potentiate growth inhibition. Finally, we will discuss recent members of the TGF-ß signaling pathway that have been discovered to be mutated in human cancers, in particular Smad-4. Throughout the discussion, we will attempt to address current questions and present challenges that await researchers in the field.
|
II. Ligands and Receptors |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
Five TGF-ß ligands have been cloned (TGF-ß1–5). These ligands are secreted as 100-kDa inactive complexes. The inactive form consists of a dimer of the N-terminal peptide, noncovalently associated with the 25-kDa dimer of the biologically active form. Structurally, the ligands exhibit a high degree of amino acid identity (64–82%) with nine invariant cysteine residues. The structures of TGF-ß1 and ß2 have been solved (11, 12, 13). Eight of the nine cysteines make four intramolecular disulfide bonds, while the ninth cysteine (amino acid 77) forms an intermolecular disulfide bond with the corresponding ninth cysteine of the other monomer. The distinguishing feature of the TGF-ß structure is the ’cysteine knot, ’ formed from three of the four intramolecular disulfide bonds that maintains structural integrity for the monomer. Aside from the nine invariant cysteines, superfamily members share less than 40% amino acid identity. This may account for the diversity of responses elicited from different superfamily members.
B. Receptors
Like other members of its superfamily, TGF-ß ligands signal by
binding to specific receptors on the cell surface. A breakthrough in
the field came with the identification (reviewed in Refs. 10, 14) and cloning of these receptors (16, 17, 18, 19, 20, 21, 22). Although four receptors
have been cloned (type I, II, III, endoglin), only two of them, the
type II and type I receptors, have been conclusively proven to mediate
TGF-ß signaling (23).
The type II receptor is a 75- to 85-kDa glycoprotein composed of 565 amino acids. It contains a signal sequence, a 136-residue hydrophilic extracellular domain, a single transmembrane domain, and a large intracellular domain of 376 amino acids. Its cytoplasmic region contains 18 of 21 conserved residues found in serine-threonine kinases. A fusion protein comprised of the cytoplasmic region linked to glutathione-S-transferase was able to autophosphorylate its serine residues, with some threonine phosphorylation.
The type I receptor is a 50- to 60-kDa protein composed of 501 amino acids. It contains a 101-amino acid hydrophilic extracellular domain, a single 23-amino acid transmembrane domain, and a 355-amino acid intracellular domain. Like the type II receptor, the type I receptor contains conserved regions resembling serine-threonine kinases. The cytoplasmic region fused to glutathione-S-transferase was able to autophosphorylate its threonine residues, with some serine phosphorylation. Unlike the type II receptor, the type I receptor does not bind independently to ligand. The cloning and identification of both mammalian type II and type I TGF-ß receptors are discussed in detail in other reviews (4, 10).
C. Receptor kinase signaling
The type II receptor is a constitutive receptor kinase that
associates with type I receptor on binding to TGF-ß. Upon
association, type I receptor kinase is phosphorylated by the
constitutively active type II receptor in a glycine-serine-rich region
known as the GS domain (24). Phosphorylation of the GS domain
subsequently activates the type I receptor kinase. A constitutively
active type I receptor has been created by mutating residues adjacent
to the site of phosphorylation. This receptor kinase can mediate the
TGF-ß growth-inhibitory effect and induce certain TGF-ß-responsive
genes, suggesting that type I receptor is a bona fide mediator of
TGF-ß signaling. At present it is not known whether type II has
physiological substrates other than type I receptor with which it
interacts.
Other TGF-ß superfamily members also signal through a type II and type I receptor kinase cascade. Previously, we and others had proposed a nomenclature whereby type II and type I receptors are grouped according to both the functional and structural characteristics they exhibit rather than by size alone, as the TGF-ß receptors were first categorized (4, 7). Structurally, type II receptors have longer extracellular domains and longer serine-threonine cytoplasmic tails than type I receptors. In addition, each receptor type contains certain conserved regions that the other does not possess. One such region is the GS domain. We further proposed that receptors containing GS domains be categorized as type I receptors whereas receptors that either bind ligand independently or lack a GS domain be categorized as type II receptors.
|
III. Smads |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
. In contrast, the most recently
identified members of the Smad family, Smad-6 and Smad-7, negatively
regulate receptor action. In this section, we will give a comprehensive
review of the role of Smads in TGF-ß signaling. First, we will
describe the cloning of the Smads from their homologs in
Drosophila and Caenorhabditis elegans. Second, we
will review findings that suggest a model for Smad activation upon
TGF-ß treatment. Possible nuclear functions for Smads as both a
transcription factor and coactivator will be presented. Third, we will
focus on the structure and possible function of the different Smad
domains. Questions that need to be addressed to further elucidate the
role of Smads in TGF-ß signal transduction will be discussed
throughout the section.
|
A. Cloning of the Smads
Smad homologs were first identified using a genetic approach in
Drosophila and C. elegans, two organisms in which
a TGF-ß-like signaling pathway is present. In Drosophila,
signaling proteins exist that are homologous to their mammalian TGF-ß
superfamily counterparts. ’Decapentaplegic’ (dpp) is the TGF-ß-like
ligand (5). The receptor homologs are ’punt’ (type II), ’thick
veins,’ and ’saxophone’ (both type I). In a genetic screen to
determine dominant enhancers of a weak dpp allele, ’mothers against
dpp’ (MAD) was isolated (25). Loss-of-function mutations in MAD result
in organisms that phenotypically resemble those with null alleles of
dpp, and a constitutively active thick veins phenotype is repressed by
null alleles of MAD, suggesting that MAD is in the TGF-ß-like
signaling pathway.
In C. elegans, the story is remarkably similar. The receptor homologs are daf-4 (type II) and daf-1 (type I). A mutant daf-4 gives rise to small worms and fused male tail rays. SMA-2, SMA-3, and SMA-4 were identified as genes that are downstream of daf-4. Individual null sma alleles result in an organism that phenotypically resembles that of the mutant daf-4. The fact that other SMA genes cannot compensate for inactivation of a single SMA gene suggests that each SMA allele may play a functionally distinct role in the pathway. These genes are homologs of MAD, further suggesting the importance of these genes in TGF-ß signaling.
With conservation of these proteins in both Drosophila and C. elegans, research turned toward identifying homologs in higher organisms. Using Drosophila MAD or C. elegans SMA genes, libraries from organisms such as Xenopus, mice, and humans were screened for homologous genes (26, 27, 28, 29, 30, 31, 32, 33). A region on chromosome 18q that is homozygously deleted in human pancreatic carcinomas was found to contain a gene called Deleted in Pancreatic Cancer-4 (DPC-4), identified as a MAD homolog (34). These homologs of the SMA and MAD genes have subsequently been called Smads.
B. Smad regulation
Since the cloning of the Smads, a working model has emerged as to
how Smads are regulated. Upon direct phosphorylation by type I TGF-ß
receptor, Smad-2 or Smad-3 binds to its Smad-4 partner to form a
heteromeric complex and translocates into the nucleus. Once in the
nucleus, the Smads regulate the transcription of genes important in the
TGF-ß response. One such gene may be the plasminogen activator
inhibitor-1 (PAI-1). This working model is the result of several
different lines of investigation that will be reviewed in this section.
Initial work by several groups showed that certain endogenous Smads are
phosphorylated upon treatment with TGF-ß or BMP (28, 29, 35).
Specifically, Smad-1 and Smad-5 are phosphorylated upon BMP treatment.
Smad-2 and Smad-3 are phosphorylated upon TGF-ß or activin treatment
(Fig. 1
). This transient phosphorylation is induced within 15 min and
peaks at about 1 h before returning to basal levels by 2 h.
Based on this initial observation, work focused on determining the
kinase responsible for phosphorylating the Smads.
The first clue on the possible candidate kinase that could phosphorylate Smads after TGF-ß addition came from a study in which Smad-3, but not Smad-4, was found to be phosphorylated and associated with the ligand-bound receptor kinase, suggesting that Smad-3 may serve as a direct substrate for the receptor kinase (31). Subsequently, overexpression of the type I receptor in COS cells led to phosphorylation of overexpressed Smad-2 (MADR2) in the presence of type II receptor. Furthermore, upon mutation of three serine residues in the C terminus of Smad-2, the phosphorylation of Smad-2 by type I receptor kinase was inhibited. Using two-dimensional tryptic peptide mapping, it was further shown that these three serines are indeed phosphorylated by type I receptor (32). The mutant Smad-2 could be coimmunoprecipitated with type I receptor, suggesting that Smad-2 is an in vivo substrate for the type I receptor kinase.
Phosphorylation of the C terminus of Smad-2 is necessary for its nuclear accumulation (32), but the exact mechanism by which nuclear localization occurs upon Smad-2 phosphorylation is still unclear. Phosphorylation may change the conformation of Smad-2 or Smad-3 so that it can form a heteromeric complex with Smad-4 and/or translocate to the nucleus and thereby exert its nuclear function.
C. Smad nuclear function
1. Smads as coactivators. A breakthrough in our understanding
of Smad nuclear function came with the discovery that Smad-2 was in a
complex with the transcription factor forkhead activin-induced signal
transducer (FAST-1) (36). Upon activin, TGF-ß, or Vg-1 treatment, a
rapidly induced complex forms on the activin response element (ARE) of
the Mix.2 promoter. This novel transcription factor, FAST-1, is the
major DNA-binding component of the activin response factor (ARF).
Smad-2 is found only in activin-induced ARF complexes, suggesting that
Smad-2 plays a role as a mediator of the activin signal. Overexpression
of the MH-2 domain or Smad-2 was sufficient to activate transcription
of the Mix.2 promoter, whereas FAST-1 overexpression could not cause an
increase in transcription, suggesting an essential role for Smad-2 in
transcriptional activation of the Mix.2 promoter. It is unclear as to
how FAST-1 DNA binding ability is induced, although Smad-2
phosphorylation and subsequent association with FAST-1 may be important
in allowing DNA binding to occur.
The same group further showed that Smad-4 is also contained in ARF complexes (37). They mapped the region of interaction of FAST-1 with Smad-2/Smad-4 and used a yeast two-hybrid system to show that FAST-1 directly interacts with Smad-2 rather than Smad-4. Furthermore, they created a putative dominant-negative FAST-1 containing 153 amino acids of the C terminus (a.a. 366–518) that interacts with the Smads. Injection of the dominant negative fragment into the animal pole of two-cell embryos inhibited the formation of ARF and the mRNA expression of early mesodermal marker brachyury. This suggests that the Smad complex may be directly involved in ARF function.
2. Smads bind DNA directly. A new development to the story came with the discovery that Smads have DNA binding ability (38). The Drosophila MAD, homolog of Smad-1, can bind to a consensus promoter sequence GCCGnCGC in the fly vg promoter. Our work in mammalian cells indicates that Smad-4 can also bind DNA in a sequence-specific manner as discussed below (39). Although the Smads have an intrinsic DNA binding ability, it is unclear as to the exact contribution of Smad DNA binding activity on transcriptional activation. One function for DNA binding might be to cooperate with other adjacent transcription factors to allow stabilization of these factors with the transcriptional machinery complex, thereby resulting in the activation of transcription. This degree of cooperativity may vary in different promoter contexts. For example, in a different Drosophila promoter, Ubx, which also contains the same MAD/Smad-1 DNA binding site, the element responsive to dpp activation was mapped to an adjacent cAMP response/regulatory element (CRE). Further investigation should clarify the role that Smad DNA binding may play in transcriptional activation.
3. Smads can activate transcription. Smad phosphorylation and translocation to the nucleus are thought to effect transcriptional regulation of TGF-ß-responsive genes, one of which may be the gene for PAI-1, a protein important in extracellular matrix deposition. Using a luciferase reporter gene under the transcriptional control of the PAI-1 promoter, it was determined that: 1) reintroduction of Smad-4 could restore TGF-ß activation of PAI-1 in a Smad-4 mutant cell line SW480.7 and 2) overexpression of Smad-3 with Smad-4 was sufficient to activate the PAI-1 promoter in a ligand-independent manner (31). Those initial observations were subsequently substantiated with similar results from studies of Smad-2 and Smad-4 coexpression in the activation of the p3TP-lux reporter in another Smad-4 deficient MDA-MB-468 human breast cancer cell line (40, 41). The p3TP-lux reporter is an artificial promoter consisting of a combination of AP-1 binding site-containing 12-O-tetradecanoylphorbol-13-acetate-responsive elements and a portion of the PAI-1 promoter (39).
Thus, Smads may potentiate transcription by multiple mechanisms: as a transcription factor capable of binding to specific sequences and/or as a coactivator. This conclusion is supported by the results of a recent study in which Smad-3/Smad-4 overexpression was able to activate transcription from a minimal reporter controlled by four sets of a 31-bp sequence derived from the 3TP-lux promoter (39). Smad-4 was shown to bind directly to DNA on a site adjacent to the AP-1 element. To dissect the individual contribution of Smad DNA binding activity and AP-1 on transcriptional activation of p3TP-lux by TGF-ß, two sets of luciferase constructs were constructed. The first set of constructs was a series of concatemerized minimal AP-1 elements from the p3TP-lux promoter. Smad-3/Smad-4 overexpression induced transcription driven by the concatemerized minimal AP-1 elements 20-fold but the construct was barely activated by TGF-ß (39). This suggests that although overexpression of Smads may allow an increase in transactivation, the ability of TGF-ß to mobilize endogenous levels of Smads may not be sufficient to potentiate transcription through a minimal AP-1 element alone. Transcriptional activation may require the presence of a Smad-4 DNA binding site to stabilize the interaction of AP-1 complex with the core transcriptional apparatus. To address this question, a second set of constructs was generated to contain the wild-type AP-1-responsive element with the adjacent Smad-4 DNA-binding sites mutated. Despite mutating the Smad-4 DNA-binding sites, TGF-ß treatment and Smad-3/Smad-4 overexpression could still drive transcription with the same fold induction as in the wild-type construct.
The unanticipated result from this study, therefore, was the apparent dispensability of the Smad binding site within the p3TP-Lux reporter (39). While the lack of correlation between the DNA binding capability of Smads and their ability to activate transcription makes the interpretation of those findings somewhat difficult, the result could be explained in several ways. Smad complex binding may cause effects that cannot be assayed in these transient transfection experiments with the use of a large quantity of plasmid DNA as the template of transcription. For example, Smad binding may play a role in the recruitment of other transcription factors to adjacent sites (e.g., AP1) or in rearrangement of chromosome structure to provide accessibility of other transcription factors to their binding sites so that an effect in the transient transfection assay may be difficult to observe. The transient nature of Smad nuclear accumulation and DNA binding would be consistent with this type of role in transcriptional activation. Alternatively, Smad binding sites may represent enhancer-like regulatory sequences that can function properly only in the context of specific promoters. The proper promoter context may be essential to allow the appropriate interactions between the Smads and the core transcription machinery. Thus, in the context of the artificial p3TP-Lux promoter constructs, Smad binding may not be required, but in the context of wild-type promoters, Smad binding may become indispensable for transcriptional activation. In this regard, the demonstration in the same study that the PAI-1 promoter contains a Smad-3/Smad-4 binding site allows an opportunity to dissect in vivo functions of the Smad-3/Smad-4 binding site and should provide insight into these important questions.
4. Smad transcriptional ability effected by other extracellular signals. In a normal cellular context, Smad transcriptional activity may also be influenced by other extracellular signals that possibly dictate its proclivity for driving TGF-ß-mediated transcription. One such target of modulation might be the ERK (extracellular stimulus responsive kinase/mitogen-activated protein kinase) PXSP phosphorylation motif found in Smad-1 (42). When four of these consensus sites are mutated to an inactive alanine, the mutant protein remains in the nucleus. When treated with epidermal growth factor (EGF), which instigates a receptor-mediated kinase cascade resulting in phosphorylation of PXSP target sites, the wild-type Smad-1 protein remains in the cytoplasm. BMP treatment leads to Smad-1 nuclear accumulation, but when treated with both BMP and EGF, Smad-1 remains in the cytoplasm. These data suggest that the phosphorylation status of Smads in vivo may dictate their responsiveness to TGF-ß superfamily signaling.
D. Smad structure
The transcriptional activation of the AP-1 element by Smads is
supported by a growing body of biochemical findings that relate the
structure of Smads to their function. The five cloned Smads have two
regions of high homology, MH-1 and MH-2, shared with each other and
with both Drosophila MAD and C. elegans SMA
genes. While evolutionarily conserved, these regions contain no known
structural or signaling motifs. Figure 1 compares the percent identity
of seven of the known mammalian Smad protein sequences with their
C. elegans and Drosophila homologs. In addition,
the specific Smads involved in a particular TGF-ß superfamily
signaling pathway are indicated.
1. Autoinhibitory regulation: MH-1 domain represses MH-2 domain. Studies with BMP have provided clues as to possible functions for the different domains of the Smads. In these experiments, the MH-2 region of Smad-1 and Smad-4 possess a constitutive transcription activity when fused to a GAL4 DNA-binding domain (30, 43). This supports the model in which Smads are recruited to specific promoters either by direct DNA binding or through transcription factors and activate transcription. Full-length Smad-4 fused to the GAL4 DNA-binding domain is normally transcriptionally silent, but can be inducibly activated upon BMP treatment. This suggests a model for Smad intramolecular regulation such that the MH-1 normally functions to silence the transcription activity of the MH-2 region of the Smad molecule. Upon ligand treatment, the MH-1 region dissociates from its MH-2 region, allowing the MH-2 region to modulate transcription.
2. Homomeric and heteromeric interactions of Smad-3 and Smad-4. Using yeast two-hybrid and coimmunoprecipitations, the domains important for homomerization and heteromerization of Smad-3 and Smad-4 were mapped (44). In Smad-3, the MH-2 domain mediates homomerization, whereas in Smad-4 both the MH-1 and MH-2 domain are involved. The heteromeric interactions between Smad-3 and Smad-4 are mediated through their MH-2 domains. This correlates with the functional data showing that Smad-3 and Smad-4 MH-2 domain overexpression is sufficient to induce PAI-1 promoter reporter activity (44, 45).
3. Crystallization of MH-2 domain of Smad-4. The recent crystallization of the C-terminal domain (CTD/MH-2) of Smad-4 helps shed light on the importance of this conserved region (46). As mentioned above, the MH-2 domain can act as an effector for the ligand-independent transactivating function. An in-depth characterization of its structure has provided insight into the mechanism by which Smad4 interacts with other Smads and how the interaction is regulated by TGF-ß. The structural information has allowed a determination of the importance of specific amino acids for Smad function through a detailed comparative analysis of sequences from mutant Smads identified in both developmental systems and human cancers.
Based on the structural analysis, the Smad-4 MH-2 domain was shown to consist of a core of five and six antiparallel ß-strands forming a ß-sandwich. These are flanked on one side by a three-helix bundle (H3–H5) and a three-loop/one-helix (L1–L3/H1) region on the other. The three-helix bundle and loop/helix region may play an important role in Smad function because mutations in these regions produce organisms with severe developmental abnormalities. In addition, five Smad-4 mutants isolated from human tumors contain mutations in these regions.
Those mutations were postulated to result in the disruption of protein-protein interactions. Therefore, the ability of wild-type and mutant Smads to homo-oligomerize or hetero-oligomerize was investigated. Crystal structures showed a strong interaction between three Smad4 proteins through an interface composed of one loop/helix region contacting the helix/bundle region of another Smad. Interestingly, only one loop (L3) is free from this homomeric interaction. When oriented, it was discovered that in the homotrimer, the L3 from each protein is positioned on the same side of the protein complex. The L3 region is mutated in both developmental organisms and some tumors. Mutations in L3 prevent hetero-oligomerization whereas mutations in other regions of the loop/helix and three-helix bundle region prevent both homo-oligomerization and hetero-oligomerization.
4. Proline linker region of Smad-4 important for restoration of TGF-ß response. Further domain studies on Smad-4 showed an essential region in Smad-4 that is important in restoring TGF-ß transcriptional activation of the promoter reporter p3TP-lux (41). In the Smad-4 defective cell line, MDA-MB468, transfected full-length Smad-4 restored TGF-ß responsiveness as assayed with the p3TP-lux reporter. Using a combination of chimeric Smad-1/Smad-4 and deletion Smad-4 constructs, an essential region in Smad-4 for restoration of TGF-ß mediated p3TP-lux activity was isolated. This 47-amino acid region of Smad-4 is from amino acids 274–321 and resides in the proline-rich linker region between the MH-1 and MH-2 domains.
E. Smads as negative regulators
Recently, two other Smad family members have been cloned,
Smad-6/DAD and Smad-7 (47, 48, 49, 50). Both these proteins were identified
using expressed sequence tag technology (48, 50). DAD was identified
using enhancer trap screening of Dpp-controlled genes (49), and Smad-7
was identified from proteins induced upon laminar fluid sheer stress
(47). These proteins share MH-2 domain homology with the other Smad
family members but lack an MH-1 domain, suggesting a potential
functional difference. The proteins lack the three regulatory serines
at the C-terminal end conserved in the other Smads (except Smad-4)
suggesting a different mode of regulation by TGF-ß and its family
members. Smad-7 RNA was shown to be induced within 30 min of TGF-ß
treatment in both HaCaT and mink lung epithelial cells before peaking
at 90 min (50). Both Smad-6 and Smad-7 may interfere with
Smad-associated gene responses such as p3TP-lux induction (47, 48, 50).
In addition, Smad-7 can block GAL4-Smad2 transactivation by TGF-ß
(47). From genetic studies in Drosophila, a distinct wing
phenotype caused by overexpression of MAD can be rescued by
overexpression of DAD, the Drosophila homolog for Smad-6
(49).
The abilities of Smad-6 and Smad-7 to associate with and become
phosphorylated by the receptor complex were investigated. In COS cells,
Flag-tagged Smad-6 or Smad-7 was overexpressed and the type I and type
II receptor complex was covalently affinity labeled with
[125I]TGF-ß1. Flag-antibodies coimmunoprecipitated the
receptor-ligand complex independent of the co-transfected type I
receptor kinase status (WT or K
R), although the integrity of type II
receptor kinase status must be maintained (48, 50). This suggests that
Smad-6 and Smad-7, when overexpressed, interact with the receptor-Smad
complex. Overexpressed Smad-6 appears to block the ability of the type
I receptor to phosphorylate Smad-1 and Smad-2 whereas Smad-7 blocks the
ability of the type I receptor to phosphorylate Smad-2 and Smad-3 (47, 48, 50). Smad-6 or Smad-7 may directly bind to the receptor complex to
prevent this phosphorylation, or interacts with the complex through
endogenous Smads. A truncated form of Smad-7, D408, is unable to block
TGF-ß-mediated induction of p3TP-lux activity, does not associate
with the receptor, and does not prevent Smad-2 interaction with the
receptor (47).
From both the biochemical and functional data it is clear that these new Smad family members, Smad-6 and Smad-7, are biochemically regulated differently by TGF-ß and its superfamily members and thus have different functions than the previously cloned Smads. Given the fact that Smad-7 can be transcriptionally activated by TGF-ß in both mink lung epithelial cells and human keratinocytes, it is possible that Smad-7 may play a role as an effector of other uncharacterized effects of TGF-ß signaling, in addition to its possible role in feedback regulation as a negative regulator of Smad-2 or Smad-3 signaling. As such, these findings represent an additional exciting new area of research for Smad signaling, and reinforce the potential importance of Smad family members as effectors of the TGF-ß signaling cascade.
F. Smad-deficient mice
Recently, homozygous Smad-4 mutant mice were created (51). The
mice were embryonic lethal before embryonic day 7.5. The mutant embryos
have developmental defects that include failure to gastrulate or
express mesodermal markers, an abnormal visceral endoderm, and a
reduced size. Mutant embryos were rescued with wild-type visceral
endoderm, suggesting that Smad-4 is required for visceral endoderm
differentiation. Smad-4-deficient embryos exhibit a similar phenotype
as those of the BMP4- and BMPR-I deficient embryos which exhibit a
reduction in size and inability to express mesodermal markers.
Because Smad-4 homozygous mutant mice were embryonic lethal, the effect of Smad-4 loss on TGF-ß signaling in adult tissues could not be studied. Other Smad-deficient mice that are being generated include Smad-1, 2, 3, and 5. Work is proceeding apace to analyze these mutants.
G. Smads: an emerging model
A model has emerged as to how Smads are activated by TGF-ß
family members. Researchers in the field have made a comparison between
Smad signaling and signal transducers and activators of transcription
(STAT) signaling. Upon phosphorylation by the receptor tyrosine
kinases, Janus kinases (JAKS) phosphorylate the STATs, which in turn
dimerize and translocate to the nucleus, where they bind to DNA and
activate transcription. Figure 2 depicts
a proposed model for TGF-ß activation of Smads from the cytoplasm to
the nucleus. TGF-ß treatment causes the formation of heteromeric type
II receptor/type I receptor/ligand complexes. Type II receptor
phosphorylates the type I receptor at the GS domain, thereby activating
it. Type I receptor, in turn, phosphorylates its Smad-2 or Smad-3
substrate. When phosphorylated, this homotrimeric Smad forms a
hexaheteromer with Smad-4. This event most likely allows entry of the
Smad complex into the nucleus where the Smads can make contacts with
specific DNA elements and/or other sequence-specific transcription
activators, perhaps as well as components of the core transcriptional
machinery, to modulate transcription. This model is highly simplified
for a number of reasons. At the cytoplasmic level, it is unclear
whether the Smads are phosphorylated by type I receptor in a homotrimer
complex. Whether this phosphorylation triggers a Smad-4 interaction
concomitant with translocation into the nucleus is also unknown. In the
nucleus, even more questions remain unanswered. What are other
Smad-inducible target genes, in addition to the PAI-1 gene? What
dictates the different modes of transcriptional activation, as a
coactivator or through DNA binding, the Smads may use? How does Smad
localization into the nucleus precipitate transcriptional activation?
Do Smads bind to members of the basal transcription machinery? With
what other transcription factors might the Smads interact? These are
just some of the more important Smad-related questions whose answers
will further define the role of Smads in TGF-ß superfamily signaling.
In addition, Smad-6/DAD and Smad-7 may represent a completely new class
of Smads with different functions in TGF-ß signaling. Analysis of
Smad-deficient mice should soon bring a great deal of insight into the
physiological role various Smads play in TGF-ß signaling.
|
|
IV. TGF-ß and the Cell Cycle |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
Cytoplasmic events that propagate the TGF-ß growth-arrest signal into the nucleus are not well understood, and the involvement of Smads in this process has not been determined (our unpublished results). The nuclear signal by which TGF-ß halts the cell cycle is well characterized and can be best explained by beginning with TGF-ß’s link to the retinoblastoma protein, Rb (52, 53).
Rb is a cell cycle protein that is differentially phosphorylated during the cell cycle. About two-thirds of the way through G1, at the restriction point, it changes from a hypophosphorylated state to a hyperphosphorylated state that is maintained through the cell cycle until the cell emerges from mitosis. Cells are only responsive to TGF-ß when they are in the window of G1 before the restriction point. Once past the restriction point, the cell is committed to completing the cell cycle (54). Thus, TGF-ß mediated growth arrest was found to be correlated with an accumulation of hypophosphorylated Rb, although the precise mechanism of action was not known at that time.
Further work in the cell cycle field showed that the kinase activity of specific enzyme complexes was required for G1 to S phase transition (55, 56, 57). Importantly, the kinase activity of these complexes was regulated in a cell-cycle dependent manner. The regulatory subunit of these complexes are known as ’cyclins,’ and their catalytic subunits are the ’cyclin-dependent kinases’ (CDKs). In addition to cyclin association, the kinase activity of the CDKs can be regulated in three other ways. First, the cyclins can be regulated through transcription. Second, activating and deactivating phosphorylations can regulate CDK activity (e.g., CDK4 has an activating phosphorylation site on threonine-160 and inactivating phosphorylation sites on threonine-14 and tyrosine-15). The kinases and phosphatases that regulate CDK phosphorylation can also be regulated. In the case of CDK4, the activating kinase is CAK, and the activating phosphatase is cdc25A. Third, a group of low molecular wt molecules known as the cyclin-dependent kinase inhibitors (CKIs) have been shown to associate with and concomitantly block CDK activity.
It was hypothesized that one of the targets for cyclin-CDK phosphorylation was Rb. Indeed, certain cyclin-CDK complexes could phosphorylate Rb. The phosphorylation of Rb prevents its association with the transcription factor, E2F. Without Rb binding to it, E2F can transcriptionally activate genes necessary for S phase progression (reviewed in Ref. 58). Recent studies suggest the existence of two rate-limiting steps during the G1 to S phase transition (59). The first step is dictated by the ability of CDK4 to phosphorylate in vivo substrates, including Rb, thereby allowing E2F to transactivate S phase-required genes. The second step is dictated by the ability of CDK2 to phosphorylate a much broader range of in vivo targets, which may include other proteins in addition to Rb and Rb family members. At present, other targets of the cyclin-CDK complexes are not known. To arrest cells in G1, TGF-ß would potentially have to block the activities of both CDKs.
A. TGF-ß induction of the CKIs
Of the four strategies by which CDK activity can be regulated as
mentioned above, TGF-ß was first shown to be able to increase the
amount of functional cyclin-dependent kinase inhibitors in the cell.
These CKIs can be grouped into two subsets based on homology and
function. The first group of proteins include p21 (WAF1/Cip1), p27
(Kip1), and p57 (Kip2). The second group of inhibitors includes p15
(INK4B/MTS2), p16 (INK4A/MTS1), p18, and p19. The first group of CKIs
bind to cyclin-CDK complexes and inhibit complex activity while the
second group of CKIs bind to CDKs alone, which sequesters the catalytic
CDKs from their regulatory cyclin partners, thereby preventing
activity.
1. In the mink lung epithelial cell line CCL64. The first CKI to be implicated in TGF-ß G1 arrest was p27. It was found to be associated with Cyclin E-CDK2 complexes in mink lung epithelial cells and subsequently cloned by both biochemical purification and the two-hybrid system in yeast (60, 61, 62). p27 Binds to and inhibits the activity of cyclin E-CDK2, indicating that it plays a negative regulatory role in the cell cycle. Recently, the crystal structure of the protein complex between Cyclin A-CDK2 and p27 has been resolved. p27 Makes contacts with both Cyclin A and CDK2 so that ATP can no longer bind to the active site of CDK2 enzyme (63). In functional assays, overexpression of p27 can arrest Saos-2 cells. Although p27 is placed functionally in the TGF-ß pathway due to its ability to inhibit cyclin E-CDK2 activity, TGF-ß does not increase its transcription or translation.
In mink lung epithelial cells, TGF-ß can potently increase the transcription of another CKI, p15 (64). This inhibitor functions by sequestering CDK4 and CDK6 (64). The current model for TGF-ß growth arrest in mink lung epithelial cells is that upon TGF-ß treatment, p15 transcription is induced. This causes an increase in p15 protein levels, leading to their association and sequestration of CDK4. p27 Protein, which normally associates with cyclin D1-CDK4, is now displaced and binds to its lower affinity partner Cyclin E-CDK2. The combined inactivation of both CDK2 and CDK4 corresponds with a decrease in Rb phosphorylation and arrest in G1.
2. In the human keratinocyte cell line HaCaT. The mechanism by which TGF-ß increases the expression of the CKIs p15 and p21 is more easily understood in human keratinocytes. TGF-ß treatment causes a rapid increase in p15 and p21 RNA levels within 1 h (64, 65, 66). For p21, this corresponds to an increase in protein levels, an increased association with cyclin E-CDK2 and cyclin D1-CDK4, and a decrease in the activity of those cyclin-CDK complexes. The association of p21 with CDK complexes is thought to prevent activating phosphorylations on CDKs (T-160). Recent crystallography studies suggest that the p21 family of inhibitors (including p21 and p27) function by masking the ATP-binding domain of CDK2. For p15, the increase in protein level allows its sequestration of CDK4 and CDK6, thereby inhibiting kinase activity.
To determine the mechanism by which TGF-ß regulates the transcription of p15 and p21, a comprehensive analysis of the TGF-ß response elements in the promoters of both p15 and p21 was conducted. Both TGF-ß response elements mapped to a 6-bp GC-rich element (67, 68). Electromobility shift analysis confirmed that Sp1 family members could bind to this element, although no changes in binding affinities were detected for any of the complexes upon TGF-ß treatment. Further functional assays with different GAL4-Sp1 fusions showed that GAL4-Sp1 proteins can confer TGF-ß responsiveness. These experiments demonstrate that, in addition to physically associating with the TGF-ß response element, Sp1 is also functionally important in modulating the TGF-ß-mediated response, although TGF-ß does not regulate its level of expression. The protein synthesis inhibitor cycloheximide does not inhibit TGF-ß’s ability to induce p21 or p15 RNA accumulation, suggesting that the level of regulation by TGF-ß is posttranslational. Overall phosphorylation of Sp1 does not change upon TGF-ß treatment, suggesting another possible scenario involving the differential modification of an adaptor molecule, which may activate Sp1 to allow transactivation of the p15 and p21 promoters.
a. E1A blocks TGF-ß-mediated growth inhibition at multiple
levels.
To determine the identity of this hypothetical adaptor
protein, the adenoviral immediate early gene product, 12S E1A, proved
to be a valuable tool. Both E1A of adenovirus and the large T antigen
of SV40 prevent TGF-ß-mediated growth arrest, presumably due to the
ability of both of these proteins to bind Rb and mimic the
hyperphosphorylation state of Rb (52, 69, 70, 71, 72, 73, 74). The dissociation of Rb
allows E2F to transactivate genes important for S phase progression
(58). Although the ability of E1A and large T to bind Rb appears to be
sufficient to abrogate TGF-ß signaling, cellular infections with a
mutant E1A adenovirus, Ad 928, which is severely attenuated in its
ability to bind Rb, demonstrate that it too is able to overcome
TGF-ß-mediated growth arrest in mink lung epithelial cells (75). If
the phosphorylation status of Rb dictates the ability of a cell to
enter S phase, it was hypothesized that E1A might be able to intercept
Rb function at a point upstream of physical sequestration of Rb, such
as at the level of the CKIs. This model was particularly attractive
since the suppression of TGF-ß-mediated transcriptional induction of
the CKIs could prevent the hyperphosphorylation of Rb, and hence
provide E1A with an additional mechanism by which to abrogate TGF-ß
signal and create an S phase environment necessary for viral
replication. This model proved to also apply in mink lung epithelial
cells, where E1A has recently been shown to bind directly to p27, but
not p21 (76).
b. E1A blocks CKI induction by binding to p300.
In HaCaT,
ribonuclease (RNase) protection analysis showed that p15 mRNA levels
are not induced upon TGF-ß treatment when HaCaT cells are infected
with adenovirus expressing 12S E1A. To further isolate the proteins
involved in abrogation of TGF-ß signaling, cotransfections were
performed with different 12S E1A mutants. Of these, an N-terminal
deletion mutant of E1A displayed an attenuated ability to block
TGF-ß-mediated induction. This implicated proteins bound to the
N-terminal region of E1A, such as p300, as part of the TGF-ß-
signaling pathway. Additional luciferase experiments with pCMV-p300
cotransfected with pCMV-12S-E1A demonstrated that expression of p300
could partially rescue the inhibition of response to TGF-ß caused by
E1A. In addition to demonstrating a novel function for E1A by blocking
CKI induction, these experiments directly point to p300 as a potential
mediator of TGF-ß activation of the p15 and p21 promoters (77).
p300 Was first discovered in anti-E1A immunoprecipitations of radioactively labeled cells. It associates with the N-terminal region of E1A, which by itself is sufficient to stimulate DNA synthesis. A different region of E1A containing the pocket-domain that binds to Rb and Rb family members is also capable of inducing DNA synthesis (78, 79, 80). However, both regions of E1A were necessary to form colonies in soft agar assays (81). Ubiquitously expressed, the p300 protein contains three cysteine-histidine rich regions (two of which are putative zinc fingers), a bromo domain, and a nuclear localization sequence (82). p300 Has a functional homolog, CREB-binding protein (CBP), which also binds to E1A (83, 84, 85). p300 And CBP can both overcome E1A-mediated expression of the SV40 enhancer activity. p300 And CBP contact both transcription factors and the basal machinery, suggesting a role for them as transcriptional adaptors or coactivators. Recently, p300 and CBP have been implicated in nuclear receptor signaling (86, 87), cAMP signaling (88, 89), STAT signaling (90), calcium-dependent differentiation (74A ), and myogenesis (91).
c. A working hypothesis for the mechanism of TGF-ß mediated
transcriptional activation of the p15 and p21 promoters.
As
discussed above, the mechanism by which TGF-ß acts to induce
transcription of the p15 and p21 promoters remains largely unknown.
Studying this mechanism could lead to further insight into what factors
are modified by TGF-ß that allow it to differentially regulate gene
transcription and thereby engineer its cellular activities. The
implication that p300 and CBP are involved in TGF-ß signaling creates
an opportunity to further dissect its signaling mechanism. Furthermore,
the Smads may be found to act as one of the effectors transducing the
TGF signal from the activated receptor complex directly to the
transcriptional apparatus assembled onto the p15 and p21 promoters,
thus linking the tumor suppressor activitity of Smads to the cell cycle
control machinery. Based on available evidence, a working hypothesis
can be proposed, which consists of the following two aspects: 1)
p300/CBP may respond to the TGF-ß signal to form a functional complex
with Sp1, thus leading to transcriptional activation of the p15 and p21
promoters. Our studies with NGF mediated transcriptional induction of
the p21 promoter through a NGF-induced interaction between p300 and Sp1
(our unpublished results) support this model in which a similar
interaction between the two proteins can be induced by TGF-ß. 2) The
Smads may form a complex with p300/CBP and/or Sp1 to activate
transcription once they translocate into the nucleus. This hypothesis
is supported by our preliminary findings that coexpression of
"dominant negative" forms of either Smad3 or Smad4 lacking their
conserved functional MH1 domain with the p15 and p21
promoter-controlled reporter constructs result in a significant
reduction in TGF-ß-induced transcriptional activity. Although we
cannot rule out the possibility that the truncated Smads can sequester
some common transcription factors other than the coactivator p300/CBP
and/or Sp1, this attractive model needs to be further explored. One
important feature of this potentially Smad-mediated transactivating
event, however, lies in the finding that overexpression of the Smads
does not lead to the activation of the p15 and p21 promoters (our
unpublished results), unlike the scenario with the AP1
elements-controlled transcriptional activation of PAI-1 promoter
mediated by the overexpression of Smads as discussed in the last
section. It is possible that Smads with proper structural conformation,
achieved through receptor-mediated phosphorylation, are required to
physically interact with p300/CBP and/or Sp1. Alternatively but not
exclusively, a separate signaling cascade, also initiated at the
receptor complex, may be required in conjuction with Smads to activate
transcription at the p15 and p21 promoters. Further analysis in this
direction will undoubtly yield valuable information in elucidating the
mechanism through which the p15 and p21 promoters are regulated by
TGF-ß.
B. TGF-ß-mediated decrease in Cdc25A levels in the breast
epithelial cell line MCF10A
Recently, a different mechanism for cell cycle arrest was observed
in a spontaneously immortalized epithelial line, MCF10A (92). In these
cells, a rapid decrease in mRNA levels of the cell cycle phosphatase,
cdc25A, is observed upon TGF-ß treatment. Consequently, protein
levels for cdc25A decrease by 8 h. As mentioned above, cdc25A is
thought to regulate CDK activity. Based on an analogous model with cdc2
and the wee1 kinase/cdc25 phosphatase in the G2/M transition,
dephosphorylation of an inhibitory tyrosine on CDK4, presumably by
cdc25A, is thought to be required for cell cycle progression. In the
MCF10A cells, the authors correlate a decrease in cdc25A activity with
an increase in tyrosine phosphorylation of CDK4 and CDK6. As such, this
may be another mechanism by which TGF-ß can cause an arrest in
G1. E1A can also increase cdc25A activity in quiescent
fibroblasts, suggesting another possible level of regulation by which
E1A can overcome TGF-ß growth inhibition (93).
C. TGF-ß-mediated cyclin-CDK inhibition may be common strategy to
arrest cells in G1
TGF-ß appears to halt cell cycle progression in a cell
type-specific manner. This may reflect the intrinsic differential
protein profiles that separate one cell type from another. Thus, the
entire cell cycle-arrest programming may vary from cell type to cell
type. One example is the difference between TGF-ß-mediated CKI
induction of mink lung epithelial cells and human keratinocytes. In
other cells, TGF-ß may employ one of the remaining three strategies
for influencing CDK activity to effect a G1 cell cycle
arrest. The breast epithelial cell line, MCF10A, is one such example.
Figure 3 is a schematic summarizing the
known strategies by which TGF-ß inhibits CDK activity. In this
section, we have reviewed some of the documented strategies by which
TGF-ß impinges on the cell cycle to effect an arrest in
G1 depending on cell type. Other cell types that are also
responsive to the TGF-ß growth-inhibitory signal, such as
fibroblasts, may not use these mechanisms. However, we suggest that to
arrest the cells in G1, TGF-ß may employ other strategies
limiting cyclin-CDK activities to reach the same growth inhibition
endpoint. Restricting cyclin-CDK kinase activity may be a common
denominator by which TGF-ß can arrest cells in G1 phase
of the cell cycle.
|
|
V. TGF-ß and Cancer |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
In this review, we have focused on characterizing the known components of the TGF-ß pathway. Components that are necessary for the negative-regulatory TGF-ß signal to propagate, such as the receptors, are commonly mutated in cancer. Blockage of TGF-ß signaling may disrupt this cellular steady state, biasing the cell toward inappropriate growth that ultimately results in tumor formation. In this section, we will briefly describe more mutations in type II receptor that have recently been discovered in cancer patients. The primary focus of this section will be on the discovery of mutated Smad family members in cancers. In addition, other cell cycle components in the TGF-ß growth arrest pathway that are mutated in cancers will be briefly mentioned.
A. Receptor mutants
In recent years, an overwhelming body of evidence shows that
mutations or loss of expression of the type II receptor are found in a
number of gastrointestinal cancer lines (94, 95, 96, 97, 98, 99, 100, 101, 102). In most of the cell
lines assayed, type II receptor was either mutated or not expressed
correlating with an unresponsiveness to the TGF-ß growth-inhibitory
signal. Southern blot analysis revealed deletions in the type II
receptor gene. Sequencing of other mutants revealed 1-bp frame shifts.
To verify the importance of the type II receptor mutant on
carcinogenesis, wild-type receptor II was transfected into cells and
the cells were assayed for fibronectin expression, foci formation, and
clonogenicity in soft agar (98). In the receptor II-transfected cells,
a TGF-ß-dependent increase in fibronectin expression and a reduction
in clonogenicity were observed, suggesting not only that the block in
TGF-ß signaling correlated directly with an increase in
clonogenicity, but that the downstream signaling components were still
functional. Similar experiments have also been published in small cell
lung cancer and breast cancer (103, 104, 105), suggesting that mutations in
receptor type II expression are not specific for gastrointestinal
cancers. These data strongly support the notion that type II receptor
dysfunction is an important event in cancer progression. Although some
cancer lines have type I receptor mutations, they are less frequent
(105).
B. Smad mutants
Many of the events downstream of TGF-ß signaling have been shown
to be mutated in cancers. Two of these are Smad-2 and Smad-4. Using
PCR, researchers showed that a region on chromosome 18q that maps to
Smad-4 is homozygously deleted in pancreatic carcinomas. This region is
specific for Smad-4 and not the gene DCC (Deleted in Colorectal
Cancer), which is located adjacent to Smad-4. The deleted fragment was
isolated from yeast artificial chromosomes (YAC) and hybridized to a
human fetal brain cDNA library leading to the cloning of Smad-4 (34).
In 84 pancreatic carcinomas screened, Smad-4 was homozygously deleted
in 25 cases. Of 27 pancreatic carcinomas without homozygous deletions,
six had additional mutations. Ninety percent of all pancreatic tumors
screened have deletions in chromosome 18, but only 50% are deletions
for Smad-4. Researchers hypothesized that another dysfunctional gene on
chromosome 18 may also contribute to pancreatic carcinomas.
The discovery that Smad-2 mapped to 18q21 just two genes away from Smad-4 suggests that Smad-2 may be the other mutational target contributing to pancreatic carcinomas (32). Based on our current knowledge of Smad signaling, it is an attractive model to suggest that only one Smad-2 or one Smad-4 homozygous deletion is necessary to abrogate their physiological role in TGF-ß signaling since their cooperation is needed for appropriate signal transduction. At this time, the loss of Smad-2 in pancreatic carcinoma has yet to be characterized. Of 66 colon carcinomas screened, four were found to contain point mutations in Smad-2. Of these, three could not be phosphorylated by TGF-ß. Since phosphorylation of Smad-2 initiates their localization to the nucleus, these mutations may prevent the proper Smad-2 homotrimerization and subsequent heterohexamerization with Smad-4 to allow nuclear entry. Recently, it was discovered that the MH1 region of Smad 2 and Smad 4 bind to and inhibit the MH2 region from signaling. Phosphorylation of the C-terminal region may prevent the Smad-2 MH2 domain from interacting with its MH1 domain, thereby activating it. Overexpression of the MH1 domain of either Smad-2 or Smad-4 can prevent interaction between full-length Smad-2 and Smad-4 (43). However, homotrimerization of the MH2 domain was not prevented by the MH1 domain of either Smad-2 or Smad-4. Functionally, the Smad-4 R100T mutant or the Smad-2 R133C mutant may be preventing wild-type Smad-2 from interacting with wild-type Smad-4, since cotransfection of wild-type Smads in conjunction with Smad mutants blocks the ability of the overexpressed Smads to transactivate the 3TP-lux reporter. In addition to pancreatic cancers, Smad-4 mutations have also been discovered in breast, ovary, head and neck, and esophagus cancers (34, 106, 107, 108, 109), whereas Smad-2 mutations have only been found in colon, head, and neck cancers (32, 33, 110).
C. Cell cycle mutants
Mutations in basic cell cycle components are a frequent target of
cancers, since a change in a single cell-cycle protein may provide a
significant growth advantage for the cell. Recently, CBP and p300
mutations have also been discovered in gastric and colon carcinomas and
leukemias (111, 112). Gastric and colon carcinomas frequently contain
loss of heterozygosity on chromosome 22q. Although p300 is mutated in
both of these cancers, the frequency is low (one of six) in colon
cancers, and thus the overall contribution of p300 to these tumors is
unclear. Other downstream targets of TGF-ß action include Rb and p15.
Mutations and deletions in these genes have been found in wide variety
of cancers (113). In addition to mutations of the negative regulators
mentioned above, oncogenic mutations in positive cell cycle regulators
can also give the cell a tremendous growth advantage. In certain
cancers, mutations have been discovered in cyclin D1, cdk4, and the
cdc25 family of phosphatases. The contribution of both types of cell
cycle mutants has been reviewed extensively in the literature.
D. Multilevel resistance to TGF-ß is important in multistep model
of carcinogenesis
Resistance to TGF-ß growth inhibition may occur at multiple
levels in a cancerous cell. Receptor mutations create an unresponsive
phenotype, and mutations in different downstream components of the
TGF-ß-signaling pathway may also generate the same phenotype.
Mutations in basic cell cycle regulators such as Rb, p15, p21, and
p300/CBP may prevent negative growth regulators in addition to TGF-ß
from regulating cell growth. These mutations may also cooperate with
oncogenic mutations such as cdk4 or cyclin D to allow rapid cellular
proliferation. At present, it is unclear whether the Smads are part of
the cell cycle machinery since their overexpression does not lead to an
increase in CKIs p15 and p21 (our unpublished results). However, the
Smads are clearly involved in extracellular matrix deposition as their
overexpression results in the transcriptional induction of PAI-1.
Mutations in this pathway could lead to differences in cell-cell
contact that may provide a growth advantage to the cell. These multiple
levels of deregulation ultimately result in an aggressive, rapidly
growing, highly metastatic tumor that is unresponsive to negative
growth signals.
|
VI. Conclusions |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
The propagation of the Smad signal from the cytoplasm to the nucleus is the only translocation event known for TGF-ß signaling. Thus, it is interesting to speculate on a possible role for the Smads as effectors for TGF-ß-mediated growth inhibition. Although Smad overexpression can lead to a higher percentage of cells arresting (28), whether this arrest pathway is molecularly identical to the well characterized TGF-ß-mediated growth inhibition, mediated by either an increase in p15 or p21 protein levels or a decrease in cdc25A levels to decrease cyclin-CDK activity, remains to be determined. Smad overexpression does not appear sufficient to cause an increase in p15 and p21 promoter activity, suggesting the requirement of another TGF-ß-mediated signal to mediate growth inhibition (see discussion in Section IV.A). Whether Smads are required for TGF-ß-mediated growth inhibition also remains to be determined. Furthermore, the molecular mechanism by which the p15 and p21 promoters are activated through the functional interactions between p300/CBP, Sp1, and possibly Smads, will be the focus of immediate intensive research.
In cancer, not only are the genes involved in the cell cycle often mutated, genes involved in cell-cell contact or extracellular matrix deposition such as Smads, when mutated, may induce a morphological change on the cell surface such that these cells have a growth advantage or are more easily able to metastasize. This may be the case with Smad-4/DPC, which has been shown to be mutated in many pancreatic cancers. Although the Smad-4 heterozygous null mice do not appear more predisposed to cancer compared with their wild-type counterparts, it is possible that other mutations are also required for tumorigenesis (51). The loss of Smad-4 may represent a relatively late event in carcinogenesis that triggers the beginning of a more malignant and aggressive form of cancer.
In addition to its potential deregulation in diseases such as cancer, TGF-ß may cause more global effects in multicellular systems such as in the immune response, wound healing, and synaptic facilitation. Whether Smad signaling and/or TGF-ß mediated growth inhibition is important in these cells remains to be determined. The Smad-deficient mice may also prove to be very useful in addressing these questions.
With the collaborative efforts of so many researchers focused on the dissection of the TGF-ß-signaling circuitry, more answers should be forthcoming. With the recent discovery of the Smad family of proteins and new insights into the mechanism of TGF-ß cell cycle arrest, a new era in the field of serine-threonine receptor kinase signaling is beginning.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Supported by a National Science Foundation Predoctoral Fellowship. 
2 Leukemia Society Scholar. Supported by NIH Grant DK-45746.
3 Abbreviations: TGF-ß, Transforming growth
factor-ß; CDK, cyclin-dependent kinase; CKI, cyclin-dependent kinase
inhibitor; BMP, bone morphogenic protein, ARE, activin response
element; ARF, activin response factor; PAI-1, plasminogen activator
inhibitor; EGF, epidermal growth factor; DPC-4, deleted in pancreatic
cancer, locus 4; DAD, daughters against decapentaplegic; dpp,
decapentaplegic; MAD, mothers against decapentaplegic.
|
References |
|---|
|
TopAbstractI. IntroductionII. Ligands and ReceptorsIII. SmadsIV. TGF-{beta} and the...V. TGF-{beta} and CancerVI. ConclusionsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. M. Rechtman, A. Nakaryakov, K. E. Shapira, M. Ehrlich, and Y. I. Henis Different Domains Regulate Homomeric and Heteromeric Complex Formation among Type I and Type II Transforming Growth Factor-{beta} Receptors J. Biol. Chem., March 20, 2009; 284(12): 7843 - 7852. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wang, L. Yang, J. Yang, K. Kuropatwinski, W. Wang, X.-Q. Liu, J. Hauser, and M. G. Brattain Transforming Growth Factor {beta} Induces Apoptosis through Repressing the Phosphoinositide 3-Kinase/AKT/Survivin Pathway in Colon Cancer Cells Cancer Res., May 1, 2008; 68(9): 3152 - 3160. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. A. Compton, D. A. Potash, C. B. Brown, and J. V. Barnett Coronary Vessel Development Is Dependent on the Type III Transforming Growth Factor {beta} Receptor Circ. Res., October 12, 2007; 101(8): 784 - 791. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K.R. Barnett, C. Schilling, C.R. Greenfeld, D. Tomic, and J.A. Flaws Ovarian follicle development and transgenic mouse models Hum. Reprod. Update, September 1, 2006; 12(5): 537 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kfir, M. Ehrlich, A. Goldshmid, X. Liu, Y. Kloog, and Y. I. Henis Pathway- and Expression Level-Dependent Effects of Oncogenic N-Ras: p27Kip1 Mislocalization by the Ral-GEF Pathway and Erk-Mediated Interference with Smad Signaling Mol. Cell. Biol., September 15, 2005; 25(18): 8239 - 8250. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Hsu, T. Yamamoto, J. Borke, D. S. Walsh, B. Singh, S. Rao, K. Takaaki, N. Nah-Do, C. Lapp, D. Lapp, et al. Green Tea Polyphenol-Induced Epidermal Keratinocyte Differentiation Is Associated with Coordinated Expression of p57/KIP2 and Caspase 14 J. Pharmacol. Exp. Ther., March 1, 2005; 312(3): 884 - 890. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. Mruk and C. Y. Cheng Sertoli-Sertoli and Sertoli-Germ Cell Interactions and Their Significance in Germ Cell Movement in the Seminiferous Epithelium during Spermatogenesis Endocr. Rev., October 1, 2004; 25(5): 747 - 806. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Tomic, K. P. Miller, H. A. Kenny, T. K. Woodruff, P. Hoyer, and J. A. Flaws Ovarian Follicle Development Requires Smad3 Mol. Endocrinol., September 1, 2004; 18(9): 2224 - 2240. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Kohli, S. Hu, E. Clelland, T. Di Muccio, J. Rothenstein, and C. Peng Cloning of Transforming Growth Factor-{beta}1 (TGF-{beta}1) and Its Type II Receptor from Zebrafish Ovary and Role of TGF-{beta}1 in Oocyte Maturation Endocrinology, May 1, 2003; 144(5): 1931 - 1941. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Hill, M. V. Davies, A. A. Pearson, J. H. Wang, R. M. Hewick, N. M. Wolfman, and Y. Qiu The Myostatin Propeptide and the Follistatin-related Gene Are Inhibitory Binding Proteins of Myostatin in Normal Serum J. Biol. Chem., October 18, 2002; 277(43): 40735 - 40741. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. N. Lee, E. Y. Jung, H. J. Kwun, H. K. Jun, D.-Y. Yu, Y. H. Choi, and K. L. Jang Hepatitis C virus core protein represses the p21 promoter through inhibition of a TGF-{beta} pathway J. Gen. Virol., September 1, 2002; 83(9): 2145 - 2151. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Giannelli, E. Fransvea, F. Marinosci, C. Bergamini, S. Colucci, O. Schiraldi, and S. Antonaci Transforming Growth Factor-{beta}1 Triggers Hepatocellular Carcinoma Invasiveness via {alpha}3{beta}1 Integrin Am. J. Pathol., July 1, 2002; 161(1): 183 - 193. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. K. Ghosh Factors Involved in the Regulation of Type I Collagen Gene Expression: Implication in Fibrosis Experimental Biology and Medicine, May 1, 2002; 227(5): 301 - 314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Tomic, S.G. Brodie, C. Deng, R.J. Hickey, J.K. Babus, L.H. Malkas, and J.A. Flaws Smad 3 May Regulate Follicular Growth in the Mouse Ovary Biol Reprod, April 1, 2002; 66(4): 917 - 923. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P. Grande, G. M. Warner, H. J. Walker, A. N. K. Yusufi, J. Cheng, C. E. Gray, J. B. Kopp, and K. A. Nath TGF-{beta}1 Is an Autocrine Mediator of Renal Tubular Epithelial Cell Growth and Collagen IV Production Experimental Biology and Medicine, March 1, 2002; 227(3): 171 - 181. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Zhang, K. Graziano, T. Pham, C. D. Logsdon, and D. M. Simeone Adenovirus-mediated gene transfer of dominant-negative Smad4 blocks TGF-{beta} signaling in pancreatic acinar cells Am J Physiol Gastrointest Liver Physiol, June 1, 2001; 280(6): G1247 - G1253. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K.-C. Choi, S. K. Kang, C.-J. Tai, N. Auersperg, and P. C. K. Leung The Regulation of Apoptosis by Activin and Transforming Growth Factor-{beta} in Early Neoplastic and Tumorigenic Ovarian Surface Epithelium J. Clin. Endocrinol. Metab., May 1, 2001; 86(5): 2125 - 2135. [Abstract] [Full Text]
|
|
|
|
|
|
B. D. Rodgers, G. M. Weber, C. V. Sullivan, and M. A. Levine Isolation and Characterization of Myostatin Complementary Deoxyribonucleic Acid Clones from Two Commercially Important Fish: Oreochromis mossambicus and Morone chrysops Endocrinology, April 1, 2001; 142(4): 1412 - 1418. [Abstract] [Full Text]
|
|
|
|
 |
|
 H. Kaji, L. Canaff, J.-J. Lebrun, D. Goltzman, and G. N. Hendy Inactivation of menin, a Smad3-interacting protein, blocks transforming growth factor type beta signaling PNAS, March 7, 2001; (2001) 61358098. [Abstract] [Full Text]
|
|
|
|
 |
|
 A Franchi, O Gallo, I Sardi, and M Santucci Downregulation of transforming growth factor {beta} type II receptor in laryngeal carcinogenesis J. Clin. Pathol., March 1, 2001; 54(3): 201 - 204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.-C. Kang, M. Ohmori, N. Harii, T. Endo, and T. Onaya Pax-8 Is Essential for Regulation of the Thyroglobulin Gene by Transforming Growth Factor-{beta}1 Endocrinology, January 1, 2001; 142(1): 267 - 275. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Qu, P. A. Godin, M. Nisolle, and J. Donnez Expression of receptors for insulin-like growth factor-I and transforming growth factor-{beta} in human follicles Mol. Hum. Reprod., February 1, 2000; 6(2): 137 - 145. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. P.-c. Hu, X. Shen, D. Huang, Y. Liu, C. Counter, and X.-F. Wang The MEK Pathway Is Required for Stimulation of p21WAF1/CIP1 by Transforming Growth Factor-beta J. Biol. Chem., December 10, 1999; 274(50): 35381 - 35387. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. P.-c. Hu, B. L. Harvat, S. S. Hook, X. Shen, X.-F. Wang, and A. R. Means c-Jun Enhancement of Cyclic Adenosine 3',5'-Monophosphate Response Element-Dependent Transcription Induced by Transforming Growth Factor-{beta} Is Independent of c-Jun Binding to DNA Mol. Endocrinol., December 1, 1999; 13(12): 2039 - 2048. [Abstract] [Full Text]
|
|
|
|
|
|
R.-H. Chen, M.-C. Chang, Y.-H. Su, Y.-T. Tsai, and M.-L. Kuo Interleukin-6 Inhibits Transforming Growth Factor-beta -induced Apoptosis through the Phosphatidylinositol 3-Kinase/Akt and Signal Transducers and Activators of Transcription 3 Pathways J. Biol. Chem., August 13, 1999; 274(33): 23013 - 23019. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. Datto, J. P. Frederick, L. Pan, A. J. Borton, Y. Zhuang, and X.-F. Wang Targeted Disruption of Smad3 Reveals an Essential Role in Transforming Growth Factor beta -Mediated Signal Transduction Mol. Cell. Biol., April 1, 1999; 19(4): 2495 - 2504. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Hayashi, E. A. McGee, G. Min, C. Klein, U. M. Rose, M. v. Duin, and A. J. W. Hsueh Recombinant Growth Differentiation Factor-9 (GDF-9) Enhances Growth and Differentiation of Cultured Early Ovarian Follicles Endocrinology, March 1, 1999; 140(3): 1236 - 1244. [Abstract] [Full Text]
|
|
|
|
 |
|
 X. Shen, P. P.-c. Hu, N. T. Liberati, M. B. Datto, J. P. Frederick, and X.-F. Wang TGF-beta -induced Phosphorylation of Smad3 Regulates Its Interaction with Coactivator p300/CREB-binding Protein Mol. Biol. Cell, December 1, 1998; 9(12): 3309 - 3319. [Abstract] [Full Text]
|
|
|
|
|
|
C.-F. Lai, X. Feng, R. Nishimura, S. L. Teitelbaum, L. V. Avioli, F. P. Ross, and S.-L. Cheng Transforming Growth Factor-beta Up-regulates the beta 5 Integrin Subunit Expression via Sp1 and Smad Signaling J. Biol. Chem., November 10, 2000; 275(46): 36400 - 36406. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Peron, M. Rahmani, Y. Zagar, A.-M. Durand-Schneider, B. Lardeux, and D. Bernuau Potentiation of Smad Transactivation by Jun Proteins during a Combined Treatment with Epidermal Growth Factor and Transforming Growth Factor-beta in Rat Hepatocytes. ROLE OF PHOSPHATIDYLINOSITOL 3-KINASE-INDUCED AP-1 ACTIVATION J. Biol. Chem., March 23, 2001; 276(13): 10524 - 10531. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. T. Liberati, M. Moniwa, A. J. Borton, J. R. Davie, and X.-F. Wang An Essential Role for Mad Homology Domain 1 in the Association of Smad3 with Histone Deacetylase Activity* J. Biol. Chem., June 15, 2001; 276(25): 22595 - 22603. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Piek, W. J. Ju, J. Heyer, D. Escalante-Alcalde, C. L. Stewart, M. Weinstein, C. Deng, R. Kucherlapati, E. P. Bottinger, and A. B. Roberts Functional Characterization of Transforming Growth Factor beta Signaling in Smad2- and Smad3-deficient Fibroblasts J. Biol. Chem., June 1, 2001; 276(23): 19945 - 19953. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Ahmed, R. A. Alcock, D. Chendil, S. Dey, A. Das, K. Venkatasubbarao, M. Mohiuddin, L. Sun, W. E. Strodel, and J. W. Freeman Restoration of Transforming Growth Factor-beta Signaling Enhances Radiosensitivity by Altering the Bcl-2/Bax Ratio in the p53 Mutant Pancreatic Cancer Cell Line MIA PaCa-2 J. Biol. Chem., January 11, 2002; 277(3): 2234 - 2246. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Pardali, A. Kurisaki, A. Moren, P. ten Dijke, D. Kardassis, and A. Moustakas Role of Smad Proteins and Transcription Factor Sp1 in p21Waf1/Cip1 Regulation by Transforming Growth Factor-beta J. Biol. Chem., September 15, 2000; 275(38): 29244 - 29256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Kaji, L. Canaff, J.-J. Lebrun, D. Goltzman, and G. N. Hendy Inactivation of menin, a Smad3-interacting protein, blocks transforming growth factor type beta signaling PNAS, March 27, 2001; 98(7): 3837 - 3842. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
