Wnt/β-catenin pathway in tissue injury: roles in pathology and therapeutic opportunities for regeneration

Abstract

The Wnt/β-catenin pathway is an evolutionarily conserved set of signals with critical roles in embryonic and neonatal development across species. In mammals the pathway is quiescent in many organs. It is reactivated in response to injury and is reported to play complex and contrasting roles in promoting regeneration and fibrosis. We review the current understanding of the role of the Wnt/β-catenin pathway in injury of various mammalian organs and discuss the current advances and potential of Wnt inhibitory therapeutics toward promoting tissue regeneration and reducing fibrosis.—Bastakoty, D., Young, P. P. Wnt/β-catenin pathway in tissue injury: roles in pathology and therapeutic opportunities for regeneration.

The Wnt signaling pathway is believed to have evolved with the first multicellular organisms (metazoa) and is considered to be an important mediator of the advent of multicellularity. The pathway is essential for embryonic and postnatal development in vertebrates and insects (1). In invertebrates and the few vertebrate species that display regeneration, such as urodele amphibians, the pathway remains active in the adult organism, playing a role in stem cell maintenance and self-renewal for normal tissue homeostasis and for regeneration after injury (2, 3). In mammals and other nonregenerating vertebrates, the role of the pathway in the adult organism becomes more complex and context dependent. In organs, such as colon, that undergo constant turnover and have a clearly defined, active pool of tissue-resident stem cells, the pathway remains active and continues to regulate stem cell homeostasis (4); but in nonregenerating organs, such as heart in higher vertebrates, the pathway is quiescent and is often activated in response to injury (5, 6). The role this pathway plays is incompletely understood and controversial, but a growing body of data suggests that rather than simply reviving the pool of quiescent stem cells and promoting regeneration in these injured organs, the pathway is also involved in fibrotic processes as part of the imperfect healing in these organs (7–11).

Because of the complexity of the Wnt pathway, deciphering the exact role of the reactivated Wnt pathway in response to injury has been difficult. The efforts at understanding the effects of Wnt pathway modulation in injury are further complicated by the limitations of classic genetic approaches that rely on Wnt modulation through specific promoters and, therefore, do not always target the complete wound milieu (12, 13) or carry unintended effects on the cells used for targeting (14, 15). Recently, however, with increasing focus on developing Wnt modulatory therapeutics aimed at Wnt-driven cancers, new studies that investigate the effects of modulating Wnt signals in regeneration are emerging (16–18). These studies are helping us to better understand the role of the pathway in regeneration and, more importantly, are pioneering the development of Wnt-modulatory therapeutics. In this review we discuss the current understanding of the complicated role of the Wnt/β-catenin pathway in mammalian organ injury, focusing on the lung, kidney, skin, and heart. We also outline the development of relevant Wnt modulators and deliberate on the challenges and promises of pharmacologic Wnt inhibition to promote regeneration.

Wnt/β-CATENIN SIGNALING PATHWAY IN DEVELOPMENT: A HISTORICAL PERSPECTIVE

The first Wnt gene, Wnt1, was identified in 1982 as a mammalian oncogene Int1 (19), a few years after identification of wingless (20, 21) as necessary for Drosophila wing development. With the finding that the 2 genes are homologous, the field of Wnt signaling was born (22). In the 3.5 decades since, the conserved roles of the Wnt family of genes in development and disease continue to be uncovered. The Wnt pathway, as it is currently described, consists of a family of 19 secreted glycoproteins (in mammals) (23), Frizzled family of transmembrane receptors (24, 25). Lipoprotein receptor-related protein (LRP) family of coreceptors, and several other downstream components and effector proteins that fall under either the canonical or 1 of the 2 major noncanonical arms of the Wnt signaling pathway. β-Catenin, which was originally identified as Armadillo (26, 27) in Drosophila, is an important effector protein that defines the canonical or β-catenin-dependent arm of the signaling pathway. This arm is involved in segment polarity, stem cell homeostasis, oncogenesis, and tissue repair, which will be the focus of this review. The noncanonical arms that also signal through Frizzled receptors, but in combination with a distinct set of coreceptors [such as the tyrosine kinase receptors ROR and RYK (28)], control cell polarity, and calcium signaling (28). These noncanonical arms of the Wnt signaling pathway are also critical in development and disease (29), but are beyond the scope of this review.

The canonical Wnt pathway (Fig. 1) is activated when a subset of the Wnt ligands bind their cognate Frizzled receptors and LRP5/6 coreceptors. This process causes phosphorylation-dependent sequestration of axis inhibitor (AXIN), a rate-limiting component of the cytoplasmic β-catenin degradation complex (30). The β-catenin degradation complex includes the kinases glycogen synthase kinase (GSK)-3β and casein kinase (CK)-1α that phosphorylate and target β-catenin for ubiquitin-mediated degradation (31–33). It is hypothesized that with AXIN2 sequestration, GSK3β, and possibly other members of the β-catenin degradation complex are recruited to the membrane to form a “signalosome,” thereby preventing β-catenin degradation (34). The stabilized β-catenin translocates to the nucleus by a yet-unknown mechanism (22). In the nucleus, β-catenin replaces Groucho/transducin-like enhancer protein in binding to T-cell factor/lymphoid enhancing factor (TCF/LEF) transcription factors (35) and in association with other transcriptional coactivators (36, 37), promotes transcription of its target genes. In addition to this classic cascade of signals, new studies describe other signaling arms of the canonical Wnt pathway. For example, during oxidative stress, β-catenin binds directly to, and promotes transcriptional activity of, forkhead box protein (FOXO) family of transcription factors (38). Further, Inoki et al. (39) showed in 2006 that the Wnt pathway can also activate translation through mammalian target of rapamycin (mTOR) signaling, independent of β-catenin-mediated transcription through interaction with GSK. In the Wnt-producing cell, important players are involved in the posttranslational modification of Wnt, which is considered necessary for the ligand secretion and activity. Works from various groups have indicated that most Wnt ligands are N-glycosylated. This posttranslational modification is not essential for Wnt secretion, but predominantly enhances palmitoylation—a lipid modification that is indispensable for Wnt ligand secretion and Frizzled receptor binding (40–42). The membrane-bound acyltransferase, Porcupine is essential for this modification in the endoplasmic reticulum and is being widely explored as a target for anti-Wnt therapeutics (18). The transport of the lipidated Wnt proteins into secretory vesicles for secretion of the ligand into the extracellular space is carried out by the multipass transmembrane protein Wntless/Evi (GPR177 in mouse/human) (43, 44).

Figure 1.

Simplified model of the canonical Wnt pathway and inhibitors. The Wnt–β-catenin pathway consists of secreted glycoproteins (WNT ligands) and Frizzled family of transmembrane receptors or LRP5/6 transmembrane receptors. In the endoplasmic reticulum (ER) of WNT producing cells, the membrane bound O-acyltransferase, Porcupine, acylates the WNT protein. This lipid modification is necessary for secretion of the WNT ligand. The secreted WNT ligand can be sequestered by secreted Frizzled-related protein (sFRP) family of secreted proteins to prevent Wnt ligand binding to receptors. DKK1, another secreted Wnt inhibitor prevents binding of WNT ligand to LRP5/6 receptors. In the absence of extracellular WNT glycoproteins, a destruction complex—including the proteins APC, GSK3β, CK1α, and AXIN—phosphorylates β-catenin, targeting it for ubiquitylation and proteasomal degradation. When WNTs bind to the Frizzled and LRP5/6 coreceptors, Disheveled helps phosphorylate and sequester AXIN to the LRP5/6 receptor, causing disassociation of the β-catenin degradation complex. The stabilized β-catenin enters the nucleus, binds to the TCF family of transcription factors and activates transcription of target genes. Distinct coactivators of transcription such as CBP, the E1A-associated protein p300, Pygopus (PYGO), BCL-9, and Brahma-related gene 1 (BRG1) are involved in the transcription of specific target genes. Independent of TCF, β-catenin can also associate with the FOXO family of transcription factors to activate transcription of genes primarily involved in aging. In the cytoplasm, inhibition of GSK3β activity by WNT ligand binding can simultaneously activate mTOR complex 1 signaling, which results in mRNA translation into proteins. Inhibitors of the pathway and their respective targets are shown in red font on yellow tabs and are further described in Table 1.

A gradient in the expression of Wnt genes controls axial patterning during embryonic development of organisms ranging from mammals to frogs, worms, birds, and echinoderms (45). Perturbation of the pathway or alteration of this gradient results in dramatic phenotypes caused by anteroposterior patterning defects (45). In later stages of embryonic development, Wnt modulators (both positive and negative) continue to be expressed and play crucial roles in development of organs. In heart development, for example, the Wnt pathway must be repressed during the early stage of heart formation, but its reactivation later is required in the development of the outflow tract and the valves (46). Similarly, an initial repression and a later reactivation of Wnt signaling in specific cells are necessary during neural crest formation (47). The strict temporal and spatial regulation of Wnt activity is an important attribute of the pathway in its various roles in development and homeostasis/regeneration/fibrosis, as will be discussed in the following sections.

Wnt IN REGENERATION AND FIBROSIS: THE GOOD, THE BAD, AND THE UGLY

Wnt pathway in invertebrates and regenerating vertebrates

The canonical Wnt signaling pathway plays diverse roles during regeneration of different organisms. In bilaterally polarized invertebrates such as planarians the pathway is reactivated in response to injury and is necessary for appropriate head vs. tail polarity during regeneration, consistent with its role in patterning during development (48). Wnt pathway enhancement does not prevent regeneration, but it promotes ectopic head formation in these organisms (49) and even in lower invertebrates, such as hydra (50). This effect of Wnt pathway modulation mirrors the observation, first made by McMahon and Moon (51), that int-1 RNA injection in Xenopus embryos results in dual axis formation.

Few vertebrate species such as zebrafish and urodele amphibians display regenerative abilities as adults. The Wnt pathway is implicated in part of the regenerative process in these organisms. In both zebrafish fin regeneration and urodele amphibian limb regeneration, the blastema, which is an organized mass of tissue with stem/progenitor cells, is thought to drive the mesenchymal regenerative process. The Wnt pathway is activated in a spatiotemporally regulated manner during blastema formation (52) and is considered to play a role in orchestrating tissue organization and differentiation or proliferation of the blastema cells (3, 53–55).

Wnt pathway in nonregenerating vertebrates: stem cell homeostasis vs. fibrosis

In nonregenerating vertebrates, including mammals, the Wnt/β-catenin pathway activity is maintained in the adult organism only in specific organs/tissues with high cell turnover, such as the hematopoietic compartment (56), the intestinal epithelium (4, 57), and the epidermis (58). Wnt pathway activity is associated with the stem cell self-renewal necessary for normal homeostasis in these organs. In the small intestine and colon, Lgr5, a common Wnt target gene, is considered a marker of a subset of stem cells (59). Knockdown of TCF4 (canonical Wnt) signal in the small intestine depletes the crypt stem cell compartment needed for self-renewal of cells in the intestinal epithelium (60). LGR5⁺ cells are also considered to be the self-renewing progenitor cells in gastric epithelium (61), mammary gland (62), and epidermis (63, 64). Accordingly, mutations in components of the Wnt pathway are implicated in epithelial cancers in these organs (65, 66), in which tumorigenesis is often driven by aberrantly proliferative cancer stem cells (67, 68). To be clear, there are distinct stem cell populations that are independent of Wnt signals: leucine-rich repeats and Ig-like domain-1 (LRIG)-1, a negative regulator of ErbB/epidermal growth factor receptor (EGFR) signaling (69), marks the more quiescent stem cells in both the colon (70) and the interfollicular epidermis (IFE) of the skin (71). LRIG1, through suppression of the proliferative EGFR signals, is known to maintain quiescence of slow-cycling stem cell populations, whereas the Wnt and Notch activation and bone morphogenic protein suppression are traditionally associated with homeostatic self-renewal of the fast-cycling leucine-rich repeat containing G protein-coupled receptor (LGR)5⁺ stem cells in the colon and skin (70, 71).

In other mammalian tissues, however, the Wnt pathway is quiescent in the adult and is reactivated in response to injury. These tissues possess either quiescent or differentiated “facultative” stem cells (72) that can dedifferentiate upon injury, as in the lungs (72), or that may lack stem cells capable of orchestrating complete regeneration (e.g., heart). The involvement of Wnt pathway in response to injury and its contributions to injury repair through regenerative (or more commonly fibrotic) processes are discussed in more detail in the following sections.

Lung

The lung is a complex organ composed of multiple specialized epithelium- and mesoderm-derived cells spatially arranged along an arborized architecture. The lung epithelium undergoes slow homeostatic turnover, replacing most of its cells from a progenitor pool [once every 4 months in rats (73)]. It is mostly quiescent, except when activated by injury. The differentiated epithelial cells of the lung are capable of dedifferentiation, proliferation, and transdifferentiation into diverse cell lineages. These epithelial cells are renewed from a pool of stem/progenitor cells distributed along multiple niches in the proximal–distal epithelium. Mediated by both the regenerative epithelial cells and the multiple progenitor cell subtypes, the process of regeneration in the lung is complex and spatially regulated (74–76). Likewise, the involvement of the Wnt pathway in the process is complicated and incompletely understood. The Wnt pathway is activated in the lung in response to injury. Chilosi et al. (77) reported increased nuclear localization of β-catenin and expression of the Wnt target genes cyclin D1 and matrilysin in bronchiolar lesions, damaged alveoli, and fibrotic foci of lung samples from patients with idiopathic pulmonary fibrosis (IPF). Similar findings of Wnt pathway activation in the alveoepithelial type II cells (AECII) in patients with IPF was reported by Koingshoff et al. (10). AECIIs are considered the facultative progenitors of the distal lung and are reported to replace the alveolar epithelium and lung parenchyma in bleomycin-induced and hyperoxic rodent lung injury models (78). Reports indicate a role of the Wnt pathway in the survival of AECs in bleomycin-induced lung injury models and increased AEC death upon selective depletion of β-catenin in them. In contrast, multiple reports have shown attenuation of bleomycin-induced pulmonary fibrosis by inhibition of the Wnt pathway through administration of small-molecule Wnt inhibitors (79, 80), although this attenuation may occur, in part, through effects on the mesenchymal rather than on the epithelial cells. The Wnt pathway was activated in the bronchioalveolar stem cells (BASCs) in response to naphthalene-based acute lung injury. Although Wnt activation induced proliferation of BASCs, inhibition of Wnt activity by GATA6 was deemed necessary for differentiation and proper regeneration of the damaged epithelium (81). In another report, Zemke and colleagues (82) showed by Cre-mediated expression of mutant β-catenin in airway epithelial cells that Wnt pathway activity is not needed for repair of bronchiolar epithelium after naphthalene-induced injury.

Despite the presence of multiple resident regenerative cell lineages, effective regeneration after injury does not always occur in the lung, as is evident with the increasing prevalence of fibrotic diseases of the lung (83). Repeated injuries and persistent inflammatory insults, and in some cases, genetic predispositions, compromise the regenerative potential of lung cells, leading to activation of fibrotic signals driving IPF, chronic obstructive pulmonary disease, and other ailments (72). The Wnt pathway is widely reported to be activated in response to injury and is shown to promote proliferation of at least a subset of regenerative cell lineages, and yet, the regeneration of the lung after injury appears insufficient to ward off fibrotic diseases. Moreover, the increasing reports citing Wnt pathway involvement in initiating profibrotic signals by mesenchymal cells (84) indicates a dual effect of Wnt activation resulting in regenerative epithelial signals, but also activation of profibrotic signaling which drives the pathology in chronic fibrotic pulmonary diseases.

Kidney

The kidney is capable of recovery from acute kidney injury (AKI) induced by ischemia or nephrotoxic drugs that commonly affect the renal tubule. Although the cells that contribute to the repair are still under investigation, a significant body of work suggests that proliferative and dedifferentiated tubular epithelial cells are the major contributors to repair (85–88). Despite this intrinsic healing potential, AKI results in significant morbidity and mortality, indicative of the inherent limitations in regeneration (88, 89).

The Wnt/β-catenin pathway is activated in response to AKI, as indicated by β-galactosidase expression in reporter mice expressing the protein in response to Wnt/β-catenin transcriptional activation or Wnt target gene Axin2 expression (90). In an ischemia-reperfusion–induced kidney injury model, Duffield and colleagues (90) showed that kidney macrophages secrete WNT ligands after injury and that Frizzled receptor (FZD)-4-expressing tubular epithelial cells are WNT responsive. The disruption of Wnt/β-catenin signals in these cells by mutation of FZD4 receptor or LRP5/LRP6 coreceptors reduced regeneration of tubular epithelium by increasing apoptosis of tubular epithelial cells, although the effect was modest even with depletion of both FZD4 and LRP5/6 activity. Another group reported aggravation of AKI through an increase in epithelial cell apoptosis upon tubule-specific depletion of β-catenin (91). However, an important caveat to these results is that β-catenin depletion in epithelial cells can alter β-catenin/E-cadherin signals, affecting the survival and regenerative potential of these cells independent of Wnt pathway activity.

CKD is characterized by progressive interstitial fibrosis, often leading to kidney failure. A recent study showed that sustained Wnt/β-catenin pathway activation after AKI is associated with progression to CKD, even though transient Wnt activity may aid epithelial repair. The authors also showed that partial blockade of Wnt/β-catenin pathway activity by pharmacologic inhibition of β-catenin binding to transcription cofactor cAMP response element binding protein could block progression to CKD (92). The effects of Wnt pathway on specific cells of the kidney have been implicated in CKD pathogenesis in multiple studies. For example, the pericytes, which are described as cells that differentiate into fibrogenic myofibroblasts (89, 93), express the myofibroblast marker α-smooth muscle actin (αSMA) in response to Wnt signaling activation (94). Forced upregulation of Wnt/β-catenin signals by stabilizing β-catenin in the interstitial cells is sufficient to induce increased myofibroblast differentiation, indicating a critical role for the Wnt pathway in interstitial fibrosis in the kidney (94). In UUO, which is the most commonly used murine model of kidney fibrosis, WNT ligand expression and β-catenin nuclear localization increased in both the kidney tubules and fibrous interstitial regions. He et al. (95) reported that blockade of Wnt/β-catenin signaling by intravenous injection of a Dickkopf1 (Dkk1)-encoding plasmid vector reduced renal acta2 mRNA expression (encoding αSMA) and interstitial fibrosis. Wnt/β-catenin pathway activation has also been reported to contribute to CKD progression by promoting podocyte dysfunction (96). In another study, both hyperactivation and ectopic down-regulation of Wnt/β-catenin signals led to podocyte dysfunction and compromised glomerular filtration (97). Taken together, these findings suggest that, in the kidney as in the lung, transient Wnt/β-catenin pathway activation may induce regeneration by putative “facultative progenitors”; however, upon prolonged injury, a hyperactivation of the Wnt pathway appears to initiate signals in multiple cell types, leading to progressive fibrosis.

Skin

As a tissue that undergoes significant turnover and harbors multiple distinct stem and progenitor cells and yet routinely heals with a scar, the skin displays a unique dichotomy of regeneration and fibrosis upon injury. The involvement of the Wnt pathway in skin injury response is similarly dichotomous, with Wnt activity playing a critical role in epidermal stem cell maintenance, hair follicle development, and regeneration (98, 99), but also in dermal fibroblast-mediated scarring (100, 101).

The stem/stem-like cells and their respective proliferative progenitors (often referred to as transient amplifying cells; TACs) in the skin are mostly distributed in niches in the basal layer of the IFE and the bulge of the hair follicles (102, 103). The IFE consists of partially differentiated progenitors that undergo asymmetric division, giving rise to TACs and ultimately to basal epidermal cells that contribute to epidermal stratification during development and homeostatic turnover (104, 105). The bulge stem cells, originally identified as slow-cycling label retaining cells based on their longer retention of bromodeoxyuridine BrdU label, are considered multipotent stem cells that can differentiate to form cells of the hair follicle or can give rise to the progenitors of the IFE upon injury (98, 103, 106). The Wnt pathway plays an integral role in the homeostasis and multipotency of both of these stem cell subtypes in the skin. Aside from being necessary for hair follicle development during embryogenesis (107, 108), Wnt-Lef1 transcription activation in bulge stem cells is specifically associated with initiation of the hair cycle in the adult skin (108). In partially committed progenitor cells, such as the IFE stem cells (13), or differentiated cells of the dermal papillae (109), Wnt pathway activation is reported to cause these cells to revert to a multipotent stem-like phenotype, wherein they form de novo hair follicles. Consistent with this observation, focal Wnt activation in IFE stem cells is reported to drive self-renewal and proliferation to promote wound closure after injury (110) and de novo hair follicle formation from IFE stem cells, most likely by enhancing reversal to multipotency (15). Hence, there is clear evidence that Wnt/β-catenin/TCF/LEF signals are important for stem cell maintenance, hair follicle formation, and turnover of epithelium during homeostasis and in response to injury. Although the role of the Wnt pathway in promoting wound resurfacing by the epidermis and its role in hair follicle regeneration has been widely studied, none of the studies has assessed the effect on dermis, which is the region that undergoes scarring. Indeed, in human wounds and in most rodent models, the limiting factor in regenerative wound repair is not reepithelialization, but scar resolution in the dermis. Because the initiating signal for hair follicle morphogenesis during development originates in the dermis [first dermal message (111)] and has not been completely decoded, the lack of hair follicle regeneration in most wound scars may be explained by the nonregenerative healing in the dermis. Dermal papillae cells are known to induce hair follicle differentiation of epithelial cells during development and in hair reconstitution assays, which indicates a need for dermal regeneration of new hair follicles to occur after repair of a full-thickness skin wound.

Studies of skin wound healing that examined scar formation in the dermis demonstrated that Wnt pathway activation plays a profibrotic role in this compartment (100, 101, 112, 113). Alman and colleagues (101) have shown that β-catenin transcriptional activity, which is upregulated in mouse dermis in response to injury, causes hyperproliferation of fibroblasts and increases wound size and fibromatosis in a rodent excisional wound model (101). Conversely, β-catenin destabilization by adenoviral Cre-mediated deletion of β-catenin exon 3 reduces wound size and reverses the fibromatosis phenotype (100). Similarly, prolonged activation of Wnt/β-catenin signaling has been observed in human hyperplastic wounds (114). Skin biopsy specimens from patients with systemic sclerosis also reveal altered expressions of Wnt pathway components, resulting in activation of Wnt signaling (115, 116).

Hence, the notion that the Wnt pathway promotes regeneration in the skin may apply to hair follicle regeneration in conditions such as alopecia or epidermal abrasion. However, for the regeneration of a full-thickness dermal wound, or burn wounds, the role of Wnt signaling should be reexamined with a focus on dermal scarring. Although Wnt activity is clearly needed for reepithelialization, Wnt inhibition in the dermis may be important in reducing scarring. For example, in our study, a small-molecule inhibitor of the Wnt pathway that caused a more appreciable reduction in β-catenin/TCF-driven transcription specifically in the dermis, promotes regenerative wound repair with less scarring (16). This finding aligns with the report by Alman’s group (100) that adenoviral Cre mediates conditional depletion of β-catenin, which inhibits the Wnt/β-catenin signaling throughout the wound and is not limited by use of an epithelial or hair follicle–specific promoter, successfully drives reduction in fibrosis and wound size. Although the same authors have reported that complete ablation of the Wnt pathway in the dermal macrophages prevents wound closure by impeding granulation tissue formation (113), the study lends credence to the idea that fine tuning of Wnt inhibition in the dermis may be key to promoting scarless wound healing in the skin. Development of effective pharmacological Wnt inhibitors and dosing strategies that result in an appropriately calibrated Wnt signaling reduction specifically targeted to the dermis may lead to new and effective wound-healing therapeutics.

Heart

The heart has historically been considered a nonregenerative organ devoid of a functional stem cell population with no replacement of dying cardiomyocytes through mitosis. However, in the past few decades, both of these views have been challenged by evidence of cardiomyocyte replenishment during the adult lifetime of humans (117, 118) and by reports of multiple cardiac-resident stem/progenitor cells in the adult heart (119–123) with the potential to regenerate cardiomyocytes after injury. Some of these studies have led to clinical trials in which autologous cardiac stem cells were infused into the coronary artery of patients who have had myocardial infarction (MI) after ex vivo culture and expansion (124). Positive outcomes in these studies have been modest at best, mirroring the initial enthusiasm followed by disinterest that cell therapy with bone-marrow–derived (CD34⁺) mononuclear cells garnered a few years prior (125). Given the limited success of these expensive therapeutic approaches, the interest in understanding signals that can augment regenerative healing, with or without the participation of endogenous cardiac stem cells, has persisted. As with its involvement in injury response in other tissues, the Wnt/β-catenin pathway has been reported by several research groups to be involved in both proreparative and profibrotic response to cardiac ischemic injury (9, 14, 126).

The Wnt/β-catenin pathway is activated in multiple cells of the heart, starting 72 h after injury (5, 127). Consistent with its role in specification of cardiac progenitors during heart development (128), reactivation of the Wnt pathway after injury is reported to affect cardiac progenitors in multiple ways. For example, Bergmann and colleagues (9) have shown that conditional depletion of β-catenin in (αMHC-expressing cells of the heart after infarction promotes recovery by increasing cardiomyogenic differentiation of αMHC⁺SCA1⁺ (cells partially committed to cardiomyocyte lineage) cardiac progenitors. Stabilization of β-catenin using the same promoter reduces cardiomyogenic differentiation (indicated by GATA4 coexpression) of the partially committed progenitors. In a more recent study, the group has used knockdown of the Kruppel-like-15 transcription factor, a negative regulator of β-catenin/TCF transcriptional activity, to show that increased Wnt/β-catenin signaling induces SCA1⁺ progenitors toward an endothelial, as opposed to a cardiomyogenic, lineage commitment in the adult mouse heart (129). Exogenous administration of secreted Frizzled-related protein (sFRP)-2, a secreted Wnt inhibitor, also promotes cardiomyogenic differentiation of SCA1⁺ progenitors in the infarcted heart through inhibition of canonical Wnt pathway and concurrent activation of noncanonical Wnt signals (130). In a study focusing on cardiac side population cells that were identified based on their ability to efflux Hoechst 33342 dye (131), Wnt pathway activation by injecting recombinant WNT3a in the peri-infarct region depleted the endogenous pool of cardiac progenitors and worsened cardiac remodeling after infarction (126). In vitro, WNT3a exerted an antiproliferative effect on these side-population progenitors through insulin growth factor binding protein-3 signaling (126).

Independent of stem/progenitor cells, the Wnt pathway is reported to affect infarct repair through many other cells in the heart. Wnt inhibition by sFRP2, for example, is credited with promoting survival of cardiomyocytes (7, 132). Conversely, injection of recombinant WNT3A in the infarct zone increases cardiomyocyte apoptosis (126). Both of these studies point to a direct effect of the Wnt pathway on cardiomyocyte apoptosis, which is an important aspect of infarction pathology. However, other studies have demonstrated an opposite prosurvival effect of the Wnt pathway in cardiomyocytes by adenovirus-mediated gene transfer of nonphosphorylatable, constitutively active β-catenin in the infarct zone (133).

The Wnt pathway and its modulation also affect cardiac fibroblasts and their activity after ischemic injury. Overexpression or exogenous addition of sFRP-1 and -2 in the infarcted heart are reported to promote healing by modulating matrix synthesis and remodeling, at least in part, through Wnt pathway inhibition; however, Wnt-independent roles of the protein are also considered essential for this effect (134, 135). Other reports suggest that a profibrotic effect of Wnt pathway activation promotes adaptive cardiac remodeling after infarction. Hyo-Soo Kim and colleagues (133) used adenoviral gene transfer of stabilized β-catenin to activate the Wnt pathway after infarction. They reported enhanced fibroblast proliferation and differentiation of fibroblasts into myofibroblasts, which unexpectedly resulted in a slight improvement in cardiac function and a reduction in infarct size and adverse remodeling. More recently, Duan et al. (14) showed that Wnt/β-catenin signaling is activated in the epicardium after infarction and leads to differentiation of epicardial cells into fibroblasts through the process of epithelial–mesenchymal transition (EMT). Disruption of this signal in either the epicardium or in fibroblasts using inducible genetic models increases cardiac dysfunction and ventricular dilatation after ischemia–reperfusion injury. Other studies have shown that Wnt inhibition by sFRP2 overexpression increases the potency of exogenous bone-marrow–derived mesenchymal stem cells in promoting cardiac repair after infarction (136, 137). Wnt signals also affect angiogenesis during infarct repair. Laeremans et al. (138) showed that blocking Wnt/Frizzled signaling with a peptide mimic of WNT3A/5A improves neovascularization, whereas another study reported a positive role for the Wnt pathway in mediating neovascularization in the infarcted heart through VEGF and ANG1 signals (139). In addition, interaction of the Wnt pathway components with other signaling networks further complicate healing outcomes. CK-Iα, which negatively regulates Wnt/β-catenin signals by phosphorylating β-catenin, also potentiates the Hedgehog signaling pathway, which is necessary for maintenance of coronary vasculature during homeostasis and injury repair (140). Similarly, interaction with the noncanonical arms of the Wnt pathway also affects healing. For example, Mirotsou and coworkers (130) showed that sFRP2 promotes cardiac repair through inhibition of Wnt/β-catenin signals and concurrent potentiation of both the Ca/Cam kinase and JNK branches of noncanonical Wnt signaling pathways. The role of the Wnt pathway in infarction pathology and repair is complex and multifaceted, and this complexity complicates efforts at understanding effects of Wnt modulation using genetic models that are limited by the promoters or cells types used for targeted mutations.

Other tissues

Other tissues illustrate the complicated, and more commonly, negative role of the Wnt pathway in injury and repair. Skeletal muscle, a partially regenerating tissue, uses satellite cells with the ability to facilitate regeneration of injured muscle. An initial inhibition of the Wnt pathway followed by a brief activation is necessary for complete regeneration of skeletal muscle (141). In keeping with the theme of Wnt as a profibrotic factor, a sustained increase in activity of the canonical Wnt pathway in myogenic progenitors is associated with increased conversion of satellite cells from myogenic to fibrogenic lineages in aging mice (142). In the liver, another tissue with facultative progenitors that successfully regenerate the tissue after injury, fibrosis can occur as a result of chronic injury from drug or alcohol toxicity and persistent viral infection. Wnt pathway activity is implicated in promoting fibrosis through myofibroblast differentiation of liver progenitor cells, known as hepatic stellate cells (143, 144). In contrast, Wnt is plays the role of mediator of homeostasis and regeneration through proliferation of both the hepatocytes (145, 146) and the more quiescent liver progenitors (147). Hence, in many mammalian tissues, regardless of their intrinsic regenerative and self-renewal potential, or lack thereof, the recurring pattern appears to be a positive, albeit complex and context- or cell-subtype–dependent, role for the Wnt pathway in stem cell homeostasis, proliferation, and differentiation and a concurrent profibrotic effect from longer term Wnt activation (112). In the face of this dichotomy, pharmacologic agents that allow targeting of the overall Wnt pathway (as opposed to specific ligands or receptors targeted by genetic models) and fine tuning of the level of Wnt inhibition in a spatiotemporally restricted manner may be crucial to achieving a therapeutically relevant wound-healing outcome.

Wnt THERAPEUTICS: SMALL MOLECULES FOR BIG GAINS IN REGENERATION

The involvement of the Wnt pathway in fibrotic diseases and, even more commonly, in cancers (65–68), has engendered widespread interest in developing Wnt inhibitory therapeutics. However, progress has been slow because of the complexity of the pathway and the wide-ranging effects it has on homeostasis of multiple tissues. Several U.S. Food and Drug Administration (FDA)–approved drugs that inhibit the Wnt pathway are in clinical use, although Wnt inhibition may not be recognized as their primary mode of action. Nonsteroidal anti-inflammatory drugs and a cyclooxygenase-2 inhibitor (celecoxib) inhibited Wnt/β-catenin–mediated transcription (148) and reduced polyp formation in patients and in mouse models of colorectal cancer (149, 150). More recently, high-throughput screening with Wnt-activated luciferase reporters has led to identification of novel compounds or repurposing opportunities of known drugs as Wnt inhibitors (Fig. 1 and Table 1).

TABLE 1.

Wnt pathway inhibitors and progress in the clinical and preclinical pipeline

Drug class	Compound	Molecular target	Drug developmental stage	Reference/company
Small molecule	Pyrvinium	Activates CKIα	FDA-approved anti-helminth	151
	Niclosamide	Inhibits Frizzled endocytosis	FDA-approved anti-helminth	152
	XAV-939	Inhibits tankyrase/stabilizes AXIN	Preclinical	153
	IWR-1	Stabilizes AXIN	Preclinical	18
	iCRT	Prevents β-catenin binding to TCF	Preclinical	156
	ICG-001	Prevents β-catenin binding to CBP	Preclinical	157
	PRI-724	Prevents β-catenin binding to CBP	Phase I for advanced solid tumors Clinicaltrials.gov: NCT01302405	Prism Pharma, New Delhi, India
	IWP	Inhibits Porcupine activity	Preclinical	18
	LGK-974	Inhibits Porcupine activity	Phase I for Wnt-driven malignancies	Novartis, Basel, Switzerland
			Clinicaltrials.gov: NCT01351103
Biologic	OTSA101 mAb	Frizzled-10	Phase I for synovial sarcoma	OncoTherapy Science, Tokyo, Japan
			Clinicaltrials.gov: NCT01469975
	Vantictumab	Frizzled	Phase I for solid tumors	OncoMed Pharmaceuticals, Redwood City, CA, USA
			Clinicaltrials.gov: NCT01345201, NCT01957007, NCT01973309, NCT02005315
	OMP-54F28	WNT scavenger	Phase I for solid tumors	OncoMed Pharmaceuticals
			Clinicaltrials.gov: NCT01608867, NCT02050178, NCT02092363, NCT02069145
	UM206 (WNT3A/5A peptide mimetic)	Blocks Frizzled	Preclinical	138, 172

Pyrvinium, an FDA-approved antihelminth was identified as a potent Wnt inhibitor in a Wnt/TCF-responsive luciferase screen (151). The compound potentiates CK1α, enabling degradation of β-catenin (151), and it has been shown to improve infarct repair (17) and promote regenerative healing of cutaneous injury (16). However, in cardiac studies, intracardiac injection of the drug resulted in significant mortality among the animals (17), curbing the potential of this drug as a therapeutic agent, unless it can be reformulated to counter the systemic toxicity. Another antihelminth, niclosamide, was identified by a similar approach as an inhibitor of the Wnt pathway. Niclosamide acts by preventing Frizzled endocytosis upon ligand binding (152). Both pyrvinium and niclosamide, although FDA-approved, have the limitation of being developed originally as orally administered compounds targeting the gastrointestinal tract and hence present a need for thorough examination of their systemic bioavailability and safety.

Another small-molecule Wnt inhibitor, XAV-939, which targets the β-catenin degradation complex by inhibiting tankyrases and thereby stabilizing AXIN2, a rate-limiting component of the β-catenin degradation complex (153). XAV-939 has been reported to suppress growth of Wnt-driven cancer cells derived from colorectal cancer (154) and neuroblastoma (155). In our study of a cutaneous injury model, topical treatment of skin injury with XAV-939 showed marked improvement in regenerative wound repair with reduction in fibrosis and regeneration of auricular cartilage (16).

Because components of the β-catenin degradation complex, particularly adenomatous polyposis coli (APCs), are mutated in several Wnt-driven cancers, some investigators have focused on inhibiting β-catenin-responsive transcription (CRT) directly in the nucleus by targeting β-catenin binding to TCF or other transcription cofactors in the nucleus (156). Besides circumventing the limitations of dysregulation of other Wnt pathway components, this approach counters the effects of altering β-catenin protein levels and hence interfering with E-cadherin/β-catenin signals. ICG-001, which inhibits β-catenin/TCF mediated transcriptional activity by binding to the transcription cofactor CBP, was identified by Emami et al. (157). Although the drug successfully reduced proliferation of colorectal carcinoma cells, improved heart function in a rat myocardial infarction model (158), reversed pulmonary fibrosis (80), attenuated renal fibrosis (159), and inhibited Wnt-driven cholangiocarcinoma (160), concerns about the off-target effect of preventing CBP from binding to its many other transcription factor partners persist among researchers (156, 161). A second generation CBP/β-catenin inhibitor PRI-724 with improved plasma half-life entered a phase I dose-escalation study for advanced solid tumors in 2011 and raised no safety concerns, but the study was terminated for unclear reasons (clinicaltrials.gov: NCT01302405). Meanwhile, other CRT inhibitors targeting β-catenin binding to its transcription cofactors have also been investigated. Shivdasani and colleagues (162) developed a cell-free β-catenin/TCF binding-based high-throughput screen and identified small molecules that inhibit β-catenin-TCF binding. Unfortunately, because the screen was performed in a cell-free system, the small molecules showed off-target inhibition of β-catenin-APC binding when tested in cell lines. Similarly, Gonsalves et al. (156) identified specific inhibitors of β-catenin–driven transcription with an RNAi-based chemical genetic screen of small molecules. The lead compounds identified in the screen inhibited Wnt/β-catenin reporter activity and reduced proliferation in myeloma cells (163).

Although these CRT inhibitors have moved toward the clinic in a slow, halting progression, the recent identification of the acyltransferase Porcupine, which is needed for WNT ligand secretion, has garnered much enthusiasm as a druggable Wnt pathway target. Consequently, new small-molecule Porcupine inhibitors have been described by multiple groups (18, 164–166). One of the Porcupine inhibitors identified in those screens, LGK974, is currently in phase I dose-escalation studies for malignancies driven by WNT ligands (clinicaltrials.gov: NCT01351103). We reported a significant improvement in cardiac recovery from infarction by temporary systemic treatment with an analog of LGK974 (167). Intravenous administration of a Porcupine inhibitor GNF-6231 for 6 d after experimental infarction in mice results in reduced adverse remodeling, improved function, and decreased infarct size. Therapeutic Wnt inhibition in the postinfarction period affects multiple aspects of infarct repair, including proliferation of nonmyocyte cells, thus inhibiting myofibroblast proliferation and matrix synthesizing activity and improving myocyte survival (167).

Although small molecules constitute the bulk of the drugs being developed to target the Wnt pathway, biologics such as blocking antibodies and WNT peptide mimetics have also been developed, and some are being tested in clinical trials (168). Antibodies against WNT5a (169) and WNT1 (170) have been described to be effective against Wnt-driven gastric and colorectal cancer in rodent models. However, none of the anti-Wnt antibodies has reached clinical studies yet. On the other hand, antibodies against Frizzled receptors are making considerable progress in the clinical and preclinical pipelines. OTSA101, a chimeric humanized monoclonal antibody against Frizzled-10, developed by OncoTherapy Science (Tokyo, Japan) has reached phase I clinical trial for synovial sarcoma, in which Frizzled-10 is uniquely expressed (168) (clinicaltrials.gov: NCT01469975). Another anti-Frizzled monoclonal antibody, vantictumab, developed by OncoMed Pharmaceuticals (Redwood City, CA, USA), has completed phase I trials for solid tumors (clinicaltrials.gov: NCT01345201) and is in ongoing phase I studies for lung cancer (clinicaltrials.gov: NCT01957007), metastatic breast cancer (clinicaltrials.gov: NCT01973309), and stage IV pancreatic cancer (clinicaltrials.gov: NCT02005315), as a combined therapy with cytotoxic chemotherapeutic agents. A different biologic by OncoMed Pharma, OMP-54F28, a Wnt scavenger formed by fusing the WNT binding region of Frizzled-8 to the Fc region of IgG (171), is also in phase I dose-escalation studies for several different solid tumors, either as a solo agent or in combination with cytotoxic chemotherapeutics (clinicaltrials.gov: NCT01608867, NCT02050178, NCT02092363, and NCT02069145).

WNT mimetic peptides constitute another class of biologics that are being developed for targeting the Wnt pathway, and they have made considerable progress in clinical or preclinical studies (168). UM206, a WNT3A/5A peptide mimic that blocks Wnt/Frizzled signaling, prevented infarct expansion and development of heart failure in both mouse (138) and swine (172) models.

These interventions that target the Wnt pathway in the clinic have great potential for promoting tissue regeneration, although many of them were originally developed as cancer therapeutics. In addition to the development of pharmacologic agents, design of new tools or repurposing of existing ones for targeting therapeutics to the tissue of interest and fine tuning their delivery are paramount for their success. Because the Wnt pathway activity is critical in homeostasis of certain tissues and the role of Wnt activation or inhibition is often biphasic and context or cell-type dependent in most tissues, the optimization of delivery, dosage, and targets is critical for the safety of patients. The need for carefully assessing safety and off-target effects of Wnt antagonists has been exemplified in the case of the Frizzled-10 mAb vantictumab, and the Wnt scavenger OMP-54F28 (173). The phase I clinical trials for these drugs were put on temporary hold by the FDA because of mild-to-moderate bone-related side effects on patients. Although the studies have resumed after revision of study protocols to mitigate side effects (174), this experience reiterates the need to optimize delivery and dosage so as to minimize the systemic exposure to the drugs and to ward off collateral effects on tissues or cells that require Wnt activity for homeostasis. In cutaneous injury, for example, controlled-release scaffolds placed in the dermis may allow targeting of the drug specifically to the dermis, where fibrogenesis occurs, without undesired effects on the epidermis or hair follicles. Similarly, in the heart, use of implantable stents to release the drug directly into the infarct region, and hence reduce the dose and systemic exposure to the drug, may aid in achieving optimal Wnt inhibition while minimizing the off-target effects associated with systemic delivery.

CONCLUSIONS

The Wnt pathway, with its various roles in regeneration/stem-cell maintenance and fibrosis, presents a conundrum for therapeutic targeting. After more than 3 decades of intense investigation, understanding the complete role of the Wnt pathway in homeostasis and disease is still a work in progress, and no Wnt therapeutics have succeeded in reaching patients (22, 175). This failure can mostly be attributed to the increasing complexity of the pathway, and its wide-ranging, disparate roles in various tissues. Based on the lessons learned from the genetic studies, wherein the effects seem to vary depending on the timing, target cell type, and specific part of the Wnt signaling cascade that is targeted (12, 139), innovative interdisciplinary approaches, such as systems biology methods and computational modeling may provide additional insights into predicting outcomes. In addition, future studies that elucidate coactivators of β-catenin transcription involved in activating transcription of specific Wnt/β-catenin target genes in a given context may allow development of targeted inhibitors with potentially fewer side effects (176, 177).

Despite the challenges in safely and effectively targeting the Wnt/β-catenin pathway for regeneration, given emerging evidence pointing to wide-ranging effects on repair of multiple tissues and pathologies, concrete progress in the area has the potential to result in big gains in regenerative medicine. Our constantly improving understanding of the intricacies of the Wnt pathway and availability of more sophisticated tools to target it in the tissue of interest present opportunities to bring to the bedside Wnt inhibitory therapeutics that can significantly improve regenerative healing.

Glossary

αMHC

α-myosin heavy chain

αSMA

αsmooth muscle actin

AEC II

alveo-epithelial cell type II

AKI

acute kidney injury

APC

adenomatous polyposis coli

AXIN

axis inhibition BASC, bronchioalveolar stem cell

BCL

B-cell lymphoma

CBP

cAMP response element–binding protein

CK

casein kinase

CKD

chronic kidney disease

COX

cyclooxygenase

Dkk

Dickkopf

FDA

U.S. Food and Drug Administration

FOXO

forkhead box protein O

Fzd

Frizzled receptor

GSK

glycogen synthase kinase

IFE

interfollicular epidermis

IPF

idiopathic pulmonary fibrosis

LEF

lymphoid enhancing factor

LRIG

leucine-rich repeats and Ig-like domain

LRP

lipoprotein receptor-related protein

MI

myocardial infarction

mTOR

mammalian target of rapamycin

Pygo

Pygopus

Sca

stem cell antigen

sFRP

secreted Frizzled-related protein

TAC

transient amplifying cell, TCF, T-cell factor

UUO

unilateral ureteral obstruction