β-catenin as a potential key target for tumor suppression

Authors

  • Yuejun Fu,

    Corresponding author
    1. Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan, People's Republic of China
    • Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan 030006, People's Republic of China
    Search for more papers by this author
    • Fax: +86-351-7011499

  • Shuhua Zheng,

    1. Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan, People's Republic of China
    Search for more papers by this author
    • Y.F. and S.Z. contributed equally to this work

  • Na An,

    1. Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan, People's Republic of China
    Search for more papers by this author
  • Takis Athanasopoulos,

    1. Department of Biochemistry, Gene Therapy Laboratory, SWAN Institute of Biomedical and Life Sciences, School of Biological Sciences, Royal Holloway University of London, Egham, Surrey, TW20 0EX, United Kingdom
    Search for more papers by this author
  • Linda Popplewell,

    1. Department of Biochemistry, Gene Therapy Laboratory, SWAN Institute of Biomedical and Life Sciences, School of Biological Sciences, Royal Holloway University of London, Egham, Surrey, TW20 0EX, United Kingdom
    Search for more papers by this author
  • Aihua Liang,

    1. Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan, People's Republic of China
    Search for more papers by this author
  • Ke Li,

    1. Zhejiang Entry-Exit Inspection and Quarantine Bureau, Hangzhou 310008, People's Republic of China
    Search for more papers by this author
  • Changchen Hu,

    1. Department of Neurosurgery, Shanxi Provincial People's Hospital, Taiyuan 030012, People's Republic of China
    2. Department of Neurosurgery, Zhujiang Hospital, Southern Medical University, Guangzhou 510282, People's Republic of China
    Search for more papers by this author
  • Yajing Zhu

    1. Key Laboratory of Chemical Biology and Molecular Engineering of Ministry of Education, Institute of Biotechnology, Shanxi University, Taiyuan, People's Republic of China
    Search for more papers by this author

Abstract

β-catenin is a multifunctional protein identified to be pivotal in embryonic patterning, organogenesis and adult homeostasis. It plays a critical structural role in mediating cadherin junctions and is also an essential transcriptional co-activator in the canonical Wnt pathway. Evidence has been documented that both the canonical Wnt pathway and cadherin junctions are deregulated or impaired in a plethora of human malignancies. In the light of this, there has been a recent surge in elucidating the mechanisms underlying the etiology of cancer development from the perspective of β-catenin. Here, we focus on the emerging roles of β-catenin in the process of tumorigenesis by discussing novel functions of old players and new proteins, mechanisms identified to mediate or interact with β-catenin and the most recently unraveled clinical implications of β-catenin regulatory pathways toward tumor suppression.

Since its identification as an essential and central component in the Wnt signaling cascade, and the subsequent finding of its pivotal role in cadherin-based cell-cell adhesion, β-catenin has been critically studied to elucidate the coordination of these two pathways. The significance of this coordination is substantiated in a plethora of metabolic processes, such as axis and mesoderm formation, stem cell differentiation and carcinogenesis.1, 2 Generally, the Wnt pathway is divided into four branches, namely the canonical Wnt/β-catenin pathway, and the noncanonical (or heretical) Wnt/Ca2+ and planar cell polarity pathways. Amongst them, the canonical Wnt pathway is the best studied and is reported to be highly conserved through evolution but is frequently altered in many human malignancies such as colorectal cancers, hepatocellular carcinomas and gastric cancers.3–6 Ca2+-dependent cell-cell adhesion in the cadherin family of proteins is typified by an extracellular segment that consists of five distinct Ca2+-binding domains and a conserved cytoplasmic domain, which interacts with β-catenin and p120 catenin (herein p120); β-catenin then provides a binding site for α-catenin. Cadherin junctions, among other cell-cell adhesion complexes, are essential for cellular processes such as cell polarity and migration.7, 8 Since an indispensible morphological hallmark of malignant tumors is reduced cell-cell adhesiveness, it is predicted that the components of cadherin junctions in many human malignancies, such as breast cancer, colorectal cancer and prostate cancer, are genetically altered.9–11 Besides cell-cell adhesion, cadherin junctions function as a potent competitor of free cytosolic β-catenin. This is supported by studies of the co-crystal structures of β-catenin/cadherin and β-catenin/TCF revealing that the two β-catenin ligands shared overlapping binding regions along β-catenin.12 The role of the cadherin junction in the subcellular distribution of β-catenin has recently been further extended as discussed below.

In this review, a myriad of recently emerged mechanisms governing the signaling and adhesive activity of β-catenin are presented.

β-catenin in the Canonical Wnt Pathway: Components, Mechanisms and Tumorigenesis

The multiple structural/functional roles of β-catenin as component of the classical cadherin junction and Wnt signaling pathway are depicted in Figure 1. The Wnt pathway is activated by the binding of a Wnt ligand to a seven-pass transmembrane Frizzled (Fz) receptor and its co-receptor LRP6 or LRP5.13 The Wnt induced complex Fz-LRP6 leads to the phosphorylation of LRP6 by CK1 and GSK3β at five conserved PP(S/T)PX(S/T) (hereafter PPPSPXS for simplicity) motifs, located in the intracellular domain of LRP6. The phosphorylated PPPSPXS motif provides an optimal binding site for Axin,14, 15 resulting in the recruitment of Axin and some other associated proteins to the Fz-LRP6 complex.16, 17 Thus, the Axin mediated phospho-degradation complex is deregulated, leading to the accumulation of stabilized cytosolic β-catenin which then travels to the nucleus to form complexes with members of the DNA binding family: TCF-1/lymphoid enhancer factors (LEF-1, 3, 4). The activation of the β-catenin/TCF complex includes the displacement of Groucho and recruitment of histone acetylases CBP/p300 as co-activators (for a review see Ref.19).18, 19 The resultant activated Wnt responsive gene products include some important oncogenic proteins such as Cyclin D1, c-Myc and matrix metalloproteinases (MMP-2, 3, 7, 9, 13).3, 20 And inhibitors such as ICG-001, PKF118-310, PKF115-584 and CGP049090 that interrogate the formation of the transcription complex have been proved efficient in inhibiting tumor progression in preclinical studies.21–24 Of note here, however, nuclear accumulation does not necessary promote tumor progression. It was found in melanoma that elevated Wnt/β-catenin signaling could drive the differentiation-related genes expression in cancer stem cells, resulting in less aggressive and proliferative cancer.25, 26

Figure 1.

An updated representation of β-catenin in classical cadherin junction and Wnt signaling pathway. The figure represents two contiguous cells. When Wnt signaling pathway is inactive, β-catenin is phosphorylated by CK1 and GSK3β sequentially and then targeted for ubiquitination and proteasomal degradation. The binding of Wnt to Frizzled and co-receptors LRP5/6 recruits Axin-GSK3β complex to cell membrane. This results in destabilization of the destruction complex and nuclear accumulation of β-catenin. Then, some Wnt responsive genes are activated. In E-cadherin junctions β-catenin bridges the cytoplasmic region of cadherin with α-catenin, which then interacts with the actin binding protein EPLIN. Interaction between cadherin and microtubule is mediated by proteins Nezha, PLEKHA7 and p120. Axin and LRP5/6 can interact with cadherin via an unknown protein. Also, interaction between tetraspanins CD82/CD9 and E-cadherin promotes the exosomal release of β-catenin. This interacting promotes phospho-degradation of β-catenin. The figure is based partly on previous articles.50, 84, 102, 109, 122 P, phosphate group; UP, unknown protein. Ub, ubiquitination; N, N-terminus; C, C-terminus. GSK3, GSK3β. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

When the Wnt ligand is absent, β-catenin is constantly phosphorylated at the N-terminal four Ser/Thr residues (Ser33, Ser37, Thr41 and Ser45) by the degradation complex, composed of the scaffolding protein Axin, tumor suppressor APC, GSK3β and CK1.27 The process of phosphorylation is conducted in a priming mechanism: β-catenin is obligatory phosphorylated at Ser45 by CK1α before it can undergo further phosphorylation by GSK3β, earmarking β-catenin for ubiquitination-dependent proteolysis by β-TrCP and the 26S proteasome and thereby keep cytosolic β-catenin at a low level.28–31 Consequently, Wnt target genes are repressed by TCF-TLE/Groucho, HDAC and CtBP.32–34 It is noteworthy that several members of the Wnt signal transduction pathway can also be regulated independently by other signaling pathways.

LRP5/6, arrows point the canonical Wnt pathway

LRP5/6 are large type I transmembrane proteins, one of which is required as a co-receptor, upon binding of a Wnt ligand to Fz receptor, to initiate the canonical Wnt pathway.13 As the phosphorylation of intracellular PPPSPXS motifs of membrane-anchored LRP5/6 is well established as an essential step in the transduction of Wnt signals,14, 15, 35 multimodal propositions have been put forward to explain how the LRP5/6 phosphorylation elicits subsequent β-catenin stabilization. The enhanced degradation of endogenous Axin promoted by Wnt signaling through LRP5/6 was reported in cultured mammalian cells and oocytes, embryos of Xenopus.17, 36 Meanwhile, Axin is the limiting factor in β-catenin phospho-destruction complex formation, partly due to its 1,000 times lower concentration relative to other components of the destruction complex.37 Thus, it was previously thought that the degradation of Axin was critical for signal transduction.38 However, this hypothesis is challenged by the fact that Wnt-induced stabilization of β-catenin in cultured mammalian cells occurs about 2 hr before decreased concentration of Axin is detectable.37 A recent postulation is that the intracellular localized Ser/Thr rich clusters near the PPPSPXS motifs have high affinity for GSK3β. Thus, Ser/Thr rich clusters and phosphoprotein Disheveled (Dvl) synergistically recruit the Axin-GSK3β complex to the Fz-LRP5/6 complex, placing GSK3β in the vicinity of PPPSPXS motifs. This complex then specifically inhibits the phosphorylation of β-catenin by GSK3β mediated interactions.39–41 Degradation of Axin may be a secondary, positive feedback mechanism, rather than the primary event that induces β-catenin stabilization.42

The correlation between deregulation of LRP5/6 and carcinogenesis has been substantiated in many cell lines. For example, the elevated expression of LRP6 was found in testicular germ cell tumors, while siRNA engineered against LRP6 was effective in turning off the constitutively activated Wnt pathway, an abnormal autocrine mechanism observed in some human cancers.43, 44 Also in a MMTV-LRP6 transgenic mice model, the overexpression of LRP6 has been demonstrated to be sufficient in inducing nuclear accumulation of β-catenin and the subsequent elevated expression of Wnt responsive genes, with resultant formation of hyperplasia, a precursor of breast cancer.45

Axin, a potent scaffolding protein

Axin, originally identified as the product of the mouse fused gene called “fused,” is a potent multimeric scaffold protein that functions in the stress-activated protein kinase, transforming growth factor β (TGFβ), p53 and Wnt signaling pathways.46, 47 In the canonical Wnt pathway, Axin provides binding sites for APC, GSK3β, CK1 and hypophosphorylation causes destabilization and degradation of Axin.48, 49 Actually, phosphorylation and dephosphorylation of Axin constitute a new level of regulation of the Wnt pathway.50 Besides protein phosphatase 2A (PP2A) and protein phosphatase 2C (PP2C), protein phosphatase 1 (PP1) was recently identified as a positive mediator of Wnt signaling pathway. Dephosphorylation of Axin by PP1 occurs at several CK1-phosphorylated Ser residues, the phosphorylation of which can induce a conformational change of Axin that enhances the affinity between Axin and GSK3.51 Importantly, Dab2, an endocytic adaptor protein, has been identified to block the interaction between Axin and PP1.52 These findings may provide intriguing therapeutic clues to disorders with deregulated β-catenin.

Using Axin as a cargo gene to restore Axin expression in cancer cells was proposed as a promising strategy.46 However, the usage of classical genetic approaches in suppressing hyperactivated Wnt pathway in many human malignancies is hampered by the uncertainty of outcome of adult tissue homeostasis in inhibiting Wnt pathway. Alternatively, small molecules suppressing Wnt pathway are accepted as promising candidates for cancer therapeutics.53 For example, XAV939 was identified to stabilize Axin through tankyrase inhibition.54 Tankyrase is a member of the poly-ADP-ribose polymerase family responsible for PARsylation of target proteins using NAD+ as a substrate.55 In Wnt pathway, tankyrase functions to interact with highly conserved N-terminal domain of Axin and promote degradation and ubiquitination of Axin putatively through PARsylation.54 Also, IWR-1 (inhibitors of Wnt response-1) was identified to bind to and stabilize Axin by a yet unknown mechanism.56 Thus, both XAV939 and IWR-1 could principally downregulate β-catenin by stabilizing Axin. The efficacy of both XAV939 and IWR-1 in inhibiting APC-deficient colorectal cancer cells DLD-1, SW480 in vitro and tailfin and intestinal tissue regeneration in zebrafish in vivo, implicates the role of β-catenin signaling in carcinogenesis.54, 56

APC, an escort for β-catenin degradation

APC is a multifunctional protein essential for processes such as actin elongation, DNA replication, cell cycle transition and DNA demethylation.57–60 The deregulation of any one of these processes constitutes a hallmark of tumorigenesis. A “ratchet” model was proposed in depicting the role of APC in the destruction of the canonical Wnt pathway. It has been reported that the phosphorylated N-terminal of β-catenin represents the recognition site for β-TrCP.61 Phosphorylation of APC by GSK3β and CK1 induces a conformational change that allows the 20-amino acid repeat region of APC to compete with Axin for binding to β-catenin, already phosphorylated by GSK3β and CK1.62 Subsequently, APC specifically protects the β-TrCP recognition site from PP2A-mediated destruction, ensuring that β-catenin is targeted to ubiquitination and degradation.61 Furthermore, in the nucleus, APC was identified to facilitate the recruitment of CtBP to the β-catenin/LEF-1 enhancer complex, with resultant downregulated transcription of Wnt target genes.63

APC dysregulation correlates with a myriad of human malignancies, and the best studied among them is colorectal adenomas.64 Genetic alterations of APC have been well substantiated to be a causative agent of both sporadic and familial (familial adenomatous polyposis, FAP) human colorectal tumors.64–66 The current scenario of APC function predicts that APC alterations and nuclear accumulation of β-catenin occur coincidently and constitute the initial stage for intestinal adenoma formation. Nonetheless, this model conflicts with findings that adenomas taken from FAP patients may lack detectable nuclear accumulation of β-catenin.67, 68 This controversy was resolved only recently. It was proposed that CtBP-1, also regulated by APC, triggers the first step of adenoma initiation when APC is altered, whereas KRAS activation and β-catenin nuclear accumulation functions to promote the deterioration of intestinal cell from adenoma to carcinomas.69 Thus, the initiation and progression of colorectal cancer may happen in a step-by-step manner.

GSK3β, a versatile kinase

GSK3 is a ubiquitous Ser/Thr kinase initially identified to have the capacity to inhibit glycogen synthesis by phosphorylating glycogen synthase. Two isoforms of GSK3, i.e., GSK3β and GSK3α, are found in mammals.70, 71 Recent studies showed that GSK3 was involved in diverse cellular processes including nutrient and energy homeostasis, proliferation and apoptosis, cell fate specification and stem cell self-renewal.72 Some recent seminal findings substantiated that GSK3 acts as a key effector in the mitochondria-mediated intrinsic apoptotic pathway and a number of apoptosis significant transcription and translation factors such as p53, heat shock factor-1 (HSF-1) and Myc, have been identified to be functionally regulated by GSK3.73 Intriguingly, GSK3 was also proven to activate hTERT gene expression and hence contributes to telomere length homeostasis.74 Given that hTERT is essential for telomerase activity, antitumors agents that specifically interrogate GSK3 activity and consequently reduce telomere lengths, may represent a new clinical strategy for tumor suppression. Of note, researchers found that in mouse embryonic stem cell lines lacking various combinations of the GSK3 alleles, increased cytosolic level of β-catenin and elevated target genes expression were observable only when total complement of GSK3 was reduced to one-quarter of normal levels.75 These results clearly indicated some functional redundancy of GSK3α and GSK3β in the phosphorylation of β-catenin in the canonical Wnt pathway. Besides functioning as a major kinase in phospho-degradation of β-catenin, GSK3β plays an essential role in linking the canonical Wnt pathway with PI3K/Akt signaling pathway. PI3K-activated AKT could inactivate/phosphorylate GSK3β (at residue Ser9) in tumor cells and leaded to enhanced nuclear accumulation of β-catenin.76

Given the pleiotropic functions of GSK3, therapeutically promising chemical agents were designed to specifically mediate the activity of GSK3. For example, in vitro studies showed that methylselenic acid exerts antitumor effects by enhancing GSK3β activity in a dose-dependent manner and consequently promotes degradation of β-catenin.77 Nonetheless, a recent clinical trial [Selenium and vitamin E cancer prevention trial (SELECT)] failed to substantiate the efficacy of selenium in preventing prostate cancer, partly due to the formulation and dosage of selenium used.78 Contrarily, enzastaurin that specifically inhibits GSK3, has entered phase II trial for patients with high-grade recurrent gliomas and has achieved, to some degree, encouraging results.79 It is intriguing that enzastaurin induced inhibition of GSK3β and consequent stabilization of β-catenin could elicit inhibitory effects on tumor growth. Intriguingly, it was proposed that the enzastaurin induced cytosolic accumulation of β-catenin was correlated with endoplasmic reticulum (ER) stress-mediated cell growth inhibition and c-Jun up-regulation induced apoptosis.80

Dvl/Dsh, at the crossroads of Wnt signaling branches

There are three Dvl isoforms in mammals named Dvl1, Dvl2 and Dvl3, all with three conserved domains: DIX (Dishevelled, Axin) binding domain located at N-terminus; PDZ (Postsynaptic density 95, Discs Large, Zonula occludens-1) domain in the mid region of Dvl; and DEP (Dvl, Egl-10, Pleckstrin) domain at the midway between the PDZ domain and the C-terminal of Dvl.81, 82 The canonical Wnt signaling is most sensitive to changes in the abundance of either Dvl3 or Dvl1, whereas it is relatively inert to the most highly expressed Dvl2, implicating some overlapping and distinct features of the proteins.83 Dvl functions in both Wnt/β-catenin signaling and other Fz/LRP dependent pathways and is generally considered as the branching point of β-catenin and noncanonical pathways, although the mechanisms underlying the signal specificities remain enigmatic.84 In the canonical Wnt pathway, Dvl facilitates the recruitment of Axin-GSK3β complex to Wnt induced Fz-LRP complex and the ensuing phosphorylation of LRP5/6, with resultant amplification of Wnt signals.85 Recent endeavors uncovered some novel functions of Dvl. Dvl3 was identified to associate with KSRP, an RNA-decay-promoting protein already substantiated to interact with the 3′-UTR of β-catenin mRNA in a PI3K-Akt mediated manner.86, 87 Stimulation of Wnt3 releases the β-catenin mRNA from Dvl3-KSRP complex, leading to enhanced half-life of the transcript.87 Upon stimulation of Wnt signals, Dvl may function to recruit Akt which phosphorylates KSRP at Ser193, resulting in impaired decay-promoting activity of KSRP.86, 87 The identification of NES and nuclear accumulation of Dvl has driven rigorous study on the nuclear function of Dvl.88 Dvl was identified to interact with TCF, bridged by the transcriptional factor c-Jun. The resultant Dvl/c-Jun/β-catenin/TCF complex was justified to be necessary for efficient transcription of Wnt responsive genes.89

Elevated expression of ATDC was observed in a number of human cancer types such as bladder, breast and endometrial cancers.90–92 However, it is only recently that a putative model was proposed in correlating tumorigenesis and deregulation of ATDC. The oncogenic function of ATDC was elicited by interacting with and stabilizing Dvl2, resulting in the dissociation of β-catenin from the destruction complex.92 Furthermore, a recent study pointed out that Dvl was fine-tuned by the process of ubiquitination and deubiquitination. K63-linked hyper-ubiquitination of the DIX domain of Dvl came along with enhanced Wnt signaling in CYLD mutated human cylindroma skin tumors. The enzyme CYLD exerted its antitumor function by interacting with and regulating K63-linked ubiquitination of Dvl.93 This finding may partially explain why CYLD loss is frequently observed in diverse tumor cell lines.94

Wnt Pathway and Cadherin Junctions: Friends or Foes?

The classical cadherin family is composed of approximately 20 members such as E-cadherin, N-cadherin and P-cadherin, all of them share a common domain organization.95 Ca2+ ions act as a switch in modulating the “on” and “off” conformation of the extracellular domain of cadherin, constituted with five repetitive subdomains, called cadherin repeats or EC domain.7 Adhesive specificity of classical cadherins was well identified as pivotal for the patterning of multicellular organisms, in a manner that the degree of selectivity differs with different partners.96, 97 However, the precise explanation of how Ca2+ induced EC domain would homophilically interact with each other remains lacking. A recent model implicated that trans binding of EC1 domains initiated the interaction, followed with strengthening lateral clustering of cadherins.98 The previously prevailing scenario that α-catenin acted as a bridge in linking cadherin/β-catenin complex with actin cytoskeleton was challenged by the finding that α-catenin could not bind β-catenin and actin filaments simultaneously in vitro, although the interaction between cadherin junctions and actin cytoskeleton was well substantiated to be essential for epithelial morphogenesis such as apico-basal polarity.99, 100 One possible reconciliation of this controversy is that some mediators such as EPLIN may bridge α-catenin and actin filaments.101, 102 Furthermore, a pleckstrin homology (PH) domain protein called PLEKHA7 and Nezha (formerly KIAA1543) were identified to bridge the minus ends of noncentrosomes anchored microtubules with p120, which played an essential role in stabilizing cadherin-catenin mediated cell-cell adhesion complex.103, 104 The role of β-catenin in cadherin junction is versatile. Besides functions as a bridge in cadherin/β-catenin/α-catenin complex, β-catenin is required to facilitate cadherin transport from the Golgi apparatus to the plasma membrane.105, 106 Also, β-catenin is essential in stabilizing cadherin junction structurally or by physically interacting with other factors.107 For example, vinculin, an essential component of cell-matrix, was identified to bind β-catenin in cadherin junction and was evidenced to strengthen cadherin-mediated cell-cell adhesion, offering a putative explanation for the observation that vinculin was frequently absent from adhesion sites in cancers originating from epithelium cells.108

β-catenin in linking cadherin junction and Wnt pathway

It is intriguing that in vertebrates a single protein β-catenin functions in both cadherin junction and Wnt signaling. Actually, elevated nuclear accumulation of β-catenin by its dissociation from cadherin junction was frequently observed during tumorigenesis and may reveal some promising therapeutic targets.109 For example, epidermal growth factor (EGF) stimulated Akt dependent phosphorylation of β-catenin at Ser552in vivo and in vitro results in dissociation of β-catenin from cadherin junctions and subsequent enhanced nuclear and cytosolic localization of β-catenin.110 Similarly, Rac1, a member of Rho family, was shown to activate JNK2 which in turn phosphorylated β-catenin at Ser191 and Ser605, and the resultant phosphorylated β-catenin had strong propensity to be translocated to the nucleus.111 Cellular Src (c-Src) represents the major kinase that phosphorylate β-catenin at tyrosine residues and constitutive tyrosine activation promoted the dissociation of β-catenin from cadherin junctions, leading to reduced cell-cell adhesion and enhanced transcription of Wnt target genes.112, 113 Bosutinib (SKI-606), an inhibitor of Src family kinases now in clinical trials in the treatment of many human cancer types, was postulated to elicit its antitumor function by modulating the phosphotyrosine status of β-catenin and consequently altered signaling and adhesive activities.114–117

Of note, it was also proposed that Wnt signaling might generate a TCF selective monomeric form of β-catenin, whereas β-catenin/α-catenin complex preferentially interacted with cadherin.118 Furthermore, the priming of phosphorylation by CK1 at Ser45 was proposed to generate a signaling specific form of β-catenin that preferentially traveled to the nucleus and activated transcription but was unable to interact with cadherin.119 Interestingly, the intersection between Wnt signaling pathway and cadherin junction may happen in a seemingly β-catenin independent manner. Axin mediated phospho-destruction complex was indentified to colocalize with cadherin junctions and this interaction promoted the N-terminal phosphorylation of β-catenin, leading to enhanced turnover of cytosolic β-catenin.120, 121 Interestingly, a recent study identified that over-expression of tetraspanin CD82 or CD9 suppressed Wnt signaling in a GSK3β and proteasome independent manner. The implication was that interaction between E-cadherin and tetraspanins facilitated the exosomal release of β-catenin and consequently suppressed Wnt signaling.122 These intriguing findings may well indicate some promising therapeutic targets for tumor suppression.

β-catenin in tumor metastasis: EMT transition and E-cadherin junction

The process of epithelial-mesenchymal transition (EMT) was critical for regulated embryonic development and for epithelial-derived tumors to become invasive and metastasize.123 Both developmental and oncogenic EMT are associated with the loss of apico-basal polarity, destabilization of intercellular adhesion complexes (gap junctions, desmosomes, tight junctions and adherens junctions) and replacement of epithelial-cell markers (E-cadherin, β-catenin) by mesenchymal-cell markers (N-cadherin, vimentin), endowing cells with migratory and invasive properties.124, 125 In early organisms, dynamic interconversion between epithelial and mesenchymal cells provides plasticity and mobility that are essential for organogenesis.126 In human pathology, however, inappropriately expressed developmental EMT regulators such as TGF-β, Six1 and Twist induce stem cell properties, resistance to apoptosis and senescence, and initiation of metastasis in tumor cells.127, 128 Of note, some recent findings and hypotheses questioned the precise role of EMT in tumor metastasis. For one thing, the oncogenic EMT transition was also reversible, contradictory with previous depiction of EMT as an irreversible, permanent process. During the initiatory stage of cancer metastasis, EMT endows carcinoma cells with migratory properties that facilitate local invasion and dissemination. However, when tumor cells reach target organs, mesenchymal to epithelial reverting transition (MErT) may be activated and the resultant epithelial phenotype would contribute to their ‘seeding’ in those distant organs.129, 130 But how can one argue that the initiatory activation and late stage reversion of EMT would lead to an instant acquisition of therapeutic value in targeting metastatic cancer? The answer lies in the possibility of identifying “ideal” prognostic markers or signaling pathways that will allow smart detection of metastatic lesions at an early stage or late stage of tumor progression, which conceivably will increase disease free survival rates. The most recently documented relationship among EMT, MErT and metastatic progression implicated E-cadherin re-expression as a critical indicator of MErT.129–132 But why and how would metastatic cells re-expressed E-cadherin in secondary, ectopic microenvironment? Regarding the former, it was hypothesized that re-expression of E-cadherin would enhance the heterophilic ligation between epithelial-like cancer cells and normal parenchymal cells, and thus surmount the metastatic inefficiency by facilitating extravasation and colonization of distant organs.129, 133 A number of studies offer support for this speculation. For example, human prostate carcinoma cells can form heterotypic adhesions with rat hepatocytes when E-cadherin re-expression is induced, whereas, intriguingly enough, when they were co-cultured prostate carcinoma cells would increase E-cadherin expression.134, 135

Given the complexity of mediators, factors in the target organ microenvironment plus that tumor metastasis may be organ-specific, mechanisms responsible for E-cadherin re-expression are unlikely to be functionally apparent. However, promisingly, in some tumor cell lines β-catenin was found to act as a major determinant in both EMT and MErT. Mechanistically, these observations are conceivable since that although β-catenin is necessitated to maintain robust cell-cell adhesion in epithelial cells by mediating E-cadherin junction, its transcriptional activity would induce Slug or Twist gene expression that could represses E-cadherin expression by binding to its promoter.136, 137 Also most recently identified convergence between Wnt/β-catenin and other EMT, MErT regulatory signaling pathways, such as the TGF-β, epidermal growth factor receptor (EGFR), Snail and DAB2IP (also known as ASK1-interacting protein-1) further highlighted the significance of β-catenin in metastatic progression138–142 (Fig. 2).

Figure 2.

The role of β-catenin in neoplastic metastasis. The figure represents three contiguous cells linked by E-cadherin junctions. The convergence between Wnt/β-catenin signaling with other pathways highlights its role in EMT transition. WNTs induced inhibition of GSK-3β would lead to nuclear accumulation of β-catenin, resulting in up-regulated expression of MMPs, Twist and Slug. Among these MMPs, MMP-3 and MMP-9 could cleave E-cadherin. As a result, membrane-bound β-catenin is released, E-cadherin is destabilized and Rac 1b is produced. Rac 1b then induce production of ROS from mitochondria, which will induce the expression of Snail. Twist, Slug and Snail can down-regulate E-cadherin expression by binding to its promoter (1). DAB2 IP can reduce the transcriptional activity of TCF/β-catenin complex and enhance the activity of GSK3β and consequently, inhibit EMT (2). Epidermal growth factors can induce the phosphorylation of E-cadherin bound β-catenin and thus relocate β-catenin from membrane to cytoplasm. EGFR can also induce the phosphorylation and deactivation of GSK3β (3). In response to TGF-β, Smad-mediate transcription of Snail can inhibit E-cadherin expression. Also, in non-Smad-mediated signaling, TGF-β can activate PI3K/Akt signaling which then deactivates GSK3β, resulting in β-catenin accumulation (4). P, phosphate group; GSK3, GSK3β; the dotted line indicates that the underlying mechanism is not well identified. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Furthermore, β-catenin mediated transcription of secreted MMPs (MMP-2, 3, 7, 9, 13) enhance EMT transition mainly by degrading extracellular matrix that subtends epithelial cells and thus confer tumor cells with migratory and invasive activities.143 Mechanistically, however, when and how MMPs would facilitate EMT still await elucidation. In this aspect, the functioning role of MMP-3 in EMT was relatively well studied. Treatment of MMP-3 in SCp2 mouse mammary epithelial cells caused expression of a highly activated splice isoform of Rac1 termed Rac1b, resulting in elevation of cellular ROS which then stimulated Snail expression and contributed to EMT transition.144 Intriguingly, since MMP-9 could also initiate the production of Rac 1b, it was extrapolated that the proteolytic activity of MMPs caused cleavage of E-cadherin, leading to the relocation of β-catenin from membrane to the nucleus in which it facilitated the alternative splicing of Rac 1b in an unknown manner.144 (Fig. 2).

Conclusion and Perspective

The work presented here emphasizes the significance of β-catenin in tumorigenesis and tumor metastasis from the aspects of its signaling pathways and adhesive functions. During last two decades, we see a surge of researches unraveling the significance of β-catenin in cancer progression and a number of therapeutically significant agents have been developed. However, we have to admit that clinical application of genetic or pharmaceutic agents targeting β-catenin remains lacking. Future endeavors that uncover those enigmatic mechanisms such as the balance between the signaling and adhesive activity of β-catenin and its delicate subcellular localization may pave new avenues in tumor suppression.

Ancillary