Crosstalk between Wnt/β-catenin and Hedgehog/Gli signaling pathways in colon cancer and implications for therapy

Abstract

Wnt/β-catenin and Hedgehog/Gli signalings play key roles in multiple biogenesis such as embryonic development and tissue homeostasis. Dysregulations of these 2 pathways are frequently found in most cancers, particularly in colon cancer. Their crosstalk has been increasingly appreciated as an important mechanism in regulating colon cancer progression. Our studies into the link between Wnt/β-catenin and Hedgehog/Gli signalings in colon cancer revealed several possible crosstalk points and suggested potential therapeutic strategies for colon cancer.

Keywords:

• β-catenin
• colon cancer
• crosstalk
• Gli
• Hedgehog/Gli signaling
• therapeutic targets
• Wnt/β-catenin signaling

Introduction

Colon cancer is one of the major malignancies worldwide. Its pathogenesis is complex and requires the accumulated alteration of multiple genes and pathways. More than 80% of colon cancer cases harbor aberrant Wnt/β-catenin signaling that regulate colon cancer progression through intracellular mechanism or interaction with tumor microenvironment and cancer stem cells. In colon cancer, Wnt/β-catenin signaling is activated via genetic or epigenetic changes of its components, and further derived by oncogenes or tumor suppressors outside canonical signaling pathway. In addition to Wnt/β-catenin signaling, Hedgehog/Gli signaling has also attracted increasing attention due to its important roles in colon tumor maintenance. Hedgehog/Gli signaling is activated by canonical and non-canonical signaling routes in colon cancer. Accumulating evidences have suggested that Hedgehog/Gli signaling acts on colon cancer in an autocrine, paracrine or cancer stem cell fashions. Recent studies have revealed that Wnt/β-catenin and Hedgehog/Gli signalings are not independent cascades in colon cancer. In light of this, we focused our studies on the current advancements on their correlation which, more importantly, could lead to potential strategies for colon cancer therapy.

Wnt/β-catenin Signaling Pathway

The discovery of Wnt gene in Drosophila and its homologous gene in vertebrates laid the foundation for an evolutionary conserved signaling cascade which is commonly referred to as canonical Wnt signaling pathway. In the absence of wnt ligands, β-catenin usually locates in cell membrane where it interacts with E-cadherin and mediates intercellular adhesion.¹ Free cytoplasm β-catenin is destabilized by a destruction complex including adenomatous polyposis coli (APC), Axin, glycogen synthase kinase-3β (GSK3β) and casein kinase 1α (CK1α).^2,3 When wnts interact with Frizzled and LRP, Wnt/β-catenin signaling is activated. Both Dishevelled (Dvl) and Axin are recruited to cell membrane, followed by GSK3β inhibition and β-catenin releasing from a destruction complex. β-catenin is transported into cell nucleus and forms complex with T-cell transcription factor (TCF)/lymphoid enhancer-binding factor (LEF). In the participation of coactivatiors cAMP response element-binding protein (CREB)-binding protein (CBP), a series of target genes are expressed, which include proteins associated with cell proliferation, apoptosis, invasion and angiogenesis.

Wnt/β-catenin Signaling and Colon Cancer

Human hereditary and sporadic colon cancers usually develop from the earliest aberrant crypt foci to larger adenomas, which will progress to carcinoma and invasive adenocarcinomas. Genetic or epigenetic changes of Wnt pathway components are frequently found in most colon cancer cases which lead to aberrant Wnt/β-catenin signaling (Fig. 1). Found in about 80% of colon tumors, APC mutations are one of the earliest events in colon cancer progression.^4,5 APC mutations usually occur in a mutation cluster region that is important for binding to Axin and forming destruction complex, and lead to produce truncated form of APC as well as the stabilization of β-catenin. In addition, wildtype APC, containing both nuclear localization sequences (NLS) and nuclear export sequences (NES), acts to shuttle β-catenin between cell nucleus and cytoplasm. Lacking nuclear export sequences (NES), truncated APC is incompetent for β-catenin shuttling. Among the 10% of colon cancer cases that contain wildtype APC, β-catenin is mutated by point or in frame deletion of serine and threonine residues. These mutations render β-catenin difficult for ubiquitination and degradation by cellular proteosomes. Moreover, Axin mutation is found in some microsatellite instability (MSI) colon tumor cases. Axin is an important component of destruction complex and its inactive mutations could disturb the process of β-catenin degradation. Epigenetic alterations have occurred in colon cancer. Components of Wnt/β-catenin signaling, such as sFRP1, WIF-1 and DKK1, are silenced by hypermethylation, which contributes to aberrant Wnt/β-catenin signaling in colon cancer.⁶ While colorectal adenomas progressing to invasive carcinoma, further genetic changes in oncogenes or tumor suppressors are found to aggravate the activation of Wnt/β-catenin signaling (Fig. 1). Oncogenic activation of KRAS is frequently observed in colon cancer which can increase β-catenin stability through inhibiting GSK-3β by KRAS/PI3K/AKT signaling.⁷ Inactive mutation of tumor suppressor p53 is also found in colorectal cancer and related to the stabilization of β-catenin.^8,9 Loss expression of phosphatase and tensin homolog (PTEN) has been associated with colorectal cancer's aggressive capacity.^10-12 PTEN knockdown can increase the accumulation of β-catenin in cell nucleus.^10,11

Figure 1. The mechanism that result in aberrant Wnt/β-catenin signaling in colon cancer. There include genetic like inactivating and activating mutations or epigenetic changes like DNA hypermethylation of Wnt pathway components; Meanwhile, another cause involves genetic and/or expression changes of oncogenes or tumor suppressors outside Wnt pathway.

Active Wnt/β-catenin signaling stimulates colon cancer development through its downstream cancer-related targets such as c-myc, Cyclin D1, Cox-2, MMPs, uPAR, VEGF, etc. In addition to intracellular mechanism, Wnt/β-catenin signaling also plays an important role in colon cancer progression by interacting with tumor microenvironment and cancer stem cells. In tumor microenvironment, growth factors secreted by stromal cells could enhance the activity of Wnt/β-catenin. Hepatocyte growth factor receptor (HGFR) forms a complex with β-catenin.¹³ When HGF binds to HGFR, β-catenin is dissociated from HGFR/β-catenin complex, followed by tyrosine phosphorylation. Finally, Wnt/β-catenin activity is enhanced in the same manner as KRAS activation mutation. Other factors such as platelet-derived growth factor (PDGF), epidermal growth factor (EGF) and transforming growth factor β (TGF β) can enhance Wnt/β-catenin signaling by p68.¹⁴ PDGF, EGF and TGF β promote tyrosine phosphorylation of p68, which inhibits β-catenin phosphorylation. Cancer stem cells (CSCs) are small parts of cancer cells, which can self–renew and differentiate to generate many types of tumor cells. In 2007, O’Brien and Ricci-Vitiani discovered colon cancer stem cells.^15,16 High Wnt activity can designate colon cancer stem cell population.¹⁷

Hedgehog/Gli Signaling Pathway

Firstly identified in Drospphila melanogaster, Hedgehog/Gli signaling derives its name from an intercellular signaling ligand called Hedgehog. In vertebrates, there are at least 3 Hh ligands: Sonic hedgehog (Shh), Indian Hh (Ihh) and Desert Hh (Dhh). Lipid modified Hh ligands are secreted in the participation of Dispatched protein (Disp) and then interact with reception system consisting of Patched (PTCH) and Smoothened (SMO). SMO is a crucial signal transducer of Hedgehog pathway, signaling to a cytoplasmic transduction cascade that ultimately regulates transcriptional activators Gli proteins. There are 3 different Gli proteins: Gli1, Gli2 and Gli3. Gli1 acts as a transcriptional activator. Gli2 and Gli3 are bifunctional transcription factors which exist in 2 forms: their full-length ones as transcriptional activators and truncated N-terminal fragment as transcriptional repressors.¹⁸ In the absence of Hh ligands, PTCH localizes to primary cilium and blocks SMO activity by preventing SMO from entering the cilium.^19,20 Consequently, Gli transcription factors in cilia are phosphorylated by PKA, GSK3β and CK1α, followed by cleavage to form Gli repressor and inhibiting target genes transcription.²¹ When hedgehog ligands bind to PTCH, SMO is transported to cilia. SMO interacts with suppressor of fused (Sufu) to relieve its inhibitory effect on Gli.^22,23 When activated, Gli triggers the transcription of target genes.

Abnormal Hedgehog/Gli Signaling in Colon Cancer

It has been suggested that aberrant Hedgehog/Gli signaling plays a crucial role in colon cancer progression. The persistent activation of Hedgehog/Gli signaling in colon cancer is driven in a ligand-dependent manner, which is mediated through either canonical or non-canonical signaling routes (Fig. 2). Hedgehog/Gli signaling related components, Shh, PTCH1, SMO and Gli, are observed to be over-expressed in colon tumor.^24-27 A positive correlation between tumor progression and expression levels of Shh, PTCH1 or SMO has been suggested.²⁷ In non-canonical signaling routes, increased Gli function is regulated by multiple oncogenes and tumor suppressors, which consequently leads to aberrant Hedgehog signaling during colon carcinogenesis. Oncogenes BRAF are frequently observed with activating mutations in colon cancer.²⁸ Oncogenic KRAS mutant could promote Gli1 by downstream RAF/MEK/ERK and PI3K/AKT signaling cascades.^29,30 In addition, inactive p53 and lost PTEN expression have been proven to result in up-regulated Gli1 activity in colon cancer cells.³⁰

Figure 2. The mechanism that result in aberrant Hedgehog/Gli signaling in colon cancer. There include canonical and non-canonical signaling routes. Canonical routes involve in the over-expression of Hh ligand, PTCH1, SMO and Gli. Non-canonical signaling routes include genetic or expression changes of multiple oncogenes and tumor suppressors such as p53, KRAS, BRAF and PTEN.

The mechanisms of aberrant Hedgehog/Gli signaling associated with cancer are diverse.³¹ In colon cancer, the precise mechanism is still unclear. The current understanding of the role of Hedgehog signaling in colon cancer can be summarized below. First, Hedgehog signaling acts on colon cancer in an autocrine manner. Tumor cell secretes Hh ligands and then binds itself to trigger the transcription of target genes, which stimulates proliferation and survival of tumor cells. This view is supported by the observation that the hyperplastic polyps, adenomas and adenocarcinomas of colon express Shh, PTCH1 and SMO.^32,33 Furthermore, cyclopamine, an inhibitor of Hedgehog signaling, inhibits the growth of colon cancer cell lines.³⁴ Active Hedgehog signaling is expressed in tumor epithelial cells rather than in the stroma.³⁵ Secondly, Hedgehog signaling acts in a paracrine fashion. Tumor cell secretes Hh ligands, which stimulates the surrounding stroma to produce a mitogen. The mitogen then promotes tumor cell growth and survival. This has been supported by data that suggest both epithelium and mesenchymal cells in gut express Hedgehog signaling.^32,33 The xenografts of tumor cell are inhibited by Hedgehog signaling antagonist in the stroma.³⁶ Finally, Hedgehog signaling is important for the maintenance of cancer stem cells. Studies showed that Hedgehog signaling activation is essential to stem cell survival and expansion.³⁵

Crosstalk between Wnt/β-catenin and Hedgehog/Gli Signalings in Colon Cancer

While Wnt/β-catenin and Hedgehog/Gli signalings are crucially involved in colon cancer processes, crosstalk between them has been identified to be important for colon cancer recurrence, invasion and metastasis.^30,37-42 To summarize, possible molecular crosstalk between 2 pathways include the following (Fig. 3):

Figure 3. Crosstalk between Wnt/β-catenin and Hedgehog/Gli signalings in colon cancer. Firstly, both pathways are regulated by common modulators such as GSK3β, CK1α, Sufu, p53, PTEN, SMO and KRAS; Secondly, β-catenin interacts with Gli1 by Gli3R and their downstream targets (such as CBD-BP and c-myc for β-catenin; Hh, Snail, sFRP1 and Wnt for Gli1). Finally, β-catenin and Gli1 have antagonistic role in regulating TCF and its downstream target genes.

First, both Wnt/β-catenin and Hedgehog/Gli signalings are regulated by molecules such as GSK3β, CK1α, Sufu, p53, PTEN, SMO and KRAS. Two protein kinases GSK3β and CK1α negatively regulate both β-catenin and Gli1. In Wnt/β-catenin signaling, GSK3β, CK1α and β-catenin bind to Axin. Axin mediates β-catenin phosphorylation in Ser 45 by CK1α, which allows GSK3β to phosphorylate Thr 41, Ser 37 and Ser 33 of β-catenin.^2,3 Phosphorylated β-catenin is recognized, ubiquitinylated by β-TrCP and ultimately degraded.⁴³ Furthermore, in Hedgehog/Gli signaling both GSK3β and CK1α phosphorylate full-length Gli3.⁴⁴ Phosphorylated Gli3 is recognized by β-TrCP and leads to the degradation of C-terminal peptides to generate Gli3R, which in turn inhibits Gli1 activity.⁴⁵ Sufu, as a suppressor of the Fused kinase, is a negative regulator of both Wnt/β-catenin and Hedgehog/Gli signaling pathways. Sufu not only interacts with Gli1 but also binds to β-catenin to control their nuclear–cytoplasmic distributions.^42,46,47 In addition, inactive mutation of p53 or loss of PTEN in colon cancer could stimulate both pathways by activating β-catenin and Gli1.^10,30 Studies have shown that high levels of functional p53 could lead to a reduction of the amount and transcriptional activity of β-catenin.^8,9 PTEN knockdown can increase the stabilization of β-catenin through activating PI3K/AKT pathway.^10,11 For Hedgehog/Gli signaling, knockdown of p53 increases Gli1 activity whereas over-expression of p53 inhibits Gli1 activity in colon cancer cells.³⁰ Either enhanced levels of PTEN or inhibition of downstream AKT could disturb Gli1 activity.³⁰ SMO is an upstream active factor of Gli1 in Hedgehog signaling.¹⁹ Studies have shown an inhibition of SMO could reduce protein levels of active β-catenin and induce its nuclear exclusion, which is independent on the effect of Gli.³⁹ Activated KRAS could stimulate the action of 2 pathways in colon cancer.^{7,29,30,48,49} KRAS mutation up-regulates β-catenin through the inhibition of GSK-3β mediated by PI3K/AKT signaling.⁷ Studies have shown that the inhibitor U0126 for RAS/RAF/MEK/ERK signaling could repress Gli1 activity in human colon cancer cells.^29,30 Overexpression of both KRAS mutants and the downstream component AKT enhance Gli1 activity in colon cancer cells.³⁰

Secondly, β-catenin interacts with Gli1 by Gli3R and each downstream target. A few studies have demonstrated the interaction between β-catenin and Gli1 in colon cancer in multiple ways. In colon cancer, Gli3R displays mutual antagonism with β-catenin and Gli1 respectively.^30,35,50 Gli3R inhibits β-catenin activity through interacting with the C-terminal transactivation domain of β-catenin and thereby antagonizes the active forms of β-catenin.⁵⁰ The effects that Gli1 enhancing and Gli3R repressing cell proliferation are mutually rescued in co-transfected cells.³⁵ An over-expression of Gli1 decreases the mRNA levels of Gli3R, and the transcription of Gli1 is also repressed by the enhanced Gli3R expression.³⁰ Thus, β-catenin and Gli1 could regulate each other by Gli3R. In addition, β-catenin could regulate Gli1 by downstream targets of CBD-BP and c-myc. β-catenin induces the expression of an RNA-binding protein CBD-BP, which in turn binds to the coding region of Gli1 mRNA and stabilizes it.³⁸ Enhanced expression of c-myc, a key target of β-catenin, by transfection of plasmids containing c-myc cDNA in colon cancer cells induces elevate Gli1 mRNA levels.³⁰ Moreover, c-myc overexpression could repress Gli3, which forms Gli3R and then exerts mutual antagonism with Gli1.³⁰ Gli1 also regulates β-catenin through its targets: Snail, Wnt, Shh and sFRP1. Gli1 could induce Snail expression, which then interacts with β-catenin and stimulates its transcriptional activity.^51,52 Wnt2b, Wnt4 and Wnt7b are shown to be target proteins of Gli1⁵³; while they are the upstream ligands of Wnt/β-catenin signaling, they could trigger downstream cascades to promote the stability of β-catenin. Another target protein, sFRP1, is an antagonist of Wnt ligands which indirectly inhibits the activity of downstream β-catenin.^54,55 In addition, Gli1 could up-regulate Shh expression, which is secreted and acts on stromal cells. Stromal cells responding to Shh enhance Foxf1 and Foxf2 expression, which inhibits mesenchymal expression of Wnt5a and leads to suppression of β-catenin⁵⁶.

Finally, β-catenin and Gli1 have antagonistic roles in regulating TCF and downstream target genes in metastatic colon cancer. Even if β-catenin is still highly expressed in metastatic colon cancer, TCF's targets are decreased in contrast to that in non-metastatic stage.³⁰ In addition, studies have suggested enhanced Gli1 could attenuate TCF activity and its downstream targets in colon cancer cells.³⁰

Implications for Colon Cancer Therapy

The crosstalk between Wnt/β-catenin and Hedgehog/Gli signalings plays an important role in colon cancer and hence is an interesting therapeutic target. The therapeutic strategies may include the following 2 aspects:

(1) Targeting common regulators of both Wnt/β-catenin and Hedgehog/Gli signalings. Targeting common activators or suppressors that lead to a blockage of both pathways is theoretically effective to colon cancer suppression. The potential targets are described as the following: i. SMO. Cyclopamine is a natural Hedgehog pathway antagonist and acts by binding to and inhibiting SMO. Cyclopamine treatment could induce colon cancer cells apoptosis in vitro and rescue the lethality and intestinal phenotypes owing to APC loss in mice.^34,35,37 Vismodegib (GDC-0449), a cyclopamine derivative, could delay colon tumor growth in mice.³⁶ ii. KRAS. KRAS suppression by antisense oligonucleotides decreases colon cancer cell viability.⁵⁷ Inhibitor of KRAS membrane localization, farnesyltransferase (FTS) could inhibit colon cancer cell growth by reducing KRAS and up-regulating p53 and p21.⁵⁸ In addition, some inhibitors of downstream RAF-MEK-ERK/MAPK cascade of KRAS such as PD 098059, U0126, and CI-1040 have shown antagonistic effects on colon cancer cells.^29,30,59 iii. CK1α and GSK3β. Small molecular pyrvinium could activate CK1α, which in turn suppresses Wnt signaling and colon cancer cell proliferation.⁶⁰ PX-316 that indirectly activates GSK3β by repressing upstream AKT could present antitumor activity against colon cancer in mice.⁶¹ iv. p53. Restoring p53 function have been developed as a novel therapeutic approaches.⁶² In colon cancer, adenovirus mediated p53 gene transfer could induce apoptosis and inhibit growth and angiogenesis.⁶³ In addition, overexpression of Sufu, PTEN or Gli3R has shown anti-tumor effects on colon cancer cells.^35,42,64 So they may be developed to be effective therapeutic molecules for colon cancer.

(2) Combined inhibition of Wnt/β-catenin and Hedgehog/Gli signalings

In addition to common regulators, there are independent mechanism that mediates Wnt/β-catenin and Hedgehog/Gli signalings respectively (Fig. 3). The combination of inhibitors that separately disturb Wnt/β-catenin or Hedgehog/Gli signaling may be a feasible approach for colon cancer treatment.

The potential targets of Wnt/β-catenin signaling include (Fig. 3): i. Upstream mediators of β-catenin: SFRP, APC and Axin. Re-expression of SFRP in colon cancer cells leads to attenuated Wnt signaling as well as cell death.⁶⁵ Recombinant adenovirus (Ad-CBR) that constitutively expresses the central third of APC could block nuclear translocation of β-catenin and inhibit β-catenin/TCF-4-mediated transactivation, which consequently induces colon cancer cells apoptosis.⁶⁶ Adenovirus mediated gene transfer of Axin could induce colorectal cancer cells apoptosis.⁶⁷ These data suggest SFRP, APC and Axin may be effective therapeutic molecules for colon cancer. ii. β-catenin. Because of the elevated β-catenin in most of colon cancers, approaches such as decreasing β-catenin expression by antisense, RNA interference, protein knockdown and disturbing β-catenin/LEF/TCF nuclear complex have been suggested.⁶⁸ The use of β-catenin antisense oligonucleotides in colon cancer decreases β-catenin level, TCF transcrption and Cyclin D1 expression. Meanwhile, there is a reduction in cell proliferation and invasiveness in vitro and inhibition of tumor growth in vivo.^69,70 Expressing siRNA targeting β-catenin decreases TCF transcription and leads to an increase in G1 cell cycle arrest and promotion of cell differentiation⁷¹. Cong et al. constructed a chimeric protein with β-catenin binding domain of E-cadherin fused to β-TrCP ubiquitin protein ligase (β-TrCP-E-cadherin), which can knock down cytosolic β-catenin.⁷² Expression of β-TrCP-E-cadherin can affect growth and clonogenic ability in colon cancer cells and lead to lose their tumorigenic ability in nude mice.⁷² Emami and his colleagues screened an inhibitor called ICG-001, which disturbs β-catenin/LEF/TCF nuclear complex by specifically targeting the coactivator CBP.⁷³ This agent could reduce colon cancer growth both in vitro and in vivo.⁷³ iii. Downstream effectors of β-catenin. Because downstream targets of β-catenin play an important role in colon cancer development, targeting these genes can be effective for tumor therapy. COX-2 inhibitors such as NSAIDS have been proven to suppress colon cancer.⁷⁴

The important role of Hedgehog/Gli signaling in various cancers has caused attention on targeting it. According to the crosstalk between 2 pathways, the potential targets apart from above common regulators include (Fig. 3): i. BRAF. Inhibition of BRAF by RNA interference increases apoptosis in colon cancer cells.⁷⁵ ii. Gli1. A few small molecular inhibitors for Gli1, GANT61, have been identified. Treatment with GANT61 induces DNA damage and cell death in colon cancer cells.^29,76

Concluding Remarks

Both Wnt/β-catenin and Hedgehog/Gli signalings are frequently activated in colon cancer. These 2 pathways have been shown to coordinate or cross-regulate at multiple aspects during colon cancer development. Despite understanding of the interaction between Wnt/β-catenin and Hedgehog/Gli signalings, detailed molecular mechanisms need to be clarified. The development and identification of selective inhibitors based on the crosstalk between Wnt/β-catenin and Hedgehog/Gli signalings provide approaches for preventing and treating colon cancer progression.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Funding

This work was supported by National Natural Sciences Foundation of China (No. 31271516, No. 21207084); Research Fund for the Doctoral Program of Higher Education of China (20111401110011)，China Postdoctoral Science Foundation (2012M521178) and Natural Sciences Foundation of Shanxi (2014011027–5).